Citation
VP1-mediated repression of alpha-amylase genes in developing maize aleurone

Material Information

Title:
VP1-mediated repression of alpha-amylase genes in developing maize aleurone
Creator:
Hoecker, Ute, 1964-
Publication Date:
Language:
English
Physical Description:
v, 104 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Corn -- Genetic engineering ( lcsh )
Dissertations, Academic -- Plant Molecular and Cellular Biology -- UF
Plant Molecular and Cellular Biology thesis, Ph. D
Corn ( jstor )
Barley ( jstor )
Embryos ( jstor )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 91-103).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Ute Hoecker.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
021925211 ( ALEPH )
33628683 ( OCLC )

Downloads

This item has the following downloads:

AA00029819_00001 ( .pdf )

vp1mediatedrepre00hoec_0091.txt

AA00029819_00001_0017.txt

AA00029819_00001_0085.txt

vp1mediatedrepre00hoec_0042.txt

AA00029819_00001_0070.txt

vp1mediatedrepre00hoec_0077.txt

vp1mediatedrepre00hoec_0028.txt

vp1mediatedrepre00hoec_0049.txt

vp1mediatedrepre00hoec_0106.txt

vp1mediatedrepre00hoec_0015.txt

AA00029819_00001_0034.txt

vp1mediatedrepre00hoec_0007.txt

vp1mediatedrepre00hoec_0000.txt

AA00029819_00001_0064.txt

AA00029819_00001_0091.txt

vp1mediatedrepre00hoec_0005.txt

AA00029819_00001_0087.txt

vp1mediatedrepre00hoec_0092.txt

vp1mediatedrepre00hoec_0078.txt

vp1mediatedrepre00hoec_0080.txt

vp1mediatedrepre00hoec_0050.txt

AA00029819_00001_0005.txt

vp1mediatedrepre00hoec_0052.txt

AA00029819_00001_0060.txt

vp1mediatedrepre00hoec_0064.txt

vp1mediatedrepre00hoec_0093.txt

vp1mediatedrepre00hoec_0051.txt

AA00029819_00001_0055.txt

AA00029819_00001_0036.txt

vp1mediatedrepre00hoec_0063.txt

AA00029819_00001_0098.txt

AA00029819_00001_0025.txt

AA00029819_00001_0088.txt

AA00029819_00001_0107.txt

vp1mediatedrepre00hoec_0096.txt

AA00029819_00001_0003.txt

vp1mediatedrepre00hoec_0094.txt

vp1mediatedrepre00hoec_0016.txt

AA00029819_00001_0108.txt

AA00029819_00001_0016.txt

vp1mediatedrepre00hoec_0107.txt

AA00029819_00001_0101.txt

AA00029819_00001_0096.txt

AA00029819_00001_0007.txt

vp1mediatedrepre00hoec_0040.txt

AA00029819_00001_0052.txt

AA00029819_00001_0079.txt

AA00029819_00001_0082.txt

AA00029819_00001_0002.txt

vp1mediatedrepre00hoec_0087.txt

vp1mediatedrepre00hoec_0010.txt

vp1mediatedrepre00hoec_0026.txt

AA00029819_00001_0035.txt

AA00029819_00001_0029.txt

AA00029819_00001_0006.txt

AA00029819_00001_0032.txt

vp1mediatedrepre00hoec_0002.txt

vp1mediatedrepre00hoec_0088.txt

AA00029819_00001_0080.txt

AA00029819_00001_0001.txt

vp1mediatedrepre00hoec_0037.txt

vp1mediatedrepre00hoec_0079.txt

AA00029819_00001_0092.txt

AA00029819_00001_0020.txt

vp1mediatedrepre00hoec_0017.txt

AA00029819_00001_0019.txt

AA00029819_00001_0022.txt

vp1mediatedrepre00hoec_0081.txt

vp1mediatedrepre00hoec_0073.txt

AA00029819_00001_0044.txt

vp1mediatedrepre00hoec_0053.txt

AA00029819_00001_0103.txt

vp1mediatedrepre00hoec_0018.txt

vp1mediatedrepre00hoec_0030.txt

vp1mediatedrepre00hoec_0035.txt

AA00029819_00001_0059.txt

vp1mediatedrepre00hoec_0067.txt

vp1mediatedrepre00hoec_0036.txt

vp1mediatedrepre00hoec_0046.txt

AA00029819_00001_0065.txt

AA00029819_00001_0042.txt

AA00029819_00001_0051.txt

vp1mediatedrepre00hoec_0021.txt

vp1mediatedrepre00hoec_0101.txt

vp1mediatedrepre00hoec_0022.txt

vp1mediatedrepre00hoec_0074.txt

AA00029819_00001_0041.txt

vp1mediatedrepre00hoec_0027.txt

AA00029819_00001_0031.txt

vp1mediatedrepre00hoec_0004.txt

vp1mediatedrepre00hoec_0047.txt

vp1mediatedrepre00hoec_0109.txt

vp1mediatedrepre00hoec_0066.txt

AA00029819_00001_0090.txt

vp1mediatedrepre00hoec_0105.txt

vp1mediatedrepre00hoec_0006.txt

vp1mediatedrepre00hoec_0003.txt

vp1mediatedrepre00hoec_0086.txt

vp1mediatedrepre00hoec_0029.txt

AA00029819_00001_0009.txt

AA00029819_00001_0061.txt

AA00029819_00001_0056.txt

vp1mediatedrepre00hoec_0025.txt

AA00029819_00001_0057.txt

AA00029819_00001_0012.txt

AA00029819_00001_0058.txt

vp1mediatedrepre00hoec_0065.txt

AA00029819_00001_0073.txt

vp1mediatedrepre00hoec_0072.txt

AA00029819_00001_0018.txt

vp1mediatedrepre00hoec_0082.txt

vp1mediatedrepre00hoec_0043.txt

vp1mediatedrepre00hoec_0084.txt

AA00029819_00001_pdf.txt

AA00029819_00001_0106.txt

vp1mediatedrepre00hoec_0031.txt

AA00029819_00001_0081.txt

AA00029819_00001_0089.txt

AA00029819_00001_0076.txt

AA00029819_00001_0072.txt

EZO2P9C2K_WRGOCV_xml.txt

AA00029819_00001_0054.txt

AA00029819_00001_0047.txt

AA00029819_00001_0066.txt

vp1mediatedrepre00hoec_0097.txt

AA00029819_00001_0027.txt

AA00029819_00001_0083.txt

AA00029819_00001_0063.txt

vp1mediatedrepre00hoec_0099.txt

vp1mediatedrepre00hoec_0011.txt

AA00029819_00001_0104.txt

AA00029819_00001_0011.txt

AA00029819_00001_0074.txt

AA00029819_00001_0014.txt

vp1mediatedrepre00hoec_0032.txt

vp1mediatedrepre00hoec_0085.txt

vp1mediatedrepre00hoec_0089.txt

AA00029819_00001_0049.txt

vp1mediatedrepre00hoec_0095.txt

vp1mediatedrepre00hoec_0024.txt

AA00029819_00001_0084.txt

vp1mediatedrepre00hoec_0001.txt

AA00029819_00001_0037.txt

AA00029819_00001_0099.txt

AA00029819_00001_0040.txt

vp1mediatedrepre00hoec_0068.txt

AA00029819_00001_0043.txt

vp1mediatedrepre00hoec_0019.txt

AA00029819_00001_0095.txt

AA00029819_00001_0069.txt

vp1mediatedrepre00hoec_0061.txt

AA00029819_00001_0024.txt

vp1mediatedrepre00hoec_0020.txt

AA00029819_00001_0046.txt

vp1mediatedrepre00hoec_0045.txt

AA00029819_00001_0053.txt

AA00029819_00001_0078.txt

AA00029819_00001_0100.txt

vp1mediatedrepre00hoec_0033.txt

vp1mediatedrepre00hoec_0054.txt

vp1mediatedrepre00hoec_0014.txt

AA00029819_00001_0109.txt

vp1mediatedrepre00hoec_0057.txt

vp1mediatedrepre00hoec_0098.txt

AA00029819_00001_0026.txt

vp1mediatedrepre00hoec_0076.txt

AA00029819_00001_0023.txt

AA00029819_00001_0067.txt

vp1mediatedrepre00hoec_0071.txt

vp1mediatedrepre00hoec_0060.txt

AA00029819_00001_0075.txt

vp1mediatedrepre00hoec_0034.txt

vp1mediatedrepre00hoec_0009.txt

AA00029819_00001_0097.txt

AA00029819_00001_0010.txt

vp1mediatedrepre00hoec_0062.txt

AA00029819_00001_0013.txt

vp1mediatedrepre00hoec_0102.txt

vp1mediatedrepre00hoec_0048.txt

vp1mediatedrepre00hoec_0083.txt

vp1mediatedrepre00hoec_0013.txt

vp1mediatedrepre00hoec_0070.txt

AA00029819_00001_0045.txt

AA00029819_00001_0015.txt

AA00029819_00001_0071.txt

vp1mediatedrepre00hoec_0103.txt

vp1mediatedrepre00hoec_0012.txt

AA00029819_00001_0111.txt

vp1mediatedrepre00hoec_0039.txt

AA00029819_00001_0068.txt

AA00029819_00001_0093.txt

AA00029819_00001_0105.txt

vp1mediatedrepre00hoec_0038.txt

AA00029819_00001_0094.txt

AA00029819_00001_0102.txt

AA00029819_00001_0062.txt

AA00029819_00001_0033.txt

vp1mediatedrepre00hoec_0100.txt

vp1mediatedrepre00hoec_0056.txt

AA00029819_00001_0039.txt

AA00029819_00001_0086.txt

AA00029819_00001_0050.txt

AA00029819_00001_0004.txt

vp1mediatedrepre00hoec_0108.txt

vp1mediatedrepre00hoec_0055.txt

vp1mediatedrepre00hoec_0104.txt

vp1mediatedrepre00hoec_0041.txt

vp1mediatedrepre00hoec_0059.txt

vp1mediatedrepre00hoec_0075.txt

AA00029819_00001_0021.txt

vp1mediatedrepre00hoec_0044.txt

AA00029819_00001_0008.txt

AA00029819_00001_0038.txt

AA00029819_00001_0028.txt

vp1mediatedrepre00hoec_0058.txt

vp1mediatedrepre00hoec_0008.txt

vp1mediatedrepre00hoec_0069.txt

AA00029819_00001_0048.txt

AA00029819_00001_0110.txt

vp1mediatedrepre00hoec_0023.txt

vp1mediatedrepre00hoec_0090.txt

AA00029819_00001_0077.txt

AA00029819_00001_0030.txt

vp1mediatedrepre00hoec_0110.txt


Full Text













VP1-MEDIATED REPRESSION OF ALPHA-AMYLASE GENES
IN DEVELOPING MAIZE ALEURONE















BY


UTE HOECKER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1995









ACKNOWLEDGEMENTS

I am deeply grateful to Dr. Donald McCarty and Dr. Indra Vasil for their valuable advice, guidance and encouragement throughout the course of this study. I wish to thank Drs. Karen Koch, Nigel Richards, Alice Harmon and Bill Gurley for serving as members of my committee. I also thank all past and present members of both laboratories for their helpful discussions as well as for providing an enjoyable and lively work atmosphere. I am especially grateful to Lennie Rosenkrans and Dr. Mark Taylor for their immeasurable technical assistance. I wish to extend my thanks to Ms. Elaine Summers for her frequent help in dealing with the administrative side of Graduate School.


II















TABLE OF CONTENTS



ACKNOW LEDGMENTS .................................................................................................. ...... .

ABSTRACT ............................................................................................................................... IV

INTRODUCTION ..................................................................................................................... .1

REVIEW OF LITERATURE .................................................................. ....... ............ ....4

Developmental and Hormonal Regulation of Seed Maturation................................... 4
Isolation of Mutants Affected in Seed Maturation .......................................... 6
Analysis of Gene Expression ...................................................................... 13

The Aleurone Germination Response in Cereal Seeds ............................................. 21
Hormonal Regulation .................................................................................. 22
The c%-Amylase Genes ................................................................................ 23
The Organization of a-Amylase Prom oters ................................................. 25
GA and ABA Signal Transduction ............................................................... 28

The Developmental Switch from Seed Maturation to Seed Germination.................... 31
Quiescent Seeds.......................................................................................... 33
Dorm ant Seeds............................................................................................ 34

MATERIALS AND METHODS .............................................................................................. 39

Plant Material............................................................................................................... 39
Plasm id Constructs................................................................................................... 41
Particle Bom bardm ent and Transient Expression ...................................................... 43

RESULTS ................................................................................................................................. 47

Repression of Hydrolase Genes by VP1 in Aleurones of Developing Maize Seeds....... 47 Interaction between VP1 and Abscisic Acid ............................................................... 52
Over-expression of VP1 in Aleurones of Germinating Maize and Barley Seeds ..... 54 Interaction between VP1 and Gibberellic Acid ........................................................... 56
Role of the Embryo in Repression of c-amylase Genes in the Aleurone ....................61
Functional Analysis of the VP1 Protein .................................................................... 64

DISCUSSION ...........................................................................................................................74

SUM MARY AND CONCLUSIONS ............................................................................................ 90

REFERENCES ......................................................................................................................... 91

BIOGRAPHICAL SKETCH...................................................................................................... 104



III
















Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy VP1 -MEDIATED REPRESSION OF ALPHA-AMYLASE GENES IN DEVELOPING MAIZE ALEURONE


By
Ute Hoecker


August 1995


Chairman: l.K. Vasil
Cochairman: D.R. McCarty
Major Department: Plant Molecular and Cellular Biology

The Viviparous-1 (VP1) transcriptional activator of maize is required for abscisic acid-induction of maturation-specific genes late in seed development. In the presented work, it is shown that, in addition, VP1 inhibits precocious induction of the germination-specific a-amylase genes in aleurone cells of the developing seed. In developing seeds of the somatically unstable vpl-m2 mutant, hydrolase activity was de-repressed specifically in endosperm sectors underlying vpl mutant aleurone. Moreover, in transient expression experiments based on particle bombardment of aleurone tissue, a barley high-pl a-amylase promoter-GUS fusion construct (Amy-GUS) was induced in developing vpl mutant aleurone cells but not in wild-type aleurone cells. A direct role of VP1 in repression of Amy-GUS is suggested from the finding that co-expression of recombinant VP1 in vpl mutant aleurone cells strongly inhibited expression of Amy-GUS. Hence, VP1 expression in the developing seed appears to integrate the control of two developmental programs, seed maturation and seed germination.

Over-expression of VP1 also inhibited Amy-GUS expression in aleurones of wild-type germinating maize and barley seeds. In barley aleurone cells, VP1 specifically repressed induction of Amy-GUS by gibberellic acid (GA), while in maize aleurone tissue, VP1 inhibited a GA-dependent as well


iv









as an apparent GA-Independent activity. Deletion of the acidic transcriptional activation domain of VP1 did not affect the inhibitory activity, indicating that VP1 has a discrete repressor function. Further deletion analysis of VP1 showed that domains essential for repression of Amy-GUS are distinguishable from domains required for activation of the maturation-related genes Em and Cf.

The role of the embryo in the expression of Amy-GUS in developing maize aleurone cells was studied. Amy-GUS was de-repressed in vpl mutant aleurone in seeds that either carried a viviparous embryo or aborted the embryo early In development but not in seeds with a normal, non-viviparous embryo. This suggests that a normal embryo contributes a diffusible signal with inhibitory effect on AmyGUS expression in the aleurone. Amy-GUS was partially de-repressed in wild-type aleurone cells of embryo-less seeds, suggesting that both Vpl expression in the aleurone and a non-viviparous embryo are required for complete repression of a-amylase genes in the developing maize aleurone.


v




fT!~


INTRODUCTION


The formation of seeds is a unique characteristic of higher plants which promotes dispersal of the species and allows interruption of the life cycle during unfavorable environmental conditions. To survive in the dehydrated state, plant embryos undergo an adaptation process during late stages of seed formation (maturation phase) which renders them tolerant to desiccation and gradually causes arrested growth. In maize and other cereals, the outermost layer of the seed endosperm (aleurone layer) also undergoes a maturation process and remains viable through desiccation.

Seed maturation Is associated with the activation of a variety of genes encoding storage proteins and various hydrophilic, late-embryogenesis-abundant (LEA) proteins which possibly function as desiccation protectants (Dure et al. 1989; Skriver and Mundy 1990). Analysis of viviparous mutants in maize has demonstrated that the developmental program of seed maturation is controlled by at least two factors, the hormone abscisic acid (ABA) and the product of the Viviparous-I (VpI) gene (Robertson 1955; Neill et al. 1986). Developing vpl mutant embryos are distinct from ABA-deficient embryos in that they exhibit a reduced sensitivity to ABA in culture (Robichaud et al., 1980; Robichaud and Sussex, 1986). In addition to causing vivipary, the vpl mutation blocks synthesis of anthocyanins in embryo and aleurone tissues (Robertson, 1955; Dooner, 1985). The vpl mutant phenotype is restricted to seed tissues. Mutant embryos rescued prior to desiccation develop into apparently normal, fully fertile plants with normal pattems of anthocyanin accumulation.

The Vpl gene was cloned by transposon tagging (McCarty et al., 1989a). It encodes a 2500-nucleotide mRNA that is expressed specifically in embryo and endosperm tissues of the developing seed. Within the endosperm, Vp1 expression may be limited to the aleurone layer because so far no mutant phenotype has been detected in the starchy endosperm. This is


1







2

consistent with the relatively low abundance of Vp1 message In whole endosperm extracts (McCarty at al., 1989a).

Vpl encodes a novel, 73 kD protein with a functional acidic transcriptional activation domain (McCarty et al., 1991). Over-expression of VP1 In maize protoplasts trans-activated reporter constructs containing late-embryogenesis-specific promoters: C1, a maize gene that encodes a transcription factor required for anthocyanin synthesis in the seed, and Em, a wheat LEA gene (Hattori et al., 1992; McCarty et al., 1991). In agreement with the phenotype of ABAdeficient mutants, VPI-activation of Em was strongly dependent on the presence of exogenous ABA (McCarty et al., 1991). These functional data confirm that VP1 plays a central role in the induction of seed maturation.



Following imbibition of mature non-dormant seeds, expression of maturation-specific genes is terminated and expression of a new set of genes related to the developmental program of seed germination is executed (Comai and Harada, 1990). In rehydrated cereal seeds, the germination-specific c-amylase genes which encode starch-hydrolyzing enzymes are induced in the aleurone cells by the hormone gibberellic acid (GA) that is secreted by the embryo early in germination (Jacobsen and Chandler, 1987). They are constitutively expressed in de-germed seeds of the barley GA-response mutant slender (Chandler, 1988; Lanahan and Ho, 1988), and their induction can be antagonistically inhibited by application of ABA (Jacobsen and Chandler, 1987).



Expression of the normally consecutive programs of seed maturation and seed germination is under strict developmental control. Precocious induction of germination-related events prior to seed maturity appears to be actively repressed. In developing seeds of cereals and maize, no c-amylase activities are found prior to seed maturity (Evans et al., 1975; Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Moreover, ciamylase genes are unresponsive to applied GA (Nicholls, 1979; Comford et al., 1986; Garcia-







3


Maya et al., 1990; OishI and BOwley, 1990). It has been suggested that the presence of ABA In developing seeds Is responsible for the Inhibition of a-amylase genes at this developmental stage (King, 1976). However, treatment of maize developing seeds with the ABA synthesis inhibitor flouridone was not sufficient to sensitize the aleurone cells to GA, suggesting the action of additional factors in repressing a-amylase genes in the developing seed (Oishl and Bewley, 1990).



The objective of this study was to elucidate a role of VP1 In the negative regulation of aamylase gene expression in the developing maize seed. It is demonstrated that VP1 in addition to activating seed maturation programs blocks precocious Induction of germinationspecific c-amylase genes in the developing maize seed. A somatically unstable vpl mutant is described that displays de-repression of hydrolase activity specifically in vpl mutant sectors of the aleurone. Using a transient expression approach, it is shown that expression of recombinant VP1 in aleurone cells of maize and barley strongly inhibits expression of an a-amylase promoterGUS reporter gene (Amy-GUS). Evidence is provided indicating that VP1 specifically represses GA-induction of Amy-GUS in aleurone of germinating barley seeds. It is also shown that deletion of the acidic activation sequence of VP1 does not affect VP1 repressor activity, indicating that VP1 has a discrete repressor function. Thus, it is suggested that the coupled activator and repressor functions of VP1 play a key role in integrating the control of the normally not simultaneously occurring maturation and germination programs in the seed.















REVIEW OF LITERATURE


Developmental and Hormonal Reaulation of Seed Maturation


The biochemical mechanisms allowing seed tissues to tolerate extreme desiccation remain unclear. Many studies have implicated soluble sugars in desiccation protection. One of the suggested functions of soluble sugars is the protection of membranes which are often considered a primary site of desiccation damage (Crowe et al., 1992). It is thought that hydroxyl constituents of sugars substitute for water during dehydration and thereby stabilize membrane structures in the dehydrated state (Crowe et al., 1992). In agreement with this, di- and oligosaccharides, especially sucrose and in some species raffinose and stachyose, increase in concentration in maturing seeds (Amuti and Pollard, 1977) and In desiccating pollen grains (Hoekstra et al., 1989). Their accumulation has been correlated with the acquisition of desiccation tolerance (Hoekstra and van Roekel, 1988; Koster and Leopold, 1988; Chen and Burris, 1990; Leprince et al., 1990; Blackman et al., 1992; Crowe et al., 1992). However, recent comparative studies using desiccation intolerant mutants or recalcitrant (desiccation intolerant) species found no positive correlation between oligo-saccharide content in the seed and the development of desiccation tolerance (Ooms et al., 1993; Still et al., 1994). Data of one of the studies suggest that a low ratio of mono- to oligo-saccharides may be the critical factor rather than the absolute amount of soluble sugars (Ooms et al., 1993) This suggests that sucrose may be involved in the formation of 'glass' during dehydration which is promoted by oligosaccharides and inhibited by monosaccharides (Ooms et al., 1993). A glasso is a liquid of high viscosity, such that it stops or slows down all chemical reactions requiring molecular diffusion, and thus, might conserve tissue structures during dehydration (Bruni and Leopold, 1991).


4







5


A second characteristic that may be Involved In rendering the seed tolerant to desiccation Is the synthesis and accumulation of specific proteins late In seed development aateImbryogenesis-jbundant proteins, LEAs). The direct function of LEAs Is unknown but based on their high degree of hydrophilicity (Dure at al., 1989) they are assumed to stabilize the structure of cellular proteins during dehydration (Skdiver and Mundy, 1990; Dure, 1993).

Equally uncertain remain the mechanisms that arrest embryo growth and prevent precocious germination prior to seed maturity. Because Immature embryos excised from the seed and placed In culture are capable of germinating readily, precocious germination of the embryo in vivo may be actively suppressed by a process that is dependent on an intact seed. Evidence from embryo culture experiments has implicated two factors in suppression of precocious germination: restricted water uptake (low water potential in the seed) and the hormone abscisic acid (ABA). Both factors, when imposed on cultured embryos, inhibited germination (reviewed in Quatrano, 1987; Kermode, 1990). Indeed, the osmotic potential of developing soybean embryos has been shown to be even more negative than that of the osmoticum used to inhibit germination of isolated embryos (Xu et al., 1990). Similarly, ABA concentrations increase early in the seed maturation phase (Quatrano, 1987) and may thus play a role in arresting embryo growth.


In summary, the physiological processes responsible for acquisition of desiccation tolerance and arrest in embryo growth are poorly understood. In the past 15 years, the focus of research has shifted to the identification of regulatory factors that control the activities of late embryogeny, especially at the level of gene expression. From its discovery in the 1950s, ABA has been a factor of interest because a rise in seed ABA concentration correlates well with the onset of maturation events. Normally, ABA concentrations peak at the time of maximum dry weight accumulation in the seed and then decrease to low concentrations towards seed maturity (Quatrano, 1987). A function of ABA in initiating maturation events and suppressing precocious germination of the developing embryo was confirmed by analyses of mutants defective in late embryogeny and studies of gene expression.







a


Isolation of Mutants Affected in Seed Maturation


Genetic deficiencies In seed maturation manifest themselves In precocious germination (vivipary) or reduced dormancy. Severely affected mutants exhibit additional features, such as intolerance to desiccation and reduced accumulation of seed storage proteins and LEAs.

Mutants have been most Intensively Isolated and analyzed in maize and Arabidopsis. Maize has been a model species for studying genetics for many years, mostly because Its monoecious flower structure in combination with self fertility has allowed easy outcrossing and selfing. Moreover, the identification of several maize transposable elements (Ac/Ds, Spm/En, Mu) has made it possible to generate transposable element-induced mutants and subsequently clone the mutated locus using the transposon as a tag. Arabidopsis has become a powerful model species for a variety of reasons (such as short generation time, small genome size (Meyerowitz, 1987; Meyerowitz, 1994). Rather recently, cloning of mutated genes has been achieved by Aqrobacterium transformation-mediated T-DNA tagging or chromosome walking (positional cloning).

Many maize viviparous mutants that arose spontaneously were collected and described already during the first half of this century (Eyster, 1931; Mangelsdorf, 1930; Robertson, 1955). Transposon-tagged mutants were induced more recently (e.g. McCarty et al., 1989a,b). In Arabidopsis, many mutants displaying vivipary or reduced dormancy have been isolated by screening chemically mutagenized seed for germination on medium containing ABA at a concentration that inhibits germination of wild-type seeds. Other mutants have been isolated by screening for the ability to germinate in the absence of GA which is normally required. Here, GA contents in the seed were reduced either by treatment with a GA-synthesis inhibitor or by using a mutant deficient in GA-biosynthesis. Recently, screening of transgenic lines produced by Agrobacterium-mediated seed transformation (Feldmann, 1991) for mutants defective in late embryogeny has revealed additional loci that control induction of seed maturation/suppression of precocious germination. Taken together, the identified mutants fall into three classes: 1) ABA-







7


deficient mutants, 2) ABA-insensitive mutants and 3) mutants affected In a thus far unknown mechanism.


ABA deficient mutants

Five ABA-deficient mutants have been Identified in maize (vp2, vp5, vp7, vp6, vp9, Neill et al., 1986). Four of these mutants (vp2, vp5, vp7, vp9) lack carotenoids In addition to ABA (Robertson, 1955). These mutants accumulate various Intermediates of the carotenoid biosynthesis pathway, Indicating that they have lesions In carotenoid biosynthetic enzymes (Robertson at al., 1978). The deficiency in both carotenoids and ABA confirms that carotenoids are the precursor for ABA-biosynthesis in plants. Phenotypically, these mutants are viviparous, display a pale yellow to white coloring of the normally orange-yellow endosperm and form a lethal, white seedling which is deficient in chlorophyll due to photobleaching caused by the lack of carotenoids (Robertson, 1955; Anderson and Robertson, 1960).

Only one viviparous mutant (vp8) Is deficient in ABA without affecting carotenoid synthesis. vp8 mutant seedlings are viable but form plants of severely dwarfed stature (Robertson, 1955). The biochemical lesion of this mutant is unknown. It may be deficient in a later step in the ABA-biosynthesis pathway that is involved in the conversion of the carotenoid xanthophyll to ABA (Zeevaart and Creelman, 1988). Whether the lack of ABA causes dwarfism has not been examined thus far.

Vivipary in maize is determined by the genotype of the embryo and is entirely independent of the genotypes of the mother plant or the endosperm. Viviparous seeds segregate at the expected ratio on a heterozygous mother plant, indicating that ABA contributed by the mother plant mainly early in seed development does not play a role in preventing vivipary. The relative contribution of embryo and endosperm in preventing vivipary can be assessed by the use of TB-translocations which make It possible to generate seeds with embryo and endosperm of different.genetic constitution (e.g. a seed with a vp5 mutant embryo and a wild-type endosperm, or vice versa; Roman, 1947; Beckett, 1993). In these experiments, seeds with a vp5 mutant embryo and a wild-type endosperm were viviparous, while seeds with a wild-







8


type embryo and a vp5 mutant endosperm were not (Robertson, 1952). Similar results were obtained with other ABA-deficient mutants (Robertson, 1955). Thus, vivipary In those mutants is entirely conditioned by the lack of ABA production In the embryo. The endosperm does not play an active role in preventing vivipary of a genetically viviparous embryo. This is consistent with later findings showing that the embryo is the major source of ABA produced In the developing seed (Zeevaart and Creelman, 1988).


In Arabidopsis, only one ABA-deficient mutant (aba) has been Identified. It was isolated in a genetic screen selecting for the ability to germinate of the normally non-germinating GAdeficient ga-1 mutant (Koomeef et al., 1982). The aba mutant Is Impaired in the epoxidation of the carotenoid zeaxanthin (Duckham et al., 1991) and thus displays normal accumulation of carotenoids. No carotenoid deficient mutants have been isolated in any genetic screen, which may reflect the predicted lethality of such mutations. The aba mutant produces plants that show increased withering of stems, leaves and siliques and an enhanced rate of water loss which is probably caused by reduced stomata closure upon water stress (Koomeef et al., 1982). aba mutant seeds exhibit strongly diminished seed dormancy. Wild-type Arabidopsis seeds normally require cold and light treatments to break imposed seed dormancy and allow germination to occur, whereas a high percentage of aba mutant seeds germinated readily without a need for dormancy-breaking treatments (Koomeef et al., 1982). However, in contrast to ABA-deficient mutants in maize, even severe aba alleles that reduce ABA levels in the seed below the level of detection produce seeds that are non-viviparous and desiccation tolerant (Koomeef et al., 1982). Hence, ABA may not be required for the induction of desiccation tolerance in Arabidopsis. However, leakiness of the aba mutant cannot be ruled out. The nature of the performed mutant screen which selected for the ability of mature, dry seeds to germinate may not allow the identification of more severely affected mutants. Possibly, residual, very low concentrations of ABA that may be present in aba mutant seed are sufficient to prevent vivipary (Koomeef et al., 1989). To test this, a mutant screen could be performed that selects for the ability of seeds to







9


germinate precociously late In seed development. Such a screen has proven successful in isolating ab13 and other mutants (Keith et al., 1994).

Wild-type developing seeds of Arabkdopsls accumulate ABA as a dual peak, an earlier maternally-derived one and a later embryo-derived one (Karssen et al., 1983). Reciprocal crosses between wild-type and aba mutant plants demonstrated normal dormancy in the absence of maternal ABA but not In the absence of embryonic ABA. Moreover, Induction of dormancy, as judged from the acquired Inability of the developing seed to germinate precociously, correlated well with the later peak of ABA accumulation (Karssen et al., 1983). Hence, acquisition of a dormant state is dependent on ABA produced by the embryo and is normally independent of ABA provided by the mother plant. ABA-insensitive mutants

ABA-insensitive mutants of maize and Arabidopsis accumulate normal or higher concentrations of ABA in developing seeds as compared to wild-type (Neill et al., 1986; 1987; Koomeef et al., 1984). However, while the mutant phenotype of ABA-deficient mutants can be complemented by exogenous application of ABA, ABA-insensitive mutants continue to display vivipary or reduced dormancy in the presence of added ABA (Raubichaud et al., 1980; Raubichaud and Sussex, 1986; Koomeef et al., 1984). In a maize mutant, it was shown that the reduced sensitivity to ABA was not caused by a deficiency in ABA transport or metabolism (Raubichaud and Sussex, 1986). Thus, ABA-insensitive mutants are likely to be affected in ABA signal transduction.

A single locus (Vp1) regulating ABA-sensitivity has been identified in maize. vp1 mutant embryos do not acquire desiccation tolerance and germinate precociously on the ear producing green seedlings. The vpl mutation affects only seed tissues. When rescued and transferred to soil prior to desiccation, mutant seedlings form a normal appearing mature plant. Interestingly, this mutation causes a pleiotropic phenotype. Besides displaying vivipary, vpl mutant seeds fail to accumulate anthocyanin pigments in embryo and aleurone tissues (Robertson, 1955). Consistent with this phenotype, activities of enzymes catalyzing anthocyanin biosynthesis were







10

not detectable in mutant seed tissues (Dooner. 1985). The lack of anthocyanin pigments is not likely to be a result of the reduced ABA-sensitivity of this mutant. ABA-deficient mutants accumulate normal amounts of anthocyanins, Implying that ABA Is not required for pigment formation. Furthermore, separation of the two phenotypes was observed In seeds carrying the vpl-Mcfhrter allele. Those seeds are unpigmented but non-viviparous (Coe at al, 1978). Therefore, the pleiotropic phenotype of the vpl mutant implies that In evolution, two processes, suppression of precocious germination and production of anthocyanins, have come under the control of a single protein.

As In ABA-deficient mutants, the viviparous phenotype of the vpl mutant is entirely determined by the genotype of the embryo (Robertson, 1955). Similarly, anthocyanin deficiency in embryo or aleurone solely reflects lack of functional VP1 in the respective tissue (Robertson, 1955). Thus, the failure to accumulate pigments in vpl mutant aleurone is not a direct or Indirect result of precocious induction of germination. Cell autonomous function of VP1 in the aleurone was demonstrated in a transposable element-induced, somatically unstable mutant (vpl-muml). Homozygous vpl-muml seeds exhibit small sectors in the aleurone that have regained VP1 function due to excision of the Robertson's Mutator transposable element. These revertant sectors, recognizable by their pigmentation, can be as small as single cells, indicating that VP1 function does not result in production of a diffusable factor that might induce anthocyanins In neighboring cells (McCarty et al., 1989a).

The Vp1 gene was cloned by transposon tagging using the vpl-muml allele (McCarty at al. 1989a). It encodes a 2500 bp mRNA that is translated into a 73 kD protein. Vp1 is expressed specifically in embryo and endosperm tissues of the developing seed. Within the endosperm, Vpl expression is likely to be restricted to the aleurone, as suggested from the low abundance of Vp1 mRNA detected in whole endosperm extracts (McCarty at al. 1989a) and the apparent absence of a mutant phenotype in vpl mutant endosperm. The Vpl transcript is present in the seed as early as 10 days after pollination (DAP), reaches maximum accumulation at 16 DAP and decreases in abundance towards seed maturity. No Vp1 expression was detected in germinating







11


seeds, root or shoot tissues (McCarty et al. 1989a; Carson, 1992). Hence, the expression pattern of Vpl Is highly consistent with the seed-specific phenotype of the vpl mutant

The Vpl gene consists of six exons and five Introns. Apart from putative VP1 homologs cloned from barley, rice and Arabidopsis (M. Stoll and D.R. McCarty, unpublished results; Hattori et al., 1994; Giraudat et al., 1992), the sequence of VP1 shows no significant homologies to any known protein sequences, suggesting that VP1 Is a novel protein. The N-terminus of VP1 Is predicted to form two negatively charged amphipathic helices, a feature which Is characteristic of many bacterial and eukaryotic transcriptional activators (Ptashne, 1988). Indeed, this region of VP1 was found capable of functionally replacing the acidic activation domain of the bacterial transcrption factor GAL4 In a eukaryotic gene expression system (McCarty at al., 1991). This confirmed that the acidic region of VP1 has transcriptional activator function and suggested that VP1 may function as a regulatory protein in controlling seed maturation and anthocyanin accumulation.


In Arabidopsis, mutants displaying reduced sensitivity to ABA have been Identified using genetic screens selecting for the ability of seeds to germinate on medium containing at least 3 M ABA, a concentration that inhibits germination of wild-type seeds. In such screens, five loci controlling ABA-sensitivity have been identified: Abil, Abi2, Abi3 (Koomeef et al., 1984), Abi4 and Abi5 (Finkelstein, 1994). Mutations in any of these loci confer reduction in seed dormancy. However, while the phenotype of abi3, abi4 and abiS mutants is restricted to seed tissues, abil and abi2 mutants are also impaired in stomatal regulation and a variety of stress responses in vegetative tissues (Koomeef et al., 1984; Finkelstein, 1994; Chandler and Robertson, 1994). Interestingly, when the abi3-1 mutant was crossed to the ABA-deficient aba mutant, seeds of the resulting double mutant were desiccation intolerant, remained green and frequently displayed vivipary (Koomeef et al., 1989). The phenotype of the double mutant suggests that the abI3-1 allele may be leaky and allow some ABA-responsiveness. Hence, additional reduction in seed ABA concentrations may be necessary to obtain a viviparous phenotype. Indeed, severe abI3 mutants were isolated that were phenotypically similar to the abI3-1/aba double mutant,




-1


12

confirming that a strong abI3 allele conferring high Insensitivity to ABA is sufficient to cause vivipary (Nambara et al., 1992; Ooms at al., 1993).

The AbI3 gene was cloned by chromosome walking (Giraudat at al., 1992). The predicted protein of 79.5 kD displays discrete regions of sequence homology to the maize VP1 protein. Since there are also phenotypic similarities between the abI3 and vpl mutant at least with respect to seed-specific insensitivity to ABA AB13 and VP1 are likely to have similar functions in regulating ABA response in the seed. However, the functions of AB13 and VP1 differ in so far that AB13 is required for seed dormancy in Arabldopsis while VP1 does not impose seed dormancy in maize. Conversely, VP1, but not AB13, induces synthesis of anthocyanins in the seed. Whether these phenotypic differences reflect differences In sequence between AB13 and VP1 or the differential involvement of other factors remains to be determined.

The Abil gene was cloned in two laboratories by chromosome walking (Leung ot al., 1994; Meyer et al., 1994). At its C-terminus, the predicted ABII protein (47.5 kD) displays sequence similarity with the 2C class of serine-threonine protein phosphatases from rat and yeast. Its N-terminus exhibits features typical for a Ca"-binding site (EF hand). Hence, ABIl may function as a Ca"-dependent protein phosphatase. Indeed, regulation of stomatal aperture by ABA involves Ca" as a second messenger and protein phosphorylation events (Blatt and Thiel, 1993; Luan et al., 1993). How ABI1 may regulate ABA-induction of seed dormancy is thus far unknown.


Mutants affected in a thus far unknown mechanism

Three mutants of Arabidopsis (le1, 1ec2, fus3) have been isolated that are non-dormant but normal in their response to ABA (Meinke, 1992; Keith et al., 1994; Meinke et al., 1994; Baumlein et al., 1994). lec1 and fus3 have similar phenotypes. Immature mutant seeds germinate readily when placed in culture and display occasional vivipary when left to mature In the siliques. Furthermore, they are intolerant to desiccation and accumulate anthocyanins late In seed development, a feature that is not characteristic of wild-type seeds. Prematurely germinated seeds give rise to viable green seedlings that appear normal except that trichomes







13

are found on the adadal surfaces of its cotyledons. Trichomes normally form only on leaves, stems and sepals, but not on cotyledons. Hence, lecl and fus3 cotyledons are considered to be partially transformed Into leaves, which gave two of the mutants their name (loc, Gleafy cotyledons', Meinke, 1992). The lec2 mutant also exhibits leafy cotyledons and accumulation of anthocyanin, but differs from lecl and fus3 in that seeds are tolerant to desiccation and nonviviparous, and seedlings often appear distorted In shape (elongated hypocotyl, curled cotyledons) (Meinke et al., 1994). Unlike the abI3 mutant, germination of lecl and fus3 mutant seeds is inhibited by ABA, indicating that they retain normal sensitivity to ABA (Keith at al., 1994; Meinke et al., 1994). Proof for normal ABA synthesis in these mutants is still lacking. However, since ABA-deficient mutants have thus far not been shown to exhibit leafy cotyledons or accumulation of anthocyanins, it is unlikely that a possible lack of ABA would be the sole cause of the mutant phenotype. Nevertheless, the role of ABA in these mutants remains to be examined.

To investigate the interaction between abi3 and leafy cotyledon mutants, double mutants were constructed. abi3/ecl and abi3us3 double mutant seeds were highly viviparous, insensitive to ABA, exhibited leafy cotyledons and accumulated large amounts of anthocyanins (Meinke et al., 1994, BAumlein et al., 1994). The additive effect of abi3 and lec/1us3 in the double mutants suggests that abi3 and lec1 or fus3, respectively, are altered in distinct pathways. Consequently, suppression of precocious germination requires at least ABA, developmental factors controlling ABA-sensitivity and the leafy cotyledon-factors whose interactions with ABA, however, remain to be analyzed.



Analysis of Gene Expression


Late stages of seed formation are correlated with the expression of characteristic genes which were analyzed first and very extensively. in cotton embryos (Galau et al., 1986, 1987; Hughes and Galau, 1991). Based on changes in the levels of specific sets of cotton mRNAs, late seed development has been categorized into several stages (Galau et al., 1991). The







14

earlier amaturation stage" comprises the longest time interval (19 days) and Is characterized by high abundance of storage protein-mRNAs. This phase coincides with the presence of high concentrations of ABA. It Is apparently terminated by abscission of the vascular connections between embryo and mother plant and is followed by the "postabscisslon stage" (5 days) during which maturation stage-specific mRNAs decline rapidly and a new set of mRNAs accumulates. To these postabscission stage-specific mRNAs belong the LEA's (this term was introduced by Galau et al., 1988) and the RAB's (responsive to ABA, a term used by other authors for similar proteins as LEA's). Subsequently, seed formation Is terminated by rapid water loss and termination of transcription.

A similar temporal patten of mRNA accumulation was reported for Arabidopsis (Parcy at al., 1994) and maize (Paiva and Kriz, 1994; Williams and Tsang, 1994). Nevertheless, for species other than cotton, late stages of seed formation are usually referred to as the Omaturation phase" which is not subdivided into two stages. It should be mentioned that in maize, the seed maturation phase also correlates with the accumulation of anthocyanin pigments in embryo and aleurone tissues.

In the following, progress in our understanding of the regulation of maturation-specific genes (genes encoding storage proteins, LEA's and RAB's, proteins of the anthocyanin pathway) will be reviewed, placing emphasis on the roles of ABA and VP1 as regulators of seed maturation in monocot seeds.


Storaae Droteins

In many species, immature embryos cultured in ABA exhibited precocious and enhanced accumulation of storage proteins and their corresponding mRNAs as compared to those cultured on ABA-free medium or left to mature on the mother plant (Quatrano, 1987). These results indicate that ABA upregulates expression of storage protein genes. However, high levels of ABA are not required for expression of the major stQrage protein genes, as shown for maize and Arabidopsis. The aba mutant of Arabidopsis was found to accumulate normal levels of identified ABA-upregulated storage proteins (2S, 12S) (Koomeef et al., 1989) and their corresponding







15

mRNAs (Parcy et I., 1994). Similarly, maize ABA-deficient mutants accumulated mRNAs corresponding to the ABA-regulated 7S globulins in only slightly reduced amounts (KrIz et al., 1990; Paiva and Krlz, 1994). Hence, ABA does not normally appear to be a limiting factor In expression of these storage protein genes. In contrast, developing embryos of the ABAinsensitive mutants vpl of maize and abi3 of Arabidopsis exhibited very low or undetectable expression of 7S globulins or 2S and 12S storage protein genes, respectively, indicating that VP1/AB13 are required for their expression (Kriz et al., 1990; Palva and KrIz. 1994; Nambara et al., 1992). Furthermore, exogenous ABA did not induce expression of storage protein genes in cultured immature vpl or abi3 mutant seeds, while it did so in cultured immature wild-type seeds (Paiva and Kriz, 1994; Finkelstein and Somerville, 1990). Thus, VP1/AB13 appear to be essential for ABA action. Since VP1/AB13 are expressed at normal levels in ABA-deficient mutants (McCarty et al., 1991; Paiva and Kriz, 1994; Parcy et al., 1994), it may be that normal accumulation of storage proteins in these mutants is mediated by VP1/AB13 either without a need for ABA or requiring residual amounts of ABA present in mutant seeds. LEAs/RABs

An extensive survey in Arabidopsis examining the accumulation kinetics of 18 marker mRNAs expressed at high levels during mid to late seed development suggested that LEAs/RABs-encoding genes fall into distinct classes with different requirements for ABA and AB13 to induce expression. For most markers, transcript levels did not solely correlate with the amounts of endogenous ABA or AB13 present in the seed, thus implicating a role of other factors in controlling temporal patterns of expression (Parcy et al., 1994).

However, abundance of several tested mRNAs was highly reduced in seeds of the aba mutant as well as the abi3 mutant (Parcy et al., 1994). Similarly in maize, expression of a wellstudied LEA gene (Em) originally isolated from wheat was undetectable in developing seeds of mutants deficient for ABA or functional VP1 (McCarty et al., 1991). Hence, ABA as well as AB13/VP1 appear to be required for expression of certain LEAs, which is consistent with a possible role of AB13NP1 in ABA perception or signal transduction in the seed.







16

To Investigate further the Interaction of VP1 and ABA In controlling expression of the Em gene In the maize seed, a transient gene expression system was used that is based on electroporation of maize protoplasts Isolated from an Immature embryo-derived suspension (Vasil et al., 1989). A plasmid containing the promoter (0.6 Kb) of the Em gene fused to the coding sequence of the bacterial p-glucuronidase (GUS) gene (Em-GUS; Marcotte et al., 1988) was used as a reporter construct. The VP1 cDNA was over-expressed from the constitutive CaMV 35S promoter enhanced by insertion of the first Intron of the maize Shi gene (Vasil et al., 1989) into the 5' untranslated leader of the VP1 cDNA (35S-Sh-VP1). In these experiments, electroporation of protoplasts with a mixture of Em-GUS and 35S-Sh-VP1 resulted in 100-300fold higher GUS activity as compared to the very low Em-GUS activity detected In the absence of co-electroporated 35S-Sh-VP1 (McCarty et al., 1991). Similar activation was obtained when protoplasts electroporated with Em-GUS were cultured in ABA (McCarty et al., 1991), which Is consistent with the reported ABA-regulation of Em-GUS in rice protoplasts (Marcotte et al., 1988; 1989). Over-expression of VP1 in maize protoplasts Interacted synergistically with ABA, resulting in 2,500-fold induction of Em-GUS (McCarty et al., 1991). The synergistic effect of VP1 and ABA underlines the importance of both, VP1 and ABA, in high-level expression of Em. However, the substantial activation of Em-GUS obtained by either over-expressing VP1 or culture in ABA might imply that VP1 and ABA can partially activate Em-GUS independently. This would be in contrast to the absence of detectable Em transcript in vpl or vp5 mutant embryos which contain normal levels of ABA or Vpl transcript, respectively (Neill et al., 1987; McCarty et al., 1991). However, action of endogenous ABA and VP1 that may be present in the wild-type protoplasts cannot be ruled out. To test this, Em-GUS was introduced into vpl and vp5 mutant seed tissue (aleurone) via particle bombardment. In these experiments, ABA did not activate Em-GUS in vpl/vp5 double mutant tissue while it did so in VPIA/p5 tissue or when cobombarded with recombinant VP1 (S. Cocciolone and D. R. McCarty, unpublished results), thus confirming that functional VP1 is required for ABA action. In vp5 mutant tissue, over-expression of VP1 slightly activated Em-GUS, though 10-fold lower than in the presence of exogenous ABA.







17


This apparent independent activity of VP1 may be caused by low levels of maternal ABA present in vp5 mutant seeds. Alternatively, the abnormally high levels of recombinant VP1 In expressing cells may allow some ABA-independent activation of the Em promoter normally not found In vivo.

The VP1 protein was subjected to functional analysis by testing deletion-derivatives for their ability to buns-activate Em-GUS. Sequence analysis and domain swapping experiments between VP1 and GAL4 had suggested that the N-terminus of VP1 contains an acidic transcriptional activation domain (McCarty et al., 1991). Indeed, deletion of this acidic domain abolished VPI's ability to activate Em-GUS. Replacing it with the acidic activation sequence of the herpes simplex virus transcription factor VP16 partially restored transcriptional activation of Em-GUS (McCarty et al., 1991). These results strongly Indicate that VPI functions as a transcriptional activator in inducing Em-GUS. Analysis of Internal deletion-constructs of VPI identified two highly basic domains that are important for activation of Em-GUS (L. Rosenkrans, V. Vasil, L.K. Vasil and D.R. McCarty, unpublished results).

Deletion analysis of the Em promoter indicated that two G-box-related sequences (Emla: ACAOGTGG; Emib: ACACQIGC) which are conserved in many promoters responsive to ABA, light or anaerobiosis are involved in VP1- and ABA-mediated activation of Em (Marcotte et al., 1989; Guiltinan et al., 1990; V. Vasil et al., unpublished results). The finding that ABAinduction of Em does not require protein synthesis (Williamson and Quatrano, 1987) suggests that VP1 and ABA trans-activate Em directly through the G-box elements rather than through activation of intermediate regulatory genes further upstream in the ABA signal transduction pathway. Thus far, no DNA-binding activity of VP1 to these putative target sequences has been detected (T. Hattori, B. U and D.R. McCarty, unpublished results). Hence, VP1 might activate Em via protein-protein interactions with G-box-binding protein(s). A bZIP-type protein binding specifically to the Emla motif has been cloned (Guiltinan at al., 1990) and may thus be a candidate.







18


The molecular mechanism underlying the synergistic Interaction of VP1 and ABA Is thus far unclear. The possibility that ABA Is required for high stability of the VP1 transcript or protein which then in turn activates Em is highly unlikely because In the ABA-deficient vp5 mutant no reduction in VP1 transcrpt and protein levels was observed (McCarty et al., 1991; C. Carson and D.R. McCarty, unpublished results). Possibly, ABA post-translationally modifies the VP1 protein and thus enhances its trans-activation function. Alternatively, VP1 and an ABA-dependent factor might be part of a complex that forms on the Em promoter and Induces transcription.


Genes encoding Droteins of the anthocyanin Dathwav

As Identified by mutants, at least eight genes are essential for accumulation of anthocyanin pigments in embryo and aleurone cells of the maturing maize seed (Fig. 1). Five of these genes (Al, AZ CZ Bzl, Bz2) encode enzymes of the anthocyanin biosynthesis pathway (Dooner et al.. 1991). Expression of these structural genes requires the coordinate action of two regulatory proteins, C1 and a member of the R/B gene family (Coe et al., 1988). Both proteins exhibit features of transcription factors. The C1 protein contains a functional acidic transcriptional activation domain (Goff et al., 1991) and a region of sequence homology to the DNA-binding domain of animal myb proto-oncogene products (Paz et al., 1990). Proteins encoded by the R/B gene family display high homology with the helix-loop-helix motif of myc proto-oncogene products (Ludwig et al., 1989). At least for the promoter of the Bzl structural gene, it has been shown that sequences homologous to the consensus binding sites of animal MYB and MYC proteins are essential for Cl- and R-mediated activation of Bzl expression (Roth et al., 1991), suggesting a direct interaction of C1 and R with these sequences in the Bzl promoter.

A third regulatory factor required for pigmentation of tissues in the developing maize seed is VP1. The lack of anthocyanin pigments in vpl mutant seed is associated with the absence of Cl transcript (McCarty et al., 1989a). Over-expression of a 35S-C1 construct in vpl






19


ABA


Em


x


VP1


Y


C1


R


1.


Maturation Genes


Al


A2


Bz1


Bz2


C2


Anthocyanin Genes




Fig. 1. Role of VP1 in activation of seed maturation-related pathways and anthocyanin biosynthesis.







20

mutant aleurone cells by particle bombardment complemented the failure to accumulate anthocyanins In a cell autonomous fashion (Hattori et al., 1992). Thus, lack of C1 appears to be responsible for the block in anthocyanin synthesis in vpl mutant seed. A direct role of VP1 in activating C1 expression was concluded from the demonstration that over-expression of VP1 in maize protoplasts activated transcription of a C1 promoter-GUS fusion gene. Most Importantly, VP1 function in activating C1 was dependent on its transcriptional activation domain (Hattol at al., 1992). Hence, VP1 and C1 are part of a regulatory hierarchy controlling activation of anthocyanin structural genes.

Several lines of evidence indicate that VPI's function in activating C1 Is distinct from its function in activating Em. First, though induction of both genes Is dependent on the acidic activation domain of VP1 and thus appears to involve transcriptional activation, other domains of VP1 involved in function differed depending on the target promoter. While sequences in the middle of the VP1 protein were required for trans-activation of Em-GUS in maize protoplasts, VP1-activation of Cl-GUS was dependent on the C-terminal end of VP1 (L. Rosenkrans, V. Vasil, L.K. Vasil and D.R. McCarty, unpublished results). This is consistent with the nonviviparous/unpigmented phenotype of the vp1-McWhirter mutant which produces a truncated VP1 protein lacking ca. 150 bp from the C-terminus (McCarty et al., 1989b).

Second, in agreement with the involvement of distinct domains of VP1, different ciselements in the C1 and Em promoters appear to be the target of VP1 function. In contrast to activation of Em which depended on two G-box sequences, activation of C1 did not require any of the two G-box-like sequences present in the promoter but the 13 bp sequence -145 TCCATGCATGCAC -158 (Hattori et al., 1992). This sequence, designated as Sph-element, is found in promoters of other seed-specific genes (Dickinson et al., 1988).

Finally, VPi-mediated activation of Em and C1 differ in their interaction with ABA. While there is a synergistic effect of VP1 and ABA in activating Em, the role of ABA in VP1mediated activation of C1 is less clear. Anthocyanins accumulate at normal levels in ABAdeficient mutants of maize, suggesting that C1 expression is hormone-independent. On the







21

other hand, ABA fran-activated Cl-GUS in maize protoplasts. Moreover, a c1 mutant (cl-p Chen and Coe, 1978) that falls to accumulate anthocyanin during seed development carries a 5 bp deletion in the promoter region of the gene that when reconstructed by site-directed mutagenesis of the C1 promoter and used In transient expression experiments, specifically abolished ABA-responsiveness without severely affecting rans-activation by VP1 (Hattorl et al., 1992). Hence, the unpigmented phenotype of cl-p may be caused by a deficiency in ABAresponse. Pigmentation in ABA-deficient mutants might be possible if activation of C1 has an ABA-requirement several orders of magnitude lower than activation of Em.






The Aleurone Germination Response in Cereal Seeds


During cereal seed development, the outermost endosperm cells differentiate into the aleurone layer which at seed maturity consists of small, thick-walled cells with plasmodesmatal connections. The aleurone cells are characteristically rich in protein and lipid bodies, mitochondria and ER, but are devoid of starch grains. In response to GA released by the germinating embryo, these highly specialized cells synthesize large amounts of hydrolytic enzymes co-translationally on the rough ER and following proper folding in the lumen of the ER

- secrete these enzymes into the endosperm. This aleurone germination response to GA can be inhibited by treatment with ABA.

The predominant hydrolytic enzyme synthesized is a-amylase, constituting ca. 15-20% of total translatable mRNA and ca. 30% of total protein synthesis in germinating barley seeds (Khursheed and Rogers, 1988). In barley and wheat, the major source of a-amylase is the aleurone layer, whereas in maize, sorghum and rice, significant contributions are also made by the embryo scutellum (Ranki and Sopanen, 1984; Dure, 1960a,b; Akazawa and Miyata, 1982).







22


Hormonal Reaulation


Extensive studies in barley and several investigations with other cereal species have shown that synthesis of cereal a-amylases is induced by GA and antagonistically Inhibited by ABA (for review see: Jacobsen and Chandler, 1987; Jones and Jacobsen, 1991; Fincher, 1989). The first reports on GA-induced a-amylase activity in germinating barley seeds appeared in 1960 (Paleg, 1960a; b; Yomo, 1960). The action of ABA as an antagonist of GA was discovered In 1966 (Chrispeels and Vamer, 1966).

Application of transcription and translation inhibitors Indicated that the GA-induced appearance of a-amylase activity was due to de novo synthesis of the enzyme (Vamer and Chandra, 1964; Filner and Vamer, 1967). Ultimate proof for an effect of GA and ABA on aamylase synthesis came from the demonstration that GA treatment drastically increased the amount of in vitro translatable a-amylase mRNA (Higgins et al., 1976), while simultaneous application of ABA blocked this effect (Mozer, 1980). Run-on transcription experiments provided evidence that GA and ABA regulate the transcription of a-amylase genes (Jacobsen and Beach, 1985; Zwar and Hooley, 1986). Eventually, the development of transient gene expression technology has provided an additional tool to elucidate GA and ABA action. It was shown that transient expression of a wheat a-amylase promoter-GUS reporter gene fusion construct in oat aleurone protoplasts was regulated in the same manner as the endogenous genes (Huttley and Baulcombe, 1989).

Relatively few studies have addressed hormonal regulation of maize a-amylase. Ingle and Hageman (1965) reported a stimulating effect of GA on catabolism of carbohydrates in excised endosperms. In a different study (Harvey and Oaks, 1974), exogenous GA applied to excised endosperms further increased (3-fold) total amylase activity in germinating seeds of a GA-deficient mutant (d5), but not in wild-type seeds. In contrast, culture of wild-type endosperms in ABA strongly reduced amylase activity. Even though no molecular data are available, these results indicate that maize a-amylase genes are probably regulated by GA and







23

ABA in a similar fashion as other cereal a-amylases. Ultimate proof for this, however, Is still lacking.

It has now been generally accepted that the embryo is the site of GA biosynthesis in the germinating barley grain. When de-embryonated seeds are imbibed, no Increase In GA levels in the endosperm and very little subsequent production of a-amylase can be detected (Jacobsen and Chandler, 1987). Further studies demonstrated that the scutellum, rather than the embryo axis, is the source of GA (Radley, 1967; MacLeod and Palmer, 1967).

In maize, evidence for the importance of the germinating embryo as a source of GA has been contradictory. in imbibed de-embryonated seeds, Dure (1960) found only -amylase (which is stored in protein bodies In the dry seed and therefore is not de-novo synthesized during germination), but no a-amylase activity, whereas whole kemels showed both activities. Two other studies (Harvey and Oaks, 1974; Goldstein and Jennings, 1978), however, demonstrated comparable total amylase activities in de-germed as well as whole seed endosperm. As only part of the activity was due to release of p-amylase from protein bodies, it was concluded that mature seeds store considerable amounts of GAs in the endosperm. Therefore, the germinating maize embryo does not appear to be an essential source of GA for a-amylase synthesis in the aleurone.



The a-Amylase Genes


a-amylases of cereals can be biochemically separated into a number of isoforms that differ in their isoelectric point (pl) but not considerably in their molecular weight. In barley, there are two families of isozymes, the low-pi a-amylases with pis of ca. 4.4-5.2 and the high-pi aamylases with pis of ca. 5.7-6.2 (Jacobsen and Chandler, 1987). These two families differ in many other biochemical characteristics while isozymes within those families are more alike. Though some of the variants are post-translational modifications of the same gene product, a genetic basis for most of the variation seen became evident when the gene(s) for the low and high pi families were mapped to different chromosomes, chromosome I and 6, respectively







24

(Brown and Jacobsen, 1982; Muthukrishnan et al., 1984). Isolation and sequencing of a number of cDNA clones verified that there are sequence differences between isoforms. Base sequence homology is 90-95% within gene families and about 75% between gene families (Jacobsen and Chandler, 1987). Southern blot analyses of two barley varieties revealed that there are at least 6-7 high-pi genes and at least 3 low-pi genes, indicating that a-amylases are encoded by two multigene families (Khursheed and Rogers, 1988; Muthukrishnan at al., 1984). While no detailed mapping has been undertaken in barley, three rice ct-amylase genes were found to be clustered within 28 kb of genomic DNA. Molecular analysis suggested gene duplication as a cause (Sutliff et al., 1991).

The two barley a-amylase gene families are regulated differently in the aleurone of germinating seeds. mRNA levels of the high-pl isoforms increase very quickly upon imbibition, reaching a maximum after two days and decreasing to low levels after four days. Synthesis of low-pl isoform-mRNAs begins later, not before three days after imbibition, but then increases rapidly so that low-pi isozymes become the dominant enzyme group after four days of imbibition (Chandler and Jacobsen, 1991). The two isoforms are also differentially responsive to GA. In some studies with isolated aleurone layers, mRNA as well as protein of the low-pl isoforms can be detected before GA is added, while those of high pl-isoforms cannot (Chandler and Jacobsen, 1991; Jacobsen and Higgins, 1982; Rogers, 1985). Others, using isolated aleurone layers or aleurone protoplasts, find no low-pi message in the absence of GA (Chandler and Jacobsen, 1991; Nolan and Ho, 1988). However, low-pi isoforms appear to be more sensitive to GA as they respond to GA-concentrations as low as 10-9 M (Nolan and Ho, 1988). Thus, low-pi aamylase genes might be either more responsive to GA or leaky in expression.

In comparison to barley a-amylases, many fewer studies on maize cx-amylases have been reported. Partial purification of a-amylases from endosperm of germinating seeds has revealed two major groups of isozymes, one with pis of 5.1-5.7, the other with pis of about 4.6 (Warner and Knutson, 1991). Other authors report the purification of a-amylase isozymes with a variety of pis (Warner et al., 1991; MacGregor et al., 1988; Chao and Scandelios, 1971). No







25

genes coding for a-amylases have been cloned so far. Thus no information about gene expression is available.



The Oruanization of a-Amvlase Promoters


Gene expression is thought to be regulated by proteins ("bns-acting factors') that bind in a sequence-specific manner to short stretches of base pairs ("cis-acting elements') located in the promoter region of the gene. With an interest to study regulation of a-amylase gene expression, genomic clones were isolated. Sequence comparisons revealed little homology between promoters of a-amylase genes belonging to different pi groups which may relate to their differential expression in response to hormones described above. However, a few blocks of sequence were found highly conserved among barley, wheat and rice a-amylase promoters:




High-pi, barley CGCCTTTTGAGCTCACCGTACCGGCCGATAACAAACTCCGGCCGACATATCCACTG -117 (Khursheed and Rogers, 1988)


Low-pI, barley GCACCTTTTCTCGTAACAGAGTCTGGTATCCATGCA -98 (Whittier et al., 1987)

Low-pI, wheat GCACCTTTTTTCGTAACAGAGTCTGGTATCCATGCA -95 (Huttley et al., 1992)




To identify cis-acting elements involved in hormone-regulated expression, functional analyses of a-amylase promoters have been performed. For this purpose, mutated promoter sequences are fused to a reporter gene (e.g. GUS, Luciferase) and are assayed for function in a transient gene expression system (electroporation of aleurone protoplasts or particle bombardment of intact aleurone layers).

Progressive 5' truncations of a-amylase promoters showed that 289 bp of a wheat low-pI a-amylase promoter (Huttley and Baulcombe, 1989) and 174 bp of a barley high-pi a-amylase promoter (Jacobsen and Close, 1991) were sufficient to direct GA- and ABA-regulated







26


expression of a reporter gene, indicating that d.-acting elements are positioned in the proximal region of the promoters. Indeed, Skriver at al. (1991) demonstrated that a chimeric construct containing 69 bp (-189 to -120) of the barley high-pi promoter fused to the 35S TATA box could impose increased transcription by GA and its suppression by ABA. Moreover, six tandemly repeated copies of the sequence GGCCGAIAMCTCCGGCC (21 bp) conferred proper GAand ABA-regulation. However, this result could not be confirmed when particle bombardment was used as the method of transformation (J. Rogers, pers. communication; U. Hoecker, unpublished results) suggesting that other cls-elements apart from TAACAAA contribute to GAregulated transcription. This was confirmed when clustered point mutations were introduced covering the proximal region of the promoters. Mutations in the pyrimidine box (CCTTTT) or in the TATCCACIT box reduced GA-induced transcription to about 20 % of minimal level in the barley high-pl and low-pi promoters (Gubler and Jacobsen, 1992; F. Gubler, pers. communication; Lanahan et al., 1992). Thus the entirety of the three conserved elements appears to be involved In mediating GA-response. Interestingly, In no case could ABAresponsive elements be separated from GA-responsive elements, suggesting that GA and ABA function through the same cis-elements in the a-amylase promoters.

Rogers and co-workers (Lanahan et al., 1992) identified an additional element in the lowpl promoter that is located between positions -152 and -134, just upstream of the pyrimidine box. Mutations in this element reduced the level of expression by 96% while retaining significant but low GA-responsiveness. This region of the promoter (termed 002SO element) shows sequence homology to two well-described motifs: the "endosperm box", a conserved element present in promoters of maize, barley and wheat endosperm protein genes, and the consensus sequence for binding of the maize Opaque-2 protein which is a leucine zipper protein (bZIP) that is necessary for transcription of the 22 kDa zein genes. Interestingly, it was found that substitution of the GA-responsive TAACAGA sequence of the low-pi promoter with an ABA response element (ABRE) from the rice Rab-1&A gene converted the promoter from a GA-upregulated one into one whose transcription was increased by ABA (Rogers and Rogers, 1992). Thus, the ABRE







27

regulated transciption in the context of the low-pi a-amylase promoter In a similar way as it does In Its native rice promoter. Importantly, Its function In the amylase promoter was highly dependent upon the presence of the 02S sequence. Thus, the 02S element appears to function as a *coupling elements that is necessary for high-level, hormone-regulated transcription from the low-pi ct-amylase promoter (Rogers and Rogers, 1992). No 02S-like sequence is evident in high-pi a-amylase promoters. However, Inserting the 02S element from a low-pI promoter into a high-pi promoter at a position upstream of the pyrimidine box enhanced transcription ca. 5-fold, suggesting that the 02S element function could Interact properly with the high-pl promoter fragment to give high-level transcription (Rogers et al., 1994).

To identify the trans-acting factors that bind to the cds-elements In the promoters and thereby confer hormone-dependent expression of the a-amylase genes, DNA-protein Interactions have been characterized using band shift assays and DNase I footprinting analyses. In a wheat low-pl promoter (a-Amy 2/54), the 02S box and the TAACAGA element, but not the TATCCAC sequence, were found protected from DNase I digestion, confirming the binding of protein factor(s) to at least two functionally important elements. However, these binding activities were not dependent upon GA (Rushton et al., 1992). Evidence for the presence of a GA-dependent factor on a barley low-pl promoter was provided by Sutliff et al. (1993). Results from band shifts performed in this study demonstrated that a GA-inducible binding activity interacted with the TAACAGA and TATCCAC elements in a sequence-specific manner. In a different report on a rice a-amylase promoter (Amy3c), GA-dependent binding to a pyrimidine box-like sequence was demonstrated (Goldman et al., 1994). However, since this promoter displays little sequence homology to barley and wheat promoters, comparisons cannot be made. In summary, GAindependent as well as GA-dependent factors appear to constitute the complex(es) on a-amylase promoters that result in GA-regulated expression. Further characterization of the protein-DNA and protein-protein interactions is needed.







28


GA and ABA Sianal Transduction


it Is generally accepted that molecules with hormonal function act as Ilgands that upon binding to their specific receptors elicit a response which ultimately can result in altered gene expression. In animals, a large number of hormone receptor-encoding genes has been cloned and characterization of their gene products has demonstrated that receptors can be found located intracellularly (e.g. steroid hormone receptors) or integrated into the cell membrane with their ligand-binding domain facing the extracellular space. Identification of plant hormone receptors has proven difficult. Though hormone-binding proteins have been identified, definite proof for the function of these proteins is still lacking.

With respect to GA and ABA signal perception in the cereal aleurone, studies have so far concentrated on the Identification of the cellular site of the receptors. Results from at least one study suggest that GA does not have to enter the cell to regulate gene expression. Hooley and colleagues (Hooley et aI., 1991) demonstrated that GA immobilized to Sepharose beads was capable of enhancing c-amylase transcription in oat aleurone protoplasts. Though these data point to a perception of the GA signal on the extemal surface of the plasma membrane, the existence of intra-cellular receptors cannot be ruled out. Clear evidence against intra-cellular receptors came from elegant experiments performed in the laboratory of R. L. Jones. A method was developed to visualize c-amylase gene expression and c-amylase secretion from individual protoplasts (Hillmer et al., 1992). This allowed to test whether or not hormones microinjected into the cytosol of aleurone protoplasts were capable of eliciting a response (Gilroy and Jones, 1993). It was found that protoplasts injected with GA did not respond to the hormone. Similarly, ABA microinjected into protoplasts was ineffective in antagonizing the stimulating effect of preapplied extemal GA. The failure to respond to microinjected hormones was not due to disruption of protoplast function by microinjection since protoplasts that had been subjected to this procedure remained responsive to externally applied GA. These results indicate that the sites of perception of GA and ABA are located on the extemal face of the plasma membrane in aleurone cells.







29

There is little concept of how the perceived GA and ABA signals are transduced from the receptors to the nucleus. The function of Ca* and calmodulin (CaM) as second messengers that regulate protein kinase activity has been characterized in many animal and a few plant systems (Roberts and Harmon, 1992; Neuhaus et al., 1993). Also in barley aleurone protoplasts, treatment with GA was found to increase cytoplasmic Ca+ and CaM concentrations (Gilroy and Jones, 1992, 1993). ABA reversed the effect of GA on [Ca"]1 (Gilroy and Jones, 1992). Even though the increase in [Ca+i and [CaM] preceded the GA-induced Increase in a-amylase activity by 2-4 h (Gilroy and Jones, 1992), direct evidence for an involvement of Ca* and CaM in the regulation of c-amylase transcription is still lacking. Ca* and CaM have been found to regulate the activity of a slow vacuolar ion channel located in the tonoplast of storage protein vacuoles in barley aleurone cells (Bethke and Jones, 1994). Moreover, CaM was shown to stimulate Ca"+ uptake into the ER where the Ca* containing cx-amylase enzyme Is synthesized (Bush et al., 1993). Thus, hormone-regulated changes in Ca* and CaM concentrations may be regulating processes such as o-amylase formation and secretion rather than having a direct effect on the transcription of cx-amylase genes.

Analysis of mutants provides a valuable tool to study the genetics underlying the regulation of hormone action. Many mutants have been isolated that display altered responses to GA, suggesting that these mutants are affected in a component of GA signal transduction. Since GA promotes stem and leaf elongation, these mutants have been identified by their altered plant height. They fall into two classes: 1) those that show a reduced sensitivity to GA (GAinsensitive mutants') and are therefore of dwarf stature, and 2) those that show an enhanced sensitivity to GA (constitutive response mutants") and are therefore excessively tall. In response mutants, the concentrations of biologically active GAs do not in accord with the phenotype. Generally, GA-insensitive mutants accumulate higher concentrations of active GAs as compared to wild-type, while tissues of constitutive response mutants contain reduced GA concentrations (Stoddart, 1984; Fujioka et al, 1988; Croker et al., 1990). These observations







30

have been attributed to feedback regulations on GA metabolism in response to altered GAsensitivity.

To the class of GA-Insensitive mutants belong the Rht (reduced height) mutants of wheat. A total of 10 Rht loci have been identified, showing varying degrees of dominance. Of these, the mutation Rht3 exerts the strongest dwarfing effect. The GA-insensitive phenotype of Rht3 is also expressed in the aleurone: germinating seeds had 75% reduced levels of amylase activity as compared to tall (rhi) varieties and showed no or little increase In amylase activity after GA treatment (Gale and Marshall, 1975; Fick and Qualset, 1975). The degree of GAinsensitivity of the aleurone was found to increase with the dosage of Rht3 alleles (Gale and Marshall, 1975). A similar GA-insensitive mutant has been described In rice (Mitsunaga et al., 1994). The finding that the failure to respond to GA is expressed in plant and seed tissues indicates that these tissues share at least in part a common signal transduction pathway.

Two dominant GA-insensitive dwarf mutants have been Identified In maize (D8, D9). Besides being of reduced stature, these mutants mimic additional characteristics of GA-deficient mutants, such as reduced apical dominance and formation of anthers on the ear (Coe and Neuffer, 1977). D8 and D9 are located on IL and 5S, respectively (Coe and Neuffer, 1977). Since the region of 5S contains duplicate loci with 1 L it is likely that D8 and D9 are duplicate loci encoding gene products with identical or similar function. X-ray-induced chromosome breakage was used to create clonal sectors of wild-type cells within D8 mutant tissue. Results from these experiments indicated that D8-mediated effects can be expressed cell autonomously at least in some tissues (Harberd and Freeling, 1989) which is consistent with the hypothesis that the wildtype gene product is part of a GA signal transduction pathway. The gain-of-function nature of the mutation in association with a GA-insensitive phenotype allows to speculate about the function of the wild-type gene product. Possibly, d8(+) (and dg(+)) encodes a negative regulator of GA response that normally is inactivated by exposure of the cell to GA. In this scenario, the mutant D9 protein would be constitutively active in this repressor activity, i.e. even in the absence of GA (Harberd and Freeling, 1989).







31


A constitutive response mutant has been Identified In barley (Foster, 1977). This recessive mutation, termed "slende( (sin), causes a plant to appear as If it had been treated with high doses of GA (Lanahan and Ho, 1988; Croker et al., 1990). Also, c-amylase genes were highly expressed in sin mutant half grains in the absence of applied GA (Chandler, 1988). Thus, the absence of functional SLN protein causes constitutive expression of GA-responses and thereby uncouples transcription of GA-regulated genes from a need for GA. This phenotype suggests that Sin encodes a negative regulator of GA-response. Importantly, x-amylase production in sin mutant aleurones was susceptible to Inhibition by ABA, Indicating that the sin mutant retains normal sensitivity to ABA (Chandler, 1988; Lanahan and Ho, 1988). Hence, ABA most probably functions at a step downstream of GA in the signal transduction pathway leading to regulation of c-amylase transcription. A function of ABA fully Independent of GA cannot be ruled out but is unlikely because GA and ABA appear to act through the same response elements in a-amylase promoters. The findings also indicate that GA and ABA do not act at the same site in the signal transduction pathway, I.e. they do not for example antagonistically phosphorylate/ de-phosphorylate an intermediate.






The Developmental Switch from Seed Maturation to Seed Germination


Desiccation is the normal terminal event in seed development, leading to a state of metabolic quiescence. In many species (e.g. maize, bean), hydration of the mature, dry seed is sufficient to initiate germination. Thus, in these seeds (termed quiescent or non-dormant seeds), the transition from seed maturation to germination is associated with the reversal of the desiccated state. Seeds of other species (e.g. cereals, Arabidopsis) develop dormancy during late stages of seed development. In these species, freshly harvested mature seeds do not germinate following imbibition but require a treatment such as light, low temperature or afterripening (dry storage) to overcome the state of dormancy and allow induction of germination.







32

Hence, dormant seeds execute the developmental switch to germination during the imposed dormancy-breaking treatment.

As described earlier, the transition from seed development to germination Is associated with major changes in gene expression. It is generally accepted that in quiescent seeds, maturation-related genes cease expression once the water content falls below a level permitting transcriptional activity, and following imbibition, expression of a new set of genes is initiated which is specific to the germinating seed. What gene expression programs are executed in imbibed, dormant seeds is less clear. However, there is evidence from studies in wheat pointing to a maintenance of maturation-specific gene expression during the state of dormancy (Ried and Walker-Simmons, 1990, 1993; Morris et al., 1991).

Most importantly, maturation and germination programs appear to be regulated coordinately in the developing seed. Not only is precocious germination of the immature embryo suppressed, but similarly, the premature induction of germination-related genes appears to be inhibited during this developmental state. Developing seeds of wheat and barley contain biologically active GAs in concentrations adequate to induce a-amylase production in the aleurone layer (Wheeler, 1972; Radley, 1976). Nevertheless, only very low levels of a-amylase enzyme activity or mRNAs were detected in immature seeds (Comford et al., 1986; Garcia-Maya et al., 1990). More compelling, neither (%-amylase activity nor a-amylase gene expression was induced when immature seeds were exposed to exogenous GA (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Similar results were obtained when treating dormant seeds with GA (Schuurink et al., 1992a). Given that immature or dormant embryos excised from the seed and placed in culture are capable of responding to GA, these data are strong evidence for active repression of the GA-response in developing or dormant seeds. Results consistent with this idea were also reported for dicot seeds (soybean, castor bean). In these species, immature seeds contained enzymes involved in the degradation of fatty acids and proteins (malate synthetase, LeuNase, isocitrate lyase) at a much lower level than germinating seeds or isolated embryos in culture (Kermode, 1990).







33

The mechanism underlying the developmental switch from seed maturation to seed germination, precisely the "tuming off' of maturation-related genes and the de-repression/ induction of germination-specific genes, Is only poorly understood. Evidence on the molecular nature of this switch is reviewed in the following for quiescent and dormant seeds.



Quiescent Seeds


Desiccation and subsequent rehydration of the seed appears to be the normal trigger to switch the developmental program from maturation to germination (Comai and Harada, 1990). Even when applied prematurely, drying resulted In the termination of maturation-related gene expression (Oliver et al., 1993) and, upon Imbibition of the dry seed, the Induction of genes specifically associated with germination (Kermode, 1990). Drying altered the developmental potential of seeds such that a-amylase production became sensitive to GA (Evans et al., 1975; Nicholls, 1979; Armstrong et al., 1982; Comford et al., 1986; Oishi and Bewley, 1990).

The nature of this switch in GA-sensitivity remains elusive. King (1976) has postulated that accumulation of ABA in the developing seed prevents precocious induction of hydrolase gene expression in the developing aleurone. Indeed, drying has been shown to cause a concomitant decline in grain ABA content (McWha, 1975; King, 1976; Oishi and Bewley, 1990). Moreover, incubation of immature grains in buffer which caused a drop in endogenous ABA to undetectable levels evoked GA-responsiveness of the aleurone (Napier et al., 1989). Hence, depletion of endogenous ABA, by drying or washing, may be responsible for the induction of GAresponsiveness. This is consistent with the finding that in maize mutants that are either deficient for embryonic ABA (vp5) or insensitive to ABA (vp1), c-amylase activity was induced late in seed development (Wilson et al., 1973).

However, results from Oishi and Bewley (1990) indicate that induction of z-amylase synthesis as a result of drying is not solely due to a reduction in ABA content. The authors compared the responses of maize kernels to premature drying and treatment with an ABA biosynthesis inhibitor (flouridone) which reduces ABA contents in the seed to a similar extent as







34

drying of immature seeds and elicits precocious germination of Immature maize kernels. If drying merely served to deplete endogenous ABA In developing seeds, then flourdone-treated kernels and dried seeds should behave similarly with respect to GA-response. However, while drying resulted in synthesis of high levels of a-amylase following Imbibition, flouridone-treated seeds produced only very low amounts of c-amylase in response to GA. Hence, drying may achieve two effects: 1) It frees seed tissues of the Inhibitory effect of ABA, and 2) it renders the aleurone competent of responding to GA. The cause of the ABA-independent GA-insensitivity in immature seeds is thus far unknown.



Dormant Seeds


Cereals

Imposed dormancy in cereal species is normally released by prolonged storage of dry seeds (afterripening). The duration of seed dormancy following seed maturity depends on a variety of factors such as the genetic constitution (cultivar), the environmental conditions during grain maturation (low temperatures and short day length increase dormancy; Schuurink et al., 1992b) and the rehydration temperature (high temperatures enhance dormancy; George, 1967). Such differences in the depth of seed dormancy have been utilized to investigate the roles of ABA concentration and ABA-sensitivity in preventing embryo germination. No clear correlation between ABA content in the mature embryo and the degree of dormancy was found (WalkerSimmons, 1987, 1988; Morris et al., 1989; Skadsen, 1993). However, there are large differences between dormant and non-dormant embryos with respect to sensitivity to ABA, as measured by the capacity of ABA to inhibit germination. Isolated embryos of a non-dormant wheat cultivar lost their sensitivity to ABA in culture as the grain entered maturation stage, whereas those of a dormant cultivar retained sensitivity to ABA beyond grain maturity (Walker-Simmons, 1987). Similarly, elevating the incubation temperature from 150C to 300C, thus inducing hightemperature dormancy, significantly enhanced the ability of ABA to block germination of isolated







35

wheat embryos (Walker-Simmons, 1988). This differential inhibitory effect of ABA depending on the degree of dormancy was also observed using intact, mature seeds (Morris at al., 1989).

In conclusion, depth of dormancy appears to be positively correlated with ABA-sensitivity with respect to inhibition of germination. Why, and if, enhanced ABA-sensitivity is the immediate cause for inhibition of germination is unclear. Because ABA has been shown to inhibit water uptake by the embryo (Schopfer and Plachy, 1984), it has been suggested that high sensitivity to ABA in dormant seeds may result in reduced water uptake in the embryo and thereby prevent radicle emergence (Walker-Simmons, 1987). Additionally, ABA may have a differential effect on gene expression in dormant and non-dormant seeds. Indeed, transcript levels of a variety of ABA-responsive genes remained high in imbibed dormant wheat seeds, whereas they declined rapidly in non-dormant seeds following imbibition (Morris et al., 1991). Similarly, maturationrelated LEA proteins were abundant in rehydrated dormant seeds but not in non-dormant seeds (Ried and Walker-Simmons, 1990, 1993). However, most of the identified ABA-responsive proteins that accumulate specifically in dormant seeds are predicted to be dehydrins and may therefore function primarily in maintaining the embryo in a desiccation-tolerant state rather than in directly inhibiting germination.

Exogenous application of ABA is known to inhibit GA-mediated activation of hydrolase genes in the aleurone of germinating cereal seeds (Jacobsen and Chandler, 1987). In this context, the following observation may be important. Imbibed dormant barley seeds showed very low expression of a-amylase genes as compared to non-dormant seeds (Morris et al., 1991; Schuurink et al., 1992a; Skadsen, 1993). Moreover, dormant grains produced less c-amylase in response to GA than non-dormant grains (Schuurink et al., 1992a; Skadsen, 1993). Hence, seed dormancy appears to be correlated with a reduced responsiveness of the aleurone to GA. Experiments with isolated aleurone layers indicated that the reduced GA-sensitivity of aleurone cells of dormant barley seeds is dependent on the presence of the starchy endosperm (Schuurink, 1992a; Skadsen, 1993), implying that the starchy endosperm may liberate an inhibitory factor that diffuses to the aleurone cells. It is suggestive that this putative diffusible




. n f.


36

factor may be ABA stored in the dry seed. However, since the dormant and non-dormant seeds used in one experiment contained similar concentrations of ABA (Skadsen, 1993), it may be that the higher ABA-sensitivity in dormant seeds relative to non-dormant seeds plays a role in inhibiting GA-response in the aleurone. However, It cannot be excluded that a factor(s) other than ABA inhibits the GA-response in the aleurone of dormant seeds. zx-amylase genes are known to be sensitive to repression by soluble carbohydrates (Yu at al., 1991; Karrer and Rodriguez, 1992). Possibly, aleurone cells of dormant seeds display a higher sensitivity to the inhibitory effect of soluble sugars supplied by the starchy endosperm. Alternatively, the inhibitory factor may be synthesized in the aleurone cells themselves and the presence of the starchy endosperm is only required to provide an environment of high osmolality which may be essential to maintain production of the putative inhibitor of GA-response in the aleurone cells. Are bidopsis

Dormant seeds of Arabidopsis require either several months of dry storage or rehydration followed by exposure to low temperatures and light in order to break dormancy and induce germination. Analyses of mutants has demonstrated that initiation of dormancy during late seed development involves the action of ABA. Even in light, wild-type seeds are normally incapable of germinating during the seed maturation phase and, for a period of time, after reaching seed maturity. In contrast, seeds of the ABA-deficient mutant aba gradually acquire germination capacity during seed development and at maturity germinate at a frequency of 100% in light and 30% in darkness (Karssen et al., 1983; Karssen and Lacka, 1985). Hence aba mutant seeds display highly reduced dormancy. A germination behavior similar to aba mutant seeds was observed for the ABA-insensitive mutants abil, abi2 and abi3 (Koomeef, 1984), suggesting that these mutants are non-dormant due to a failure to respond to ABA. However, as mentioned earlier, strong alleles of abi3 do not only cause lack of seed dormancy but also vivipary, whereas seeds carrying strong alleles of aba have thus far not been shown to be viviparous. Thus, an ABA-independent function of AB13 cannot be ruled out.







37

While ABA is dearly Involved in the Induction of seed dormancy, Its role In the maintenance of a dormant state beyond seed maturity Is less clear. Late In seed development, ABA concentrations decline rapidly to a very low amount present in the dry seed. This amount has been considered insufficient to inhibit germination (Karssen et al., 1983). Also, aba mutant and wild-type seeds were equally sensitive to applied ABA (Koomeef et al., 1982, 1984), indicating that the difference in germination capacity between aba mutant and wild-type seeds Is not due to a difference in ABA-sensitivity. It was therefore thought unlikely that ABA or ABAsensitivity are involved in maintaining the state of dormancy (Karssen and Lacka, 1985). Instead, Karssen and Lacka (1985) proposed that the maintenance of dormancy is, at least In part, mediated by an insensitivity of the seed to GA. This was concluded from evidence showing that a gradual relief of dormancy by afterripening, cold or light treatments was correlated with an increased sensitivity of the seed to the germination-promoting effect of applied GA (Karssen and Lacka, 1985; Derkx and Karssen, 1993). GA is normally absolutely required for induction of seed germination, as evident from the fact that seeds of the GA-deficient mutant ga-1 do not germinate under any condition, unless GA is supplied exogenously (Koomeef and van der Veen, 1980). Consequently, insensitivity to GA may present a strict measure to inhibit seed germination. Consistent with this hypothesis, reduced germination frequencies were reported for the partially GA-insensitive mutant Gai (Koomeef et al., 1985).

In conclusion, the present evidence implies that GA and ABA do not normally interact directly at any stage of seed development. ABA in concert with high ABA-sensitivity appears to be responsible for the induction of dormancy during seed development, and GA in concert with GA-sensitivity induced by dormancy-breaking treatments appears to stimulate germination. Nevertheless, reduction in seed dormancy as a result of low concentrations of ABA (aba) or insensitivity to ABA (abil, abi2, abi3) partially relieved the mature seed form a need for GA to induce germination. Seeds of double mutants between ga-1 and aba, abil, ab12 or abi3, respectively, were capable of germinating (Karssen et al., 1983; Koomeef et al., 1984, Nambara et al., 1992), whereas ga-1 single mutants have, thus far, never been shown to germinate without







38

application of GA (Hilhorst and Karssen, 1992). Since mature seeds contain very low concentrations of ABA, it was not considered likely that the ge-1 single mutant required GA to directly oppose the action of endogenous ABA present In the seed. In contrast, a remote control model was suggested in which the GA requirement for germination depends on the depth of dormancy induced during seed development. Deeply dormant seeds, as wild-type seeds, have a high GA-requirement to promote germination, while seeds with little dormancy (ABA mutants) have a low GA-requirement which may be satisfied by low concentrations of GA present in the possibly leaky ge-1 mutant.















MATERIALS AND METHODS


Plant Material


Except for immature dl mutant kernels, which were obtained from greenhouse-grown plants, all maize developing ears were harvested from field-grown plants. Under the local environmental conditions, kernels typically begin accumulation of anthocyanins at day 17 postpollination and reach seed maturity after ca. 30-33 DAP. The wild-type maize stock used In this study was a color-converted W22 inbred line carrying all factors required for anthocyanin pigmentation of the aleurone. The vpl-R allele (Robertson 1955) segregated In a colorconverted W22 inbred line carrying all other factors required for anthocyanin pigmentation of the aleurone. This line is routinely maintained by selfing. The vpl-m2 allele (originally named vplmum2, McCarty et al., 1989b) arose in Robertson's Mutator transposable element stocks (Robertson 1978), but was confirmed to carry an Mpi transposable element insertion (D.R. McCarty, unpublished results). Seed segregating for the vp5 mutation was obtained from the Maize Genetics Corporation Stock Center (University of Illinois, Urbana-Champaign). To produce vpl,vp5 double mutant seeds, heterozygous vp5 mutant plants were crossed with heterozygous vpl-R mutant plants. A mutation conferring embryo abortion at early globular stage ("germless") arose in a Robertson's Mutator-induced mutant screen (D.R. McCarty et al., unpublished results). Germless mutant seed were backcrossed into W22 background for at least two generations.

vpl-non-concordant seed was generated using a TB translocation stock (Fig. 2). TB3La seed carrying a BA-translocation on the long arm of chromosome 3, the location of the Vp1 gene, was obtained from the stock center. This seed contains an extra, normally heterochromatic chromosome, called B-chromosome, in addition to the normal set of A


39




1:1114 T


40


Microspore Sperm Nuclei


Heterochromatin







CDt


vp vp


+


I+ vp


Genotype of Macrospore (Female Parent): fvp


Fig. 2. Generation of vpl-non-concordant seed using a TB3La translocation stock.


vp






41


chromosomes. Due to a translocation event between A-chromosome 3 and the B-chromosome, the 3La part of the A-chromosome 3 Is carried by the B-chromosome, while heterochromatic DNA is found on the 3La part of the A-chromosome 3 (Fig. 2). The TB3La stock was crossed to vpl-R at least once. Hence, resulting TB3La, AvP1-R+ B+ plants carry one 3A-chromosome segregating for the vpl mutation, the homologous 3A chromosome with heterochromatic DNA (thus confering a vpl mutant phenotype) and a B-chromosome carrying the wild-type Vpl gene. To obtain vpl non-concordant kernels, pollen from TB3La, Av1-R* B+ plants Is crossed onto segregating vpl-R females. During the second pollen mitosis, the replicated B-chromosomes undergo non-disjunction forming one sperm nucleus with two B-chromosomes and one sperm nucleus without a B-chromosome (Fig. 2). Hence, following double fertilization, non-concordant seeds are produced carrying a vpl mutant embryo and a wild-type endosperm or vice versa (Fig. 2).


For experiments with germinating wild-type seeds of maize, seeds of the variety NK508 were used (kindly provided by Northrup-King).

Wild-type barley seeds c.v. Himalaya were obtained from Washington State University, Pullman, WA (harvests 1988, 1991 and 1992). Seed segregating for the slender mutation (Himalaya background) was kindly provided by P. Chandler. As with seed segregating for D8, wild-type and slender mutant endosperms were genotyped by germination of the excised embryo.



Plasmid Constructs


In all experiments, JR254 (Amy-GUS) or JR303 were used as reporter constructs (see Fig. 3). Amy-GUS and JR303 were kindly provided by J. Rogers and T.H.D. Ho, respectively. Amy-GUS contains ca. 1,800 bp of the 5' flanking sequence of a barley high pi a-amylase gene (Amy6-4; Kursheed and Rogers, 1988), the first intron of Arny6-4, the GUS reporter gene and the Amy6-4 3' terminator. JR303, containing a low-pl z-amylase promoter, was derived from






42


Reporter Plasmids:


Amy 6-4 inon 1


a


High-pi Amy-GUS
(JR254)


Low-pi Amy-GUS2
(JR303)


GUS


Amy 0-4 3


Amy 6-4 promoter


GUS


CM-M I*
Amy 32 poxtr


35S-Sh-VP1


Shi Intron


CaMV 35S promoter


Shi Intron


35S-Sh-CAT
(No-Vp1 Control)


CaMV 35S promoter


Internal Standard:


Ubiquitin-Luciferase
(pAHC18)


Ubiqitin promoter


Fig. 3. Plasmid constructs.


MW 3


Effector Plasmids:


Vpl cDNA


NOS


CAT


NOS 3


intron


Lux


NOSY






43


Amy32b (Lanahan et al., 1992). For its structure, see Fig 3. As effector construct, 35S-Sh-VPI was used (McCarty et al., 1991). For no-VP1 control treatments, 35S-Sh-CAT (Vasil et al.. 1989) was added instead of 35S-Sh-VP1 to maintain a constant amount of total DNA and 35S promoter in the bombardment mixtures. To normalize for transformation efficiency, a Ubiquitin-Luciferase construct (Ubi-LUC; Bruce et al., 1989) was included into each bombardment mixture. Hence, expression data are presented as Amy-GUS / Ubi-LUC or JR303 / Ubi-LUC ratios. It was confirmed that co-expression of 35S-Sh-VP1 has no effect on expression of Ubi-LUC.

Construction of plasmids carrying an activation domain-deletion derivative of VP1 or a replacement with Herpes simplex virus transcription factor VP16 activation domain was described in McCarty et al. (1991). Intemal deletion constructs were made by introduction of two Ncol restriction sites and subsequent deletion of the insert and religation of the backbone (Fig. 4). Ncol sites were Introduced by site-directed mutagenesis using the Altered Sites in vru Mutagenesis System from Promega. Briefly, mutant oligonucleotides and an ampicillin repair oligonucleotide which restores the function of a defective ampicillin resistance gene in the phagemid provided were annealed to single-stranded DNA template. Following DNA synthesis and ligation, the resulting double-stranded phagemid was transformed into a repair-deficient strain of E.coli which is subsequently grown in Ampicillin-containing liquid medium for selection. From the obtained bacterial suspension, plasmid DNA was isolated and transformed into an E. coli strain conventionally used for transformations (JM83). Colonies growing on Ampicillin were tested for the presence of the desired mutation by restriction enzyme digestion.

The constructs 86/87, 86/85, 85/87, 87/88, 93/95, 103/104, 101/100 and VP1-McW were made available by L. Rosenkrans and D.R. McCarty.



Particle Bombardment and Transient Expression Tissue preparation

Maize developing ears or dry, mature seeds were surface sterilized in 70% ethanol for I min followed by 0.525% NaOO for 10 min. Dry seeds were germinated in a solution containing








44


P a K G G D x G G A P A D IF N 7 9 I a G II


T.CWTCCACTCCACCTTCOTCCAACTCCTTCAeCTCCIeaemACT S P S S S T F S S N S S S N S S S A Y T N T






GATCGCCGAGACTCG1CW3CTO0ACT -------- MCCCeCTGACA D A L D D I D Q L L D F A S L S N P W D S 9



SV G D G M 9 1 1 K A V P 1 G T T G G 9 1 A



T S N R I N I S A 2 D L R G I R L R I S T I

83: CCLTOG 07: CATGG
VGCTCAT cTm
K L I L T N V 0 N H N L Q It K R P R D V N 9

32: CCATOG

Y S F P A G G Q D N A A G G G T S W N P N 0


A



P



C





K
m.


A



G IF



N



A






Ae
A


R A P 0 D A A A A G


wT0Te TUAAC CCAmWTTC~ cV w


A G 0 Q Y S F N Q G P S


CAGTACTGCCC CTTCC M
Q Y V P V P P P G A S T

104: CCWGG
CCCCWACCATGGCAGGCGTGGAGGCCT M P Q R M A G V E A S A T


CAGATCCAeCCGTCCACCTGCAGGAGCCGTC Q IQ T S V H L Q I P S


I



I



I


T 5 S V V V N S Q P 1 3 P P


CTACCCTATOCCGCACMCTTCCeCCCATTCO G S Y P N P Q P F S P G G




e~a~aaCMCRGAGCAGGACAcWTCC K E A A X K R S A R Q R R L CCCOGTCCACGCACTCCGGCCCGGTCACGorTCGA P I S T H S G P V T P S A G


C L SG












Q L LrCCACCJ
P r


9 D T F P S L P D F P



0 G 9 P S I P A S A G W: CCATG

V S N K L I N A N S A



S I G 9 9 L P R F F N



A R L 0 6 G R 0 0 T N



L R V 9 L P S P V A N

"s: CCLIM lft:CCATOO P P A A Y G G D A V Y

93: CCASM T
1"6: CCATOO

P V G D N H G A N N A AILGG XQGT




91: CCAT



G Q Y A G A G A G H L 103: C:T

3 C L Q 0 0 R S Q Q L


CGGCTGGATTCTGAGCMAGCAGCCAGC G W G 7 W S P S S Q Q


AACCTCTCCAAGTCCAATT~CTAGGeCeCCGMCCT COAGCGGCGMCC CAGarAAAGCCeCCGCrGGTGrCGGCAGACATCACCACeCrC
N P L S K S N S S R A P P S S L E A A A A A P 0 T K P A P A G A R Q D D I H H R


CTeCCAG~cGTTCAGATAAe AGecA~eGcG~c~cTCeTGCTGCAAGTGCGAAGCAGAG~r.ATWC e0CecCCATcGCTCCvCAAA
L A A A S D K R 0 G A K A D K N L R F L L Q K V L K Q S D V G S L G R I V L P K

101: CCAI.
AAGGAAGCGGAGGTTCACCTGCCGGAGCTGAAGAmGAGOATGGCALTCTCCATCCCCATGGAGGACATCOGAACGC GGGAAAGGT K E A 9 V H L P I L K T R D G I S I P N I D I G T S R V W N N R Y R F W P N N K

100: CCATOG
AGCAGARTGTATCTGCTGGAAAACACAGGGGAATTTGTTCGTTCCAACGAGCTTCAGGAGmonATTTCATAGTGATCTACTCCGATGCAAGTCGGGCAAATATCTGATArvOGGCGTG S R M Y L L S N T G I F V R S N I L Q 9 G D 7 I V I Y S D V K S G K Y L I R G V


AAGGTA CC-NGCOnGGAGGCAAGCACAGGCCCCTCTGTCCAGCAGGTCC CCTGAAGAMCC


K V R P P P A Q 9 9 G S G S S G G G K H R P L C P A G P 9 R A A A A G A P 9 D A GTCATCGAQ KQ-CGCGGGCCTGCASAGnGSTCTCCGGAAGGCGTGGGOGTTAA AGA CTGAGAGCAGCAGG
V V D G V S G A C K G R S P I G V R R V R Q Q G A G A M S Q N A V S 1


CTCCCCATATATTGATCGAT ACCAATCMTCTTAGTTCTCCAWTTACTATTAGCTAGATGCCGTCGTGTGTGTGC


TAAGCATGTAG.CGTGCTAGGAGATGATATATTAAATATAATmrAGTAGTArhr-*CTACCCrCTGTGTr.&CrTAATTT~c1 m.Tan:.~mo


CTGT0TCATGTCWTGOTA2TAGCTATACTATCCTTCTTAGA. P1khflplhlkkkppkpllfkppt 2498











Fig. 4. Location of Ncol (C/CATGG) restriction sites introduced by site-directed mutagenesis

(adapted from McCarty et al., 1991).


120
1s


S












L



L



G



A


N 9 A S S a S S P P II S Q a II P


240 56 340
9'
tso 136 600 176 720
216 640
256 960
296



1080 336


1200 376


1320 416 1440 456 1560
496 1680 536


1800 576


1920 616


2040 656 2160 691


2280 2400


4


w P 0 I: MAT1 CAGTU0CC
S V A



S L G


GCAGGTCCAG
0 V Q


GCAGrvCCAACAGTACTCTTT






45


MS salts and MS vitamins (Sigma, cat# M5519) on a gyratory shaker in the dark for ca. 36 hrs, while developing seeds were used immediately. The embryo as well as pericarp and testa tissues were removed from the seeds to expose the aleurone layer of the endosperm. Prepared endosperms were placed on Geirite-solidified salt medium and then subjected to particle bombardment.

Barley seeds were de-embryonated prior to surface sterilization in 70% ethanol for 1 min followed by 10 min in 1.75 2.9 % NaOCL. A minimum of 1.75% NaOCI (z30% Clorox) was necessary to allow easy removal of pericarp and testa layers prior to bombardment. Sterilized half-seeds were imbibed ovemight and prepared for particle bombardment as described above for maize seeds.


Particle Bombardment

Particle bombardment was performed as described in Taylor and Vasil (1991) using a DuPont PDS-1000 particle gun. Briefly, 35 to 50 pl of a sterile 50 mg/mi gold stock solution (Biorad, 1.0 or 1.6 pm particle diameter; prepared in water) was mixed with premixed plasmid DNA in a 250 pl-Eppendorf tube and vortexed on maximum speed for 10 s. Immediately, the tube was shifted sideways and 10 pl of 0.1 M spermidine (free base) and 25 pI of 2.5 M CaC2 were placed onto the side of the tube without allowing it to mix with the gold/DNA solution. Then, the tube was placed upright and subjected to vortexing for 10 s. The precipitated gold/DNA particles were allowed to settle for ca. 3 min, after which part of the supernatant was removed leaving 35-45 pi of liquid behind. The tubes were placed on ice until further use in particle bombardment. For particle bombardment, 2 pi of sonicated gold/DNA solution (containing ca. 80 pg of gold) were used for individual shots.


The bombardment procedure had frequently to be adjusted to the gold characteristics which varied substantially from batch to batch. Modifying the amount of gold used per shot was found most successful in improving bombardment efficiency.


--T1






46


Incubation and Extraction of Endosoerms Followlna Bombardment

Following bombardment, I ml of a solution containing MS salts and MS vitamins supplemented with no hormones, 104 M GA3 or 10-6 M GA3 and 104 M (or 10 M) ABA was dripped over the endosperms. After 24 h of Incubation In darkness, maize endosperms were ground either individually (in experiments using developing seeds) or in bulk from each bombardment (when germinating seeds were used) with mortar and pestle aided by the addition of silicon carbide powder in 200-1,000 pl of extraction buffer (0.1 M potassium phosphate (pH 7.8), 2 mM EDTA (pH 8), 2 mM DTT, 5% glycerol). The homogenates were centrifuged to recover the cell extract To obtain barley aleurone extract, the aleurone layers were separated from the endosperms and ground In bulk for each replicate in 200 pi of extraction buffer. The homogenates had to undergo two rounds of centrifugation to obtain clear cell extract.


Quantification of Transient Expression

Quantitative measurement of GUS activities was performed as described in Jefferson et al. (1987) with the modification that the substrate MUG was dissolved in the extraction buffer described above. For determination of luciferase activities, 10 i aliquots of the extract and 200 pd of reaction buffer (25 mM tricine (pH 7.8), 15 mM MgCI2, 5 mM ATP, 0.05% BSA) were placed in cuvettes and immediately assayed using a Monolight 2010 luminometer. The luminometer automatically injects 100 ;d of 1 mM luciferin and then counts the emitted photons for 15 s. The unit of measurement is the Relative Light Unit (RLU).















RESULTS


Repression of Hydrolase Genes by VP1 In Aleurones of Developina Maize Seeds


Phenotvoic analysis of vyl-m2 kernels

The vpl-m2 allele of Vpl carries a transposon insertion in the third intron which causes somatic Instability of the gene during endosperm development (McCarty et al., 1989b). As a result, mosaic kernels develop with clonal vpl mutant and wild-type sectors. In these kernels, a striking pattern of endosperm remobilization is often evident. Endosperm tissue underlying vpl mutant aleurone cells is frequently softened and depressed in surface while wild-type sectors are raised relative to adjacent mutant sectors. This produces kemels with a distinctive etched appearance (Fig. Sa). The softening response was also observed when only a small fraction of the endosperm was comprised of mutant tissue (Fig. 5b), indicating that expression of this phenotype is cell autonomous. The softening of starchy endosperm tissues that underlie islands of vpl mutant aleurone cells appears to be attributed to precocious induction and secretion of hydrolytic enzymes caused by the loss of VP1 function. Thus, repression of hydrolases in developing maize kernels is evidently dependent on the presence of functional VP1.


Transient expression of Amy-GUS in maize aleurone

In order to more directly address the role of VP1 in repressing hydrolase activity, a quantitative transient gene expression assay was developed that is based on particle bombardment of aleurone tissue with a barley high-pi a.-amylase promoter-GUS fusion construct (Amy-GUS). It was first interesting to determine whether the observed differential activity of hydrolases in vpl mutant and wild-type sectors was due to transcriptional control. For this purpose, Amy-GUS was introduced into aleurone cells of developing vpl mutant and


47



































14


Fig. 5. Cell autonomous de-repression of the aleurone germination response in vpl-m2 mutant aleurone sectors. (A) The kernel shown is a mosaic: regions pigmented with purple anthocyanin are wild-type; yellow, anthocyanin-deficient regions are clonal sectors of aleurone that have lost Vpl function. (B) Magnification of a vpl-m2 kernel.






49


wild-type maize seeds. Table I shows that during mid-late development Amy-GUS was not expressed in developing wild-type aleurone, even In the presence of exogenous GA. In contrast, GA-induction of Amy-GUS was detected In vpl mutant aleurones as early as 20 days after pollination (DAP). These data indicate that in developing wild-type aleurone tissue c-amylase genes are insensitive to GA while In vpl mutant aleurone cells a-amylase expression Is transcriptionally de-repressed.

Amy-GUS expression in vpl mutant aleurone was found to be under developmental and hormonal control. Prior to approx. 18 DAP, Amy-GUS was inactive in GA-treated as well as untreated aleurone, Indicating that ealy in seed development the aleurone Is unresponsive to GA even in the absence of VP1 protein. At 20 DAP, Amy-GUS was Induced by exogenous GA, whereas its activity remained low in untreated aleurone (Fig. 6). Late in seed development (24 DAP), Amy-GUS was constitutively active in the absence of GA, indicating a greatly reduced dependence on exogenous hormone. Two observations, however, Indicate that Amy-GUS expression was not fully constitutive at this stage: 1) GA treatment significantly enhanced AMYGUS expression (as much as 3-fold over that of untreated aleurones) in some, but not all experiments (Table 1, Fig. 6). 2) GA treated aleurones typically exhibited less quantitative variation in Amy-GUS expression than non-treated sibling aleurones. The latter effect suggests that developmental or spatial variation affecting endogenous hormone concentrations within the ear or seed might contribute to the large variation observed in the absence of exogenous GA.

The differential expression of Amy-GUS in developing vpl mutant and wild-type aleurone cells confirms a role of VP1 in the repression of a-amylase genes during seed development. In order to test whether expression of recombinant VP1 could evoke inhibition of a-amylase transcription in vpl mutant aleurones, aleurones were bombarded with a mixture of Amy-GUS and 35S-Sh-VP1 plasmids. Co-expression of VP1 strongly inhibited Amy-GUS expression in vpl mutant aleurone in the presence as well as absence of exogenous GA (Fig. 6), indicating that recombinant VP1 effectively restored the wild-type phenotype. We can rule out the possibility that over-expression of VP1 causes non-specific squelching of general





50


Amy-GUS / LUC *104 [pmoles MU/h/RLU] vpl-R mutant Aleurones Wild-type Aleurones

Days after -GA +GA -GA +GA
Pollination Range Mean S.E.M. Range Mean S.E.M. Mean Mean

18 <1 <1 <1 <1

20 <1 4-35 21 *7 <1 <1

24 114-531 263 130 41-150 110 35 <1 <1


Table 1. Amy-GUS is inducible in vpl-R mutant aleurone cells but not in wild-type aleurone cells. Aleurones of developing vpl-R mutant and wild-type kernels at 18, 20 and 24 DAP were bombarded with a mixture of 10 pg of Amy-GUS and 5 pg of Ubi-LUC. Post-bombardment, kernels were treated with a solution containing no hormones or 10-6 M GA3. Data represent mean ( S.E.M) of three to five replicates.






51


a


no hormones




+GA




+GA +ABA


0


75

1


247 F3


11

0.3
I I I I p


50


100 150 200
Amy-GUS / LUC *10 4 Ipmoles MU/h/RLU]


250


[i-V1


300


Fig. 6. Effect of VP1 over-expression and ABA on Amy-GUS expression in vpl-R mutant aleurone. vpl-R mutant aleurones from kernels harvested 26 DAP were bombarded with 10 pg of Amy-GUS, 5 gg of Ubi-LUC, and 5 gg of 35S-Sh-VP1 or 35S-Sh-CAT (for no-VP1 controls). Following bombardment, a solution containing no hormones, 10-6 M GA3 or 10-6 M GA3 and 104 M ABA was applied to the kernels. Numbers behind bars represent means of five replicates. Error bars show S.E.M.






52


transcuiption factors because no inhibitory effect of co-bombarded VP1 on 35S-Sh-GUS or Ubiquitin-Luciferase expression was observed (data not shown). Moreover, VP1 caused fransactivation of positively regulated reporter constructs, Em-GUS and C1-Sh-GUS, In aleurone cells using similar bombardment conditions (S. Cocciolone and D.R. McCarty, unpublished results).



Interaction between VP1 and Abscisic Acid


In concert with VP1, the hormone ABA plays an important role during seed maturation (McCarty and Carson. 1991). Moreover, ABA functions as an inhibitor of a-amylase expression in germinating cereal seeds (Jacobsen and Chandler, 1987). This suggests that ABA might also be involved in repression of c%-amylase genes in the developing seed. Therefore, possible interactions between ABA and VP1 in repressing Amy-GUS were analyzed.

ABA was effective In blocking Amy-GUS expression in vpl mutant aleurone (Fig. 6). This indicates that repression by ABA does not require VP1. In combination, ABA and VP1 overexpression produced a roughly additive effect (Fig. 6).

To test whether z-amylase repression by VP1 is dependent on ABA, recombinant VP1 was over-expressed in aleurone of developing vpl,vp5 double mutant kernels that are deficient for ABA biosynthesis. Figure 7 shows that VP1 was highly effective in repressing Amy-GUS in vp5 mutant background. While it cannot by ruled out that matemal ABA derived from the vp5/+ parent plant may be sufficient for VP1 function, it is suggested that VPI-mediated repression of Amy-GUS expression does not require ABA. This would be consistent with the finding that VP1 also functions in aleurone of germinating seeds (see below) where ABA levels are very low (Oishi and Bewley, 1990). Taken together, these data suggest that ABA and VP1 inhibit AmyGUS expression independently.






53


in


-a


a
.3


w
0




I


7'



5


2


5



0






n


-GA


+GA


+ 35S-Sh-CAT M + 35S-Sh-VP1


Fig. 7. Co-expressed VP1 inhibited Amy-GUS in aleurone of developing vpl/vp5 double mutant seeds that are deficient for ABA biosynthesis. Kemels were harvested 24 DAP. Bombardments were performed as described in Fig. 6. Following bombardment, kernels were incubated In no hormones or 10-6 M GA3. Data represent mean ( S.E.M) of 7-8 replicates.


Effector Construct Amylase-GUS / Ubi-LUC 104 (*S.E.M.)
[pmoles MU/hr/RLU]

Maize Seeds Barley Seeds

Amy-GUS JR303 Amy-GUS JR303


35S-Sh-CAT (Control) 91 (*17) 1.19 (0_3M 247 (3 65 (45.2)


35S-Sh-Vpl 10.5 (-6) 0.07 (10.014) 70 (16) 25 (64)




Fig. 8. Co-expressed VP1 inhibited Amy-GUS and JR303 in aleurone of germinating maize and barley seeds. Aleurones of imbibed seeds were bombarded with Amy-GUS (maize: 4 pg; barley: 2 pg) or JR303 (maize: 10 pg; barley: 5 pg), 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35SSh-CAT. Post-bombardment, kernels were incubated in 10-6 M GA3. Data represent mean ( S.E.M) of 3-5 replicates.






54


Over-expression of VP1 in Aleurones of Germinatina Maize and Barlev Seeds


Endogenous expression of Vp1 In embryo and aleurone tissues is under strict developmental control. Vp1 mRNA peaks at 16 DAP and then gradually decreases as the seed reaches maturity (McCarty et al., 1991). Germinating seeds, in contrast, display no Vp1 expression or detectable levels of VP1 protein (Carson, 1992). Thus, VPI function in maize Is limited to the maturing seed. To test whether VP1 can function in germinating seeds in a way equivalent to maturing seeds, we co-expressed 35S-Sh-VP1 and Amy-GUS In aleurones of germinating wild-type seeds of maize. In the presence of exogenous GA. VP1 reduced AmyGUS expression by ca. 90% (Fig. 8). Thus, VP1 also functions in germinating seeds, apparently without the need for additional developmental factors. Furthermore, VP1 also repressed expression directed by the barley low-pl a-amylase promoter (JR303) which shows considerable sequence divergence from Amy-GUS. This indicates that expression of high- as well as low-pi L

-amylase genes Is under control of VP1.

Because germination-specific responses are well characterized in barley, VPI-mediated repression was tested in aleurones of germinating barley seeds. Though not as effective as in maize, VP1 also inhibited Amy-GUS and JR303 expression in barley (Fig. 8).

Moreover, variation within as well as between experiments was significantly reduced in aleurones of germinating wild-type seeds of maize and barley as compared to aleurones of developing vpl mutant maize seeds. Therefore bombardment of germinating seeds constitutes a useful experimental system to further characterize VP1 function.

A VP1 dose-response curve was generated to determine the amount of co-expressed VP1 necessary to achieve maximum repression of Amy-GUS in barley aleurone cells. Figure 9 shows that repression was already evident when 1.25 pg of 35S-Sh-VP1 were co-transferred with Amy-GUS, while increasing the amount of 35S-Sh-VPI beyond 2.5 pg did not lead to further repression of Amy-GUS. Hence, comparatively low amonts of recombinant VP1 are sufficient to achieve inhibition of Amy-GUS. To confirm that the inhibitory effect of VP1 on Amy-GUS in barley aleurone is promoter-specific, 35S-Sh-VPI was co-expressed with an Em-GUS reporter






55


I


100


8060


4020-


U*4


0.0


. I
2.5 5 .0


7.5 10.0


Amount of 35S-Sh-VPI added (ug)


Fig. 9: VP1 dose-response for repression of Amy-GUS in aleurone of germinating barley seeds. 0 to 10 pg of 35S-Sh-VP1 were co-precipitated with 0.5 pg of Amy-GUS and 5 pg of Ubi-LUC. Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment, endosperms were cultured in 10-6 M GA3. Data represent mean ( S.E.M) of five replicates.


100-


5.


I
0

iE


755025-


0-'-


T


-r-


U


1


Amount of 35S-Sh-VP1 added (ug)

Fig. 10: Co-expression of VP1 activated Em-GUS in barley aleuronb. Aleurones of germinating barley seeds were bombarded with 2 pg of Em-GUS, 5 pg of Ubi-LUC and 0, 1 or 5 pg of 35SSh-VP1. Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment, endosperms were cultured in no hormones. Data represent mean ( S.E.M) of five replicates.






56


construct containing the full length promoter of the wheat Em gene fused to the GUS coding sequence (Marcotte at al., 1989). Recombinant VP1 Increased expression of Em-GUS by ca. 5fold (Fig. 10) which Is consistent with the function of VP1 as a transcriptional activator of Em (McCarty et al., 1991). In conclusion, these data show that VP1 can repress or activate gene transcription depending on the promoter context.



Interaction between VP1 and Gibberellic Acid


The well characterized hormonal responses in Himalaya barley aleurone facilitated further studies regarding the interaction between VP1 and GA. For this purpose, GA response curves of Amy-GUS expression were determined in aleurones of de-germed Imbibed Himalaya "half seeds" (Fig. 11). In the absence of co-expressed VP1, Amy-GUS expression showed a typical GA-induction. In contrast, when a mixture of Amy-GUS and recombinant VP1 was introduced into aleurone cells, GA-induction of Amy-GUS expression was reduced by ca. 80%. Most noticeably, the clearly detectable basal activity of Amy-GUS was not significantly affected by co-expression of VP1 (Fig. 11, insert). Thus, VP1 only inhibited the GA-dependent activity of the a-amylase promoter. This implies that VP1 may interfere with the GA signalling pathway.

Recessive mutations that cause constitutive GA-response have been identified in barley and a few other species (Ross, 1994). Barley slender (sin) mutant plants are characterized by excessive elongation of stem and leaf tissues and constitutive expression of hydrolytic enzymes in the aleurone of imbibed half seeds in the absence of exogenous GA. The mutant phenotype suggests that the Sin gene encodes a negative regulator that is normally inactivated by GA (Chandler, 1988; Lanahan and Ho, 1988). To test whether VP1 inhibitory function depends on the presence of the SLN protein, aleurones of sin mutant half seeds were co-bombarded with Amy-GUS and recombinant VP1. Figure 12 shows that VP1-mediated repression of Amy-GUS was as effective in sin mutant aleurones as in wild-type aleurones.






57


40 at 0 M GA3 30
20 10

0 VP +VP1


1200 1000800600

400200

0


- 35S-Sh-VP1
--
+ 35S-Sh-VP1


0.


-j
cc (D

E
CL LLI


0




E
<


GA23-Concentration (M)





Fig. 11. Co-expression of VP1 inhibited GA3-induction of Amy-GUS but did not affect its basal activity in aleurone of germinating barley half seeds Aleurones were bombarded with 2 pg of Amy-GUS, 5 ;ig of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT. Following bombardment, 3 replicates of 5 kemels each were incubated in 0-10-5 M GA3. Data represent mean of three replicates ( S.E.M). The insert shows activities in the absence of GA3-


-8 -7 -6
10 10 10


-5
10


0 10


i -






58


I


slender, -GA


WT, +GA


M+ 35S-Sh-CAT M+ 35S-Sh-VP1


Fig. 12. VP1 function does not require the Slender gene product. Aleurones of imbibed slender
(sn) mutant and wild-type barley half seeds harvested from a plant segregating for the sin mutation were bombarded with 5 pLg of Amy-GUS, 5 pg of Ubi-LUC, and 5 ;g of 35S-Sh-VP1 or 35S-Sh-CAT. Following bombardment, sin mutant or wild-type seeds were incubated in a solution containing no hormones or 10-6 M GA3, respectively Data represent mean ( S.E.M.) of 10 replicates.


I


1000 750500250-


-J

I
U

i
w
0
4




C,


n 1 1 ........... .... I. -


...........






59


This Indicates that VP1 Is likely to act further downstream In, or Independently of, the SLN pathway.


While GA is a strong inducer of a-amylase genes in aleurones of Himalaya barley and other cereals, the importance of GA in the regulation of maize ci-amylase genes is less clear. aamylase activities were found high in isolated endosperms that had been de-embryonated prior to imbibition (Harvey and Oaks 1974; Goldstein and Jennings, 1978). Moreover, application of exogenous GA to Isolated endosperms did not further enhance ci-amylase activities (OlshI and Black, 1990). Hence, it was argued that mature endosperms store high concentrations of GA (Harvey and Oaks 1974; Goldstein and Jennings, 1978; Oishi and Black, 1990). However, seeds of the GA-deficient, extremely dwarfed mutant d5 displayed considerable ci-amylase activity that was only 3-fold lower than in wild-type seeds, implying a GA-Independent component in maize ciamylase production. To investigate this, Amy-GUS was introduced into aleurones of germinating wild-type and GA1-deficient dl mutant seeds. Both genotypes displayed similar, high Amy-GUS activities in the absence of exogenous GA (Figs.13, 14). Furthermore, application of GA to dl mutant seeds increased Amy-GUS expression by less than two-fold, thus to a similar extent as in wild-type seeds (Figs.13, 14). Hence, the dl mutation did not appear to alter Amy-GUS expression in the aleurone, suggesting that deficiency in the highly active gibberellin GA1 does not severely affect high-level expression of a-amylase genes. Thus, consistent with the data on the mutant d5 (Harvey and Oaks, 1974), the possibility of a constitutive activity of ci-amylase genes in the absence of GA appears likely. The finding that co-expression of VP1 and application of ABA reduced Amy-GUS activity to a very low level (Figs. 13, 14) indicates that VP1 and ABA repress the GA-dependent as well as the putative constitutive activity of AmyGUS in maize.






60


i of .


5.







0


77~


+ABA


+VP1


+ABA +VP


Fig. 13. Effect of GA, ABA and recombinant VP1 on Amy-GUS expression In aleurone of germinating wild-type (NK508) seeds. Aleurones were bombarded with 4 pg Amy-GUS, 5 pg of Ubi-LUC and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT and then incubated in a solution containing no hormones, 10-6 M GA3, 104 M ABA, or 10-6 M GA3 and 10-5 M ABA. Data represent mean ( S.E.M.) of five replicates.


=3No Hormones
M + GA













..... ....


+ABA +VP


Fig. 14. Effect of GA, ABA and recombinant VP1 germinating dl mutant seeds. Methods as in Fig. 13.


on Amy-GUS expression in aleurone of


= No Hormones
-+ GA


750500250-


0-i


Af)[ .i'


25





w


T


I


200100-


-A
-ABA


0-L-


+ABA


+VP1






61


Role of the Embryo in Reoression of a-Amylase Genes in the Aleurone


The differential response of mutant and wild-type aleurone tissue in developing vp1-m2 kernels (Fig. 5a,b) did not appear to be fully independent of the state of the embryo. Embryos of vpl-mn2 seed are frequently non-viviparous and survive desiccation. It was observed that mosaic aleurones that were associated with non-viviparous embryos very rarely exhibited precocious endosperm remobilization while those with viviparous embryos typically did. Moreover, aleurone near the germinal face and crown of the kernel was more strongly affected than aleurone on the abgerminal face. This phenotype suggests that signalling from the embryo as well as responsiveness of the aleurone cells contribute to the softening response.

In order to assess the impact of the physiological state of the embryo on a-amylase expression in aleurone cells, TB3La translocation stocks were used which allow the generation of vpl non-concordant seeds with embryo and endosperm of different genetic constitution (for a brief description of the system see Materials and Methods). As a result, ears developed that segregated four genotypes: 1) seeds with a vpl mutant embryo and a wild-type endosperm, 2) seeds with a wild-type embryo and a vpl mutant endosperm, 3) concordant vpl mutant seeds and 4) concordant wild-type seeds. Aleurones of these four genotype combinations were bombarded with Amy-GUS and postbombardment cultured in the absence of added hormones. Amy-GUS was only expressed in aleurones of concordant vpl mutant kemels and not in any of the other three genotype combinations (Fig. 15). These data are consistent with the observation made in vpl-m2 seeds that endosperm tissue underlying vpl mutant aleurone was remobilized predominantly in viviparous seeds. Thus, c-amylase genes appear to be de-repressed in aleurone cells that lack functional VP1 predominantly if the embryo is also viviparous. In some experiments in which kernels very late in development were used, both vpl non-concordant genotypes displayed significant Amy-GUS activities, while no Amy-GUS expression was detected in aleurones of concordant wild-type seeds (data not shown). This indicates that partial de-repression of Amy-GUS in aleurone cells is facilitated if either the embryo is viviparous or the aleurone cells lack functional VP1.







62









11. 109.


5.
-







7
r

S 51
3.




2




0

Embryo: vp1 WT vp1 WT

Endosperm: vpl vp1 WT WT







Fig. 15. Amy-GUS expression in aleurone of developing vpl non-concordant maize seeds. Aleurones (31 DAP, fall season) were bombarded with 10 pg of Amy-GUS and 5 pg of Ubi-LUC and cultured post-bombardment in no hormones. Data represent mean ( S.E.M.) of six replicates.






63









60


50 -+


40


30 +


20 +


10 +


0 1
vp1 vp1, gmless gmless WT




Fig. 16. Amy-GUS expression in aleurone of developing germless seeds. Aleurones (29 DAP, fall season) of an ear segregating for the mutations vpl and germiess were bombarded and cultured as described in Fig 15. Crosses represent single data points of six replicates (exception: four replicates in the vpl/germless double mutant).






64


There are at least two possibile scenarios that would explain how the state of the embryo might influence de-repression of z-amylase genes in vpl mutant aleurone: 1) a viviparous embryo might secrete an Inductive signal required for a-amylase expression in the aleurone of developing seeds (e.g. GA), and/or 2) a non-viviparous embryo might contribute a diffusible inhibitory signal that prevents a-amylase expression in the aleurone (e.g. ABA). In order to test these hypotheses, Amy-GUS was introduced into aleurone cells of a germless mutant in which the embryo aborts during the early globular state (P. Becraft and D.R. McCarty, pers. communication). Hence, use of ears that segregate for the germless and vpl mutations allows assessment of aleurone responsiveness in the absence of a signal from the embryo. Amy-GUS was highly de-repressed in vpl-mutant aleurone of germless seeds. Aleurones of all four seeds bombarded expressed Amy-GUS (Fig. 16). This indicates that a viviparous embryo per se Is not required for de-repression of a-amylase genes in vpl mutant aleurone cells. Rather, it appears that the lack of a normal (non-viviparous) embryo caused induction of Amy-GUS in the germless,vpl double mutant. This suggests that a wild-type embryo secretes a signal with inhibitory function on a-amylase expression in the aleurone.

In the single mutants vpl and germless, Amy-GUS was partially de-repressed (2-3 seeds of a total of six bombarded expressed Amy-GUS, Fig. 16). Only wild-type seeds displayed complete repression of Amy-GUS (Fig. 16). Hence, both VP1 expression in the aleurone and a normal embryo appear to be required for complete inhibition of a-amylase genes in the aleurone. This is consistent with the Amy-GUS activities found in vpl non-concordant kemels very late in development, as described above.



Functional Analysis of the VP1 Protein


The acidic activation domain

We considered two models of how VP1 may function in repression of the aleurone germination response. 1) VP1 might be a transcriptional activator of an intermediate repressor gene that in tum inhibits expression of a-amylase genes (Figure 17a). 2) VP1 itself might





65


B.


vP1



REPRESSOR


I
Amytase


VP1
or


ACTI


/
Am ase


vP1



VATOR


Effector Construct


Rel. Amy-GUS / LUC


Maize


Barley


35S-Sh-CAT (Control)


100 (*M 100 (*17)


WT-VP1


n,28-121 0 1


3x (VP16 act)


* *.am om*


Fig. 17. Mode of action of VP1 in repressing Amy-GUS.
(A),(B). Alternative models for VP1 action as described in the text.
(C). Effect of deletion and substitution derivatives of VP1 on Amy GUS expression in aleurone tissue of germinating maize and barley seeds cultured in GA3. In A28-121, the activation domain of VP1 was deleted. In 3x (VP16 act), the activation domain of VP1 was replaced by three copies of the Herpes simplex VP16 activation domain. Data represent activities (mean S.E.M) relative to control (=100). Black boxes show sequence homology between VP1 and AB13. (n.d.: not determined).


A.


C.


activator


13 ( 4)


42 (8) 17 4*)

12 (")


n.d.


tom F.-I !=






66


function as a repressor of the a-amylase genes or of an intermediate gene that is required for activation of the a-amylase promoter (Figure 17b). In order to distinguish between these models, we determined whether the transcriptional activation domain of VP1 which is essential for activation of the Em and C1 genes in maize cells is also required for inhibition of z-amylase. Figure 17c shows that a deletion derivative of VP1 that lacks the N-terminal activation domain was as effective in repressing Amy-GUS expression in maize and badey as the full-length protein. In addition, a VP16NP1 hybrid protein that contains three copies of the VP16 acidic activation domain and has a restored capacity to activate Em-GUS and C1-Sh-GUS (McCarty et al., 1991; Rosenkrans and McCarty, unpublished results) was not more effective than the activator deletion mutant in causing repression of Amy-GUS. The lack of a requirement for the activation sequence clearly distinguishes the mechanism of VP1-mediated repression from the mechanism of activation of diverse maturation related genes by VP1. These results strongly indicate that the VP1 protein has a discrete repressor function. Identification of Sequences Essential for the Repressor Function of VP1

A number of internal deletion constructs were tested for their ability to repress Amy-GUS in maize and barley aleurone (Fig. 18). A large ca. 350 bp deletion (86/87) entirely abolished VP1 repressor function, indicating that the deletion-derivative may lack a functionally important domain, or the deletion may affect the spacing and thereby the function of domains present outside this sequence. An only slightly smaller deletion in this region (86/85) did not affect repression in maize nor barley aleurone, suggesting that altered spacing is not likely the reason for the failure of 86/87 to repress. Indeed, a small deletion of 42 bp in the C-terminal half of 86/87 (85/87) abolished repression in barley and consistently reduced activity in maize aleurone by ca. 50%. Hence, this region (hereafter referred to as the RED domain) appears to be essential for VP1 repressor function. Consistent with this conclusion, a large part of the sequence of the RED domain (W V Q N H* H+ L Q R+ K* R* P R* D-) is highly charged, predicting this domain to be positioned on the surface of the folded VP1 protein, thus accessible for interactions with other molecules.





87


Rel. Amy-GUS / LUC Effector Construct Maize Barley


Control (no VPI)


100


4-12


100 18-33


8687


86185 E 85/87 =KI


I 133 (*29) 87 (8)


] 7 (*2) 13 (*3)


Ipwln-am n dw Wns) (RED)


87/88


87/92 [11U { 92/181

93/95


I'I[


92/88 E mil 11 MU. I


196/88


45(4) 99 (*21)


85-162 90-231

9 (*1) 39 (*10)


6 (*2) 29 (*7)

14 (*2) 33 (*8)


27-100 125 (*17)

84 (*5)


Fig. 18. Deletion analysis of the VPI protein: Part 1. For Materials and Methods see Fig. 19. Black boxes indicate sequence homology to the VP1 homolog from barley.


VP1-WT


t I EF





68


Rel. Amy-GUS / LUC

Effector Construct Maize Barley


Control (no VP1) 100 100

VPI-WT 4-12 15-30

103/104 21 (4) 27 (*3)


101/100 18 (*2) 23 (*4)


VPI-McW I E I|t N MILE 32 (49) 13 (*0.5)



Fig. 19. Deletion analysis of the VP1 protein: Part II. Aleurones of maize and badey germinating seeds were bombarded with 2-5 pg of Amy-GUS, 5 pg of Ubi-LUC and 5 pg of effector construct and then cultured in 106 M GA3. Data represent mean ( S.E.M.) of 3-5
replicates






69


A second large. ca. 400 bp deletion (87/88) also rendered the VP1 protein Incapable of repressing Amy-GUS (FIg. 18). The activity of this construct varied to an unusual extent, from slight, but non-significant repression in some experiments to more than two-fold, statistically significant activation of Amy-GUS in others. Similarly, a construct with a slightly smaller deletion of this region (92/88) displayed a highly variable effect. Subsequently, four adjacent subdeletions within the 87/88 domain were constructed. Each of the N-terminal three deletions (87/92, 92/181, 93/95) eliminates one region that is conserved in the barley VP1 homolog, but none of these deletions severely diminished repression. In contrast, the C-terminal, ca. 140 bp deletion 19/88 eliminating a non-conserved stretch of VP1 almost entirely abolished repressor function, implying that this region may contain an important site involved In Amy-GUS repression.

Deletions in the C-terminal portion of VP1 did not strongly affect repression of Amy-GUS (Fig. 19). Truncation of the C-terminal 450 bp (VPI-McW) generating the product of the vplMc~hkrter allele which confers a non-viviparous, anthocyanin-deficient phenotype had only a slight effect in maize, while not affecting repression in barley. Similarly, deletion of the domains 101/100 and 103/104 did not strongly diminish Amy-GUS repression.


The 87/88 deletion mutant

It was shown that in the absence of GA, co-expression of VP1 had no effect on AmyGUS expression in aleurone of germinating barley half seeds (see Fig. 11). In contrast, when the 87/88 deletion-derivative of VP1 was over-expressed with Amy-GUS in the absence of GA, Amy-GUS expression was activated (Fig. 20). This activation was highly variable, ranging from 2-fold in some experiments to up to 12-fold in others. It appeared specific to the 87/88 deletion mutant and was not found for any other tested constructs containing deletions outside this region of VP1 (data not shown). Interestingly, activation of Amy-GUS by 87/88 was also observed in the presence of ABA (Fig. 21). In the presence of GA, a slight activation of Amy-GUS (max. 2fold) by 87/88 was observed in some, but not all experiments (Figs. 18, 20).













150



100s0.


2500

t200015001000e 5000-


70





no GA


I


4....


~1

I






C,

t


'1


CAT VP1 -Act 87/88 -Act 87/88







+ GA













CAT VP1 -Act 87/88 -Act 87/88


Fig. 20. The VP1 deletion-derivative 87/88 activates Amy-GUS in the absence of GA in aleurones of germinating barley seeds. Aleurones were bombarded with 2 pg of Amy-GUS, 5 pg of Ubi-LUC and 10 pg of either 35S-Sh-CAT, 35S-Sh-VP1, the activation domain-deletion mutant described in Fig. 18 (-Act), 87/88 or the double mutant that carries deletions of the activation domain and 87/88 (-Act:87/88), respectively. After bombardment, endosperms were cultured in no hormones (top graph) or 10-6 M GA3 (bottom graph). Data represent mean ( S.E.M.) of three replicates.


......... ...... ............. ....... .............. ....... .......
......... ....... .......






























CAT VP1 87/88


=3 no ABAxms
-.+ ABA


T


CAT VP 67/88


Fig. 21. ABA does not inhibit 87/88-mediated activation of Amy-GUS found in the absence of GA. Aleurones of germinating barley seeds were bombarded as described in Fig. 20 and then cultured in no hormones or 10-5 M ABA.


T


T


T


UAI ao=


Vri Vra187/88


Fig. 22. The VP1 deletion mutant 87/88 displays a dominant negative effect on VP1-mediated repression of Amy-GUS. Aleurones of germinating barley seeds were bombarded with 6 pg of Amy-GUS, 5 pg of Ubi-LUC and either 25 pg of CAT, 20 pg of 87/88, 5 pg of VP1 or both 20 pg of 87/88 and 5 pg of VP1. To all mixtures, CAT plasmid was added to obtain a total amount of 36 pg of plasmid DNA. After bombardment, endosperms were cultured in 10-6 M GA3. Data represent mean ( S.E.M.) of five replicates.


71


5.




-.


1501251007550 25


I


II-'


150-


5.










2 D:


100-


50-


o-J-


sib






72


.3






C,


4030-


100-


T


T


CAT VP1


vP1+ vP1+ vP1+ S7/88 WM 87/88-KW


Fig. 23. Effect of the double deletion mutants 85/88 and 87/88:McW on Inhibition of Amy-GUS by over-expressed VP1. Materials and Methods as described in Fig. 22. The double mutants 85/88 and 87/88:McW were constructed by restriction enzyme digestion and subsequent ligation. Data represent mean ( S.E.M.) of 5-6 replicates.


w 1250a 750.. 500 25


T


T


T


-I--


0 1-


67/88 87/104 87/18:0110


Fig. 24. Effect of the double deletion mutants 87/104 and 87/88:101/100 on inhibition of AmyGUS by over-expressed VP1. Materials and Methods as described in Fig. 22.






73


The finding that 87/88 activates Amy-GUS Indicates that the mutant protein is not fully non-functional. The data suggest that 87/88 may be capable of interacting with a normal component of the repression mechanism but unable to cause repression. In doing so, there are at least two possibilities as to how it might activate Amy-GUS. 1) The acidic activation domain of 87/88 might elicit transcriptional activation of c,-amylase genes or an intermediate gene. In the wild-type VP1 protein, this activity might either be not accessible or masked by the repressor function. 2) In producing a non-functional complex, 87/88 might compete with, or titrate out, an endogenous repressor (e.g. endogenous barley VPI-homolog possibly present in aleurone of germinating barley seeds) and thus exert a dominant negative effect.

In order to test the first possibility, a double deletion mutant was constructed that deletes the 87/88 domain and the acidic activation sequence. This double mutant was as effective in activating Amy-GUS as the 87/88 single mutant (Fig. 20), suggesting that the transcriptional activation domain is not involved. Therefore, it was tested whether 87/88 is capable of inhibiting the effect of recombinant VP1. 87/88 and recombinant VP1 were expressed by themselves and in combination (ratio 4:1) together with Amy-GUS in barley aleurone. Co-expression of 87/88 reduced VPI-mediated repression of Amy-GUS by ca. 75% (Fig. 22). This is consistent with the view of a dominant negative effect of 87/88.

To identify domains involved in mediating the dominant negative effect of 87/88, doubledeletion mutants deleting 87/88 and other sequences of the VP1 protein were constructed and tested for their ability to reduce repression of Amy-GUS by co-expressed VP1. Double mutants deleting the domains 87/88 and 85/87 (the RED domain) or 101/100, respectively, were as effective in competing with recombinant VP1 as the 87/88 single mutant (Figs. 23, 24). In contrast, the double mutants deleting 87/88 and either the C-terminal 450 bp of VP1 (87/88:McW) or the domain 103/104 (87/104) did not show a dominant negative effect on AmyGUS repression by co-expressed VP1 (Figs. 23, 24). suggesting that these domains may be important for the inhibitory role of 87/88 on VP1 repressor function.















DISCUSSION


VP1 of maize Is a transcription factor that Is specifically expressed in the developing seed (McCarty et al., 1989a, 1991). It was shown previously that VP1 is required for ABAinduced activation of a variety of genes associated with seed maturation (McCarty et al., 1991). Results of this work show that, in addition to its transcriptional activator function, VP1 has a specific role in blocking precocious induction of germination-specific c-amylase genes during seed development.



VP1 Represses a-Amylase Genes


This study provides at least three lines of evidence that indicate a function of VP1 in repression of c-amylase genes in the developing seed. First, somatically unstable vpl-m2 seeds containing both vpl mutant and wild-type sectors displayed cell autonomous de-repression of endosperm remobilization specifically in sectors underlying vpl mutant aleurone (Fig. 5a,b). Second, in transient expression experiments Amy-GUS was inducible or constitutively active in developing vpl mutant aleurone cells but not in wild-type aleurone cells (Table 1). Third, coexpression of recombinant VP1 with Amy-GUS in vpl mutant aleurone cells inhibited Amy-GUS expression by >95% (Fig. 6). These results are consistent with findings that cc-amylase genes are not expressed in the developing seed (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Hence, cessation of VP1 expression prior to germination may be necessary to allow induction of cL-amylase genes in the germinating seed.


74






75


Gene Reression Is a Discrete Function of VP1


In contrast to the mechanism of transcriptional activation of maturation-specific genes, VP1-mediated repression of a-amylase genes does not require the transcriptional activation function located at the N-terminal domain of VP1 (Fig. 17). This indicates that VP1 has a discrete repressor function that is mechanistically distinct from the transcriptional activation function. Several systems in which a single transcription factor functions as both an activator and a repressor depending on the target promoter have been described in animals (Miner and Yamamoto, 1991; Tsai and O'Malley, 1994). Direct structural homologs of VP1 are thus far known only in plants, suggesting that this strategy has evolved independently in plants and animals.



Functional Analysis of the VP1 Protein


To identify domains in the VP1 protein that are important for repressor function, mutant derivatives containing deletions covering ca. 80 % of the total protein were tested for their ability to inhibit Amy-GUS. Deletion of very highly conserved sequences in the C-terminal half of VP1 (103/104, 101/100, McW) did not, or only slightly, reduce repressor function (Fig. 19). In contrast, two constructs deleting sequences in the middle of the VP1 protein (85/87, 87/88) were strongly affected in repression of Amy-GUS (Figs. 18). While disruption of VP1 function is one possibility for lack of Amy-GUS repression, low stability of mutant mRNA or protein could be an altemative explanation. However, this possibility is unlikely for two reasons: 1) both constructs were capable of activating a C1-Sh-GUS reporter gene in maize protoplasts: 85/87 and 87/88 activated C1-Sh-GUS at a level of 77-84% or 56% of the wild-type VP1 construct, respectively (V. Vasil, L Rosenkrans et al., unpublished results). 2) co-expression of 87/88 as well as the double-deletion mutant 85/88 exhibited a dominant negative effect on repression of Amy-GUS by wild-type VP1 in barley aleurone, indicating presence of mutant protein in transformed cells.






78


The strongly positively charged, 15-amino acid-domain (RED domain), deleted In 85/87, is highly conserved (ca. 80%) among maize, barley and rice genes (Fig. 25). Between maize and barley, 11 out of 15 amino acids are Identical and one amino acid constitutes a conservative substitution (R to K) also found in rice. The high degree of sequence conservation Is consistent with the finding that this region is of functional importance for repression of Amy-GUS. Interestingly, deletion of the RED domain exhibited a differential effect In maize and barley aleurone. While 85/87 was incapable of inhibiting Amy-GUS in barley aleurone, it retained ca. 50% of wild-type VPI activity in maize aleurone cells (Fig. 18). In maize, deletion of additional sequences 5' to the 85/87 deletion (construct 86/87) was necessary to eliminate repressor function (Fig. 18). This indicates that in the barley cell the RED domain is absolutely essential for proper execution of the repression mechanism, whereas In the more concordant system of the maize aleurone cell other possibly less conserved regions in the VP1 protein partially compensate in function for the RED domain.

The construct 87/88 that deletes a large but poorly conserved stretch of VP1 was incapable of repressing Amy-GUS in maize or barley cells. To further analyze this domain of VP1, four smaller deletions within this region were constructed and tested (Fig. 18). The three deletion constructs 87/92 (deleting 29 aa), 92/181 (deleting 26 aa) and 93/95 (deleting 40 aa) of which each lacks a short stretch of conserved sequence were not severely impaired in repressing Amy-GUS. In contrast, the construct 196/88 deleting 48 amino acids at the C-terminal end of the 87/88 deletion was incapable of repression in barley (no data for this construct were obtained in maize). These results allow at least two interpretations: 1) 196/88 deletes a domain essential for repression, while the sequences located between the deletion points 87 and 95 are not required for repressor function. 2) The partially conserved sequences between 87 and 95 are of redundant function. Therefore, deletion of two or more conserved blocks may be necessary to lose repression of Amy-GUS. In this interpretation, 196/88 may delete an additional important domain or, alternatively, affect proper spacing between further N-termnal and C-terminal sequences.







77


Z.m. MEA-SSGSSPPHSQENPPEH GGD ------M-GG ------AP-AZZI GGZAA -------DDF 39
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
H.v. MDA-SAGPPPPRHPQGSALRRGKG-----------------P-AVEIRHGE---------DDF 34
III III I I I I I 1I1 II 1I
O.S. MDA-SAGSSAPHSHGNPGK-GGG-------GGGGGGRGKAP-AAZIR-GEAAR------DDV 46
1 1 1 1 1 1 1 1
A. t. MKSLHVANAGDLAEDCGIL-GGDADDTVLMDGIDEVGREIWLDD---HGGDNNHVHGHODDL 60


Z.m. MFAED--TF----PSLPDFPCLSSPSSSTFSSN------SSSNSSSAYTNTAGRA-G-----G 86
liII I I I ll llllll 1
H.v. MFAQD--TF----PAFPDFPCLSSPSSSAADIV----------------------------LCG 64
1I 1 I1 I l lllIll ll
O.s. FFADD--TF----PLLPDFPCLSSPSSSTFSSS------SSSNSSSAFTTAAGGGCG-----G 94
1 1 I l1lll ll Ill I I1 lill
A.t. IVHHDPSIFYGDLPTLPDFPCMSSSSSSSTSPAPVNAIVSSASSSSAASSSTSSAASMALRS 123


Z.m. EPSEPASAGEGFDA---LDDIDLLDFASLSM--PWDSEP------------------------- 125
111111 II I I I II 11 I1 11 I 1
H.v. EPSEPAAAGDGMDD---LSDIDHLLDLASINDDVPWDDE-PL------------------------ 102
HIM1 1 11 1 Ill lIl ll II I 1
0.s. EPSEPASAADGFGE---LADIDQLLDLASLS--VPWAEQPL------------------------ 135
I I I
A.t. DGEDPTPNQNQYASGNCDDSSGALQSTASMEIPLDSSGFGCGEGGGDCIDMMTFGYMDLLD 186


86
Z .m. ---FP-G VSMMENAMSAPPQPVGD--GMSEEKA--VPEGTT---GGEEACM-DAS--EG-EE 163
1 1 1 1 1 1 1 I I I I 1I1 I I
H.v. ---FP-DVGMMLEDVISEQQQQQQQHPLAGHGAGGRVASDTAGG-GGEDAFMGGGGSGSAADD 160
11 H il11 1 1I I I I I II 11 1 111
0.3. ---FPDDVGMMIEDAMSGOPHQADDCTGDGDTKA--VMEAAGGGDDAGDACM-E-GS-DAPDD 179
1 1 1 1 1 I I I 1 1 1
A.t. SNEFFDTSAIFSODDDTQNPNLMDQTLERQEDQV-VVPMMENNS-GODMQMM--NSSLEQDDD 240

85
Z.m. LPRFFMEWLTSNRENISAEDLRGIRLRRSTIEAAAARLGGGRQGTMQLLKLILTWVQNHHLQR 230
||||1111i 1 111111| IT l mll il i ml i m ml il llim ilm
H.v. LPRFFMEWLTNIRDCISAEDLLSIRLRRSTIETTTALLGGGRQDTMQLLILTWVQSHHLQK 223
11 1111111 11 111 11 MT mi l Ti I T lim i m mlll llim i lil i
0.s. LPAFFHEWLTSNREYISADDLRSIRIRRSTIEAAAARLGGGRQGTMQLLJILTWVQNHHLQK 247
1 1 111 1 I 11 111 I ITI i Mi II imiiiii iIi III
A. t. LAAVFLEWLNNKETVSAEDLRKVKIKKATIESAARRLGG1EAMKQLLKLILEWVQTNHLR 308


87 92 93
Z.m. KRPRDVMEE-EA-GLHVQLPSPVANPPGYEFPAGGQDMAAGGGTSWM---PHQQAFTPPAAYG 288
111 iii I I1 I iii I 1 1 11
H.v. KRPRVGAMDOEAPPAGGQLPSPGANPS-YEFPT---ETGAAAATSWM---PY-QAFSPTASYG 278
Ii i 1 I 111111111 1 1 1 1 1111111 I I 11 1 1 11
0.s. KRPRTAIDDGAA-SSDPQLPSPGANP-GYEFPSGGQEMGSAAATSWM---PYQ-AFTPPAAYG 304
I I I I I 1
A. t. RRTTTTTTNLSY-QQSFQQDPFQNPNPNNNNLIPPSDQTCFSPSTWVPPPPQQQAFVSDPGFG 370


181 196,95
Z.m. GDAVYPSAAGQQYSFHQGPSTSSVVVNSQPFSPP---PVGDMH---GANMAWPQQYVPFPPPG 345
1 1 II IIII IIIIIIII I I I I 11111 II
H.v. GEAMYPFQ--------QGCSTSSVAVSSQPFSPP--AAA-DMHA--G---AWPLYAAFVPAG 325
11111 1 111111 111111111 111 1111 1 I 1 1 1
0.s. GDAMYPGAAG-PFPFQQSCSKSSVVVSSQPFSPPTAAAAGDMHASGGGNMAWQQFAPF--PV 364
I I I I II I 1 1 1 1
A.t. ----YMPAP--NYP--PQPEFLPLLESPPSWPPP---PQ-------SGPMP-HQQF-PM-PPT 412







78


88 104 **** 103
Z.m. ASTGS---YPMPQPFSPGFGQYAGAGAGHLSVAPORMAGVEASATKLAJKaRMARQRRLSCL 405
I 11111 I I I liii I lIlIIIIIllIlIlIIII
H.v. ATSAGTQTYPMPPP-GPV-POPFAAPGFA--GQFPRM---EPAATREARKKRMARQRRLSCL 381

0.s. SSTSS---YTMPSVVPPPTAGFPGQYSGGHAMCSPRLAGvEPSSTKEARKKRMARQRRLSCL 424
I I I I II I lI III II II ll l I
A.t. SQYNQFG DPTGFNGYNMNPYQYPYVPAGOMRDRLLRLCSSATKEARKKRMARORRL--L 470

Z.m. QQQRSQQLSLGQIQTSVHLQEPSPRSTHSGPVTPSAGGWGFWSPSSQ----QQVQNPLS-KSN 463
1 111 1 1111 1i i 11 1111 111
H.v. Q???????????IQTGGFPQQPSPRAAHSAPVWG?HWSPPAVQAQPHGQLIQVPNPLSTKSN 444
I 1 1 I I I IlIl II 1111 11
O.s. QQQRSQQLNLSQIHISGHPOEPSPRAAHSAPVTPSSAGCRSWGIWPP--AAQIIQNPLSNXPN 485
I I I I I I I II
A.t. SHHHRH--NNNNNNNNNNQQNQTQIGETCAAVAPQLN--------PV--ATTATGGTWMYWPN 521

Z.m. SSRAPPSSLEAAAAAPQTIPAP-AGARQDDIHHRLAAASDKRQGAKADKNLRFLLQKVLKQSD 526
III III I I11111I lII llllI l
H.v. SSRQKQQKPSPDAAAR-PPSGGGASQQRQGQ----AAASDKQRQQ---KLRFLLQKVLKQSD 499
1 1 1 1 1 1 I I I l IIl1 1Il
O.S. PPPAT--SKQPKPSPEKIKPKQAAATAGASLQRSTASEKRO-AKTDKNLRFLLQKVLKQSD 545
11 1 1 1 1 1 1 1 1 1 1 IIIIIIIII IIII
A. t. -VPAV--PPQLPPVMETQLPTMDRAGSASAMPRQQVVP-DRRQGWKPEKNLRFLLOKVLKQSD 581


/MoW 101
Z.m. VGSLGRIVLPKEAEVHLPELKTRDGISIPMEDIGTSRVWNMRYRFWPNKSRMYLLENTGEF 589
I1 1l 11111 1111 1 111111 IM H I 11 1111 II l IIII
H.v. VGTLGRIVLPIOEAETHLPELKTGDGISIPIEDIGTSQVWSMRYRFRPNNKSRMYLLENTGEF 562
I 1Illllllll 1111111 I1 11 111111111 1IIIIIIIIIIIIIIIII I
O.S. VGSLGRIVLPKEAEVHLPELKTRDGVSIPMEDIGTSQVWNMRYRFWPNNKSRMYLLENTGDF 600
II lilllltIi l 111 liii| 1111111| IIIIlIlI IIIIllIllIIIIII
A.t. VGNLGRIVLPKKEAETHLPELEARDGISLAMEDIGTSRVWNMRYRFWPNNKSRMYLLENTGDF 644


100
Z.m. VRSNELQEGDFIVIYSDVKSGKYLIRGVKVR-PPAQEQGS--GSSG-GGi------RP-LC- 640
11111 1lii HIM11 HIM1 III III I 1 I1 1
H.v. VRSNE????DFIVLYSDVKSQKYLIRGYKVR--AAQELASTRWQSREGGA??V-------LAO 616
11111 1111 11 II 11 H IM III III I I
0.s. VRSNELQEGDFIVIYSDIKSGKYLIRGVKVRR-AAQEQGN---SSGAVGKHKHGSPEKPGVSS 667
I I lIll llllll I ll llIll I 11I
A.t. VKTNGLQEGDFIVIYSDVKCGKYLIRGVKVRPSGQKPEA-PPSSAATKR------------- 693


Z.m. PAGPERAAAAGAPEDAVVDGV--------SGACKGRSPEGVRRVRQQGAGA--MSQMAVSI 691

H.v. TAAD 620
I I Il||ll| ll| lli II 11111111
O.S. NTIAAGAEDGTGGDDSAEAAAAAAAGkDGGGCKGKSPHGVRRSRQEAAAAASMSQMAVSI 728
I1 1
A.t. --QNKSQRNINNNSPSA-NVVVA-----------SPTSQTVK----------------------- 720





Fig. 25. Alignment of the amino acid sequences of VP1 from maize (Z.m.), HVVP1 from barley (H.v.), OSVP1 from rice (O.s.) and AB13 from Arabidopsis (A.t.). Identical amino acids are shown by vertical lines. Conserved regions among the four species are boxed. Italic numbers above sequence show site of deletion points. Underlined amino acids indicate location of putative leucine zipper. Asterisks indicate putative NLS.






79


in summary, using single-deletion mutants, amino acid residues within the region from 222 to 374 of VP1 were found to be Important for repression of Amy-GUS. Further functional analysis of sequences outside this region was conducted by taking advantage of the dominant negative effect of the 87/88 mutant on repression of Amy-GUS by wild-type VP1: In the presence of GA, co-expression of 87/88 with VP1 and Amy-GUS severely reduced repression of Amy-GUS by VP1 (Fig. 22). Domains essential for mediating the dominant negative effect of 87/88 are most likely also involved in function of the wild-type VP1 protein. 87/88 may exert a dominant negative effect for instance by competing with wild-type VP1 for binding to a component of the repression mechanism or If VP1 functions as a dimer by forming nonfunctional heterodimers with wild-type VP1. To identify domains in the 87/88 mutant that are required for expressing the dominant negative effect, double-deletion mutants between 87/88 and other deletion mutants were constructed and tested. The RED domain deleted In the 85/87 mutant was considered a putative domain because it has an important function in repression. However, the double-deletion mutant 85/88 was as effective in causing a dominant negative effect as the 87/88 single mutant, indicating that the RED domain is not essential for this effect (Fig. 23). Similar results were obtained for the domain 101/100 (Fig. 24). In contrast, the double mutants deleting 87/88 and either the C-terminal 450 bp of VP1 (87/88:McW) or the highly basic domain 103/104 (87/103) did not exhibit a dominant negative effect on Amy-GUS repression by co-expressed VP1 (Figs. 23, 24). Although it cannot be ruled out that these double mutant constructs express instable proteins, these data suggest that the C-terminus and the domain 103/104 may be required for mediating the dominant negative effect of 87/88. Deletion of these domains displayed a clearly measurable effect only if the 87/88 domain was deleted also, while the single mutants 103/104 and McW retained almost wild-type repressor function (Fig. 19). This suggests that in the single mutants other sequences can compensate in function for the deleted domains, while this is not possible in the 87/88 mutant. However, stability of the mutant proteins needs to be confirmed, especially for the 87/88:McW double mutant which deletes ca. 42% of the VP1 sequence.






80


in summary, deletion analysis of the VP1 protein has allowed the identification of several domains that are essential for repressor function: 1) the conserved, highly charged RED domain 85/87 and 2) the poorly conserved region 87/88. Moreover, though not essential, the domain 103/104 and the C-terminus may play a role in repression.


When comparing the repressor domains of VP1 with the domains required for transcriptional activation of the Em or C1 genes, it is evident that different functions of VP1 map to different sequences in the protein (Fig. 26). Apart from the differential requirement of the acidic activation domain at the N-terminus of VP1, a-amylase repression and Em activation differed in the need for the highly positively charged domain 103/104 of VP1. While deletion of 103/104 reduced Em-GUS activation by 98% in maize protoplasts (L. Rosenkrans et al., unpublished results), It did not severely affect repression of Amy-GUS. Ukewise, a-amyfase repression and C1 activation displayed a differential requirement for the C-terminal part of VP1. Deletion of the C-terminal ca. 150 amino acids of VP1 entirely eliminated activation of CI-ShGUS (L. Rosenkrans et al., unpublished results). In contrast, this domain was not found essential for repression of Amy-GUS.

Similarly, different domains of VP1 are essential for activation of C1 and activation of Em. Overall, sequences required for induction of anthocyanin biosynthesis map to the highly conserved C-terminal end of VP1, while sequences essential for activating Em are located in the central part of VP1 (85/87, 87/88 and 103/104). These findings are consistent with the phenotype of mutant alleles in maize and Arabidopsis. The maize vpl-McW allele truncating the C-terminal ca. 150 amino acids of VP1 produces seeds exhibiting nearly normal developmental arrest but lack of anthocyanin accumulation, indicating that the C-terminal part of VP1 is not essential for preventing vivipary but is essential for anthocyanin production (McCarty et al., 1989b). In Arabidopsis, two abi-3 alleles that produce a viviparous phenotype have been sequenced: abi3-6 contains an intemal deletion of ca. 750 bp between positions 1,073 and 1,944 (Nambara et al., 1994), thus deleting the domains corresponding to 196/88 and 103/104. The mutation in abi3-3 induces a premature stop codon at Gin417 (Giraudat et al., 1992), thus deleting









Amylase repression Em activation C1 activation


86 85 87 88 103 104


-n.d. LI ||!|


U


VP1 1


Exons 1 2 345 6


non-essential


essential


Fig. 26. Summary on domains of VP1 Involved In repression of a-amylase genes, activation of C1 and activation of Em.


w


1 -1

EMEM






82


the domain 103/104 and the C-terminus.

VP1 shows no significant sequence homology to other proteins. Therefore, the function of domains other than the acidic activation sequence Is thus far unknown. It was suggested that the region from amino acid 208 to 235 of the rice VP1 may form a leucine zipper-like structure (Hattori et al., 1994). In this region, Lou or lie residues are located at every seventh residue (5 repeats) with an exception for the fourth position (see Fig. 25). However, no severe loss-offunction phenotype has been observed when this domain was deleted in the 8W/85 construct. 86/85 effected full repression of Amy-GUS and ca. 50% of wild-type activity with respect to activation of Em-GUS. These data do not support an important role of this domain in proteinprotein interactions.

Based on the evidence that VP1 transcriptionally activates C1 and Em (McCarty et al., 1991; Hattori et al., 1992), nuclear targeting is likely to be a requirement for VP1 function. A 100% conserved putative nuclear localization sequence (NLS), RKKR, exists in the domain from amino acid 392 to 395 of VP1 (see Fig. 25), as mentioned by Giraudat et al. (1992). Consistent with these views, deletion of this putative NLS (construct 103/104) fully eliminated activation of Em-GUS (L. Rosenkrans et al., unpublished results). However, 103/104 retained capacity to activate CI-Sh-GUS (31% of wild-type VP1), suggesting that the mutant protein is targeted to the nucleus though possibly with reduced efficiency. Assuming that RKKR is a functional NLS, these data indicate that VP1 contains two or more NLSs with at least partially redundant function, a feature not uncommon among nuclear proteins (Raikhel, 1992). The differential effects of 103/104 on activation of Em and C1 may reflect different threshold levels of VP1 protein required for these activator functions. Hence, the extent of nuclear localization of 103/104 may be sufficient for partial activation of C1 but not for activation of Em. Altematively, 103/104 may serve an additional function in activation of Em. A dual role of a domain in both nuclear targeting and DNA binding was reported for the regulatory protein 02 (Varagona et al., 1994).

With respect to repression of Amy-GUS by VP1, it is unknown whether nuclear localization of VP1 is required for function. The construct 103/104 was not affected in repression






83


of Amy-GUS. This may Indicate that nuclear targeting of VP1 is not required or that other domains with redundant function may compensate in function for the deletion in 103/104. Interestingly, the double-deletion mutant 87/103 lost the ability to exert a dominant negative effect on Amy-GUS repression by VP1. A failure of the double mutant protein to be targeted to the nucleus would be consistent with the observed loss of function.


Apart from the activation domain, the biochemical functions of sequences In the VP1 protein remain unclear. However, the deletion analysis demonstrated very clearly that different domains are involved in the different functions of VP1, thus underlining the multifunctional nature of this transcription factor.



Interactions Between VPI and Plant Hormones


One can envision at least three models of how VP1 might function in repressing oamylase genes: 1) VP1 might mediate ABA antagonism of GA signalling during seed development. ABA is known to antagonize GA-action in the regulation of cx-amylase genes in germinating cereal seeds (Jacobsen and Chandler, 1987). Because VP1 is required for ABAinduced gene expression associated with seed maturation (McCarty et al., 1991), it might also be essential in ABA-mediated repression of a-amylase genes (Fig. 27A). Consequently, the vpl mutant might allow de-repression of a-amylase genes by failing to respond to ABA present In the developing seed. 2) VP1 might specifically inhibit the GA-response pathway independently of ABA (Fig. 27B). 3) VP1 might repress a-amylase genes via a pathway that functions independently of both GA and ABA (Fig. 27C).

The results presented in this work do not support the first model. ABA was effective in blocking Amy-GUS expression in vpl mutant aleurone cells (Fig. 6), indicating that ABA action in this instance does not depend on the presence of VP1. In combination, VP1 and ABA effects were roughly additive. This stands in contrast to evidence showing that VP1 is required for ABAinduced expression of the maize Em gene (McCarty et al., 1991). Thus, there appear to be at






84


GA


}-- VPI ABA


GA


ABA VP1


AmyISe


B.


A. VPI may be required for ABA
function


VPI may block GA response pathway independently of ABA


*1


GA
Activator
ABA -- VPI


I Amyase


C. VP1 may repress amylase genes
independently of hormones


Fig. 27. Alternative models for VP1 function as described in the text


Amylase






85


least two modes of ABA action in the maize seed, a VPI-dependent pathway and a VP1Independent pathway. Multiple ABA transduction pathways are also Indicated by interactions between ABA-insensitive mutants of Arabidopsis (Finkelstein and Somerville, 1990; Finkelstein, 1994). This suggests that ABA modulates the activity of diverse regulatory cascades in the seed.

The second scenario In which VP1 could specifically block GA signal transduction Is supported by the evidence that over-expression of VPI in aleurone of imbibed barley half seeds severely reduced GA-induction of Amy-GUS without affecting the basal activity of the a-amylase promoter (Fig. 11). This suggests that expression of VP1 In the developing seed may be, at least in part, responsible for the observed GA-insensitivity of cereal and maize C-amylase genes prior to seed maturity (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). VP1 displayed full repressing activity in slender (sin) mutant barley seeds (Fig. 12) which are constitutive in GA response of the aleurone (Chandler, 1988; Lanahan and Ho, 1988), suggesting that VPI functions at a point downstream of the Sin gene product.

With respect to the maize seed, the data do not rule out the possibility that VP1 acts independently of GA as a developmental repressor of a-amylase genes. Although we have shown that Amy-GUS is GA-inducible in vpl mutant aleurones early in development (Table 1), it is not clear that the high constitutive activities found later in development are entirely attributable to changes in GA concentration. In contrast to the situation of Himalaya barley seed and other cereal grains, studies of a-amylase regulation in normal and GA-deficient (d5 mutant) genotypes of maize indicate that a-amylase induction in germinating maize seeds is largely independent of GA (Harvey and Oaks, 1974). Consistent with these studies, it was found in the present work that during germination Amy-GUS is constitutively active in the GA1-deficient dl mutant of maize. Because Amy-GUS was fully VP1-repressible in aleurones of developing vpl mutant, germinating wild-type and germinating dl-mutant seeds of maize, it is suggested that VP1mediated repression is not necessarily restricted to, nor solely defined by, inhibition of the GA response. Though the significance of GA in the expression of a-amylase genes needs to be






88


investigated further, it is likely that GA, ABA and VP1 are three among several factors that regulate the activity of constituents required for expression of a-amylase genes in the maize seed (Fig. 27C).


Expression experiments in developing maize seeds have shown that VP1 and ABA are likely to act independently in repressing Amy-GUS. Further evidence regarding the relative positions of VPI and ABA in the regulatory network allowed the characterization of the dominant negative effect caused by the deletion mutant 87/88. Co-expression of 87/88 with wild-type VPI and Amy-GUS has shown that 87/88 inhibits repression of Amy-GUS by VP1 ('dominant negative effect", Fig. 22). This suggests that the observed 87/88-mediated activation of AmyGUS in the absence of GA (Fig. 20) may be caused by the presence of residual amounts of endogenous barley VPI homolog in the wild-type aleurones. Because ABA was incapable of inhibiting the 87/88-mediated activation of Amy-GUS (Fig. 21), it is suggestive that VP1 functions either downstream of ABA or via a different signalling pathway than ABA. It therefore will be interesting to map the cis-elements in a-amylase promoters that are responsible for VP1 and ABA action.



The Role of the Embryo


Although the vpl mutant phenotype was cell autonomous within the aleurone of vpl-m2 seeds (Fig. Sa,b), de-repression of a-amylase genes was not fully independent of the physiological state of the embryo: 1) in vpl-m2 seeds, precocious hydrolization of endosperm reserves in sectors underlying vpl mutant aleurone was predominantly observed in seeds carrying a viviparous embryo. 2) Amy-GUS was de-repressed in vp1 mutant aleurone of concordant vpl mutant seeds but not in vpl mutant aleurone of non-concordant seeds exhibiting a wild-type embryo (Fig. 15). These observations suggest that a viviparous embryo facilitates expression of a-amylase genes in vpl mutant aleurone. However, the finding that Amy-GUS is highly induced in vpl mutant aleurone of germless seeds (Fig. 16) indicates that a viviparous


I e* ,






87


embryo per se Is not required for de-repression of Amy-GUS In vpl mutant aleurone. Instead, it rather appears to be the lack of a normal embryo that facilitates expression of Amy-GUS, suggesting that a wild-type embryo contributes a diffusable signal with inhibitory effect on aamylase gene expression In the aleurone. Experimental evidence suggests that developing embryos are the major source of ABA present in the maturing seed (King, 1979; Jones and Brenner, 1987). Because Amy-GUS remains sensitive to Inhibition by ABA in vp1 mutant aleurone, ABA produced by the wild-type embryo may be responsible for the observed repression of Amy-GUS in vpl mutant aleurone of non-concordant seeds. This is consistent with the finding that Amy-GUS was de-repressed to a similar extent in developing aleurones of the ABA-deficient mutant vp5 as in aleurones of the germless mutant (data not shown).

In concordant vpl mutant seeds, ABA concentrations are equal to, or only ca. two-fold lower than, those present in wild-type seeds (Neill et al., 1986, 1987; Palva and Kriz, 1994). This suggests that the viviparous embryo also contributes an inductive signal (e.g. GA) that counteracts the effect of ABA and therefore uncovers de-repression of a-amylase genes in vpl mutant aleurone. However, the observed strong expression of Amy-GUS in vpl mutant aleurone of germless seeds clearly shows that GA production by the embryo is not required for x-amylase expression in maize aleurone. Supported by the evidence that the endosperm of mature cereal seeds is not a source of GA (Jacobsen and Chandler, 1987), these data suggest that Amy-GUS expression is largely independent of GA. Moreover, this interpretation is consistent with other studies (Harvey and Oaks, 1974) and findings in this work indicating that in mature seeds of GAdeficient mutants of maize a-amylase genes are expressed at high levels.

Complete repression of Amy-GUS in aleurones of developing seeds was observed only if the embryo as well as the endosperm were of wild-type genetic constitution (Fig. 16). Lack of either a normal, arrested embryo or VP1 expression in the aleurone lead to partial de-repression of Amy-GUS in aleurone cells. This indicates that neither factor expression of VP1 In the aleurone cells or the presence of a normal embryo is sufficient for total inhibition of ax-amylase genes.






88


VP1 intearates the Control of Seed Maturation and Germination Proarams


it has been shown In this work that VP1 participates In the regulation of two developmental pathways in the developing maize seed. As a transcriptional activator it Is required for activation of maturation-specific genes (McCarty et al., 1991) and as a repressor it prevents precocious induction of the normally germination-specific c,-amylase genes (data presented herin). Hence, expression of VP1 specifically during seed development appears to be involved in ensuring proper ordering of maturation and germination programs. Physically combining activation and repression function in one protein appears to provide one mechanism for directly integrating control of mutually exclusive developmental pathways in the plant embryo. The importance of a tight control of maturation and germination programs for seed survival is evident in the phenotype of vpl-m2 seeds.

Premature Induction of postgerminative development was also reported for the lecl (leafy cotyledon 1) mutant of Arabidopsis. In this ABA-sensitive, viviparous mutant, developing embryos expressed isocitrate lyase genes and a gene encoding a lipid transfer protein at levels that are normally characteristic of seedlings (West et al., 1994). Double mutant analysis suggested that the putative Arabidopsis VP1 homolog, ABI3, and LEC1 function in different pathways (Meinke et al., 1994). Hence, it appears that multiple mechanisms have evolved in flowering plants to prevent precocious induction of normally germination-specific genes in the developing embryo.

Thus far, the evidence that VP1 inhibits germination-specific genes is limited to hydrolase genes in aleurone cells. It is unknown to what extent this repressor activity of VP1 is also involved in preventing precocious germination of the embryo. Further insight into the inhibitory role of VP1 during seed development may be provided by stable transformation of vpl mutant plants with VPI-derivatives that are mutated specifically in the activator or repressor function.

Cloning of the Vpl related genes from barley (M. Stoll and D.R. McCarty, unpublished results), rice (Hattori et al., 1994), Arabidopsis (Giraudat et al., 1992) and tobacco (Phillips and






89


Conrad, 1994) Indicates that the Vpl gene Is conserved among flowering plants. Loss of ABI3 function in Arabidopsis causes a similar viviparous phenotype as the vpl mutation In maize (Nambarm et al., 1992). The functions of AB13 and VP1, however, diverge In so far that AB13 is required for seed dormancy in Arabidopsis while VP1 does not impose seed dormancy in maize. Because ABI3 mRNA is stored in the dry seed (Parcy et al., 1994), whereas VP1 transcript and protein are non-detectable in the mature seed (McCarty et al., 1989; Carson, 1992), one can speculate that dormancy in Arabidopsis may reflect an extended timing of ABI3 expression after seed maturity rather than a functional difference in the proteins. This view is supported by the results showing that over-expression of VP1 in aleurone of germinating maize seeds was effective in repressing Amy-GUS. A role of VP1 in maintaining seed dormancy is also consistent with the finding that dormancy in barley is correlated with a reduced GA-Inducibility of a,-amylase genes in the aleurone (Schuurink et al., 1992; Skadsen, 1993). Hence, it is suggested that VP1 plays a role in integrating the control of seed maturation, dormancy and germination programs.















SUMMARY AND CONCLUSIONS


The Viviparous-1 (VP1) transcriptional activator of maize is required for abscislc acidinduction of maturation-specific genes late In seed development leading to acquisition of desiccation tolerance and arrest in embryo growth (McCarty at al., 1991). The presented research extends these findings by showing that VP1, in addition to Its transcriptional activation function, inhibits precocious induction of the germination-specific z-amylase genes in aleurone cells of the developing seed. Functional analysis of deletion-derivatives of VP1 In a transient gene expression system indicated that VP1 has a discrete repressor function that Is separable from its transcriptional activation function. It is therefore suggested that physically combining activator and repressor functions in one protein provides one mechanism for directly integrating control of the mutually exclusive developmental pathways, seed maturation and seed germination, in the plant embryo.


90















REFERENCES


Akazawa, T. and S. Miyata. 1982. Biosynthesis and secretion of a-amylase and other
hydrolases in germinating cereal seeds. Essays Biochem. 18: 41-78.

Amuti. K.S. and C.J. Pollard. 1977. Soluble carbohydrates of dry and developing seeds.
Phytochem. 16: 529-532.

Anderberg, R.J. and M.K. Walker-Simmons. 1992. Isolation of a wheat cDNA clone for an
abscisic acid-inducible transcript with homology to protein kinases. Proc. Nati. Acad.
Sci. USA 89: 19183-10187.

Anderson, l.C. and D.S. Robertson. 1960. Role of carotenoids in protecting chlorophyll from
photodestruction. Plant Physiol. 35: 531-534.

Armstrong, C., M. Black, J.M. Chapman, H.A. Norman and R. Angold. 1982. The Induction of
sensitivity to gibberellin in aleurone tissue of developing wheat grains. I. The effect of
dehydration. Planta 154: 573-577.

BAumlein, H., S. Mis6ra, H. Luerpen, K. KOlle, C. Horstmann, U. Wobus and A.J. MOller. 1994.
The FUS3 gene of Arabidopsis thaliana is a regulator of gene expressin during late
embryogenesis. Plant J. 6: 379-387.

Beckett, J.B. 1993. Locating recessive genes to chromosome arm with B-A translocations. In:
The Maize Handbook, M. Freeling and V. Walbot, eds., pp. 308-314.

Bethke, P.C. and R.L. Jones. 1994. Ca2+-calmodulin modulates ion channel activity in storage
protein vacuoles of barley aleurone cells. Plant Cell 6: 277-285.

Blackman, S.A., R.L. Obendorf and A.C. Leopold. 1992. Maturation proteins and sugars In
desiccation tolerance of developing soybean seeds. Plant Physiol. 100: 225-230.

Blatt, M.R. and G. Thiel. 1993. Hormonal control of ion channel gating. Ann. Rev. Plant
Physiol. Mol. Biol. 44: 543-567.

Brown, A.H.D. and J.V. Jacobsen. 1982. Genetic basis and natural variation of cx-amylase
isozymes in barley. Planta 157: 493-501.

Bruni, F. and A.C. Leopold. 1991. Glass transitions in soybean seed: Relevance to anhydrous
biology. Plant Physiol. 96: 660-663.

Bush, D.S., A.K. Biwas and R.L. Jones. 1993. Hormonal regulation of Ca2+ transport in the
endomembrane system of the barley aleurone. Planta 189: 507-515.

Carson, C.B. 1992. Characterization of the structure and expression of the mutant alleles of the
viviparous-1 gene of maize. U.S. dissertation, University of Florida.


91






92


Chandler, P.M. 1988. Hormonal regulation of gene expression In the slendero mutant of barley
(Hordeum vulgare L.). Planta 175: 115-120.

Chandler, P.M. and J.V. Jacobsen. 1991. Primer extension studies on ci-amylase mRNAs in
barley aleurone. II. Hormonal regulation of expression. Plant Mol. Biol. 16: 637-645.

Chandler, P.M. and M. Robertson. 1994. Gene expression regulated by abscisic acid and its
relation to stress tolerance. Ann. Rev. Plant Physiol. Mol. Biol. 45: 113-141.

Chao, S.E. and J.G. Scandalios. 1969. Identification and genetic control of starch-degrading
enzymes in maize endosperm. Biochem. Genet. 3: 537-547.

Chao, S.E. and J.G. Scandalios. 1971. Alpha-amylase of maize: Differential allelic expression
at the Amy-1 gene locus, and some physiological properties of the isozymes. Genetics
69: 47-61.

Chao, S.E. and J.G. Scandalios. 1972. Developmentally dependent expression of tissue
specific amylases in maize. Mol. Gen. Genet. 115: 1-9.

Chen, S.M. and E.H. Coe Jr. 1978. Control of anthocyanin synthesis by the C locus in maize.
Biochem. Genet. 15: 333-346.

Chen, Y. and J.S. Burris. 1990. Role of carbohydrates in desiccation tolerance and membrane
behavior in maturing maize seed. Crop Sci. 30: 971-975.

Coe Jr., E.H. and M.G. Neuffer. 1977. The genetics of com. In: Corn and Com Improvement,
Ed. 2, G.F. Sprague ed., American Society of Agronomy, (Madison, WI), pp. 111-223.

Coe Jr., E.H., M.G. Neuffer and D.A Hoisington. 1988. The genetics of com. In: Com and Com
Improvement, G.F. Sprague and J.W. Dudley, eds, American Society of Agronomy
(Madison, WI), pp. 81-258.

Comai, L. and J.J. Harada. 1990. Transcriptional activities in dry seed nuclei indicate the timing
of the transition from embryogeny to germination. Proc. Nati. Acad. Sci. USA 87: 26712674.

Comford, C.A., M. Black, J.M. Chapman and D.C. Baulcombe. 1986. Expression of c-amylase
and other gibberellin-regulated genes in aleurone tissue of developing wheat grains.
Plant 169: 420-428.

Crispeels, M.J. and J.E. Varner. 1966. Inhibition of gibberellic acid induced formation of cxamylase by abscision 11. Nature 212: 1066-1067.

Croker, S.J., P. Hedden, J.R. Lenton and J.L. Stoddart. 1990. Comparison of gibberellins in
normal and slender barley seedlings. Plant Physiol. 94: 194-200.

Crowe, J.H., F.A. Hoekstra and L.M. Crowe. 1992. Anhydrobiosis. Ann. Rev. Physiol. 54:579599.

Derkx, M.P.M. and C.M. Karssen. 1993. Effects of light and temperature on seed dormancy and
gibberellin-stimulated germination in Arabidopsis thaliana: studies with gibberellindeficient and -insensitive mutants. Physiol. Plant. 89: 360-368.






93


Dickinson, C.D., R.P. Evans and N.C. Neilsen. 1988. RY repeats are conserved in the 5"flanking regions of legume see protein genes. Nucleic Acids Res. 16: 371-378.

Dooner, H.K. 1985. Viviparous-1 mutation in maize conditions pleiotroplc enzyme deficiencies
in the aleurone. Plant Physiol. 77: 486-488.

Dooner, H.K., T.P. Robbins and R.A. Jorgensen. 1991. Genetic and developmental control of
anthocyanin biosynthesis. Ann. Rev. Genet. 25: 173-199.

Duckham, S.C., R.S.T. Unforth and I.B. Taylor. 1991. Abscisic acid-deficient mutants at the
aba gene locus of Arabidopsis thaliana are impaired in the epoxidation of zeaxanthin.
Plant Cell Envir. 14: 601-606.

Dure, L.S. 1960a. Gross nutritional contributions of maize endosperm and scutellum to
germination growth of maize axis. Plant Physiol. 35: 919-925.

Dure, L.S. 1960b. Site of origin and extent of activity of amylases in maize germination. Plant
Physiol. 35: 925-934.

Dure IlIl, L. 1993. A repeating 1 1-mer amino acid motif and plant desiccation Plant J. 3: 383369.

Dure Ill, L., M. Crouch, J. Harada, T.H.D. Ho, J. Mundy, R. Quatrano, T. Thomas and Z.R. Sung.
1989. Common amino acid sequence domains among the LEA proteins of higher plants.
Plant Mol. Biol. 12: 475-486.

Evans, M., M. Black and J. Chapman. 1975. Induction of hormone sensitivity by dehydration is
one positive role for drying in cereal seed. Nature 258: 144-145.

Eyster, W.H. 1931. Vivipary in maize. Genetics 16: 574-590.

Feldmann, K.A. 1991. T-DNA insertion mutagenesis in Arabidopsis: Mutational spectrum. Plant
J. 1: 71-82.

Fick, G.N. and C.O. Qualset. 1975. Genetic control of endosperm amylase activity and
gibberellic acid responses in standard-height and short-statured wheats. Proc. NatI.
Acad. Sci. USA 72: 892-895.

Filner, P., and J.E. Vamer. 1967. A test for de novo synthesis of enzymes: density labeling with
H2018 of barley a-amylase induced by gibberellic acid. Proc. Natl. Acad. Sci. USA
58:1520-1526.

Fincher, G.B. 1989. Molecular and cellular biology associated with endosperm mobilization in
germinating cereal grains. Ann. Rev. Plant Physiol. Mol. Biol. 40: 305-346.

Finkelstein, R.R. 1994. Mutations at two new Arabidopsis ABA response loci are similar to the
abi3 mutations. Plant J. 5: 765-771.

Finkelstein, R.R. and C. Somerville. 1990. Three classes of abscisic acid (ABA)-insensitive
mutations of Arabidopsis define genes that control overlapping subsets of ABA
responses. Plant Physiol. 94: 1172-1179.






94


Foster, CA. 1977. Slender an accelerated extension growth mutant of barley. Barley Genet.
Newslett. 7: 24-27.

Fujioka, S., H. Yamane, C.R. Spray, M. Katsuml, B.O. Phinney, P. Gaskin, J. MacMillan and N.
Takahashi. 1988. The dominant non-gibberellin-responding dwarf mutant (DO) of maize
accumulates native gibberellins. Proc. Nati. Acad. Sci. USA 85: 9031-9035.

Galau, G.A., N. Bijaisordat and D.W. Hughes. 1987. Accumulation kinetics of cotton late
embryogenesis-abundant mRNAs and storage protein mRNAs: Coordinate regulation
during embryogenesis and the role of abscisic acid. Dev. Biol. 123: 198-212.

Galau, G.A., D.W. Hughes and L. Dure Ill. 1986. Abscisic acid induction of cloned cotton late
embryogenesis-abundant (Lea) mRNAs. Plant Mol. Blol. 7:155-170.

Galau, G.A., K.S. Jakobsen and D.W. Hughes. 1991. The controls of late dicot embryogenesis
and early germination. Physiol. Plant. 81: 280-288.

Gale, M.D. and GA. Marshall. 1975. The nature and genetic control of gibberellin insensitivity
in dwarf wheat grain. Heredity 35: 55-65.

Garcia-Maya, M., J.M. Chapman and M. Black. 1990. Regulation of a-amylase formation and
gene expression In the developing wheat embryo. Plant 181: 296-303.

George, D.W. 1967. High temperature seed dormancy in wheat (Triticum aestivum L.) Crop
Sci. 7: 249-253.

Gilroy, S. and R.L. Jones. 1992. Gibberellic acid and abscisic acid coordinately regulate
cytoplasmic calcium and secretory activity in barley aleurone protoplasts. Proc. NatI.
Acad. Sci. USA 89: 3591-3595.

Gilroy, S. and R.L. Jones. 1993. Calmodulin stimulation of unidirectional calcium uptake by the
endoplasmic reticulum of barley aleurone. Plant 190: 289-296.

Giraudat, J., B.M. Hauge, C. Valon, J. Smalle, F. Parcy and H. Goodman. 1992. Isolation of the
Arabidopsis ABI3 gene by positional cloning. Plant Cell 4: 1251-1261.

Goldman, S., Y.R. Mawal, I. Tanida and R Wu. 1994. Studies of a gibberellin-dependent DNAbinding protein related to the expression of a rice a-amylase gene. Plant Sci. 99: 75-88.

Goldstein, L.D. and P.H. Jennings. 1978. Amylase enzymes isolated from incubated deembryonated maize kernels. New Phytol. 81: 233-242.

Gubler, F., and J.V. Jacobsen. 1992. Gibberellin-responsive elements in the promoter of a
barley high-pi (%-amylase gene. Plant Cell 4: 1435-1441.

Guiltinan, M.J., W.R. Marcotte, Jr., and R.S. Quattrano. 1990. A plant leucine zipper protein that recognizes an abscisic acid response element. Science 250: 267-271.

Harberd, N.P. and M. Freeling. 1989. Genetics of dominant gibberellin-insensitive dwarfism in maize. Genetics 121: 827-838.

Harvey, B.M.R. and A. Oaks. 1974. The role of gibberellic acid in the hydrolysis of endosperm reserves in Zea mays. Plant 121: 67-74.






95


Hattori, T., T. Terada and S.T. Hamasuna. 1994. Sequence and functional analyses of the rice
gene homologous to the maize Vpl. Plant Mol. Biol. 24: 805-810.

Hattori, T., V. Vasil, L Rosenkrans, L.C. Hannah, D.R. McCarty and L.K. Vasil. 1992. The
Viviparous-1 gene and abscisic acid activate the C1 regulatory gene for anthocyanin
biosynthesis during seed maturation in maize. Genes and Dev. 6: 609-618.

Higgins, T.J., JA. Zwar and J.V. Jacobsen. 1976. Gibberellic acid enhances the level of
translatable mRNA for a-amylase in barley aleurone layers. Nature 260: 166-169.

Hilhorst, H.W.M. and C.M. Karssen. 1992. Seed dormancy and germination: the role of abscisic
acid and gibberellins and the importance of hormone mutants. Plant Growth Reg. 11:
225-238.

Hilimer, S., S. Gilroy and R.L. Jones. 1992. Visualizing enzyme secretion from individual barley
(Hordeum vulgare) aleurone protoplasts. Plant Physiol. 102: 279-286.

Hoekstra, FA., L.M. Crowe and J.H. Crowe. 1989. Differential desiccation sensitivity of com
and Pennisetum pollen linked to their sucrose contents. Plant Cell Envir. 12: 83-91.

Hoekstra, FA. and T. van Roekel. 1988. Desiccation tolerance of Papaver dubium during its
development in the anther Possible role of phospholipid composition and sucrose
content. Plant Physiol. 88: 626-632.

Hooley, R., M.H. Beale and S.J. Smith. 1991. Gibberellin perception at the plasma membrane
of Avena fatua aleurone protoplasts. Planta 183: 274-280.

Hughes, D.W., and G.W. Galau. 1991. Developmental and environmental induction of Lea and
LeaA mRNAs and the postabscission program during embryo culture. Plant Cell 3: 605618.

Huttly, A.K. and D.C. Baulcombe. 1989. A wheat a-Amy2 promoter is regulated by gibberellin in
transformed oat aleurone protoplasts. EMBO J. 8: 1907-1913.

Huttly, A.K., A.L. Phillips and J.W. Tregear. 1992. Localization of cis elements in the promoter
of a wheat a-amy2 gene. Plant Mol. Biol. 19: 903-911.

Ingle, J. and R.H. Hageman. 1965. Metabolic changes associated with the germination of com.
IlIl. Effects of gibberellic acid on endosperm metabolism. Plant Physiol. ??: 672-675.

Jacobsen, J.V. and L.R. Beach. 1985. Control of transcription of c-amylase and rRNA genes in
barley aleurone protoplasts by gibberellin and abscisic acid. Nature 316: 275-277.

Jacobsen, J.V. and P.M. Chandler. 1987. Gibberellin and abscisic acid in germinating cereals.
In Plant Hormones and Their Role in Plant Growth and Development (ed. P.J. Davies),
Dordrecht: Martinus Nijhoff. pp. 164-193.

Jacobsen, J.V. and T.J. Close. 1991. Control of transient expression of chimaeric genes by
gibberellic acid and abscisic acid in protoplasts prepared from mature barley aleurone
layers. Plant Mol. Biol. 16: 713-724.




Full Text
12
confirming that a strong abi3 allele conferring high insensitivity to ABA is sufficient to cause
vivipary (Nambara et al., 1992; Ooms et al., 1993).
The Abi3 gene was cloned by chromosome walking (Giraudat et al., 1992). The
predicted protein of 79.5 kD displays discrete regions of sequence homology to the maize VP1
protein. Since there are also phenotypic similarities between the abi3 and vp1 mutant at least
with respect to seed-specific insensitivity to ABA ABI3 and VP1 are likely to have similar
functions in regulating ABA response in the seed. However, the functions of ABI3 and VP1 differ
in so far that ABI3 is required for seed dormancy in Arabidopsis while VP1 does not Impose seed
dormancy in maize. Conversely, VP1, but not ABI3, induces synthesis of anthocyanins in the
seed. Whether these phenotypic differences reflect differences in sequence between ABI3 and
VP1 or the differential involvement of other factors remains to be determined.
The Abi 1 gene was doned in two laboratories by chromosome walking (Leung et al.,
1994; Meyer et al., 1994). At its C-terminus, the predided ABM protein (47.5 kD) displays
sequence similarity with the 2C dass of serine-threonine protein phosphatases from rat and
yeast. Its N-terminus exhibits features typical for a Ca++-binding site (EF hand). Hence, ABI1
may fundion as a Ca++-dependent protein phosphatase. Indeed, regulation of stomatal aperture
by ABA involves Ca++ as a second messenger and protein phosphorylation events (Blatt and
Thiel, 1993; Luan et al., 1993). How ABI1 may regulate ABA-indudion of seed dormancy is thus
far unknown.
Mutants affeded in a thus far unknown mechanism
Three mutants of Arabidopsis (Iec1, tec2, fus3) have been isolated that are non-donmant
but normal in their response to ABA (Meinke, 1992; Keith et al., 1994; Meinke et al., 1994;
Bumlein et al., 1994). Iec1 and fus3 have similar phenotypes. Immature mutant seeds
germinate readily when placed in culture and display occasional vivipary when left to mature in
the siliques. Furthermore, they are intolerant to desiccation and accumulate anthocyanins late in
seed development, a feature that is not charaderistic of wild-type seeds. Prematurely
germinated seeds give rise to viable green seedlings that appear normal except that trichomes


80
In summary, deletion analysis of the VP1 protein has allowed the Identification of several
domains that are essential for repressor function: 1) the conserved, highly charged RED domain
85/87 and 2) the poorly conserved region 87/88. Moreover, though not essential, the domain
103/104 and the C-terminus may play a role in repression.
When comparing the repressor domains of VP1 with the domains required for
transcriptional activation of the Em or C1 genes, it is evident that different functions of VP1 map
to different sequences in the protein (Fig. 26). Apart from the differential requirement of the
acidic activation domain at the N-terminus of VP1, a-amylase repression and Em activation
differed in the need for the highly positively charged domain 103/104 of VP1. While deletion of
103/104 reduced Em-GUS activation by 98% in maize protoplasts (L. Rosenkrans et al.,
unpublished results), it did not severely affect repression of Amy-GUS. Likewise, a-amylase
repression and C1 activation displayed a differential requirement for the C-terminal part of VP1.
Deletion of the C-terminal ca. 150 amino acids of VP1 entirely eliminated activation of C1-Sh-
GUS (L. Rosenkrans et al., unpublished results). In contrast, this domain was not found
essential for repression of Amy-GUS.
Similarly, different domains of VP1 are essential for activation of C1 and activation of
Em. Overall, sequences required for induction of anthocyanin biosynthesis map to the highly
conserved C-terminal end of VP1, while sequences essential for activating Em are located in the
central part of VP1 (85/87, 87/88 and 103/104). These findings are consistent with the
phenotype of mutant alleles in maize and Arabidopsis. The maize vp1-McW allele truncating the
C-terminal ca. 150 amino acids of VP1 produces seeds exhibiting nearly normal developmental
arrest but lack of anthocyanin accumulation, indicating that the C-terminal part of VP1 is not
essential for preventing vivipary but is essential for anthocyanin production (McCarty et al.,
1989b). In Arabidopsis, two abi-3 alleles that produce a viviparous phenotype have been
sequenced: ab/3-6 contains an internal deletion of ca. 750 bp between positions 1,073 and 1,944
(Nambara et al., 1994), thus deleting the domains corresponding to 196/88 and 103/104. The
mutation in ab/3-3 induces a premature stop codon at Gin417 (Giraudat et al., 1992), thus deleting


65
c.
Effector Construct
Rel. Amy-GUS / LUC
Maize Barley
35S-Sh-CAT (Control)
100<25>
100 <417)
activator
WT-VP1
i ta mm 1
i mmmS1
16(2)
42 (8>
^28-121
I M I
1
|
&
13(4)
17<>
3x (VP16 act)
mm
m wmm1
I n.d.
12 (t7)
Fig. 17. Mode of action of VP1 in repressing Amy-GUS.
(A),(B). Alternative models for VP1 action as described in the text.
(C). Effect of deletion and substitution derivatives of VP1 on Amy GUS expression in aleurone
tissue of germinating maize and barley seeds cultured in GA3. In A28-121, the activation
domain of VP1 was deleted. In 3x (VP16 act), the activation domain of VP1 was replaced by
three copies of the Herpes simplex VP16 activation domain. Data represent activities (mean
S.E.M) relative to control (=100). Black boxes show sequence homology between VP1 and
ABI3. (n.d.: not determined).


29
There is little concept of how the perceived GA and ABA signals are transduced from the
receptors to the nucleus. The function of Ca** and calmodulin (CaM) as second messengers
that regulate protein kinase activity has been characterized in many animal and a few plant
systems (Roberts and Harmon, 1992; Neuhaus et al.. 1993). Also in barley aleurone protoplasts,
treatment with GA was found to increase cytoplasmic Ca~ and CaM concentrations (Gilroy and
Jones, 1992, 1993). ABA reversed the effect of GA on (Ca~], (Gilroy and Jones, 1992). Even
though the increase in [Ca~], and [CaM] preceded the GA-induced increase in a-amylase
activity by 2-4 h (Gilroy and Jones, 1992), direct evidence for an involvement of Ca++ and CaM
in the regulation of a-amylase transcription is still lacking. Ca++ and CaM have been found to
regulate the activity of a slow vacuolar ion channel located in the tonoplast of storage protein
vacuoles in barley aleurone cells (Bethke and Jones, 1994). Moreover, CaM was shown to
stimulate Ca++ uptake into the ER where the Ca~ containing a-amylase enzyme is synthesized
(Bush et al.. 1993). Thus, hormone-regulated changes in Ca** and CaM concentrations may be
regulating processes such as a-amylase formation and secretion rather than having a direct
effect on the transcription of a-amylase genes.
Analysis of mutants provides a valuable tool to study the genetics underlying the
regulation of hormone action. Many mutants have been isolated that display altered responses
to GA, suggesting that these mutants are affected in a component of GA signal transduction.
Since GA promotes stem and leaf elongation, these mutants have been identified by their altered
plant height. They fall into two classes: 1) those that show a reduced sensitivity to GA (*GA-
insensitive mutants') and are therefore of dwarf stature, and 2) those that show an enhanced
sensitivity to GA ("constitutive response mutants') and are therefore excessively tall. In
response mutants, the concentrations of biologically active GAs do not in accord with the
phenotype. Generally, GA-insensitive mutants accumulate higher concentrations of active GAs
as compared to wild-type, while tissues of constitutive response mutants contain reduced GA
concentrations (Stoddart, 1984; Fujioka et al, 1988; Croker et al., 1990). These observations


59
This indicates that VP1 is likely to act further downstream in, or independently of, the SLN
pathway.
While GA is a strong inducer of a-amylase genes in aleurones of Himalaya barley and
other cereals, the importance of GA in the regulation of maize a-amylase genes is less clear, a-
amylase activities were found high in isolated endosperms that had been de-embryonated prior
to imbibition (Harvey and Oaks 1974; Goldstein and Jennings, 1978). Moreover, application of
exogenous GA to isolated endosperms did not further enhance a-amyiase activities (Oishi and
Black, 1990). Hence, it was argued that mature endosperms store high concentrations of GA
(Harvey and Oaks 1974; Goldstein and Jennings, 1978; Oishi and Black, 1990). However, seeds
of the GA-deficient, extremely dwarfed mutant dS displayed considerable a-amylase activity that
was only 3-fold lower than in wild-type seeds, implying a GA-independent component in maize a-
amylase production. To investigate this, Amy-GUS was introduced into aleurones of germinating
wild-type and GA1-deficient d1 mutant seeds. Both genotypes displayed similar, high Amy-GUS
activities in the absence of exogenous GA (Figs. 13, 14). Furthermore, application of GA to d1
mutant seeds increased Amy-GUS expression by less than two-fold, thus to a similar extent as in
wild-type seeds (Figs.13, 14). Hence, the d1 mutation did not appear to alter Amy-GUS
expression in the aleurone, suggesting that deficiency in the highly active gibberellin GA1 does
not severely affect high-level expression of a-amylase genes. Thus, consistent with the data on
the mutant d5 (Harvey and Oaks, 1974), the possibility of a constitutive activity of a-amylase
genes in the absence of GA appears likely. The finding that co-expression of VP1 and
application of ABA reduced Amy-GUS activity to a very low level (Figs. 13, 14) indicates that
VP1 and ABA repress the GA-dependent as well as the putative constitutive activity of Amy-
GUS in maize.


86
investigated further, it is likely that GA, ABA and VP1 are three among several factors that
regulate the activity of constituents required for expression of a-amylase genes in the maize
seed (Fig. 27C).
Expression experiments in developing maize seeds have shown that VP1 and ABA are
likely to act independently in repressing Amy-GUS. Further evidence regarding the relative
positions of VP1 and ABA in the regulatory network allowed the characterization of the dominant
negative effect caused by the deletion mutant 87/88. Co-expression of 87/88 with wild-type VP1
and Amy-GUS has shown that 87/88 inhibits repression of Amy-GUS by VP1 ("dominant
negative effect*, Fig. 22). This suggests that the observed 87/88-mediated activation of Amy-
GUS in the absence of GA (Fig. 20) may be caused by the presence of residual amounts of
endogenous barley VP1 homolog in the wild-type aleurones. Because ABA was incapable of
inhibiting the 87/88-mediated activation of Amy-GUS (Fig. 21), it is suggestive that VP1
functions either downstream of ABA or via a different signalling pathway than ABA. It therefore
will be interesting to map the cis-elements in a-amylase promoters that are responsible for VP1
and ABA action.
The Role of the Embrvo
Although the vp1 mutant phenotype was cell autonomous within the aleurone of vp1-w2
seeds (Fig. 5a,b), de-repression of a-amylase genes was not fully independent of the
physiological state of the embryo: 1) in vp1-m2 seeds, precocious hydrolization of endosperm
reserves in sectors underlying vp1 mutant aleurone was predominantly observed in seeds
carrying a viviparous embryo. 2) Amy-GUS was de-repressed in vp1 mutant aleurone of
concordant vp1 mutant seeds but not in vp1 mutant aleurone of non-concordant seeds exhibiting
a wild-type embryo (Fig. 15). These observations suggest that a viviparous embryo facilitates
expression of a-amylase genes in vp1 mutant aleurone. However, the finding that Amy-GUS is
highly induced in vp1 mutant aleurone of germless seeds (Fig. 16) indicates that a viviparous


82
the domain 103/104 and the C-terminus.
VP1 shows no significant sequence homology to other proteins. Therefore, the function
of domains other than the acidic activation sequence is thus far unknown, it was suggested that
the region from amino acid 208 to 235 of the rice VP1 may form a leucine zipper-like structure
(Hattori et al., 1994). In this region, Leu or lie residues are located at every seventh residue (5
repeats) with an exception for the fourth position (see Fig. 25). However, no severe loss-of-
function phenotype has been observed when this domain was deleted in the 86/85 construct.
86/85 effected full repression of Amy-GUS and ca. 50% of wild-type activity with respect to
activation of Em-GUS. These data do not support an important role of this domain in protein-
protein interactions.
Based on the evidence that VP1 transcriptionally activates C1 and Em (McCarty et al.,
1991; Hattori et al., 1992), nuclear targeting is likely to be a requirement for VP1 function. A
100% conserved putative nuclear localization sequence (NLS), RKKR, exists in the domain from
amino acid 392 to 395 of VP1 (see Fig. 25), as mentioned by Giraudat et al. (1992). Consistent
with these views, deletion of this putative NLS (construct 103/104) fully eliminated activation of
Em-GUS (L. Rosenkrans et al., unpublished results). However, 103/104 retained capacity to
activate C1-Sh-GUS (31% of wild-type VP1), suggesting that the mutant protein is targeted to
the nucleus though possibly with reduced efficiency. Assuming that RKKR is a functional NLS,
these data indicate that VP1 contains two or more NLSs with at least partially redundant function,
a feature not uncommon among nuclear proteins (Raikhel, 1992). The differential effects of
103/104 on activation of Em and C1 may reflect different threshold levels of VP1 protein
required for these activator functions. Hence, the extent of nuclear localization of 103/104 may
be sufficient for partial activation of C1 but not for activation of Em. Alternatively, 103/104 may
serve an additional function in activation of Em. A dual role of a domain in both nuclear
targeting and DNA binding was reported for the regulatory protein 02 (Varagona et al., 1994).
With respect to repression of Amy-GUS by VP1, it is unknown whether nuclear
localization of VP1 is required for function. The construct 103/104 was not affected in repression


as an apparent GA-independent activity. Deletion of the addle transcriptional activation domain of VP1
did not affect the inhibitory activity, indicating that VP1 has a discrete repressor function. Further
deletion analysis of VP1 showed that domains essential for repression of Amy-GUS are distinguishable
from domains required for activation of the maturation-related genes Em and C1.
The role of the embryo in the expression of Amy-GUS in developing maize aleurone cells was
studied. Amy-GUS was de-re pressed in vp1 mutant aleurone in seeds that either carried a viviparous
embryo or aborted the embryo earty in development but not in seeds with a normal, non-viviparous
embryo. This suggests that a normal embryo contributes a diffusible signal with inhibitory effect on Amy-
GUS expression in the aleurone. Amy-GUS was partially de-re pressed in wild-type aleurone cells of
embryo-less seeds, suggesting that both Vp1 expression in the aleurone and a non-viviparous embryo
are required for complete repression of a-amylase genes in the developing maize aleurone.
v


55
Fig. 9: VP1 dose-response for repression of Amy-GUS in aleurone of germinating barley seeds.
0 to 10 pg of 35S-Sh-VP1 were co-precipitated with 0.5 pg of Amy-GUS and 5 pg of Ubi-LUC.
Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment,
endosperms were cultured in 10*6 M GA3. Data represent mean ( S.E.M) of five replicates.
Amount of 35S-Sh-VP1 added (ug)
Fig. 10: Co-expression of VP1 activated Em-GUS in barley aleurone. Aleurones of germinating
barley seeds were bombarded with 2 pg of Em-GUS, 5 pg of Ubi-LUC and 0, 1 or 5 pg of 35S-
Sh-VP1. Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment,
endosperms were cultured in no hormones. Data represent mean ( S.E.M) of five replicates.


50
Amy-GUS / LUC *104 [pmoles MU/h/RLU]
vp1-R mutant Aleurones
Wild-type Aleurones
Days after
Pollination
-GA
+GA
-GA
+GA
Range
Mean S.E.M.
Range Mean S.E.M.
Mean
Mean
18
<1
<1
<1
<1
20
<1
4-35 21
7
<1
<1
24
114-531
263 130
41-150 110
35
<1
<1
Table 1. Amy-GUS is inducible in vp1-R mutant aleurone cells but not in wild-type aleurone
cells. Aleurones of developing vp1-R mutant and wild-type kernels at 18, 20 and 24 DAP were
bombarded with a mixture of 10 pg of Amy-GUS and 5 pg of Ubi-LUC. Post-bombardment,
kernels were treated with a solution containing no hormones or 10' M GA3. Data represent
mean ( S.E.M) of three to five replicates.


31
A constitutive response mutant has been identified in barley (Foster, 1977). This
recessive mutation, termed slender" (sin), causes a plant to appear as if It had been treated with
high doses of GA (Lanahan and Ho, 1988; Croker et al., 1990). Also, a-amylase genes were
highly expressed in sin mutant half grains in the absence of applied GA (Chandler, 1988). Thus,
the absence of functional SLN protein causes constitutive expression of GA-responses and
thereby uncouples transcription of GA-regulated genes from a need for GA. This phenotype
suggests that Sin encodes a negative regulator of GA-response. Importantly, a-amylase
production in sin mutant aleurones was susceptible to inhibition by ABA, indicating that the sin
mutant retains normal sensitivity to ABA (Chandler, 1988; Lanahan and Ho, 1988). Hence, ABA
most probably functions at a step downstream of GA in the signal transduction pathway leading
to regulation of a-amylase transcription. A function of ABA fully independent of GA cannot be
ruled out but is unlikely because GA and ABA appear to act through the same response
elements in a-amylase promoters. The findings also indicate that GA and ABA do not act at the
same site in the signal transduction pathway, i.e. they do not for example antagonistically
phosphorylate/ de-phosphorylate an intermediate.
The Developmental Switch from Seed Maturation to Seed Germination
Desiccation is the normal terminal event in seed development, leading to a state of
metabolic quiescence. In many species (e.g. maize, bean), hydration of the mature, dry seed is
sufficient to initiate germination. Thus, in these seeds (termed quiescent or non-dormant seeds),
the transition from seed maturation to germination is associated with the reversal of the
desiccated state. Seeds of other species (e.g. cereals, Arabidopsis) develop dormancy during
late stages of seed development. In these species, freshly harvested mature seeds do not
germinate following imbibition but require a treatment such as light, low temperature or after
ripening (dry storage) to overcome the state of dormancy and allow induction of germination.


93
Dickinson, C.D., R.P. Evans and N.C. Neilsen. 1988. RY repeats are conserved in the 5-
flanking regions of legume see protein genes. Nucleic Acids Res. 10: 371-378.
Dooner, H.K. 1985. Vtviparous-1 mutation in maize conditions pieiotropic enzyme deficiencies
in the aieurone. Plant Physiol. 77:486-488.
Dooner, H.K., T.P. Robbins and R.A. Jorgensen. 1991. Genetic and developmental control of
anthocyanin biosynthesis. Ann. Rev. Genet. 25:173-199.
Duckham, S.C., R.S.T. Linforth and I.B. Taylor. 1991. Abscisic acid-deficient mutants at the
aba gene locus of Arabidopsis thaliana are impaired in the epoxidation of zeaxanthin.
Plant Cell Envir. 14: 601-606.
Dure, L.S. 1960a. Gross nutritional contributions of maize endosperm and scutellum to
germination growth of maize axis. Plant Physiol. 35: 919-925.
Dure, L.S. 1960b. Site of origin and extent of activity of amylases in maize germination. Plant
Physiol. 35: 925-934.
Dure III, L. 1993. A repeating 11-mer amino acid motif and plant desiccation Plant J. 3:363-
369.
Dure III, L, M. Crouch, J. Harada, T.H.D. Ho, J. Mundy, R. Quatrano, T. Thomas and Z.R. Sung.
1989. Common amino acid sequence domains among the LEA proteins of higher plants.
Plant Mol. Biol. 12: 475-486.
Evans, M., M. Black and J. Chapman. 1975. Induction of hormone sensitivity by dehydration is
one positive role for drying in cereal seed. Nature 258:144-145.
Eyster, W.H. 1931. Vivipary in maize. Genetics 16: 574-590.
Feldmann, K.A. 1991. T-DNA insertion mutagenesis in Arabidopsis: Mutational spectrum. Plant
J. 1:71-82.
Fick, G.N. and C.O. Qualset. 1975. Genetic control of endosperm amylase activity and
gibberellic acid responses in standard-height and short-statured wheats. Proc. Natl.
Acad. Sci. USA 72: 892-895.
Filner, P., and J.E. Varner. 1967. A test for de novo synthesis of enzymes: density labeling with
H2O18 of barley a-amylase induced by gibberellic acid. Proc. Natl. Acad. Sci. USA
58:1520-1526.
Fincher, G.B. 1989. Molecular and cellular biology associated with endosperm mobilization in
germinating cereal grains. Ann. Rev. Plant Physiol. Mol. Biol. 40: 305-346.
Finkelstein, R.R. 1994. Mutations at two new Arabidopsis ABA response loci are similar to the
ab/3 mutations. Plant J. 5: 765-771.
Finkelstein, R.R. and C. Somerville. 1990. Three classes of abscisic acid (ABA)-insensitive
mutations of Arabidopsis define genes that control overlapping subsets of ABA
responses. Plant Physiol. 94:1172-1179.


20
mutant aleurone cells by particle bombardment complemented the failure to accumulate
anthocyanins in a cell autonomous fashion (Hattorl et al., 1992). Thus, lack of C1 appears to be
responsible for the block in anthocyanln synthesis in vp1 mutant seed. A direct role of VP1 in
activating C1 expression was concluded from the demonstration that over-expression of VP1 in
maize protoplasts activated transcription of a C1 promoter-GUS fusion gene. Most importantly,
VP1 function in activating C1 was dependent on its transcriptional activation domain (Hattori et
al., 1992). Hence, VP1 and C1 are part of a regulatory hierarchy controlling activation of
anthocyanin structural genes.
Several lines of evidence indicate that VP1's function in activating C1 is distinct from its
function in activating Em. First, though induction of both genes is dependent on the acidic
activation domain of VP1 and thus appears to involve transcriptional activation, other domains of
VP1 involved in function differed depending on the target promoter. While sequences in the
middle of the VP1 protein were required for trans-activation of Em-GUS in maize protoplasts,
VP1 -activation of C1-GUS was dependent on the C-terminal end of VP1 (L. Rosenkrans, V.
Vasil, I.K. Vasil and D.R. McCarty, unpublished results). This is consistent with the non-
viviparous/unpigmented phenotype of the vp1-McWhirter mutant which produces a truncated
VP1 protein lacking ca. 150 bp from the C-terminus (McCarty et al., 1989b).
Second, in agreement with the involvement of distinct domains of VP1, different cr's-
elements in the C1 and Em promoters appear to be the target of VP1 function. In contrast to
activation of Em which depended on two G-box sequences, activation of C1 did not require any
of the two G-box-like sequences present in the promoter but the 13 bp sequence -145
TCCATGCATGCAC -158 (Hattori et al., 1992). This sequence, designated as Sph-element, is
found in promoters of other seed-specific genes (Dickinson et al., 1988).
Finally, VP1 -mediated activation of Em and C1 differ in their interaction with ABA.
While there is a synergistic effect of VP1 and ABA in activating Em, the role of ABA in VP1-
mediated activation of C1 is less clear. Anthocyanins accumulate at normal levels in ABA-
deficient mutants of maize, suggesting that C1 expression is hormone-independent. On the


83
of Amy-GUS. This may indicate that nuclear targeting of VP1 is not required or that other
domains with redundant function may compensate in function for the deletion in 103/104.
Interestingly, the double-deletion mutant 87/103 lost the ability to exert a dominant negative
effect on Amy-GUS repression by VP1. A failure of the double mutant protein to be targeted to
the nucleus would be consistent with the observed loss of function.
Apart from the activation domain, the biochemical functions of sequences in the VP1
protein remain unclear. However, the deletion analysis demonstrated very clearly that different
domains are involved in the different functions of VP1, thus underlining the multifunctional
nature of this transcription factor.
Interactions Between VP1 and Plant Hormones
One can envision at least three models of how VP1 might function in repressing a-
amylase genes: 1) VP1 might mediate ABA antagonism of GA signalling during seed
development. ABA is known to antagonize GA-action in the regulation of a-amylase genes in
germinating cereal seeds (Jacobsen and Chandler, 1987). Because VP1 is required for ABA-
induced gene expression associated with seed maturation (McCarty et al 1991), it might also be
essential in ABA-mediated repression of a-amylase genes (Fig. 27A). Consequently, the vp1
mutant might allow de-repression of a-amylase genes by failing to respond to ABA present in the
developing seed. 2) VP1 might specifically inhibit the GA-response pathway independently of
ABA (Fig. 27B). 3) VP1 might repress a-amylase genes via a pathway that functions
independently of both GA and ABA (Fig. 27C).
The results presented in this work do not support the first model. ABA was effective in
blocking Amy-GUS expression in vp1 mutant aleurone cells (Fig. 6), indicating that ABA action in
this instance does not depend on the presence of VP1. In combination, VP1 and ABA effects
were roughly additive. This stands in contrast to evidence showing that VP1 is required for ABA-
induced expression of the maize Em gene (McCarty et al., 1991). Thus, there appear to be at


102
Taylor, M.G. and I.K. Vasil. 1991. Histology of, and physical factors affecting transient GUS
expression in pearl millet (Pennisetum glaucum (L.) R. Br.) embryos following
microprojectile bombardment. Plant Cell Rep. 10:120-125.
Tsai, M. and B.W. OMalley. 1994. Molecular mechanisms of action of sterokl/thyroid receptor
superfamily members. Ann. Rev. Biochem. 03: 451-486.
Varner, J.E. and G.R. Chandra. 1964. Hormonal control of enzyme synthesis in barley
endosperm. Proc. Natl. Acad. Sci. USA 52:100-106.
Varagona, M.J. and N. Raikhel. 1994. The basic domain in the bZIP regulatory protein
Opaque2 serves two independent functions: DNA binding and nuclear localization.
Plant J. 5: 207-214.
Vasil, V., M. Clancy, R.J. Fed, I.K. Vasil and L.C. Hannah. 1989. Increased gene expression by
the first intron of the maize sh1 locus in grass species. Plant Physiol. 91:1575-1579.
Walker-Simmons, M. 1987. ABA levels and sensitivity in developing wheat embryos of
sprouting resistant and susceptible cultivare. Plant Physiol. 84: 61-66.
Walker-Simmons, M. 1988. Enhancement of ABA responsiveness in wheat embryos by high
temperature. Plant Cell Envir. 11: 769-775.
Warner, DA, M.J. Grove and CA Knutson. 1991. Isolation and characterization of a-
amylases from endosperm of germinating maize. Cereal Chem. 68: 383-390.
Warner, D.A. and CA Knutson. 1991. Isolation of a-amylases and other starch degrading
enzymes from endosperm of germinating maize. Plant Science 78:143-150.
West, M.A., K.M.Yee, J. Danao, J.L. Zimmerman, R.L. Fischer, R.B. Goldberg and J.J. Harada.
1994. Leafy cotyledon 1 is an essential regulator of late embryogenesis and cotyledon
identity in Arabidopsis. Plant Cell 6:1731-1745.
Wheeler, A.W. 1972. Changes in growth-substance contents during growth of wheat grains.
Ann. Appl. Biol. 72: 327-334.
Whittier, R.F., D.A. Dean and J.C. Rogers. 1987. Sequence analysis of alpha-amylase and thiol
protease genes that are hormonally regulated in barley aleurone cells.. Nuci. Acids Res.
15: 2515-2535.
Williams, B.A. and A. Tsang. 1994. Analysis of multiple classes of abscisic acid-responsive
genes during embryogenesis in Zea mays. Dev. Genet. 15: 415-424.
Williamson, J.D. and R.S. Quattrano. 1988. ABA regulation of two classes of embryo-specific
sequences in mature wheat embryos. Plant Physiol. 86: 208-215.
Wilson, G.F., A.M. Rhodes and A.M. Dickinson. 1973. Some physiological effects of viviparous
genes vp1 and vpS on developing maize kernels. Plant Physiol. 52: 350-356.
Xu, N., K.M. Coulter and J.D. Bewley. 1990. Abscisic acid and osmoticum prevent germination
of developing alfalfa embryos, but only osmoticum maintains the synthesis of
developmental proteins. Planta 182: 382-390.


TABLE OF CONTENTS
page
ACKNOWLEDGMENTS H
ABSTRACT iv
INTRODUCTION 1
REVIEW OF LITERATURE 4
Developmental and Hormonal Regulation of Seed Maturation 4
Isolation of Mutants Affected In Seed Maturation 6
Analysis of Gene Expression 13
The Aleurone Germination Response in Cereal Seeds 21
Hormonal Regulation 22
The a-Amylase Genes 23
The Organization of a-Amylase Promoters 25
GA and ABA Signal Transduction 28
The Developmental Switch from Seed Maturation to Seed Germination 31
Quiescent Seeds 33
Dormant Seeds 34
MATERIALS AND METHODS 39
Plant Material 39
Plasmid Constructs 41
Particle Bombardment and Transient Expression 43
RESULTS 47
Repression of Hydrolase Genes by VP1 in Aleurones of Developing Maize Seeds 47
Interaction between VP1 and Abscisic Acid 52
Over-expression of VP1 in Aleurones of Germinating Maize and Barley Seeds 54
Interaction between VP1 and Gibberellic Acid 56
Role of the Embryo in Repression of a-amylase Genes in the Aleurone 61
Functional Analysis of the VP1 Protein 64
DISCUSSION 74
SUMMARY AND CONCLUSIONS 90
REFERENCES 91
BIOGRAPHICAL SKETCH
104


103
Yomo, H. 1960. Studies on the a-amylase activating substance. IV. On the amylase activating
action of gibbereilin. Hakko Kyokaishi. 18:600-602.
Yu. S.M., Y.H. Kuo. G. Sheu, Y.J. Sheu and LF. Uu. 1991. Metabolic de-repression of a-
amylase gene expression in suspension-cultered cells of rice. J. Biol. Chem. 266:
21131-21137.
Zeevaart, JAD. and RA Creelman. 1988. Metabolism and physiology of abscisic acid. Ann.
Rev. Plant Physiol. Mol. Biol. 39: 439-73.
Zwar, JA, and R. Hooley. 1986. Hormonal regulation of a-amylase gene transcription in wild
oat (Avena fatua L.) aleurone protoplasts. Plant Physiol. 80: 459-463.


11
seeds, root or shoot tissues (McCarty et at. 1989a; Carson, 1992). Hence, the expression pattern
of Vp1 is highly consistent with the seed-specific phenotype of the vpf mutant.
The Vp1 gene consists of six exons and five introns. Apart from putative VP1 homologs
cloned from barley, rice and Arabidopsis (M. Stoll and D.R. McCarty, unpublished results; Hattori
et al., 1994; Giraudat et al., 1992), the sequence of VP1 shows no significant homologies to any
known protein sequences, suggesting that VP1 is a novel protein. The N-terminus of VP1 is
predicted to form two negatively charged amphipathic helices, a feature which is characteristic of
many bacterial and eukaryotic transcriptional activators (Ptashne, 1988). Indeed, this region of
VP1 was found capable of functionally replacing the addic activation domain of the bacterial
transcription factor GAL4 in a eukaryotic gene expression system (McCarty et al., 1991). This
confirmed that the acidic region of VP1 has transcriptional activator function and suggested that
VP1 may function as a regulatory protein in controlling seed maturation and anthocyanin
accumulation.
In Arabidopsis, mutants displaying reduced sensitivity to ABA have been identified using
genetic screens selecting for the ability of seeds to germinate on medium containing at least
3 |j.M ABA, a concentration that inhibits germination of wild-type seeds. In such screens, five loci
controlling ABA-sensitivity have been identified: Abi1, Abi2, Abi3 (Koomeef et al., 1984), Abi4
and Abi5 (Finkelstein, 1994). Mutations in any of these loci confer reduction in seed dormancy.
However, while the phenotype of ab¡3, abi4 and abi5 mutants is restricted to seed tissues, abi1
and abi2 mutants are also impaired in stomatal regulation and a variety of stress responses in
vegetative tissues (Koomeef et al., 1984; Finkelstein, 1994; Chandler and Robertson, 1994).
Interestingly, when the abi3-1 mutant was crossed to the ABA-defident aba mutant, seeds of the
resulting double mutant were desiccation intolerant, remained green and frequently displayed
vivipary (Koomeef et al., 1989). The phenotype of the double mutant suggests that the abi3-1
allele may be leaky and allow some ABA-responsiveness. Hence, additional reduction in seed
ABA concentrations may be necessary to obtain a viviparous phenotype. Indeed, severe abi3
mutants were isolated that were phenotypically similar to the abi3-1/aba double mutant,


96
Jacobsen, J.V. and T.J.V. Higgins. 1982. Characterization of the a-amylases synthesized by
aieurone layers of Himalaya barley in response to gibberellic acid. Plant Physiol. 70:
1647-1653.
Jacobsen, J.V., JA Zwar and P.M. Chandler. 1985. Gibberellic-acid-responsive protoplasts
from mature aieurone of Himalaya barley. Planta 163: 430-438.
Jones, R.J. and M.L. Brenner. 1987. Distribution of abscisic acid In maize kernel during grain
filling. Plant Physiol. 83: 905-909.
Jones, R.L. and J.V. Jacobsen. 1991. Regulation of synthesis and transport of secreted proteins
in cereal aieurone. int. Rev. Cytol. 126: 49-88.
Karrer, E.E. and R.R. Rodriguez. 1992. Metabolic regulation of rice a-amylase and sucrose
synthase genes in planta. Plant J. 2: 517-523.
Karssen, C.M., D.L.C. Brinkhorst-van der Swan, A.E. Breekland and M. Koomeef. 1983.
Induction of dormancy during seed development by endogenous abscisic acid: studies
on abscisic acid-deficient genotypes of Arabidopsis thaliana (L.) Heynh. Planta 157:
158-165.
Karssen, C.M., and E. Lapka. 1985. A revision of the hormone balance theory of seed
dormancy: studies on gibberellin and/or abscisic acid-deficient mutants of Arabidopsis
thaliana. In: Plant Growth Substances 1985, M. Bopp ed., Springer Vertag, Berlin, pp.
315-323.
Keith, K., M. Kraml, N.G. Dengler and P. McCourt. 1994. fusca3: a heterochronic mutation
affecting late embryo development in Arabidopsis. Plant Ceil 6: 589-600.
Kermode, A.R. 1990. Regulatory mechanisms involved in the transition from seed development
to germination. CRC Crit. Rev. Plant Sci. 9:155-195
Khursheed, B., and J.C. Rogers. 1988. Barley a-amylase genes: Quantitative comparison of
steady-state mRNA levels from individual members of the two different families
expressed in aieurone cells. J. Biol. Chem. 263:18953-18960.
King, R.W. 1976. Abscisic acid in developing wheat grains and its relationship to grain growth
and maturation. Planta 132: 43-51.
King, R.W. 1979. Abscisic acid synthesis and metabolism in wheat ears. Aust. J. Plant Physiol.
6:99-108.
Koomeef, M., A. Elgersma, C.J. Hanhart, E.P. van Loenen-Martinet, L. van Rijn and JAD.
Zeevaart. 1985. A gibberellin-insensitive mutant of Arabidopsis thaliana. Physiol.
Plant. 65: 33-39.
Koomeef, M. C.J. Hanhart, H.W.M. Hilhorst and C.M. Karssen. 1989. In vivo inhibition of seed
development and reserve protein accumulation in recombinants of abscisic acid
biosynthesis and responsiveness mutants in Arabidopsis thaliana. Plant Physiol. 90:
463-469.
Koomeef, M., M.L. Joma, D.L.C. Brinkhorst-van der Swan and C.M. Karssen. 1982. The
isolation of abscisic acid (ABA) deficient mutants by selection of induced revertants in


REFERENCES
Akazawa, T. and S. Miyata. 1982. Biosynthesis and secretion of a-amylase and other
hydrolases in germinating cereal seeds. Essays Biochem. 18:41-78.
Amuti. K.S. and C.J. Pollard. 1977. Soluble carbohydrates of dry and developing seeds.
Phytochem. 16: 529-532.
Anderberg, R.J. and M.K. Walker-Simmons. 1992. Isolation of a wheat cDNA clone for an
abscisic acid-inducible transcript with homology to protein kinases. Proc. Natl. Acad.
Sci. USA 89: 19183-10187.
Anderson, I.C. and D.S. Robertson. 1960. Role of carotenoids in protecting chlorophyll from
photodestruction. Plant Physiol. 35: 531-534.
Armstrong, C., M. Black, J.M. Chapman, H.A. Norman and R. Angold. 1982. The induction of
sensitivity to gibbereliin in aleurone tissue of developing wheat grains. I. The effect of
dehydration. Planta 154: 573-577.
Bflumlein, H., S. Misra, H. Luerpen, K. Kfllle, C. Horstmann, U. Wobus and A.J. Mller. 1994.
The FUS3 gene of Arabidopsis thaliana is a regulator of gene expressin during late
embryogenesis. Plant J. 6: 379-387.
Beckett, J.B. 1993. Locating recessive genes to chromosome arm with B-A translocations. In:
The Maize Handbook, M. Freeling and V. Walbot, eds., pp. 308-314.
Bethke, P.C. and R.L. Jones. 1994. Ca2+-calmodulin modulates ion channel activity in storage
protein vacuoles of barley aleurone cells. Plant Cell 6: 277-285.
Blackman, S.A., R.L. Obendorf and A.C. Leopold. 1992. Maturation proteins and sugars in
desiccation tolerance of developing soybean seeds. Plant Physiol. 100: 225-230.
Blatt, M.R. and G. Thiel. 1993. Hormonal control of ion channel gating. Ann. Rev. Plant
Physiol. Mol. Biol. 44: 543-567.
Brown, A.H.D. and J.V. Jacobsen. 1982. Genetic basis and natural variation of a-amylase
isozymes in barley. Planta 157: 493-501.
Bruni, F. and A.C. Leopold. 1991. Glass transitions in soybean seed: Relevance to anhydrous
biology. Plant Physiol. 96: 660-663.
Bush, D.S., A.K. Biwas and R.L. Jones. 1993. Hormonal regulation of Ca2* transport in the
endomembrane system of the barley aleurone. Planta 189: 507-515.
Carson, C.B. 1992. Characterization of the structure and expression of the mutant alleles of the
viviparous-1 gene of maize. U.S. dissertation, University of Florida.
91


2
consistent with the relatively low abundance of Vp1 message in whole endosperm extracts
(McCarty et al., 1989a).
Vp1 encodes a novel, 73 kD protein with a functional acidic transcriptional activation
domain (McCarty et al., 1991). Over-expression of VP1 in maize protoplasts frans-activated
reporter constructs containing late-embryogenesis-specific promoters: C1, a maize gene that
encodes a transcription factor required for anthocyanin synthesis in the seed, and Em, a wheat
LEA gene (Hattori et al., 1992; McCarty et al., 1991). In agreement with the phenotype of ABA-
deficient mutants, VP1-activation of Em was strongly dependent on the presence of exogenous
ABA (McCarty et al., 1991). These functional data confirm that VP1 plays a central role in the
induction of seed maturation.
Following imbibition of mature non-dormant seeds, expression of maturation-specific
genes is terminated and expression of a new set of genes related to the developmental program
of seed germination is executed (Comai and Harada, 1990). In rehydrated cereal seeds, the
germination-specific a-amylase genes which encode starch-hydrolyzing enzymes are induced in
the aleurone cells by the hormone gibberellic acid (GA) that is secreted by the embryo early in
germination (Jacobsen and Chandler, 1987). They are constitutively expressed in de-germed
seeds of the barley GA-response mutant slender (Chandler, 1988; Lanahan and Ho, 1988), and
their induction can be antagonistically inhibited by application of ABA (Jacobsen and Chandler,
1987).
Expression of the normally consecutive programs of seed maturation and seed
germination is under strict developmental control. Precocious induction of germination-related
events prior to seed maturity appears to be actively repressed. In developing seeds of cereals
and maize, no a-amylase activities are found prior to seed maturity (Evans et al., 1975; Nicholls,
1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Moreover, a-
amylase genes are unresponsive to applied GA (Nicholls, 1979; Comford et al., 1986; Garcia-


RESULTS
Repression of Hydrolase Genes bv VP1 In Aleurones of Developing Maize Seeds
Phenotypic analysis of vp1-m2 kernels
The vp1-m2 allele of Vp1 carries a transposon insertion in the third intron which causes
somatic instability of the gene during endosperm development (McCarty et al 1989b). As a
result, mosaic kernels develop with clonal vp1 mutant and wild-type sectors. In these kernels, a
striking pattern of endosperm remobilization is often evident. Endosperm tissue underlying vp1
mutant aleurone cells is frequently softened and depressed in surface while wild-type sectors are
raised relative to adjacent mutant sectors. This produces kernels with a distinctive etched
appearance (Fig. 5a). The softening response was also observed when only a small fraction of
the endosperm was comprised of mutant tissue (Fig. 5b), indicating that expression of this
phenotype is cell autonomous. The softening of starchy endosperm tissues that underlie islands
of vp1 mutant aleurone cells appears to be attributed to precocious induction and secretion of
hydrolytic enzymes caused by the loss of VP1 function. Thus, repression of hydrolases in
developing maize kernels is evidently dependent on the presence of functional VP1.
Transient expression of Amv-GUS in maize aleurone
In order to more directly address the role of VP1 in repressing hydrolase activity, a
quantitative transient gene expression assay was developed that is based on particle
bombardment of aleurone tissue with a barley high-pl a-amylase promoter-GUS fusion construct
(Amy-GUS). It was first interesting to determine whether the observed differential activity of
hydrolases in vp1 mutant and wild-type sectors was due to transcriptional control. For this
purpose, Amy-GUS was introduced into aleurone cells of developing vp1 mutant and
47


DISCUSSION
VP1 of maize is a transcription factor that is specifically expressed in the developing
seed (McCarty et al., 1989a, 1991). It was shown previously that VP1 is required for ABA-
induced activation of a variety of genes associated with seed maturation (McCarty et al., 1991).
Results of this work show that, in addition to its transcriptional activator function, VP1 has a
specific role in blocking precocious induction of germination-specific a-amylase genes during
seed development.
VP1 Represses a-AmvIase Genes
This study provides at least three lines of evidence that indicate a function of VP1 in
repression of a-amylase genes in the developing seed. First, somatically unstable vp1-m2 seeds
containing both vp1 mutant and wild-type sectors displayed cell autonomous de-repression of
endosperm remobilization specifically in sectors underlying vp1 mutant aleurone (Fig. 5a,b).
Second, in transient expression experiments Amy-GUS was inducible or constitutively active in
developing vp1 mutant aleurone cells but not in wild-type aleurone cells (Table 1). Third, co-
expression of recombinant VP1 with Amy-GUS in vp1 mutant aleurone cells inhibited Amy-GUS
expression by >95% (Fig. 6). These results are consistent with findings that a-amylase genes
are not expressed in the developing seed (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et
al., 1990; Oishi and Bewley, 1990). Hence, cessation of VP1 expression prior to germination
may be necessary to allow induction of a-amylase genes in the germinating seed.
74


77
Z.m. MEA-SSGSSPPHSQENPPEH GGD M-GG AP-AEEI GGEAA DDF 39
I I I I It I I I II II Ml
H.V. MDA-SAGPPPPRHPQGSALRRGKG P-AVEIRHGE DDF 34
III III I II I M
O.s. MDA-SAGS SAP HSHGNPGKQ-GGG GGGGGGRGKAP-AAEIR-GEAAR DDV 46
I I I II I I M
A.t. MKSLHVAANAGDLAEDCGIL-GGDADDTVLMDGIDEVGREIWLDD HGGDNNHVHGHQDDL 60
Z.m. MFAEDTF PSLPDFPCLSSPSSSTFSSN S S S NS S S AYTNT AGRA- G G 86
I I I I I I I I I I I I I I I I I I I I
H.V. MFAQD--TF PAFPDFPCLSSPSSSAADIV LCG 64
I I I I I I I I I I I I I I I I I I I
0.3. FFADDTF PLLPDFPCLSSPSSSTFSSS SSSNSSSAFTTAAGGGCG G 94
I I I I I I I I I I I I I I I II I I I I
A.t. IVHHDPSIFYGDLPTLPDFPCMSSSSSSSTSPAPVNAIVSSASSSSAASSSTSSAASWAILRS 123
Z.m. EPSEPASAGEGFDA LDDIDQLLDFASLSMPWDSE-P 125
I I I I I I I I I I I I I I I I I I I I I I I I
H.V. EPSEPAAAGDGMDD LS DIDHLLDLASINDDVPWDDE PL 102
I I I I I I I I I I I I I I I I I I I III III
0.3. EPSEPASAADGFGE LADIDQLLDLASLSVPWEAEQPL 135
I III
A.t. DGEDPTPNQNQYASGNCDDSSGALQSTASMEIPLDSSQGFGCGEGGGDCIDMMETFGYMDLLD 186
86
Z.m. FP-GVSMMLENAMSAPPQPVGDGMSEEKAVPEGTT GGEEACH-DASEG-EE 163
I I I I I I I I I I I I I I I I
H.v. FP-DVGMMLEDVIS E QOQQQQQH P LAGHGAGGRVAS DTAGG- GGE DAFMGGGGS GS AADD 160
II Mill II I I I I I II II I II I III
0.3. FPDDVGMMIEDAMSGQPHQADDCTGDGDTKAVMEAAGGGDDAGDACM-E-GS-DAPDD 179
II III I I I I I I
A.t. SNEFFDTSAIFSQDDDTQNPNLMDQTLERQEDQV-WPMMENNS-GQDMQMMNSSLEQDDD 240
85
Z.m. LPRFFMEWLTSNRENISAEDLRGIRLRRSTIEAAAARX;GGGRQGTMQLLKLILTWVQNHHLQR 230
1111111111 i 111111 Ti 1111 ill Ti 1111 111111 ni 1111 1111
H.V. LPRFFMEWLTNIRDCIS AEDLLSIRLRRS TIETTTALLGGGRQDTMQLLKLILTWVQSHHLQK 223
ll 1111111 ll III ll iTi 111ill l Tl 1111 111111iTi1111 11111
0.3. LPAFFMEWLTSNREYISADDLRSIRIRRSTIEAAAARLGGGRQGTMQLLKLILTWVQNHHLQK 247
I I I I I I II I I I I I ~ I III I I iTi I I I I I IiTi I I I III
A.t. LAAVFLEWLKNNKETVSAEDLRKVKIKKATIESAARRLGGGKEAMKQLLKLILEWVQTNHLQR 308
87 92 93
Z.m. KRP RDVMEE EA- GLHVQLP S PVANP PGYE FP AGGQDMAAGGGT SWM PHQQAFTPPAAYG 288
I I I I II I I I I I I I I I I I I I I I I I I I I I I I I I
H.v. KRPRVGAMDQEAPPAGGQLPSPGANPS-YEFPT ETGAAAATSWM PY-QAFSPTASYG 278
I I I I I I I I I I I I I I I I I I I I I I I I I I I I II I I I I I I
0.3. KRPRTAIDDGAA-SSDPQLPSPGANP-GYEFPSGGQEMGSAAATSWM PYQ-AFTPPAAYG 304
II I I I II I
A.t. RRTTTTTTNLSY-QQSFQQDPFQNPNPNNNNLIPPSDQTCFSPSTWPPPPQQQAFVSDPGFG 370
181 196,95
Z.m. GDAVYPSAAGQQYSFHQGPSTSSVWNSQPFSPP PVGDMH GANMAWPQQYVPFPPPG 345
I I I I I I I I I I I I I I I I I I I III I I I I I I III
H.v. GEAMYPFQ QGCSTSSVAVSSQPFSPPAAA-DMHAG AWPLQYAAFVPAG 325
I I I I I I II I I I I I I I I I I I I I I I I I I I I I I I I I I I
0.3. GDAMYPGAAG-PFPFQQSCSKSSVWSSQPFSPPTAAAAGDMHASGGGNMAWPQQFAPFPV 364
II I I II I I I I I I I
A.t. YMPAPNYPPQPEFLPLLESPPSWPPP PQ SGPMP-HQQF-PM-PPT 412


ACKNOWLEDGEMENTS
I am deeply grateful to Dr. Donald McCarty and Dr. Indra Vasil for their valuable advice,
guidance and encouragement throughout the course of this study. I wish to thank Drs. Karen
Koch, Nigel Richards, Alice Harmon and Bill Gurley for serving as members of my committee. I
also thank all past and present members of both laboratories for their helpful discussions as well
as for providing an enjoyable and lively work atmosphere. I am especially grateful to Lennie
Rosenkrans and Dr. Mark Taylor for their immeasurable technical assistance. I wish to extend
my thanks to Ms. Elaine Summers for her frequent help in dealing with the administrative side of
Graduate School.
ii


30
have been attributed to feedback regulations on GA metabolism in response to altered GA-
sensitivity.
To the class of GA-insensitive mutants belong the Rht (reduced height) mutants of
wheat. A total of 10 Rht loci have been identified, showing varying degrees of dominance. Of
these, the mutation Rht3 exerts the strongest dwarfing effect. The GA-insensitive phenotype of
Rht3 is also expressed in the aleurone: germinating seeds had 75% reduced levels of amylase
activity as compared to tall (rht) varieties and showed no or little increase in amylase activity
after GA treatment (Gale and Marshall, 1975; Fick and Qualset, 1975). The degree of GA-
insensitivity of the aleurone was found to increase with the dosage of Rht3 alleles (Gale and
Marshall, 1975). A similar GA-insensitive mutant has been described in rice (Mitsunaga et al.,
1994). The finding that the failure to respond to GA is expressed in plant and seed tissues
indicates that these tissues share at least in part a common signal transduction pathway.
Two dominant GA-insensitive dwarf mutants have been identified in maize (D8, D9).
Besides being of reduced stature, these mutants mimic additional characteristics of GA-deficient
mutants, such as reduced apical dominance and formation of anthers on the ear (Coe and
Neuffer, 1977). D8 and D9 are located on 1L and 5S, respectively (Coe and Neuffer, 1977).
Since the region of 5S contains duplicate loci with 1L it is likely that D8 and D9 are duplicate loci
encoding gene products with identical or similar function. X-ray-induced chromosome breakage
was used to create clonal sectors of wild-type cells within D8 mutant tissue. Results from these
experiments indicated that 08-mediated effects can be expressed cell autonomously at least in
some tissues (Harberd and Freeling, 1989) which is consistent with the hypothesis that the wild-
type gene product is part of a GA signal transduction pathway. The gain-of-function nature of
the mutation in association with a GA-insensitive phenotype allows to speculate about the
function of the wild-type gene product. Possibly, cf8(+) (and d9(+)) encodes a negative regulator
of GA response that normally is inactivated by exposure of the cell to GA. in this scenario, the
mutant D9 protein would be constitutively active in this repressor activity, i.e. even in the
absence of GA (Harberd and Freeling, 1989).


24
(Brown and Jacobsen, 1982; Muthukrishnan et al., 1984). Isolation and sequencing of a number
of cDNA clones verified that there are sequence differences between Isoforms. Base sequence
homology Is 90-95% within gene families and about 75% between gene families (Jacobsen and
Chandler, 1987). Southern blot analyses of two barley varieties revealed that there are at least
6-7 high-pl genes and at least 3 low-pl genes, indicating that a-amylases are encoded by two
multigene families (Khursheed and Rogers, 1988; Muthukrishnan et al.. 1984). While no detailed
mapping has been undertaken in barley, three rice a-amylase genes were found to be clustered
within 28 kb of genomic DNA. Molecular analysis suggested gene duplication as a cause (Sutliff
et al., 1991).
The two barley a-amylase gene families are regulated differently in the aleurone of
germinating seeds. mRNA levels of the high-pl isoforms increase very quickly upon imbibition,
reaching a maximum after two days and decreasing to low levels after four days. Synthesis of
low-pl isoform-mRNAs begins later, not before three days after imbibition, but then increases
rapidly so that low-pl isozymes become the dominant enzyme group after four days of imbibition
(Chandler and Jacobsen, 1991). The two isoforms are also differentially responsive to GA. In
some studies with isolated aleurone layers, mRNA as well as protein of the low-pl isoforms can
be detected before GA is added, while those of high pl-isoforms cannot (Chandler and Jacobsen,
1991; Jacobsen and Higgins, 1982; Rogers, 1985). Others, using isolated aleurone layers or
aleurone protoplasts, find no low-pl message in the absence of GA (Chandler and Jacobsen,
1991; Nolan and Ho, 1988). However, low-pl isoforms appear to be more sensitive to GA as
they respond to GA-concentrations as low as 10'9 M (Nolan and Ho, 1988). Thus, low-pl a-
amylase genes might be either more responsive to GA or leaky in expression.
In comparison to barley a-amylases, many fewer studies on maize a-amylases have
been reported. Partial purification of a-amylases from endosperm of germinating seeds has
revealed two major groups of isozymes, one with pis of 5.1-5.7, the other with pis of about 4.6
(Warner and Knutson, 1991). Other authors report the purification of a-amylase isozymes with a
variety of pis (Warner et al., 1991; MacGregor et al., 1988; Chao and Scandelios, 1971). No


INTRODUCTION
The formation of seeds is a unique characteristic of higher plants which promotes
dispersal of the species and allows interruption of the life cycle during unfavorable environmental
conditions. To survive in the dehydrated state, plant embryos undergo an adaptation process
during late stages of seed formation (maturation phase) which renders them tolerant to
desiccation and gradually causes arrested growth. In maize and other cereals, the outermost
layer of the seed endosperm (aleurone layer) also undergoes a maturation process and remains
viable through desiccation.
Seed maturation is associated with the activation of a variety of genes encoding storage
proteins and various hydrophilic, late-embryogenesis-abundant (LEA) proteins which possibly
function as desiccation protectants (Dure et al. 1989; Skriver and Mundy 1990). Analysis of
viviparous mutants in maize has demonstrated that the developmental program of seed
maturation is controlled by at least two factors, the hormone abscisic acid (ABA) and the product
of the Viviparous-1 (Vp1) gene (Robertson 1955; Neill et al. 1986). Developing vpf mutant
embryos are distinct from ABA-deficient embryos in that they exhibit a reduced sensitivity to
ABA in culture (Robichaud et al., 1980; Robichaud and Sussex, 1986). In addition to causing
vivipary, the vp1 mutation blocks synthesis of anthocyanins in embryo and aleurone tissues
(Robertson, 1955; Dooner, 1985). The vpf mutant phenotype is restricted to seed tissues.
Mutant embryos rescued prior to desiccation develop into apparently normal, fully fertile plants
with normal patterns of anthocyanin accumulation.
The Vp1 gene was cloned by transposon tagging (McCarty et al., 1989a). It encodes a
2500-nucleotide mRNA that is expressed specifically in embryo and endosperm tissues of the
developing seed. Within the endosperm, Vp1 expression may be limited to the aleurone layer
because so far no mutant phenotype has been detected in the starchy endosperm. This is
1


27
regulated transcription In the context of the low-pl a-amylase promoter In a similar way as it does
in its native rice promoter. Importantly, its function in the amylase promoter was highly
dependent upon the presence of the 02S sequence. Thus, the 02S element appears to function
as a "coupling element" that is necessary for high-level, hormone-regulated transcription from
the low-pl a-amylase promoter (Rogers and Rogers, 1992). No 02S-like sequence Is evident in
high-pl a-amylase promoters. However, inserting the 02S element from a low-pl promoter into a
high-pl promoter at a position upstream of the pyrimidine box enhanced transcription ca. 5-fold,
suggesting that the 02S element function could interact properly with the high-pl promoter
fragment to give high-level transcription (Rogers et al., 1994).
To identify the trans-acting factors that bind to the crs-elements in the promoters and
thereby confer hormone-dependent expression of the a-amylase genes, DNA-protein interactions
have been characterized using band shift assays and DNase I footprinting analyses. In a wheat
low-pl promoter (a-Amy 2/54), the 02S box and the TAACAGA element, but not the TATCCAC
sequence, were found protected from DNase I digestion, confirming the binding of protein
factor(s) to at least two functionally important elements. However, these binding activities were
not dependent upon GA (Rushton et al., 1992). Evidence for the presence of a GA-dependent
factor on a barley low-pl promoter was provided by Sutliff et al. (1993). Results from band shifts
performed in this study demonstrated that a GA-inducible binding activity interacted with the
TAACAGA and TATCCAC elements in a sequence-specific manner. In a different report on a
rice a-amylase promoter (Amy3c), GA-dependent binding to a pyrimidine box-like sequence was
demonstrated (Goldman et al., 1994). However, since this promoter displays little sequence
homology to barley and wheat promoters, comparisons cannot be made. In summary, GA-
independent as well as GA-dependent factors appear to constitute the complex(es) on a-amylase
promoters that result in GA-regulated expression. Further characterization of the protein-DNA
and protein-protein interactions is needed.


75
Gene Repression is a Discrete Function of VP1
In contrast to the mechanism of transcriptional activation of maturation-specific genes,
VP1 -mediated repression of a-amylase genes does not require the transcriptional activation
function located at the N-terminal domain of VP1 (Fig. 17). This indicates that VP1 has a
discrete repressor function that is mechanistically distinct from the transcriptional activation
function. Several systems in which a single transcription factor functions as both an activator
and a repressor depending on the target promoter have been described in animals (Miner and
Yamamoto, 1991; Tsai and OMalley, 1994). Direct structural homologs of VP1 are thus far
known only in plants, suggesting that this strategy has evolved independently in plants and
animals.
Functional Analysis of the VP1 Protein
To identify domains in the VP1 protein that are important for repressor function, mutant
derivatives containing deletions covering ca. 80 % of the total protein were tested for their ability
to inhibit Amy-GUS. Deletion of very highly conserved sequences in the C-terminal half of VP1
(103/104, 101/100, McW) did not, or only slightly, reduce repressor function (Fig. 19). In
contrast, two constructs deleting sequences in the middle of the VP1 protein (85/87, 87/88) were
strongly affected in repression of Amy-GUS (Figs. 18). While disruption of VP1 function is one
possibility for lack of Amy-GUS repression, low stability of mutant mRNA or protein could be an
alternative explanation. However, this possibility is unlikely for two reasons: 1) both constructs
were capable of activating a C1-Sh-GUS reporter gene in maize protoplasts: 85/87 and 87/88
activated C1-Sh-GUS at a level of 77-84% or 56% of the wild-type VP1 construct, respectively
(V. Vasil, L. Rosenkrans et al., unpublished results). 2) co-expression of 87/88 as well as the
double-deletion mutant 85/88 exhibited a dominant negative effect on repression of Amy-GUS
by wild-type VP1 in barley aleurone, indicating presence of mutant protein in transformed cells.


VP1-MEDIATED REPRESSION OF ALPHA-AMYLASE GENES
IN DEVELOPING MAIZE ALEURONE
BY
UTE HOECKER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1995


87
embryo per se is not required for de-repression of Amy-GUS in vp1 mutant aleurone. Instead, it
rather appears to be the lack of a normal embryo that facilitates expression of Amy-GUS,
suggesting that a wild-type embryo contributes a diffusable signal with inhibitory effect on a-
amylase gene expression in the aleurone. Experimental evidence suggests that developing
embryos are the major source of ABA present in the maturing seed (King, 1979; Jones and
Brenner, 1987). Because Amy-GUS remains sensitive to inhibition by ABA in vp1 mutant
aleurone, ABA produced by the wild-type embryo may be responsible for the observed
repression of Amy-GUS in vp1 mutant aleurone of non-concordant seeds. This is consistent with
the finding that Amy-GUS was de-repressed to a similar extent in developing aleurones of the
ABA-deficient mutant vpS as in aleurones of the germless mutant (data not shown).
In concordant vp1 mutant seeds, ABA concentrations are equal to, or only ca. two-fold
lower than, those present in wild-type seeds (Neill et al., 1986,1987; Paiva and Kriz, 1994). This
suggests that the viviparous embryo also contributes an inductive signal (e.g. GA) that
counteracts the effect of ABA and therefore uncovers de-repression of a-amylase genes in vp1
mutant aleurone. However, the observed strong expression of Amy-GUS in vp1 mutant aleurone
of germless seeds clearly shows that GA production by the embryo is not required for a-amylase
expression in maize aleurone. Supported by the evidence that the endosperm of mature cereal
seeds is not a source of GA (Jacobsen and Chandler, 1987), these data suggest that Amy-GUS
expression is largely independent of GA. Moreover, this interpretation is consistent with other
studies (Harvey and Oaks, 1974) and findings in this work indicating that in mature seeds of GA-
deficient mutants of maize a-amylase genes are expressed at high levels.
Complete repression of Amy-GUS in aleurones of developing seeds was observed only
if the embryo as well as the endosperm were of wild-type genetic constitution (Fig. 16). Lack of
either a normal, arrested embryo or VP1 expression in the aleurone lead to partial de-repression
of Amy-GUS in aleurone cells. This indicates that neither factor expression of VP1 in the
aleurone cells or the presence of a normal embryo is sufficient for total inhibition of a-amylase
genes.


15
mRIMAs (Parcy et al., 1894). Similarly, maize ABA-defldent mutants accumulated mRNAs
corresponding to the ABA-regulated 7S globulins in only slightly reduced amounts (Krtz et al.,
1990; Paiva and Kriz, 1994). Hence. ABA does not normally appear to be a limiting factor in
expression of these storage protein genes. In contrast, developing embryos of the ABA-
insensitive mutants vp1 of maize and ab/3 of Arabidopsis exhibited very low or undetectable
expression of 7S globulins or 2S and 12S storage protein genes, respectively, indicating that
VP1/ABI3 are required for their expression (Kriz et al., 1990; Paiva and Krtz, 1994; Nambara et
al., 1992). Furthermore, exogenous ABA did not induce expression of storage protein genes in
cultured immature vp1 or ab¡3 mutant seeds, while it did so in cultured immature wild-type seeds
(Paiva and Kriz, 1994; Finkelstein and Somerville, 1990). Thus, VP1/ABI3 appear to be
essential for ABA action. Since VP1/ABI3 are expressed at normal levels in ABA-defident
mutants (McCarty et al., 1991; Paiva and Kriz, 1994; Parcy et al., 1994), It may be that normal
accumulation of storage proteins in these mutants is mediated by VP1/ABI3 either without a
need for ABA or requiring residual amounts of ABA present In mutant seeds.
LEAs/RABs
An extensive survey in Arabidopsis examining the accumulation kinetics of 18 marker
mRNAs expressed at high levels during mid to late seed development suggested that
LEAs/RABs-encoding genes fall into distinct classes with different requirements for ABA and
ABI3 to induce expression. For most markers, transcript levels did not solely correlate with the
amounts of endogenous ABA or ABI3 present in the seed, thus implicating a role of other factors
in controlling temporal patterns of expression (Parcy et al., 1994).
However, abundance of several tested mRNAs was highly reduced in seeds of the aba
mutant as well as the abi3 mutant (Parcy et al., 1994). Similarly in maize, expression of a well-
studied LEA gene (Em) originally isolated from wheat was undetectable in developing seeds of
mutants deficient for ABA or functional VP1 (McCarty et al., 1991). Hence, ABA as well as
ABI3/VP1 appear to be required for expression of certain LEAs, which Is consistent with a
possible role of ABI3/VP1 in ABA perception or signal transduction in the seed.


14
earlier 'maturation stage' comprises the longest time Interval (19 days) and is characterized by
high abundance of storage protein-mRNAs. This phase coincides with the presence of high
concentrations of ABA. It is apparently terminated by abscission of the vascular connections
between embryo and mother plant and is followed by the 'postabsdssion stage' (5 days) during
which maturation stage-specific mRNAs decline rapidly and a new set of mRNAs accumulates.
To these postabscission stage-specific mRNAs belong the LEA'S (this term was introduced by
Galau et al., 1986) and the RAB's (responsive to ASA. a term used by other authors for similar
proteins as LEA'S). Subsequently, seed formation is terminated by rapid water loss and
termination of transcription.
A similar temporal pattern of mRNA accumulation was reported for Arabidopsis (Farcy et
al.. 1994) and maize (Paiva and Kriz, 1994; Williams and Tsang, 1994). Nevertheless, for
species other than cotton, late stages of seed formation are usually refened to as the maturation
phase* which is not subdivided into two stages. It should be mentioned that in maize, the seed
maturation phase also correlates with the accumulation of anthocyanin pigments in embryo and
aleurone tissues.
In the following, progress in our understanding of the regulation of maturation-specific
genes (genes encoding storage proteins, LEA'S and RAB's, proteins of the anthocyanin pathway)
will be reviewed, placing emphasis on the roles of ABA and VP1 as regulators of seed
maturation in monocot seeds.
Storage proteins
In many species, immature embryos cultured in ABA exhibited precocious and enhanced
accumulation of storage proteins and their corresponding mRNAs as compared to those cultured
on ABA-free medium or left to mature on the mother plant (Quatrano, 1987). These results
indicate that ABA upregulates expression of storage protein genes. However, high levels of ABA
are not required for expression of the major storage protein genes, as shown for maize and
Arabidopsis. The aba mutant of Arabidopsis was found to accumulate normal levels of identified
ABA-upregulated storage proteins (2S, 12S) (Koomeef et al., 1989) and their corresponding


17
This apparent independent activity of VP1 may be caused by low levels of maternal ABA present
in vp5 mutant seeds. Alternatively, the abnormally high levels of recombinant VP1 In expressing
cells may allow some ABA-independent activation of the Em promoter normally not found In
vivo.
The VP1 protein was subjected to functional analysis by testing deletion-derivatives for
their ability to irans-activate Em-GUS. Sequence analysis and domain swapping experiments
between VP1 and GAL4 had suggested that the N-terminus of VP1 contains an acidic
transcriptional activation domain (McCarty et al., 1991). Indeed, deletion of this acidic domain
abolished VP1s ability to activate Em-GUS. Replacing it with the acidic activation sequence of
the herpes simplex virus transcription factor VP16 partially restored transcriptional activation of
Em-GUS (McCarty et al., 1991). These results strongly indicate that VP1 functions as a
transcriptional activator in inducing Em-GUS. Analysis of internal deletion-constructs of VP1
identified two highly basic domains that are important for activation of Em-GUS (L. Rosenkrans,
V. Vasil, I.K. Vasil and D.R. McCarty, unpublished results).
Deletion analysis of the Em promoter indicated that two G-box-related sequences
(Em1a: ACACGTGG: Em1b: ACACGTGC) which are conserved in many promoters responsive
to ABA, light or anaerobiosis are involved in VP1- and ABA-mediated activation of Em (Marcotte
et al., 1989; Guiltinan et al., 1990; V. Vasil et al., unpublished results). The finding that ABA-
induction of Em does not require protein synthesis (Williamson and Quatrano, 1987) suggests
that VP1 and ABA frans-activate Em directly through the G-box elements rather than through
activation of intermediate regulatory genes further upstream in the ABA signal transduction
pathway. Thus far, no DNA-binding activity of VP1 to these putative target sequences has been
detected (T. Hattori, B. Li and D.R. McCarty, unpublished results). Hence, VP1 might activate
Em via protein-protein interactions with G-box-binding protein(s). A bZIP-type protein binding
specifically to the Em1a motif has been cloned (Guiltinan et al., 1990) and may thus be a
candidate.


MATERIALS AND METHODS
Plant Material
Except for immature d1 mutant kernels, which were obtained from greenhouse-grown
plants, all maize developing ears were harvested from field-grown plants. Under the local
environmental conditions, kernels typically begin accumulation of anthocyanins at day 17
postpollination and reach seed maturity after ca. 30-33 DAP. The wild-type maize stock used in
this study was a color-converted W22 inbred line carrying all factors required for anthocyanin
pigmentation of the aleurone. The vp1-R allele (Robertson 1955) segregated in a color-
converted W22 inbred line carrying all other factors required for anthocyanin pigmentation of the
aleurone. This line is routinely maintained by selfing. The vp1-m2 allele (originally named vp1-
mum2, McCarty et al., 1989b) arose in Robertsons Mutator transposable element stocks
(Robertson 1978), but was confirmed to carry an Mpi transposable element insertion (D.R.
McCarty, unpublished results). Seed segregating for the vp5 mutation was obtained from the
Maize Genetics Corporation Stock Center (University of Illinois, Urbana-Champaign). To
produce vp1,vp5 double mutant seeds, heterozygous vp5 mutant plants were crossed with
heterozygous vp1-R mutant plants. A mutation conferring embryo abortion at early globular
stage (germless*) arose in a Robertsons Mutator-induced mutant screen (D.R. McCarty et al.,
unpublished results). Germless mutant seed were backcrossed into W22 background for at least
two generations.
vpf-non-concordant seed was generated using a TB translocation stock (Fig. 2). TB3La
seed carrying a BA-translocation on the long arm of chromosome 3, the location of the Vp1
gene, was obtained from the stock center. This seed contains an extra, normally
heterochromatic chromosome, called B-chromosome, in addition to the normal set of A
39


98
Marcotte Jr., W.R., S.H. Russell and R.S. Quatrano. 1989. Absdsic add-responsive sequences
from the Em gene of wheat. Plant Cell 1:969-976.
McCarty, D.R. and C.B. Carson. 1991. The molecular genetics of seed maturation in maize.
Physiol. Plant. 81: 267-272.
McCarty, D.R., C.B. Carson, M. Lazar and S.C. Slmonds. 1989b. Transposable element-
induced mutations of the viviparous-1 gene in maize. Dev. Genet. 10: 473-481.
McCarty. D.R., C.B. Carson, P.S. Stinard and D.S. Robertson. 1989a. Molecular analysis of
viviparous-1: an abscisic acid-insensitive mutant of maize. Plant Cell 1: 523-532.
McCarty, D.R., T. Hattori, C.B. Carson, V. Vasil, M. Lazar and I.K. Vasil. 1991. The Viviparous-
1 developmental gene of maize encodes a novel transcriptional activator. Cell 66: 895-
905.
McWha, J.A. 1975. Changes in abscisic acid levels in developing grains of wheat (Triticum
aestivum L.). J. Exp. Bot. 26: 823-833.
Meinke, D.W. 1992. A homeotic mutant of Arabidopsis thaliana with leafy cotyledons. Science
258: 1647-1650.
Meinke, D.W., L.H. Franzmann, T.C. Nickle and E.C. Yeung. 1994. Leafy cotyledon mutants of
Arabidopsis. Plant Cell 6:1049-1064.
Meurs, C., A.S. Basra, C.M. Karssen and L.C. van Loon. 1992. Role of abscisic acid in the
induction of desiccation tolerance in developing seeds of Arabidopsis thaliana. Plant
Physiol. 98: 1484-1493.
Meyer, K M.P. Leube and E. Grill. 1994. A protein phosphatase 2C involved in ABA signal
transduction in Arabidopsis thaliana. Science 264:1452-1455.
Meyerowitz, E.M. 1987. Arabidopsis. Ann. Rev. Genet. 21: 93-111.
Meyerowitz, E.M. 1994. Structure and organization of the Arabidopsis thaliana nuclear genome.
In: Arabidopsis, E.M. Meyerowitz and C.R. Scmerville, eds., Cold Spring Harbor
Laboratory Press, New York.
Miner, J.N. and K.R. Yamamoto. 1991. Regulatory crosstalk at composite response elements.
Trends Biochem. Sci. 16: 423-426.
Mitsunaga, S., T. Tashiro and J. Yamaguchi. 1994. Identification and characterization of
gibberellin-insensitive mutants selected from among dwarf mutants of rice. Theoret.
Appl. Genet. 87: 705-712.
Morris, C.F. J.M. Moffatt, R.G. Sears and G.M. Paulsen. 1989. Seed dormancy and response of
caryopses, embryos, and calli to abscisic acid in wheat. Plant. Physiol. 90: 643-647.
Morris, F.M., R.J. Anderberg, P.J. Goldmark and M.K. Walker-Simmons. 1991. Molecular
cloning and expression of abscisic acid-responsive genes in embryos of dormant wheat
seeds. Plant Physiol. 95: 814-821.
Mozer, T.J. 1980. Control of protein synthesis in barley aleurone layers by the plant hormones
gibberellic acid and abscisic acid. Cell 20: 479-485.


54
Over-expression of VP1 in Aleurones of Germinating Maize and Barley Seeds
Endogenous expression of Vp1 in embryo and aleurone tissues is under strict
developmental control. Vp1 mRNA peaks at 16 DAP and then gradually decreases as the seed
reaches maturity (McCarty et al., 1991). Germinating seeds, in contrast, display no Vp1
expression or detectable levels of VP1 protein (Carson, 1992). Thus, VP1 function in maize Is
limited to the maturing seed. To test whether VP1 can function in germinating seeds in a way
equivalent to maturing seeds, we co-expressed 35S-Sh-VP1 and Amy-GUS In aleurones of
germinating wild-type seeds of maize. In the presence of exogenous GA, VP1 reduced Amy-
GUS expression by ca. 90% (Fig. 8). Thus, VP1 also functions in germinating seeds, apparently
without the need for additional developmental factors. Furthermore, VP1 also repressed
expression directed by the barley low-pl a-amylase promoter (JR303) which shows considerable
sequence divergence from Amy-GUS. This indicates that expression of high- as well as low-pl a
-amylase genes is under control of VP1.
Because germination-specific responses are well characterized in barley, VP1-mediated
repression was tested in aleurones of germinating barley seeds. Though not as effective as in
maize, VP1 also inhibited Amy-GUS and JR303 expression in barley (Fig. 8).
Moreover, variation within as well as between experiments was significantly reduced in
aleurones of germinating wild-type seeds of maize and barley as compared to aleurones of
developing vp1 mutant maize seeds. Therefore bombardment of germinating seeds constitutes
a useful experimental system to further characterize VP1 function.
A VP1 dose-response curve was generated to determine the amount of co-expressed
VP1 necessary to achieve maximum repression of Amy-GUS in barley aleurone cells. Figure 9
shows that repression was already evident when 1.25 pg of 35S-Sh-VP1 were co-transferred with
Amy-GUS, while increasing the amount of 35S-Sh-VP1 beyond 2.5 pg did not lead to further
repression of Amy-GUS. Hence, comparatively low amonts of recombinant VP1 are sufficient to
achieve inhibition of Amy-GUS. To confirm that the inhibitory effect of VP1 on Amy-GUS in
barley aleurone is promoter-specific, 35S-Sh-VP1 was co-expressed with an Em-GUS reporter


60
1000
Fig. 13. Effect of GA, ABA and recombinant VP1 on Amy-GUS expression in aleurone of
germinating wild-type (NK508) seeds. Aleurones were bombarded with 4 pg Amy-GUS, 5 pg of
Ubi-LUC and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT and then incubated in a solution containing no
hormones, 10- M GA3, 10-5 M ABA, or 10-6 M GAg and 10- M ABA. Data represent mean (
S.E.M.) of five replicates.
300
Fig. 14. Effect of GA, ABA and recombinant VP1 on Amy-GUS expression in aleurone of
germinating d1 mutant seeds. Methods as in Fig. 13.


37
While ABA is clearly involved in the induction of seed dormancy, its role in the
maintenance of a dormant state beyond seed maturity is less dear. Late In seed development,
ABA concentrations dedine rapidly to a very low amount present in the dry seed. This amount
has been considered insuffident to inhibit germination (Karssen et al., 1983). Also, aba mutant
and wild-type seeds were equally sensitive to applied ABA (Koomeef et al., 1982, 1984),
indicating that the difference in germination capacity between aba mutant and wild-type seeds is
not due to a difference in ABA-sensitivity. It was therefore thought unlikely that ABA or ABA-
sensitivity are involved in maintaining the state of dormancy (Karssen and Lacka, 1985).
Instead, Karssen and Lacka (1985) proposed that the maintenance of dormancy is, at least in
part, mediated by an insensitivity of the seed to GA. This was conduded from evidence showing
that a gradual relief of dormancy by aftenripening, cold or light treatments was correlated with an
increased sensitivity of the seed to the germination-promoting effed of applied GA (Karssen and
Lacka, 1985; Dericx and Karssen, 1993). GA is normally absolutely required for induction of
seed germination, as evident from the fad that seeds of the GA-defident mutant ga-1 do not
germinate under any condition, unless GA is supplied exogenously (Koomeef and van der Veen,
1980). Consequently, insensitivity to GA may present a strict measure to inhibit seed
germination. Consistent with this hypothesis, reduced germination frequendes were reported for
the partially GA-insensitive mutant Gai (Koomeef et al., 1985).
In condusion, the present evidence implies that GA and ABA do not normally interad
diredly at any stage of seed development. ABA in concert with high ABA-sensitivity appears to
be responsible for the indudion of dormancy during seed development, and GA in concert with
GA-sensitivity induced by dormancy-breaking treatments appears to stimulate germination.
Nevertheless, redudion in seed dormancy as a result of low concentrations of ABA (aba) or
insensitivity to ABA (abi1, ab¡2, ab¡3) partially relieved the mature seed form a need for GA to
induce germination. Seeds of double mutants between ga-1 and aba, ab¡1, abi2 or ab/3,
respedively, were capable of germinating (Karssen et al., 1983; Koomeef et al., 1984, Nambara
et al., 1992), whereas ga-1 single mutants have, thus far, never been shown to germinate without


46
Incubation and Extraction of Endosperms Following Bombardment
Following bombardment, 1 ml of a solution containing MS salts and MS vitamins
supplemented with no hormones, 10" M GAj or 10" M GA3 and 10-4 M (or 10-5 M) ABA was
dripped over the endosperms. After 24 h of incubation in darkness, maize endosperms were
ground either individually On experiments using developing seeds) or in bulk from each
bombardment (when germinating seeds were used) with mortar and pestle aided by the addition
of silicon carbide powder in 200-1,000 pi of extraction buffer (0.1 M potassium phosphate (pH
7.8), 2 mM EDTA (pH 8), 2 mM DTT, 5% glycerol). The homogenates were centrifuged to
recover the cell extract. To obtain barley aleurone extract, the aleurone layers were separated
from the endosperms and ground in bulk for each replicate in 200 pi of extraction buffer. The
homogenates had to undergo two rounds of centrifugation to obtain clear cell extract.
Quantification of Transient Expression
Quantitative measurement of GUS activities was performed as described in Jefferson et
al. (1987) with the modification that the substrate MUG was dissolved in the extraction buffer
described above. For determination of luciferase activities, 10 pi aliquots of the extract and 200
pi of reaction buffer (25 mM tricine (pH 7.8), 15 mM MgCI2, 5 mM ATP, 0.05% BSA) were placed
in cuvettes and immediately assayed using a Monolight 2010 luminometer. The luminometer
automatically injects 100 pi of 1 mM luciferin and then counts the emitted photons for 15 s. The
unit of measurement is the Relative Light Unit (RLU).


36
factor may be ABA stored in the dry seed. However, since the dormant and non-dormant seeds
used in one experiment contained similar concentrations of ABA (Skadsen, 1993), it may be that
the higher ABA-sensitivity in dormant seeds relative to non-dormant seeds plays a role in
inhibiting GA-response in the aleurone. However, it cannot be excluded that a factors) other
than ABA inhibits the GA-response in the aleurone of dormant seeds, a-amylase genes are
known to be sensitive to repression by soluble carbohydrates (Yu et al., 1991; Karrer and
Rodriguez, 1992). Possibly, aleurone cells of dormant seeds display a higher sensitivity to the
inhibitory effect of soluble sugars supplied by the starchy endosperm. Alternatively, the
inhibitory factor may be synthesized in the aleurone cells themselves and the presence of the
starchy endosperm is only required to provide an environment of high osmolality which may be
essential to maintain production of the putative inhibitor of GA-response in the aleurone cells.
Arabidoosis
Dormant seeds of Arabidopsis require either several months of dry storage or
rehydration followed by exposure to low temperatures and light in order to break dormancy and
induce germination. Analyses of mutants has demonstrated that initiation of dormancy during
late seed development involves the action of ABA. Even in light, wild-type seeds are normally
incapable of germinating during the seed maturation phase and, for a period of time, after
reaching seed maturity. In contrast, seeds of the ABA-deficient mutant aba gradually acquire
germination capacity during seed development and at maturity germinate at a frequency of
100% in light and 30% in darkness (Karssen et al., 1983; Karssen and Lacka, 1985). Hence aba
mutant seeds display highly reduced dormancy. A germination behavior similar to aba mutant
seeds was observed for the ABA-insensitive mutants abi1, ab¡2 and ab¡3 (Koomeef, 1984),
suggesting that these mutants are non-dormant due to a failure to respond to ABA. However, as
mentioned earlier, strong alleles of ab¡3 do not only cause lack of seed dormancy but also
vivipary, whereas seeds carrying strong alleles of aba have thus far not been shown to be
viviparous. Thus, an ABA-independent function of ABI3 cannot be ruled out.


26
expression of a reporter gene, indicating that cto-acting elements are positioned in the proximal
region of the promoters. Indeed. Skriver et al. (1991) demonstrated that a chimeric construct
containing 69 bp (-189 to -120) of the barley high-pl promoter fused to the 35S TATA box could
impose increased transcription by GA and its suppression by ABA. Moreover, six tandemly
repeated copies of the sequence GGCCGATAACAAACTCCGGCC (21 bp) conferred proper GA-
and ABA-regillation. However, this result could not be confirmed when particle bombardment
was used as the method of transformation (J. Rogers, pers. communication; U. Hoecker,
unpublished results) suggesting that other ris-elements apart from TAACAAA contribute to GA-
regulated transcription. This was confirmed when clustered point mutations were introduced
covering the proximal region of the promoters. Mutations in the pyrimidine box (CCTTTT) or in
the TATCCAC/T box reduced GA-induced transcription to about 20 % of minimal level in the
barley high-pl and low-pi promoters (Gubler and Jacobsen, 1992; F. Gubier, pers.
communication; Lanahan et al., 1992). Thus the entirety of the three conserved elements
appears to be involved in mediating GA-response. Interestingly, in no case could ABA-
responsive elements be separated from GA-responsive elements, suggesting that GA and ABA
function through the same c/s-elements in the a-amylase promoters.
Rogers and co-workers (Lanahan et al., 1992) identified an additional element in the low-
pi promoter that is located between positions -152 and -134, just upstream of the pyrimidine box.
Mutations in this element reduced the level of expression by 96% while retaining significant but
low GA-responsiveness. This region of the promoter (termed 02S element) shows sequence
homology to two well-described motifs: the endosperm box, a conserved element present in
promoters of maize, barley and wheat endosperm protein genes, and the consensus sequence
for binding of the maize Opaque-2 protein which is a leucine zipper protein (bZIP) that is
necessary for transcription of the 22 kDa zein genes. Interestingly, it was found that substitution
of the GA-responsive TAACAGA sequence of the low-pi promoter with an ABA response
element (ABRE) from the rice Rab-16A gene converted the promoter from a GA-upregulated one
into one whose transcription was increased by ABA (Rogers and Rogers, 1992). Thus, the ABRE


85
least two modes of ABA action in the maize seed, a VP1-dependent pathway and a VP1-
independent pathway. Multiple ABA transduction pathways are also Indicated by interactions
between ABA-insensitive mutants of Arabidopsis (Finkelstein and Somerville, 1990; Finkelstein,
1994). This suggests that ABA modulates the activity of diverse regulatory cascades in the
seed.
The second scenario in which VP1 could specifically block GA signal transduction is
supported by the evidence that over-expression of VP1 in aleurone of imbibed barley half seeds
severely reduced GA-induction of Amy-GUS without affecting the basal activity of the a-amylase
promoter (Fig. 11). This suggests that expression of VP1 in the developing seed may be, at
least in part, responsible for the observed GA-insensitivity of cereal and maize a-amylase genes
prior to seed maturity (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al 1990; Oishi and
Bewley, 1990). VP1 displayed full repressing activity in slender (sin) mutant barley seeds (Fig.
12) which are constitutive in GA response of the aleurone (Chandler, 1988; Lanahan and Ho,
1988), suggesting that VP1 functions at a point downstream of the Sin gene product.
With respect to the maize seed, the data do not rule out the possibility that VP1 acts
independently of GA as a developmental repressor of a-amylase genes. Although we have
shown that Amy-GUS is GA-inducible in vp1 mutant aleurones early in development (Table 1), it
is not clear that the high constitutive activities found later in development are entirely attributable
to changes in GA concentration. In contrast to the situation of Himalaya barley seed and other
cereal grains, studies of a-amylase regulation in normal and GA-deficient (d5 mutant) genotypes
of maize indicate that a-amylase induction in germinating maize seeds is largely independent of
GA (Harvey and Oaks, 1974). Consistent with these studies, it was found in the present work
that during germination Amy-GUS is constitutively active in the GA1-deficient d1 mutant of
maize. Because Amy-GUS was fully VP1-repressible in aleurones of developing vpf mutant,
germinating wild-type and germinating df-mutant seeds of maize, it is suggested that VP1-
mediated repression is not necessarily restricted to, nor solely defined by, inhibition of the GA
response. Though the significance of GA in the expression of a-amylase genes needs to be


REVIEW OF LITERATURE
Developmental and Hormonal Regulation of Seed Maturation
The biochemical mechanisms allowing seed tissues to tolerate extreme desiccation
remain unclear. Many studies have implicated soluble sugars in desiccation protection. One of
the suggested functions of soluble sugars is the protection of membranes which are often
considered a primary site of desiccation damage (Crowe et al., 1992). It is thought that hydroxyl
constituents of sugars substitute for water during dehydration and thereby stabilize membrane
structures in the dehydrated state (Crowe et al., 1992). In agreement with this, di- and oligo
saccharides, especially sucrose and in some species raffinose and stachyose, increase in
concentration In maturing seeds (Amuti and Pollard, 1977) and in desiccating pollen grains
(Hoekstra et al., 1989). Their accumulation has been correlated with the acquisition of
desiccation tolerance (Hoekstra and van Roekel, 1988; Koster and Leopold, 1988; Chen and
Burris, 1990; Leprince et al., 1990; Blackman et al., 1992; Crowe et al., 1992). However, recent
comparative studies using desiccation intolerant mutants or recalcitrant (desiccation intolerant)
species found no positive correlation between oligo-saccharide content in the seed and the
development of desiccation tolerance (Ooms et al., 1993; Still et al., 1994). Data of one of the
studies suggest that a low ratio of mono- to oligo-saccharides may be the critical factor rather
than the absolute amount of soluble sugars (Ooms et al., 1993) This suggests that sucrose may
be involved in the formation of glass' during dehydration which is promoted by oligosaccharides
and inhibited by monosaccharides (Ooms et al., 1993). A glass* is a liquid of high viscosity,
such that it stops or slows down all chemical reactions requiring molecular diffusion, and thus,
might conserve tissue structures during dehydration (Bruni and Leopold, 1991).
4


56
construct containing the full length promoter of the wheat Em gene fused to the GUS coding
sequence (Marcotte et al 1989). Recombinant VP1 increased expression of Em-GUS by ca. 5-
fold (Fig. 10) which is consistent with the function of VP1 as a transcriptional activator of Em
(McCarty et al., 1991). In conclusion, these data show that VP1 can repress or activate gene
transcription depending on the promoter context.
Interaction between VP1 and Gibberellic Acid
The well characterized hormonal responses in Himalaya barley aleurone facilitated
further studies regarding the interaction between VP1 and GA. For this purpose, GA response
curves of Amy-GUS expression were determined in aleurones of de-germed imbibed Himalaya
half seeds* (Fig. 11). In the absence of co-expressed VP1, Amy-GUS expression showed a
typical GA-induction. In contrast, when a mixture of Amy-GUS and recombinant VP1 was
introduced into aleurone cells, GA-induction of Amy-GUS expression was reduced by ca. 80%.
Most noticeably, the clearly detectable basal activity of Amy-GUS was not significantly affected
by co-expression of VP1 (Fig. 11, insert). Thus, VP1 only inhibited the GA-dependent activity of
the a-amylase promoter. This implies that VP1 may interfere with the GA signalling pathway.
Recessive mutations that cause constitutive GA-response have been identified in barley
and a few other species (Ross. 1994). Barley slender (sin) mutant plants are characterized by
excessive elongation of stem and leaf tissues and constitutive expression of hydrolytic enzymes
in the aleurone of imbibed half seeds in the absence of exogenous GA. The mutant phenotype
suggests that the Sin gene encodes a negative regulator that is normally inactivated by GA
(Chandler, 1988; Lanahan and Ho, 1988). To test whether VP1 inhibitory function depends on
the presence of the SLN protein, aleurones of sin mutant half seeds were co-bombarded with
Amy-GUS and recombinant VP1. Figure 12 shows that VP1-mediated repression of Amy-GUS
was as effective in sin mutant aleurones as in wild-type aleurones.


51
no hormones
GA
+GA +ABA
Amy-GUS / LUC *10 4 [pmoles MU/h/RLU]
Fig. 6. Effect of VP1 over-expression and ABA on Amy-GUS expression in vp1-R mutant
aleurone. vp1-R mutant aleurones from kernels harvested 26 DAP were bombarded with 10 pg
of Amy-GUS, 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT (for no-VP1 controls).
Following bombardment, a solution containing no hormones, 106 M GA3 or 10'6 M GA3 and 10*
4 M ABA was applied to the kernels. Numbers behind bars represent means of five replicates.
Error bars show S.E.M.


52
transcription factors because no inhibitory effect of co-bombarded VP1 on 35S-Sh-GUS or
UbiquHin-Luciferase expression was observed (data not shown). Moreover, VP1 caused trans-
activation of positively regulated reporter constructs, Em-GUS and C1-Sh-GUS, in aleurone cells
using similar bombardment conditions (S. Cocciolone and D.R. McCarty, unpublished results).
Interaction between VP1 and Abscisic Acid
in concert with VP1, the hormone ABA plays an important role during seed maturation
(McCarty and Carson, 1991). Moreover, ABA functions as an inhibitor of a-amylase expression
in germinating cereal seeds (Jacobsen and Chandler, 1987). This suggests that ABA might also
be involved in repression of a-amylase genes in the developing seed. Therefore, possible
interactions between ABA and VP1 in repressing Amy-GUS were analyzed.
ABA was effective in blocking Amy-GUS expression in vp1 mutant aleurone (Fig. 6).
This indicates that repression by ABA does not require VP1. In combination, ABA and VP1 over
expression produced a roughly additive effect (Fig. 6).
To test whether a-amylase repression by VP1 is dependent on ABA, recombinant VP1
was over-expressed in aleurone of developing vp1,vp5 double mutant kernels that are deficient
for ABA biosynthesis. Figure 7 shows that VP1 was highly effective in repressing Amy-GUS in
vp5 mutant background. While it cannot by ruled out that maternal ABA derived from the vp5/+
parent plant may be sufficient for VP1 function, it is suggested that VP1-mediated repression of
Amy-GUS expression does not require ABA. This would be consistent with the finding that VP1
also functions in aleurone of germinating seeds (see below) where ABA levels are very low
(Oishi and Bewley, 1990). Taken together, these data suggest that ABA and VP1 inhibit Amy-
GUS expression independently.


7
deficient mutants, 2) ABA-insensitive mutants and 3) mutants affected in a thus far unknown
mechanism.
ABA deficient mutants
Five ABA-deficient mutants have been identified in maize (vp2, vp5, vp7, vp8, vp9, Neill
et al., 1986). Four of these mutants (vp2, vpS, vp7, vp0) lack carotenoids in addition to ABA
(Robertson, 1955). These mutants accumulate various intermediates of the carotenoid
biosynthesis pathway, indicating that they have lesions in carotenoid biosynthetic enzymes
(Robertson et al., 1978). The deficiency in both carotenoids and ABA confirms that carotenoids
are the precursor for ABA-biosynthesis in plants. Phenotypically, these mutants are viviparous,
display a pale yellow to white coloring of the normally orange-yellow endosperm and form a
lethal, white seedling which is deficient in chlorophyll due to photobleaching caused by the lack
of carotenoids (Robertson, 1955; Anderson and Robertson, 1960).
Only one viviparous mutant (vp8) is deficient in ABA without affecting carotenoid
synthesis. vp8 mutant seedlings are viable but form plants of severely dwarfed stature
(Robertson, 1955). The biochemical lesion of this mutant is unknown. It may be deficient in a
later step in the ABA-biosynthesis pathway that is involved in the conversion of the carotenoid
xanthophyll to ABA (Zeevaart and Creelman, 1988). Whether the lack of ABA causes dwarfism
has not been examined thus far.
Vivipary in maize is determined by the genotype of the embryo and is entirely
independent of the genotypes of the mother plant or the endosperm. Viviparous seeds
segregate at the expected ratio on a heterozygous mother plant, indicating that ABA contributed
by the mother plant mainly early in seed development does not play a role in preventing
vivipary. The relative contribution of embryo and endosperm in preventing vivipary can be
assessed by the use of TB-translocations which make it possible to generate seeds with embryo
and endosperm of different, genetic constitution (e.g. a seed with a vp5 mutant embryo and a
wild-type endosperm, or vice versa; Roman, 1947; Beckett, 1993). In these experiments, seeds
with a vp5 mutant embryo and a wild-type endosperm were viviparous, while seeds with a wild-


53
EHHD-t- 35S-Sh-CAT
Mi + 35S-Sh-VP1
Fig. 7. Co-expressed VP1 inhibited Amy-GUS in aleurone of developing vp1lvp5 double mutant
seeds that are deficient for ABA biosynthesis. Kernels were harvested 24 DAP. Bombardments
were performed as described in Fig. 6. Following bombardment, kernels were incubated in no
hormones or 10'6 M GA3. Data represent mean ( S.E.M) of 7-8 replicates.
Effector Construct
Amylase-GUS / Ubi-LUC 104 (S.E.M.)
[pmoles MU/hr/RLU]
Maize Seeds
Barley Seeds
Amy-GUS
JR303
Amy-GUS
JR303
35S-Sh-CAT (Control)
91 <17>
1.19 i*0 35)
247 (36)
65 (5.2)
35S-Sh-Vp1
10.5 i*06*
0.07 <*014)
70(16)
25 i*6-4)
Fig. 8. Co-expressed VP1 inhibited Amy-GUS and JR303 in aleurone of germinating maize and
barley seeds. Aleurones of imbibed seeds were bombarded with Amy-GUS (maize: 4 pg; barley:
2 M9) or JR303 (maize: 10 pg; barley: 5 pg), 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35S-
Sh-CAT. Post-bombardment, kernels were incubated in 10'6 M GA3. Data represent mean (
S.E.M) of 3-5 replicates.
I


68
Re. Amy-GUS / LUC
Effector Construct Maize Barley
VP1-WT
103/104
101/100
VP1-McW
Control (no VP1)
H
A
II

III
100
100
4-12
15-30
21 (*4>
27 <*3)
18(2)
23 (4)
32 (*9)
13 (*-5)
Fig. 19. Deletion analysis of the VP1 protein: Part II. Aleurones of maize and barley
germinating seeds were bombarded with 2-5 pg of Amy-GUS, 5 pg of Ubi-LUC and 5 pg of
effector construct and then cultured in 10-6 M GA3. Data represent mean ( S.E.M.) of 3-5
replicates


13
are found on the adaxial surfaces of its cotyledons. Trichomes normally form only on leaves,
stems and sepals, but not on cotyledons. Hence, tecf and fus3 cotyledons are considered to be
partially transformed into leaves, which gave two of the mutants their name (fee, 'leafy
cotyledons*, Meinke, 1992). The tec2 mutant also exhibits leafy cotyledons and accumulation of
anthocyanin, but differs from tecf and fus3 in that seeds are tolerant to desiccation and non-
viviparous, and seedlings often appear distorted in shape (elongated hypocotyl, curled
cotyledons) (Meinke et al., 1994). Unlike the abi3 mutant, germination of tecf and fus3 mutant
seeds is inhibited by ABA, indicating that they retain normal sensitivity to ABA (Keith et al., 1994;
Meinke et al., 1994). Proof for normal ABA synthesis in these mutants is still lacking. However,
since ABA-deficient mutants have thus far not been shown to exhibit leafy cotyledons or
accumulation of anthocyanins, it is unlikely that a possible lack of ABA would be the sole cause
of the mutant phenotype. Nevertheless, the role of ABA in these mutants remains to be
examined.
To investigate the interaction between ab/3 and leafy cotyledon mutants, double mutants
were constructed. abi3/led and ab¡3/fus3 double mutant seeds were highly viviparous,
insensitive to ABA, exhibited leafy cotyledons and accumulated large amounts of anthocyanins
(Meinke et al., 1994, BSumlein et al., 1994). The additive effect of ab/3 and Iec1/fus3 in the
double mutants suggests that ab/3 and tecf or fus3, respectively, are altered in distinct pathways.
Consequently, suppression of precocious germination requires at least ABA, developmental
factors controlling ABA-sensitivity and the leafy cotyledon-factors whose interactions with ABA,
however, remain to be analyzed.
Analysis of Gene Expression
Late stages of seed formation are correlated with the expression of characteristic genes
which were analyzed first and very extensively, in cotton embryos (Galau et al., 1986, 1987;
Hughes and Galau, 1991). Based on changes in the levels of specific sets of cotton mRNAs,
late seed development has been categorized into several stages (Galau et al., 1991). The


VP1-MEDIATED REPRESSION OF ALPHA-AMYLASE GENES
IN DEVELOPING MAIZE ALEURONE
BY
UTE HOECKER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1995

ACKNOWLEDGEMENTS
I am deeply grateful to Dr. Donald McCarty and Dr. Indra Vasil for their valuable advice,
guidance and encouragement throughout the course of this study. I wish to thank Drs. Karen
Koch, Nigel Richards, Alice Harmon and Bill Gurley for serving as members of my committee. I
also thank all past and present members of both laboratories for their helpful discussions as well
as for providing an enjoyable and lively work atmosphere. I am especially grateful to Lennie
Rosenkrans and Dr. Mark Taylor for their immeasurable technical assistance. I wish to extend
my thanks to Ms. Elaine Summers for her frequent help in dealing with the administrative side of
Graduate School.
ii

TABLE OF CONTENTS
page
ACKNOWLEDGMENTS H
ABSTRACT iv
INTRODUCTION 1
REVIEW OF LITERATURE 4
Developmental and Hormonal Regulation of Seed Maturation 4
Isolation of Mutants Affected In Seed Maturation 6
Analysis of Gene Expression 13
The Aleurone Germination Response in Cereal Seeds 21
Hormonal Regulation 22
The a-Amylase Genes 23
The Organization of a-Amylase Promoters 25
GA and ABA Signal Transduction 28
The Developmental Switch from Seed Maturation to Seed Germination 31
Quiescent Seeds 33
Dormant Seeds 34
MATERIALS AND METHODS 39
Plant Material 39
Plasmid Constructs 41
Particle Bombardment and Transient Expression 43
RESULTS 47
Repression of Hydrolase Genes by VP1 in Aleurones of Developing Maize Seeds 47
Interaction between VP1 and Abscisic Acid 52
Over-expression of VP1 in Aleurones of Germinating Maize and Barley Seeds 54
Interaction between VP1 and Gibberellic Acid 56
Role of the Embryo in Repression of a-amylase Genes in the Aleurone 61
Functional Analysis of the VP1 Protein 64
DISCUSSION 74
SUMMARY AND CONCLUSIONS 90
REFERENCES 91
BIOGRAPHICAL SKETCH
104

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
VP1-MEDIATED REPRESSION OF ALPHA-AMYLASE GENES
IN DEVELOPING MAIZE ALEURONE
By
Ute Hoecker
August 1995
Chairman: I.K. Vasil
Cochairman: D.R. McCarty
Major Department: Plant Molecular and Cellular Biology
The Viviparous-1 (VP1) transcriptional activator of maize is required for abscisic acid-induction
of maturation-specific genes late in seed development. In the presented work, it is shown that, in
addition, VP1 inhibits precocious induction of the germination-specific a-amylase genes in aleurone cells
of the developing seed. In developing seeds of the somatically unstable vp1-m2 mutant, hydrolase
activity was de-repressed specifically in endosperm sectors underlying vp1 mutant aleurone. Moreover,
in transient expression experiments based on particle bombardment of aleurone tissue, a barley high-pl
a-amylase promoter-GUS fusion construct (Amy-GUS) was induced in developing vp1 mutant aleurone
cells but not in wild-type aleurone cells. A direct role of VP1 in repression of Amy-GUS is suggested
from the finding that co-expression of recombinant VP1 in vp1 mutant aleurone cells strongly inhibited
expression of Amy-GUS. Hence, VP1 expression in the developing seed appears to integrate the control
of two developmental programs, seed maturation and seed germination.
Over-expression of VP1 also inhibited Amy-GUS expression in aleurones of wild-type
germinating maize and barley seeds. In barley aleurone cells, VP1 specifically repressed induction of
Amy-GUS by gibberellic acid (GA), while in maize aleurone tissue, VP1 inhibited a GA-dependent as well
IV

as an apparent GA-independent activity. Deletion of the addle transcriptional activation domain of VP1
did not affect the inhibitory activity, indicating that VP1 has a discrete repressor function. Further
deletion analysis of VP1 showed that domains essential for repression of Amy-GUS are distinguishable
from domains required for activation of the maturation-related genes Em and C1.
The role of the embryo in the expression of Amy-GUS in developing maize aleurone cells was
studied. Amy-GUS was de-re pressed in vp1 mutant aleurone in seeds that either carried a viviparous
embryo or aborted the embryo earty in development but not in seeds with a normal, non-viviparous
embryo. This suggests that a normal embryo contributes a diffusible signal with inhibitory effect on Amy-
GUS expression in the aleurone. Amy-GUS was partially de-re pressed in wild-type aleurone cells of
embryo-less seeds, suggesting that both Vp1 expression in the aleurone and a non-viviparous embryo
are required for complete repression of a-amylase genes in the developing maize aleurone.
v

INTRODUCTION
The formation of seeds is a unique characteristic of higher plants which promotes
dispersal of the species and allows interruption of the life cycle during unfavorable environmental
conditions. To survive in the dehydrated state, plant embryos undergo an adaptation process
during late stages of seed formation (maturation phase) which renders them tolerant to
desiccation and gradually causes arrested growth. In maize and other cereals, the outermost
layer of the seed endosperm (aleurone layer) also undergoes a maturation process and remains
viable through desiccation.
Seed maturation is associated with the activation of a variety of genes encoding storage
proteins and various hydrophilic, late-embryogenesis-abundant (LEA) proteins which possibly
function as desiccation protectants (Dure et al. 1989; Skriver and Mundy 1990). Analysis of
viviparous mutants in maize has demonstrated that the developmental program of seed
maturation is controlled by at least two factors, the hormone abscisic acid (ABA) and the product
of the Viviparous-1 (Vp1) gene (Robertson 1955; Neill et al. 1986). Developing vpf mutant
embryos are distinct from ABA-deficient embryos in that they exhibit a reduced sensitivity to
ABA in culture (Robichaud et al., 1980; Robichaud and Sussex, 1986). In addition to causing
vivipary, the vp1 mutation blocks synthesis of anthocyanins in embryo and aleurone tissues
(Robertson, 1955; Dooner, 1985). The vpf mutant phenotype is restricted to seed tissues.
Mutant embryos rescued prior to desiccation develop into apparently normal, fully fertile plants
with normal patterns of anthocyanin accumulation.
The Vp1 gene was cloned by transposon tagging (McCarty et al., 1989a). It encodes a
2500-nucleotide mRNA that is expressed specifically in embryo and endosperm tissues of the
developing seed. Within the endosperm, Vp1 expression may be limited to the aleurone layer
because so far no mutant phenotype has been detected in the starchy endosperm. This is
1

2
consistent with the relatively low abundance of Vp1 message in whole endosperm extracts
(McCarty et al., 1989a).
Vp1 encodes a novel, 73 kD protein with a functional acidic transcriptional activation
domain (McCarty et al., 1991). Over-expression of VP1 in maize protoplasts frans-activated
reporter constructs containing late-embryogenesis-specific promoters: C1, a maize gene that
encodes a transcription factor required for anthocyanin synthesis in the seed, and Em, a wheat
LEA gene (Hattori et al., 1992; McCarty et al., 1991). In agreement with the phenotype of ABA-
deficient mutants, VP1-activation of Em was strongly dependent on the presence of exogenous
ABA (McCarty et al., 1991). These functional data confirm that VP1 plays a central role in the
induction of seed maturation.
Following imbibition of mature non-dormant seeds, expression of maturation-specific
genes is terminated and expression of a new set of genes related to the developmental program
of seed germination is executed (Comai and Harada, 1990). In rehydrated cereal seeds, the
germination-specific a-amylase genes which encode starch-hydrolyzing enzymes are induced in
the aleurone cells by the hormone gibberellic acid (GA) that is secreted by the embryo early in
germination (Jacobsen and Chandler, 1987). They are constitutively expressed in de-germed
seeds of the barley GA-response mutant slender (Chandler, 1988; Lanahan and Ho, 1988), and
their induction can be antagonistically inhibited by application of ABA (Jacobsen and Chandler,
1987).
Expression of the normally consecutive programs of seed maturation and seed
germination is under strict developmental control. Precocious induction of germination-related
events prior to seed maturity appears to be actively repressed. In developing seeds of cereals
and maize, no a-amylase activities are found prior to seed maturity (Evans et al., 1975; Nicholls,
1979; Comford et al., 1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Moreover, a-
amylase genes are unresponsive to applied GA (Nicholls, 1979; Comford et al., 1986; Garcia-

3
Maya at al., 1990; Olshi and Bewley, 1990). It ha been suggested that the presence of ABA in
developing seeds is responsible for the inhibition of a-amylase genes at this developmental
stage (King, 1976). However, treatment of maize developing seeds with the ABA synthesis
inhibitor flouridone was not sufficient to sensitize the aleurone cells to GA, suggesting the action
of additional factors in repressing a-amylase genes in the developing seed (Oishi and Bewley,
1990).
The objective of this study was to elucidate a role of VP1 in the negative regulation of a-
amylase gene expression in the developing maize seed. It is demonstrated that VP1 in
addition to activating seed maturation programs blocks precocious induction of germination-
specific a-amylase genes in the developing maize seed. A somatically unstable vp1 mutant is
described that displays de-repression of hydrolase activity specifically in vp1 mutant sectors of
the aleurone. Using a transient expression approach, it is shown that expression of recombinant
VP1 in aleurone cells of maize and barley strongly inhibits expression of an a-amylase promoter-
GUS reporter gene (Amy-GUS). Evidence is provided indicating that VP1 specifically represses
GA-induction of Amy-GUS in aleurone of germinating barley seeds. It is also shown that deletion
of the acidic activation sequence of VP1 does not affect VP1 repressor activity, indicating that
VP1 has a discrete repressor function. Thus, it is suggested that the coupled activator and
repressor functions of VP1 play a key role in integrating the control of the normally not
simultaneously occurring maturation and germination programs in the seed.

REVIEW OF LITERATURE
Developmental and Hormonal Regulation of Seed Maturation
The biochemical mechanisms allowing seed tissues to tolerate extreme desiccation
remain unclear. Many studies have implicated soluble sugars in desiccation protection. One of
the suggested functions of soluble sugars is the protection of membranes which are often
considered a primary site of desiccation damage (Crowe et al., 1992). It is thought that hydroxyl
constituents of sugars substitute for water during dehydration and thereby stabilize membrane
structures in the dehydrated state (Crowe et al., 1992). In agreement with this, di- and oligo
saccharides, especially sucrose and in some species raffinose and stachyose, increase in
concentration In maturing seeds (Amuti and Pollard, 1977) and in desiccating pollen grains
(Hoekstra et al., 1989). Their accumulation has been correlated with the acquisition of
desiccation tolerance (Hoekstra and van Roekel, 1988; Koster and Leopold, 1988; Chen and
Burris, 1990; Leprince et al., 1990; Blackman et al., 1992; Crowe et al., 1992). However, recent
comparative studies using desiccation intolerant mutants or recalcitrant (desiccation intolerant)
species found no positive correlation between oligo-saccharide content in the seed and the
development of desiccation tolerance (Ooms et al., 1993; Still et al., 1994). Data of one of the
studies suggest that a low ratio of mono- to oligo-saccharides may be the critical factor rather
than the absolute amount of soluble sugars (Ooms et al., 1993) This suggests that sucrose may
be involved in the formation of glass' during dehydration which is promoted by oligosaccharides
and inhibited by monosaccharides (Ooms et al., 1993). A glass* is a liquid of high viscosity,
such that it stops or slows down all chemical reactions requiring molecular diffusion, and thus,
might conserve tissue structures during dehydration (Bruni and Leopold, 1991).
4

5
A second characteristic that may be involved in rendering the seed tolerant to
desiccation is the synthesis and accumulation of specific proteins late in seed development Qate-
gmbryogenesis-flbundant proteins, LEAs). The direct function of LEAs is unknown but based on
their high degree of hydrophilicity (Dure et al., 1989) they are assumed to stabilize the structure
of cellular proteins during dehydration (Skriver and Mundy, 1990; Dure, 1993).
Equally uncertain remain the mechanisms that arrest embryo growth and prevent
precocious germination prior to seed maturity. Because immature embryos excised from the
seed and placed in culture are capable of germinating readily, precocious germination of the
embryo in vivo may be actively suppressed by a process that is dependent on an intact seed.
Evidence from embryo culture experiments has implicated two factors in suppression of
precocious germination: restricted water uptake (low water potential in the seed) and the
hormone abscisic acid (ABA). Both factors, when imposed on cultured embryos, inhibited
germination (reviewed in Quatrano, 1987; Kermode, 1990). Indeed, the osmotic potential of
developing soybean embryos has been shown to be even more negative than that of the
osmoticum used to inhibit germination of isolated embryos (Xu et al., 1990). Similarly, ABA
concentrations increase early in the seed maturation phase (Quatrano, 1987) and may thus play
a role in arresting embryo growth.
In summary, the physiological processes responsible for acquisition of desiccation
tolerance and arrest in embryo growth are poorly understood. In the past 15 years, the focus of
research has shifted to the identification of regulatory factors that control the activities of late
embryogeny, especially at the level of gene expression. From its discovery in the 1950s, ABA
has been a factor of interest because a rise in seed ABA concentration correlates well with the
onset of maturation events. Normally, ABA concentrations peak at the time of maximum dry
weight accumulation in the seed and then decrease to low concentrations towards seed maturity
(Quatrano, 1987). A function of ABA in initiating maturation events and suppressing precocious
germination of the developing embryo was confirmed by analyses of mutants defective in late
embryogeny and studies of gene expression.

6
Isolation of Mutants Affected In Seed Maturation
Genetic deficiencies in seed maturation manifest themselves In precocious germination
(vivipary) or reduced dormancy. Severely affected mutants exhibit additional features, such as
intolerance to desiccation and reduced accumulation of seed storage proteins and LEAs.
Mutants have been most intensively isolated and analyzed in maize and Arabktopsis.
Maize has been a model species for studying genetics for many years, mostly because its
monoecious flower structure in combination with self fertility has allowed easy outcrossing and
selfing. Moreover, the identification of several maize transposable elements (Ac/Ds, Spm/En,
Mu) has made it possible to generate transposable element-induced mutants and subsequently
clone the mutated locus using the transposon as a tag. Arabidopsis has become a powerful
model species for a variety of reasons (such as short generation time, small genome size
(Meyerowitz, 1987; Meyerowitz, 1994). Rather recently, cloning of mutated genes has been
achieved by Agrobacterium transformation-mediated T-DNA tagging or chromosome walking
(positional cloning).
Many maize viviparous mutants that arose spontaneously were collected and described
already during the first half of this century (Eyster, 1931; Mangelsdorf, 1930; Robertson, 1955).
Transposon-tagged mutants were induced more recently (e.g. McCarty et al., 1989a,b). In
Arabidopsis, many mutants displaying vivipary or reduced dormancy have been isolated by
screening chemically mutagenized seed for germination on medium containing ABA at a
concentration that inhibits germination of wild-type seeds. Other mutants have been isolated by
screening for the ability to germinate in the absence of GA which is normally required. Here, GA
contents in the seed were reduced either by treatment with a GA-synthesis inhibitor or by using a
mutant deficient in GA-biosynthesis. Recently, screening of transgenic lines produced by
Agrobacterium-mediated seed transformation (Feldmann, 1991) for mutants defective in late
embryogeny has revealed additional loci that control induction of seed maturation/suppression of
precocious germination. Taken together, the identified mutants fall into three classes: 1) ABA-

7
deficient mutants, 2) ABA-insensitive mutants and 3) mutants affected in a thus far unknown
mechanism.
ABA deficient mutants
Five ABA-deficient mutants have been identified in maize (vp2, vp5, vp7, vp8, vp9, Neill
et al., 1986). Four of these mutants (vp2, vpS, vp7, vp0) lack carotenoids in addition to ABA
(Robertson, 1955). These mutants accumulate various intermediates of the carotenoid
biosynthesis pathway, indicating that they have lesions in carotenoid biosynthetic enzymes
(Robertson et al., 1978). The deficiency in both carotenoids and ABA confirms that carotenoids
are the precursor for ABA-biosynthesis in plants. Phenotypically, these mutants are viviparous,
display a pale yellow to white coloring of the normally orange-yellow endosperm and form a
lethal, white seedling which is deficient in chlorophyll due to photobleaching caused by the lack
of carotenoids (Robertson, 1955; Anderson and Robertson, 1960).
Only one viviparous mutant (vp8) is deficient in ABA without affecting carotenoid
synthesis. vp8 mutant seedlings are viable but form plants of severely dwarfed stature
(Robertson, 1955). The biochemical lesion of this mutant is unknown. It may be deficient in a
later step in the ABA-biosynthesis pathway that is involved in the conversion of the carotenoid
xanthophyll to ABA (Zeevaart and Creelman, 1988). Whether the lack of ABA causes dwarfism
has not been examined thus far.
Vivipary in maize is determined by the genotype of the embryo and is entirely
independent of the genotypes of the mother plant or the endosperm. Viviparous seeds
segregate at the expected ratio on a heterozygous mother plant, indicating that ABA contributed
by the mother plant mainly early in seed development does not play a role in preventing
vivipary. The relative contribution of embryo and endosperm in preventing vivipary can be
assessed by the use of TB-translocations which make it possible to generate seeds with embryo
and endosperm of different, genetic constitution (e.g. a seed with a vp5 mutant embryo and a
wild-type endosperm, or vice versa; Roman, 1947; Beckett, 1993). In these experiments, seeds
with a vp5 mutant embryo and a wild-type endosperm were viviparous, while seeds with a wild-

8
type embryo and a vp5 mutant endosperm were not (Robertson, 1952). Similar results were
obtained with other ABA-deficient mutants (Robertson, 1955). Thus, vivipary in these mutants is
entirely conditioned by the lack of ABA production in the embryo. The endosperm does not play
an active role in preventing vivipary of a genetically viviparous embryo. This is consistent with
later findings showing that the embryo is the major source of ABA produced in the developing
seed (Zeevaart and Creelman, 1988).
In Arabidopsis, only one ABA-deficient mutant (aba) has been identified. It was isolated
in a genetic screen selecting for the ability to germinate of the normally non-germinating GA-
deficient ga-1 mutant (Koomeef et al., 1982). The aba mutant is impaired in the epoxidation of
the carotenoid zeaxanthin (Duckham et al., 1991) and thus displays normal accumulation of
carotenoids. No carotenoid deficient mutants have been isolated in any genetic screen, which
may reflect the predicted lethality of such mutations. The aba mutant produces plants that show
increased withering of stems, leaves and siliques and an enhanced rate of water loss which is
probably caused by reduced stomata closure upon water stress (Koomeef et al., 1982). aba
mutant seeds exhibit strongly diminished seed dormancy. Wild-type Arabidopsis seeds normally
require cold and light treatments to break imposed seed dormancy and allow germination to
occur, whereas a high percentage of aba mutant seeds germinated readily without a need for
dormancy-breaking treatments (Koomeef et al., 1982). However, in contrast to ABA-deficient
mutants in maize, even severe aba alleles that reduce ABA levels in the seed below the level of
detection produce seeds that are non-viviparous and desiccation tolerant (Koomeef et al., 1982).
Hence, ABA may not be required for the induction of desiccation tolerance in Arabidopsis.
However, leakiness of the aba mutant cannot be ruled out. The nature of the performed mutant
screen which selected for the ability of mature, dry seeds to germinate may not allow the
identification of more severely affected mutants. Possibly, residual, very low concentrations of
ABA that may be present in aba mutant seed are sufficient to prevent vivipary (Koomeef et al.,
1989). To test this, a mutant screen could be performed that selects for the ability of seeds to

9
germinate precociously late in seed development. Such a screen has proven successful in
isolating abi3 and other mutants (Keith et al.t 1994).
Wild-type developing seeds of Arabidopsis accumulate ABA as a dual peak, an earlier
maternally-derived one and a later embryo-derived one (Karssen et al., 1983). Reciprocal
crosses between wild-type and aba mutant plants demonstrated normal dormancy in the absence
of maternal ABA but not in the absence of embryonic ABA. Moreover, induction of dormancy,
as judged from the acquired inability of the developing seed to germinate precociously,
correlated well with the later peak of ABA accumulation (Karssen et al., 1983). Hence,
acquisition of a dormant state is dependent on ABA produced by the embryo and is normally
independent of ABA provided by the mother plant.
ABA-insensitive mutants
ABA-insensitive mutants of maize and Arabidopsis accumulate normal or higher
concentrations of ABA in developing seeds as compared to wild-type (Neill et al., 1986; 1987;
Koomeef et al., 1984). However, while the mutant phenotype of ABA-deficient mutants can be
complemented by exogenous application of ABA, ABA-insensitive mutants continue to display
vivipary or reduced dormancy in the presence of added ABA (Raubichaud et al., 1980;
Raubichaud and Sussex, 1986; Koomeef et al., 1984). In a maize mutant, it was shown that the
reduced sensitivity to ABA was not caused by a deficiency in ABA transport or metabolism
(Raubichaud and Sussex, 1986). Thus, ABA-insensitive mutants are likely to be affected in ABA
signal transduction.
A single locus (Vp1) regulating ABA-sensitivity has been identified in maize. vp1 mutant
embryos do not acquire desiccation tolerance and germinate precociously on the ear producing
green seedlings. The vp1 mutation affects only seed tissues. When rescued and transferred to
soil prior to desiccation, mutant seedlings form a normal appearing mature plant. Interestingly,
this mutation causes a pieiotropic phenotype. Besides displaying vivipary, vp1 mutant seeds fail
to accumulate anthocyanin pigments in embryo and aleurone tissues (Robertson, 1955).
Consistent with this phenotype, activities of enzymes catalyzing anthocyanin biosynthesis were

10
not detectable In mutant seed tissues (Dooner, 1985). The lack of anthocyanin pigments Is not
likely to be a result of the reduced ABA-sensItivity of this mutant. ABA-defldent mutants
accumulate normal amounts of anthocyanins, implying that ABA is not required for pigment
formation. Furthermore, separation of the two phenotypes was observed in seeds carrying the
vp1-McWhirter allele. Those seeds are unpigmented but non-viviparous (Coe et al, 1978).
Therefore, the pleiotropic phenotype of the vp1 mutant implies that in evolution, two processes,
suppression of precocious germination and production of anthocyanins, have come under the
control of a single protein.
As in ABA-deficient mutants, the viviparous phenotype of the vp1 mutant is entirely
determined by the genotype of the embryo (Robertson, 1955). Similarly, anthocyanin deficiency
in embryo or aleurone solely reflects lack of functional VP1 in the respective tissue (Robertson,
1955). Thus, the failure to accumulate pigments in vp1 mutant aleurone is not a direct or indirect
result of precocious induction of germination. Cell autonomous function of VR1 in the aleurone
was demonstrated in a transposable element-induced, somatically unstable mutant (vp1-mum1).
Homozygous vp1-mum1 seeds exhibit small sectors in the aleurone that have regained VP1
function due to excision of the Robertsons Mutator transposable element. These revertant
sectors, recognizable by their pigmentation, can be as small as single cells, indicating that VP1
function does not result in production of a diffusable factor that might induce anthocyanins in
neighboring cells (McCarty et al., 1989a).
The Vp1 gene was cloned by transposon tagging using the vp1-mum1 allele (McCarty et
al. 1989a). It encodes a 2500 bp mRNA that is translated into a 73 kD protein. Vp1 is expressed
specifically in embryo and endosperm tissues of the developing seed. Within the endosperm,
Vp1 expression is likely to be restricted to the aleurone, as suggested from the low abundance of
Vp1 mRNA detected in whole endosperm extracts (McCarty et al. 1989a) and the apparent
absence of a mutant phenotype in vp1 mutant endosperm. The Vp1 transcript is present in the
seed as early as 10 days after pollination (DAP), reaches maximum accumulation at 16 DAP and
decreases in abundance towards seed maturity. No Vp1 expression was detected in germinating

11
seeds, root or shoot tissues (McCarty et at. 1989a; Carson, 1992). Hence, the expression pattern
of Vp1 is highly consistent with the seed-specific phenotype of the vpf mutant.
The Vp1 gene consists of six exons and five introns. Apart from putative VP1 homologs
cloned from barley, rice and Arabidopsis (M. Stoll and D.R. McCarty, unpublished results; Hattori
et al., 1994; Giraudat et al., 1992), the sequence of VP1 shows no significant homologies to any
known protein sequences, suggesting that VP1 is a novel protein. The N-terminus of VP1 is
predicted to form two negatively charged amphipathic helices, a feature which is characteristic of
many bacterial and eukaryotic transcriptional activators (Ptashne, 1988). Indeed, this region of
VP1 was found capable of functionally replacing the addic activation domain of the bacterial
transcription factor GAL4 in a eukaryotic gene expression system (McCarty et al., 1991). This
confirmed that the acidic region of VP1 has transcriptional activator function and suggested that
VP1 may function as a regulatory protein in controlling seed maturation and anthocyanin
accumulation.
In Arabidopsis, mutants displaying reduced sensitivity to ABA have been identified using
genetic screens selecting for the ability of seeds to germinate on medium containing at least
3 |j.M ABA, a concentration that inhibits germination of wild-type seeds. In such screens, five loci
controlling ABA-sensitivity have been identified: Abi1, Abi2, Abi3 (Koomeef et al., 1984), Abi4
and Abi5 (Finkelstein, 1994). Mutations in any of these loci confer reduction in seed dormancy.
However, while the phenotype of ab¡3, abi4 and abi5 mutants is restricted to seed tissues, abi1
and abi2 mutants are also impaired in stomatal regulation and a variety of stress responses in
vegetative tissues (Koomeef et al., 1984; Finkelstein, 1994; Chandler and Robertson, 1994).
Interestingly, when the abi3-1 mutant was crossed to the ABA-defident aba mutant, seeds of the
resulting double mutant were desiccation intolerant, remained green and frequently displayed
vivipary (Koomeef et al., 1989). The phenotype of the double mutant suggests that the abi3-1
allele may be leaky and allow some ABA-responsiveness. Hence, additional reduction in seed
ABA concentrations may be necessary to obtain a viviparous phenotype. Indeed, severe abi3
mutants were isolated that were phenotypically similar to the abi3-1/aba double mutant,

12
confirming that a strong abi3 allele conferring high insensitivity to ABA is sufficient to cause
vivipary (Nambara et al., 1992; Ooms et al., 1993).
The Abi3 gene was cloned by chromosome walking (Giraudat et al., 1992). The
predicted protein of 79.5 kD displays discrete regions of sequence homology to the maize VP1
protein. Since there are also phenotypic similarities between the abi3 and vp1 mutant at least
with respect to seed-specific insensitivity to ABA ABI3 and VP1 are likely to have similar
functions in regulating ABA response in the seed. However, the functions of ABI3 and VP1 differ
in so far that ABI3 is required for seed dormancy in Arabidopsis while VP1 does not Impose seed
dormancy in maize. Conversely, VP1, but not ABI3, induces synthesis of anthocyanins in the
seed. Whether these phenotypic differences reflect differences in sequence between ABI3 and
VP1 or the differential involvement of other factors remains to be determined.
The Abi 1 gene was doned in two laboratories by chromosome walking (Leung et al.,
1994; Meyer et al., 1994). At its C-terminus, the predided ABM protein (47.5 kD) displays
sequence similarity with the 2C dass of serine-threonine protein phosphatases from rat and
yeast. Its N-terminus exhibits features typical for a Ca++-binding site (EF hand). Hence, ABI1
may fundion as a Ca++-dependent protein phosphatase. Indeed, regulation of stomatal aperture
by ABA involves Ca++ as a second messenger and protein phosphorylation events (Blatt and
Thiel, 1993; Luan et al., 1993). How ABI1 may regulate ABA-indudion of seed dormancy is thus
far unknown.
Mutants affeded in a thus far unknown mechanism
Three mutants of Arabidopsis (Iec1, tec2, fus3) have been isolated that are non-donmant
but normal in their response to ABA (Meinke, 1992; Keith et al., 1994; Meinke et al., 1994;
Bumlein et al., 1994). Iec1 and fus3 have similar phenotypes. Immature mutant seeds
germinate readily when placed in culture and display occasional vivipary when left to mature in
the siliques. Furthermore, they are intolerant to desiccation and accumulate anthocyanins late in
seed development, a feature that is not charaderistic of wild-type seeds. Prematurely
germinated seeds give rise to viable green seedlings that appear normal except that trichomes

13
are found on the adaxial surfaces of its cotyledons. Trichomes normally form only on leaves,
stems and sepals, but not on cotyledons. Hence, tecf and fus3 cotyledons are considered to be
partially transformed into leaves, which gave two of the mutants their name (fee, 'leafy
cotyledons*, Meinke, 1992). The tec2 mutant also exhibits leafy cotyledons and accumulation of
anthocyanin, but differs from tecf and fus3 in that seeds are tolerant to desiccation and non-
viviparous, and seedlings often appear distorted in shape (elongated hypocotyl, curled
cotyledons) (Meinke et al., 1994). Unlike the abi3 mutant, germination of tecf and fus3 mutant
seeds is inhibited by ABA, indicating that they retain normal sensitivity to ABA (Keith et al., 1994;
Meinke et al., 1994). Proof for normal ABA synthesis in these mutants is still lacking. However,
since ABA-deficient mutants have thus far not been shown to exhibit leafy cotyledons or
accumulation of anthocyanins, it is unlikely that a possible lack of ABA would be the sole cause
of the mutant phenotype. Nevertheless, the role of ABA in these mutants remains to be
examined.
To investigate the interaction between ab/3 and leafy cotyledon mutants, double mutants
were constructed. abi3/led and ab¡3/fus3 double mutant seeds were highly viviparous,
insensitive to ABA, exhibited leafy cotyledons and accumulated large amounts of anthocyanins
(Meinke et al., 1994, BSumlein et al., 1994). The additive effect of ab/3 and Iec1/fus3 in the
double mutants suggests that ab/3 and tecf or fus3, respectively, are altered in distinct pathways.
Consequently, suppression of precocious germination requires at least ABA, developmental
factors controlling ABA-sensitivity and the leafy cotyledon-factors whose interactions with ABA,
however, remain to be analyzed.
Analysis of Gene Expression
Late stages of seed formation are correlated with the expression of characteristic genes
which were analyzed first and very extensively, in cotton embryos (Galau et al., 1986, 1987;
Hughes and Galau, 1991). Based on changes in the levels of specific sets of cotton mRNAs,
late seed development has been categorized into several stages (Galau et al., 1991). The

14
earlier 'maturation stage' comprises the longest time Interval (19 days) and is characterized by
high abundance of storage protein-mRNAs. This phase coincides with the presence of high
concentrations of ABA. It is apparently terminated by abscission of the vascular connections
between embryo and mother plant and is followed by the 'postabsdssion stage' (5 days) during
which maturation stage-specific mRNAs decline rapidly and a new set of mRNAs accumulates.
To these postabscission stage-specific mRNAs belong the LEA'S (this term was introduced by
Galau et al., 1986) and the RAB's (responsive to ASA. a term used by other authors for similar
proteins as LEA'S). Subsequently, seed formation is terminated by rapid water loss and
termination of transcription.
A similar temporal pattern of mRNA accumulation was reported for Arabidopsis (Farcy et
al.. 1994) and maize (Paiva and Kriz, 1994; Williams and Tsang, 1994). Nevertheless, for
species other than cotton, late stages of seed formation are usually refened to as the maturation
phase* which is not subdivided into two stages. It should be mentioned that in maize, the seed
maturation phase also correlates with the accumulation of anthocyanin pigments in embryo and
aleurone tissues.
In the following, progress in our understanding of the regulation of maturation-specific
genes (genes encoding storage proteins, LEA'S and RAB's, proteins of the anthocyanin pathway)
will be reviewed, placing emphasis on the roles of ABA and VP1 as regulators of seed
maturation in monocot seeds.
Storage proteins
In many species, immature embryos cultured in ABA exhibited precocious and enhanced
accumulation of storage proteins and their corresponding mRNAs as compared to those cultured
on ABA-free medium or left to mature on the mother plant (Quatrano, 1987). These results
indicate that ABA upregulates expression of storage protein genes. However, high levels of ABA
are not required for expression of the major storage protein genes, as shown for maize and
Arabidopsis. The aba mutant of Arabidopsis was found to accumulate normal levels of identified
ABA-upregulated storage proteins (2S, 12S) (Koomeef et al., 1989) and their corresponding

15
mRIMAs (Parcy et al., 1894). Similarly, maize ABA-defldent mutants accumulated mRNAs
corresponding to the ABA-regulated 7S globulins in only slightly reduced amounts (Krtz et al.,
1990; Paiva and Kriz, 1994). Hence. ABA does not normally appear to be a limiting factor in
expression of these storage protein genes. In contrast, developing embryos of the ABA-
insensitive mutants vp1 of maize and ab/3 of Arabidopsis exhibited very low or undetectable
expression of 7S globulins or 2S and 12S storage protein genes, respectively, indicating that
VP1/ABI3 are required for their expression (Kriz et al., 1990; Paiva and Krtz, 1994; Nambara et
al., 1992). Furthermore, exogenous ABA did not induce expression of storage protein genes in
cultured immature vp1 or ab¡3 mutant seeds, while it did so in cultured immature wild-type seeds
(Paiva and Kriz, 1994; Finkelstein and Somerville, 1990). Thus, VP1/ABI3 appear to be
essential for ABA action. Since VP1/ABI3 are expressed at normal levels in ABA-defident
mutants (McCarty et al., 1991; Paiva and Kriz, 1994; Parcy et al., 1994), It may be that normal
accumulation of storage proteins in these mutants is mediated by VP1/ABI3 either without a
need for ABA or requiring residual amounts of ABA present In mutant seeds.
LEAs/RABs
An extensive survey in Arabidopsis examining the accumulation kinetics of 18 marker
mRNAs expressed at high levels during mid to late seed development suggested that
LEAs/RABs-encoding genes fall into distinct classes with different requirements for ABA and
ABI3 to induce expression. For most markers, transcript levels did not solely correlate with the
amounts of endogenous ABA or ABI3 present in the seed, thus implicating a role of other factors
in controlling temporal patterns of expression (Parcy et al., 1994).
However, abundance of several tested mRNAs was highly reduced in seeds of the aba
mutant as well as the abi3 mutant (Parcy et al., 1994). Similarly in maize, expression of a well-
studied LEA gene (Em) originally isolated from wheat was undetectable in developing seeds of
mutants deficient for ABA or functional VP1 (McCarty et al., 1991). Hence, ABA as well as
ABI3/VP1 appear to be required for expression of certain LEAs, which Is consistent with a
possible role of ABI3/VP1 in ABA perception or signal transduction in the seed.

16
To investigate further the interaction of VP1 and ABA In controlling expression of the Em
gene in the maize seed, a transient gene expression system was used that is based on
electroporation of maize protoplasts isolated from an immature embryo-derived suspension
(Vasil et al., 1989). A plasmid containing the promoter (0.6 Kb) of the Em gene fused to the
coding sequence of the bacterial p-glucuronidase (GUS) gene (Em-GUS; Marcotte et al., 1988)
was used as a reporter construct. The VP1 cDNA was over-expressed from the constitutive
CaMV 35S promoter enhanced by insertion of the first intron of the maize Sh1 gene (Vasil et al.,
1989) into the 5' untranslated leader of the VP1 cDNA (35S-Sh-VP1). In these experiments,
electroporation of protoplasts with a mixture of Em-GUS and 35S-Sh-VP1 resulted in 100-300-
fold higher GUS activity as compared to the very low Em-GUS activity detected in the absence
of co-electroporated 35S-Sh-VP1 (McCarty et al., 1991). Similar activation was obtained when
protoplasts electroporated with Em-GUS were cultured in ABA (McCarty et al., 1991), which is
consistent with the reported ABA-regulation of Em-GUS in rice protoplasts (Marcotte et al., 1988;
1989). Over-expression of VP1 in maize protoplasts interacted synergistically with ABA,
resulting in 2,500-fold induction of Em-GUS (McCarty et al., 1991). The synergistic effect of VP1
and ABA underlines the importance of both, VP1 and ABA, in high-level expression of Em.
However, the substantial activation of Em-GUS obtained by either over-expressing VP1 or
culture in ABA might imply that VP1 and ABA can partially activate Em-GUS independently.
This would be in contrast to the absence of detectable Em transcript in vp1 or vp5 mutant
embryos which contain normal levels of ABA or Vp1 transcript, respectively (Neill et al., 1987;
McCarty et al., 1991). However, action of endogenous ABA and VP1 that may be present in the
wild-type protoplasts cannot be ruled out. To test this, Em-GUS was introduced into vp1 and vp5
mutant seed tissue (aleurone) via particle bombardment. In these experiments, ABA did not
activate Em-GUS in vp1A/p5 double mutant tissue while it did so in VP1/vp5 tissue or when co
bombarded with recombinant VP1 (S. Cocciolone and D. R. McCarty, unpublished results), thus
confirming that functional VP1 is required for ABA action. In vp5 mutant tissue, over-expression
of VP1 slightly activated Em-GUS, though 10-fold lower than in the presence of exogenous ABA.

17
This apparent independent activity of VP1 may be caused by low levels of maternal ABA present
in vp5 mutant seeds. Alternatively, the abnormally high levels of recombinant VP1 In expressing
cells may allow some ABA-independent activation of the Em promoter normally not found In
vivo.
The VP1 protein was subjected to functional analysis by testing deletion-derivatives for
their ability to irans-activate Em-GUS. Sequence analysis and domain swapping experiments
between VP1 and GAL4 had suggested that the N-terminus of VP1 contains an acidic
transcriptional activation domain (McCarty et al., 1991). Indeed, deletion of this acidic domain
abolished VP1s ability to activate Em-GUS. Replacing it with the acidic activation sequence of
the herpes simplex virus transcription factor VP16 partially restored transcriptional activation of
Em-GUS (McCarty et al., 1991). These results strongly indicate that VP1 functions as a
transcriptional activator in inducing Em-GUS. Analysis of internal deletion-constructs of VP1
identified two highly basic domains that are important for activation of Em-GUS (L. Rosenkrans,
V. Vasil, I.K. Vasil and D.R. McCarty, unpublished results).
Deletion analysis of the Em promoter indicated that two G-box-related sequences
(Em1a: ACACGTGG: Em1b: ACACGTGC) which are conserved in many promoters responsive
to ABA, light or anaerobiosis are involved in VP1- and ABA-mediated activation of Em (Marcotte
et al., 1989; Guiltinan et al., 1990; V. Vasil et al., unpublished results). The finding that ABA-
induction of Em does not require protein synthesis (Williamson and Quatrano, 1987) suggests
that VP1 and ABA frans-activate Em directly through the G-box elements rather than through
activation of intermediate regulatory genes further upstream in the ABA signal transduction
pathway. Thus far, no DNA-binding activity of VP1 to these putative target sequences has been
detected (T. Hattori, B. Li and D.R. McCarty, unpublished results). Hence, VP1 might activate
Em via protein-protein interactions with G-box-binding protein(s). A bZIP-type protein binding
specifically to the Em1a motif has been cloned (Guiltinan et al., 1990) and may thus be a
candidate.

18
The molecular mechanism underlying the synergistic interaction of VP1 and ABA is thus
far unclear. The possibility that ABA is required for high stability of the VP1 transcript or protein
which then in turn activates Em is highly unlikely because in the ABA-defident vp5 mutant no
reduction in VP1 transcript and protein levels was observed (McCarty et al., 1991; C. Carson and
D.R. McCarty, unpublished results). Possibly, ABA post-translationally modifies the VP1 protein
and thus enhances its trans-activation function. Alternatively, VP1 and an ABA-dependent factor
might be part of a complex that forms on the Em promoter and induces transcription.
Genes encoding proteins of the anthocvanin pathway
As identified by mutants, at least eight genes are essential for accumulation of
anthocyanin pigments in embryo and aleurone cells of the maturing maize seed (Fig. 1). Five of
these genes (A1, A2, C2, Bz1, Bz2) encode enzymes of the anthocyanin biosynthesis pathway
(Dooner et al., 1991). Expression of these structural genes requires the coordinate action of two
regulatory proteins, C1 and a member of the R/B gene family (Coe et al., 1988). Both proteins
exhibit features of transcription factors. The C1 protein contains a functional acidic
transcriptional activation domain (Goff et al., 1991) and a region of sequence homology to the
DNA-binding domain of animal myb proto-oncogene products (Paz et al., 1990). Proteins
encoded by the R/B gene family display high homology with the helix-loop-helix motif of myc
proto-oncogene products (Ludwig et al., 1989). At least for the promoter of the Bz1 structural
gene, it has been shown that sequences homologous to the consensus binding sites of animal
MYB and MYC proteins are essential for C1- and R-mediated activation of Bz1 expression (Roth
et al., 1991), suggesting a direct interaction of C1 and R with these sequences in the Bz1
promoter.
A third regulatory factor required for pigmentation of tissues in the developing maize
seed is VP1. The lack of anthocyanin pigments in vp1 mutant seed is associated with the
absence of C1 transcript (McCarty et al., 1989a). Over-expression of a 35S-C1 construct in vp1

19
I
ABA VP1
Maturation Genes
Anthocyanin Genes
Fig. 1. Role of VP1 in activation of seed maturation-related pathways and anthocyanin
biosynthesis.

20
mutant aleurone cells by particle bombardment complemented the failure to accumulate
anthocyanins in a cell autonomous fashion (Hattorl et al., 1992). Thus, lack of C1 appears to be
responsible for the block in anthocyanln synthesis in vp1 mutant seed. A direct role of VP1 in
activating C1 expression was concluded from the demonstration that over-expression of VP1 in
maize protoplasts activated transcription of a C1 promoter-GUS fusion gene. Most importantly,
VP1 function in activating C1 was dependent on its transcriptional activation domain (Hattori et
al., 1992). Hence, VP1 and C1 are part of a regulatory hierarchy controlling activation of
anthocyanin structural genes.
Several lines of evidence indicate that VP1's function in activating C1 is distinct from its
function in activating Em. First, though induction of both genes is dependent on the acidic
activation domain of VP1 and thus appears to involve transcriptional activation, other domains of
VP1 involved in function differed depending on the target promoter. While sequences in the
middle of the VP1 protein were required for trans-activation of Em-GUS in maize protoplasts,
VP1 -activation of C1-GUS was dependent on the C-terminal end of VP1 (L. Rosenkrans, V.
Vasil, I.K. Vasil and D.R. McCarty, unpublished results). This is consistent with the non-
viviparous/unpigmented phenotype of the vp1-McWhirter mutant which produces a truncated
VP1 protein lacking ca. 150 bp from the C-terminus (McCarty et al., 1989b).
Second, in agreement with the involvement of distinct domains of VP1, different cr's-
elements in the C1 and Em promoters appear to be the target of VP1 function. In contrast to
activation of Em which depended on two G-box sequences, activation of C1 did not require any
of the two G-box-like sequences present in the promoter but the 13 bp sequence -145
TCCATGCATGCAC -158 (Hattori et al., 1992). This sequence, designated as Sph-element, is
found in promoters of other seed-specific genes (Dickinson et al., 1988).
Finally, VP1 -mediated activation of Em and C1 differ in their interaction with ABA.
While there is a synergistic effect of VP1 and ABA in activating Em, the role of ABA in VP1-
mediated activation of C1 is less clear. Anthocyanins accumulate at normal levels in ABA-
deficient mutants of maize, suggesting that C1 expression is hormone-independent. On the

21
other hand, ABA firans-activated C1-GUS in maize protoplasts. Moreover, a c1 mutant (cf-p;
Chen and Coe, 1978) that fails to accumulate anthocyanin during seed development carries a 5
bp deletion in the promoter region of the gene that, when reconstructed by site-directed
mutagenesis of the C1 promoter and used in transient expression experiments, specifically
abolished ABA-responsiveness without severely affecting frans-activation by VP1 (Hattori et al.,
1992). Hence, the unpigmented phenotype of c1-p may be caused by a deficiency in ABA-
response. Pigmentation in ABA-deficient mutants might be possible if activation of C1 has an
ABA-requirement several orders of magnitude lower than activation of Em.
The Aleurone Germination Response in Cereal Seeds
During cereal seed development, the outermost endosperm cells differentiate into the
aleurone layer which at seed maturity consists of small, thick-walled cells with plasmodesmatal
connections. The aleurone cells are characteristically rich in protein and lipid bodies,
mitochondria and ER, but are devoid of starch grains. In response to GA released by the
germinating embryo, these highly specialized cells synthesize large amounts of hydrolytic
enzymes co-translationally on the rough ER and following proper folding in the lumen of the ER
- secrete these enzymes into the endosperm. This aleurone germination response to GA can be
inhibited by treatment with ABA.
The predominant hydrolytic enzyme synthesized is a-amylase, constituting ca. 15-20%
of total translatable mRNA and ca. 30% of total protein synthesis in germinating barley seeds
(Khursheed and Rogers, 1988). In barley and wheat, the major source of a-amylase is the
aleurone layer, whereas in maize, sorghum and rice, significant contributions are also made by
the embryo scutellum (Ranki and Sopanen, 1984; Dure, 1960a,b; Akazawa and Miyata, 1982).

22
Hormonal Regulation
Extensive studies in barley and several investigations with other cereal species have
shown that synthesis of cereal a-amylases is induced by GA and antagonistically inhibited by
ABA (for review see: Jacobsen and Chandler, 1987; Jones and Jacobsen, 1991; Fincher, 1989).
The first reports on GA-induced a-amylase activity in germinating barley seeds appeared in 1960
(Paleg, 1960a; b; Yomo, 1960). The action of ABA as an antagonist of GA was discovered in
1966 (Chrispeels and Varner, 1966).
Application of transcription and translation inhibitors indicated that the GA-induced
appearance of a-amylase activity was due to de novo synthesis of the enzyme (Varner and
Chandra, 1964; Filner and Varner, 1967). Ultimate proof for an effect of GA and ABA on a-
amylase synthesis came from the demonstration that GA treatment drastically increased the
amount of in vitro translatable a-amylase mRNA (Higgins et al., 1976), while simultaneous
application of ABA blocked this effect (Mozer, 1980). Run-on transcription experiments provided
evidence that GA and ABA regulate the transcription of a-amylase genes (Jacobsen and Beach,
1985; Zwar and Hooley, 1986). Eventually, the development of transient gene expression
technology has provided an additional tool to elucidate GA and ABA action. It was shown that
transient expression of a wheat a-amylase promoter-GUS reporter gene fusion construct in oat
aleurone protoplasts was regulated in the same manner as the endogenous genes (Huttley and
Baulcombe, 1989).
Relatively few studies have addressed hormonal regulation of maize a-amylase. Ingle
and Hageman (1965) reported a stimulating effect of GA on catabolism of carbohydrates in
excised endosperms. In a different study (Harvey and Oaks, 1974), exogenous GA applied to
excised endosperms further increased (3-fold) total amylase activity in germinating seeds of a
GA-deficient mutant (dS), but not in wild-type seeds. In contrast, culture of wild-type
endosperms in ABA strongly reduced amylase activity. Even though no molecular data are
available, these results indicate that maize a-amylase genes are probably regulated by GA and

23
ABA in a similar fashion as other cereal a-amylases. Ultimate proof for this, however, is still
lacking.
It has now been generally accepted that the embryo is the site of GA biosynthesis in the
germinating barley grain. When de-embryonated seeds are imbibed, no increase in GA levels in
the endosperm and very little subsequent production of a-amylase can be detected (Jacobsen
and Chandler, 1987). Further studies demonstrated that the scutellum, rather than the embryo
axis, is the source of GA (Radley, 1967; MacLeod and Palmer, 1967).
In maize, evidence for the importance of the germinating embryo as a source of GA has
been contradictory. In imbibed de-embryonated seeds, Dure (1960) found only ^-amylase
(which is stored in protein bodies in the dry seed and therefore is not de-novo synthesized during
germination), but no a-amylase activity, whereas whole kernels showed both activities. Two
other studies (Harvey and Oaks, 1974; Goldstein and Jennings, 1978), however, demonstrated
comparable total amylase activities in de-germed as well as whole seed endosperm. As only
part of the activity was due to release of fi-amylase from protein bodies, it was concluded that
mature seeds store considerable amounts of GAs in the endosperm. Therefore, the germinating
maize embryo does not appear to be an essential source of GA for a-amylase synthesis in the
aleurone.
The g-Amviase Genes
a-amylases of cereals can be biochemically separated into a number of isoforms that
differ in their isoelectric point (pi) but not considerably in their molecular weight. In barley, there
are two families of isozymes, the low-pi a-amylases with pis of ca. 4.4-5.2 and the high-pl a-
amylases with pis of ca. 5.7-6.2 (Jacobsen and Chandler, 1987). These two families differ in
many other biochemical characteristics while isozymes within those families are more alike.
Though some of the variants are post-translational modifications of the same gene product, a
genetic basis for most of the variation seen became evident when the gene(s) for the low and
high pi families were mapped to different chromosomes, chromosome 1 and 6, respectively

24
(Brown and Jacobsen, 1982; Muthukrishnan et al., 1984). Isolation and sequencing of a number
of cDNA clones verified that there are sequence differences between Isoforms. Base sequence
homology Is 90-95% within gene families and about 75% between gene families (Jacobsen and
Chandler, 1987). Southern blot analyses of two barley varieties revealed that there are at least
6-7 high-pl genes and at least 3 low-pl genes, indicating that a-amylases are encoded by two
multigene families (Khursheed and Rogers, 1988; Muthukrishnan et al.. 1984). While no detailed
mapping has been undertaken in barley, three rice a-amylase genes were found to be clustered
within 28 kb of genomic DNA. Molecular analysis suggested gene duplication as a cause (Sutliff
et al., 1991).
The two barley a-amylase gene families are regulated differently in the aleurone of
germinating seeds. mRNA levels of the high-pl isoforms increase very quickly upon imbibition,
reaching a maximum after two days and decreasing to low levels after four days. Synthesis of
low-pl isoform-mRNAs begins later, not before three days after imbibition, but then increases
rapidly so that low-pl isozymes become the dominant enzyme group after four days of imbibition
(Chandler and Jacobsen, 1991). The two isoforms are also differentially responsive to GA. In
some studies with isolated aleurone layers, mRNA as well as protein of the low-pl isoforms can
be detected before GA is added, while those of high pl-isoforms cannot (Chandler and Jacobsen,
1991; Jacobsen and Higgins, 1982; Rogers, 1985). Others, using isolated aleurone layers or
aleurone protoplasts, find no low-pl message in the absence of GA (Chandler and Jacobsen,
1991; Nolan and Ho, 1988). However, low-pl isoforms appear to be more sensitive to GA as
they respond to GA-concentrations as low as 10'9 M (Nolan and Ho, 1988). Thus, low-pl a-
amylase genes might be either more responsive to GA or leaky in expression.
In comparison to barley a-amylases, many fewer studies on maize a-amylases have
been reported. Partial purification of a-amylases from endosperm of germinating seeds has
revealed two major groups of isozymes, one with pis of 5.1-5.7, the other with pis of about 4.6
(Warner and Knutson, 1991). Other authors report the purification of a-amylase isozymes with a
variety of pis (Warner et al., 1991; MacGregor et al., 1988; Chao and Scandelios, 1971). No

25
genes coding for a-amyiases have been doned so far. Thus no information about gene
expression is available.
The Organization of g-Amviase Promoters
Gene expression is thought to be regulated by proteins ffrans-ading factors*) that bind
in a sequence-spedfic manner to short stretches of base pairs f cfc-acting elements") located in
the promoter region of the gene. With an interest to study regulation of a-amylase gene
expression, genomic dones were isolated. Sequence comparisons revealed little homology
between promoters of a-amylase genes belonging to different pi groups which may relate to their
differential expression in response to hormones described above. However, a few blocks of
sequence were found highly conserved among barley, wheat and rice a-amyiase promoters:
High-pl, barley cgccttttgagctcaccgtaccggccgataacaaactccggccgacatatccactg -H7
(Khursheed and Rogers, 1988)
Low-pl, barley gcaccttttctcgtaacagagtctggtatccatgca -98
(Whittier et at., 1987)
Low-pl, wheat gcaccttttttcgtaacagagtctggtatccatgca -95
(Huttley et al 1992)
To identify c/s-ading elements involved in hormone-regulated expression, fundional
analyses of a-amylase promoters have been performed. For this purpose, mutated promoter
sequences are fused to a reporter gene (e.g. GUS, Ludferase) and are assayed for function in a
transient gene expression system (eledroporation of aleurone protoplasts or partide
bombardment of intad aleurone layers).
Progressive 5' truncations of a-amylase promoters showed that 289 bp of a wheat low-pl
a-amylase promoter (Huttley and Baulcombe, 1989) and 174 bp of a barley high-pl a-amylase
promoter (Jacobsen and Close, 1991) were suffident to dired GA- and ABA-regulated

26
expression of a reporter gene, indicating that cto-acting elements are positioned in the proximal
region of the promoters. Indeed. Skriver et al. (1991) demonstrated that a chimeric construct
containing 69 bp (-189 to -120) of the barley high-pl promoter fused to the 35S TATA box could
impose increased transcription by GA and its suppression by ABA. Moreover, six tandemly
repeated copies of the sequence GGCCGATAACAAACTCCGGCC (21 bp) conferred proper GA-
and ABA-regillation. However, this result could not be confirmed when particle bombardment
was used as the method of transformation (J. Rogers, pers. communication; U. Hoecker,
unpublished results) suggesting that other ris-elements apart from TAACAAA contribute to GA-
regulated transcription. This was confirmed when clustered point mutations were introduced
covering the proximal region of the promoters. Mutations in the pyrimidine box (CCTTTT) or in
the TATCCAC/T box reduced GA-induced transcription to about 20 % of minimal level in the
barley high-pl and low-pi promoters (Gubler and Jacobsen, 1992; F. Gubier, pers.
communication; Lanahan et al., 1992). Thus the entirety of the three conserved elements
appears to be involved in mediating GA-response. Interestingly, in no case could ABA-
responsive elements be separated from GA-responsive elements, suggesting that GA and ABA
function through the same c/s-elements in the a-amylase promoters.
Rogers and co-workers (Lanahan et al., 1992) identified an additional element in the low-
pi promoter that is located between positions -152 and -134, just upstream of the pyrimidine box.
Mutations in this element reduced the level of expression by 96% while retaining significant but
low GA-responsiveness. This region of the promoter (termed 02S element) shows sequence
homology to two well-described motifs: the endosperm box, a conserved element present in
promoters of maize, barley and wheat endosperm protein genes, and the consensus sequence
for binding of the maize Opaque-2 protein which is a leucine zipper protein (bZIP) that is
necessary for transcription of the 22 kDa zein genes. Interestingly, it was found that substitution
of the GA-responsive TAACAGA sequence of the low-pi promoter with an ABA response
element (ABRE) from the rice Rab-16A gene converted the promoter from a GA-upregulated one
into one whose transcription was increased by ABA (Rogers and Rogers, 1992). Thus, the ABRE

27
regulated transcription In the context of the low-pl a-amylase promoter In a similar way as it does
in its native rice promoter. Importantly, its function in the amylase promoter was highly
dependent upon the presence of the 02S sequence. Thus, the 02S element appears to function
as a "coupling element" that is necessary for high-level, hormone-regulated transcription from
the low-pl a-amylase promoter (Rogers and Rogers, 1992). No 02S-like sequence Is evident in
high-pl a-amylase promoters. However, inserting the 02S element from a low-pl promoter into a
high-pl promoter at a position upstream of the pyrimidine box enhanced transcription ca. 5-fold,
suggesting that the 02S element function could interact properly with the high-pl promoter
fragment to give high-level transcription (Rogers et al., 1994).
To identify the trans-acting factors that bind to the crs-elements in the promoters and
thereby confer hormone-dependent expression of the a-amylase genes, DNA-protein interactions
have been characterized using band shift assays and DNase I footprinting analyses. In a wheat
low-pl promoter (a-Amy 2/54), the 02S box and the TAACAGA element, but not the TATCCAC
sequence, were found protected from DNase I digestion, confirming the binding of protein
factor(s) to at least two functionally important elements. However, these binding activities were
not dependent upon GA (Rushton et al., 1992). Evidence for the presence of a GA-dependent
factor on a barley low-pl promoter was provided by Sutliff et al. (1993). Results from band shifts
performed in this study demonstrated that a GA-inducible binding activity interacted with the
TAACAGA and TATCCAC elements in a sequence-specific manner. In a different report on a
rice a-amylase promoter (Amy3c), GA-dependent binding to a pyrimidine box-like sequence was
demonstrated (Goldman et al., 1994). However, since this promoter displays little sequence
homology to barley and wheat promoters, comparisons cannot be made. In summary, GA-
independent as well as GA-dependent factors appear to constitute the complex(es) on a-amylase
promoters that result in GA-regulated expression. Further characterization of the protein-DNA
and protein-protein interactions is needed.

26
GA and ABA Signal Transduction
It is generally accepted that molecules with hormonal function act as ligands that upon
binding to their specific receptors elicit a response which ultimately can result in altered gene
expression. In animals, a large number of hormone receptor-encoding genes has been cloned
and characterization of their gene products has demonstrated that receptors can be found
located intracellularly (e.g. steroid hormone receptors) or integrated into the cell membrane with
their ligand-binding domain facing the extracellular space. Identification of plant hormone
receptors has proven difficult. Though hormone-binding proteins have been identified, definite
proof for the function of these proteins is still lacking.
With respect to GA and ABA signal perception in the cereal aleurone, studies have so
far concentrated on the identification of the cellular site of the receptors. Results from at least
one study suggest that GA does not have to enter the cell to regulate gene expression. Hooley
and colleagues (Hooley et al 1991) demonstrated that GA immobilized to Sepharose beads was
capable of enhancing a-amylase transcription in oat aleurone protoplasts. Though these data
point to a perception of the GA signal on the external surface of the plasma membrane, the
existence of intra-cellular receptors cannot be ruled out. Clear evidence against intra-celluiar
receptors came from elegant experiments performed in the laboratory of R. L. Jones. A method
was developed to visualize a-amylase gene expression and a-amylase secretion from individual
protoplasts (Hillmer et al., 1992). This allowed to test whether or not hormones microinjected
into the cytosol of aleurone protoplasts were capable of eliciting a response (Gilroy and Jones,
1993). It was found that protoplasts injected with GA did not respond to the hormone. Similarly,
ABA microinjected into protoplasts was ineffective in antagonizing the stimulating effect of pre
applied external GA. The failure to respond to microinjected hormones was not due to disruption
of protoplast function by microinjection since protoplasts that had been subjected to this
procedure remained responsive to externally applied GA. These results indicate that the sites of
perception of GA and ABA are located on the external face of the plasma membrane in aleurone
cells.

29
There is little concept of how the perceived GA and ABA signals are transduced from the
receptors to the nucleus. The function of Ca** and calmodulin (CaM) as second messengers
that regulate protein kinase activity has been characterized in many animal and a few plant
systems (Roberts and Harmon, 1992; Neuhaus et al.. 1993). Also in barley aleurone protoplasts,
treatment with GA was found to increase cytoplasmic Ca~ and CaM concentrations (Gilroy and
Jones, 1992, 1993). ABA reversed the effect of GA on (Ca~], (Gilroy and Jones, 1992). Even
though the increase in [Ca~], and [CaM] preceded the GA-induced increase in a-amylase
activity by 2-4 h (Gilroy and Jones, 1992), direct evidence for an involvement of Ca++ and CaM
in the regulation of a-amylase transcription is still lacking. Ca++ and CaM have been found to
regulate the activity of a slow vacuolar ion channel located in the tonoplast of storage protein
vacuoles in barley aleurone cells (Bethke and Jones, 1994). Moreover, CaM was shown to
stimulate Ca++ uptake into the ER where the Ca~ containing a-amylase enzyme is synthesized
(Bush et al.. 1993). Thus, hormone-regulated changes in Ca** and CaM concentrations may be
regulating processes such as a-amylase formation and secretion rather than having a direct
effect on the transcription of a-amylase genes.
Analysis of mutants provides a valuable tool to study the genetics underlying the
regulation of hormone action. Many mutants have been isolated that display altered responses
to GA, suggesting that these mutants are affected in a component of GA signal transduction.
Since GA promotes stem and leaf elongation, these mutants have been identified by their altered
plant height. They fall into two classes: 1) those that show a reduced sensitivity to GA (*GA-
insensitive mutants') and are therefore of dwarf stature, and 2) those that show an enhanced
sensitivity to GA ("constitutive response mutants') and are therefore excessively tall. In
response mutants, the concentrations of biologically active GAs do not in accord with the
phenotype. Generally, GA-insensitive mutants accumulate higher concentrations of active GAs
as compared to wild-type, while tissues of constitutive response mutants contain reduced GA
concentrations (Stoddart, 1984; Fujioka et al, 1988; Croker et al., 1990). These observations

30
have been attributed to feedback regulations on GA metabolism in response to altered GA-
sensitivity.
To the class of GA-insensitive mutants belong the Rht (reduced height) mutants of
wheat. A total of 10 Rht loci have been identified, showing varying degrees of dominance. Of
these, the mutation Rht3 exerts the strongest dwarfing effect. The GA-insensitive phenotype of
Rht3 is also expressed in the aleurone: germinating seeds had 75% reduced levels of amylase
activity as compared to tall (rht) varieties and showed no or little increase in amylase activity
after GA treatment (Gale and Marshall, 1975; Fick and Qualset, 1975). The degree of GA-
insensitivity of the aleurone was found to increase with the dosage of Rht3 alleles (Gale and
Marshall, 1975). A similar GA-insensitive mutant has been described in rice (Mitsunaga et al.,
1994). The finding that the failure to respond to GA is expressed in plant and seed tissues
indicates that these tissues share at least in part a common signal transduction pathway.
Two dominant GA-insensitive dwarf mutants have been identified in maize (D8, D9).
Besides being of reduced stature, these mutants mimic additional characteristics of GA-deficient
mutants, such as reduced apical dominance and formation of anthers on the ear (Coe and
Neuffer, 1977). D8 and D9 are located on 1L and 5S, respectively (Coe and Neuffer, 1977).
Since the region of 5S contains duplicate loci with 1L it is likely that D8 and D9 are duplicate loci
encoding gene products with identical or similar function. X-ray-induced chromosome breakage
was used to create clonal sectors of wild-type cells within D8 mutant tissue. Results from these
experiments indicated that 08-mediated effects can be expressed cell autonomously at least in
some tissues (Harberd and Freeling, 1989) which is consistent with the hypothesis that the wild-
type gene product is part of a GA signal transduction pathway. The gain-of-function nature of
the mutation in association with a GA-insensitive phenotype allows to speculate about the
function of the wild-type gene product. Possibly, cf8(+) (and d9(+)) encodes a negative regulator
of GA response that normally is inactivated by exposure of the cell to GA. in this scenario, the
mutant D9 protein would be constitutively active in this repressor activity, i.e. even in the
absence of GA (Harberd and Freeling, 1989).

31
A constitutive response mutant has been identified in barley (Foster, 1977). This
recessive mutation, termed slender" (sin), causes a plant to appear as if It had been treated with
high doses of GA (Lanahan and Ho, 1988; Croker et al., 1990). Also, a-amylase genes were
highly expressed in sin mutant half grains in the absence of applied GA (Chandler, 1988). Thus,
the absence of functional SLN protein causes constitutive expression of GA-responses and
thereby uncouples transcription of GA-regulated genes from a need for GA. This phenotype
suggests that Sin encodes a negative regulator of GA-response. Importantly, a-amylase
production in sin mutant aleurones was susceptible to inhibition by ABA, indicating that the sin
mutant retains normal sensitivity to ABA (Chandler, 1988; Lanahan and Ho, 1988). Hence, ABA
most probably functions at a step downstream of GA in the signal transduction pathway leading
to regulation of a-amylase transcription. A function of ABA fully independent of GA cannot be
ruled out but is unlikely because GA and ABA appear to act through the same response
elements in a-amylase promoters. The findings also indicate that GA and ABA do not act at the
same site in the signal transduction pathway, i.e. they do not for example antagonistically
phosphorylate/ de-phosphorylate an intermediate.
The Developmental Switch from Seed Maturation to Seed Germination
Desiccation is the normal terminal event in seed development, leading to a state of
metabolic quiescence. In many species (e.g. maize, bean), hydration of the mature, dry seed is
sufficient to initiate germination. Thus, in these seeds (termed quiescent or non-dormant seeds),
the transition from seed maturation to germination is associated with the reversal of the
desiccated state. Seeds of other species (e.g. cereals, Arabidopsis) develop dormancy during
late stages of seed development. In these species, freshly harvested mature seeds do not
germinate following imbibition but require a treatment such as light, low temperature or after
ripening (dry storage) to overcome the state of dormancy and allow induction of germination.

32
Hence, dormant seeds execute the developmental switch to germination during the imposed
dormancy-breaking treatment.
As described earlier, the transition from seed development to germination is associated
with major changes in gene expression. It is generally accepted that in quiescent seeds,
maturation-related genes cease expression once the water content falls below a level permitting
transcriptional activity, and following imbibition, expression of a new set of genes is initiated
which is specific to the germinating seed. What gene expression programs are executed in
imbibed, dormant seeds is less clear. However, there is evidence from studies in wheat pointing
to a maintenance of maturation-specific gene expression during the state of dormancy (Ried and
Walker-Simmons, 1990,1993; Morris et al., 1991).
Most importantly, maturation and germination programs appear to be regulated
coordinately in the developing seed. Not only is precocious germination of the immature embryo
suppressed, but similarly, the premature induction of germination-related genes appears to be
inhibited during this developmental state. Developing seeds of wheat and barley contain
biologically active GAs in concentrations adequate to induce a-amylase production in the
aleurone layer (Wheeler, 1972; Radley, 1976). Nevertheless, only very low levels of a-amylase
enzyme activity or mRNAs were detected in immature seeds (Comford et al., 1986; Garcia-Maya
et al., 1990). More compelling, neither a-amylase activity nor a-amylase gene expression was
induced when immature seeds were exposed to exogenous GA (Nicholls, 1979; Comford et al.,
1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Similar results were obtained when
treating dormant seeds with GA (Schuurink et al., 1992a). Given that immature or dormant
embryos excised from the seed and placed in culture are capable of responding to GA, these
data are strong evidence for active repression of the GA-response in developing or dormant
seeds. Results consistent with this idea were also reported for dicot seeds (soybean, castor
bean). In these species, immature seeds contained enzymes involved in the degradation of fatty
acids and proteins (malate synthetase, LeuNase, isocitrate lyase) at a much lower level than
germinating seeds or isolated embryos in culture (Kermode, 1990).

33
The mechanism underlying the developmental switch from seed maturation to seed
germination, precisely the turning off* of maturation-related genes and the de-repression/
induction of germination-specific genes, is only poorly understood. Evidence on the molecular
nature of this switch is reviewed in the following for quiescent and dormant seeds.
Quiescent Seeds
Desiccation and subsequent rehydration of the seed appears to be the normal trigger to
switch the developmental program from maturation to germination (Comai and Harada, 1990).
Even when applied prematurely, drying resulted in the termination of maturation-related gene
expression (Oliver et al., 1993) and, upon imbibition of the dry seed, the induction of genes
specifically associated with germination (Kermode, 1990). Drying altered the developmental
potential of seeds such that a-amylase production became sensitive to GA (Evans et al., 1975;
Nicholls, 1979; Armstrong et al., 1982; Comford et al., 1986; Oishi and Bewley, 1990).
The nature of this switch in GA-sensitivity remains elusive. King (1976) has postulated
that accumulation of ABA in the developing seed prevents precocious induction of hydrolase
gene expression in the developing aleurone. Indeed, drying has been shown to cause a
concomitant decline in grain ABA content (McWha, 1975; King, 1976; Oishi and Bewtey, 1990).
Moreover, incubation of immature grains in buffer which caused a drop in endogenous ABA to
undetectable levels evoked GA-responsiveness of the aleurone (Napier et al., 1989). Hence,
depletion of endogenous ABA, by drying or washing, may be responsible for the induction of GA-
responsiveness. This is consistent with the finding that in maize mutants that are either deficient
for embryonic ABA (vp5) or insensitive to ABA (vpf), a-amylase activity was induced late in seed
development (Wilson et al., 1973).
However, results from Oishi and Bewley (1990) indicate that induction of a-amylase
synthesis as a result of drying is not solely due to a reduction in ABA content. The authors
compared the responses of maize kernels to premature drying and treatment with an ABA
biosynthesis inhibitor (flouridone) which reduces ABA contents in the seed to a similar extent as

34
drying of immature seeds and elicits precocious germination of immature maize kernels. If
drying merely served to deplete endogenous ABA in developing seeds, then flouridone-treated
kernels and dried seeds should behave similarly with respect to GA-response. However, while
drying resulted in synthesis of high levels of a-amylase following imbibition, flouridone-treated
seeds produced only very low amounts of a-amylase in response to GA. Hence, drying may
achieve two effects: 1) it frees seed tissues of the inhibitory effect of ABA, and 2) it renders the
aleurone competent of responding to GA. The cause of the ABA-independent GA-insensitivity in
immature seeds is thus far unknown.
Dormant Seeds
Cereals
Imposed dormancy in cereal species is normally released by prolonged storage of dry
seeds (afterripening). The duration of seed dormancy following seed maturity depends on a
variety of factors such as the genetic constitution (cultivar), the environmental conditions during
grain maturation (low temperatures and short day length increase dormancy; Schuurink et al.,
1992b) and the rehydration temperature (high temperatures enhance dormancy; George, 1967).
Such differences in the depth of seed dormancy have been utilized to investigate the roles of
ABA concentration and ABA-sensitivity in preventing embryo germination. No clear correlation
between ABA content in the mature embryo and the degree of dormancy was found (Walker-
Simmons, 1987,1988; Morris et al., 1989; Skadsen, 1993). However, there are large differences
between dormant and non-dormant embryos with respect to sensitivity to ABA, as measured by
the capacity of ABA to inhibit germination. Isolated embryos of a non-dormant wheat cultivar
lost their sensitivity to ABA in culture as the grain entered maturation stage, whereas those of a
dormant cultivar retained sensitivity to ABA beyond grain maturity (Walker-Simmons, 1987).
Similarly, elevating the incubation temperature from 15C to 30C, thus inducing high-
temperature dormancy, significantly enhanced the ability of ABA to block germination of isolated

35
wheat embryos (Walker-Simmons, 1988). This differential inhibitory effect of ABA depending on
the degree of dormancy was also observed using intact, mature seeds (Morris et al., 1989).
In conclusion, depth of dormancy appears to be positively correlated with ABA-sensitivity
with respect to inhibition of germination. Why, and if, enhanced ABA-sensitivity is the immediate
cause for inhibition of germination is unclear. Because ABA has been shown to inhibit water
uptake by the embryo (Schopfer and Plachy, 1984), it has been suggested that high sensitivity to
ABA in dormant seeds may result in reduced water uptake in the embryo and thereby prevent
radicle emergence (Walker-Simmons, 1987). Additionally, ABA may have a differential effect on
gene expression in dormant and non-dormant seeds. Indeed, transcript levels of a variety of
ABA-responsive genes remained high in imbibed dormant wheat seeds, whereas they declined
rapidly in non-dormant seeds following imbibition (Morris et al., 1991). Similarly, maturation-
related LEA proteins were abundant in rehydrated dormant seeds but not in non-dormant seeds
(Ried and Walker-Simmons, 1990, 1993). However, most of the identified ABA-responsive
proteins that accumulate specifically in dormant seeds are predicted to be dehydrins and may
therefore function primarily in maintaining the embryo in a desiccation-tolerant state rather than
in directly inhibiting germination.
Exogenous application of ABA is known to inhibit GA-mediated activation of hydrolase
genes in the aleurone of germinating cereal seeds (Jacobsen and Chandler, 1987). In this
context, the following observation may be important. Imbibed dormant barley seeds showed
very low expression of a-amylase genes as compared to non-dormant seeds (Morris et al., 1991;
Schuurink et al., 1992a; Skadsen, 1993). Moreover, dormant grains produced less a-amylase in
response to GA than non-dormant grains (Schuurink et al., 1992a; Skadsen, 1993). Hence, seed
dormancy appears to be correlated with a reduced responsiveness of the aleurone to GA.
Experiments with isolated aleurone layers indicated that the reduced GA-sensitivity of aleurone
cells of dormant barley seeds is dependent on the presence of the starchy endosperm
(Schuurink, 1992a; Skadsen, 1993), implying that the starchy endosperm may liberate an
inhibitory factor that diffuses to the aleurone cells. It is suggestive that this putative diffusible

36
factor may be ABA stored in the dry seed. However, since the dormant and non-dormant seeds
used in one experiment contained similar concentrations of ABA (Skadsen, 1993), it may be that
the higher ABA-sensitivity in dormant seeds relative to non-dormant seeds plays a role in
inhibiting GA-response in the aleurone. However, it cannot be excluded that a factors) other
than ABA inhibits the GA-response in the aleurone of dormant seeds, a-amylase genes are
known to be sensitive to repression by soluble carbohydrates (Yu et al., 1991; Karrer and
Rodriguez, 1992). Possibly, aleurone cells of dormant seeds display a higher sensitivity to the
inhibitory effect of soluble sugars supplied by the starchy endosperm. Alternatively, the
inhibitory factor may be synthesized in the aleurone cells themselves and the presence of the
starchy endosperm is only required to provide an environment of high osmolality which may be
essential to maintain production of the putative inhibitor of GA-response in the aleurone cells.
Arabidoosis
Dormant seeds of Arabidopsis require either several months of dry storage or
rehydration followed by exposure to low temperatures and light in order to break dormancy and
induce germination. Analyses of mutants has demonstrated that initiation of dormancy during
late seed development involves the action of ABA. Even in light, wild-type seeds are normally
incapable of germinating during the seed maturation phase and, for a period of time, after
reaching seed maturity. In contrast, seeds of the ABA-deficient mutant aba gradually acquire
germination capacity during seed development and at maturity germinate at a frequency of
100% in light and 30% in darkness (Karssen et al., 1983; Karssen and Lacka, 1985). Hence aba
mutant seeds display highly reduced dormancy. A germination behavior similar to aba mutant
seeds was observed for the ABA-insensitive mutants abi1, ab¡2 and ab¡3 (Koomeef, 1984),
suggesting that these mutants are non-dormant due to a failure to respond to ABA. However, as
mentioned earlier, strong alleles of ab¡3 do not only cause lack of seed dormancy but also
vivipary, whereas seeds carrying strong alleles of aba have thus far not been shown to be
viviparous. Thus, an ABA-independent function of ABI3 cannot be ruled out.

37
While ABA is clearly involved in the induction of seed dormancy, its role in the
maintenance of a dormant state beyond seed maturity is less dear. Late In seed development,
ABA concentrations dedine rapidly to a very low amount present in the dry seed. This amount
has been considered insuffident to inhibit germination (Karssen et al., 1983). Also, aba mutant
and wild-type seeds were equally sensitive to applied ABA (Koomeef et al., 1982, 1984),
indicating that the difference in germination capacity between aba mutant and wild-type seeds is
not due to a difference in ABA-sensitivity. It was therefore thought unlikely that ABA or ABA-
sensitivity are involved in maintaining the state of dormancy (Karssen and Lacka, 1985).
Instead, Karssen and Lacka (1985) proposed that the maintenance of dormancy is, at least in
part, mediated by an insensitivity of the seed to GA. This was conduded from evidence showing
that a gradual relief of dormancy by aftenripening, cold or light treatments was correlated with an
increased sensitivity of the seed to the germination-promoting effed of applied GA (Karssen and
Lacka, 1985; Dericx and Karssen, 1993). GA is normally absolutely required for induction of
seed germination, as evident from the fad that seeds of the GA-defident mutant ga-1 do not
germinate under any condition, unless GA is supplied exogenously (Koomeef and van der Veen,
1980). Consequently, insensitivity to GA may present a strict measure to inhibit seed
germination. Consistent with this hypothesis, reduced germination frequendes were reported for
the partially GA-insensitive mutant Gai (Koomeef et al., 1985).
In condusion, the present evidence implies that GA and ABA do not normally interad
diredly at any stage of seed development. ABA in concert with high ABA-sensitivity appears to
be responsible for the indudion of dormancy during seed development, and GA in concert with
GA-sensitivity induced by dormancy-breaking treatments appears to stimulate germination.
Nevertheless, redudion in seed dormancy as a result of low concentrations of ABA (aba) or
insensitivity to ABA (abi1, ab¡2, ab¡3) partially relieved the mature seed form a need for GA to
induce germination. Seeds of double mutants between ga-1 and aba, ab¡1, abi2 or ab/3,
respedively, were capable of germinating (Karssen et al., 1983; Koomeef et al., 1984, Nambara
et al., 1992), whereas ga-1 single mutants have, thus far, never been shown to germinate without

38
application of GA (Hilhorst and Karssen, 1992). Since mature seeds contain very low
concentrations of ABA, it was not considered likely that the ga-1 single mutant required GA to
directly oppose the action of endogenous ABA present in the seed. In contrast, a remote
control* model was suggested in which the GA requirement for germination depends on the
depth of dormancy induced during seed development. Deeply dormant seeds, as wild-type
seeds, have a high GA-requirement to promote germination, while seeds with little dormancy
(ABA mutants) have a low GA-requirement which may be satisfied by low concentrations of GA
present in the possibly leaky ga-1 mutant.

MATERIALS AND METHODS
Plant Material
Except for immature d1 mutant kernels, which were obtained from greenhouse-grown
plants, all maize developing ears were harvested from field-grown plants. Under the local
environmental conditions, kernels typically begin accumulation of anthocyanins at day 17
postpollination and reach seed maturity after ca. 30-33 DAP. The wild-type maize stock used in
this study was a color-converted W22 inbred line carrying all factors required for anthocyanin
pigmentation of the aleurone. The vp1-R allele (Robertson 1955) segregated in a color-
converted W22 inbred line carrying all other factors required for anthocyanin pigmentation of the
aleurone. This line is routinely maintained by selfing. The vp1-m2 allele (originally named vp1-
mum2, McCarty et al., 1989b) arose in Robertsons Mutator transposable element stocks
(Robertson 1978), but was confirmed to carry an Mpi transposable element insertion (D.R.
McCarty, unpublished results). Seed segregating for the vp5 mutation was obtained from the
Maize Genetics Corporation Stock Center (University of Illinois, Urbana-Champaign). To
produce vp1,vp5 double mutant seeds, heterozygous vp5 mutant plants were crossed with
heterozygous vp1-R mutant plants. A mutation conferring embryo abortion at early globular
stage (germless*) arose in a Robertsons Mutator-induced mutant screen (D.R. McCarty et al.,
unpublished results). Germless mutant seed were backcrossed into W22 background for at least
two generations.
vpf-non-concordant seed was generated using a TB translocation stock (Fig. 2). TB3La
seed carrying a BA-translocation on the long arm of chromosome 3, the location of the Vp1
gene, was obtained from the stock center. This seed contains an extra, normally
heterochromatic chromosome, called B-chromosome, in addition to the normal set of A
39

40
Genotype of Macrospore (Female Parent): Tvp
Fig. 2. Generation of vpf-non-concordant seed using a TB3La translocation stock.

41
chromosomes. Due to a translocation event between A-chromosome 3 and the B-chromosome,
the 3La part of the A-chromosome 3 is carried by the B-chromosome, while heterochromatlc
DNA is found on the 3La part of the A-chromosome 3 (Fig. 2). The TB3La stock was crossed to
vp1-R at least once. Hence, resulting TB3La, AVP*-** B+ plants carry one 3A-chromosome
segregating for the vp1 mutation, the homologous 3A chromosome with heterochromatic DNA
(thus conferring a vp1 mutant phenotype) and a B-chromosome carrying the wild-type Vp1 gene .
To obtain vp1 non-concordant kernels, pollen from TB3La, A*?1'** B* plants is crossed onto
segregating vp1-R females. During the second pollen mitosis, the replicated B-chromosomes
undergo non-disjunction forming one sperm nucleus with two B-chromosomes and one sperm
nucleus without a B-chromosome (Fig. 2). Hence, following double fertilization, non-concordant
seeds are produced carrying a vp1 mutant embryo and a wild-type endosperm or vice versa (Fig.
2).
For experiments with germinating wild-type seeds of maize, seeds of the variety NK508
were used (kindly provided by Northrup-King).
Wild-type barley seeds c.v. Himalaya were obtained from Washington State University,
Pullman, WA (harvests 1988, 1991 and 1992). Seed segregating for the slender mutation
(Himalaya background) was kindly provided by P. Chandler. As with seed segregating for D8,
wild-type and slender mutant endosperms were genotyped by germination of the excised
embryo.
Plasmid Constructs
In all experiments, JR254 (Amy-GUS) or JR303 were used as reporter constructs (see
Fig. 3). Amy-GUS and JR303 were kindly provided by J. Rogers and T.H.D. Ho, respectively.
Amy-GUS contains ca. 1,800 bp of the 5 flanking sequence of a barley high pi a-amylase gene
(Amy6-4] Kursheed and Rogers, 1988), the first intron of Amy6-4, the GUS reporter gene and the
Amy6-4 3 terminator. JR303, containing a low-pl a-amylase promoter, was derived from

42
Reporter Plasmids:
High-pl Amy-GUS
(JR254)
Amy 6-4 Intron 1
GUS
Amy 6-4 3^
Amy 6-4 promoter
Low-pl Amy-GUS2
(JR303)
Intron
GUS
Amy 32b 3"
Amy 32b promoter
Effector Plasmids:
35S-Sh-VP1
35S-Sh-CAT
(No-Vp1 Control)
Sh1 Intron
Vp1 cDNA
CaMV 35S promoter
Sh1 Intron
NOS 3
CaMV 35S promoter
NOS 3*

Internal Standard:
Ubiquitin-Luciferase
(pAHC18)
Intron
Ubiquitin promoter
Fig. 3. Plasmid constructs.

43
Amy32b (Lanahan et al., 1992). For its structure, see Fig 3. As effector construct. 35S-Sh-VP1
was used (McCarty et al., 1991). For no-VP1 control treatments, 35S-Sh-CAT (Vasil et al., 1989)
was added instead of 35S-Sh-VP1 to maintain a constant amount of total DNA and 35S promoter
in the bombardment mixtures. To normalize for transformation efficiency, a Ubiquitin-Luciferase
construct (Ubi-LUC; Bruce et al., 1989) was included into each bombardment mixture. Hence,
expression data are presented as Amy-GUS / Ubi-LUC or JR303 / Ubi-LUC ratios. It was
confirmed that co-expression of 35S-Sh-VP1 has no effect on expression of Ubi-LUC.
Construction of plasmids carrying an activation domain-deletion derivative of VP1 or a
replacement with Herpes simplex virus transcription factor VP16 activation domain was
described in McCarty et al. (1991). Internal deletion constructs were made by introduction of two
A/col restriction sites and subsequent deletion of the insert and religation of the backbone (Fig.
4). Nco\ sites were introduced by site-directed mutagenesis using the Altered Sites in vitro
Mutagenesis System from Promega. Briefly, mutant oligonucleotides and an ampicillin repair
oligonucleotide which restores the function of a defective ampicillin resistance gene in the
phagemid provided were annealed to single-stranded DNA template. Following DNA synthesis
and ligation, the resulting double-stranded phagemid was transformed into a repair-deficient
strain of E.coli which is subsequently grown in Ampiclllin-containing liquid medium for selection.
From the obtained bacterial suspension, plasmid DNA was isolated and transformed into an E.
coli strain conventionally used for transformations (JM83). Colonies growing on Ampicillin were
tested for the presence of the desired mutation by restriction enzyme digestion.
The constructs 86/87, 86/85, 85/87, 87/88, 93/95, 103/104, 101/100 and VP1-McW were
made available by L. Rosenkrans and D.R. McCarty.
Particle Bombardment and Transient Expression
Tissue preparation
Maize developing ears or dry, mature seeds were surface sterilized in 70% ethanol for 1
min followed by 0.525% NaOCI for 10 min. Dry seeds were germinated in a solution containing

44
1 irriTLTlTU-TlO:! 11U1UliCOTWXTCTCtOC^WOCCTCCTOOBOC^COCCCaaCCTCCCAaWCeCO 120
r^^^/TrirryTr^^i^Tf^^rraT-rT-rTyTrLraiaiTrraj^xraAayTreaXlATOACTTCATMTaArrSAAflACAUil UXO-'llXCTCCOOGA^l 1^1 l^l 11^ 210
2*HGODOOAADrM*IOO*AADAIDtFSL2DfCl* S
TCOCCgTCCJ^CCA^CTCGTCCAACTCCTa2TCAAACTCCTCCAOCOCCTACACaU>CAgKXaaaAABI>OCaiGOaOCGOCCeTCOBASCCICCTTOOgCOOQaAAOOOm 1*0
SPSSSTFSSMSSSMSSSAYTMTAGRAQOEPSEPASAGEGF 96
94: CCATG0
?TTirir?rr?^T7Tl^Jl MPWDSBPFPGVSMMLENAHSAPPQ 136
CCTOTOGGTGAni?GrATGAf?TGAA^AGAAAGCnTTf;CCT?^A*?f?frArcAmaQQQG*flRGGRcccc7t3CATGGATGCGTCGGAGQQOGAGGAGCTGCCGCQGTTCTTCATOGAGTOGCTC 600
PVGDGMSKEKAVP1GTTGGKKACH0ASIGBILPRFFMKHL 176
YEFPAGGQDMAAGGGTSWMPHQQAFTPPAAYGGDAVYPSA 296
93: CGAXGG
194: CCATOfl
PQRMAGVEASATKEARKKRMARQRRLS CLQQQRSQQLS LG 16
QIQTSVHLQEPSPRSTHSGPVTPSAGGHGFWSPSSQQQVQ 56
NPLSKSNSSRAPPSSLEAAAAAPQTKPAPAGARQDDIHHR 96
KEAEVHLPELKTRDGISIPNEDIGTSRVHNMRYRFHPNNK 576
100: CCATGG
AGCAGAATGT AT CTGCTGGAAAACACAGGGGAAT T TGT T CGT T CCAACGAGCTT CAGGAGGGGGATTT CAT AGTGAT CT ACT CCGATGT CAAGT CGGGCAAAT AT CTGAT ACGGGGCG TG 1920
SRMYLLENTGEFVRSNELQEGDFIVIYSDVKSGKYLIRGV 616
GTCGTCGACGGGGTCAGCGGCGCCTGCAAGGGGAGGTCTCOGGAAGGCGTGCGGCGGGTTOGGCAGCAGGGAGCCGGOGCCATGAGCCAGATGGCGGTGAGCATCTGAAAGAGCAGCAGG 2160
VVDGVSGACKGRS P BGVRRVRQQGAGAMSQMAVS I 691
CTCCGCCATATATTGATCGATCGACCAATCGATCGTTAGTTCTCCAAGTTACTATTAGCTAGCTATAGCCCGAAACAGCTGAACTGATGATGACGATGGTAACCTCCGTOGTGTGTGTGC 2280
T AAGCATGT AGCGTGCT AGGAGATGAT AT ATT AAATAT AATCGAGT AGT AGAGCCT ACCCGCTGTGTGACGCT AAAT T TGTGTGCATTTGGTTTGGTT TGTGAGTTGGGCCCGTGCGTGG 2 00
CTGTGTCATGTCGTGGTT AATT AGCT AT ACTAGTCCTGTCTGTACATGCATGGACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 24 98
Fig. 4. Location of Nco\ (C/CATGG) restriction sites introduced by site-directed mutagenesis
(adapted from McCarty et al., 1991).

45
MS salts and MS vitamins (Sigma, cat# M5519) on a gyratory shaker in the darte for ca. 36 hrs,
while developing seeds were used immediately. The embryo as well as pericarp and testa
tissues were removed from the seeds to expose the aleurone layer of the endosperm. Prepared
endosperms were placed on Gelrite-solidified salt medium and then subjected to particle
bombardment.
Barley seeds were de-embryonated prior to surface sterilization in 70% ethanol for 1 min
followed by 10 min in 1.75 2.9 % NaOCI. A minimum of 1.75% NaOCI (=30% Clorox) was
necessary to allow easy removal of pericarp and testa layers prior to bombardment. Sterilized
half-seeds were imbibed overnight and prepared for particle bombardment as described above
for maize seeds.
Particle Bombardment
Particle bombardment was performed as described in Taylor and Vasil (1991) using a
DuPont PDS-1000 particle gun. Briefly, 35 to 50 pi of a sterile 50 mg/ml gold stock solution
(Biorad, 1.0 or 1.6 pm particle diameter; prepared in water) was mixed with premixed plasmid
DNA in a 250 pl-Eppendorf tube and vortexed on maximum speed for 10 s. Immediately, the
tube was shifted sideways and 10 pi of 0.1 M spermidine (free base) and 25 pi of 2.5 M CaCI2
were placed onto the side of the tube without allowing it to mix with the gold/DIMA solution. Then,
the tube was placed upright and subjected to vortexing for 10 s. The precipitated gold/DNA
particles were allowed to settle for ca. 3 min, after which part of the supernatant was removed
leaving 35-45 pi of liquid behind. The tubes were placed on ice until further use in particle
bombardment. For particle bombardment, 2 pi of sonicated gold/DNA solution (containing ca. 80
pg of gold) were used for individual shots.
The bombardment procedure had frequently to be adjusted to the gold characteristics
which varied substantially from batch to batch. Modifying the amount of gold used per shot was
found most successful in improving bombardment efficiency.

46
Incubation and Extraction of Endosperms Following Bombardment
Following bombardment, 1 ml of a solution containing MS salts and MS vitamins
supplemented with no hormones, 10" M GAj or 10" M GA3 and 10-4 M (or 10-5 M) ABA was
dripped over the endosperms. After 24 h of incubation in darkness, maize endosperms were
ground either individually On experiments using developing seeds) or in bulk from each
bombardment (when germinating seeds were used) with mortar and pestle aided by the addition
of silicon carbide powder in 200-1,000 pi of extraction buffer (0.1 M potassium phosphate (pH
7.8), 2 mM EDTA (pH 8), 2 mM DTT, 5% glycerol). The homogenates were centrifuged to
recover the cell extract. To obtain barley aleurone extract, the aleurone layers were separated
from the endosperms and ground in bulk for each replicate in 200 pi of extraction buffer. The
homogenates had to undergo two rounds of centrifugation to obtain clear cell extract.
Quantification of Transient Expression
Quantitative measurement of GUS activities was performed as described in Jefferson et
al. (1987) with the modification that the substrate MUG was dissolved in the extraction buffer
described above. For determination of luciferase activities, 10 pi aliquots of the extract and 200
pi of reaction buffer (25 mM tricine (pH 7.8), 15 mM MgCI2, 5 mM ATP, 0.05% BSA) were placed
in cuvettes and immediately assayed using a Monolight 2010 luminometer. The luminometer
automatically injects 100 pi of 1 mM luciferin and then counts the emitted photons for 15 s. The
unit of measurement is the Relative Light Unit (RLU).

RESULTS
Repression of Hydrolase Genes bv VP1 In Aleurones of Developing Maize Seeds
Phenotypic analysis of vp1-m2 kernels
The vp1-m2 allele of Vp1 carries a transposon insertion in the third intron which causes
somatic instability of the gene during endosperm development (McCarty et al 1989b). As a
result, mosaic kernels develop with clonal vp1 mutant and wild-type sectors. In these kernels, a
striking pattern of endosperm remobilization is often evident. Endosperm tissue underlying vp1
mutant aleurone cells is frequently softened and depressed in surface while wild-type sectors are
raised relative to adjacent mutant sectors. This produces kernels with a distinctive etched
appearance (Fig. 5a). The softening response was also observed when only a small fraction of
the endosperm was comprised of mutant tissue (Fig. 5b), indicating that expression of this
phenotype is cell autonomous. The softening of starchy endosperm tissues that underlie islands
of vp1 mutant aleurone cells appears to be attributed to precocious induction and secretion of
hydrolytic enzymes caused by the loss of VP1 function. Thus, repression of hydrolases in
developing maize kernels is evidently dependent on the presence of functional VP1.
Transient expression of Amv-GUS in maize aleurone
In order to more directly address the role of VP1 in repressing hydrolase activity, a
quantitative transient gene expression assay was developed that is based on particle
bombardment of aleurone tissue with a barley high-pl a-amylase promoter-GUS fusion construct
(Amy-GUS). It was first interesting to determine whether the observed differential activity of
hydrolases in vp1 mutant and wild-type sectors was due to transcriptional control. For this
purpose, Amy-GUS was introduced into aleurone cells of developing vp1 mutant and
47

48
Fig. 5. Cell autonomous de-repression of the aleurone germination response in vp1-m2 mutant
aieurone sectors. (A) The kernel shown is a mosaic: regions pigmented with purple anthocyanin
are wild-type; yellow, anthocyanin-deficient regions are clonal sectors of aleurone that have lost
Vp1 function. (B) Magnification of a vp1-m2 kernel.

49
wild-type maize seeds. Table 1 shows that during mid-late development Amy-GUS was not
expressed In developing wild-type aleurone, even in the presence of exogenous GA. In contrast,
GA-induction of Amy-GUS was detected in vp1 mutant aleurones as early as 20 days after
pollination PAP). These data indicate that in developing wild-type aleurone tissue a-amylase
genes are insensitive to GA while in vp1 mutant aleurone cells a-amylase expression is
transcriptionally de-re pressed.
Amy-GUS expression in vp1 mutant aleurone was found to be under developmental and
hormonal control. Prior to approx. 18 DAP, Amy-GUS was inactive in GA-treated as well as
untreated aleurone, indicating that early in seed development the aleurone is unresponsive to
GA even in the absence of VP1 protein. At 20 DAP, Amy-GUS was induced by exogenous GA,
whereas its activity remained low in untreated aleurone (Fig. 6). Late in seed development (24
DAP), Amy-GUS was constitutively active in the absence of GA, indicating a greatly reduced
dependence on exogenous hormone. Two observations, however, indicate that Amy-GUS
expression was not fully constitutive at this stage: 1) GA treatment significantly enhanced AMY-
GUS expression (as much as 3-fold over that of untreated aleurones) in some, but not all
experiments (Table 1, Fig. 6). 2) GA treated aleurones typically exhibited less quantitative
variation in Amy-GUS expression than non-treated sibling aleurones. The latter effect suggests
that developmental or spatial variation affecting endogenous hormone concentrations within the
ear or seed might contribute to the large variation observed in the absence of exogenous GA.
The differential expression of Amy-GUS in developing vp1 mutant and wild-type
aleurone cells confirms a role of VP1 in the repression of a-amylase genes during seed
development. In order to test whether expression of recombinant VP1 could evoke inhibition of
a-amylase transcription in vp1 mutant aleurones, aleurones were bombarded with a mixture of
Amy-GUS and 35S-Sh-VP1 plasmids. Co-expression of VP1 strongly inhibited Amy-GUS
expression in vp1 mutant aleurone in the presence as well as absence of exogenous GA (Fig. 6),
indicating that recombinant VP1 effectively restored the wild-type phenotype. We can rule out
the possibility that over-expression of VP1 causes non-specific squelching of general

50
Amy-GUS / LUC *104 [pmoles MU/h/RLU]
vp1-R mutant Aleurones
Wild-type Aleurones
Days after
Pollination
-GA
+GA
-GA
+GA
Range
Mean S.E.M.
Range Mean S.E.M.
Mean
Mean
18
<1
<1
<1
<1
20
<1
4-35 21
7
<1
<1
24
114-531
263 130
41-150 110
35
<1
<1
Table 1. Amy-GUS is inducible in vp1-R mutant aleurone cells but not in wild-type aleurone
cells. Aleurones of developing vp1-R mutant and wild-type kernels at 18, 20 and 24 DAP were
bombarded with a mixture of 10 pg of Amy-GUS and 5 pg of Ubi-LUC. Post-bombardment,
kernels were treated with a solution containing no hormones or 10' M GA3. Data represent
mean ( S.E.M) of three to five replicates.

51
no hormones
GA
+GA +ABA
Amy-GUS / LUC *10 4 [pmoles MU/h/RLU]
Fig. 6. Effect of VP1 over-expression and ABA on Amy-GUS expression in vp1-R mutant
aleurone. vp1-R mutant aleurones from kernels harvested 26 DAP were bombarded with 10 pg
of Amy-GUS, 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT (for no-VP1 controls).
Following bombardment, a solution containing no hormones, 106 M GA3 or 10'6 M GA3 and 10*
4 M ABA was applied to the kernels. Numbers behind bars represent means of five replicates.
Error bars show S.E.M.

52
transcription factors because no inhibitory effect of co-bombarded VP1 on 35S-Sh-GUS or
UbiquHin-Luciferase expression was observed (data not shown). Moreover, VP1 caused trans-
activation of positively regulated reporter constructs, Em-GUS and C1-Sh-GUS, in aleurone cells
using similar bombardment conditions (S. Cocciolone and D.R. McCarty, unpublished results).
Interaction between VP1 and Abscisic Acid
in concert with VP1, the hormone ABA plays an important role during seed maturation
(McCarty and Carson, 1991). Moreover, ABA functions as an inhibitor of a-amylase expression
in germinating cereal seeds (Jacobsen and Chandler, 1987). This suggests that ABA might also
be involved in repression of a-amylase genes in the developing seed. Therefore, possible
interactions between ABA and VP1 in repressing Amy-GUS were analyzed.
ABA was effective in blocking Amy-GUS expression in vp1 mutant aleurone (Fig. 6).
This indicates that repression by ABA does not require VP1. In combination, ABA and VP1 over
expression produced a roughly additive effect (Fig. 6).
To test whether a-amylase repression by VP1 is dependent on ABA, recombinant VP1
was over-expressed in aleurone of developing vp1,vp5 double mutant kernels that are deficient
for ABA biosynthesis. Figure 7 shows that VP1 was highly effective in repressing Amy-GUS in
vp5 mutant background. While it cannot by ruled out that maternal ABA derived from the vp5/+
parent plant may be sufficient for VP1 function, it is suggested that VP1-mediated repression of
Amy-GUS expression does not require ABA. This would be consistent with the finding that VP1
also functions in aleurone of germinating seeds (see below) where ABA levels are very low
(Oishi and Bewley, 1990). Taken together, these data suggest that ABA and VP1 inhibit Amy-
GUS expression independently.

53
EHHD-t- 35S-Sh-CAT
Mi + 35S-Sh-VP1
Fig. 7. Co-expressed VP1 inhibited Amy-GUS in aleurone of developing vp1lvp5 double mutant
seeds that are deficient for ABA biosynthesis. Kernels were harvested 24 DAP. Bombardments
were performed as described in Fig. 6. Following bombardment, kernels were incubated in no
hormones or 10'6 M GA3. Data represent mean ( S.E.M) of 7-8 replicates.
Effector Construct
Amylase-GUS / Ubi-LUC 104 (S.E.M.)
[pmoles MU/hr/RLU]
Maize Seeds
Barley Seeds
Amy-GUS
JR303
Amy-GUS
JR303
35S-Sh-CAT (Control)
91 <17>
1.19 i*0 35)
247 (36)
65 (5.2)
35S-Sh-Vp1
10.5 i*06*
0.07 <*014)
70(16)
25 i*6-4)
Fig. 8. Co-expressed VP1 inhibited Amy-GUS and JR303 in aleurone of germinating maize and
barley seeds. Aleurones of imbibed seeds were bombarded with Amy-GUS (maize: 4 pg; barley:
2 M9) or JR303 (maize: 10 pg; barley: 5 pg), 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or 35S-
Sh-CAT. Post-bombardment, kernels were incubated in 10'6 M GA3. Data represent mean (
S.E.M) of 3-5 replicates.
I

54
Over-expression of VP1 in Aleurones of Germinating Maize and Barley Seeds
Endogenous expression of Vp1 in embryo and aleurone tissues is under strict
developmental control. Vp1 mRNA peaks at 16 DAP and then gradually decreases as the seed
reaches maturity (McCarty et al., 1991). Germinating seeds, in contrast, display no Vp1
expression or detectable levels of VP1 protein (Carson, 1992). Thus, VP1 function in maize Is
limited to the maturing seed. To test whether VP1 can function in germinating seeds in a way
equivalent to maturing seeds, we co-expressed 35S-Sh-VP1 and Amy-GUS In aleurones of
germinating wild-type seeds of maize. In the presence of exogenous GA, VP1 reduced Amy-
GUS expression by ca. 90% (Fig. 8). Thus, VP1 also functions in germinating seeds, apparently
without the need for additional developmental factors. Furthermore, VP1 also repressed
expression directed by the barley low-pl a-amylase promoter (JR303) which shows considerable
sequence divergence from Amy-GUS. This indicates that expression of high- as well as low-pl a
-amylase genes is under control of VP1.
Because germination-specific responses are well characterized in barley, VP1-mediated
repression was tested in aleurones of germinating barley seeds. Though not as effective as in
maize, VP1 also inhibited Amy-GUS and JR303 expression in barley (Fig. 8).
Moreover, variation within as well as between experiments was significantly reduced in
aleurones of germinating wild-type seeds of maize and barley as compared to aleurones of
developing vp1 mutant maize seeds. Therefore bombardment of germinating seeds constitutes
a useful experimental system to further characterize VP1 function.
A VP1 dose-response curve was generated to determine the amount of co-expressed
VP1 necessary to achieve maximum repression of Amy-GUS in barley aleurone cells. Figure 9
shows that repression was already evident when 1.25 pg of 35S-Sh-VP1 were co-transferred with
Amy-GUS, while increasing the amount of 35S-Sh-VP1 beyond 2.5 pg did not lead to further
repression of Amy-GUS. Hence, comparatively low amonts of recombinant VP1 are sufficient to
achieve inhibition of Amy-GUS. To confirm that the inhibitory effect of VP1 on Amy-GUS in
barley aleurone is promoter-specific, 35S-Sh-VP1 was co-expressed with an Em-GUS reporter

55
Fig. 9: VP1 dose-response for repression of Amy-GUS in aleurone of germinating barley seeds.
0 to 10 pg of 35S-Sh-VP1 were co-precipitated with 0.5 pg of Amy-GUS and 5 pg of Ubi-LUC.
Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment,
endosperms were cultured in 10*6 M GA3. Data represent mean ( S.E.M) of five replicates.
Amount of 35S-Sh-VP1 added (ug)
Fig. 10: Co-expression of VP1 activated Em-GUS in barley aleurone. Aleurones of germinating
barley seeds were bombarded with 2 pg of Em-GUS, 5 pg of Ubi-LUC and 0, 1 or 5 pg of 35S-
Sh-VP1. Total amount of DNA was balanced by addition of 35S-Sh-CAT. Post-bombardment,
endosperms were cultured in no hormones. Data represent mean ( S.E.M) of five replicates.

56
construct containing the full length promoter of the wheat Em gene fused to the GUS coding
sequence (Marcotte et al 1989). Recombinant VP1 increased expression of Em-GUS by ca. 5-
fold (Fig. 10) which is consistent with the function of VP1 as a transcriptional activator of Em
(McCarty et al., 1991). In conclusion, these data show that VP1 can repress or activate gene
transcription depending on the promoter context.
Interaction between VP1 and Gibberellic Acid
The well characterized hormonal responses in Himalaya barley aleurone facilitated
further studies regarding the interaction between VP1 and GA. For this purpose, GA response
curves of Amy-GUS expression were determined in aleurones of de-germed imbibed Himalaya
half seeds* (Fig. 11). In the absence of co-expressed VP1, Amy-GUS expression showed a
typical GA-induction. In contrast, when a mixture of Amy-GUS and recombinant VP1 was
introduced into aleurone cells, GA-induction of Amy-GUS expression was reduced by ca. 80%.
Most noticeably, the clearly detectable basal activity of Amy-GUS was not significantly affected
by co-expression of VP1 (Fig. 11, insert). Thus, VP1 only inhibited the GA-dependent activity of
the a-amylase promoter. This implies that VP1 may interfere with the GA signalling pathway.
Recessive mutations that cause constitutive GA-response have been identified in barley
and a few other species (Ross. 1994). Barley slender (sin) mutant plants are characterized by
excessive elongation of stem and leaf tissues and constitutive expression of hydrolytic enzymes
in the aleurone of imbibed half seeds in the absence of exogenous GA. The mutant phenotype
suggests that the Sin gene encodes a negative regulator that is normally inactivated by GA
(Chandler, 1988; Lanahan and Ho, 1988). To test whether VP1 inhibitory function depends on
the presence of the SLN protein, aleurones of sin mutant half seeds were co-bombarded with
Amy-GUS and recombinant VP1. Figure 12 shows that VP1-mediated repression of Amy-GUS
was as effective in sin mutant aleurones as in wild-type aleurones.

Amy-GUS/LUC*1 E04 (pmoles/hr/RLU)
57
GA 3-Concentration (M)
Fig. 11. Co-expression of VP1 inhibited GA3*induction of Amy-GUS but did not affect its basal
activity in aleurone of germinating barley half seeds Aleurones were bombarded with 2 ng of
Amy-GUS, 5 pg of Ubi-LUC, and 5 ng of 35S-Sh-VP1 or 35S-Sh-CAT. Following bombardment,
3 replicates of 5 kernels each were incubated in 0-10'5 M GA3. Data represent mean of three
replicates ( S.E.M). The insert shows activities in the absence of GA3.

s
o
EHH3+35S-Sh-CAT
mm+ 35S-Sh-vpi
Fig. 12. VP1 function does not require the Slender gene product. Aleurones of imbibed slender
(sin) mutant and wild-type barley half seeds harvested from a plant segregating for the sin
mutation were bombarded with 5 ng of Amy-GUS, 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or
35S-Sh-CAT. Following bombardment, sin mutant or wild-type seeds were incubated in a
solution containing no hormones or 10'6 M GA-j, respectively Data represent mean ( S.E.M.)
of 10 replicates.

59
This indicates that VP1 is likely to act further downstream in, or independently of, the SLN
pathway.
While GA is a strong inducer of a-amylase genes in aleurones of Himalaya barley and
other cereals, the importance of GA in the regulation of maize a-amylase genes is less clear, a-
amylase activities were found high in isolated endosperms that had been de-embryonated prior
to imbibition (Harvey and Oaks 1974; Goldstein and Jennings, 1978). Moreover, application of
exogenous GA to isolated endosperms did not further enhance a-amyiase activities (Oishi and
Black, 1990). Hence, it was argued that mature endosperms store high concentrations of GA
(Harvey and Oaks 1974; Goldstein and Jennings, 1978; Oishi and Black, 1990). However, seeds
of the GA-deficient, extremely dwarfed mutant dS displayed considerable a-amylase activity that
was only 3-fold lower than in wild-type seeds, implying a GA-independent component in maize a-
amylase production. To investigate this, Amy-GUS was introduced into aleurones of germinating
wild-type and GA1-deficient d1 mutant seeds. Both genotypes displayed similar, high Amy-GUS
activities in the absence of exogenous GA (Figs. 13, 14). Furthermore, application of GA to d1
mutant seeds increased Amy-GUS expression by less than two-fold, thus to a similar extent as in
wild-type seeds (Figs.13, 14). Hence, the d1 mutation did not appear to alter Amy-GUS
expression in the aleurone, suggesting that deficiency in the highly active gibberellin GA1 does
not severely affect high-level expression of a-amylase genes. Thus, consistent with the data on
the mutant d5 (Harvey and Oaks, 1974), the possibility of a constitutive activity of a-amylase
genes in the absence of GA appears likely. The finding that co-expression of VP1 and
application of ABA reduced Amy-GUS activity to a very low level (Figs. 13, 14) indicates that
VP1 and ABA repress the GA-dependent as well as the putative constitutive activity of Amy-
GUS in maize.

60
1000
Fig. 13. Effect of GA, ABA and recombinant VP1 on Amy-GUS expression in aleurone of
germinating wild-type (NK508) seeds. Aleurones were bombarded with 4 pg Amy-GUS, 5 pg of
Ubi-LUC and 5 pg of 35S-Sh-VP1 or 35S-Sh-CAT and then incubated in a solution containing no
hormones, 10- M GA3, 10-5 M ABA, or 10-6 M GAg and 10- M ABA. Data represent mean (
S.E.M.) of five replicates.
300
Fig. 14. Effect of GA, ABA and recombinant VP1 on Amy-GUS expression in aleurone of
germinating d1 mutant seeds. Methods as in Fig. 13.

61
Role of the Embfvo in Repression ofa-Amylase Genes in the Aleurone
The differential response of mutant and wild-type aleurone tissue in developing vp1-m2
kernels (Fig. 5a,b) did not appear to be fully independent of the state of the embryo. Embryos of
vp1-m2 seed are frequently non-viviparous and survive desiccation. It was observed that
mosaic aleurones that were associated with non-viviparous embryos very rarely exhibited
precocious endosperm remobilization while those with viviparous embryos typically did.
Moreover, aleurone near the germinal face and crown of the kernel was more strongly affected
than aleurone on the abgerminal face. This phenotype suggests that signalling from the embryo
as well as responsiveness of the aleurone cells contribute to the softening response.
In order to assess the impact of the physiological state of the embryo on a-amylase
expression in aleurone cells, TB3La translocation stocks were used which allow the generation of
vp1 non-concordant seeds with embryo and endosperm of different genetic constitution (for a
brief description of the system see Materials and Methods). As a result, ears developed that
segregated four genotypes: 1) seeds with a vp1 mutant embryo and a wild-type endosperm, 2)
seeds with a wild-type embryo and a vp1 mutant endosperm, 3) concordant vp1 mutant seeds
and 4) concordant wild-type seeds. Aleurones of these four genotype combinations were
bombarded with Amy-GUS and postbombardment cultured in the absence of added hormones.
Amy-GUS was only expressed in aleurones of concordant vp1 mutant kernels and not in any of
the other three genotype combinations (Fig. 15). These data are consistent with the observation
made in vp1-m2 seeds that endosperm tissue underlying vp1 mutant aleurone was remobilized
predominantly in viviparous seeds. Thus, a-amylase genes appear to be de-repressed in
aleurone cells that lack functional VP1 predominantly if the embryo is also viviparous. In some
experiments in which kernels very late in development were used, both vp1 non-concordant
genotypes displayed significant Amy-GUS activities, while no Amy-GUS expression was
detected in aleurones of concordant wild-type seeds (data not shown). This indicates that partial
de-repression of Amy-GUS in aleurone cells is facilitated if either the embryo is viviparous or the
aleurone cells lack functional VP1.

62
111
10-
9-
Embryo:
vp1
WT
vp1
WT
Endosperm:
vp1
vp1
WT
WT
Fig. 15. Amy-GUS expression in aleurone of developing vp1 non-concordant maize seeds.
Aleurones (31 DAP, fall season) were bombarded with 10 pg of Amy-GUS and 5 pg of Ubi-LUC
and cultured post-bombardment in no hormones. Data represent mean ( S.E.M.) of six
replicates.

63
Fig. 16. Amy-GUS expression in aleurone of developing germless seeds. Aleurones (29 DAP,
fall season) of an ear segregating for the mutations vp1 and germless were bombarded and
cultured as described in Fig 15. Crosses represent single data points of six replicates
(exception: four replicates in the vp1/germless double mutant).

64
There are at least two possible scenarios that would explain how the state of the embryo
might influence de-repression of a-amylase genes in vpf mutant aleurone: 1) a viviparous
embryo might secrete an inductive signal required for a-amylase expression in the aleurone of
developing seeds (e.g. GA), and/or 2) a non-viviparous embryo might contribute a diffusible
inhibitory signal that prevents a-amylase expression in the aleurone (e.g. ABA). In order to test
these hypotheses, Amy-GUS was introduced into aleurone cells of a germless mutant in which
the embryo aborts during the early globular state (P. Becraft and D.R. McCarty, pers.
communication). Hence, use of ears that segregate for the germless and vpf mutations allows
assessment of aleurone responsiveness in the absence of a signal from the embryo. Amy-GUS
was highly de-repressed in vpf-mutant aleurone of germless seeds. Aleurones of all four seeds
bombarded expressed Amy-GUS (Fig. 16). This indicates that a viviparous embryo per se is not
required for de-repression of a-amylase genes in vpf mutant aleurone cells. Rather, it appears
that the lack of a normal (non-viviparous) embryo caused induction of Amy-GUS in the
germless,vp1 double mutant. This suggests that a wild-type embryo secretes a signal with
inhibitory function on a-amylase expression in the aleurone.
In the single mutants vpf and germless, Amy-GUS was partially de-repressed (2-3 seeds
of a total of six bombarded expressed Amy-GUS, Fig. 16). Only wild-type seeds displayed
complete repression of Amy-GUS (Fig. 16). Hence, both VP1 expression in the aleurone and a
normal embryo appear to be required for complete inhibition of a-amylase genes in the aleurone.
This is consistent with the Amy-GUS activities found in vpf non-concordant kernels very late in
development, as described above.
Functional Analysis of the VP1 Protein
The acidic activation domain
We considered two models of how VP1 may function in repression of the aleurone
germination response. 1) VP1 might be a transcriptional activator of an intermediate repressor
gene that in turn inhibits expression of a-amylase genes (Figure 17a). 2) VP1 itself might

65
c.
Effector Construct
Rel. Amy-GUS / LUC
Maize Barley
35S-Sh-CAT (Control)
100<25>
100 <417)
activator
WT-VP1
i ta mm 1
i mmmS1
16(2)
42 (8>
^28-121
I M I
1
|
&
13(4)
17<>
3x (VP16 act)
mm
m wmm1
I n.d.
12 (t7)
Fig. 17. Mode of action of VP1 in repressing Amy-GUS.
(A),(B). Alternative models for VP1 action as described in the text.
(C). Effect of deletion and substitution derivatives of VP1 on Amy GUS expression in aleurone
tissue of germinating maize and barley seeds cultured in GA3. In A28-121, the activation
domain of VP1 was deleted. In 3x (VP16 act), the activation domain of VP1 was replaced by
three copies of the Herpes simplex VP16 activation domain. Data represent activities (mean
S.E.M) relative to control (=100). Black boxes show sequence homology between VP1 and
ABI3. (n.d.: not determined).

66
function as a repressor of the a-amylase genes or of an intermediate gene that is required for
activation of the a-amylase promoter (Figure 17b). In order to distinguish between these
models, we determined whether the transcriptional activation domain of VP1 which is essential
for activation of the Em and C1 genes in maize cells is also required for inhibition of a-amylase.
Figure 17c shows that a deletion derivative of VP1 that lacks the N-terminal activation domain
was as effective in repressing Amy-GUS expression in maize and barley as the full-length
protein. In addition, a VP16/VP1 hybrid protein that contains three copies of the VP16 acidic
activation domain and has a restored capacity to activate Em-GUS and C1-Sh-GUS (McCarty et
al., 1991; Rosenkrans and McCarty, unpublished results) was not more effective than the
activator deletion mutant in causing repression of Amy-GUS. The lack of a requirement for the
activation sequence clearly distinguishes the mechanism of VP1 -mediated repression from the
mechanism of activation of diverse maturation related genes by VP1. These results strongly
indicate that the VP1 protein has a discrete repressor function.
Identification of Sequences Essential for the Repressor Function of VP1
A number of internal deletion constructs were tested for their ability to repress Amy-GUS
in maize and barley aleurone (Fig. 18). A large ca. 350 bp deletion (86/87) entirely abolished
VP1 repressor function, indicating that the deletion-derivative may lack a functionally important
domain, or the deletion may affect the spacing and thereby the function of domains present
outside this sequence. An only slightly smaller deletion in this region (86/85) did not affect
repression in maize nor barley aleurone, suggesting that altered spacing is not likely the reason
for the failure of 86/87 to repress. Indeed, a small deletion of 42 bp in the C-terminal half of
86/87 (85/87) abolished repression in barley and consistently reduced activity in maize aleurone
by ca. 50%. Hence, this region (hereafter referred to as the RED domain) appears to be
essential for VP1 repressor function. Consistent with this conclusion, a large part of the
sequence of the RED domain (W V Q N H+ H* L Q Rf K* R* P R* D) is highly charged,
predicting this domain to be positioned on the surface of the folded VP1 protein, thus accessible
for interactions with other molecules.

87
Re. Amy-GUS / LUC
Effector Construct Maize Barley
VP1-WT
86/87
86/85
85/87
87/88
87/92
92/181
93/95
92/88
196/88
Control (no VP1)
100
100
4-12
18-33
133 (29)
87 (*8)
7 (2)
13(*3)
45 (*)
99(*21)
85-162
90-231
9(*D
39 (1)
6 (2)
29 (*7)
14(*2)
33 (*)
27-100
125 (17)
?
84 (*s)
Fig. 18. Deletion analysis of the VP1 protein: Part I. For Materials and Methods see Fig. 19.
Black boxes indicate sequence homology to the VP1 homolog from barley.

68
Re. Amy-GUS / LUC
Effector Construct Maize Barley
VP1-WT
103/104
101/100
VP1-McW
Control (no VP1)
H
A
II

III
100
100
4-12
15-30
21 (*4>
27 <*3)
18(2)
23 (4)
32 (*9)
13 (*-5)
Fig. 19. Deletion analysis of the VP1 protein: Part II. Aleurones of maize and barley
germinating seeds were bombarded with 2-5 pg of Amy-GUS, 5 pg of Ubi-LUC and 5 pg of
effector construct and then cultured in 10-6 M GA3. Data represent mean ( S.E.M.) of 3-5
replicates

69
A second large, ca. 400 bp deletion (87/88) also rendered the VP1 protein incapable of
repressing Amy-GUS (Fig. 18). The activity of this construct varied to an unusual extent, from
slight, but non-significant repression in some experiments to more than two-fold, statistically
significant activation of Amy-GUS in others. Similarly, a construct with a slightly smaller deletion
of this region (92/88) displayed a highly variable effect. Subsequently, four adjacent sub
deletions within the 87/88 domain were constructed. Each of the N-terminal three deletions
(87/92, 92/181, 93/95) eliminates one region that is conserved in the barley VP1 homolog, but
none of these deletions severely diminished repression. In contrast, the C-terminal, ca. 140 bp
deletion 196/88 eliminating a non-conserved stretch of VP1 almost entirely abolished repressor
function, implying that this region may contain an important site involved in Amy-GUS
repression.
Deletions in the C-terminal portion of VP1 did not strongly affect repression of Amy-GUS
(Fig. 19). Truncation of the C-terminal 450 bp (VP1-McW) generating the product of the vp1-
McWhirter allele which confers a non-viviparous, anthocyanin-deficient phenotype had only a
slight effect in maize, while not affecting repression in barley. Similarly, deletion of the domains
101/100 and 103/104 did not strongly diminish Amy-GUS repression.
The 87/88 deletion mutant
It was shown that in the absence of GA, co-expression of VP1 had no effect on Amy-
GUS expression in aleurone of germinating barley half seeds (see Fig. 11). In contrast, when
the 87/88 deletion-derivative of VP1 was over-expressed with Amy-GUS in the absence of GA,
Amy-GUS expression was activated (Fig. 20). This activation was highly variable, ranging from
2-fold in some experiments to up to 12-fold in others. It appeared specific to the 87/88 deletion
mutant and was not found for any other tested constructs containing deletions outside this region
of VP1 (data not shown). Interestingly, activation of Amy-GUS by 87/88 was also observed in
the presence of ABA (Fig. 21). In the presence of GA, a slight activation of Amy-GUS (max. 2-
fold) by 87/88 was observed in some, but not all experiments (Figs. 18, 20).

70
Fig. 20. The VP1 deletion-derivative 87/88 activates Amy-GUS in the absence of GA in
aleurones of germinating barley seeds. Aleurones were bombarded with 2 pg of Amy-GUS, 5 pg
of Ubi-LUC and 10 pg of either 35S-Sh-CAT, 35S-Sh-VP1, the activation domain-deletion
mutant described in Fig. 18 (-Act), 87/88 or the double mutant that carries deletions of the
activation domain and 87/88 (-Act:87/88), respectively. After bombardment, endosperms were
cultured in no hormones (top graph) or 10-6 M GA3 (bottom graph). Data represent mean (
S.E.M.) of three replicates.

71
Emms no Hormones
¡ABA
Fig. 21. ABA does not inhibit 87/88-mediated activation of Amy-GUS found in the absence of
GA. Aleurones of germinating barley seeds were bombarded as described in Fig. 20 and then
cultured in no hormones or 10*5 M ABA.
il.i
CAT 07/88 VP1 VP1+
87/88
Fig. 22. The VP1 deletion mutant 87/88 displays a dominant negative effect on VP1-mediated
repression of Amy-GUS. Aleurones of germinating barley seeds were bombarded with 6 pg of
Amy-GUS, 5 pg of Ubi-LUC and either 25 pg of CAT, 20 pg of 87/88, 5 pg of VP1 or both 20 pg
of 87/88 and 5 pg of VP1. To all mixtures, CAT plasmid was added to obtain a total amount of
36 pg of plasmid DNA. After bombardment, endosperms were cultured in 10-6 M GAg. Data
represent mean ( S.E.M.) of five replicates.

72
CAT VP1 VP1* VPU VP1*
87/88 85/88 87/88-kb W
Fig. 23. Effect of the double deletion mutants 85/88 and 87/88:McW on inhibition of Amy-GUS
by over-expressed VP1. Materials and Methods as described in Fig. 22. The double mutants
85/88 and 87/88:McW were constructed by restriction enzyme digestion and subsequent ligation.
Data represent mean ( S.E.M.) of 5-6 replicates.
Fig. 24. Effect of the double deletion mutants 87/104 and 87/88:101/100 on inhibition of Amy-
GUS by over-expressed VP1. Materials and Methods as described in Fig. 22.

73
The finding that 87/88 activates Amy-GUS indicates that the mutant protein is not fully
non-functional. The data suggest that 87/88 may be capable of interacting with a normal
component of the repression mechanism but unable to cause repression. In doing so, there are
at least two possibilities as to how it might activate Amy-GUS. 1) The acidic activation domain
of 87/88 might elicit transcriptional activation of a-amylase genes or an intermediate gene. In
the wild-type VP1 protein, this activity might either be not accessible or masked by the repressor
function. 2) In producing a non-functional complex, 87/88 might compete with, or titrate out, an
endogenous repressor (e.g. endogenous barley VP1 -homolog possibly present in aleurone of
germinating barley seeds) and thus exert a dominant negative effect.
In order to test the first possibility, a double deletion mutant was constructed that deletes
the 87/88 domain and the acidic activation sequence. This double mutant was as effective in
activating Amy-GUS as the 87/88 single mutant (Fig. 20), suggesting that the transcriptional
activation domain is not involved. Therefore, it was tested whether 87/88 is capable of inhibiting
the effect of recombinant VP1. 87/88 and recombinant VP1 were expressed by themselves and
in combination (ratio 4:1) together with Amy-GUS in barley aleurone. Co-expression of 87/88
reduced VP1 -mediated repression of Amy-GUS by ca. 75% (Fig. 22). This is consistent with the
view of a dominant negative effect of 87/88.
To identify domains involved in mediating the dominant negative effect of 87/88, double
deletion mutants deleting 87/88 and other sequences of the VP1 protein were constructed and
tested for their ability to reduce repression of Amy-GUS by co-expressed VP1. Double mutants
deleting the domains 87/88 and 85/87 (the RED domain) or 101/100, respectively, were as
effective in competing with recombinant VP1 as the 87/88 single mutant (Figs. 23, 24). In
contrast, the double mutants deleting 87/88 and either the C-terminal 450 bp of VP1
(87/88:McW) or the domain 103/104 (87/104) did not show a dominant negative effect on Amy-
GUS repression by co-expressed VP1 (Figs. 23, 24), suggesting that these domains may be
important for the inhibitory role of 87/88 on VP1 repressor function.

DISCUSSION
VP1 of maize is a transcription factor that is specifically expressed in the developing
seed (McCarty et al., 1989a, 1991). It was shown previously that VP1 is required for ABA-
induced activation of a variety of genes associated with seed maturation (McCarty et al., 1991).
Results of this work show that, in addition to its transcriptional activator function, VP1 has a
specific role in blocking precocious induction of germination-specific a-amylase genes during
seed development.
VP1 Represses a-AmvIase Genes
This study provides at least three lines of evidence that indicate a function of VP1 in
repression of a-amylase genes in the developing seed. First, somatically unstable vp1-m2 seeds
containing both vp1 mutant and wild-type sectors displayed cell autonomous de-repression of
endosperm remobilization specifically in sectors underlying vp1 mutant aleurone (Fig. 5a,b).
Second, in transient expression experiments Amy-GUS was inducible or constitutively active in
developing vp1 mutant aleurone cells but not in wild-type aleurone cells (Table 1). Third, co-
expression of recombinant VP1 with Amy-GUS in vp1 mutant aleurone cells inhibited Amy-GUS
expression by >95% (Fig. 6). These results are consistent with findings that a-amylase genes
are not expressed in the developing seed (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et
al., 1990; Oishi and Bewley, 1990). Hence, cessation of VP1 expression prior to germination
may be necessary to allow induction of a-amylase genes in the germinating seed.
74

75
Gene Repression is a Discrete Function of VP1
In contrast to the mechanism of transcriptional activation of maturation-specific genes,
VP1 -mediated repression of a-amylase genes does not require the transcriptional activation
function located at the N-terminal domain of VP1 (Fig. 17). This indicates that VP1 has a
discrete repressor function that is mechanistically distinct from the transcriptional activation
function. Several systems in which a single transcription factor functions as both an activator
and a repressor depending on the target promoter have been described in animals (Miner and
Yamamoto, 1991; Tsai and OMalley, 1994). Direct structural homologs of VP1 are thus far
known only in plants, suggesting that this strategy has evolved independently in plants and
animals.
Functional Analysis of the VP1 Protein
To identify domains in the VP1 protein that are important for repressor function, mutant
derivatives containing deletions covering ca. 80 % of the total protein were tested for their ability
to inhibit Amy-GUS. Deletion of very highly conserved sequences in the C-terminal half of VP1
(103/104, 101/100, McW) did not, or only slightly, reduce repressor function (Fig. 19). In
contrast, two constructs deleting sequences in the middle of the VP1 protein (85/87, 87/88) were
strongly affected in repression of Amy-GUS (Figs. 18). While disruption of VP1 function is one
possibility for lack of Amy-GUS repression, low stability of mutant mRNA or protein could be an
alternative explanation. However, this possibility is unlikely for two reasons: 1) both constructs
were capable of activating a C1-Sh-GUS reporter gene in maize protoplasts: 85/87 and 87/88
activated C1-Sh-GUS at a level of 77-84% or 56% of the wild-type VP1 construct, respectively
(V. Vasil, L. Rosenkrans et al., unpublished results). 2) co-expression of 87/88 as well as the
double-deletion mutant 85/88 exhibited a dominant negative effect on repression of Amy-GUS
by wild-type VP1 in barley aleurone, indicating presence of mutant protein in transformed cells.

76
The strongly positively charged, 15-amino acid-domain (RED domain), deleted in 85/87,
is highly conserved (ca. 80%) among maize, barley and rice genes (Fig. 25). Between maize
and barley, 11 out of 15 amino acids are identical and one amino add constitutes a conservative
substitution (R to K) also found in rice. The high degree of sequence conservation is consistent
with the finding that this region is of functional importance for repression of Amy-GUS.
Interestingly, deletion of the RED domain exhibited a differential effect in maize and barley
aleurone. While 85/87 was incapable of inhibiting Amy-GUS in barley aleurone, it retained ca.
50% of wild-type VP1 activity in maize aleurone cells (Fig. 18). In maize, deletion of additional
sequences 5' to the 85/87 deletion (construct 86/87) was necessary to eliminate repressor
function (Fig. 18). This indicates that in the barley cell the RED domain is absolutely essential
for proper execution of the repression mechanism, whereas in the more concordant system of
the maize aleurone cell other possibly less conserved regions in the VP1 protein partially
compensate in function for the RED domain.
The construct 87/88 that deletes a large but poorly conserved stretch of VP1 was
incapable of repressing Amy-GUS in maize or barley cells. To further analyze this domain of
VP1, four smaller deletions within this region were constructed and tested (Fig. 18). The three
deletion constructs 87/92 (deleting 29 aa), 92/181 (deleting 26 aa) and 93/95 (deleting 40 aa) of
which each lacks a short stretch of conserved sequence were not severely impaired in repressing
Amy-GUS. In contrast, the construct 196/88 deleting 48 amino acids at the C-terminal end of the
87/88 deletion was incapable of repression in barley (no data for this construct were obtained in
maize). These results allow at least two interpretations: 1) 196/88 deletes a domain essential for
repression, while the sequences located between the deletion points 87 and 95 are not required
for repressor function. 2) The partially conserved sequences between 87 and 95 are of
redundant function. Therefore, deletion of two or more conserved blocks may be necessary to
lose repression of Amy-GUS. In this interpretation, 196/88 may delete an additional important
domain or, alternatively, affect proper spacing between further N-termnal and C-terminal
sequences.

77
Z.m. MEA-SSGSSPPHSQENPPEH GGD M-GG AP-AEEI GGEAA DDF 39
I I I I It I I I II II Ml
H.V. MDA-SAGPPPPRHPQGSALRRGKG P-AVEIRHGE DDF 34
III III I II I M
O.s. MDA-SAGS SAP HSHGNPGKQ-GGG GGGGGGRGKAP-AAEIR-GEAAR DDV 46
I I I II I I M
A.t. MKSLHVAANAGDLAEDCGIL-GGDADDTVLMDGIDEVGREIWLDD HGGDNNHVHGHQDDL 60
Z.m. MFAEDTF PSLPDFPCLSSPSSSTFSSN S S S NS S S AYTNT AGRA- G G 86
I I I I I I I I I I I I I I I I I I I I
H.V. MFAQD--TF PAFPDFPCLSSPSSSAADIV LCG 64
I I I I I I I I I I I I I I I I I I I
0.3. FFADDTF PLLPDFPCLSSPSSSTFSSS SSSNSSSAFTTAAGGGCG G 94
I I I I I I I I I I I I I I I II I I I I
A.t. IVHHDPSIFYGDLPTLPDFPCMSSSSSSSTSPAPVNAIVSSASSSSAASSSTSSAASWAILRS 123
Z.m. EPSEPASAGEGFDA LDDIDQLLDFASLSMPWDSE-P 125
I I I I I I I I I I I I I I I I I I I I I I I I
H.V. EPSEPAAAGDGMDD LS DIDHLLDLASINDDVPWDDE PL 102
I I I I I I I I I I I I I I I I I I I III III
0.3. EPSEPASAADGFGE LADIDQLLDLASLSVPWEAEQPL 135
I III
A.t. DGEDPTPNQNQYASGNCDDSSGALQSTASMEIPLDSSQGFGCGEGGGDCIDMMETFGYMDLLD 186
86
Z.m. FP-GVSMMLENAMSAPPQPVGDGMSEEKAVPEGTT GGEEACH-DASEG-EE 163
I I I I I I I I I I I I I I I I
H.v. FP-DVGMMLEDVIS E QOQQQQQH P LAGHGAGGRVAS DTAGG- GGE DAFMGGGGS GS AADD 160
II Mill II I I I I I II II I II I III
0.3. FPDDVGMMIEDAMSGQPHQADDCTGDGDTKAVMEAAGGGDDAGDACM-E-GS-DAPDD 179
II III I I I I I I
A.t. SNEFFDTSAIFSQDDDTQNPNLMDQTLERQEDQV-WPMMENNS-GQDMQMMNSSLEQDDD 240
85
Z.m. LPRFFMEWLTSNRENISAEDLRGIRLRRSTIEAAAARX;GGGRQGTMQLLKLILTWVQNHHLQR 230
1111111111 i 111111 Ti 1111 ill Ti 1111 111111 ni 1111 1111
H.V. LPRFFMEWLTNIRDCIS AEDLLSIRLRRS TIETTTALLGGGRQDTMQLLKLILTWVQSHHLQK 223
ll 1111111 ll III ll iTi 111ill l Tl 1111 111111iTi1111 11111
0.3. LPAFFMEWLTSNREYISADDLRSIRIRRSTIEAAAARLGGGRQGTMQLLKLILTWVQNHHLQK 247
I I I I I I II I I I I I ~ I III I I iTi I I I I I IiTi I I I III
A.t. LAAVFLEWLKNNKETVSAEDLRKVKIKKATIESAARRLGGGKEAMKQLLKLILEWVQTNHLQR 308
87 92 93
Z.m. KRP RDVMEE EA- GLHVQLP S PVANP PGYE FP AGGQDMAAGGGT SWM PHQQAFTPPAAYG 288
I I I I II I I I I I I I I I I I I I I I I I I I I I I I I I
H.v. KRPRVGAMDQEAPPAGGQLPSPGANPS-YEFPT ETGAAAATSWM PY-QAFSPTASYG 278
I I I I I I I I I I I I I I I I I I I I I I I I I I I I II I I I I I I
0.3. KRPRTAIDDGAA-SSDPQLPSPGANP-GYEFPSGGQEMGSAAATSWM PYQ-AFTPPAAYG 304
II I I I II I
A.t. RRTTTTTTNLSY-QQSFQQDPFQNPNPNNNNLIPPSDQTCFSPSTWPPPPQQQAFVSDPGFG 370
181 196,95
Z.m. GDAVYPSAAGQQYSFHQGPSTSSVWNSQPFSPP PVGDMH GANMAWPQQYVPFPPPG 345
I I I I I I I I I I I I I I I I I I I III I I I I I I III
H.v. GEAMYPFQ QGCSTSSVAVSSQPFSPPAAA-DMHAG AWPLQYAAFVPAG 325
I I I I I I II I I I I I I I I I I I I I I I I I I I I I I I I I I I
0.3. GDAMYPGAAG-PFPFQQSCSKSSVWSSQPFSPPTAAAAGDMHASGGGNMAWPQQFAPFPV 364
II I I II I I I I I I I
A.t. YMPAPNYPPQPEFLPLLESPPSWPPP PQ SGPMP-HQQF-PM-PPT 412

78
88 104 **** 103
Z.m. ASTGS YPMPQPFSPGFGCQYAGAGAGHLSVAPQRMAGVEASATKEAWOCRMARQRRLSCL 405
i mi ii ii mi i 111111 n 11 nnn111
H.V. ATSAGTQTYPMPPP-GPV-PQPFAAPGFAGQFPQRM EPAATREARKKRMARQRRLSCL 381
| || I I lili I I I I I I I I I I I I I I I I
O 3. SSTSS YTMPSWPPPrTAGFPGQYSGGHAMCSPRIAGVEPSSTKEARKKRMARQRRLSCL 424
| | I I II I II I I I I I I I I I I I I I I
A. t. SQYNQFG DPTGFNGYNMNPYQYPYVPAGQMRDQRLLRLCSSATKEARKKRMARQRRLL 470
Z.m. QQQRSQQLSLGQIQTSVHLQEPSPRSTHSGPVTPSAGGWGFWSPSSQ QQVQNPLS-K5N 463
i in i nn n n n ii
H.v. Q???????????IQTGGFPQQPS PRAAHSAFVWG7HWS PPAVQAQPHGQLMIQVPNPLSTKSN 444
I I I II I i I I I I I I I I I I II II I I
O.s. QQQRSQQLNLSQIHISGHPQEPSPRAAHSAPVTPSSAGCRSWGIWPPAAQIIQNPLSNKPN 485
I I III II II
A. t. SHHHRHNNNNNNNNNNQQNQTQIGETCAAVAPQLN PVATTATGGTWMYWPN 521
Z.m. SSRAPPSSLEAAAAAPQTKPAP-AGARQDDIHHRLAAASDKRQGAKADKNLRFLLQKVLKQSD 526
III III I lilil I I I I I I I I I I I I I I
H.v. SSRQKQQKPSPDAAAR-PPSGGGASQQRQGQ AAASDKQRQQ K? LRFLLQKVLKQS D 499
I I I III I I I I II II I II I II II
O.s. PPPATSKQPKPSPEKPKPKPQAAATAGAESLQRSTASEKRQ-AKTDKNLRFL1.QKVLKQSD 545
IIIII I III I I I I II I II II II I I II I
A. t. VPAV PPQLP PVMETQLPTMDRAGSASAMP RQQWP DRRQGWKPEKNLRFLLQKVLKQS D 581
/McW 101
Z.m. VGSLGRIVLPKKEAEVHLPELKTRDGISIPMEDIGTSRVWNMRYRFWPNNKSRMYLLENTGEF 589
II II IIII II II II I II II II II II I I I I II I I II IIII I II II II II II II II II
H.v. VGTLGRIVLPKKEAETHLPELKTGDGISIPIEDIGTSQVWSMRYRFRPNNKSRMYLLENTGEF 562
II II II IIII II II II II I II II III I II II II II IIII II II II II II II II II I
O.s. VGSLGRIVLPKKEAEVHLPELKTRDGVSIPMEDIGTSQVWNMRYRFWPNNKSRMYLLENTGDF 600
II IIII I I II I I I I I I II I III I II IIII I II I IIIIII I I II I I I I II I II I II
A.t. VGNLGRIVLPKKEAETHLPELEARDGISLAMEDIGTSRVWNMRYRFWPNNKSRMYLLENTGDF 644
100
Z.m. VRSNELQEGDFIVTYSDVKSGKYLIRGVKVR-PPAQEQGSGSSG-GGKH RP-LC- 640
II II I IIII II II II II II II II I III I I II I
H.V. VRSNE ? ? ? 7DFIVLYSDVKSQKYLIRGYKVRAAQELASTRWQSREGGA? ?V LAQ 616
II I II II II II I II II II II II I II II I I
O.s. VRSNELQEGDFIVIYSDIKSGKYLIRGVKVRR-AAQEQGN SSGAVGKHKHGSPEKPGVSS 667
I I II II II II I I II I I II I I II II I I I III
A.t. VKTNGLQEGDFIVIYSDVKCGKYLIRGVKVRQPSGQKPEA-PPSSAATKR 693
Z.m. P AGP E RAAAAGAP E DAWDGV SGACKGRSPEGVRRVRQQGAGAMSQMAVSI
H.v. TAAD
I I I II I II II II I I II II I II II I
O.s. NTKAAGAEDGTGGDDSAEAAAAAAAGKADGGGCKGKSPHGVRRSRQEAAAAAS MSQMAVSI
III I
A.t. QNKSQRNINNNSPSA-NVWA SPTSQTVK
691
620
728
720
Fig. 25. Alignment of the amino add sequences of VP1 from maize (Z.m.), HWP1 from barley
(H.v.), OSVP1 from rice (O.s.) and ABI3 from Arabidopsis (A.t.). Identical amino adds are
shown by vertical lines. Conserved regions among the four spedes are boxed. Italic numbers
above sequence show site of deletion points. Underlined amino adds indicate location of
putative leudne zipper. Asterisks indicate putative NLS.

79
In summary, using single-deletion mutants, amino add residues within the region from
222 to 374 of VP1 were found to be important for repression of Amy-GUS. Further functional
analysis of sequences outside this region was conducted by taking advantage of the dominant
negative effect of the 87/88 mutant on repression of Amy-GUS by wild-type VP1: In the
presence of GA, co-expression of 87/88 with VP1 and Amy-GUS severely reduced repression of
Amy-GUS by VP1 (Fig. 22). Domains essential for mediating the dominant negative effect of
87/88 are most likely also involved in function of the wild-type VP1 protein. 87/88 may exert a
dominant negative effect for instance by competing with wild-type VP1 for binding to a
component of the repression mechanism or if VP1 functions as a dimer by forming non
functional heterodimers with wild-type VP1. To identify domains in the 87/88 mutant that are
required for expressing the dominant negative effect, double-deletion mutants between 87/88
and other deletion mutants were constructed and tested. The RED domain deleted in the 85/87
mutant was considered a putative domain because it has an important function in repression.
However, the double-deletion mutant 85/88 was as effective in causing a dominant negative
effect as the 87/88 single mutant, indicating that the RED domain is not essential for this effect
(Fig. 23). Similar results were obtained for the domain 101/100 (Fig. 24). In contrast, the double
mutants deleting 87/88 and either the C-terminal 450 bp of VP1 (87/88:McW) or the highly basic
domain 103/104 (87/103) did not exhibit a dominant negative effect on Amy-GUS repression by
co-expressed VP1 (Figs. 23, 24). Although it cannot be ruled out that these double mutant
constructs express instable proteins, these data suggest that the C-terminus and the domain
103/104 may be required for mediating the dominant negative effect of 87/88. Deletion of these
domains displayed a clearly measurable effect only if the 87/88 domain was deleted also, while
the single mutants 103/104 and McW retained almost wild-type repressor function (Fig. 19). This
suggests that in the single mutants other sequences can compensate in function for the deleted
domains, while this is not possible in the 87/88 mutant. However, stability of the mutant proteins
needs to be confirmed, especially for the 87/88:McW double mutant which deletes ca. 42% of
the VP1 sequence.

80
In summary, deletion analysis of the VP1 protein has allowed the Identification of several
domains that are essential for repressor function: 1) the conserved, highly charged RED domain
85/87 and 2) the poorly conserved region 87/88. Moreover, though not essential, the domain
103/104 and the C-terminus may play a role in repression.
When comparing the repressor domains of VP1 with the domains required for
transcriptional activation of the Em or C1 genes, it is evident that different functions of VP1 map
to different sequences in the protein (Fig. 26). Apart from the differential requirement of the
acidic activation domain at the N-terminus of VP1, a-amylase repression and Em activation
differed in the need for the highly positively charged domain 103/104 of VP1. While deletion of
103/104 reduced Em-GUS activation by 98% in maize protoplasts (L. Rosenkrans et al.,
unpublished results), it did not severely affect repression of Amy-GUS. Likewise, a-amylase
repression and C1 activation displayed a differential requirement for the C-terminal part of VP1.
Deletion of the C-terminal ca. 150 amino acids of VP1 entirely eliminated activation of C1-Sh-
GUS (L. Rosenkrans et al., unpublished results). In contrast, this domain was not found
essential for repression of Amy-GUS.
Similarly, different domains of VP1 are essential for activation of C1 and activation of
Em. Overall, sequences required for induction of anthocyanin biosynthesis map to the highly
conserved C-terminal end of VP1, while sequences essential for activating Em are located in the
central part of VP1 (85/87, 87/88 and 103/104). These findings are consistent with the
phenotype of mutant alleles in maize and Arabidopsis. The maize vp1-McW allele truncating the
C-terminal ca. 150 amino acids of VP1 produces seeds exhibiting nearly normal developmental
arrest but lack of anthocyanin accumulation, indicating that the C-terminal part of VP1 is not
essential for preventing vivipary but is essential for anthocyanin production (McCarty et al.,
1989b). In Arabidopsis, two abi-3 alleles that produce a viviparous phenotype have been
sequenced: ab/3-6 contains an internal deletion of ca. 750 bp between positions 1,073 and 1,944
(Nambara et al., 1994), thus deleting the domains corresponding to 196/88 and 103/104. The
mutation in ab/3-3 induces a premature stop codon at Gin417 (Giraudat et al., 1992), thus deleting

Amylase
repression
Em activation
C1 activation
VP1
Exons
1
2 345
non-essential
essential
Fig. 26. Summary on domains of VP1 involved in repression of a-amylase genes, activation of C1 and activation of Em.

82
the domain 103/104 and the C-terminus.
VP1 shows no significant sequence homology to other proteins. Therefore, the function
of domains other than the acidic activation sequence is thus far unknown, it was suggested that
the region from amino acid 208 to 235 of the rice VP1 may form a leucine zipper-like structure
(Hattori et al., 1994). In this region, Leu or lie residues are located at every seventh residue (5
repeats) with an exception for the fourth position (see Fig. 25). However, no severe loss-of-
function phenotype has been observed when this domain was deleted in the 86/85 construct.
86/85 effected full repression of Amy-GUS and ca. 50% of wild-type activity with respect to
activation of Em-GUS. These data do not support an important role of this domain in protein-
protein interactions.
Based on the evidence that VP1 transcriptionally activates C1 and Em (McCarty et al.,
1991; Hattori et al., 1992), nuclear targeting is likely to be a requirement for VP1 function. A
100% conserved putative nuclear localization sequence (NLS), RKKR, exists in the domain from
amino acid 392 to 395 of VP1 (see Fig. 25), as mentioned by Giraudat et al. (1992). Consistent
with these views, deletion of this putative NLS (construct 103/104) fully eliminated activation of
Em-GUS (L. Rosenkrans et al., unpublished results). However, 103/104 retained capacity to
activate C1-Sh-GUS (31% of wild-type VP1), suggesting that the mutant protein is targeted to
the nucleus though possibly with reduced efficiency. Assuming that RKKR is a functional NLS,
these data indicate that VP1 contains two or more NLSs with at least partially redundant function,
a feature not uncommon among nuclear proteins (Raikhel, 1992). The differential effects of
103/104 on activation of Em and C1 may reflect different threshold levels of VP1 protein
required for these activator functions. Hence, the extent of nuclear localization of 103/104 may
be sufficient for partial activation of C1 but not for activation of Em. Alternatively, 103/104 may
serve an additional function in activation of Em. A dual role of a domain in both nuclear
targeting and DNA binding was reported for the regulatory protein 02 (Varagona et al., 1994).
With respect to repression of Amy-GUS by VP1, it is unknown whether nuclear
localization of VP1 is required for function. The construct 103/104 was not affected in repression

83
of Amy-GUS. This may indicate that nuclear targeting of VP1 is not required or that other
domains with redundant function may compensate in function for the deletion in 103/104.
Interestingly, the double-deletion mutant 87/103 lost the ability to exert a dominant negative
effect on Amy-GUS repression by VP1. A failure of the double mutant protein to be targeted to
the nucleus would be consistent with the observed loss of function.
Apart from the activation domain, the biochemical functions of sequences in the VP1
protein remain unclear. However, the deletion analysis demonstrated very clearly that different
domains are involved in the different functions of VP1, thus underlining the multifunctional
nature of this transcription factor.
Interactions Between VP1 and Plant Hormones
One can envision at least three models of how VP1 might function in repressing a-
amylase genes: 1) VP1 might mediate ABA antagonism of GA signalling during seed
development. ABA is known to antagonize GA-action in the regulation of a-amylase genes in
germinating cereal seeds (Jacobsen and Chandler, 1987). Because VP1 is required for ABA-
induced gene expression associated with seed maturation (McCarty et al 1991), it might also be
essential in ABA-mediated repression of a-amylase genes (Fig. 27A). Consequently, the vp1
mutant might allow de-repression of a-amylase genes by failing to respond to ABA present in the
developing seed. 2) VP1 might specifically inhibit the GA-response pathway independently of
ABA (Fig. 27B). 3) VP1 might repress a-amylase genes via a pathway that functions
independently of both GA and ABA (Fig. 27C).
The results presented in this work do not support the first model. ABA was effective in
blocking Amy-GUS expression in vp1 mutant aleurone cells (Fig. 6), indicating that ABA action in
this instance does not depend on the presence of VP1. In combination, VP1 and ABA effects
were roughly additive. This stands in contrast to evidence showing that VP1 is required for ABA-
induced expression of the maize Em gene (McCarty et al., 1991). Thus, there appear to be at

84
GA
| VP1-^-ABA
Amylase
A. VP1 may be required for ABA
function
GA
ABA
Acti
vator
VP1
Amylase
C. VP1 may repress amylase genes
independently of hormones
GA
ABA
VP1
Amylase
B. VP1 may block GA response
pathway independently of ABA
Fig. 27. Alternative models for VP1 function as described in the text

85
least two modes of ABA action in the maize seed, a VP1-dependent pathway and a VP1-
independent pathway. Multiple ABA transduction pathways are also Indicated by interactions
between ABA-insensitive mutants of Arabidopsis (Finkelstein and Somerville, 1990; Finkelstein,
1994). This suggests that ABA modulates the activity of diverse regulatory cascades in the
seed.
The second scenario in which VP1 could specifically block GA signal transduction is
supported by the evidence that over-expression of VP1 in aleurone of imbibed barley half seeds
severely reduced GA-induction of Amy-GUS without affecting the basal activity of the a-amylase
promoter (Fig. 11). This suggests that expression of VP1 in the developing seed may be, at
least in part, responsible for the observed GA-insensitivity of cereal and maize a-amylase genes
prior to seed maturity (Nicholls, 1979; Comford et al., 1986; Garcia-Maya et al 1990; Oishi and
Bewley, 1990). VP1 displayed full repressing activity in slender (sin) mutant barley seeds (Fig.
12) which are constitutive in GA response of the aleurone (Chandler, 1988; Lanahan and Ho,
1988), suggesting that VP1 functions at a point downstream of the Sin gene product.
With respect to the maize seed, the data do not rule out the possibility that VP1 acts
independently of GA as a developmental repressor of a-amylase genes. Although we have
shown that Amy-GUS is GA-inducible in vp1 mutant aleurones early in development (Table 1), it
is not clear that the high constitutive activities found later in development are entirely attributable
to changes in GA concentration. In contrast to the situation of Himalaya barley seed and other
cereal grains, studies of a-amylase regulation in normal and GA-deficient (d5 mutant) genotypes
of maize indicate that a-amylase induction in germinating maize seeds is largely independent of
GA (Harvey and Oaks, 1974). Consistent with these studies, it was found in the present work
that during germination Amy-GUS is constitutively active in the GA1-deficient d1 mutant of
maize. Because Amy-GUS was fully VP1-repressible in aleurones of developing vpf mutant,
germinating wild-type and germinating df-mutant seeds of maize, it is suggested that VP1-
mediated repression is not necessarily restricted to, nor solely defined by, inhibition of the GA
response. Though the significance of GA in the expression of a-amylase genes needs to be

86
investigated further, it is likely that GA, ABA and VP1 are three among several factors that
regulate the activity of constituents required for expression of a-amylase genes in the maize
seed (Fig. 27C).
Expression experiments in developing maize seeds have shown that VP1 and ABA are
likely to act independently in repressing Amy-GUS. Further evidence regarding the relative
positions of VP1 and ABA in the regulatory network allowed the characterization of the dominant
negative effect caused by the deletion mutant 87/88. Co-expression of 87/88 with wild-type VP1
and Amy-GUS has shown that 87/88 inhibits repression of Amy-GUS by VP1 ("dominant
negative effect*, Fig. 22). This suggests that the observed 87/88-mediated activation of Amy-
GUS in the absence of GA (Fig. 20) may be caused by the presence of residual amounts of
endogenous barley VP1 homolog in the wild-type aleurones. Because ABA was incapable of
inhibiting the 87/88-mediated activation of Amy-GUS (Fig. 21), it is suggestive that VP1
functions either downstream of ABA or via a different signalling pathway than ABA. It therefore
will be interesting to map the cis-elements in a-amylase promoters that are responsible for VP1
and ABA action.
The Role of the Embrvo
Although the vp1 mutant phenotype was cell autonomous within the aleurone of vp1-w2
seeds (Fig. 5a,b), de-repression of a-amylase genes was not fully independent of the
physiological state of the embryo: 1) in vp1-m2 seeds, precocious hydrolization of endosperm
reserves in sectors underlying vp1 mutant aleurone was predominantly observed in seeds
carrying a viviparous embryo. 2) Amy-GUS was de-repressed in vp1 mutant aleurone of
concordant vp1 mutant seeds but not in vp1 mutant aleurone of non-concordant seeds exhibiting
a wild-type embryo (Fig. 15). These observations suggest that a viviparous embryo facilitates
expression of a-amylase genes in vp1 mutant aleurone. However, the finding that Amy-GUS is
highly induced in vp1 mutant aleurone of germless seeds (Fig. 16) indicates that a viviparous

87
embryo per se is not required for de-repression of Amy-GUS in vp1 mutant aleurone. Instead, it
rather appears to be the lack of a normal embryo that facilitates expression of Amy-GUS,
suggesting that a wild-type embryo contributes a diffusable signal with inhibitory effect on a-
amylase gene expression in the aleurone. Experimental evidence suggests that developing
embryos are the major source of ABA present in the maturing seed (King, 1979; Jones and
Brenner, 1987). Because Amy-GUS remains sensitive to inhibition by ABA in vp1 mutant
aleurone, ABA produced by the wild-type embryo may be responsible for the observed
repression of Amy-GUS in vp1 mutant aleurone of non-concordant seeds. This is consistent with
the finding that Amy-GUS was de-repressed to a similar extent in developing aleurones of the
ABA-deficient mutant vpS as in aleurones of the germless mutant (data not shown).
In concordant vp1 mutant seeds, ABA concentrations are equal to, or only ca. two-fold
lower than, those present in wild-type seeds (Neill et al., 1986,1987; Paiva and Kriz, 1994). This
suggests that the viviparous embryo also contributes an inductive signal (e.g. GA) that
counteracts the effect of ABA and therefore uncovers de-repression of a-amylase genes in vp1
mutant aleurone. However, the observed strong expression of Amy-GUS in vp1 mutant aleurone
of germless seeds clearly shows that GA production by the embryo is not required for a-amylase
expression in maize aleurone. Supported by the evidence that the endosperm of mature cereal
seeds is not a source of GA (Jacobsen and Chandler, 1987), these data suggest that Amy-GUS
expression is largely independent of GA. Moreover, this interpretation is consistent with other
studies (Harvey and Oaks, 1974) and findings in this work indicating that in mature seeds of GA-
deficient mutants of maize a-amylase genes are expressed at high levels.
Complete repression of Amy-GUS in aleurones of developing seeds was observed only
if the embryo as well as the endosperm were of wild-type genetic constitution (Fig. 16). Lack of
either a normal, arrested embryo or VP1 expression in the aleurone lead to partial de-repression
of Amy-GUS in aleurone cells. This indicates that neither factor expression of VP1 in the
aleurone cells or the presence of a normal embryo is sufficient for total inhibition of a-amylase
genes.

88
VP1 Integrates the Control of Seed Maturation and Germination Programs
It has been shown in this work that VP1 participates in the regulation of two
developmental pathways in the developing maize seed. As a transcriptional activator K is
required for activation of maturation-specific genes (McCarty et al., 1991) and as a repressor it
prevents precocious induction of the normally germination-specific a-amylase genes (data
presented herin). Hence, expression of VP1 specifically during seed development appears to be
involved in ensuring proper ordering of maturation and germination programs. Physically
combining activation and repression function in one protein appears to provide one mechanism
for directly integrating control of mutually exclusive developmental pathways in the plant
embryo. The importance of a tight control of maturation and germination programs for seed
survival is evident in the phenotype of vp1-m2 seeds.
Premature induction of postgerminative development was also reported for the Iec1
(leafy cotyledon 1) mutant of Arabidopsis. In this ABA-sensitive, viviparous mutant, developing
embryos expressed isocitrate lyase genes and a gene encoding a lipid transfer protein at levels
that are normally characteristic of seedlings (West et al., 1994). Double mutant analysis
suggested that the putative Arabidopsis VP1 homolog, ABI3, and LEC1 function in different
pathways (Meinke et al., 1994). Hence, it appears that multiple mechanisms have evolved in
flowering plants to prevent precocious induction of normally germination-specific genes in the
developing embryo.
Thus far, the evidence that VP1 inhibits germination-specific genes is limited to
hydrolase genes in aleurone cells. It is unknown to what extent this repressor activity of VP1 is
also involved in preventing precocious germination of the embryo. Further insight into the
inhibitory role of VP1 during seed development may be provided by stable transformation of vp1
mutant plants with VP1-derivatives that are mutated specifically in the activator or repressor
function.
Cloning of the Vp1 related genes from barley (M. Stoll and D.R. McCarty, unpublished
results), rice (Hatton et al., 1994), Arabidopsis (Giraudat et al., 1992) and tobacco (Phillips and

89
Conrad, 1994) indicates that the Vp1 gene is conserved among flowering plants. Loss of ABI3
function in Arabidopsis causes a similar viviparous phenotype as the vp1 mutation in maize
(Nambara et al., 1992). The functions of ABI3 and VP1, however, diverge in so far that ABI3 is
required for seed dormancy in Arabidopsis while VP1 does not impose seed dormancy in maize.
Because ABI3 mRNA is stored in the dry seed (Parcy et al., 1994), whereas VP1 transcript and
protein are non-detectable in the mature seed (McCarty et al., 1989; Carson, 1992), one can
speculate that dormancy in Arabidopsis may reflect an extended timing of ABI3 expression after
seed maturity rather than a functional difference in the proteins. This view is supported by the
results showing that over-expression of VP1 in aleurone of germinating maize seeds was
effective in repressing Amy-GUS. A role of VP1 in maintaining seed dormancy is also consistent
with the finding that dormancy in barley is correlated with a reduced GA-inducibiiity of a-amyiase
genes in the aleurone (Schuurink et al., 1992; Skadsen, 1993). Hence, it is suggested that VP1
plays a role in integrating the control of seed maturation, dormancy and germination programs.

SUMMARY AND CONCLUSIONS
The Viv¡parous-1 (VP1) transcriptional activator of maize is required for abscisic acid-
induction of maturation-specific genes late in seed development leading to acquisition of
desiccation tolerance and arrest in embryo growth (McCarty et al., 1991). The presented
research extends these findings by showing that VP1, in addition to its transcriptional activation
function, inhibits precocious induction of the germination-specific a-amylase genes in aleurone
cells of the developing seed. Functional analysis of deletion-derivatives of VP1 in a transient
gene expression system indicated that VP1 has a discrete repressor function that is separable
from its transcriptional activation function. It is therefore suggested that physically combining
activator and repressor functions in one protein provides one mechanism for directly integrating
control of the mutually exclusive developmental pathways, seed maturation and seed
germination, in the plant embryo.
90

REFERENCES
Akazawa, T. and S. Miyata. 1982. Biosynthesis and secretion of a-amylase and other
hydrolases in germinating cereal seeds. Essays Biochem. 18:41-78.
Amuti. K.S. and C.J. Pollard. 1977. Soluble carbohydrates of dry and developing seeds.
Phytochem. 16: 529-532.
Anderberg, R.J. and M.K. Walker-Simmons. 1992. Isolation of a wheat cDNA clone for an
abscisic acid-inducible transcript with homology to protein kinases. Proc. Natl. Acad.
Sci. USA 89: 19183-10187.
Anderson, I.C. and D.S. Robertson. 1960. Role of carotenoids in protecting chlorophyll from
photodestruction. Plant Physiol. 35: 531-534.
Armstrong, C., M. Black, J.M. Chapman, H.A. Norman and R. Angold. 1982. The induction of
sensitivity to gibbereliin in aleurone tissue of developing wheat grains. I. The effect of
dehydration. Planta 154: 573-577.
Bflumlein, H., S. Misra, H. Luerpen, K. Kfllle, C. Horstmann, U. Wobus and A.J. Mller. 1994.
The FUS3 gene of Arabidopsis thaliana is a regulator of gene expressin during late
embryogenesis. Plant J. 6: 379-387.
Beckett, J.B. 1993. Locating recessive genes to chromosome arm with B-A translocations. In:
The Maize Handbook, M. Freeling and V. Walbot, eds., pp. 308-314.
Bethke, P.C. and R.L. Jones. 1994. Ca2+-calmodulin modulates ion channel activity in storage
protein vacuoles of barley aleurone cells. Plant Cell 6: 277-285.
Blackman, S.A., R.L. Obendorf and A.C. Leopold. 1992. Maturation proteins and sugars in
desiccation tolerance of developing soybean seeds. Plant Physiol. 100: 225-230.
Blatt, M.R. and G. Thiel. 1993. Hormonal control of ion channel gating. Ann. Rev. Plant
Physiol. Mol. Biol. 44: 543-567.
Brown, A.H.D. and J.V. Jacobsen. 1982. Genetic basis and natural variation of a-amylase
isozymes in barley. Planta 157: 493-501.
Bruni, F. and A.C. Leopold. 1991. Glass transitions in soybean seed: Relevance to anhydrous
biology. Plant Physiol. 96: 660-663.
Bush, D.S., A.K. Biwas and R.L. Jones. 1993. Hormonal regulation of Ca2* transport in the
endomembrane system of the barley aleurone. Planta 189: 507-515.
Carson, C.B. 1992. Characterization of the structure and expression of the mutant alleles of the
viviparous-1 gene of maize. U.S. dissertation, University of Florida.
91

92
Chandler, P.M. 1988. Hormonal regulation of gene expression in the 'slender* mutant of barley
(Hordeum vulgare L). Planta 175:115-120.
Chandler, P.M. and J.V. Jacobsen. 1991. Primer extension studies on a-amylase mRNAs in
barley aleurone. II. Hormonal regulation of expression. Plant Mol. Biol. 16: 637-645.
Chandler, P.M. and M. Robertson. 1994. Gene expression regulated by abscisic acid and its
relation to stress tolerance. Ann. Rev. Plant Physiol. Mol. Biol. 45:113-141.
Chao, S.E. and J.G. Scandalios. 1969. Identification and genetic control of starch-degrading
enzymes in maize endosperm. Biochem. Genet. 3: 537-547.
Chao, S.E. and J.G. Scandalios. 1971. Alpha-amylase of maize: Differential allelic expression
at the Amy-1 gene locus, and some physiological properties of the isozymes. Genetics
69: 47-61.
Chao, S.E. and J.G. Scandalios. 1972. Developmentally dependent expression of tissue
specific amylases in maize. Mol. Gen. Genet. 115:1-9.
Chen, S.M. and E.H. Coe Jr. 1978. Control of anthocyanin synthesis by the C locus in maize.
Biochem. Genet. 15: 333-346.
Chen, Y. and J.S. Burris. 1990. Role of carbohydrates in desiccation tolerance and membrane
behavior in maturing maize seed. Crop Sci. 30: 971-975.
Coe Jr., E.H. and M.G. Neuffer. 1977. The genetics of com. In: Com and Com Improvement,
Ed. 2, G.F. Sprague ed., American Society of Agronomy, (Madison, Wl), pp. 111-223.
Coe Jr., E.H., M.G. Neuffer and D.A Hoisington. 1988. The genetics of com. In: Com and Com
Improvement, G.F. Sprague and J.W. Dudley, eds, American Society of Agronomy
(Madison, Wl), pp. 81-258.
Comai, L. and J.J. Harada. 1990. Transcriptional activities in dry seed nuclei indicate the timing
of the transition from embryogeny to germination. Proc. Natl. Acad. Sci. USA 87: 2671-
2674.
Comford, C.A., M. Black, J.M. Chapman and D.C. Baulcombe. 1986. Expression of a-amylase
and other gibberellin-regulated genes in aleurone tissue of developing wheat grains.
Planta 169: 420-428.
Crispeels, M.J. and J.E. Varner. 1966. Inhibition of gibberellic acid induced formation of a-
amylase by abscision II. Nature 212:1066-1067.
Croker, S.J., P. Hedden, J.R. Lenton and J.L. Stoddart. 1990. Comparison of gibberellins in
normal and slender barley seedlings. Plant Physiol. 94:194-200.
Crowe, J.H., F.A. Hoekstra and L.M. Crowe. 1992. Anhydrobiosis. Ann. Rev. Physiol. 54:579-
599.
Derkx, M.P.M. and C.M. Karssen. 1993. Effects of light and temperature on seed dormancy and
gibberellin-stimulated germination in Arabidopsis thaliana: studies with gibberellin-
deficient and -insensitive mutants. Physiol. Plant. 89: 360-368.

93
Dickinson, C.D., R.P. Evans and N.C. Neilsen. 1988. RY repeats are conserved in the 5-
flanking regions of legume see protein genes. Nucleic Acids Res. 10: 371-378.
Dooner, H.K. 1985. Vtviparous-1 mutation in maize conditions pieiotropic enzyme deficiencies
in the aieurone. Plant Physiol. 77:486-488.
Dooner, H.K., T.P. Robbins and R.A. Jorgensen. 1991. Genetic and developmental control of
anthocyanin biosynthesis. Ann. Rev. Genet. 25:173-199.
Duckham, S.C., R.S.T. Linforth and I.B. Taylor. 1991. Abscisic acid-deficient mutants at the
aba gene locus of Arabidopsis thaliana are impaired in the epoxidation of zeaxanthin.
Plant Cell Envir. 14: 601-606.
Dure, L.S. 1960a. Gross nutritional contributions of maize endosperm and scutellum to
germination growth of maize axis. Plant Physiol. 35: 919-925.
Dure, L.S. 1960b. Site of origin and extent of activity of amylases in maize germination. Plant
Physiol. 35: 925-934.
Dure III, L. 1993. A repeating 11-mer amino acid motif and plant desiccation Plant J. 3:363-
369.
Dure III, L, M. Crouch, J. Harada, T.H.D. Ho, J. Mundy, R. Quatrano, T. Thomas and Z.R. Sung.
1989. Common amino acid sequence domains among the LEA proteins of higher plants.
Plant Mol. Biol. 12: 475-486.
Evans, M., M. Black and J. Chapman. 1975. Induction of hormone sensitivity by dehydration is
one positive role for drying in cereal seed. Nature 258:144-145.
Eyster, W.H. 1931. Vivipary in maize. Genetics 16: 574-590.
Feldmann, K.A. 1991. T-DNA insertion mutagenesis in Arabidopsis: Mutational spectrum. Plant
J. 1:71-82.
Fick, G.N. and C.O. Qualset. 1975. Genetic control of endosperm amylase activity and
gibberellic acid responses in standard-height and short-statured wheats. Proc. Natl.
Acad. Sci. USA 72: 892-895.
Filner, P., and J.E. Varner. 1967. A test for de novo synthesis of enzymes: density labeling with
H2O18 of barley a-amylase induced by gibberellic acid. Proc. Natl. Acad. Sci. USA
58:1520-1526.
Fincher, G.B. 1989. Molecular and cellular biology associated with endosperm mobilization in
germinating cereal grains. Ann. Rev. Plant Physiol. Mol. Biol. 40: 305-346.
Finkelstein, R.R. 1994. Mutations at two new Arabidopsis ABA response loci are similar to the
ab/3 mutations. Plant J. 5: 765-771.
Finkelstein, R.R. and C. Somerville. 1990. Three classes of abscisic acid (ABA)-insensitive
mutations of Arabidopsis define genes that control overlapping subsets of ABA
responses. Plant Physiol. 94:1172-1179.

94
Foster, CA 1977. Slender an accelerated extension growth mutant of barley. Barley Genet.
Newslett. 7:24-27.
Fujioka, S., H. Yamane, C.R. Spray. M. Katsumi, B.O. Phinney, P. Gaskin, J. MacMillan and N.
Takahashi. 1988. The dominant non-gibberellin-responding dwarf mutant (Dfl) of maize
accumulates native gibberellins. Proc. Natl. Acad. Sci. USA 85: 9031-9035.
Galau, G.A., N. Bijaisordat and D.W. Hughes. 1987. Accumulation kinetics of cotton late
embryogenesis-abundant mRNAs and storage protein mRNAs: Coordinate regulation
during embryogenesis and the role of abscisic acid. Dev. Biol. 123:198-212.
Galau, G.A., D.W. Hughes and L. Dure III. 1986. Abscisic acid induction of cloned cotton late
embryogenesis-abundant (Lea) mRNAs. Plant Mol. Biol. 7:155-170.
Galau, G.A., K.S. Jakobsen and D.W. Hughes. 1991. The controls of late dicot embryogenesis
and early germination. Physiol. Plant. 81:280-288.
Gale, M.D. and GA Marshall. 1975. The nature and genetic control of gibberellin insensitivity
in dwarf wheat grain. Heredity 35: 55-65.
Garcia-Maya, M J.M. Chapman and M. Black. 1990. Regulation of a-amylase formation and
gene expression in the developing wheat embryo. Planta 181:296-303.
George, D.W. 1967. High temperature seed dormancy in wheat (Trticum aestivum L.) Crop
Sci. 7: 249-253.
Gilroy, S. and R.L. Jones. 1992. Gibberellic acid and abscisic acid coordinate^ regulate
cytoplasmic calcium and secretory activity in barley aleurone protoplasts. Proc. Natl.
Acad. Sci. USA 89: 3591-3595.
Gilroy, S. and R.L. Jones. 1993. Calmodulin stimulation of unidirectional calcium uptake by the
endoplasmic reticulum of barley aleurone. Planta 190: 289-296.
Giraudat, J., B.M. Hauge, C. Valon, J. Smalle, F. Parcy and H. Goodman. 1992. Isolation of the
Arabidopsis ABI3 gene by positional cloning. Plant Cell 4:1251-1261.
Goldman, S., Y.R. Mawal, I. Taida and R Wu. 1994. Studies of a gibberellin-dependent DNA-
binding protein related to the expression of a rice a-amylase gene. Plant Sci. 99: 75-88.
Goldstein, L.D. and P.H. Jennings. 1978. Amylase enzymes isolated from incubated de-
embryonated maize kernels. New Phytol. 81: 233-242.
Gubler, F., and J.V. Jacobsen. 1992. Gibberellin-responsive elements in the promoter of a
barley high-pl a-amylase gene. Plant Cell 4:1435-1441.
GuiKinan, M.J., W.R. Marcotte, Jr., and R.S. Quattrano. 1990. A plant leucine zipper protein
that recognizes an abscisic acid response element. Science 250: 267-271.
Harberd, N.P. and M. Freeling. 1989. Genetics of dominant gibberellin-insensitive dwarfism in
maize. Genetics 121: 827-838.
Harvey, B.M.R. and A. Oaks. 1974. The role of gibberellic acid in the hydrolysis of endosperm
reserves in Zea mays. Planta 121:67-74.

95
Hattorl, T., T. Terada and S.T. Hamasuna. 1994. Sequence and functional analyses of the rice
gene homologous to the maize Vp1. Plant Mol. Biol. 24: 805-810.
Hattori, T., V. Vasil. L. Rosenkrans, LC. Hannah, D.R. McCarty and I.K. Vasil. 1992. The
VMparous-1 gene and abscisic acid activate the C1 regulatory gene for anthocyanin
biosynthesis during seed maturation in maize. Genes and Dev. 6: 609-618.
Higgins, T.J., J.A. Zwar and J.V. Jacobsen. 1976. Gibberellic acid enhances the level of
translatable mRNA for a-amylase in barley aleurone layers. Nature 260:166-169.
Hilhorst, H.W.M. and C.M. Karssen. 1992. Seed dormancy and germination: the role of abscisic
acid and gibberellins and the importance of hormone mutants. Plant Growth Reg. 11:
225-238.
Hillmer, S., S. Gilroy and R.L. Jones. 1992. Visualizing enzyme secretion from individual barley
(¡Hordeum vulgare) aleurone protoplasts. Plant Physiol. 102: 279-286.
Hoekstra, FA. L.M. Crowe and J.H. Crowe. 1989. Differential desiccation sensitivity of com
and Pennisetum pollen linked to their sucrose contents. Plant Cell Envir. 12: 83-91.
Hoekstra, FA and T. van Roekel. 1988. Desiccation tolerance of Papaver dubium during its
development in the anther Possible role of phospholipid composition and sucrose
content. Plant Physiol. 88: 626-632.
Hooley, R., M.H. Beale and S.J. Smith. 1991. Gibberellin perception at the plasma membrane
of Avena fatua aleurone protoplasts. Planta 183: 274-280.
Hughes, D.W., and G.W. Galau. 1991. Developmental and environmental induction of Lea and
LeaA mRNAs and the postabscission program during embryo culture. Plant Cell 3:605-
618.
Huttly, A.K. and D.C. Baulcombe. 1989. A wheat a-Amy2 promoter is regulated by gibberellin in
transformed oat aleurone protoplasts. EMBO J. 8: 1907-1913.
Huttly, A.K., A.L. Phillips and J.W. Tregear. 1992. Localization of cis elements in the promoter
of a wheat a-amy2 gene. Plant Mol. Biol. 19: 903-911.
Ingle, J. and R.H. Hageman. 1965. Metabolic changes associated with the germination of com.
III. Effects of gibberellic acid on endosperm metabolism. Plant Physiol. ??: 672-675.
Jacobsen, J.V. and L.R. Beach. 1985. Control of transcription of a-amylase and rRNA genes in
barley aleurone protoplasts by gibberellin and abscisic acid. Nature 316:275-277.
Jacobsen, J.V. and P.M. Chandler. 1987. Gibberellin and abscisic acid in germinating cereals.
In Plant Hormones and Their Role in Plant Growth and Development (ed. P.J. Davies),
Dordrecht: Martinus Nijhoff. pp. 164-193.
Jacobsen, J.V. and T.J. Close. 1991. Control of transient expression of chimaeric genes by
gibberellic acid and abscisic acid in protoplasts prepared from mature barley aleurone
layers. Plant Mol. Biol. 16: 713-724.

96
Jacobsen, J.V. and T.J.V. Higgins. 1982. Characterization of the a-amylases synthesized by
aieurone layers of Himalaya barley in response to gibberellic acid. Plant Physiol. 70:
1647-1653.
Jacobsen, J.V., JA Zwar and P.M. Chandler. 1985. Gibberellic-acid-responsive protoplasts
from mature aieurone of Himalaya barley. Planta 163: 430-438.
Jones, R.J. and M.L. Brenner. 1987. Distribution of abscisic acid In maize kernel during grain
filling. Plant Physiol. 83: 905-909.
Jones, R.L. and J.V. Jacobsen. 1991. Regulation of synthesis and transport of secreted proteins
in cereal aieurone. int. Rev. Cytol. 126: 49-88.
Karrer, E.E. and R.R. Rodriguez. 1992. Metabolic regulation of rice a-amylase and sucrose
synthase genes in planta. Plant J. 2: 517-523.
Karssen, C.M., D.L.C. Brinkhorst-van der Swan, A.E. Breekland and M. Koomeef. 1983.
Induction of dormancy during seed development by endogenous abscisic acid: studies
on abscisic acid-deficient genotypes of Arabidopsis thaliana (L.) Heynh. Planta 157:
158-165.
Karssen, C.M., and E. Lapka. 1985. A revision of the hormone balance theory of seed
dormancy: studies on gibberellin and/or abscisic acid-deficient mutants of Arabidopsis
thaliana. In: Plant Growth Substances 1985, M. Bopp ed., Springer Vertag, Berlin, pp.
315-323.
Keith, K., M. Kraml, N.G. Dengler and P. McCourt. 1994. fusca3: a heterochronic mutation
affecting late embryo development in Arabidopsis. Plant Ceil 6: 589-600.
Kermode, A.R. 1990. Regulatory mechanisms involved in the transition from seed development
to germination. CRC Crit. Rev. Plant Sci. 9:155-195
Khursheed, B., and J.C. Rogers. 1988. Barley a-amylase genes: Quantitative comparison of
steady-state mRNA levels from individual members of the two different families
expressed in aieurone cells. J. Biol. Chem. 263:18953-18960.
King, R.W. 1976. Abscisic acid in developing wheat grains and its relationship to grain growth
and maturation. Planta 132: 43-51.
King, R.W. 1979. Abscisic acid synthesis and metabolism in wheat ears. Aust. J. Plant Physiol.
6:99-108.
Koomeef, M., A. Elgersma, C.J. Hanhart, E.P. van Loenen-Martinet, L. van Rijn and JAD.
Zeevaart. 1985. A gibberellin-insensitive mutant of Arabidopsis thaliana. Physiol.
Plant. 65: 33-39.
Koomeef, M. C.J. Hanhart, H.W.M. Hilhorst and C.M. Karssen. 1989. In vivo inhibition of seed
development and reserve protein accumulation in recombinants of abscisic acid
biosynthesis and responsiveness mutants in Arabidopsis thaliana. Plant Physiol. 90:
463-469.
Koomeef, M., M.L. Joma, D.L.C. Brinkhorst-van der Swan and C.M. Karssen. 1982. The
isolation of abscisic acid (ABA) deficient mutants by selection of induced revertants in

97
non-germinating gibberellin sensitive lines of Arabidopsis thaRana (L) Heynh. Theor.
Appl. Genet. 61: 385-393.
Koomeef, M., G. Reuling and C.M. Karssen. 1984. The isolation and characterization of
abscisic acid-insensitive mutants of Arabidopsis thaliana. Physiol. Plant. 61: 377-383.
Koomeef, M. and J.H. Van der Veen. 1980. Induction and analysis of gibberellin sensitive
mutants in Arabidopsis thaliana (L.) Heynh. Theoret. Appl. Genet. 58: 257-263.
Koster, K.L. and A.C. Leopold. 1988. Sugars and desiccation tolerance in seeds. Plant Physiol.
88: 829-832.
Kriz, A.R., M.S. Wallace and R. Paiva. 1990. Globulin gene expression in embryos of maize
viviparous mutants: evidence for regulation of the Glb1 gene by ABA. Plant Physiol. 92:
538-542.
Lanahan, M.B. and T.H.D. Ho. 1988. Slender barley: a constitutive gibberellin-response mutant.
Planta 175:107-114.
Lanahan, M.B., T.H.D. Ho, S.W. Rogers and J.C. Rogers. 1992. A gibberellin response
complex in cereal a-amylase gene promoters. Plant Cell 4:203-211.
Leprince, O., R. Bronchart and R. Deltour. 1990. Changes in starch and soluble sugars in
relation to the acquisition of desiccation tolerance during maturation of Brassica
campestris seed. Plant Cell Envir. 13: 539-546.
Leung, J., M. Bouvier-Durand, P.-C. Morris, D. Guerrier, F. Chefdor and J. Giraudat. 1994.
Arabidopsis ABA response gene ABI1: Features of a calcium-modulated protein
phosphatase. Science 264:1448-1452
Luan, S W. Li, F. Rusnak, S.M. Assman and S.L. Schreiber. 1993. Immunosuppressants
implicate protein phosphatase regulation of K+ channels in guard cells. Proc. Natl.
Acad. Sci. USA 90: 2202-2206.
Ludwig, S.R., L.F. Habera, S.L. Dellaporta and S.R. Wessier. 1989. Lc, a member of the maize
R gene family responsible for tissue-specific anthocyanin production, encodes a protein
similar to transcriptional activators and contains the myc homology region. Proc. Natl.
Acad. Sci. USA 86: 7092-7096.
MacGregor, A.W., B.A. Marchylo and J.E. Kruger. 1988. Multiple a-amyiase components in
germinating cereal grains determined by isoelectric focusing and chromatofocusing.
Cereal Chem. 65: 326-333.
MacLeod, A.M. and G.H. Palmer. 1967. Gibberellin from barley embryos. Nature 216:1342-
1343.
Mangelsdorf, P.C. 1930. The inheritance of dormancy and premature germination in maize.
Genetics 15: 462-494.
Marcotte, Jr., W.R., C.C. Bayley and R.S. Quatrano. 1988. Regulation of a wheat promoter by
abscisic acid in rice protoplasts. Nature 335: 454-457.

98
Marcotte Jr., W.R., S.H. Russell and R.S. Quatrano. 1989. Absdsic add-responsive sequences
from the Em gene of wheat. Plant Cell 1:969-976.
McCarty, D.R. and C.B. Carson. 1991. The molecular genetics of seed maturation in maize.
Physiol. Plant. 81: 267-272.
McCarty, D.R., C.B. Carson, M. Lazar and S.C. Slmonds. 1989b. Transposable element-
induced mutations of the viviparous-1 gene in maize. Dev. Genet. 10: 473-481.
McCarty. D.R., C.B. Carson, P.S. Stinard and D.S. Robertson. 1989a. Molecular analysis of
viviparous-1: an abscisic acid-insensitive mutant of maize. Plant Cell 1: 523-532.
McCarty, D.R., T. Hattori, C.B. Carson, V. Vasil, M. Lazar and I.K. Vasil. 1991. The Viviparous-
1 developmental gene of maize encodes a novel transcriptional activator. Cell 66: 895-
905.
McWha, J.A. 1975. Changes in abscisic acid levels in developing grains of wheat (Triticum
aestivum L.). J. Exp. Bot. 26: 823-833.
Meinke, D.W. 1992. A homeotic mutant of Arabidopsis thaliana with leafy cotyledons. Science
258: 1647-1650.
Meinke, D.W., L.H. Franzmann, T.C. Nickle and E.C. Yeung. 1994. Leafy cotyledon mutants of
Arabidopsis. Plant Cell 6:1049-1064.
Meurs, C., A.S. Basra, C.M. Karssen and L.C. van Loon. 1992. Role of abscisic acid in the
induction of desiccation tolerance in developing seeds of Arabidopsis thaliana. Plant
Physiol. 98: 1484-1493.
Meyer, K M.P. Leube and E. Grill. 1994. A protein phosphatase 2C involved in ABA signal
transduction in Arabidopsis thaliana. Science 264:1452-1455.
Meyerowitz, E.M. 1987. Arabidopsis. Ann. Rev. Genet. 21: 93-111.
Meyerowitz, E.M. 1994. Structure and organization of the Arabidopsis thaliana nuclear genome.
In: Arabidopsis, E.M. Meyerowitz and C.R. Scmerville, eds., Cold Spring Harbor
Laboratory Press, New York.
Miner, J.N. and K.R. Yamamoto. 1991. Regulatory crosstalk at composite response elements.
Trends Biochem. Sci. 16: 423-426.
Mitsunaga, S., T. Tashiro and J. Yamaguchi. 1994. Identification and characterization of
gibberellin-insensitive mutants selected from among dwarf mutants of rice. Theoret.
Appl. Genet. 87: 705-712.
Morris, C.F. J.M. Moffatt, R.G. Sears and G.M. Paulsen. 1989. Seed dormancy and response of
caryopses, embryos, and calli to abscisic acid in wheat. Plant. Physiol. 90: 643-647.
Morris, F.M., R.J. Anderberg, P.J. Goldmark and M.K. Walker-Simmons. 1991. Molecular
cloning and expression of abscisic acid-responsive genes in embryos of dormant wheat
seeds. Plant Physiol. 95: 814-821.
Mozer, T.J. 1980. Control of protein synthesis in barley aleurone layers by the plant hormones
gibberellic acid and abscisic acid. Cell 20: 479-485.

99
Muthukrtshnan, S., B.S. Gill, M. Swegle and G.R. Chandra. 1984. Structural oenes for amylases are located on barley chromosome 1 and 6. J. Biol. Chem. 259:13837-13639.
Nambara, E.. K. Keith, P. McCourt and S. Naito. 1994. Isolation of an internal deletion mutant
of the Arabidopsis thaliana ABI3 gene. Plant Cell Physiol. 35: 509-513.
Nambara, E., S. Naito, and P. McCourt. 1992. A mutant of Arabidopsis which is defective in
seed development and storage protein accumulation is a new ab/3 allele. Plant J. 2:
435-441.
Napier, JA, J.M. Chapman and M. Black. 1989. Calcium-dependent induction of novel
proteins by abscisic acid in wheat aleurone tissue of different developmental stages.
Planta 179:156-164.
Neill, S.J., R. Horgan and A.D. Parcy. 1986. The carotenoid and abscisic acid content of
viviparous kernels and seedlings of Zea mays L. Planta 169: 87-96.
Neill S.J., R. Horgan and A.F. Rees. 1987. Seed development and vivipary in Zea mays L.
Planta 171: 358-364.
Neuhaus, G., C. Bowler, R. Kern and N.-H. Chua. 1993. Calcium/calmodulin-dependent and -
independent phytochrome signal transduction pathways. Cell 73: 937-952.
Nicholls, P.B. 1979. Induction of sensitivity to gibberellic acid in developing wheat caryopses:
effect of rate of desiccation. Aust. J. Plant Physiol. 6:229-240.
Nicholls, P.B., A.W. MacGregor and BA Marchylo. 1986. Production of a-amyiase isozymes in
barley caryopses in the absence of embryo and exogenous gibberellic acid. Aust. J.
Plant Physiol. 13: 239-247.
Nolan, R.C. and T.H.D. Ho. 1988. Hormonal regulation of gene expression in barley aleurone
layers. Induction and suppression of specific genes. Planta 174: 551-560.
Oishi, M.Y. and J.D. Bewley. 1990. Distinction between the responses of developing maize
kernels to flouridone and desiccation in relation to germinability, a-amylase activity, and
abscisic acid content. Plant Physiol. 94: 592-598.
Oliver, M.J., J. Armstrong and J.D. Bewley. 1993. Desiccation and the control of expression of
p-phaseolin in transgenic tobacco seeds. J. Exp. Bot. 44:1239-1244.
Ooms, J.J.J., K.M. Leon-Kloosterziel, D. Bartels, M. Koomeef and C.M. Karssen. 1993.
Acquisition of desiccation tolerance and longevity in seeds of Arabidopsis thaliana. Plant
Physiol. 102:1185-1191.
Paiva, R. and A.L. Kriz. 1994. Effect of abscisic acid on embryo-specific gene expression
during normal and precocious germination in normal and viviparous maize (Zea mays)
embryos. Planta 192: 332-339.
Paleg, L.G. 1960a. Physiological effects of gibberellic acid: I. On carbohydrate metabolism and
amylase activity of barley endosperm. Plant Physiol. 35: 293-299.
Paleg, L.G. 1960b. Physiological effects of gibberellic acid: II. On starch hydrolyzing enzymes
of barley endosperm. Plant Physiol. 35: 902-906.

100
Parcy, F.t C. Valon, M. Raynal, P. Gaubler-Comella. M. Delseny, and J. Giraudat. 1994.
Regulation of gene expression programs during Arabidopsis seed development: roles of
the ABJ3 locus and of endogenous abscisic add. Plant Cell 6:1567-1582.
Paz-Ares. J., D. Ghosal and H. Saedler. 1990. Molecular analysis of the C1-I allele from Zea
mays', a dominant mutant of the regulatory C1 locus. EMBO J. 9: 315-321.
Ptashne, M. 1988. How eukaryotic transcriptional activators work. Nature 335: 683-689.
Quatrano, R.S. 1987. The role of hormones during seed development. In: Plant Hormones and
Their Role in Plant Growth and Development, P.J. Davies ed., Dordrecht: Martinus
Nijhoff,. pp. 494-514.
Radley, M. 1967. Site of production of gibberellin-like substances in germinating barley
embryos. Planta 75:164-171.
Radley, M. 1976. The development of wheat grain in relation to endogenous growth
substances. J. Exp. Bot. 27:1009-1021.
Raikhel, N. 1992. Nudear targeting in plants. Plant Physiol. 100:1627-1632.
Ranki. H and T. Sopanen. 1984. Secretion of a-amylase by the aleurone layer and the
scutellum of germinating barley grain. Plant Physiol. 75: 710-715.
Ried, J.L. and M.K. Walker-Simmons. 1990. Synthesis of absdsic acid-responsive, heat-stable
proteins in embryonic axes of dormant wheat grain. Plant Physiol. 93:662-667.
Ried, J.L. and M.K. Walker-Simmons. 1993. Group 3 late embryogenesis abundant proteins in
desiccation-tolerant seedlings of wheat (Trticum aestivum L). Plant Physiol. 102:125-
131.
Roberts, D.M. and A.C. Harmon. 1992. Calcium-modulated proteins: targets of intracellular
caldum signals in higher plants. Ann. Rev. Plant Physiol. Mol. Biol. 43: 375-414.
Robertson, D.S. 1952. The genotype of the endosperm and embryo as it influences vivipary in
maize. Proc. Natl. Acad. Sd. USA 38: 580-583.
Robertson, D.S. 1955. The genetics of vivipary in maize. Genetics 40:745-760.
Robertson, D.S. 1978. Charaderization of a mutator system in maize. Mutation Res. 51: 21-28.
Robichaud, C.S. and I.M. Sussex. 1986. The response of viviparous-1 and wild-type embryos
of Zea mays to culture in the presence of abscisic add. J. Plant Physiol. 126: 235-242.
Robichaud, C.S. and I.M. Sussex. 1987. The uptake and metabolism of [2-14C}-ABA by
exdsed wild type and viviparous-1 embryos of Zea mays. L. J. Plant Physiol. 130:1 SI-
188.
Robichaud, C.S., J. Wong and I.M. Sussex. 1980. Control and growth of viviparous embryo
mutants of maize by absdsic add. Dev. Genet 1: 325-330.
Rogers, J.C. 1985. Two barley a-amylase gene families are regulated differently in aleurone
cells. J. Biol. Chem. 260: 3731-3738.

101
Rogers, J.C., M.B. Lanahan and S.W. Rogers. 1894. The cte-acting gibberellin response
complex In hlgh-pl a-amylase gene promoters. Plant Physiol. 105:151-158.
Rogers, J.C. and S.W. Rogers. 1992. Definition and functional implications of gibberellin and
absdsic acid exacting hormone response complexes. Plant Cell 4:1443-1451.
Roman, H. 1947. Mitotic nondisjunction in the case of interchanges involving the B-type
chromosomes in maize. Genetics 32: 391-409.
Roth, BA., SA. Goff, T.M. Klein and M.E. Fromm. 1991. Cl- and R-dependent expression of
the maize Bz1 gene requires sequences with homology to mammalian myb and myc
binding sites. Plant Cell 3: 317-325.
Rushton, P.J., R. Hooley and C.M. Lazarus. 1992. Aleurone nuclear proteins bind to similar
elements in the promoter regions of two gibberellin-regulated a-amylase genes. Plant
Mol. Biol. 19: 891-901.
Schopfer, P. and C. Plachy. 1984. Control of seed germination by abscisic acid. II. Effect on
embryo water uptake in Brassica napus L. Plant Physiol. 76:155-160.
Schuurink, R.C., N.J.A. Sedee and M. Wang. 1992a. Dormancy of the barley grain is correlated
with gibberellic acid responsiveness of the isolated aleurone layer. Plant Physiol. 100:
1834-1839.
Schuurink, R.C., J.M.M. van Beckum and F. Heidekamp. 1992b. Modulation of grain dormancy
in barley by variation of plant growth conditions. Hereditas 117:137-143.
Skadsen, R.W. 1993. Aleurones from a barley with low a-amylase activity become highly
responsive to gibberellin when detached from the starchy endosperm. Plant Physiol.
102: 195-203.
Skriver, K. and J. Mundy. 1990. Gene expression in response to abscisic acid and osmotic
stress. Plant Cell 2: 503-512.
Skriver, K., F.L. Olsen, J.C. Rogers and J. Mundy. 1991. Cis-acting DNA elements responsive
to gibberellin and its antagonist abscisic acid. Proc. Natl. Acad. Sci. USA 88: 7266-
7270.
Still, D.W., D.A. Kovach and K.J. Bradford. 1994. Development of desiccation tolerance during
embryogenesis in rce (Oryza sativa) and Wild Rice (Zizania palustris). Plant Physiol.
104: 431-438.
Stoddart, J.L. 1984. Growth and gibberellin-A1 metabolism in normal and gibberellin-insensitive
(Rht3) wheat (Jriticum aestivum L.) seedlings. Planta 161: 432-438.
Sutliff, T.D., N. Huang, J.C. Littsand R.L. Rodriguez. 1991. Characterization of an a-amylase
multigene duster in rice. Plant Mol. Biol. 16: 579-591.
Sutliff, T.D., M.B. Lanahan and T.H.D. Ho. 1993. Gibberellin treatment stimulates nudear
fador binding to the gibberellin response complex in a barley a-amylase promoter.
Plant Cell 5:1681-1692.

102
Taylor, M.G. and I.K. Vasil. 1991. Histology of, and physical factors affecting transient GUS
expression in pearl millet (Pennisetum glaucum (L.) R. Br.) embryos following
microprojectile bombardment. Plant Cell Rep. 10:120-125.
Tsai, M. and B.W. OMalley. 1994. Molecular mechanisms of action of sterokl/thyroid receptor
superfamily members. Ann. Rev. Biochem. 03: 451-486.
Varner, J.E. and G.R. Chandra. 1964. Hormonal control of enzyme synthesis in barley
endosperm. Proc. Natl. Acad. Sci. USA 52:100-106.
Varagona, M.J. and N. Raikhel. 1994. The basic domain in the bZIP regulatory protein
Opaque2 serves two independent functions: DNA binding and nuclear localization.
Plant J. 5: 207-214.
Vasil, V., M. Clancy, R.J. Fed, I.K. Vasil and L.C. Hannah. 1989. Increased gene expression by
the first intron of the maize sh1 locus in grass species. Plant Physiol. 91:1575-1579.
Walker-Simmons, M. 1987. ABA levels and sensitivity in developing wheat embryos of
sprouting resistant and susceptible cultivare. Plant Physiol. 84: 61-66.
Walker-Simmons, M. 1988. Enhancement of ABA responsiveness in wheat embryos by high
temperature. Plant Cell Envir. 11: 769-775.
Warner, DA, M.J. Grove and CA Knutson. 1991. Isolation and characterization of a-
amylases from endosperm of germinating maize. Cereal Chem. 68: 383-390.
Warner, D.A. and CA Knutson. 1991. Isolation of a-amylases and other starch degrading
enzymes from endosperm of germinating maize. Plant Science 78:143-150.
West, M.A., K.M.Yee, J. Danao, J.L. Zimmerman, R.L. Fischer, R.B. Goldberg and J.J. Harada.
1994. Leafy cotyledon 1 is an essential regulator of late embryogenesis and cotyledon
identity in Arabidopsis. Plant Cell 6:1731-1745.
Wheeler, A.W. 1972. Changes in growth-substance contents during growth of wheat grains.
Ann. Appl. Biol. 72: 327-334.
Whittier, R.F., D.A. Dean and J.C. Rogers. 1987. Sequence analysis of alpha-amylase and thiol
protease genes that are hormonally regulated in barley aleurone cells.. Nuci. Acids Res.
15: 2515-2535.
Williams, B.A. and A. Tsang. 1994. Analysis of multiple classes of abscisic acid-responsive
genes during embryogenesis in Zea mays. Dev. Genet. 15: 415-424.
Williamson, J.D. and R.S. Quattrano. 1988. ABA regulation of two classes of embryo-specific
sequences in mature wheat embryos. Plant Physiol. 86: 208-215.
Wilson, G.F., A.M. Rhodes and A.M. Dickinson. 1973. Some physiological effects of viviparous
genes vp1 and vpS on developing maize kernels. Plant Physiol. 52: 350-356.
Xu, N., K.M. Coulter and J.D. Bewley. 1990. Abscisic acid and osmoticum prevent germination
of developing alfalfa embryos, but only osmoticum maintains the synthesis of
developmental proteins. Planta 182: 382-390.

103
Yomo, H. 1960. Studies on the a-amylase activating substance. IV. On the amylase activating
action of gibbereilin. Hakko Kyokaishi. 18:600-602.
Yu. S.M., Y.H. Kuo. G. Sheu, Y.J. Sheu and LF. Uu. 1991. Metabolic de-repression of a-
amylase gene expression in suspension-cultered cells of rice. J. Biol. Chem. 266:
21131-21137.
Zeevaart, JAD. and RA Creelman. 1988. Metabolism and physiology of abscisic acid. Ann.
Rev. Plant Physiol. Mol. Biol. 39: 439-73.
Zwar, JA, and R. Hooley. 1986. Hormonal regulation of a-amylase gene transcription in wild
oat (Avena fatua L.) aleurone protoplasts. Plant Physiol. 80: 459-463.

BIOGRAPHICAL SKETCH
Ute Hoecker was bom on August 29, 1964, in Aachen, Germany. She completed high
school in Bonn, Germany, in 1984. She enrolled in the Friedrich Wilhelm University of Bonn and
received a 'Vordiplom' (B.S.) in agricultural sciences in 1986. Following a year of practical
training on a laboratory farm in Bonn, she transferred to the University of Hohenheim in
Stuttgart-Hohenheim, Germany, and earned a "Diplom" (M.S.) in agricultural sciences with
specialization in plant breeding and population genetics in 1990. She moved to Gainesville,
Florida, in August, 1990, to begin her doctoral studies.
104

I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fullyNadequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
Indra K. Vasil, Chair
Graduate Research' Professor of Plant
Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
Donald R. McCarty,/Cochair //
Associate Professor of Horticultural Science and
Plant Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
K&ren E'. Koch 7 ~
Professor of Horticultural Science and Plant
Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
William B. Gurley /
Associate Professor of Microbiology and Cell
Science, and Plant Molecular and Cellular
Biology
I certify that I have read this study and that in
standards of scholarly presentation and is fully adequate,
for the degree of Doctor of Philosophy.
opinion it conforms to acceptable
n scope and quality, as a dissertation
Wl
Nigel G. Richards
Assistant Professo
Chemistry

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and
to the Graduate School and was accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
August 1995
Dean, College of Agriculture
Dean, Graduate School



101
Rogers, J.C., M.B. Lanahan and S.W. Rogers. 1894. The cte-acting gibberellin response
complex In hlgh-pl a-amylase gene promoters. Plant Physiol. 105:151-158.
Rogers, J.C. and S.W. Rogers. 1992. Definition and functional implications of gibberellin and
absdsic acid exacting hormone response complexes. Plant Cell 4:1443-1451.
Roman, H. 1947. Mitotic nondisjunction in the case of interchanges involving the B-type
chromosomes in maize. Genetics 32: 391-409.
Roth, BA., SA. Goff, T.M. Klein and M.E. Fromm. 1991. Cl- and R-dependent expression of
the maize Bz1 gene requires sequences with homology to mammalian myb and myc
binding sites. Plant Cell 3: 317-325.
Rushton, P.J., R. Hooley and C.M. Lazarus. 1992. Aleurone nuclear proteins bind to similar
elements in the promoter regions of two gibberellin-regulated a-amylase genes. Plant
Mol. Biol. 19: 891-901.
Schopfer, P. and C. Plachy. 1984. Control of seed germination by abscisic acid. II. Effect on
embryo water uptake in Brassica napus L. Plant Physiol. 76:155-160.
Schuurink, R.C., N.J.A. Sedee and M. Wang. 1992a. Dormancy of the barley grain is correlated
with gibberellic acid responsiveness of the isolated aleurone layer. Plant Physiol. 100:
1834-1839.
Schuurink, R.C., J.M.M. van Beckum and F. Heidekamp. 1992b. Modulation of grain dormancy
in barley by variation of plant growth conditions. Hereditas 117:137-143.
Skadsen, R.W. 1993. Aleurones from a barley with low a-amylase activity become highly
responsive to gibberellin when detached from the starchy endosperm. Plant Physiol.
102: 195-203.
Skriver, K. and J. Mundy. 1990. Gene expression in response to abscisic acid and osmotic
stress. Plant Cell 2: 503-512.
Skriver, K., F.L. Olsen, J.C. Rogers and J. Mundy. 1991. Cis-acting DNA elements responsive
to gibberellin and its antagonist abscisic acid. Proc. Natl. Acad. Sci. USA 88: 7266-
7270.
Still, D.W., D.A. Kovach and K.J. Bradford. 1994. Development of desiccation tolerance during
embryogenesis in rce (Oryza sativa) and Wild Rice (Zizania palustris). Plant Physiol.
104: 431-438.
Stoddart, J.L. 1984. Growth and gibberellin-A1 metabolism in normal and gibberellin-insensitive
(Rht3) wheat (Jriticum aestivum L.) seedlings. Planta 161: 432-438.
Sutliff, T.D., N. Huang, J.C. Littsand R.L. Rodriguez. 1991. Characterization of an a-amylase
multigene duster in rice. Plant Mol. Biol. 16: 579-591.
Sutliff, T.D., M.B. Lanahan and T.H.D. Ho. 1993. Gibberellin treatment stimulates nudear
fador binding to the gibberellin response complex in a barley a-amylase promoter.
Plant Cell 5:1681-1692.


76
The strongly positively charged, 15-amino acid-domain (RED domain), deleted in 85/87,
is highly conserved (ca. 80%) among maize, barley and rice genes (Fig. 25). Between maize
and barley, 11 out of 15 amino acids are identical and one amino add constitutes a conservative
substitution (R to K) also found in rice. The high degree of sequence conservation is consistent
with the finding that this region is of functional importance for repression of Amy-GUS.
Interestingly, deletion of the RED domain exhibited a differential effect in maize and barley
aleurone. While 85/87 was incapable of inhibiting Amy-GUS in barley aleurone, it retained ca.
50% of wild-type VP1 activity in maize aleurone cells (Fig. 18). In maize, deletion of additional
sequences 5' to the 85/87 deletion (construct 86/87) was necessary to eliminate repressor
function (Fig. 18). This indicates that in the barley cell the RED domain is absolutely essential
for proper execution of the repression mechanism, whereas in the more concordant system of
the maize aleurone cell other possibly less conserved regions in the VP1 protein partially
compensate in function for the RED domain.
The construct 87/88 that deletes a large but poorly conserved stretch of VP1 was
incapable of repressing Amy-GUS in maize or barley cells. To further analyze this domain of
VP1, four smaller deletions within this region were constructed and tested (Fig. 18). The three
deletion constructs 87/92 (deleting 29 aa), 92/181 (deleting 26 aa) and 93/95 (deleting 40 aa) of
which each lacks a short stretch of conserved sequence were not severely impaired in repressing
Amy-GUS. In contrast, the construct 196/88 deleting 48 amino acids at the C-terminal end of the
87/88 deletion was incapable of repression in barley (no data for this construct were obtained in
maize). These results allow at least two interpretations: 1) 196/88 deletes a domain essential for
repression, while the sequences located between the deletion points 87 and 95 are not required
for repressor function. 2) The partially conserved sequences between 87 and 95 are of
redundant function. Therefore, deletion of two or more conserved blocks may be necessary to
lose repression of Amy-GUS. In this interpretation, 196/88 may delete an additional important
domain or, alternatively, affect proper spacing between further N-termnal and C-terminal
sequences.


84
GA
| VP1-^-ABA
Amylase
A. VP1 may be required for ABA
function
GA
ABA
Acti
vator
VP1
Amylase
C. VP1 may repress amylase genes
independently of hormones
GA
ABA
VP1
Amylase
B. VP1 may block GA response
pathway independently of ABA
Fig. 27. Alternative models for VP1 function as described in the text


71
Emms no Hormones
¡ABA
Fig. 21. ABA does not inhibit 87/88-mediated activation of Amy-GUS found in the absence of
GA. Aleurones of germinating barley seeds were bombarded as described in Fig. 20 and then
cultured in no hormones or 10*5 M ABA.
il.i
CAT 07/88 VP1 VP1+
87/88
Fig. 22. The VP1 deletion mutant 87/88 displays a dominant negative effect on VP1-mediated
repression of Amy-GUS. Aleurones of germinating barley seeds were bombarded with 6 pg of
Amy-GUS, 5 pg of Ubi-LUC and either 25 pg of CAT, 20 pg of 87/88, 5 pg of VP1 or both 20 pg
of 87/88 and 5 pg of VP1. To all mixtures, CAT plasmid was added to obtain a total amount of
36 pg of plasmid DNA. After bombardment, endosperms were cultured in 10-6 M GAg. Data
represent mean ( S.E.M.) of five replicates.


87
Re. Amy-GUS / LUC
Effector Construct Maize Barley
VP1-WT
86/87
86/85
85/87
87/88
87/92
92/181
93/95
92/88
196/88
Control (no VP1)
100
100
4-12
18-33
133 (29)
87 (*8)
7 (2)
13(*3)
45 (*)
99(*21)
85-162
90-231
9(*D
39 (1)
6 (2)
29 (*7)
14(*2)
33 (*)
27-100
125 (17)
?
84 (*s)
Fig. 18. Deletion analysis of the VP1 protein: Part I. For Materials and Methods see Fig. 19.
Black boxes indicate sequence homology to the VP1 homolog from barley.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZO2P9C2K_WRGOCV INGEST_TIME 2015-03-31T17:58:06Z PACKAGE AA00029819_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


49
wild-type maize seeds. Table 1 shows that during mid-late development Amy-GUS was not
expressed In developing wild-type aleurone, even in the presence of exogenous GA. In contrast,
GA-induction of Amy-GUS was detected in vp1 mutant aleurones as early as 20 days after
pollination PAP). These data indicate that in developing wild-type aleurone tissue a-amylase
genes are insensitive to GA while in vp1 mutant aleurone cells a-amylase expression is
transcriptionally de-re pressed.
Amy-GUS expression in vp1 mutant aleurone was found to be under developmental and
hormonal control. Prior to approx. 18 DAP, Amy-GUS was inactive in GA-treated as well as
untreated aleurone, indicating that early in seed development the aleurone is unresponsive to
GA even in the absence of VP1 protein. At 20 DAP, Amy-GUS was induced by exogenous GA,
whereas its activity remained low in untreated aleurone (Fig. 6). Late in seed development (24
DAP), Amy-GUS was constitutively active in the absence of GA, indicating a greatly reduced
dependence on exogenous hormone. Two observations, however, indicate that Amy-GUS
expression was not fully constitutive at this stage: 1) GA treatment significantly enhanced AMY-
GUS expression (as much as 3-fold over that of untreated aleurones) in some, but not all
experiments (Table 1, Fig. 6). 2) GA treated aleurones typically exhibited less quantitative
variation in Amy-GUS expression than non-treated sibling aleurones. The latter effect suggests
that developmental or spatial variation affecting endogenous hormone concentrations within the
ear or seed might contribute to the large variation observed in the absence of exogenous GA.
The differential expression of Amy-GUS in developing vp1 mutant and wild-type
aleurone cells confirms a role of VP1 in the repression of a-amylase genes during seed
development. In order to test whether expression of recombinant VP1 could evoke inhibition of
a-amylase transcription in vp1 mutant aleurones, aleurones were bombarded with a mixture of
Amy-GUS and 35S-Sh-VP1 plasmids. Co-expression of VP1 strongly inhibited Amy-GUS
expression in vp1 mutant aleurone in the presence as well as absence of exogenous GA (Fig. 6),
indicating that recombinant VP1 effectively restored the wild-type phenotype. We can rule out
the possibility that over-expression of VP1 causes non-specific squelching of general


42
Reporter Plasmids:
High-pl Amy-GUS
(JR254)
Amy 6-4 Intron 1
GUS
Amy 6-4 3^
Amy 6-4 promoter
Low-pl Amy-GUS2
(JR303)
Intron
GUS
Amy 32b 3"
Amy 32b promoter
Effector Plasmids:
35S-Sh-VP1
35S-Sh-CAT
(No-Vp1 Control)
Sh1 Intron
Vp1 cDNA
CaMV 35S promoter
Sh1 Intron
NOS 3
CaMV 35S promoter
NOS 3*

Internal Standard:
Ubiquitin-Luciferase
(pAHC18)
Intron
Ubiquitin promoter
Fig. 3. Plasmid constructs.


61
Role of the Embfvo in Repression ofa-Amylase Genes in the Aleurone
The differential response of mutant and wild-type aleurone tissue in developing vp1-m2
kernels (Fig. 5a,b) did not appear to be fully independent of the state of the embryo. Embryos of
vp1-m2 seed are frequently non-viviparous and survive desiccation. It was observed that
mosaic aleurones that were associated with non-viviparous embryos very rarely exhibited
precocious endosperm remobilization while those with viviparous embryos typically did.
Moreover, aleurone near the germinal face and crown of the kernel was more strongly affected
than aleurone on the abgerminal face. This phenotype suggests that signalling from the embryo
as well as responsiveness of the aleurone cells contribute to the softening response.
In order to assess the impact of the physiological state of the embryo on a-amylase
expression in aleurone cells, TB3La translocation stocks were used which allow the generation of
vp1 non-concordant seeds with embryo and endosperm of different genetic constitution (for a
brief description of the system see Materials and Methods). As a result, ears developed that
segregated four genotypes: 1) seeds with a vp1 mutant embryo and a wild-type endosperm, 2)
seeds with a wild-type embryo and a vp1 mutant endosperm, 3) concordant vp1 mutant seeds
and 4) concordant wild-type seeds. Aleurones of these four genotype combinations were
bombarded with Amy-GUS and postbombardment cultured in the absence of added hormones.
Amy-GUS was only expressed in aleurones of concordant vp1 mutant kernels and not in any of
the other three genotype combinations (Fig. 15). These data are consistent with the observation
made in vp1-m2 seeds that endosperm tissue underlying vp1 mutant aleurone was remobilized
predominantly in viviparous seeds. Thus, a-amylase genes appear to be de-repressed in
aleurone cells that lack functional VP1 predominantly if the embryo is also viviparous. In some
experiments in which kernels very late in development were used, both vp1 non-concordant
genotypes displayed significant Amy-GUS activities, while no Amy-GUS expression was
detected in aleurones of concordant wild-type seeds (data not shown). This indicates that partial
de-repression of Amy-GUS in aleurone cells is facilitated if either the embryo is viviparous or the
aleurone cells lack functional VP1.


22
Hormonal Regulation
Extensive studies in barley and several investigations with other cereal species have
shown that synthesis of cereal a-amylases is induced by GA and antagonistically inhibited by
ABA (for review see: Jacobsen and Chandler, 1987; Jones and Jacobsen, 1991; Fincher, 1989).
The first reports on GA-induced a-amylase activity in germinating barley seeds appeared in 1960
(Paleg, 1960a; b; Yomo, 1960). The action of ABA as an antagonist of GA was discovered in
1966 (Chrispeels and Varner, 1966).
Application of transcription and translation inhibitors indicated that the GA-induced
appearance of a-amylase activity was due to de novo synthesis of the enzyme (Varner and
Chandra, 1964; Filner and Varner, 1967). Ultimate proof for an effect of GA and ABA on a-
amylase synthesis came from the demonstration that GA treatment drastically increased the
amount of in vitro translatable a-amylase mRNA (Higgins et al., 1976), while simultaneous
application of ABA blocked this effect (Mozer, 1980). Run-on transcription experiments provided
evidence that GA and ABA regulate the transcription of a-amylase genes (Jacobsen and Beach,
1985; Zwar and Hooley, 1986). Eventually, the development of transient gene expression
technology has provided an additional tool to elucidate GA and ABA action. It was shown that
transient expression of a wheat a-amylase promoter-GUS reporter gene fusion construct in oat
aleurone protoplasts was regulated in the same manner as the endogenous genes (Huttley and
Baulcombe, 1989).
Relatively few studies have addressed hormonal regulation of maize a-amylase. Ingle
and Hageman (1965) reported a stimulating effect of GA on catabolism of carbohydrates in
excised endosperms. In a different study (Harvey and Oaks, 1974), exogenous GA applied to
excised endosperms further increased (3-fold) total amylase activity in germinating seeds of a
GA-deficient mutant (dS), but not in wild-type seeds. In contrast, culture of wild-type
endosperms in ABA strongly reduced amylase activity. Even though no molecular data are
available, these results indicate that maize a-amylase genes are probably regulated by GA and


78
88 104 **** 103
Z.m. ASTGS YPMPQPFSPGFGCQYAGAGAGHLSVAPQRMAGVEASATKEAWOCRMARQRRLSCL 405
i mi ii ii mi i 111111 n 11 nnn111
H.V. ATSAGTQTYPMPPP-GPV-PQPFAAPGFAGQFPQRM EPAATREARKKRMARQRRLSCL 381
| || I I lili I I I I I I I I I I I I I I I I
O 3. SSTSS YTMPSWPPPrTAGFPGQYSGGHAMCSPRIAGVEPSSTKEARKKRMARQRRLSCL 424
| | I I II I II I I I I I I I I I I I I I I
A. t. SQYNQFG DPTGFNGYNMNPYQYPYVPAGQMRDQRLLRLCSSATKEARKKRMARQRRLL 470
Z.m. QQQRSQQLSLGQIQTSVHLQEPSPRSTHSGPVTPSAGGWGFWSPSSQ QQVQNPLS-K5N 463
i in i nn n n n ii
H.v. Q???????????IQTGGFPQQPS PRAAHSAFVWG7HWS PPAVQAQPHGQLMIQVPNPLSTKSN 444
I I I II I i I I I I I I I I I I II II I I
O.s. QQQRSQQLNLSQIHISGHPQEPSPRAAHSAPVTPSSAGCRSWGIWPPAAQIIQNPLSNKPN 485
I I III II II
A. t. SHHHRHNNNNNNNNNNQQNQTQIGETCAAVAPQLN PVATTATGGTWMYWPN 521
Z.m. SSRAPPSSLEAAAAAPQTKPAP-AGARQDDIHHRLAAASDKRQGAKADKNLRFLLQKVLKQSD 526
III III I lilil I I I I I I I I I I I I I I
H.v. SSRQKQQKPSPDAAAR-PPSGGGASQQRQGQ AAASDKQRQQ K? LRFLLQKVLKQS D 499
I I I III I I I I II II I II I II II
O.s. PPPATSKQPKPSPEKPKPKPQAAATAGAESLQRSTASEKRQ-AKTDKNLRFL1.QKVLKQSD 545
IIIII I III I I I I II I II II II I I II I
A. t. VPAV PPQLP PVMETQLPTMDRAGSASAMP RQQWP DRRQGWKPEKNLRFLLQKVLKQS D 581
/McW 101
Z.m. VGSLGRIVLPKKEAEVHLPELKTRDGISIPMEDIGTSRVWNMRYRFWPNNKSRMYLLENTGEF 589
II II IIII II II II I II II II II II I I I I II I I II IIII I II II II II II II II II
H.v. VGTLGRIVLPKKEAETHLPELKTGDGISIPIEDIGTSQVWSMRYRFRPNNKSRMYLLENTGEF 562
II II II IIII II II II II I II II III I II II II II IIII II II II II II II II II I
O.s. VGSLGRIVLPKKEAEVHLPELKTRDGVSIPMEDIGTSQVWNMRYRFWPNNKSRMYLLENTGDF 600
II IIII I I II I I I I I I II I III I II IIII I II I IIIIII I I II I I I I II I II I II
A.t. VGNLGRIVLPKKEAETHLPELEARDGISLAMEDIGTSRVWNMRYRFWPNNKSRMYLLENTGDF 644
100
Z.m. VRSNELQEGDFIVTYSDVKSGKYLIRGVKVR-PPAQEQGSGSSG-GGKH RP-LC- 640
II II I IIII II II II II II II II I III I I II I
H.V. VRSNE ? ? ? 7DFIVLYSDVKSQKYLIRGYKVRAAQELASTRWQSREGGA? ?V LAQ 616
II I II II II II I II II II II II I II II I I
O.s. VRSNELQEGDFIVIYSDIKSGKYLIRGVKVRR-AAQEQGN SSGAVGKHKHGSPEKPGVSS 667
I I II II II II I I II I I II I I II II I I I III
A.t. VKTNGLQEGDFIVIYSDVKCGKYLIRGVKVRQPSGQKPEA-PPSSAATKR 693
Z.m. P AGP E RAAAAGAP E DAWDGV SGACKGRSPEGVRRVRQQGAGAMSQMAVSI
H.v. TAAD
I I I II I II II II I I II II I II II I
O.s. NTKAAGAEDGTGGDDSAEAAAAAAAGKADGGGCKGKSPHGVRRSRQEAAAAAS MSQMAVSI
III I
A.t. QNKSQRNINNNSPSA-NVWA SPTSQTVK
691
620
728
720
Fig. 25. Alignment of the amino add sequences of VP1 from maize (Z.m.), HWP1 from barley
(H.v.), OSVP1 from rice (O.s.) and ABI3 from Arabidopsis (A.t.). Identical amino adds are
shown by vertical lines. Conserved regions among the four spedes are boxed. Italic numbers
above sequence show site of deletion points. Underlined amino adds indicate location of
putative leudne zipper. Asterisks indicate putative NLS.


s
o
EHH3+35S-Sh-CAT
mm+ 35S-Sh-vpi
Fig. 12. VP1 function does not require the Slender gene product. Aleurones of imbibed slender
(sin) mutant and wild-type barley half seeds harvested from a plant segregating for the sin
mutation were bombarded with 5 ng of Amy-GUS, 5 pg of Ubi-LUC, and 5 pg of 35S-Sh-VP1 or
35S-Sh-CAT. Following bombardment, sin mutant or wild-type seeds were incubated in a
solution containing no hormones or 10'6 M GA-j, respectively Data represent mean ( S.E.M.)
of 10 replicates.


99
Muthukrtshnan, S., B.S. Gill, M. Swegle and G.R. Chandra. 1984. Structural oenes for amylases are located on barley chromosome 1 and 6. J. Biol. Chem. 259:13837-13639.
Nambara, E.. K. Keith, P. McCourt and S. Naito. 1994. Isolation of an internal deletion mutant
of the Arabidopsis thaliana ABI3 gene. Plant Cell Physiol. 35: 509-513.
Nambara, E., S. Naito, and P. McCourt. 1992. A mutant of Arabidopsis which is defective in
seed development and storage protein accumulation is a new ab/3 allele. Plant J. 2:
435-441.
Napier, JA, J.M. Chapman and M. Black. 1989. Calcium-dependent induction of novel
proteins by abscisic acid in wheat aleurone tissue of different developmental stages.
Planta 179:156-164.
Neill, S.J., R. Horgan and A.D. Parcy. 1986. The carotenoid and abscisic acid content of
viviparous kernels and seedlings of Zea mays L. Planta 169: 87-96.
Neill S.J., R. Horgan and A.F. Rees. 1987. Seed development and vivipary in Zea mays L.
Planta 171: 358-364.
Neuhaus, G., C. Bowler, R. Kern and N.-H. Chua. 1993. Calcium/calmodulin-dependent and -
independent phytochrome signal transduction pathways. Cell 73: 937-952.
Nicholls, P.B. 1979. Induction of sensitivity to gibberellic acid in developing wheat caryopses:
effect of rate of desiccation. Aust. J. Plant Physiol. 6:229-240.
Nicholls, P.B., A.W. MacGregor and BA Marchylo. 1986. Production of a-amyiase isozymes in
barley caryopses in the absence of embryo and exogenous gibberellic acid. Aust. J.
Plant Physiol. 13: 239-247.
Nolan, R.C. and T.H.D. Ho. 1988. Hormonal regulation of gene expression in barley aleurone
layers. Induction and suppression of specific genes. Planta 174: 551-560.
Oishi, M.Y. and J.D. Bewley. 1990. Distinction between the responses of developing maize
kernels to flouridone and desiccation in relation to germinability, a-amylase activity, and
abscisic acid content. Plant Physiol. 94: 592-598.
Oliver, M.J., J. Armstrong and J.D. Bewley. 1993. Desiccation and the control of expression of
p-phaseolin in transgenic tobacco seeds. J. Exp. Bot. 44:1239-1244.
Ooms, J.J.J., K.M. Leon-Kloosterziel, D. Bartels, M. Koomeef and C.M. Karssen. 1993.
Acquisition of desiccation tolerance and longevity in seeds of Arabidopsis thaliana. Plant
Physiol. 102:1185-1191.
Paiva, R. and A.L. Kriz. 1994. Effect of abscisic acid on embryo-specific gene expression
during normal and precocious germination in normal and viviparous maize (Zea mays)
embryos. Planta 192: 332-339.
Paleg, L.G. 1960a. Physiological effects of gibberellic acid: I. On carbohydrate metabolism and
amylase activity of barley endosperm. Plant Physiol. 35: 293-299.
Paleg, L.G. 1960b. Physiological effects of gibberellic acid: II. On starch hydrolyzing enzymes
of barley endosperm. Plant Physiol. 35: 902-906.


6
Isolation of Mutants Affected In Seed Maturation
Genetic deficiencies in seed maturation manifest themselves In precocious germination
(vivipary) or reduced dormancy. Severely affected mutants exhibit additional features, such as
intolerance to desiccation and reduced accumulation of seed storage proteins and LEAs.
Mutants have been most intensively isolated and analyzed in maize and Arabktopsis.
Maize has been a model species for studying genetics for many years, mostly because its
monoecious flower structure in combination with self fertility has allowed easy outcrossing and
selfing. Moreover, the identification of several maize transposable elements (Ac/Ds, Spm/En,
Mu) has made it possible to generate transposable element-induced mutants and subsequently
clone the mutated locus using the transposon as a tag. Arabidopsis has become a powerful
model species for a variety of reasons (such as short generation time, small genome size
(Meyerowitz, 1987; Meyerowitz, 1994). Rather recently, cloning of mutated genes has been
achieved by Agrobacterium transformation-mediated T-DNA tagging or chromosome walking
(positional cloning).
Many maize viviparous mutants that arose spontaneously were collected and described
already during the first half of this century (Eyster, 1931; Mangelsdorf, 1930; Robertson, 1955).
Transposon-tagged mutants were induced more recently (e.g. McCarty et al., 1989a,b). In
Arabidopsis, many mutants displaying vivipary or reduced dormancy have been isolated by
screening chemically mutagenized seed for germination on medium containing ABA at a
concentration that inhibits germination of wild-type seeds. Other mutants have been isolated by
screening for the ability to germinate in the absence of GA which is normally required. Here, GA
contents in the seed were reduced either by treatment with a GA-synthesis inhibitor or by using a
mutant deficient in GA-biosynthesis. Recently, screening of transgenic lines produced by
Agrobacterium-mediated seed transformation (Feldmann, 1991) for mutants defective in late
embryogeny has revealed additional loci that control induction of seed maturation/suppression of
precocious germination. Taken together, the identified mutants fall into three classes: 1) ABA-


69
A second large, ca. 400 bp deletion (87/88) also rendered the VP1 protein incapable of
repressing Amy-GUS (Fig. 18). The activity of this construct varied to an unusual extent, from
slight, but non-significant repression in some experiments to more than two-fold, statistically
significant activation of Amy-GUS in others. Similarly, a construct with a slightly smaller deletion
of this region (92/88) displayed a highly variable effect. Subsequently, four adjacent sub
deletions within the 87/88 domain were constructed. Each of the N-terminal three deletions
(87/92, 92/181, 93/95) eliminates one region that is conserved in the barley VP1 homolog, but
none of these deletions severely diminished repression. In contrast, the C-terminal, ca. 140 bp
deletion 196/88 eliminating a non-conserved stretch of VP1 almost entirely abolished repressor
function, implying that this region may contain an important site involved in Amy-GUS
repression.
Deletions in the C-terminal portion of VP1 did not strongly affect repression of Amy-GUS
(Fig. 19). Truncation of the C-terminal 450 bp (VP1-McW) generating the product of the vp1-
McWhirter allele which confers a non-viviparous, anthocyanin-deficient phenotype had only a
slight effect in maize, while not affecting repression in barley. Similarly, deletion of the domains
101/100 and 103/104 did not strongly diminish Amy-GUS repression.
The 87/88 deletion mutant
It was shown that in the absence of GA, co-expression of VP1 had no effect on Amy-
GUS expression in aleurone of germinating barley half seeds (see Fig. 11). In contrast, when
the 87/88 deletion-derivative of VP1 was over-expressed with Amy-GUS in the absence of GA,
Amy-GUS expression was activated (Fig. 20). This activation was highly variable, ranging from
2-fold in some experiments to up to 12-fold in others. It appeared specific to the 87/88 deletion
mutant and was not found for any other tested constructs containing deletions outside this region
of VP1 (data not shown). Interestingly, activation of Amy-GUS by 87/88 was also observed in
the presence of ABA (Fig. 21). In the presence of GA, a slight activation of Amy-GUS (max. 2-
fold) by 87/88 was observed in some, but not all experiments (Figs. 18, 20).


9
germinate precociously late in seed development. Such a screen has proven successful in
isolating abi3 and other mutants (Keith et al.t 1994).
Wild-type developing seeds of Arabidopsis accumulate ABA as a dual peak, an earlier
maternally-derived one and a later embryo-derived one (Karssen et al., 1983). Reciprocal
crosses between wild-type and aba mutant plants demonstrated normal dormancy in the absence
of maternal ABA but not in the absence of embryonic ABA. Moreover, induction of dormancy,
as judged from the acquired inability of the developing seed to germinate precociously,
correlated well with the later peak of ABA accumulation (Karssen et al., 1983). Hence,
acquisition of a dormant state is dependent on ABA produced by the embryo and is normally
independent of ABA provided by the mother plant.
ABA-insensitive mutants
ABA-insensitive mutants of maize and Arabidopsis accumulate normal or higher
concentrations of ABA in developing seeds as compared to wild-type (Neill et al., 1986; 1987;
Koomeef et al., 1984). However, while the mutant phenotype of ABA-deficient mutants can be
complemented by exogenous application of ABA, ABA-insensitive mutants continue to display
vivipary or reduced dormancy in the presence of added ABA (Raubichaud et al., 1980;
Raubichaud and Sussex, 1986; Koomeef et al., 1984). In a maize mutant, it was shown that the
reduced sensitivity to ABA was not caused by a deficiency in ABA transport or metabolism
(Raubichaud and Sussex, 1986). Thus, ABA-insensitive mutants are likely to be affected in ABA
signal transduction.
A single locus (Vp1) regulating ABA-sensitivity has been identified in maize. vp1 mutant
embryos do not acquire desiccation tolerance and germinate precociously on the ear producing
green seedlings. The vp1 mutation affects only seed tissues. When rescued and transferred to
soil prior to desiccation, mutant seedlings form a normal appearing mature plant. Interestingly,
this mutation causes a pieiotropic phenotype. Besides displaying vivipary, vp1 mutant seeds fail
to accumulate anthocyanin pigments in embryo and aleurone tissues (Robertson, 1955).
Consistent with this phenotype, activities of enzymes catalyzing anthocyanin biosynthesis were


44
1 irriTLTlTU-TlO:! 11U1UliCOTWXTCTCtOC^WOCCTCCTOOBOC^COCCCaaCCTCCCAaWCeCO 120
r^^^/TrirryTr^^i^Tf^^rraT-rT-rTyTrLraiaiTrraj^xraAayTreaXlATOACTTCATMTaArrSAAflACAUil UXO-'llXCTCCOOGA^l 1^1 l^l 11^ 210
2*HGODOOAADrM*IOO*AADAIDtFSL2DfCl* S
TCOCCgTCCJ^CCA^CTCGTCCAACTCCTa2TCAAACTCCTCCAOCOCCTACACaU>CAgKXaaaAABI>OCaiGOaOCGOCCeTCOBASCCICCTTOOgCOOQaAAOOOm 1*0
SPSSSTFSSMSSSMSSSAYTMTAGRAQOEPSEPASAGEGF 96
94: CCATG0
?TTirir?rr?^T7Tl^Jl MPWDSBPFPGVSMMLENAHSAPPQ 136
CCTOTOGGTGAni?GrATGAf?TGAA^AGAAAGCnTTf;CCT?^A*?f?frArcAmaQQQG*flRGGRcccc7t3CATGGATGCGTCGGAGQQOGAGGAGCTGCCGCQGTTCTTCATOGAGTOGCTC 600
PVGDGMSKEKAVP1GTTGGKKACH0ASIGBILPRFFMKHL 176
YEFPAGGQDMAAGGGTSWMPHQQAFTPPAAYGGDAVYPSA 296
93: CGAXGG
194: CCATOfl
PQRMAGVEASATKEARKKRMARQRRLS CLQQQRSQQLS LG 16
QIQTSVHLQEPSPRSTHSGPVTPSAGGHGFWSPSSQQQVQ 56
NPLSKSNSSRAPPSSLEAAAAAPQTKPAPAGARQDDIHHR 96
KEAEVHLPELKTRDGISIPNEDIGTSRVHNMRYRFHPNNK 576
100: CCATGG
AGCAGAATGT AT CTGCTGGAAAACACAGGGGAAT T TGT T CGT T CCAACGAGCTT CAGGAGGGGGATTT CAT AGTGAT CT ACT CCGATGT CAAGT CGGGCAAAT AT CTGAT ACGGGGCG TG 1920
SRMYLLENTGEFVRSNELQEGDFIVIYSDVKSGKYLIRGV 616
GTCGTCGACGGGGTCAGCGGCGCCTGCAAGGGGAGGTCTCOGGAAGGCGTGCGGCGGGTTOGGCAGCAGGGAGCCGGOGCCATGAGCCAGATGGCGGTGAGCATCTGAAAGAGCAGCAGG 2160
VVDGVSGACKGRS P BGVRRVRQQGAGAMSQMAVS I 691
CTCCGCCATATATTGATCGATCGACCAATCGATCGTTAGTTCTCCAAGTTACTATTAGCTAGCTATAGCCCGAAACAGCTGAACTGATGATGACGATGGTAACCTCCGTOGTGTGTGTGC 2280
T AAGCATGT AGCGTGCT AGGAGATGAT AT ATT AAATAT AATCGAGT AGT AGAGCCT ACCCGCTGTGTGACGCT AAAT T TGTGTGCATTTGGTTTGGTT TGTGAGTTGGGCCCGTGCGTGG 2 00
CTGTGTCATGTCGTGGTT AATT AGCT AT ACTAGTCCTGTCTGTACATGCATGGACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 24 98
Fig. 4. Location of Nco\ (C/CATGG) restriction sites introduced by site-directed mutagenesis
(adapted from McCarty et al., 1991).


79
In summary, using single-deletion mutants, amino add residues within the region from
222 to 374 of VP1 were found to be important for repression of Amy-GUS. Further functional
analysis of sequences outside this region was conducted by taking advantage of the dominant
negative effect of the 87/88 mutant on repression of Amy-GUS by wild-type VP1: In the
presence of GA, co-expression of 87/88 with VP1 and Amy-GUS severely reduced repression of
Amy-GUS by VP1 (Fig. 22). Domains essential for mediating the dominant negative effect of
87/88 are most likely also involved in function of the wild-type VP1 protein. 87/88 may exert a
dominant negative effect for instance by competing with wild-type VP1 for binding to a
component of the repression mechanism or if VP1 functions as a dimer by forming non
functional heterodimers with wild-type VP1. To identify domains in the 87/88 mutant that are
required for expressing the dominant negative effect, double-deletion mutants between 87/88
and other deletion mutants were constructed and tested. The RED domain deleted in the 85/87
mutant was considered a putative domain because it has an important function in repression.
However, the double-deletion mutant 85/88 was as effective in causing a dominant negative
effect as the 87/88 single mutant, indicating that the RED domain is not essential for this effect
(Fig. 23). Similar results were obtained for the domain 101/100 (Fig. 24). In contrast, the double
mutants deleting 87/88 and either the C-terminal 450 bp of VP1 (87/88:McW) or the highly basic
domain 103/104 (87/103) did not exhibit a dominant negative effect on Amy-GUS repression by
co-expressed VP1 (Figs. 23, 24). Although it cannot be ruled out that these double mutant
constructs express instable proteins, these data suggest that the C-terminus and the domain
103/104 may be required for mediating the dominant negative effect of 87/88. Deletion of these
domains displayed a clearly measurable effect only if the 87/88 domain was deleted also, while
the single mutants 103/104 and McW retained almost wild-type repressor function (Fig. 19). This
suggests that in the single mutants other sequences can compensate in function for the deleted
domains, while this is not possible in the 87/88 mutant. However, stability of the mutant proteins
needs to be confirmed, especially for the 87/88:McW double mutant which deletes ca. 42% of
the VP1 sequence.


32
Hence, dormant seeds execute the developmental switch to germination during the imposed
dormancy-breaking treatment.
As described earlier, the transition from seed development to germination is associated
with major changes in gene expression. It is generally accepted that in quiescent seeds,
maturation-related genes cease expression once the water content falls below a level permitting
transcriptional activity, and following imbibition, expression of a new set of genes is initiated
which is specific to the germinating seed. What gene expression programs are executed in
imbibed, dormant seeds is less clear. However, there is evidence from studies in wheat pointing
to a maintenance of maturation-specific gene expression during the state of dormancy (Ried and
Walker-Simmons, 1990,1993; Morris et al., 1991).
Most importantly, maturation and germination programs appear to be regulated
coordinately in the developing seed. Not only is precocious germination of the immature embryo
suppressed, but similarly, the premature induction of germination-related genes appears to be
inhibited during this developmental state. Developing seeds of wheat and barley contain
biologically active GAs in concentrations adequate to induce a-amylase production in the
aleurone layer (Wheeler, 1972; Radley, 1976). Nevertheless, only very low levels of a-amylase
enzyme activity or mRNAs were detected in immature seeds (Comford et al., 1986; Garcia-Maya
et al., 1990). More compelling, neither a-amylase activity nor a-amylase gene expression was
induced when immature seeds were exposed to exogenous GA (Nicholls, 1979; Comford et al.,
1986; Garcia-Maya et al., 1990; Oishi and Bewley, 1990). Similar results were obtained when
treating dormant seeds with GA (Schuurink et al., 1992a). Given that immature or dormant
embryos excised from the seed and placed in culture are capable of responding to GA, these
data are strong evidence for active repression of the GA-response in developing or dormant
seeds. Results consistent with this idea were also reported for dicot seeds (soybean, castor
bean). In these species, immature seeds contained enzymes involved in the degradation of fatty
acids and proteins (malate synthetase, LeuNase, isocitrate lyase) at a much lower level than
germinating seeds or isolated embryos in culture (Kermode, 1990).


94
Foster, CA 1977. Slender an accelerated extension growth mutant of barley. Barley Genet.
Newslett. 7:24-27.
Fujioka, S., H. Yamane, C.R. Spray. M. Katsumi, B.O. Phinney, P. Gaskin, J. MacMillan and N.
Takahashi. 1988. The dominant non-gibberellin-responding dwarf mutant (Dfl) of maize
accumulates native gibberellins. Proc. Natl. Acad. Sci. USA 85: 9031-9035.
Galau, G.A., N. Bijaisordat and D.W. Hughes. 1987. Accumulation kinetics of cotton late
embryogenesis-abundant mRNAs and storage protein mRNAs: Coordinate regulation
during embryogenesis and the role of abscisic acid. Dev. Biol. 123:198-212.
Galau, G.A., D.W. Hughes and L. Dure III. 1986. Abscisic acid induction of cloned cotton late
embryogenesis-abundant (Lea) mRNAs. Plant Mol. Biol. 7:155-170.
Galau, G.A., K.S. Jakobsen and D.W. Hughes. 1991. The controls of late dicot embryogenesis
and early germination. Physiol. Plant. 81:280-288.
Gale, M.D. and GA Marshall. 1975. The nature and genetic control of gibberellin insensitivity
in dwarf wheat grain. Heredity 35: 55-65.
Garcia-Maya, M J.M. Chapman and M. Black. 1990. Regulation of a-amylase formation and
gene expression in the developing wheat embryo. Planta 181:296-303.
George, D.W. 1967. High temperature seed dormancy in wheat (Trticum aestivum L.) Crop
Sci. 7: 249-253.
Gilroy, S. and R.L. Jones. 1992. Gibberellic acid and abscisic acid coordinate^ regulate
cytoplasmic calcium and secretory activity in barley aleurone protoplasts. Proc. Natl.
Acad. Sci. USA 89: 3591-3595.
Gilroy, S. and R.L. Jones. 1993. Calmodulin stimulation of unidirectional calcium uptake by the
endoplasmic reticulum of barley aleurone. Planta 190: 289-296.
Giraudat, J., B.M. Hauge, C. Valon, J. Smalle, F. Parcy and H. Goodman. 1992. Isolation of the
Arabidopsis ABI3 gene by positional cloning. Plant Cell 4:1251-1261.
Goldman, S., Y.R. Mawal, I. Taida and R Wu. 1994. Studies of a gibberellin-dependent DNA-
binding protein related to the expression of a rice a-amylase gene. Plant Sci. 99: 75-88.
Goldstein, L.D. and P.H. Jennings. 1978. Amylase enzymes isolated from incubated de-
embryonated maize kernels. New Phytol. 81: 233-242.
Gubler, F., and J.V. Jacobsen. 1992. Gibberellin-responsive elements in the promoter of a
barley high-pl a-amylase gene. Plant Cell 4:1435-1441.
GuiKinan, M.J., W.R. Marcotte, Jr., and R.S. Quattrano. 1990. A plant leucine zipper protein
that recognizes an abscisic acid response element. Science 250: 267-271.
Harberd, N.P. and M. Freeling. 1989. Genetics of dominant gibberellin-insensitive dwarfism in
maize. Genetics 121: 827-838.
Harvey, B.M.R. and A. Oaks. 1974. The role of gibberellic acid in the hydrolysis of endosperm
reserves in Zea mays. Planta 121:67-74.


35
wheat embryos (Walker-Simmons, 1988). This differential inhibitory effect of ABA depending on
the degree of dormancy was also observed using intact, mature seeds (Morris et al., 1989).
In conclusion, depth of dormancy appears to be positively correlated with ABA-sensitivity
with respect to inhibition of germination. Why, and if, enhanced ABA-sensitivity is the immediate
cause for inhibition of germination is unclear. Because ABA has been shown to inhibit water
uptake by the embryo (Schopfer and Plachy, 1984), it has been suggested that high sensitivity to
ABA in dormant seeds may result in reduced water uptake in the embryo and thereby prevent
radicle emergence (Walker-Simmons, 1987). Additionally, ABA may have a differential effect on
gene expression in dormant and non-dormant seeds. Indeed, transcript levels of a variety of
ABA-responsive genes remained high in imbibed dormant wheat seeds, whereas they declined
rapidly in non-dormant seeds following imbibition (Morris et al., 1991). Similarly, maturation-
related LEA proteins were abundant in rehydrated dormant seeds but not in non-dormant seeds
(Ried and Walker-Simmons, 1990, 1993). However, most of the identified ABA-responsive
proteins that accumulate specifically in dormant seeds are predicted to be dehydrins and may
therefore function primarily in maintaining the embryo in a desiccation-tolerant state rather than
in directly inhibiting germination.
Exogenous application of ABA is known to inhibit GA-mediated activation of hydrolase
genes in the aleurone of germinating cereal seeds (Jacobsen and Chandler, 1987). In this
context, the following observation may be important. Imbibed dormant barley seeds showed
very low expression of a-amylase genes as compared to non-dormant seeds (Morris et al., 1991;
Schuurink et al., 1992a; Skadsen, 1993). Moreover, dormant grains produced less a-amylase in
response to GA than non-dormant grains (Schuurink et al., 1992a; Skadsen, 1993). Hence, seed
dormancy appears to be correlated with a reduced responsiveness of the aleurone to GA.
Experiments with isolated aleurone layers indicated that the reduced GA-sensitivity of aleurone
cells of dormant barley seeds is dependent on the presence of the starchy endosperm
(Schuurink, 1992a; Skadsen, 1993), implying that the starchy endosperm may liberate an
inhibitory factor that diffuses to the aleurone cells. It is suggestive that this putative diffusible


38
application of GA (Hilhorst and Karssen, 1992). Since mature seeds contain very low
concentrations of ABA, it was not considered likely that the ga-1 single mutant required GA to
directly oppose the action of endogenous ABA present in the seed. In contrast, a remote
control* model was suggested in which the GA requirement for germination depends on the
depth of dormancy induced during seed development. Deeply dormant seeds, as wild-type
seeds, have a high GA-requirement to promote germination, while seeds with little dormancy
(ABA mutants) have a low GA-requirement which may be satisfied by low concentrations of GA
present in the possibly leaky ga-1 mutant.


SUMMARY AND CONCLUSIONS
The Viv¡parous-1 (VP1) transcriptional activator of maize is required for abscisic acid-
induction of maturation-specific genes late in seed development leading to acquisition of
desiccation tolerance and arrest in embryo growth (McCarty et al., 1991). The presented
research extends these findings by showing that VP1, in addition to its transcriptional activation
function, inhibits precocious induction of the germination-specific a-amylase genes in aleurone
cells of the developing seed. Functional analysis of deletion-derivatives of VP1 in a transient
gene expression system indicated that VP1 has a discrete repressor function that is separable
from its transcriptional activation function. It is therefore suggested that physically combining
activator and repressor functions in one protein provides one mechanism for directly integrating
control of the mutually exclusive developmental pathways, seed maturation and seed
germination, in the plant embryo.
90


64
There are at least two possible scenarios that would explain how the state of the embryo
might influence de-repression of a-amylase genes in vpf mutant aleurone: 1) a viviparous
embryo might secrete an inductive signal required for a-amylase expression in the aleurone of
developing seeds (e.g. GA), and/or 2) a non-viviparous embryo might contribute a diffusible
inhibitory signal that prevents a-amylase expression in the aleurone (e.g. ABA). In order to test
these hypotheses, Amy-GUS was introduced into aleurone cells of a germless mutant in which
the embryo aborts during the early globular state (P. Becraft and D.R. McCarty, pers.
communication). Hence, use of ears that segregate for the germless and vpf mutations allows
assessment of aleurone responsiveness in the absence of a signal from the embryo. Amy-GUS
was highly de-repressed in vpf-mutant aleurone of germless seeds. Aleurones of all four seeds
bombarded expressed Amy-GUS (Fig. 16). This indicates that a viviparous embryo per se is not
required for de-repression of a-amylase genes in vpf mutant aleurone cells. Rather, it appears
that the lack of a normal (non-viviparous) embryo caused induction of Amy-GUS in the
germless,vp1 double mutant. This suggests that a wild-type embryo secretes a signal with
inhibitory function on a-amylase expression in the aleurone.
In the single mutants vpf and germless, Amy-GUS was partially de-repressed (2-3 seeds
of a total of six bombarded expressed Amy-GUS, Fig. 16). Only wild-type seeds displayed
complete repression of Amy-GUS (Fig. 16). Hence, both VP1 expression in the aleurone and a
normal embryo appear to be required for complete inhibition of a-amylase genes in the aleurone.
This is consistent with the Amy-GUS activities found in vpf non-concordant kernels very late in
development, as described above.
Functional Analysis of the VP1 Protein
The acidic activation domain
We considered two models of how VP1 may function in repression of the aleurone
germination response. 1) VP1 might be a transcriptional activator of an intermediate repressor
gene that in turn inhibits expression of a-amylase genes (Figure 17a). 2) VP1 itself might


19
I
ABA VP1
Maturation Genes
Anthocyanin Genes
Fig. 1. Role of VP1 in activation of seed maturation-related pathways and anthocyanin
biosynthesis.


41
chromosomes. Due to a translocation event between A-chromosome 3 and the B-chromosome,
the 3La part of the A-chromosome 3 is carried by the B-chromosome, while heterochromatlc
DNA is found on the 3La part of the A-chromosome 3 (Fig. 2). The TB3La stock was crossed to
vp1-R at least once. Hence, resulting TB3La, AVP*-** B+ plants carry one 3A-chromosome
segregating for the vp1 mutation, the homologous 3A chromosome with heterochromatic DNA
(thus conferring a vp1 mutant phenotype) and a B-chromosome carrying the wild-type Vp1 gene .
To obtain vp1 non-concordant kernels, pollen from TB3La, A*?1'** B* plants is crossed onto
segregating vp1-R females. During the second pollen mitosis, the replicated B-chromosomes
undergo non-disjunction forming one sperm nucleus with two B-chromosomes and one sperm
nucleus without a B-chromosome (Fig. 2). Hence, following double fertilization, non-concordant
seeds are produced carrying a vp1 mutant embryo and a wild-type endosperm or vice versa (Fig.
2).
For experiments with germinating wild-type seeds of maize, seeds of the variety NK508
were used (kindly provided by Northrup-King).
Wild-type barley seeds c.v. Himalaya were obtained from Washington State University,
Pullman, WA (harvests 1988, 1991 and 1992). Seed segregating for the slender mutation
(Himalaya background) was kindly provided by P. Chandler. As with seed segregating for D8,
wild-type and slender mutant endosperms were genotyped by germination of the excised
embryo.
Plasmid Constructs
In all experiments, JR254 (Amy-GUS) or JR303 were used as reporter constructs (see
Fig. 3). Amy-GUS and JR303 were kindly provided by J. Rogers and T.H.D. Ho, respectively.
Amy-GUS contains ca. 1,800 bp of the 5 flanking sequence of a barley high pi a-amylase gene
(Amy6-4] Kursheed and Rogers, 1988), the first intron of Amy6-4, the GUS reporter gene and the
Amy6-4 3 terminator. JR303, containing a low-pl a-amylase promoter, was derived from


48
Fig. 5. Cell autonomous de-repression of the aleurone germination response in vp1-m2 mutant
aieurone sectors. (A) The kernel shown is a mosaic: regions pigmented with purple anthocyanin
are wild-type; yellow, anthocyanin-deficient regions are clonal sectors of aleurone that have lost
Vp1 function. (B) Magnification of a vp1-m2 kernel.


73
The finding that 87/88 activates Amy-GUS indicates that the mutant protein is not fully
non-functional. The data suggest that 87/88 may be capable of interacting with a normal
component of the repression mechanism but unable to cause repression. In doing so, there are
at least two possibilities as to how it might activate Amy-GUS. 1) The acidic activation domain
of 87/88 might elicit transcriptional activation of a-amylase genes or an intermediate gene. In
the wild-type VP1 protein, this activity might either be not accessible or masked by the repressor
function. 2) In producing a non-functional complex, 87/88 might compete with, or titrate out, an
endogenous repressor (e.g. endogenous barley VP1 -homolog possibly present in aleurone of
germinating barley seeds) and thus exert a dominant negative effect.
In order to test the first possibility, a double deletion mutant was constructed that deletes
the 87/88 domain and the acidic activation sequence. This double mutant was as effective in
activating Amy-GUS as the 87/88 single mutant (Fig. 20), suggesting that the transcriptional
activation domain is not involved. Therefore, it was tested whether 87/88 is capable of inhibiting
the effect of recombinant VP1. 87/88 and recombinant VP1 were expressed by themselves and
in combination (ratio 4:1) together with Amy-GUS in barley aleurone. Co-expression of 87/88
reduced VP1 -mediated repression of Amy-GUS by ca. 75% (Fig. 22). This is consistent with the
view of a dominant negative effect of 87/88.
To identify domains involved in mediating the dominant negative effect of 87/88, double
deletion mutants deleting 87/88 and other sequences of the VP1 protein were constructed and
tested for their ability to reduce repression of Amy-GUS by co-expressed VP1. Double mutants
deleting the domains 87/88 and 85/87 (the RED domain) or 101/100, respectively, were as
effective in competing with recombinant VP1 as the 87/88 single mutant (Figs. 23, 24). In
contrast, the double mutants deleting 87/88 and either the C-terminal 450 bp of VP1
(87/88:McW) or the domain 103/104 (87/104) did not show a dominant negative effect on Amy-
GUS repression by co-expressed VP1 (Figs. 23, 24), suggesting that these domains may be
important for the inhibitory role of 87/88 on VP1 repressor function.


95
Hattorl, T., T. Terada and S.T. Hamasuna. 1994. Sequence and functional analyses of the rice
gene homologous to the maize Vp1. Plant Mol. Biol. 24: 805-810.
Hattori, T., V. Vasil. L. Rosenkrans, LC. Hannah, D.R. McCarty and I.K. Vasil. 1992. The
VMparous-1 gene and abscisic acid activate the C1 regulatory gene for anthocyanin
biosynthesis during seed maturation in maize. Genes and Dev. 6: 609-618.
Higgins, T.J., J.A. Zwar and J.V. Jacobsen. 1976. Gibberellic acid enhances the level of
translatable mRNA for a-amylase in barley aleurone layers. Nature 260:166-169.
Hilhorst, H.W.M. and C.M. Karssen. 1992. Seed dormancy and germination: the role of abscisic
acid and gibberellins and the importance of hormone mutants. Plant Growth Reg. 11:
225-238.
Hillmer, S., S. Gilroy and R.L. Jones. 1992. Visualizing enzyme secretion from individual barley
(¡Hordeum vulgare) aleurone protoplasts. Plant Physiol. 102: 279-286.
Hoekstra, FA. L.M. Crowe and J.H. Crowe. 1989. Differential desiccation sensitivity of com
and Pennisetum pollen linked to their sucrose contents. Plant Cell Envir. 12: 83-91.
Hoekstra, FA and T. van Roekel. 1988. Desiccation tolerance of Papaver dubium during its
development in the anther Possible role of phospholipid composition and sucrose
content. Plant Physiol. 88: 626-632.
Hooley, R., M.H. Beale and S.J. Smith. 1991. Gibberellin perception at the plasma membrane
of Avena fatua aleurone protoplasts. Planta 183: 274-280.
Hughes, D.W., and G.W. Galau. 1991. Developmental and environmental induction of Lea and
LeaA mRNAs and the postabscission program during embryo culture. Plant Cell 3:605-
618.
Huttly, A.K. and D.C. Baulcombe. 1989. A wheat a-Amy2 promoter is regulated by gibberellin in
transformed oat aleurone protoplasts. EMBO J. 8: 1907-1913.
Huttly, A.K., A.L. Phillips and J.W. Tregear. 1992. Localization of cis elements in the promoter
of a wheat a-amy2 gene. Plant Mol. Biol. 19: 903-911.
Ingle, J. and R.H. Hageman. 1965. Metabolic changes associated with the germination of com.
III. Effects of gibberellic acid on endosperm metabolism. Plant Physiol. ??: 672-675.
Jacobsen, J.V. and L.R. Beach. 1985. Control of transcription of a-amylase and rRNA genes in
barley aleurone protoplasts by gibberellin and abscisic acid. Nature 316:275-277.
Jacobsen, J.V. and P.M. Chandler. 1987. Gibberellin and abscisic acid in germinating cereals.
In Plant Hormones and Their Role in Plant Growth and Development (ed. P.J. Davies),
Dordrecht: Martinus Nijhoff. pp. 164-193.
Jacobsen, J.V. and T.J. Close. 1991. Control of transient expression of chimaeric genes by
gibberellic acid and abscisic acid in protoplasts prepared from mature barley aleurone
layers. Plant Mol. Biol. 16: 713-724.


BIOGRAPHICAL SKETCH
Ute Hoecker was bom on August 29, 1964, in Aachen, Germany. She completed high
school in Bonn, Germany, in 1984. She enrolled in the Friedrich Wilhelm University of Bonn and
received a 'Vordiplom' (B.S.) in agricultural sciences in 1986. Following a year of practical
training on a laboratory farm in Bonn, she transferred to the University of Hohenheim in
Stuttgart-Hohenheim, Germany, and earned a "Diplom" (M.S.) in agricultural sciences with
specialization in plant breeding and population genetics in 1990. She moved to Gainesville,
Florida, in August, 1990, to begin her doctoral studies.
104


21
other hand, ABA firans-activated C1-GUS in maize protoplasts. Moreover, a c1 mutant (cf-p;
Chen and Coe, 1978) that fails to accumulate anthocyanin during seed development carries a 5
bp deletion in the promoter region of the gene that, when reconstructed by site-directed
mutagenesis of the C1 promoter and used in transient expression experiments, specifically
abolished ABA-responsiveness without severely affecting frans-activation by VP1 (Hattori et al.,
1992). Hence, the unpigmented phenotype of c1-p may be caused by a deficiency in ABA-
response. Pigmentation in ABA-deficient mutants might be possible if activation of C1 has an
ABA-requirement several orders of magnitude lower than activation of Em.
The Aleurone Germination Response in Cereal Seeds
During cereal seed development, the outermost endosperm cells differentiate into the
aleurone layer which at seed maturity consists of small, thick-walled cells with plasmodesmatal
connections. The aleurone cells are characteristically rich in protein and lipid bodies,
mitochondria and ER, but are devoid of starch grains. In response to GA released by the
germinating embryo, these highly specialized cells synthesize large amounts of hydrolytic
enzymes co-translationally on the rough ER and following proper folding in the lumen of the ER
- secrete these enzymes into the endosperm. This aleurone germination response to GA can be
inhibited by treatment with ABA.
The predominant hydrolytic enzyme synthesized is a-amylase, constituting ca. 15-20%
of total translatable mRNA and ca. 30% of total protein synthesis in germinating barley seeds
(Khursheed and Rogers, 1988). In barley and wheat, the major source of a-amylase is the
aleurone layer, whereas in maize, sorghum and rice, significant contributions are also made by
the embryo scutellum (Ranki and Sopanen, 1984; Dure, 1960a,b; Akazawa and Miyata, 1982).


18
The molecular mechanism underlying the synergistic interaction of VP1 and ABA is thus
far unclear. The possibility that ABA is required for high stability of the VP1 transcript or protein
which then in turn activates Em is highly unlikely because in the ABA-defident vp5 mutant no
reduction in VP1 transcript and protein levels was observed (McCarty et al., 1991; C. Carson and
D.R. McCarty, unpublished results). Possibly, ABA post-translationally modifies the VP1 protein
and thus enhances its trans-activation function. Alternatively, VP1 and an ABA-dependent factor
might be part of a complex that forms on the Em promoter and induces transcription.
Genes encoding proteins of the anthocvanin pathway
As identified by mutants, at least eight genes are essential for accumulation of
anthocyanin pigments in embryo and aleurone cells of the maturing maize seed (Fig. 1). Five of
these genes (A1, A2, C2, Bz1, Bz2) encode enzymes of the anthocyanin biosynthesis pathway
(Dooner et al., 1991). Expression of these structural genes requires the coordinate action of two
regulatory proteins, C1 and a member of the R/B gene family (Coe et al., 1988). Both proteins
exhibit features of transcription factors. The C1 protein contains a functional acidic
transcriptional activation domain (Goff et al., 1991) and a region of sequence homology to the
DNA-binding domain of animal myb proto-oncogene products (Paz et al., 1990). Proteins
encoded by the R/B gene family display high homology with the helix-loop-helix motif of myc
proto-oncogene products (Ludwig et al., 1989). At least for the promoter of the Bz1 structural
gene, it has been shown that sequences homologous to the consensus binding sites of animal
MYB and MYC proteins are essential for C1- and R-mediated activation of Bz1 expression (Roth
et al., 1991), suggesting a direct interaction of C1 and R with these sequences in the Bz1
promoter.
A third regulatory factor required for pigmentation of tissues in the developing maize
seed is VP1. The lack of anthocyanin pigments in vp1 mutant seed is associated with the
absence of C1 transcript (McCarty et al., 1989a). Over-expression of a 35S-C1 construct in vp1


62
111
10-
9-
Embryo:
vp1
WT
vp1
WT
Endosperm:
vp1
vp1
WT
WT
Fig. 15. Amy-GUS expression in aleurone of developing vp1 non-concordant maize seeds.
Aleurones (31 DAP, fall season) were bombarded with 10 pg of Amy-GUS and 5 pg of Ubi-LUC
and cultured post-bombardment in no hormones. Data represent mean ( S.E.M.) of six
replicates.


70
Fig. 20. The VP1 deletion-derivative 87/88 activates Amy-GUS in the absence of GA in
aleurones of germinating barley seeds. Aleurones were bombarded with 2 pg of Amy-GUS, 5 pg
of Ubi-LUC and 10 pg of either 35S-Sh-CAT, 35S-Sh-VP1, the activation domain-deletion
mutant described in Fig. 18 (-Act), 87/88 or the double mutant that carries deletions of the
activation domain and 87/88 (-Act:87/88), respectively. After bombardment, endosperms were
cultured in no hormones (top graph) or 10-6 M GA3 (bottom graph). Data represent mean (
S.E.M.) of three replicates.


92
Chandler, P.M. 1988. Hormonal regulation of gene expression in the 'slender* mutant of barley
(Hordeum vulgare L). Planta 175:115-120.
Chandler, P.M. and J.V. Jacobsen. 1991. Primer extension studies on a-amylase mRNAs in
barley aleurone. II. Hormonal regulation of expression. Plant Mol. Biol. 16: 637-645.
Chandler, P.M. and M. Robertson. 1994. Gene expression regulated by abscisic acid and its
relation to stress tolerance. Ann. Rev. Plant Physiol. Mol. Biol. 45:113-141.
Chao, S.E. and J.G. Scandalios. 1969. Identification and genetic control of starch-degrading
enzymes in maize endosperm. Biochem. Genet. 3: 537-547.
Chao, S.E. and J.G. Scandalios. 1971. Alpha-amylase of maize: Differential allelic expression
at the Amy-1 gene locus, and some physiological properties of the isozymes. Genetics
69: 47-61.
Chao, S.E. and J.G. Scandalios. 1972. Developmentally dependent expression of tissue
specific amylases in maize. Mol. Gen. Genet. 115:1-9.
Chen, S.M. and E.H. Coe Jr. 1978. Control of anthocyanin synthesis by the C locus in maize.
Biochem. Genet. 15: 333-346.
Chen, Y. and J.S. Burris. 1990. Role of carbohydrates in desiccation tolerance and membrane
behavior in maturing maize seed. Crop Sci. 30: 971-975.
Coe Jr., E.H. and M.G. Neuffer. 1977. The genetics of com. In: Com and Com Improvement,
Ed. 2, G.F. Sprague ed., American Society of Agronomy, (Madison, Wl), pp. 111-223.
Coe Jr., E.H., M.G. Neuffer and D.A Hoisington. 1988. The genetics of com. In: Com and Com
Improvement, G.F. Sprague and J.W. Dudley, eds, American Society of Agronomy
(Madison, Wl), pp. 81-258.
Comai, L. and J.J. Harada. 1990. Transcriptional activities in dry seed nuclei indicate the timing
of the transition from embryogeny to germination. Proc. Natl. Acad. Sci. USA 87: 2671-
2674.
Comford, C.A., M. Black, J.M. Chapman and D.C. Baulcombe. 1986. Expression of a-amylase
and other gibberellin-regulated genes in aleurone tissue of developing wheat grains.
Planta 169: 420-428.
Crispeels, M.J. and J.E. Varner. 1966. Inhibition of gibberellic acid induced formation of a-
amylase by abscision II. Nature 212:1066-1067.
Croker, S.J., P. Hedden, J.R. Lenton and J.L. Stoddart. 1990. Comparison of gibberellins in
normal and slender barley seedlings. Plant Physiol. 94:194-200.
Crowe, J.H., F.A. Hoekstra and L.M. Crowe. 1992. Anhydrobiosis. Ann. Rev. Physiol. 54:579-
599.
Derkx, M.P.M. and C.M. Karssen. 1993. Effects of light and temperature on seed dormancy and
gibberellin-stimulated germination in Arabidopsis thaliana: studies with gibberellin-
deficient and -insensitive mutants. Physiol. Plant. 89: 360-368.


5
A second characteristic that may be involved in rendering the seed tolerant to
desiccation is the synthesis and accumulation of specific proteins late in seed development Qate-
gmbryogenesis-flbundant proteins, LEAs). The direct function of LEAs is unknown but based on
their high degree of hydrophilicity (Dure et al., 1989) they are assumed to stabilize the structure
of cellular proteins during dehydration (Skriver and Mundy, 1990; Dure, 1993).
Equally uncertain remain the mechanisms that arrest embryo growth and prevent
precocious germination prior to seed maturity. Because immature embryos excised from the
seed and placed in culture are capable of germinating readily, precocious germination of the
embryo in vivo may be actively suppressed by a process that is dependent on an intact seed.
Evidence from embryo culture experiments has implicated two factors in suppression of
precocious germination: restricted water uptake (low water potential in the seed) and the
hormone abscisic acid (ABA). Both factors, when imposed on cultured embryos, inhibited
germination (reviewed in Quatrano, 1987; Kermode, 1990). Indeed, the osmotic potential of
developing soybean embryos has been shown to be even more negative than that of the
osmoticum used to inhibit germination of isolated embryos (Xu et al., 1990). Similarly, ABA
concentrations increase early in the seed maturation phase (Quatrano, 1987) and may thus play
a role in arresting embryo growth.
In summary, the physiological processes responsible for acquisition of desiccation
tolerance and arrest in embryo growth are poorly understood. In the past 15 years, the focus of
research has shifted to the identification of regulatory factors that control the activities of late
embryogeny, especially at the level of gene expression. From its discovery in the 1950s, ABA
has been a factor of interest because a rise in seed ABA concentration correlates well with the
onset of maturation events. Normally, ABA concentrations peak at the time of maximum dry
weight accumulation in the seed and then decrease to low concentrations towards seed maturity
(Quatrano, 1987). A function of ABA in initiating maturation events and suppressing precocious
germination of the developing embryo was confirmed by analyses of mutants defective in late
embryogeny and studies of gene expression.


8
type embryo and a vp5 mutant endosperm were not (Robertson, 1952). Similar results were
obtained with other ABA-deficient mutants (Robertson, 1955). Thus, vivipary in these mutants is
entirely conditioned by the lack of ABA production in the embryo. The endosperm does not play
an active role in preventing vivipary of a genetically viviparous embryo. This is consistent with
later findings showing that the embryo is the major source of ABA produced in the developing
seed (Zeevaart and Creelman, 1988).
In Arabidopsis, only one ABA-deficient mutant (aba) has been identified. It was isolated
in a genetic screen selecting for the ability to germinate of the normally non-germinating GA-
deficient ga-1 mutant (Koomeef et al., 1982). The aba mutant is impaired in the epoxidation of
the carotenoid zeaxanthin (Duckham et al., 1991) and thus displays normal accumulation of
carotenoids. No carotenoid deficient mutants have been isolated in any genetic screen, which
may reflect the predicted lethality of such mutations. The aba mutant produces plants that show
increased withering of stems, leaves and siliques and an enhanced rate of water loss which is
probably caused by reduced stomata closure upon water stress (Koomeef et al., 1982). aba
mutant seeds exhibit strongly diminished seed dormancy. Wild-type Arabidopsis seeds normally
require cold and light treatments to break imposed seed dormancy and allow germination to
occur, whereas a high percentage of aba mutant seeds germinated readily without a need for
dormancy-breaking treatments (Koomeef et al., 1982). However, in contrast to ABA-deficient
mutants in maize, even severe aba alleles that reduce ABA levels in the seed below the level of
detection produce seeds that are non-viviparous and desiccation tolerant (Koomeef et al., 1982).
Hence, ABA may not be required for the induction of desiccation tolerance in Arabidopsis.
However, leakiness of the aba mutant cannot be ruled out. The nature of the performed mutant
screen which selected for the ability of mature, dry seeds to germinate may not allow the
identification of more severely affected mutants. Possibly, residual, very low concentrations of
ABA that may be present in aba mutant seed are sufficient to prevent vivipary (Koomeef et al.,
1989). To test this, a mutant screen could be performed that selects for the ability of seeds to


40
Genotype of Macrospore (Female Parent): Tvp
Fig. 2. Generation of vpf-non-concordant seed using a TB3La translocation stock.


10
not detectable In mutant seed tissues (Dooner, 1985). The lack of anthocyanin pigments Is not
likely to be a result of the reduced ABA-sensItivity of this mutant. ABA-defldent mutants
accumulate normal amounts of anthocyanins, implying that ABA is not required for pigment
formation. Furthermore, separation of the two phenotypes was observed in seeds carrying the
vp1-McWhirter allele. Those seeds are unpigmented but non-viviparous (Coe et al, 1978).
Therefore, the pleiotropic phenotype of the vp1 mutant implies that in evolution, two processes,
suppression of precocious germination and production of anthocyanins, have come under the
control of a single protein.
As in ABA-deficient mutants, the viviparous phenotype of the vp1 mutant is entirely
determined by the genotype of the embryo (Robertson, 1955). Similarly, anthocyanin deficiency
in embryo or aleurone solely reflects lack of functional VP1 in the respective tissue (Robertson,
1955). Thus, the failure to accumulate pigments in vp1 mutant aleurone is not a direct or indirect
result of precocious induction of germination. Cell autonomous function of VR1 in the aleurone
was demonstrated in a transposable element-induced, somatically unstable mutant (vp1-mum1).
Homozygous vp1-mum1 seeds exhibit small sectors in the aleurone that have regained VP1
function due to excision of the Robertsons Mutator transposable element. These revertant
sectors, recognizable by their pigmentation, can be as small as single cells, indicating that VP1
function does not result in production of a diffusable factor that might induce anthocyanins in
neighboring cells (McCarty et al., 1989a).
The Vp1 gene was cloned by transposon tagging using the vp1-mum1 allele (McCarty et
al. 1989a). It encodes a 2500 bp mRNA that is translated into a 73 kD protein. Vp1 is expressed
specifically in embryo and endosperm tissues of the developing seed. Within the endosperm,
Vp1 expression is likely to be restricted to the aleurone, as suggested from the low abundance of
Vp1 mRNA detected in whole endosperm extracts (McCarty et al. 1989a) and the apparent
absence of a mutant phenotype in vp1 mutant endosperm. The Vp1 transcript is present in the
seed as early as 10 days after pollination (DAP), reaches maximum accumulation at 16 DAP and
decreases in abundance towards seed maturity. No Vp1 expression was detected in germinating


66
function as a repressor of the a-amylase genes or of an intermediate gene that is required for
activation of the a-amylase promoter (Figure 17b). In order to distinguish between these
models, we determined whether the transcriptional activation domain of VP1 which is essential
for activation of the Em and C1 genes in maize cells is also required for inhibition of a-amylase.
Figure 17c shows that a deletion derivative of VP1 that lacks the N-terminal activation domain
was as effective in repressing Amy-GUS expression in maize and barley as the full-length
protein. In addition, a VP16/VP1 hybrid protein that contains three copies of the VP16 acidic
activation domain and has a restored capacity to activate Em-GUS and C1-Sh-GUS (McCarty et
al., 1991; Rosenkrans and McCarty, unpublished results) was not more effective than the
activator deletion mutant in causing repression of Amy-GUS. The lack of a requirement for the
activation sequence clearly distinguishes the mechanism of VP1 -mediated repression from the
mechanism of activation of diverse maturation related genes by VP1. These results strongly
indicate that the VP1 protein has a discrete repressor function.
Identification of Sequences Essential for the Repressor Function of VP1
A number of internal deletion constructs were tested for their ability to repress Amy-GUS
in maize and barley aleurone (Fig. 18). A large ca. 350 bp deletion (86/87) entirely abolished
VP1 repressor function, indicating that the deletion-derivative may lack a functionally important
domain, or the deletion may affect the spacing and thereby the function of domains present
outside this sequence. An only slightly smaller deletion in this region (86/85) did not affect
repression in maize nor barley aleurone, suggesting that altered spacing is not likely the reason
for the failure of 86/87 to repress. Indeed, a small deletion of 42 bp in the C-terminal half of
86/87 (85/87) abolished repression in barley and consistently reduced activity in maize aleurone
by ca. 50%. Hence, this region (hereafter referred to as the RED domain) appears to be
essential for VP1 repressor function. Consistent with this conclusion, a large part of the
sequence of the RED domain (W V Q N H+ H* L Q Rf K* R* P R* D) is highly charged,
predicting this domain to be positioned on the surface of the folded VP1 protein, thus accessible
for interactions with other molecules.


This dissertation was submitted to the Graduate Faculty of the College of Agriculture and
to the Graduate School and was accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
August 1995
Dean, College of Agriculture
Dean, Graduate School


63
Fig. 16. Amy-GUS expression in aleurone of developing germless seeds. Aleurones (29 DAP,
fall season) of an ear segregating for the mutations vp1 and germless were bombarded and
cultured as described in Fig 15. Crosses represent single data points of six replicates
(exception: four replicates in the vp1/germless double mutant).


88
VP1 Integrates the Control of Seed Maturation and Germination Programs
It has been shown in this work that VP1 participates in the regulation of two
developmental pathways in the developing maize seed. As a transcriptional activator K is
required for activation of maturation-specific genes (McCarty et al., 1991) and as a repressor it
prevents precocious induction of the normally germination-specific a-amylase genes (data
presented herin). Hence, expression of VP1 specifically during seed development appears to be
involved in ensuring proper ordering of maturation and germination programs. Physically
combining activation and repression function in one protein appears to provide one mechanism
for directly integrating control of mutually exclusive developmental pathways in the plant
embryo. The importance of a tight control of maturation and germination programs for seed
survival is evident in the phenotype of vp1-m2 seeds.
Premature induction of postgerminative development was also reported for the Iec1
(leafy cotyledon 1) mutant of Arabidopsis. In this ABA-sensitive, viviparous mutant, developing
embryos expressed isocitrate lyase genes and a gene encoding a lipid transfer protein at levels
that are normally characteristic of seedlings (West et al., 1994). Double mutant analysis
suggested that the putative Arabidopsis VP1 homolog, ABI3, and LEC1 function in different
pathways (Meinke et al., 1994). Hence, it appears that multiple mechanisms have evolved in
flowering plants to prevent precocious induction of normally germination-specific genes in the
developing embryo.
Thus far, the evidence that VP1 inhibits germination-specific genes is limited to
hydrolase genes in aleurone cells. It is unknown to what extent this repressor activity of VP1 is
also involved in preventing precocious germination of the embryo. Further insight into the
inhibitory role of VP1 during seed development may be provided by stable transformation of vp1
mutant plants with VP1-derivatives that are mutated specifically in the activator or repressor
function.
Cloning of the Vp1 related genes from barley (M. Stoll and D.R. McCarty, unpublished
results), rice (Hatton et al., 1994), Arabidopsis (Giraudat et al., 1992) and tobacco (Phillips and


100
Parcy, F.t C. Valon, M. Raynal, P. Gaubler-Comella. M. Delseny, and J. Giraudat. 1994.
Regulation of gene expression programs during Arabidopsis seed development: roles of
the ABJ3 locus and of endogenous abscisic add. Plant Cell 6:1567-1582.
Paz-Ares. J., D. Ghosal and H. Saedler. 1990. Molecular analysis of the C1-I allele from Zea
mays', a dominant mutant of the regulatory C1 locus. EMBO J. 9: 315-321.
Ptashne, M. 1988. How eukaryotic transcriptional activators work. Nature 335: 683-689.
Quatrano, R.S. 1987. The role of hormones during seed development. In: Plant Hormones and
Their Role in Plant Growth and Development, P.J. Davies ed., Dordrecht: Martinus
Nijhoff,. pp. 494-514.
Radley, M. 1967. Site of production of gibberellin-like substances in germinating barley
embryos. Planta 75:164-171.
Radley, M. 1976. The development of wheat grain in relation to endogenous growth
substances. J. Exp. Bot. 27:1009-1021.
Raikhel, N. 1992. Nudear targeting in plants. Plant Physiol. 100:1627-1632.
Ranki. H and T. Sopanen. 1984. Secretion of a-amylase by the aleurone layer and the
scutellum of germinating barley grain. Plant Physiol. 75: 710-715.
Ried, J.L. and M.K. Walker-Simmons. 1990. Synthesis of absdsic acid-responsive, heat-stable
proteins in embryonic axes of dormant wheat grain. Plant Physiol. 93:662-667.
Ried, J.L. and M.K. Walker-Simmons. 1993. Group 3 late embryogenesis abundant proteins in
desiccation-tolerant seedlings of wheat (Trticum aestivum L). Plant Physiol. 102:125-
131.
Roberts, D.M. and A.C. Harmon. 1992. Calcium-modulated proteins: targets of intracellular
caldum signals in higher plants. Ann. Rev. Plant Physiol. Mol. Biol. 43: 375-414.
Robertson, D.S. 1952. The genotype of the endosperm and embryo as it influences vivipary in
maize. Proc. Natl. Acad. Sd. USA 38: 580-583.
Robertson, D.S. 1955. The genetics of vivipary in maize. Genetics 40:745-760.
Robertson, D.S. 1978. Charaderization of a mutator system in maize. Mutation Res. 51: 21-28.
Robichaud, C.S. and I.M. Sussex. 1986. The response of viviparous-1 and wild-type embryos
of Zea mays to culture in the presence of abscisic add. J. Plant Physiol. 126: 235-242.
Robichaud, C.S. and I.M. Sussex. 1987. The uptake and metabolism of [2-14C}-ABA by
exdsed wild type and viviparous-1 embryos of Zea mays. L. J. Plant Physiol. 130:1 SI-
188.
Robichaud, C.S., J. Wong and I.M. Sussex. 1980. Control and growth of viviparous embryo
mutants of maize by absdsic add. Dev. Genet 1: 325-330.
Rogers, J.C. 1985. Two barley a-amylase gene families are regulated differently in aleurone
cells. J. Biol. Chem. 260: 3731-3738.


89
Conrad, 1994) indicates that the Vp1 gene is conserved among flowering plants. Loss of ABI3
function in Arabidopsis causes a similar viviparous phenotype as the vp1 mutation in maize
(Nambara et al., 1992). The functions of ABI3 and VP1, however, diverge in so far that ABI3 is
required for seed dormancy in Arabidopsis while VP1 does not impose seed dormancy in maize.
Because ABI3 mRNA is stored in the dry seed (Parcy et al., 1994), whereas VP1 transcript and
protein are non-detectable in the mature seed (McCarty et al., 1989; Carson, 1992), one can
speculate that dormancy in Arabidopsis may reflect an extended timing of ABI3 expression after
seed maturity rather than a functional difference in the proteins. This view is supported by the
results showing that over-expression of VP1 in aleurone of germinating maize seeds was
effective in repressing Amy-GUS. A role of VP1 in maintaining seed dormancy is also consistent
with the finding that dormancy in barley is correlated with a reduced GA-inducibiiity of a-amyiase
genes in the aleurone (Schuurink et al., 1992; Skadsen, 1993). Hence, it is suggested that VP1
plays a role in integrating the control of seed maturation, dormancy and germination programs.


97
non-germinating gibberellin sensitive lines of Arabidopsis thaRana (L) Heynh. Theor.
Appl. Genet. 61: 385-393.
Koomeef, M., G. Reuling and C.M. Karssen. 1984. The isolation and characterization of
abscisic acid-insensitive mutants of Arabidopsis thaliana. Physiol. Plant. 61: 377-383.
Koomeef, M. and J.H. Van der Veen. 1980. Induction and analysis of gibberellin sensitive
mutants in Arabidopsis thaliana (L.) Heynh. Theoret. Appl. Genet. 58: 257-263.
Koster, K.L. and A.C. Leopold. 1988. Sugars and desiccation tolerance in seeds. Plant Physiol.
88: 829-832.
Kriz, A.R., M.S. Wallace and R. Paiva. 1990. Globulin gene expression in embryos of maize
viviparous mutants: evidence for regulation of the Glb1 gene by ABA. Plant Physiol. 92:
538-542.
Lanahan, M.B. and T.H.D. Ho. 1988. Slender barley: a constitutive gibberellin-response mutant.
Planta 175:107-114.
Lanahan, M.B., T.H.D. Ho, S.W. Rogers and J.C. Rogers. 1992. A gibberellin response
complex in cereal a-amylase gene promoters. Plant Cell 4:203-211.
Leprince, O., R. Bronchart and R. Deltour. 1990. Changes in starch and soluble sugars in
relation to the acquisition of desiccation tolerance during maturation of Brassica
campestris seed. Plant Cell Envir. 13: 539-546.
Leung, J., M. Bouvier-Durand, P.-C. Morris, D. Guerrier, F. Chefdor and J. Giraudat. 1994.
Arabidopsis ABA response gene ABI1: Features of a calcium-modulated protein
phosphatase. Science 264:1448-1452
Luan, S W. Li, F. Rusnak, S.M. Assman and S.L. Schreiber. 1993. Immunosuppressants
implicate protein phosphatase regulation of K+ channels in guard cells. Proc. Natl.
Acad. Sci. USA 90: 2202-2206.
Ludwig, S.R., L.F. Habera, S.L. Dellaporta and S.R. Wessier. 1989. Lc, a member of the maize
R gene family responsible for tissue-specific anthocyanin production, encodes a protein
similar to transcriptional activators and contains the myc homology region. Proc. Natl.
Acad. Sci. USA 86: 7092-7096.
MacGregor, A.W., B.A. Marchylo and J.E. Kruger. 1988. Multiple a-amyiase components in
germinating cereal grains determined by isoelectric focusing and chromatofocusing.
Cereal Chem. 65: 326-333.
MacLeod, A.M. and G.H. Palmer. 1967. Gibberellin from barley embryos. Nature 216:1342-
1343.
Mangelsdorf, P.C. 1930. The inheritance of dormancy and premature germination in maize.
Genetics 15: 462-494.
Marcotte, Jr., W.R., C.C. Bayley and R.S. Quatrano. 1988. Regulation of a wheat promoter by
abscisic acid in rice protoplasts. Nature 335: 454-457.


Amy-GUS/LUC*1 E04 (pmoles/hr/RLU)
57
GA 3-Concentration (M)
Fig. 11. Co-expression of VP1 inhibited GA3*induction of Amy-GUS but did not affect its basal
activity in aleurone of germinating barley half seeds Aleurones were bombarded with 2 ng of
Amy-GUS, 5 pg of Ubi-LUC, and 5 ng of 35S-Sh-VP1 or 35S-Sh-CAT. Following bombardment,
3 replicates of 5 kernels each were incubated in 0-10'5 M GA3. Data represent mean of three
replicates ( S.E.M). The insert shows activities in the absence of GA3.


26
GA and ABA Signal Transduction
It is generally accepted that molecules with hormonal function act as ligands that upon
binding to their specific receptors elicit a response which ultimately can result in altered gene
expression. In animals, a large number of hormone receptor-encoding genes has been cloned
and characterization of their gene products has demonstrated that receptors can be found
located intracellularly (e.g. steroid hormone receptors) or integrated into the cell membrane with
their ligand-binding domain facing the extracellular space. Identification of plant hormone
receptors has proven difficult. Though hormone-binding proteins have been identified, definite
proof for the function of these proteins is still lacking.
With respect to GA and ABA signal perception in the cereal aleurone, studies have so
far concentrated on the identification of the cellular site of the receptors. Results from at least
one study suggest that GA does not have to enter the cell to regulate gene expression. Hooley
and colleagues (Hooley et al 1991) demonstrated that GA immobilized to Sepharose beads was
capable of enhancing a-amylase transcription in oat aleurone protoplasts. Though these data
point to a perception of the GA signal on the external surface of the plasma membrane, the
existence of intra-cellular receptors cannot be ruled out. Clear evidence against intra-celluiar
receptors came from elegant experiments performed in the laboratory of R. L. Jones. A method
was developed to visualize a-amylase gene expression and a-amylase secretion from individual
protoplasts (Hillmer et al., 1992). This allowed to test whether or not hormones microinjected
into the cytosol of aleurone protoplasts were capable of eliciting a response (Gilroy and Jones,
1993). It was found that protoplasts injected with GA did not respond to the hormone. Similarly,
ABA microinjected into protoplasts was ineffective in antagonizing the stimulating effect of pre
applied external GA. The failure to respond to microinjected hormones was not due to disruption
of protoplast function by microinjection since protoplasts that had been subjected to this
procedure remained responsive to externally applied GA. These results indicate that the sites of
perception of GA and ABA are located on the external face of the plasma membrane in aleurone
cells.


34
drying of immature seeds and elicits precocious germination of immature maize kernels. If
drying merely served to deplete endogenous ABA in developing seeds, then flouridone-treated
kernels and dried seeds should behave similarly with respect to GA-response. However, while
drying resulted in synthesis of high levels of a-amylase following imbibition, flouridone-treated
seeds produced only very low amounts of a-amylase in response to GA. Hence, drying may
achieve two effects: 1) it frees seed tissues of the inhibitory effect of ABA, and 2) it renders the
aleurone competent of responding to GA. The cause of the ABA-independent GA-insensitivity in
immature seeds is thus far unknown.
Dormant Seeds
Cereals
Imposed dormancy in cereal species is normally released by prolonged storage of dry
seeds (afterripening). The duration of seed dormancy following seed maturity depends on a
variety of factors such as the genetic constitution (cultivar), the environmental conditions during
grain maturation (low temperatures and short day length increase dormancy; Schuurink et al.,
1992b) and the rehydration temperature (high temperatures enhance dormancy; George, 1967).
Such differences in the depth of seed dormancy have been utilized to investigate the roles of
ABA concentration and ABA-sensitivity in preventing embryo germination. No clear correlation
between ABA content in the mature embryo and the degree of dormancy was found (Walker-
Simmons, 1987,1988; Morris et al., 1989; Skadsen, 1993). However, there are large differences
between dormant and non-dormant embryos with respect to sensitivity to ABA, as measured by
the capacity of ABA to inhibit germination. Isolated embryos of a non-dormant wheat cultivar
lost their sensitivity to ABA in culture as the grain entered maturation stage, whereas those of a
dormant cultivar retained sensitivity to ABA beyond grain maturity (Walker-Simmons, 1987).
Similarly, elevating the incubation temperature from 15C to 30C, thus inducing high-
temperature dormancy, significantly enhanced the ability of ABA to block germination of isolated


Amylase
repression
Em activation
C1 activation
VP1
Exons
1
2 345
non-essential
essential
Fig. 26. Summary on domains of VP1 involved in repression of a-amylase genes, activation of C1 and activation of Em.


45
MS salts and MS vitamins (Sigma, cat# M5519) on a gyratory shaker in the darte for ca. 36 hrs,
while developing seeds were used immediately. The embryo as well as pericarp and testa
tissues were removed from the seeds to expose the aleurone layer of the endosperm. Prepared
endosperms were placed on Gelrite-solidified salt medium and then subjected to particle
bombardment.
Barley seeds were de-embryonated prior to surface sterilization in 70% ethanol for 1 min
followed by 10 min in 1.75 2.9 % NaOCI. A minimum of 1.75% NaOCI (=30% Clorox) was
necessary to allow easy removal of pericarp and testa layers prior to bombardment. Sterilized
half-seeds were imbibed overnight and prepared for particle bombardment as described above
for maize seeds.
Particle Bombardment
Particle bombardment was performed as described in Taylor and Vasil (1991) using a
DuPont PDS-1000 particle gun. Briefly, 35 to 50 pi of a sterile 50 mg/ml gold stock solution
(Biorad, 1.0 or 1.6 pm particle diameter; prepared in water) was mixed with premixed plasmid
DNA in a 250 pl-Eppendorf tube and vortexed on maximum speed for 10 s. Immediately, the
tube was shifted sideways and 10 pi of 0.1 M spermidine (free base) and 25 pi of 2.5 M CaCI2
were placed onto the side of the tube without allowing it to mix with the gold/DIMA solution. Then,
the tube was placed upright and subjected to vortexing for 10 s. The precipitated gold/DNA
particles were allowed to settle for ca. 3 min, after which part of the supernatant was removed
leaving 35-45 pi of liquid behind. The tubes were placed on ice until further use in particle
bombardment. For particle bombardment, 2 pi of sonicated gold/DNA solution (containing ca. 80
pg of gold) were used for individual shots.
The bombardment procedure had frequently to be adjusted to the gold characteristics
which varied substantially from batch to batch. Modifying the amount of gold used per shot was
found most successful in improving bombardment efficiency.


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
VP1-MEDIATED REPRESSION OF ALPHA-AMYLASE GENES
IN DEVELOPING MAIZE ALEURONE
By
Ute Hoecker
August 1995
Chairman: I.K. Vasil
Cochairman: D.R. McCarty
Major Department: Plant Molecular and Cellular Biology
The Viviparous-1 (VP1) transcriptional activator of maize is required for abscisic acid-induction
of maturation-specific genes late in seed development. In the presented work, it is shown that, in
addition, VP1 inhibits precocious induction of the germination-specific a-amylase genes in aleurone cells
of the developing seed. In developing seeds of the somatically unstable vp1-m2 mutant, hydrolase
activity was de-repressed specifically in endosperm sectors underlying vp1 mutant aleurone. Moreover,
in transient expression experiments based on particle bombardment of aleurone tissue, a barley high-pl
a-amylase promoter-GUS fusion construct (Amy-GUS) was induced in developing vp1 mutant aleurone
cells but not in wild-type aleurone cells. A direct role of VP1 in repression of Amy-GUS is suggested
from the finding that co-expression of recombinant VP1 in vp1 mutant aleurone cells strongly inhibited
expression of Amy-GUS. Hence, VP1 expression in the developing seed appears to integrate the control
of two developmental programs, seed maturation and seed germination.
Over-expression of VP1 also inhibited Amy-GUS expression in aleurones of wild-type
germinating maize and barley seeds. In barley aleurone cells, VP1 specifically repressed induction of
Amy-GUS by gibberellic acid (GA), while in maize aleurone tissue, VP1 inhibited a GA-dependent as well
IV


16
To investigate further the interaction of VP1 and ABA In controlling expression of the Em
gene in the maize seed, a transient gene expression system was used that is based on
electroporation of maize protoplasts isolated from an immature embryo-derived suspension
(Vasil et al., 1989). A plasmid containing the promoter (0.6 Kb) of the Em gene fused to the
coding sequence of the bacterial p-glucuronidase (GUS) gene (Em-GUS; Marcotte et al., 1988)
was used as a reporter construct. The VP1 cDNA was over-expressed from the constitutive
CaMV 35S promoter enhanced by insertion of the first intron of the maize Sh1 gene (Vasil et al.,
1989) into the 5' untranslated leader of the VP1 cDNA (35S-Sh-VP1). In these experiments,
electroporation of protoplasts with a mixture of Em-GUS and 35S-Sh-VP1 resulted in 100-300-
fold higher GUS activity as compared to the very low Em-GUS activity detected in the absence
of co-electroporated 35S-Sh-VP1 (McCarty et al., 1991). Similar activation was obtained when
protoplasts electroporated with Em-GUS were cultured in ABA (McCarty et al., 1991), which is
consistent with the reported ABA-regulation of Em-GUS in rice protoplasts (Marcotte et al., 1988;
1989). Over-expression of VP1 in maize protoplasts interacted synergistically with ABA,
resulting in 2,500-fold induction of Em-GUS (McCarty et al., 1991). The synergistic effect of VP1
and ABA underlines the importance of both, VP1 and ABA, in high-level expression of Em.
However, the substantial activation of Em-GUS obtained by either over-expressing VP1 or
culture in ABA might imply that VP1 and ABA can partially activate Em-GUS independently.
This would be in contrast to the absence of detectable Em transcript in vp1 or vp5 mutant
embryos which contain normal levels of ABA or Vp1 transcript, respectively (Neill et al., 1987;
McCarty et al., 1991). However, action of endogenous ABA and VP1 that may be present in the
wild-type protoplasts cannot be ruled out. To test this, Em-GUS was introduced into vp1 and vp5
mutant seed tissue (aleurone) via particle bombardment. In these experiments, ABA did not
activate Em-GUS in vp1A/p5 double mutant tissue while it did so in VP1/vp5 tissue or when co
bombarded with recombinant VP1 (S. Cocciolone and D. R. McCarty, unpublished results), thus
confirming that functional VP1 is required for ABA action. In vp5 mutant tissue, over-expression
of VP1 slightly activated Em-GUS, though 10-fold lower than in the presence of exogenous ABA.


3
Maya at al., 1990; Olshi and Bewley, 1990). It ha been suggested that the presence of ABA in
developing seeds is responsible for the inhibition of a-amylase genes at this developmental
stage (King, 1976). However, treatment of maize developing seeds with the ABA synthesis
inhibitor flouridone was not sufficient to sensitize the aleurone cells to GA, suggesting the action
of additional factors in repressing a-amylase genes in the developing seed (Oishi and Bewley,
1990).
The objective of this study was to elucidate a role of VP1 in the negative regulation of a-
amylase gene expression in the developing maize seed. It is demonstrated that VP1 in
addition to activating seed maturation programs blocks precocious induction of germination-
specific a-amylase genes in the developing maize seed. A somatically unstable vp1 mutant is
described that displays de-repression of hydrolase activity specifically in vp1 mutant sectors of
the aleurone. Using a transient expression approach, it is shown that expression of recombinant
VP1 in aleurone cells of maize and barley strongly inhibits expression of an a-amylase promoter-
GUS reporter gene (Amy-GUS). Evidence is provided indicating that VP1 specifically represses
GA-induction of Amy-GUS in aleurone of germinating barley seeds. It is also shown that deletion
of the acidic activation sequence of VP1 does not affect VP1 repressor activity, indicating that
VP1 has a discrete repressor function. Thus, it is suggested that the coupled activator and
repressor functions of VP1 play a key role in integrating the control of the normally not
simultaneously occurring maturation and germination programs in the seed.


33
The mechanism underlying the developmental switch from seed maturation to seed
germination, precisely the turning off* of maturation-related genes and the de-repression/
induction of germination-specific genes, is only poorly understood. Evidence on the molecular
nature of this switch is reviewed in the following for quiescent and dormant seeds.
Quiescent Seeds
Desiccation and subsequent rehydration of the seed appears to be the normal trigger to
switch the developmental program from maturation to germination (Comai and Harada, 1990).
Even when applied prematurely, drying resulted in the termination of maturation-related gene
expression (Oliver et al., 1993) and, upon imbibition of the dry seed, the induction of genes
specifically associated with germination (Kermode, 1990). Drying altered the developmental
potential of seeds such that a-amylase production became sensitive to GA (Evans et al., 1975;
Nicholls, 1979; Armstrong et al., 1982; Comford et al., 1986; Oishi and Bewley, 1990).
The nature of this switch in GA-sensitivity remains elusive. King (1976) has postulated
that accumulation of ABA in the developing seed prevents precocious induction of hydrolase
gene expression in the developing aleurone. Indeed, drying has been shown to cause a
concomitant decline in grain ABA content (McWha, 1975; King, 1976; Oishi and Bewtey, 1990).
Moreover, incubation of immature grains in buffer which caused a drop in endogenous ABA to
undetectable levels evoked GA-responsiveness of the aleurone (Napier et al., 1989). Hence,
depletion of endogenous ABA, by drying or washing, may be responsible for the induction of GA-
responsiveness. This is consistent with the finding that in maize mutants that are either deficient
for embryonic ABA (vp5) or insensitive to ABA (vpf), a-amylase activity was induced late in seed
development (Wilson et al., 1973).
However, results from Oishi and Bewley (1990) indicate that induction of a-amylase
synthesis as a result of drying is not solely due to a reduction in ABA content. The authors
compared the responses of maize kernels to premature drying and treatment with an ABA
biosynthesis inhibitor (flouridone) which reduces ABA contents in the seed to a similar extent as


23
ABA in a similar fashion as other cereal a-amylases. Ultimate proof for this, however, is still
lacking.
It has now been generally accepted that the embryo is the site of GA biosynthesis in the
germinating barley grain. When de-embryonated seeds are imbibed, no increase in GA levels in
the endosperm and very little subsequent production of a-amylase can be detected (Jacobsen
and Chandler, 1987). Further studies demonstrated that the scutellum, rather than the embryo
axis, is the source of GA (Radley, 1967; MacLeod and Palmer, 1967).
In maize, evidence for the importance of the germinating embryo as a source of GA has
been contradictory. In imbibed de-embryonated seeds, Dure (1960) found only ^-amylase
(which is stored in protein bodies in the dry seed and therefore is not de-novo synthesized during
germination), but no a-amylase activity, whereas whole kernels showed both activities. Two
other studies (Harvey and Oaks, 1974; Goldstein and Jennings, 1978), however, demonstrated
comparable total amylase activities in de-germed as well as whole seed endosperm. As only
part of the activity was due to release of fi-amylase from protein bodies, it was concluded that
mature seeds store considerable amounts of GAs in the endosperm. Therefore, the germinating
maize embryo does not appear to be an essential source of GA for a-amylase synthesis in the
aleurone.
The g-Amviase Genes
a-amylases of cereals can be biochemically separated into a number of isoforms that
differ in their isoelectric point (pi) but not considerably in their molecular weight. In barley, there
are two families of isozymes, the low-pi a-amylases with pis of ca. 4.4-5.2 and the high-pl a-
amylases with pis of ca. 5.7-6.2 (Jacobsen and Chandler, 1987). These two families differ in
many other biochemical characteristics while isozymes within those families are more alike.
Though some of the variants are post-translational modifications of the same gene product, a
genetic basis for most of the variation seen became evident when the gene(s) for the low and
high pi families were mapped to different chromosomes, chromosome 1 and 6, respectively


43
Amy32b (Lanahan et al., 1992). For its structure, see Fig 3. As effector construct. 35S-Sh-VP1
was used (McCarty et al., 1991). For no-VP1 control treatments, 35S-Sh-CAT (Vasil et al., 1989)
was added instead of 35S-Sh-VP1 to maintain a constant amount of total DNA and 35S promoter
in the bombardment mixtures. To normalize for transformation efficiency, a Ubiquitin-Luciferase
construct (Ubi-LUC; Bruce et al., 1989) was included into each bombardment mixture. Hence,
expression data are presented as Amy-GUS / Ubi-LUC or JR303 / Ubi-LUC ratios. It was
confirmed that co-expression of 35S-Sh-VP1 has no effect on expression of Ubi-LUC.
Construction of plasmids carrying an activation domain-deletion derivative of VP1 or a
replacement with Herpes simplex virus transcription factor VP16 activation domain was
described in McCarty et al. (1991). Internal deletion constructs were made by introduction of two
A/col restriction sites and subsequent deletion of the insert and religation of the backbone (Fig.
4). Nco\ sites were introduced by site-directed mutagenesis using the Altered Sites in vitro
Mutagenesis System from Promega. Briefly, mutant oligonucleotides and an ampicillin repair
oligonucleotide which restores the function of a defective ampicillin resistance gene in the
phagemid provided were annealed to single-stranded DNA template. Following DNA synthesis
and ligation, the resulting double-stranded phagemid was transformed into a repair-deficient
strain of E.coli which is subsequently grown in Ampiclllin-containing liquid medium for selection.
From the obtained bacterial suspension, plasmid DNA was isolated and transformed into an E.
coli strain conventionally used for transformations (JM83). Colonies growing on Ampicillin were
tested for the presence of the desired mutation by restriction enzyme digestion.
The constructs 86/87, 86/85, 85/87, 87/88, 93/95, 103/104, 101/100 and VP1-McW were
made available by L. Rosenkrans and D.R. McCarty.
Particle Bombardment and Transient Expression
Tissue preparation
Maize developing ears or dry, mature seeds were surface sterilized in 70% ethanol for 1
min followed by 0.525% NaOCI for 10 min. Dry seeds were germinated in a solution containing


I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fullyNadequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
Indra K. Vasil, Chair
Graduate Research' Professor of Plant
Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
Donald R. McCarty,/Cochair //
Associate Professor of Horticultural Science and
Plant Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
K&ren E'. Koch 7 ~
Professor of Horticultural Science and Plant
Molecular and Cellular Biology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation
for the degree of Doctor of Philosophy.
William B. Gurley /
Associate Professor of Microbiology and Cell
Science, and Plant Molecular and Cellular
Biology
I certify that I have read this study and that in
standards of scholarly presentation and is fully adequate,
for the degree of Doctor of Philosophy.
opinion it conforms to acceptable
n scope and quality, as a dissertation
Wl
Nigel G. Richards
Assistant Professo
Chemistry


72
CAT VP1 VP1* VPU VP1*
87/88 85/88 87/88-kb W
Fig. 23. Effect of the double deletion mutants 85/88 and 87/88:McW on inhibition of Amy-GUS
by over-expressed VP1. Materials and Methods as described in Fig. 22. The double mutants
85/88 and 87/88:McW were constructed by restriction enzyme digestion and subsequent ligation.
Data represent mean ( S.E.M.) of 5-6 replicates.
Fig. 24. Effect of the double deletion mutants 87/104 and 87/88:101/100 on inhibition of Amy-
GUS by over-expressed VP1. Materials and Methods as described in Fig. 22.


25
genes coding for a-amyiases have been doned so far. Thus no information about gene
expression is available.
The Organization of g-Amviase Promoters
Gene expression is thought to be regulated by proteins ffrans-ading factors*) that bind
in a sequence-spedfic manner to short stretches of base pairs f cfc-acting elements") located in
the promoter region of the gene. With an interest to study regulation of a-amylase gene
expression, genomic dones were isolated. Sequence comparisons revealed little homology
between promoters of a-amylase genes belonging to different pi groups which may relate to their
differential expression in response to hormones described above. However, a few blocks of
sequence were found highly conserved among barley, wheat and rice a-amyiase promoters:
High-pl, barley cgccttttgagctcaccgtaccggccgataacaaactccggccgacatatccactg -H7
(Khursheed and Rogers, 1988)
Low-pl, barley gcaccttttctcgtaacagagtctggtatccatgca -98
(Whittier et at., 1987)
Low-pl, wheat gcaccttttttcgtaacagagtctggtatccatgca -95
(Huttley et al 1992)
To identify c/s-ading elements involved in hormone-regulated expression, fundional
analyses of a-amylase promoters have been performed. For this purpose, mutated promoter
sequences are fused to a reporter gene (e.g. GUS, Ludferase) and are assayed for function in a
transient gene expression system (eledroporation of aleurone protoplasts or partide
bombardment of intad aleurone layers).
Progressive 5' truncations of a-amylase promoters showed that 289 bp of a wheat low-pl
a-amylase promoter (Huttley and Baulcombe, 1989) and 174 bp of a barley high-pl a-amylase
promoter (Jacobsen and Close, 1991) were suffident to dired GA- and ABA-regulated