Transgenic expression of the alcohol dehydrogenase-1 gene of maize in sunflower


Material Information

Transgenic expression of the alcohol dehydrogenase-1 gene of maize in sunflower
Physical Description:
vi, 161 leaves : ill., photos. ; 29 cm.
Ingersoll, John Charles, 1953-
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1990.
Includes bibliographical references (leaves 149-160).
Statement of Responsibility:
by John Charles Ingersoll.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001583935
notis - AHK7875
oclc - 23011928
System ID:

Full Text







Dedicated to Barbara Saffer


I wish to extend my thanks to fellow graduate students and colleagues who have

been helpful throughout my graduate career. Special thanks goes to my friend Jack

Shelton for always being available when I needed advice and/or a helping hand. Also to

my friends, co-workers, and roommates, Beth and Kevin O'Grady, for their help and

support. I especially wish to thank Dr. William Gurley who gave me a chance to learn and

develop in his laboratory.


ACKNOWLEDGMENTS....................................................................... iii

ABSTRACT ......................................................................................v


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

IN SUNFLOWER TUMORS .... .... ........ ............. ................19

Introduction..................................................................... 19
Materials and Methods............................................................... 21
Results......................................... ................................ 24
Discussion.................................................................. 45

SUNFLOWER TUMORS...................................................... 54

Introduction....................................................... ......... 54
Materials and Methods................................... .................... 55
Results.................................................................... 57

4. SUMMARY ......................................................................146

REFERENCES .................................................................................... 149

BIOGRAPHICAL SKETCH ............................................................. 161

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



John Charles Ingersoll

May, 1990

Chairman: William B. Gurley
Major Department: Microbiology and Cell Science

The Adhl gene from maize was incorporated into the genome of dicotyledonous

sunflower hypocotyl cells via Agrobacterium tumnefaciens mediated transfer. Transcription

of this foreign gene in sunflower tumors was induced by anaerobic stress in a similar

fashion as in its natural host. Splicing of the nine introns from the primary transcripts

proved to be aberrant in this heterologous system. Analysis of the non-processed mRNA

by S 1 nuclease hybridization indicated that initiation of transcription occurred at the

recognized Adhl start site and a possible second site 500 base pairs further upstream.

The 5' flanking sequences of Adhl were positioned upstream of a T-DNA reporter

gene (780) and an internal reference gene system was employed to analyze regulatory

regions within the promoter. The T-DNA sequences of the vector plasmid had a positive

effect on transcription when located upstream of the chimeric test genes. This positive

effect was overcome by positioning the shuttle vector T-DNA downstream of these genes.

Specific promoter mutations indicated that the anaerobic response element (ARE) of Adhl

was recognized and required for efficient transcription upon anaerobic induction in

sunflower tumors. Deletion of this element did not convert this inducible gene to a

constitutively active gene. Flanking the ARE with 8 (at -140) and 12 (at -91) base pair

insertions reduced the anaerobic response to -20% of wild type, while inverting the

orientation severely decreased the response to <5%. Re-positioning the ARE upstream

from its accustomed position abolished the response. The internal region of the promoter

(between -410 and -140) was responsible for 60% of the transcriptional activity when

positioned in its accustomed location in the promoter with respect to the ARE. When the

intermediate region was separated from the ARE by 14 base pairs, transcriptional activity

decreased almost 90%, while inverting this region increased activity 45% above wild type

activity. The distal region of the promoter (between -1094 and -410) contributed about

10% to total transcriptional activity while adversely influencing the TATA-proximal region

when the intermediate region was absent.


In recent years much attention has been focused on the interaction between

transcription factors and regulatory DNA elements of transcriptionally active plant genes.

A number of methods have been employed to identify the DNA sequences involved in

these interactions. These methods include band shift experiments (gel mobility retardation

assays) (Wu and Crothers, 1984), DNase I footprinting (Galus and Schmitz, 1978), and

in vivo dimethyl sulfate (DMS) genomic footprinting (Maxam and Gilbert, 1980). The

binding sites identified in this way have subsequently been characterized with regard to

their potential for transcriptional regulation. A more direct approach to the identification of

cis-regulatory elements has involved mutation of the 5' flanking region of the gene under

study. Important sequence elements that act as positive or negative regulatory elements

affecting transcription in a transgenic system may be identified by correlations between a

specific mutation and loss of promoter activity. Criteria for the regulatory analysis of a

gene requires a suitable transient or transgenic host and an accurate method of evaluating

activity of the gene, be it at the transcriptional or protein product level.

The relatively recent development of gene transfer techniques for plants has greatly

enhanced our ability to elucidate the regulatory mechanisms of a number of genes. The

different methods of gene transfer include microinjection of nucleic acid into protoplasts

(Crossway et al., 1986), electroporation into protoplasts (Fromm et al., 1985),

polyethelene glycol (PEG) induced uptake by protoplasts (Paskowski et al., 1984), DNA

coated microprojectiles (Klein et al., 1987, 1988), and Agrobacterium tumefaciens

mediated transformation into a wide range of dicotyledonous plants in addition to some

monocotyledonous species (Nester et al., 1984; Hooykaas et al., 1984; Binns and

Thomashaw, 1988, for review). The analysis of promoter structure in heterologous plant

systems may utilize the natural transcription unit, or may be facilitated by fusing the

promoter to one of a number of reporter genes, whose expression may be easily

monitored and quantified (Kuhlemeier et al., 1987, for review). These reporter genes

include the coding regions for two Agrobacterium T-DNA genes, octopine synthase (ocs)

and nopoline synthase (nos) (Kemp, 1982; Tempe and Goldman, 1982), and the coding

regions for other bacterial genes such as neomycin phosphotransferase (NPT) (Fraley et

al., 1983), chloramphenicol acetyl transferase (CAT) (Gorman et al., 1982; An, 1986),
and P-glucuronidase (GUS) (Jefferson, 1987; Jefferson et al., 1987).

This review focuses on the experimental investigation of a representative sampling

of differently regulated plant gene promoters and the analysis of some monocotyledonous

gene promoters in dicotyledonous plant systems. Important experiments involving the

localization and elucidation of important cis-acting sequences of a transcription unit

requires the investigation of the expression strengths delineated by in vitro mutated

promoters. Different plant systems at the transient or transgenic levels are employed for

these investigations. The promoters discussed in this survey represent some constitutively

active plant gene promoters and cis-acting regulatory sequences responsible for the

induced expression of a gene by environmental factors. Regeneration of transgenic plants

also allow for analysis of those regions responsible for regulating organ specific


Constitutive Regulation

The Constitutively Regulated Cauliflower Mosaic Virus (CaMV) 35S Promoter.

The promoter of the CaMV 35S transcription unit exemplifies a particularly well

studied example of promoter analysis in plant systems. This promoter has been

commonly used in transgenic plant experiments due to its ability to function in a wide

variety of plant systems. It is highly active in transient assays using cell lines from both

dicotyledenous and monocotyledenous plant species (Fromm et al., 1985; Ou-Lee et al.,

1986; Nagata et al., 1987; Dekeyser et al., 1989) and is constitutively expressed in

regenerated transgenic plants (Koziel et al., 1984; Odell et al., 1985; Shah et al., 1986).

To localize important cis-regulatory regions within the 35S promoter, deletion mutation

experiments have been performed with subsequent analysis of gene products. Odell et al.

(1985) fused the 35S promoter to the coding region of a human growth factor gene and by

monitoring mRNA levels in tobacco call showed that this promoter was fully active when

the 5' flanking sequences of this chimeric gene extended 343 base pairs upstream from the

transcription initiation site. Gene products from this construct were observed in roots,

petals, and stems of transgenic tobacco plants (Odell et al., 1985). Further analysis of the

promoter in tobacco calli involved additional 5' deletions to position -105 and -46 which

reduced promoter activity 67% and 95%, respectively. This group also showed that the

promoter sequences extending 46 bp upstream from the transcript start site, which

includes the TATA box, were sufficient to initiate transcription at the start site normally

utilized in the transcription of the intact CaMV 35S gene. Thus the promoter sequences

situated between -343 and -46 were shown to be necessary for full transcriptional activity

in tobacco and transcription did not appear to be tissue specific (Odell et al., 1985).

Similar deletion experiments of the CaMV 35S promoter when fused to the coding region

of a firefly luciferase gene were constructed and analyzed in carrot protoplasts (Ow et al.,

1987). The results of these experiments showed that the promoter was fully active when

it extended upstream only 148 bp from the CAP site. However, an additional 40 bp

deletion to -108 reduced activity 80%. This lower transcript level was not altered by

additional 5' deletions to -104 and -89. A further deletion of 16 base pairs to -73 did

result in a reduced level of expression of the chimeric gene to less than 1% of the

maximum expression (Ow et al., 1987).

The CaMV 35S promoter deletion studies in calli and protoplasts provided

information that enabled investigators to focus their attention on specific regions of the

promoter. The CaMV 35S promoter is on average 30-fold more transcriptionally active

than the A. tumefaciens T-DNA nopaline synthase (nos) promoter in transgenic petunia

(Sanders et al., 1987). Sequences previously shown to be important, between -392 and

-55 of the 35S promoter were positioned directly upstream of a truncated (to position

-150) nos promoter fused to a CAT reporter gene (Odell et al., 1988). The addition of

these sequences, in both orientations, resulted in increased expression of the truncated

chimeric gene in tobacco, tobacco calli, and soybean protoplasts, to levels equal to the

"wild type" 35S promoter (-392 to +1) (Odell et al., 1988). These results indicate that this

region contains promoter elements that can function in association with a heterologous

promoter in a bidirectional manner. This enhancer-like effect is position dependent in that

an inverse correlation was observed between distance from the TATA box and

transcriptional activity (Odell et al., 1988). The promoter region between -392 to -90 of

the 35S gene has also been shown to stimulate transcription in maize protoplasts when

placed upstream of the maize alcohol dehydrogenase 1 (Adhl) promoter truncated to the

-140 position (Ellis et al., 1987). These sequences also stimulate an increase in

transcription of two T-DNA genes (genes 5 and 7; Willmitzer et al., 1982) in transgenic

tobacco plants and transcriptional activity increases 10-fold when this region is duplicated

in tandem within the natural 35S promoter positioned upstream of NPT II coding

sequences (Kay et al., 1987).

An additional study in transgenic tobacco leaves analyzing the effects of 5' and

internal deletions of the 35S promoter localized three regions necessary for maximum

transcription efficiency (Fang et al., 1989). Whereas Odell et al. (1985) did not observe

any significant drop in transcription upon deletion of sequences between -343 and -168 in

tobacco, Fang et al. (1989) found that the region between -343 and -208 contributed 50%

of the activity to the downstream promoter sequences. A second region, -208 to -90, was

found to contribute 40% of the activity to downstream sequences and was enhancer-like in

its activity. A third region between -90 and -46 of the 35S promoter (which does not

include the TATA box) did not contribute to promoter activity when sequences 5' to this

region were absent, but its presence, in either orientation, in association with the upstream

sequences allowed for increased transcriptional activity (Fang et al., 1989).
Histochemical localization of different 35S promoter mutations fused to a P-

glucuronidase (GUS) reporter gene was used to identify two domains of the 35S

promoter responsible for tissue specificity not previously observed in regenerated tobacco

(Benfey et al., 1989). Domain A (-90 to +8), which does not contribute to increased

transcriptional expression in tobacco leaves (Fang et al., 1989), confers strong expression

to roots. It was observed that nuclear protein factors bind to the DNA within this region

between -90 and -59 (Benfey and Chua, 1989). Domain B (-343 to -90) is not active in

roots, but is mainly responsible for transcriptional activity in vasculated tissues such as

leaves, cotyledons, and hypocotyls (Benfey et al., 1989). These two promoter domains

work best when associated with each other and are both necessary for constitutive

expression. Analysis of mutations identified cis-elements which in turn became targets for

subsequent studies that mapped protein:DNA interactions. DNase I footprinting

experiments performed on the sequences -90 to -46 localized the binding sites for the

activation sequence factor I (ASF-I) between -85 and -58 (Lam et al., 1989).

The experiments utilized to characterize the CaMV 35S promoter illustrates the

contribution made by specific plant gene promoter mutations and their analysis in

transgenic plant systems to target and identify important regulatory regions.

Inducible Genes

Expression of Thermally Induced Plant Genes.

Different plant systems have been exploited to analyze the regulated expression of

environmentally inducible genes such as heat shock, wounding, light, and anaerobic

stress (Benfey and Chua, 1989, for review). Examples of this type of investigation

include the transformation of soybean heat shock (HS) genes in heterologous sunflower

and tobacco systems. Expression of the HS multigene family in soybean may be

necessary for survival of the plant when subjected to thermal stress (Key et al., 1987, for

review). In an effort to delineate regulatory elements of a soybean HS gene, Gmhspl7.3-

B has been incorporated into sunflower seedling hypocotyl tissue via Agrobacterium

mediated transfer (Schoffl and Baumann 1985). The transferred DNA contained the

protein coding region and about 1 kb of 5' flanking sequences. In sunflower tumor tissue

Gmhspl7.3-B was shown to be transcriptionally expressed upon thermoinduction,

though at very low levels compared to soybean. A number of 5' deletion mutations of the

Gmhspl7.3-B promoter, the longest extending 439 bp upstream of the start site, were

then constructed in vitro. Expression of each deletion was analyzed in regenerated

tobacco plants (Baumann et al., 1987). In this heterologous transgenic system, those

deletion mutants that were expressed were induced at the mRNA level by heat shock and

utilized the correct transcription initiation site. Sequences within the promoter between

180 and 154 bp upstream of the translation start were found capable of conveying thermo-

inducibility upon this gene. This region contains overlapping heat shock elements (HSEs)

first described in HS genes of Drosophila (for review see Pelham, 1985) which show a

common spacing and location in relation to start of transcription in all soybean heat shock

genes thus far described (Raschke et al., 1988). Baumann et al. (1987) also determined

that distal HSE-like sequences between -298 and -181 (relative to the translation start

codon) also make a positive contribution to thermoinducibility. An additional sequence

element which was AT-rich was identified between -439 and -298. This element was

shown to be capable of bidirectional activity when placed upstream of the truncated

Gmhspl 7.5-B promoter.

Another soybean HS gene, Gmhspl7.5-E, has also been incorporated into

sunflower tumors via Agrobacterium transfer (Gurley et al., 1986). The transgenic tumor

tissue was heat shocked at 400C whereupon transcription levels increase at least 10-fold

from a copy of the gene containing 3,250 bp of upstream sequences. Mutational analysis

of the promoter by 5' deletions revealed that sequences between -1175 and -95 were

necessary for maximal induction of the gene (Gurley et al., 1986). The identification of

important regions of the Gmhspl7.5-E promoter by these initial deletions allowed for a

more specific mutational analysis. Additional manipulations of the promoter by 5' and

internal deletions revealed that the distal region of the promoter between -1175 and -259

donates a modest 25% activity to the transcriptional activity. Further removal of

sequences to -179 decreased transcription to 58% compared to the promoter extending

1175 bp upstream (Czarecka et al., 1989). Both 5' and internal deletion mutants enabled

identification of two regions near the TATA box, site 1 (-49 to -72) and site 2 (-81 to

-102), which harbor HSEs and are involved in the transcription response to heat stress.

Removal of either of these two regions abolished thermoinducibility The sequences

between -274 and -189 contributed 66% of the induced transcriptional expression of the

gene (Czarnecka et al., 1989). Within these sequences is a TATA dyad

(TATAAAGAATITC) that is conserved in both Drosophila and soybean heat shock genes

and may also contribute thermoinducibility.

Heat shock genes appear to be ubiquitous (see Schlesinger et al., 1982) and have

similar promoter elements (Pelham, 1985, for review) among eucaryotes. It is interesting

to note that the Drosophila hsp70 promoter (extending upstream 258 bp from the

transcription initiation site) when fused to the coding region of a neomycin

phosphotransferase II gene proved to be thermally inducible in tobacco calli (Spena et al.,

1985). Similarly, upon regeneration of the call into transgenic tobacco plants the chimeric

gene was induced by heat shock in roots, stems, and leaves, but was not found expressed

in pollen (Spena and Schell, 1987).

Expression of a Wound-Inducible Plant Gene.

The proteinase inhibitor II (PI II) gene of potato is a member of a small gene

family that is normally expressed in the tubers of the plant in a tissue specific manner

(Sanchez-Serrano et al., 1986) but is induced by mechanical wounding or insect attack in

leaves (Green and Ryan, 1972). To investigate cis-regulatory elements of this gene, 5'

and 3' flanking sequences were fused to the CAT gene coding region and transformed into

tobacco (Thornburg et al., 1987). A 7.8-fold increase in CAT activity was detected as

compared to basal levels when the leaves were subjected to mechanical wounding. When

the 3' flanking sequences of the T-DNA 6b gene were substututed for those of PI I for

this chimeric construct no activity was observed. Therefore it was determined that the 3'

flanking sequences were required for the wound response of this gene in tobacco

(Thomburg et al., 1987). Transcripts of a potato PI II gene were systemically induced

and accurately spliced in tobacco plants (Johnson et al., 1989) in a similar manner as in its

natural host (Sanchez-Serrano et al., 1987). Induction levels were strongest within the

region of vascular tissue in tobacco as it is in potato (Keil et al., 1989). Information

gathered by these experiments suggest that specific in vitro mutations of regulatory

regions and analysis in tobacco may now be utilized as a means to identify important

sequences of these genes.

Light Regulated Plant Genes.

The nuclear genes of pea encoding the small subunit of ribulose-1,5 bisphosphate

(RuBP) carboxylase (Coruzzi et al., 1984) represent examples of plant genes that are

transcriptionally regulated by the presence or absence of light. Different plant systems

have been utilized to identify cis-acting regulatory elements of these genes. The pea rbcS-

E9 gene has been transformed into petunia protoplasts which were subsequently grown as

calli (Broglie et al., 1984). A construct containing 1052 bp of sequences 5' to its start site

was transcribed in a similar fashion as in its natural host. Though less than in pea leaves,

transcription in petunia was shown to be induced up to 50-fold in calli upon exposure to

light. The resulting transcripts were correctly initiated, polyadenylated, and efficiently

spliced removing the two intervening introns of the immature message (Broglie et al.,

1984). The promoter of another light- regulated pea gene of the rbcS gene family, ss3.6

(Cashmore, 1983), was fused to a bacterial CAT reporter gene and transformed into

tobacco calli (Herrera-Estrella et al., 1984). This light regulated promoter extended from

-4 to -973 bp relative to the transcription start site. The promoter of ss3.6 was also shown

to regulate transcription in soybean callus tissue (Facciotti et al., 1985). Thus, the cis-

regulatory elements of these light regulated genes from pea were independently recognized

by nuclear factors of both tobacco and soybean.

Once the efficiency of transgenic expression was demonstrated for these light

regulated genes, mutational analyses were employed to identify and localize important

regulatory sequences. The promoter of the pea rbcS-E9 gene was subjected to a number

of 5' deletions and one internal deletion. Transgenic calli tissue harboring each construct

were pooled and the transcripts analyzed (Morelli et al., 1985). In this manner it was

shown that sequences between -1052 and -352 are necessary for maximum transcriptional

activity and that sequences involved in light regulation are localized to the TATA box

region from -35 to -2. Similar manipulations were performed on the pea rbcS ss3.6 gene

promoter situated upstream of the CAT reporter gene and transformed into tobacco tumor

tissue (Timko et al., 1985). The promoter sequences from -973 to -90 were found to

convey maximum transcription levels and possessed the photoregulatory information

regardless of orientation. This ability to respond to light was also seen when the 5'

flanking sequences were fused to a heterologous nopaline synthase (nos) promoter and

assayed in tobacco cells. This upstream fragment did not, however, influence

transcription when placed 3' to the coding region of a chimeric nos/CAT construct (Timko

et al., 1985).

The light inducible rbcS-E9 gene with a promoter extending 1,052 bp upstream

from the start of transcription was induced in the presence of light in transgenic petunia

and tobacco plants with most of the activity in leaves which is similar to its expression in

pea (Nagy et al., 1985). Two additional members of the rbcS multigene family of pea,

rbcS-3A and rbcS-3C, were transferred into petunia plants and assayed for light induced

expression (Fluhr et al., 1986b; Fluhr and Chua 1986). In petunia the pea rbcS-3A gene

needed only 410 bp of its 5' flanking sequence to be efficiently transcribed with light-

mediated phytochrome control in an organ specific manner (most highly expressed in

leaves). The sequence requirement for maximum transcription efficiency in regenerated

plants differs from that observed for rbcS-E9 in petunia calli (Morelli et al., 1985). The

-327 to -48 region of the rbcS-3A promoter in regenerated petunia plants were defined as

enhancer-like in that they convey elevated, light inducible and organ specific transcription

levels when placed upstream of a truncated heterologous CaMV 35S promoter regardless

of orientation. Similar results were observed when the region -317 to -82 of the closely

related pea rbcS-E9 gene was placed in both orientations upstream of a truncated nos

promoter and analyzed in transgenic petunia plants (Fluhr et al., 1986a). Regulatory

expression levels of pea rbcS gene promoters were further investigated by additional in

vitro manipulations and analysis in transgenic tobacco. RbcS-3A and rbcS-E9 are closely

related genes, but rbcS-3A is expressed to a much higher level in pea. Sequences between

-50 and +30, including the TATA box, were physically switched between the two genes

by cloning procedures. In tobacco each chimeric was transcriptionally expressed at levels

corresponding to activity obtained with the homologous TATA region (-50 to +30)

(Kuhlemeier et al., 1988) inferring that these sequences are not responsible for the

difference in transcription levels. Sequences extending from -330 to -50 of the rbcS-3A

gene were then inserted in the same position upstream of an rbcS-E9 5' truncated deletion

to -50. This chimeric promoter resulted in high levels of induced expression similar to

that of rbcS-3A. These two experiments indicate that the amplitude of expression is

determined by sequences upstream of -50 (Kuhlemeier et al., 1988). These same pea

rbcS-3A promoter sequences, between -330 and -50, also confer cell specificity of

expression in transgenic tobacco (Aoyagi et al., 1988). Similar sequence block exchange

experiments were performed between the highly expressed petunia rbcS gene SSU301

and the weakly transcribed SSU911 in transgenic tobacco plants. In this case, promoter

sequences located between -285 and -204 of SSU301 were shown to convey the

difference in expression levels (Dean et al., 1989).

The pea rbcS-3A promoter had also been examined by a number of 5' and

internal promoter deletions when fused to a CAT reporter gene and transformed into

tobacco plants (Kuhlemeier et al., 1987a). A promoter extending upstream to the -169

position was induced by light but with 3- to 5-fold less expression than the promoter of a

similar chimeric that extended to -410, whereas a 5' deletion to -149 inhibited light

induced transcription. Thus, a positive light response element (LRE) was mapped

between positions -166 and -100. This region of rbcS-3A contains three domains,

designated boxes I (-156 to -162), II (-130 to -151), and III (-114 to -125) which are also

found within the promoter of other photoregulated pea rbcS genes. Results from fusing

this region upstream of a truncated CaMV 35S promoter/CAT coding region construct,

followed by transformation and regeneration of tobacco plants, indicated the presence of a

negative LRE which suppresses transcription of this chimeric gene in the dark

(Kuhlemeier et al., 1987a).

General repositioning and deletion studies of cis-acting regulatory elements of the

light regulated pea gene promoters by in vitro mutations and analysis in transgenic plants

localized regions within the promoter which allowed for a more fine-tuned analysis by

more specific promoter mutation experiments. A 5' deletion into box II of the rbcS-3A

promoter (to -149) drastically reduced light regulated expression in tobacco plants

(Kuhlemeier et al., 1987a). This promoter with sequences extending upstream to -175

conveyed light regulation, but point mutations within box II abolished transcription and

mutations within box III reduced transcription by 95% in transgenic tobacco (Kuhlemeier

et al., 1988). A second construct possessing the pea rbcS-3A promoter truncated to -410

expresses light regulated gene activity when boxes II and III are internally deleted

(Kuhlemeier et al., 1987a). This paradox can be explained by the presence of redundant

sequences of boxes II and III, designated boxes II* and III*, which are found within the

-220 region of the promoter. A pea nuclear transcription factor, GT-1, binds to boxes II

and III and to the redundant boxes II* and III* (Green et al., 1987). Results of these

experiments indicate the same transcription factor binds to either set of sequences within

the pea rbcS-3A promoter and deletion or mutation of each region separately does not

disrupt light regulation, but deletion of both simultaneously abolishes expression in

transgenic tobacco. Although the pea rbcS-3A gene has two sets of positive LREs within

its promoter (Kuhlemeier et al., 1988), sequences within the region harboring boxes II

and III are also responsible for repressing transcription of the constitutive CaMV 35S

promoter in the dark and therefore possess a negative LRE (Kuhlemeier et al., 1987a).

Monocotyledonous Gene Promoters

Expression of Monocotyledonous Promoters in Dicotyledonous Plants.

A number of monocot genes have been transferred and analyzed in different

dicotyledonous plant systems. Among these, an oat phytochrome gene was incorporated

into regenerated tobacco plants where it was determined that its own light repressible

promoter could not function (Keller et al., 1989). Only when this gene was placed under

the control of the constitutive CaMV 35S promoter were mRNA and proteins detected.

Similarly, the rbcS gene of wheat could not utilize its own promoter in transgenic tobacco

where again transcription products were only observed when the coding sequence was

fused to CaMV 35S promoter (Keith and Chua, 1986). Alternatively, CAT activity was

observed in transgenic tobacco in an endosperm specific manner when the promoters of

two wheat endosperm protein genes, low molecular weight (LMW) and high molecular

weight (HMW) glutenin, were fused to a CAT reporter gene and incorporated into tobacco

(Colot et al., 1987). In addition, a light inducible and tissue specific cab (chlorophyll a/b

binding protein) gene from wheat was accurately regulated and transcribed in transgenic

tobacco and petunia plants (Lamppa et al., 1985).

The promoter of the wheat cab-1 gene having 1.8 kb of the 5' flanking sequences

fused to CAT coding sequences was shown to confer light and organ specific expression

of this chimeric in transgenic tobacco plants (Nagy et al., 1986). Promoter mutations by

in vitro 5' deletions and subsequent transfer into dicotyledonous tobacco resulted in

comparable expression of chimerics having either 1.8 kb or 357 bp of 5' flanking

sequences, whereas further deletion to -124 resulted in undetectable transcript levels

(Nagy et al., 1987). Further analysis of the wheat cab-1 promoter sequences identified a

region between -357 and -89 as conveying enhancer-like activity upon the heterologous

CaMV 35S promoter in transgenic tobacco (Nagy et al., 1987). Another regulated

monocot gene that is inducible in dicots utilizing its own promoter is the hsp70 gene of

maize in transgenic petunia. Upon thermoinduction this gene is transcribed 40- to 60-fold

higher than in transgenic plants not placed under heat stress (Rochester et al., 1986).

Proper induction and expression of maize hsp70 in petunia is not surprising since the heat

shock genes appear to be highly conserved between eucaryotes as was seen with

expression of the Drosophila hsp70 promoter in tobacco (Spena and Schell, 1987). It is

apparent that the promoters of different monocot genes may be recognized by some

transgenic dicots but not by others, and in those cases where expression is seen, the

heterologous system may be exploited to identify important regulatory elements within

these promoters.

Maize zein genes code for endosperm specific storage proteins that are produced

during seed development (Higgins, 1984). This class of genes consists of about 100

copies per haploid genome and comprise several multigene families (Hagan and

Rubenstein, 1981). Early studies concerned with the cis-acting region of zein genes have

taken advantage of the fact that when members of this gene family are transferred into a

dissimilar plant system by A. tumefaciens, they are fortuitously expressed in the crown

gall tumor cells. An advantage of tumor expression studies is its avoidance of lengthy

periods needed for regeneration of transgenic plants. A second advantage is the

reproducibility obtained by pooling large numbers of tumors. Promoter mutation analysis

in pooled transformed primary tumor tissue can be performed with little concern to

position effect as seen in transgenic plants (Kuhlemeir et al., 1987; Willmitzer, 1988, for

reviews). Data gathered from these types of experiments which map promoter elements

may then be correlated with regions of the promoter identified by in vitro DNA binding

assays. This avenue of experimentation has been utilized in the investigation of the

monocotyledonous maize zein genes in Agrobacterium transformed dicots. Matske et al.,

(1984) transformed sunflower stemlet hypocotyl tissue with the zein gene (Z4) of maize

using an A. tumefaciens based vector. This highly regulated gene was transcribed at low

levels in sunflower tumor cells utilizing the same two transcription start sites recognized in

maize endosperm, but no protein products were detectable (Matzke et al., 1984). Two

additional zein genes were similarly incorporated into sunflower plantlets where again

each was transcribed at low levels utilizing the recognized transcription start and

termination sites in cloned tumor tissue (Goldbrough et al., 1986). Again, no protein was

detected (Goldbrough et al., 1986). Rousell et al., (1988) analyzed the upstream flanking

region of a zein gene (gZl9abl) with a number of 5' deletions fused to a CAT reporter

gene in sunflower tumor tissue. They reported that promoter sequences between -337 and

-125 relative to the CAP site were necessary for efficient transcription, whereas deletions

within 125 bp upstream of the CAP site greatly reduced transcriptional activity. The

assignment of a cis-element to this region of the zein promoter correlated with the

identification of a 22 bp protein binding site centered at -320. The binding site had been

identified by in vitro DNA-protein interactions utilizing maize endosperm nuclear extracts

(Maier et al., 1987). A region of 15 bp was identified as being a conserved promoter

sequence similarly located in all zein genes (Brown et al., 1987). From these results a

correlation can be seen between promoter deletion experiments of a monocot gene in dicot

tumor tissue and the factor binding assays with homologous nuclear extracts. This

correlation suggests that analogues of nuclear factors responsible for endosperm-specific

expression of zein in maize are present in dicots where they presumably regulate unrelated

gene families.

The Anaerobically Induced Maize Alcohol Dehydrogenase-1 (Adhl) Gene of Maize.

When a maize plant is subjected to anaerobic stress there is a switch from oxidative

respiration to ethanolic fermentation (for reviews see Dennis et al., 1987; Kloeckener-

Guissen and Freeling, 1987). Activity of the aerobic proteins decline in maize roots and

within one hour of hypoxia there is an accumulation of transitional proteins. The

abundance of these transitional proteins quickly declines with the subsequent accumulation

of 10 major and 10 minor anaerobic proteins (ANPs) which make up over 70% of the total

protein in the root (Sachs et al., 1980). The ANPs of the major class thus far identified

are associated with the glycolytic pathway. These include sucrose synthase (Springer et

al., 1986), pyruvate decarboxylase (Laszlo and St. Lawrence, 1983), glucose phosphate

isomerase (Kelley and Freeling, 1984a), fructose 1,6-diphosphate aldolase (Kelley and

Freeling, 1984b), and alcohol dehydrogenase (Freeling, 1973). There are two alcohol

dehydrogenase (Adh) genes of maize, Adhl and Adh2. Adhl is located in single copy on

the long arm of chromosome 1 and Adh2 in single copy on the short arm of chromosome

4 (Schwartz, 1971). The transcription products of each gene encodes for a polypeptide

which combine to form either enzymatically active homodimers (ADH1-ADH1 or ADH2-

ADH2) or heterodimer (ADH1-ADH2) (Freeling, 1974; Ferl et al, 1979).

The expression of Adhl is essential for the survival of a maize plant subjected to

extended periods of hypoxia. Regulation of the Adhl gene is at the mRNA level (Ferl et

al., 1980; Rowland and Strommer, 1986). Within two hours of anaerobic stress it is

transcriptionally expressed in roots. When plants are continuously stressed, transcript

levels increase 50-fold after about 5 hours and remain at steady state levels until plant

death (Gerlach et al., 1982, 1983; for review see Freeling and Bennett, 1985). ADH1 is

not seen in leaves (Okimoto et al., 1980) even after anaerobic stress, but is present in

pollen and the scutellum within maize kernels in both aerobic and anaerobic environments

(Freeling and Schwartz, 1973). The Adhl-S gene (S refers to slow electrophoretic

mobility) encodes a 3.4 kb transcript possessing nine A-T rich introns which are spliced to

form a final 1,650 bp mRNA (Dennis et al., 1984). This gene has a TATA box and a

putative CAAT box common to many eukaryotic genes (Dynan and Tjian, 1985).
Transgenic plant expression systems have been utilized in an attempt to identify

regulatory elements responsible for the anaerobic induction and transcript stability of

Adhl. Llewellyn et al. (1985) utilized Agrobacterium tunefaciens to transfer maize Adhl

into Nicotiana tobacum and Nicotiana plubagimfolia. The transferred Adhl gene

possessed 1.1 kb of 5' flanking sequences, its coding region, and about 3.0 kb of 3'

flanking sequences. Negligible expression of this monocot gene was observed in tumor

tissue of tobacco that had been subjected to anaerobic stress by argon for four days. This

same promoter, including its untranslated leader sequence, was fused to a CAT reporter

gene and transformed tobacco plants (Nicotiana tobacum) were regenerated (Ellis et al.,

1987). Again, the Adhl promoter showed only minimal expression in tobacco after 18

hours of anaerobic stress. Upstream sequences possessing enhancer-like activity from

either the octopine synthase (ocs) gene or the CaMV 35S gene were then placed upstream

of the -140 position of the Adhl/CAT fusion construct. Upon anaerobic induction, CAT

activity could be detected in transformed tobacco with transcription initiating at the proper

start site. Thus, the sequences between -140 and +106 of Adhl were identified as

harboring the regulatory region responsible for the anaerobic response. This response

element is weakly recognized in transgenic tobacco plants and only upon addition of an

upstream enhancer-like element could the induced response mimic that of Adhl in maize

(Peacock et al., 1987).

The entire maize Adhl gene was transformed into maize protoplasts by

electroporation where it was efficiently expressed at both transcription and protein levels

(Callis et al., 1987). An Adhl promoter/CAT coding region construct has also been

electroporated into maize protoplasts and activity determined (Howard et al., 1987)

whereupon CAT activity was observed upon anaerobiosis in this homologous transgenic

system. While these early studies merely demonstrated activity of Adhl in transgenic

systems, Walker et al. (1987) enlisted this homologous transgenic system to perform a 5'

deletion series study on the Adhl promoter. The promoter deletion extending 1094 bp

upstream from the transcription initiation site as previously observed by Howard et al.,

(1987) allowed for increased levels of transcription upon anaerobic induction. The 5'

deletion to the -140 site was still induced by anaerobiosis but the quantity of the gene

products were reduced 25%. Further deletions to -124 and -112 resulted in anaerobic

induction, but to a lesser extent than the two previous deletions with transcriptional

expression at very low levels. A final deletion to -99 eliminated both anaerobic induction

and constitutive expression of the gene. Lee et al., (1987) performed a similar set of 5'

deletions of the Adhl promoter fused to a CAT reporter gene and transformed into maize

protoplasts. Their 5' deletion series included 3 constructs extending 1,094 bp, 410 bp,

and 140 bp upstream of the initiation site. The -410 deletion decreased expression of this

chimeric 35% with the further deletion to -140 resulting in 18% activity compared to the

-1094 promoter. This last value differs from the 75% activity of the same construct

observed by Walker et al. (1987). Linker scan experiments were performed through the

Adhl promoter which identified a region (-140 to -99) responsible for increased

transcription levels upon anaerobic induction (Walker et al., 1987). The anaerobic

response element (ARE) is subdivided into two functional domains, regions I (-140 to

-133) and II (-133 to -99).

Representative examples of constitutive and inducible plant gene promoters and

their analysis in different plant systems have been presented. Data gathered by these

experiments have been valuable in identifying DNA sequences that comprise promoter

elements that mediate regulated transcriptional control of plant genes. The ability of some

monocotyledonous gene promoters to be recognized and properly expressed in

dicotyledonous plant systems, particularly the analysis of the cis-acting regulation of

maize zein genes in sunflower tumors suggests that similar experiments were feasible with

the maize Adhl gene. The present study entails the transformation of maize Adhl into

sunflower seedling hypocotyls via an Agrobacterium vector and characterization of

transgenic expression in response to anaerobic stress. The objective was to detect

transcription mediated by this foreign gene in dicotyledonous sunflower and investigate

any regulatory similarities between maize and sunflower. A comparison was made

regarding the ability of the gene to be induced by anaerobiosis, efficiency of expression,

and the length of time required for maximum expression. The second phase of this project

localized important regulatory regions of the promoter recognized in sunflower by in vitro

mutagenesis of the promoter and analysis of transcription in response to anaerobic stress.



Agrobacterium tumefaciens is a Gram negative bacterium that has the ability to

integrate a specific portion (T-DNA) of its tumor inducing plasmid (pTi) into the genome

of a wide variety of dicotyledonous plants (Hooykaas et al., 1984; Nester et al., 1984;

Binns and Thomashow, 1989; for review). This natural gene transfer system between

A.tumefaciens and higher plants has been exploited as a means of introducing foreign

DNA into plant cells (Matzke and Chilton, 1981; Caplan et al., 1983). The experimental

strategy often involves the insertion of the cloned gene of interest into an intermediate

shuttle vector which is used to transfer genes from Escherichia coli to Agrobacterium by

triparental conjugation (Comai et al., 1983; Fraley et al., 1983) where it is incorporated

into the T-DNA by cointegration via homologous recombination. The modified T-DNA is

transferred to the plant cell by Agrobacterium and subsequently incorporated into the


Transgenic expression of a number of plant genes has been obtained using T-DNA

based vector systems. In most cases the foreign gene is transcriptionally expressed in the

heterologous plant system in a manner that closely resembles its expression in the original

plant. There are numerous cases illustrating this faithfulness of transcription of genes in

different plant species. One example of the fidelity of transgenic expression is seen in the

tissue and developmentally specific expression of the seed storage gene for phaseolin from

the French Broad Bean (Phaseolus vulgaris) in transformed tobacco seedlings

(Sengupta-Gopalan et al., 1985). In this case, transcript levels for phaseolin were 1000-

fold lower in the leaves and stem, but quite abundant in seeds where it is naturally

expressed in Phaseolus.

T-DNA based vector systems have been used to transfer genes from

monocotyledonous plants into the genomes of dicotyledonous hosts. Although

Agrobacterium predominantly infects dicotyledonous plants, there is no a priori reason to

expect that individual genes from these two plant groups will exhibit fundamentally

different mechanisms for the regulation of gene expression at the molecular level. The

best example of a monocot gene functioning in a dicot is seen with the wheat gene

whAB1.6 (Lamppa et al., 1985) which encodes the chlorophyll a/b binding protein.

Expression of this gene proved to be both light regulated and organ-specific in regenerated

plants of both tobacco and petunia. Another example of a monocot gene transferred into a

dicot host is a zein gene of maize in sunflower tumors (Matzke et al., 1984). Low level

transcriptional expression of this highly regulated gene was seen to occur in tumor tissue

cloned from a single transformed cell.

The use of tumor tissue (or tumor-derived callus cultures) to study heterologous

expression of plant genes is especially well suited for genes showing little, if any, tissue

or developmental specificity in their regulation. Tumors may also serve as an

experimental system to characterize the inducible aspects of heterologous gene expression

separately from tissue and developmental specificity. An example of a gene that may

possess both tissue specific and general inducibility is maize Adhl. In maize this gene is

transcriptionally inducible by anaerobic stress, such as flooding (Freeling, 1973), in a

tissue specific manner. In addition, its gene products are present at constitutive levels in

the scutellum of the kernel (Freeling and Schwartz, 1973), whereas in leaves activity is

not present in nonstressed conditions, nor is it inducible by anaerobiosis (Okimoto et al.,

1980). The Adhl gene is also induced by anaerobiosis in maize roots (Ferl et al., 1980).

The mRNA increases about 50-fold after induction (Dennis et al., 1984) reaching this

level about 5 hours after initiation of anaerobic stress and remains at this high level for at

least 48 hours.

In the initial stages of this study, the Ti plasmid of Agrobacterium tumefaciens

strain 15955 was used as a vector to introduce the alcohol dehydrogenase 1 gene (Adhl)

of maize into the genome of sunflower. Transcriptional expression was characterized in

uncloned tumor tissue. Primary tumors from infected sunflower hypocotyls were placed

under anaerobic stress (flooding) for various lengths of time and analyzed for

transcriptional expression. We have shown by RNA blot hybridization and S 1 nuclease

hybrid protection mapping that the maize Adhl gene was incorporated into transformed

sunflower cells, and transcribed specifically in response to anaerobic stress utilizing its

natural promoter.

Materials and Methods

Transfer and Incorporation of Adhl into the Ti Plasmid.

The recombinant plasmid pB428-W9 containing the maize Adhl gene was

transferred into A.tumefaciens by tripartite conjugation (Comai et al., 1983; Fraley et al.,

1983). Three Gram negative bacteria were mated: E. coli strain SK1590 (Kushner, 1978)

containing the chloramphenicol resistant intermediate vector pB428-W9, E. coli LE392

(Murray et al., 1977) containing the kanamycin resistant helper plasmid pRK2013 (Ditta et

al., 1980), and the streptomycin resistant A.tumefaciens strain possessing its wild-type

pTi. Cultures of each were grown in 50 ml of Luria broth (LB) containing appropriate

antibiotics: overnight cultures of E. coli SK1590 harboring pB428-W9 (chloramphenicol,

20 tg/ml), and E. coli LE 392 containing pRK2013 (kanamycin, 50 gg/ml), and a two

day old culture of A. tumefaciens strain 15955 (streptomycin, 250 g.g/ml). Cells from 3

ml of each culture were pelleted by centrifugation at 4,000 xg for 10 min, washed in 10 ml

of sterile 10 mM MgSO4, and resuspended in 3 ml of LB containing no antibiotics. The

cultures were then mixed in equal proportions, and 100 gl to 300 tl of the mixture

transferred to an LB agar plate for mating (48 hr, 280C). The bacterial cells from the
tripartite mating were suspended in 2 to 3 ml of 10 mM MgSO4 and 100 to 200 ptl was

spread onto plates of LB containing streptomycin (250 pgg/ml) and chloramphenicol (20

gg/ml). The selection plates were incubated at 280C for 5 days to allow integration of the

shuttle plasmid with the pTi and colony growth to occur. Confirmation that colonies were

A. tumefaciens was obtained by a positive ketolactose test (Bernaerts and DeLay, 1963).

Incorporation of the Adhl gene into the pTi was demonstrated by Southern hybridization

(Southern, 1975) of immobilized pTi using a radioactively-labeled restriction fragment

(XbaI to KpnI of pB428; Dennis et al., 1984) corresponding to the coding region of Adhl

as the probe.

Tumor Growth.

Sunflower (Helianthus annus, cv. Large Grey) seedlings were inoculated with A.

tumefaciens containing the engineered pTi (pTi15955:pB428-W9) plasmid. A.

tumefaciens was grown in 50 ml of LB containing antibiotics for 2 days at 280C and used

to inoculate (syringe wounding) the hypocotyls of 6 to 8 day old seedlings. Plants were

grown indoors under fluorescent lights with a 16 hr light/8 hr dark cycle. Tumors were

harvested from 12 to 16 days after inoculation.

Induction of Adhl by Anaerobic Stress.

Sunflower tumors and maize roots were placed under low oxygen conditions in an

attempt to induce activity of Adhl. Immediately upon harvesting, the tumors were placed

into incubation media consisting of 1% sucrose, 1 mM KPO4 buffer (pH 6.0), and 100

pg/ml of rifamycin (modification of Gurley et al., 1986). Argon gas was bubbled through

the media for about 20 min to reduce the level of 02 before the containers were sealed.

After 1 to 24 hr of anaerobic stress, the tumors were frozen in liquid nitrogen and stored at

-800C. Maize (Dekalb XL-80) roots from 3 to 4 day old seedling were used as a control

for induction. The maize seedlings were fully submerged in water in the dark for 24 hr.

The root tips were then harvested and immediately frozen.

Analysis of Inducible Transgenic Expression.

Inducible expression of maize Adhl in stressed and unstressed tumors was

examined by RNA blot hybridization and Si nuclease hybrid protection mapping of

polyadenylated RNA. Total RNA was extracted by the

triisopropylnaphthalenesulfonate/para-aminosalicylate (TNA/PAS) method (Jackson and

Ingle, 1973) as modified by Czamecka et al., (1985). Poly(A)+RNA was fractionated by

oligo(dT)-cellulose chromatography (Silflow et al., 1979).

Northern blot hybridization of poly(A)+RNA was used to evaluate the length of

the Adhl message and to assess inducibility of the gene after exposure to stress for
various lengths of time. The poly(A)+RNA (10 gig to 20 gjg) was fractionated by

electrophoresis on a 2% (w/v) agarose, 6% (v/v) formaldehyde gel (Rave et al., 1979) and

directly transferred to nitrocellulose filter paper (Thomas, 1980). The transferred

poly(A)+RNA was hybridized with either nick translated maize Adhl cDNA from

pZML793 (Dennis et al., 1984), or an internally labeled single stranded bacteriophage

M13 (mp8) probe (Church and Gilbert, 1984) specific to the 5' portion (HaeIIl/TaqI

fragment; -90 to +85) of the gene.

Sl nuclease hybrid protection mapping (Favaloro et al., 1980) was used to

analyze the transcription start site of maize Adhl in both tumor cells and maize roots. A

Hinfl fragment from the 5' region of the gene was 5' end labeled (Maxam and Gilbert,

1980) and then digested with PstI leaving a 215 bp fragment, PstI to Hinfl (-140 to +77),
labeled at the Hinfl site. This probe was mixed with 40 to 50 p.g of poly(A)+RNA from

sunflower tumors in a total volume of 10 til of hybridization solution (40 mM piperazine-

N,N'-bis[2-ethane-sulfonic acid] buffer, pH 6.0, 400 mM NaCI, 1 mM EDTA, and 80%

formaldehyde) overlayed with 15 11 of mineral oil, and heat denatured at 850C for 15 min.

Hybridization was at 450C for 12 to 18 hr which was followed by digestion with S1

nuclease (200 units/ml) at 150C for 30 min (Czarnecka et al., 1985). The resulting

RNA:DNA hybrids were then fractionated by electrophoresis on an 8% urea DNA

sequencing gel and autoradiographed with X-ray film (Kodak SB-5) with intensifier

screens at -800C for 24 to 60 hours. Analysis of Adhl promoter 5' deletions by S1

hybridization was done by the excision of radioactive bands from the gel and

quantification by Cerekov counting.


Construction of the Intermediate Vector Plasmid.

The Adhl gene has been isolated and characterized by Dennis et al., (1984), and

has been shown by genetic studies (Schwartz, 1966) and Southern blot hybridization

analysis (Dennis et al., 1984) to exist in single copy in maize. The complete gene is

contained in plasmid pB428 (Dennis et al., 1984) which consists of an 11.4 kb fragment

inserted into the BamHI site of pBR322 (Bolivar et al., 1977). This restriction fragment

includes 1.5 kb of DNA 5' to the transcription start site, the coding region (with 9 introns)

and about 6 kb of downstream sequences (Fig. 2-1). Excess sequences 3' to the gene

were removed by subcloning an 8 kb BamHI to KpnI fragment containing Adhl into an

intermediate vector (shuttle vector).

The intermediate vector used to transfer the Adhl gene to A. tumefaciens was

designated pW9 (Bruce and Gurley, 1987) and consisted of a 4,158 bp fragment of

pT15955 T-left DNA (sites 9062 to 13220 bp, Barker et al., 1983) ligated into the BamHI

and SphI restriction sites of the chloramphenicol resistant, broad host range plasmid

pACYC184 (Chang and Cohen, 1978). The T-DNA fragment was derived from the

pBR322-based plasmid p223G (provided by Dennis Sutton, Agrigenetics Advanced


1 23 4 56 78 9 10
SOl moo IIOm

pBR322 %
%% %


I- I-



Bamrn H1


Fig. 2-1. Schematic representation of pB428 (Dennis et al., 1984). This plasmid
contains an 11.4 kb BamHI fragment harboring the maize Adhl gene in the unique
BamHI site of pBR322 (Bolivar et al., 1977). Exons are represented by black boxes with
corresponding number designation listed above each. Introns are depicted by white boxes
between the exons. Important restriction sites and their position with respect to the start of
transcription are presented.




Research Division, Madison, WI) in which the Smal site in the gene corresponding to

transcript 6b was changed to BglII by linker addition. After removal of a 776 bp BamHI

to KpnI fragment from pW9, the Adhl BamHI to KpnI fragment was inserted by

ligation. The resulting 15.4 kb intermediate vector containing the Adhl gene was

designated pB428-W9 and is shown in Figure 2-2.

Transfer and Incorporation of Adhl into the Ti Plasmid.

Tripartite conjugation (Comai et al., 1978; Fraley et al., 1983) results in the

eventual transfer of the recombinant plasmid from E. coli into the recipient A.

tumefaciens. Since the intermediate vector plasmid was a derivative of pACYC184, it

could not transfer unassisted into A. tumefaciens. The plasmid pRK2013 containing the

transfer genes from pRK2 (Figurski and Helinski, 1979) was utilized to achieve conjugal

transfer of pB428-W9 into Agrobacterium. Tripartite matings resulted in colonies of

transconjugant A. tumefaciens which possessed the ability to grow in the presence of

chloramphenicol. Since the shuttle plasmid has an E. coli replicon with a limited host

range, it cannot replicate autonomously in A. tumefaciens. The simplest explanation for

the acquisition of chloramphenicol resistance by A. tumefaciens was by co-integration of

the shuttle plasmid into the T-DNA of the pTi (Fig. 2-3). Confirmation of shuttle

plasmid:Ti plasmid cointegration was obtained by Southern hybridization analysis of the

Ti plasmid with an Adhl probe (Fig. 2-4).

Single homologous recombination of pB428-W9 with pTil5955 resulted in

incorporation of the shuttle plasmid in the T-left region of the T-DNA (Fig. 2-3). This

incorporation generated a duplication of the T-DNA genes 6a and 6b and a partial

duplication of the octopine synthase gene (lacking promoter sequences). The extra copy

of the gene corresponding to transcript 6b was probably not functional in the plant since

the reading frame was disrupted by the BglII linker at the former Smal site.

Bam H1 Bam H1
Kpn I Kpn I
* Isolate fragment Isolate fragment

Kpn I

7.7 kb

Sph I

Bam H1


3.4 kb

3.4 kb

Fig. 2-2. Cloning scheme for shuttle vector pB428-W9. Location and
transcriptional orientation of represented genes are designated by arrows within plasmids.
Origin of plasmid DNA and important restriction sites are as labeled. Numbers within
parenthesis designate position of restriction sites within octopine-type Agrobacterium
15955 T-DNA (Barker et al., 1983). Legend: Apr, ampecillin resistance; Cpr,
chloramphenicol resistance; ocs (octopine synthase), 4, 6a, and 6b represent T-DNA
genes of Agrobacterium 15955 (Willmitzer et al., 1981). Dotted lines (gene 4 and ocs)
portray lack of promoter sequences.

Q r=
u d H d 3 u, 6 -o- ~


0 >:


Z~r (A 00U
o -o.O o'-0
H1 0 v-.4 0~

"o 0

OJ 0
0: o U r

r- 0
S 0.
bo"O X.O

n C) U~i ~ O ;

a C)
04- s Zn

GU 0
c M u .O U 0 = i:
"t "?~ cud (A o Cd o

El C a E 8> CQc

tu o C cl t >
c ea f3 (U
rZ 0o

> -'r.

r. cl 00 W



4 U CO v o c) Oc0'
c o<
o =0~an ,y

boo r.~;3 v ,

F 00v o C
'o, = 0









3 4 5 6 7 8 9 10
-fqvP~E -I



^ *& **soft


, Chromosomal

- plasmid


Fig. 2-4. Confirmation of Adhl incorporation into Ti-plasmid. (A) Ethidium
bromide stained DNA from nine different colonies (lanes 1-9) of Agrobacterium that grew
on selection plates after tripartite matings (see Materials and Methods). Non-transformed
Agrobacterium control (lane 10). (B) Southern blot analysis of samples described in A.
Hybridization was to a nick-translated (Rigby et al., 1977) DNA probe harboring the
Adhl genomic clone (see Materials and Methods). Intact Ti-plasmids and bacterial
chromosomal DNA are designated.

4 5 6 7 8 9 10


1 2 3

Analysis of Transcriptional Expression of Maize Adhl in Primary Tumors.

Hypocotyls of sunflower seedlings were infected with A. tumefaciens harboring

pTil5955:pB428-W9. Tumor tissue developed at the wound sites indicating that the T-

DNA, and thus the maize Adhl gene, had transferred into the genomes of sunflower cells.

These tumors, which were 70 to 80% the size of tumors that formed on seedlings infected

with wild-type A. tumefaciens strain 15955, were harvested and anaerobically stressed.

Poly(A)+RNA was isolated from the tumor tissue and assessed for levels of maize Adhl


Sunflower tumors synthesized endogenous Adh mRNA in response to anaerobic

stress. This conclusion was based on the hybridization (Fig. 2-5) of immobilized

poly(A)+RNA from tumors inoculated with A. tumefaciens harboring either the naturally

occurring pTi15955 or the engineered pTi15955:pB428-W9 with nick translated maize

Adhl cDNA from pZML793 (Dennis et al., 1984). The sunflower transcripts were

slightly smaller than the 1,650 bp maize Adhl mRNA (Gerlach, et al., 1982) (Fig. 2-5,

lanes labeled M). The cross-hybridization of the maize cDNA with the mRNA of the

endogenous sunflower Adh gene gave an indication of the level of sequence conservation

between these two species of plants, and also indicated the level of transcriptional

induction of Adhl in tumors.

There was low level of hybridization to poly(A)+RNA from uninduced tumors

(Fig. 2-5, lanes A and B; 0 hours) incited by wild-type pTi and by pTi15955:pB428-W9.

After 1 hr of anaeobic stress, only tumors incited by pTi15955:pB428-W9 showed an

increase of Adhl mRNA levels. A dramatic increase in Adhl RNA levels was seen in

both types of tumors after flooding for 3, 6, 9, and 12 hr. Although the earlier induction

of Adhl in tumors incited by pTil5955:pB428-W9 compared to tumors containing

pTi15955 was not expected, the overall results indicate that sunflower tumors provide a

suitable experimental system to study the anaerobic response.

0 1 3 20 6 9 12 20
rA Bi 5A BA 5A BB M iA BI 'A BI 5A BI M

4 1
1 ? r



Fig. 2-5. Northern blot and time course analysis of poly(A)+RNA. Messenger
RNA (14 pg) from control tumors formed after inoculation with wild type Agrobacterium
15955 (lanes A) and tumors infected with Agrobacterium possessing the Adhl gene in T-
DNA (pTi15955:B428-W9) (lanes B). Lanes M contain poly(A)+RNA (1 tg) from roots
of anaerobically stressed maize seedlings. Time (hours) that tumors and roots were
subjected to anaerobic stress (see Materials and Methods) are indicated. Blots were
probed with nick-translated maize Adhl cDNA from pZm793 (Dennis et al., 1984). Solid
arrow indicates location of processed mRNA. Open arrows indicate location of larger
transcripts (see text).



t i:

The cDNA probe also hybridized to two large transcripts (3 to 4 kb and 6 to 10 kb)

present in the poly(A)+RNA isolated from anaerobically stressed tumors containing the

maize gene (Fig. 2-5; lanes labeled B; hrs. 3, 6, 9 and 12). These larger transcripts were

not seen in lanes containing poly(A)+RNA from control tumors or maize roots. From this

type of analysis using the cross-hybridizing cDNA probe, it was unclear whether the

introduced maize Adhl gene was transcriptionally expressed in the heterologous system.

In order to specifically detect the transcripts derived from the maize gene in the

presence of endogenous Adh RNA, a second RNA blot analysis using an M13-derived

(Church and Gilbert, 1984) strand specific probe was utilized. The internally labeled

probe consisted of a HaeII (-90) to TaqI (+85) fragment of the maize Adhl gene cloned

into M13 bacteriophage mp8 (Fig. 2-6). This probe was complementary to 85 bases of

untranslated 5' leader sequence of maize Adhl RNA.

The ability of this probe to hybridize specifically to maize Adhl transcripts was

substantiated by RNA blot analysis. Hybridization was seen between the leader specific

probe and poly(A)+RNA of anaerobically stressed tumors harboring the foreign Adhl

gene; there appeared to be only slight hybridization with transcripts that were the same size

as processed maize Adhl mRNA (1,650 bases) (Fig. 2-6). In contrast, prominent

hybridization occurred which corresponded to the two larger transcripts previously

observed (Fig. 2-5) with the nonspecific cDNA probe (Fig.2-6; lanes labeled B; hrs. 1, 3,

6, 9, and 12). The specificity of the leader probe was confirmed, since poly(A)+RNA

from sunflower tumor tissue incited by control A. tumefaciens 15955 did not hybridize

with the probe (Fig. 2-6; A lanes), whereas a 1,650 base message from maize roots

hybridized quite readily (Fig. 2-6; lane M).

The RNA blot hybridizations indicate that the transgenic maize Adhl gene in

primary sunflower tumor tissue was transcriptionally expressed in an anaerobically

inducible manner. Processing of the RNA (intron removal or proper termination),

however, appeared to be aberrant. Failure to splice the nine introns of this transcript is

1 3 20
A 1 A B1 "

6 9 12 20
iA BI IA B1 A B1 M 0







*1 +85 ATG





Fig. 2-6. Northern blot probed with maize Adhl specific probe. Experimental
parameters are as in Figure 2-5. Internally labeled (Church and Gilbert, 1984) maize
Adhl specific probe harboring 90 bp of 5' flanking and 85 bp of leader sequences is
illustrated. Lane "(" contains end labeled HaelII generated OX174 markers with size





predicted to yield a 3.4 kb message which roughly corresponds to the size of the smaller

of the two large transcripts. Further experiments that address the identity and origin of the

largest hybridizing RNA are discussed below.

Sl hybridization protection mapping was performed on poly(A)+RNA isolated

from primary tumors and maize roots in order to determine if the transgenically expressed

Adhl gene used the same transcriptional initiation site preferred in maize. An end-labeled

probe that consisted of 140 bp of DNA upstream of the reported start site for the maize

gene (Dennis et al., 1984) and 77 bp of DNA complimentary to the mRNA leader

sequence was utilized. No bands were detectable by S1 nuclease hybrid protection

mapping of RNA from control tumors (Fig. 2-7; A lanes) indicating the specific nature of

the probe. In contrast, an S 1 nuclease protected band from RNA isolated from stressed

tumors containing the maize Adhl gene was obtained (Fig. 2-7; B lanes; 1, 6, and 9 hr)

that precisely coincided in length with the protected fragment obtained from maize root

RNA (Fig. 2-7; lane M). Additional higher molecular weight transcripts which formed an

RNA:DNA hybrid over the entire 217 bp DNA probe were also present suggesting the

presence of anaerobically inducible readthrough transcription initiating upstream of the

natural start site.

Splicing of Adhl transcripts in primary tumors.

S hybridization protection analysis was employed to investigate the splicing of

Adhl transcripts in sunflower tumor cells A DNA probe was labeled at a HindII site

within the first intron and consisted of 1,094 bp of untranscribed 5' flanking sequences

and the first 213 bp of the transcribed region of the gene (Fig.2-8, DNA probe "a", panel

C) (Dennis et al., 1984). No protection was observed when hybridized to mRNA from

maize roots subjected to flooding (Fig. 2-8; panel A, lane 5). Conversely, a 213 bp

fragment was protected from S1 nuclease digestion by mRNA from tumors with

incorporated Adhl stressed for six hours (Fig. 2-8; panel A, lane 4) indicating that at least

20 0 1
0M t rA BA B'



118* -

6 9 20
-A B'A B'M

full length
"<" probe

-p I


Pst I


"- fragment

Hinf I

DNA probe *


Fig. 2-7. S1 nuclease hybridization protection and time course analysis of
poly(A)*RNA. The 217 bp 5' end labeled (*) hybridization probe is illustrated. Poly
(A)+RNA (1 ag) from maize roots (lane M) and from primary tumors (16 lg) and lengths
of time tissues were subjected to anaerobic stress are as previously described. Control
lane "t" contains 22.4 plg of tRNA from Baker's yeast (Sigma).

Fig. 2-8. S1 nuclease hybridization protection analysis of maize Adhl in
sunflower tumors. (A) Hybridization to poly(A)+RNA with DNA probe (a) 5' end
labeled at +213 within intron 1. Lane ")" is as previously described. Lanes 1 and 2
contain 15 gLg of poly(A)+RNA isolated from control sunflower tumors formed with wild
type Agrobacterium 15955, uninduced or anaerobically stressed for 6 hours, respectively
(see Materials and Methods). Lanes 3 and 4 contain 15 gg of poly(A)+RNA isolated from
tumors formed after infection with Agrobacterium harboring pTi15955-pB428-W9,
uninduced or anaerobically stressed for 6 hours, respectively. Lane 5 contains 2 pg of
poly(A)+RNA isolated from maize roots subjected to 22 hours of flooding (see Materials
and Methods). (B) Hybridization to poly(A)+RNA with DNA probe (b) 5' end labeled
within exon 2. Lane designations are as described for panel A. (C) Representation of S
hybridization strategy. Properly processed polyadenylated RNA protect an 118 bp
fragment of probe (b) and those transcripts that maintain intro 1 protect 213 bp of DNA
probe (a) and 773 bp of probe (b). Significant transcript sites are represented. End
labeled (*) probes and predicted protected fragments for each are illustrated. Arrows
indicate protected fragments: solid, predicted fragment; open, non-predicted protected



ct -700


-<- 792

S -N. 213


Hin dill
+1 +213

exon 1

Hin dill

213 bases


intron xon 2

xon I

Acc I

792 bases


0 1

o *


194. *-

118. w -

Bam HI

DNA probe (a)

DNA probe (b)

118 baes

some polyadenylated transcripts had not processed intron 1. The RNA from control

tumors and the uninduced tumors with maize Adhl showed no protected bands (Fig. 2-8;

panel A, lanes 1-3).

The failure of sunflower tumors to properly splice maize Adhl transcripts was

further indicated by a similar experiment using a DNA probe 5' end labeled within the

second exon (AccI site at +792) (Fig. 2-8). Poly(A)+RNA from the stressed maize roots

resulted in a 118 base protected fragment (Fig. 2-8; panel B, lane 5) that corresponded to

the intron 1:exon 2 splice junction (Dennis et al., 1984) indicating that intron 1 had been

properly spliced from the mature message as expected. Messenger RNA from

anaerobically induced tumors (6 hours) harboring Adhl failed to protect the 118 bp, but

rather hybridized and protected a fragment approximately 800 bp in length (Fig. 2-8; panel

B, lane 4). The 5' terminus of this hybrid protected RNA corresponded with the

transcription start site indicating that the intron was still present in most of the tumor

transcripts of maize Adhl.

These data suggest that little or no transcription products from maize Adhl in

sunflower cells had spliced intron 1 from the mature message. Keith and Chua (1986)

had similar results when analyzing a chimeric construct containing maize Adhl intron 6 in

transgenic tobacco plants. Intron 6 had been inserted between chloramphenicol acetyl

transferase (CAT) coding sequences and the pea rbc-E9 polyadenylation signal of a

construct driven by the cauliflower mosaic virus (CaMV) 35S promoter. They showed

that in tobacco about 70% of the transcripts from this chimeric failed to splice the

intervening intron. This information, together with the banding pattern of the Adhl

transcripts from stressed sunflower tumor tissue in the Northern blots (Figs. 2-5 and 2-6),

imply that all nine introns within RNA derived from this monocot gene are inefficiently

spliced in dicotyledonous plants, at least sunflower and tobacco.

Possible upstream start site.

An interesting observation developed from the Sl hybridization experiments

performed to look at splicing activity. In those lanes addressing transcript processing

from anaerobically stressed tumors, protected fragments appeared that were indicative of a

possible upstream transcription start site. The hybridization experiment using the probe

end labeled within intron 1 not only protected 213 bp of the probe from S1 digestion, but

a fragment of about 700 bp (Fig. 2-8, panel A, lane 4). The second probe end labeled

within exon 2 may have also hybridized and protected a long stretch of DNA extending

1,200 to 1,300 bp upstream (Fig. 2-8, panel B, lane 4). These surprising results suggest

that a second upstream promoter, also under the control of environmental stress, may be

present further upstream (approximately 500 bp) from the known Adhl transcription start


To further investigate the possibility of an upstream start site, a 690 bp DNA

fragment end labeled at an XbaI restriction site, 410 bp upstream of the recognized

initiation site (Dennis et al., 1984), was used in another S1 nuclease mapping analysis.

No protection was observed with transcripts from transformed tumors that had not been

stressed (Fig. 2-9; lane C), nor from maize roots subjected to flooding (Fig. 2-9; lane M).

However, poly(A)+RNA from tumors stressed for six hours protected about 100 bp of the

end labeled probe (Fig. 2-9; lane I) supporting the existence of a second initiation site

approximately 510 bp upstream of the known transcript initiation site. Consistent with

evidence of an upstream start is the presence of TATATTAA sequence approximately 25

to 35 bp further upstream of this possible start. This sequence conforms to the 5'-

TATA(A/T)A(A/T)-3' consensus sequence (Breathnach and Chambon, 1981) usually

found 21 to 35 bp upstream of a transcription start site that functions in positioning the

start of transcription in many eukaryotes (Dynan and Tjian, 1985). On the other hand, no

sequences resembling other proximal promoter elements, such as a CAAT box, are

present upstream of the TATA sequence; instead, this region is highly AT rich (Dennis et





- .90

Bam HI

Xba I

DNA probe

Fig. 2-9. Examination of possible upstream transcription initiation site by S1
hybridization mapping. The 684 bp Adhl upstream 5' end labeled (*) probe is illustrated.
Hybridization experiments are with 20 gLg each of poly(A)+RNA from tumors harboring
the Adhl gene; uninduce (Lane C) or anaerobically stressed for 6 hours (lane I). Control
mRNA (2 g.g) is from maize roots (lane M). Lane "m" contains pBR322 Hpall end
labeled fragments (lengths are presented). Lane "t" is as previously described.

al., 1984). Primer extension experiments of poly(A)+RNA from sunflower tumors or

maize roots were inconclusive in confirming the existence of an upstream start site (data

not shown).

5' deletions of the maize Adhl promoter and analysis in sunflower.

The Adhl gene of maize has been shown to utilize its own promoter in transgenic

sunflower cells and retain its ability to respond to anaerobic stress. Similar 5' deletion

mutations of the promoter were constructed in an analogous attempt to localize important

sequences involved in transcriptional regulation of the gene in this transgenic system. Due

to the presence of certain restriction sites within the maize Adhl gene, a number of

manipulations were required to construct 5' deletion mutants to positions -410 and -140.

The 8 kb BamHI to Kpnl fragment utilized in previous experiments was subcloned into

pUC19 (Vierra and Messing, 1982) where an EcoRI restriction site from the pUC19

polylinker is proximal to the KpnI site. The BamHI to EcoRI fragment from this

intermediate construct containing Adhl was then subcloned into pBR322. These

manipulations allowed the fragment of interest to reside in a vector void of HindIII

restriction sites. The gene was then "gutted" by removal of a 2.5 kb HindIII fragment

internal to the gene resulting in pAdh-AH (Fig. 2-10). The actual promoter mutations

were constructed by subcloning fragments consisting of Adhl 5' flanking sequences the

5' untranslated leader sequence, and the first 213 bases of the coding region, into pUC19.

The first of these mutations involved the XbaI (-410) to HindIII (+213) fragment which

lacked 684 bp of the most distal 5' sequences. Similarly, the second construct consisted

of a PstI (-140) to HindIII (+213) fragment minus 954 base pairs upstream of the -140

site of the promoter. Each of these promoter deletions subcloned into pUC19 acquired a

BamHI restriction site from the vector's polylinker located immediately 5' to the promoter

deletions. The promoter mutants were then subcloned into the BamHI and HindIII

restriction sites of pAdh-AH which had the wild type promoter excised. Finally, the 2.5


Bam H1
Kpn I

Bam H1




* isolate fragment

Bam H'


Bam I

Fig. 2-10. Construction of "gutted" pAdh-AH. Stippled region of plasmids
indicative of genomic sequences from maize containing Adhl; white region represents
designated vector. Location and transcription orientation of Adhl is illustrated by arrow
within bacterial plasmids. Perpendicular line through arrow of pAdh-AH indicates internal
deletion of a 2.5 kb HindIII fragment.

kb HindIII internal fragment was re-inserted in its proper orientation into the pAdh-AH

derivatives resulting in two Adhl genes with intact coding sequences having the promoter

5' deletion to position -410 and -140 (Fig. 2-11). Each of these constructs possessed a

BamHI site at their 5' ends that was essential for ligation of these mutated genes into the

pW9 shuttle vector.

Six hr anaerobically stressed sunflower tumor tissue which contained Adhl 5'

deletion mutations were examined for transcriptional expression by S1 transcript mapping

experiments. Each showed a difference in transcriptional activity compared to that of wild

type (Fig. 2-12; lane -1094-1). The -410 mutant expressed transcription levels

approximately 40% of wild type (Fig. 2-12; lane -410-1). Removal of additional

sequences to position -140 resulted in further reduction of transcript levels to

approximately 15% that of wild type (Fig. 2-12; lane -140-1). Constitutive activity of each

of the constructs was less than 1.5% of the maximum value after induction (Fig. 2-12,

lanes U). Interestingly, these results show a correlation to those of Lee et al. (1987) who

examined expression in homologous maize protoplasts. Also, both 5' deletion mutants

initiated transcription at the proper start site and expression of each gene was inducible.

This indicated that the ARE identified in maize protoplasts (Walker et al., 1987), which is

still present in these deletion mutants, may work in a similar fashion in sunflower.


Maize Adhl is a highly regulated, organ specific gene that was able to utilize its

own promoter when induced in transgenic sunflower tumors. Transcription was activated

in response to anaerobic stress and reached steady state levels between 1 and 6 hours after

induction and remained at this level for at least an additional 3 hr. Although it is apparent

that sunflower tumor cells possess transcription factors that are recognized and properly

utilized by regulatory elements of this monocot gene both RNA blot hybridizations and S1

(-140)Pst I


Barn H1/Hn dill

ga tio Pst I

Hin dll Adh-

Pst I 8.81 kb

(410)Xb I H Ar Kpn I
Barn Eco R1
Pst IpBR322
Snsert Hin dill/Hin dill

Hindill Hindill

Pst 11.3 kb

(-410) Xba I Ap Kpn I
Barn H1 Eco R1
Pst I pBR322



\j .ligation


8.54 kb

Kpn I
140)Pst I r co R1
Barn H1
Psti -pBR322|

Fig. 2-11. Construction of the two genomic Adhl 5' deletion promoter mutants,
pAdh-410 and pAdh-140. Origin of sequences within plasmids and Adhl location and
orientation are as previously described. Important restriction sites and position with
respect to the Adhl start of transcription are shown in parenthesis.

(-140) Pst I

(-410) Xba I
Barn H1



194- me

118. e-


full length
-l W- probe


" -* "" fragment

72- *

Fig. 2-12. Sl nuclease protection analysis of Adhl 5' deletion mutants from
sunflower tumors. Hybridization of the end labeled probe previously described in figure
2-7 was with poly(A)+RNA isolated from tumors formed upon infection with
Agrobacterium harboring maize Adhl transcription units with 5' flanking sequences
extending 1,094 (-1094), 410 (-410), or 140 (-140) base pairs upstream from the
recognized initiation of transcription site. Results of Sl mapping from control
(uninduced) tumors containing 20 gpg of poly(A)+RNA are presented in lanes "U" and
from induced tumors anaerobicallyy stressed for 6 hours) containing 20 .g of
poly(A)+RNA in lanes I. The 77 bp protected fragment is designated. Lanes ")" and "t"
are as previously described.

ru- -- I

t r

nuclease mapping experiments indicated that the heterologous Adhl transcript was poorly

spliced in sunflower tumors. Proper removal of intron 1 was shown to be all but non-

existent. Northern blot analysis indicated that the other eight introns were also

inefficiently spliced since the most prominent bands correspond to a full length 3.4 kb

unspliced transcript (Figs. 2-5 and 2-6). In the study by Keith and Chua (1986), a

chimeric gene consisting of the wheat rbc coding region under the control of the CaMV

35S promoter was incorporated into transgenic tobacco plants. They demonstrated that

the single intron remained unspliced 10 to 30% of the time, although those that were

spliced did so at the correct splice junctions. To further compare aberrant splicing of a

monocot transcript in a dicot system, this group developed a chimeric gene having either

intron 6 of maize Adhl or the pea rbcS-3A first intron positioned between CAT coding

sequences and the pea rbc-E9 polyadenylation signal. These chimeric genes, driven by

the CaMV 35S promoter, were incorporated into transgenic tobacco. Only about 30% of

the transcripts containing the Adhl intron were spliced, whereas roughly 80% of the

introns from pea were processed from the transcripts.

The results presented in this study differ from those obtained when maize Adhl

was incorporated into the genome of tobacco. Analysis of tumors incited by T-DNA

harboring maize Adhl resulted in no detectable transcript levels (Llewellyn et al., 1985).

When the Adhl promoter was linked to a chloramphenicol acetyl transferase (CAT) gene

and incorporated into transgenic tobacco plants, again there was little evidence of promoter

function (Ellis et al., 1987a). Anaerobically induced CAT activity and proper start site

utilization were observed in tobacco only when the upstream sequences of octopine

synthase (ocs) or cauliflower mosaic virus (CaMV) 35S genes were positioned 5' to -140

of this chimeric gene. Similarly, the developmentally regulated zein promoter showed

activity in cloned sunflower tumor tissue. Although transcripts were detectable, no

proteins were observed (Matzke et al., 1984; Goldbrough et al., 1986). After

incorporation of a maize zein gene into tobacco, poly(A)+RNA could only be detected at

very low levels in the seeds of transgenic plants (Schemthaner et al., 1988). The

promoter was shown to function in a tissue specific manner after it was linked to a

glucuronidase synthase (GUS) reporter gene whose product can react histochemically to

localize minute amounts of activity. Thus, it appears that transcription factors are present

in sunflower tumor tissue that can be recognized and utilized by maize Adhl and zein

promoters. In tobacco the transcript initiation sites of these promoters are recognized but

transcription efficiency seems to be much lower than in sunflower indicating differences in

the abundance, or recognition efficiency, of analogous transcription factors in these two

dicot systems.

Although in the present study splicing of maize Adhl transcription products was

aberrant in sunflower tumor cells, Adhl promoter function was unimpaired. Two 5'

deletions of the promoter and analysis of gene product levels by S1 hybridization indicated

similar promoter strengths in this dicotyledonous system as found by Lee et al. (1987) in

maize protoplasts. In the present study deletions of sequences upstream of position -410

and -140 did not release the promoter from anaerobic regulation but resulted in a

substantial reduction in activity to 40% and 15% of wild type, respectively. Adhl

promoter analysis in homologous maize protoplasts showed about 65% and 18% activity

for the same promoter mutations compared to wild type activity (Lee et al., 1987).

Results of the -410 deletion in protoplasts superficially appear to differ from sunflower,

but the activity in the protoplast study was actually an average of 7 different experiments

in which the data fluctuated anywhere from 100% to 42% of wild type activity. Walker et

al. (1987) also performed a similar promoter analysis of the Adhl promoter linked to a

CAT reporter gene in maize protoplasts. They reported that a 5' deletion to the -140

position was still induced by low oxygen tension, but the transcript level was 75% of wild

type. They also localized the region that is necessary for anaerobic induction (anaerobic

response element; ARE) between -140 and -99.

Analysis of RNA derived from the maize Adhl gene suggested that in sunflower

tumors a small portion of the transcripts initiated approximately 520 bp upstream of the

proximal start site. The DNA sequences at this putative upstream start site are relatively

AT-rich and may contribute to the production of spurious bands in an S1 nuclease hybrid

protection assay due to the reannealing of the DNA probe and subsequent cleavage at

regions of partial denaturation. This possibility seems unlikely in this case, however,

since under identical hybridization conditions the appearance of the high molecular weight

bands were strictly correlated with RNA from anaerobically stressed tumor cells harboring

maize Adhl. We are left with the possibility that maize Adhl has a second promoter

corresponding to the upstream start site.

The existence of linked dual regulatory sites for Adhl (F allele) has previously

been hypothesized by Schwartz (1976a, 1976b) and dual promoters have been observed

in other eukaryotic genes that are developmentally or tissue specific in their regulation (for

review, Shibler and Sierra, 1986). Of particular interest is the alcohol dehydrogenase

gene in Drosophila melanogaster. A proximal promoter is utilized in the larval stage to

produce mRNA of about 1,100 bases in length with a 70 base untranslated leader. The

major transcript in the adult stage utilizes a distal promoter 654 bp upstream to the

proximal promoter. This upstream promoter has a TATA site 31 bp 5' to the start of

transcription and produces a transcript which, after processing, yields an approximately

1,150 base mRNA initiated at the upstream start site. Both adult and larval stage mRNAs

utilize the same coding region and differ only at the extreme 5' non-coding region.

(Benyajati et al., 1983)

The occurrence of an upstream start site (distal promoter) in sunflower tumors

raises the possibility that a dual promoter system similar to that present in Drosophila Adh

may also exist in maize Adhl. The inefficiency of intron removal in sunflower

fortuitously assisted in the identification of the distal promoter since S1 nuclease mapping

probes originally selected for the presence of 5' labelling sites internal to the coding region

were able to detect a continuous RNA:DNA hybrid from the labelled 5' terminus of the

probe to the distal start site. Of possible significance is the occurrence of sequences

showing homology to the D. melanogaster Adh distal promoter in the region of the

putative maize distal promoter (Kreitman, 1983) (Fig. 2-13). Other DNA sequences in the

region of the upstream start site which are consistent with its proposed role as a distal

promoter include a TATA-like motif (5'-TATATTAA-3') similarly spaced from a 5'-CAT-

3' sequence which is recognized as the transcription initiation site in D. melanogaster.

Transcripts initiated at the putative distal promoter of maize Adhl would be 3.9 kb

in size before processing. In the case of Drosophila Adh, transcripts initiated by the distal

promoter are spliced in the untranslated leader sequence, a process which removes

spurious ATG triplets that may reduce translational efficiency if present in the mRNA.

The putative distal promoter of maize Adhl may also follow a similar pattern of intron

splicing. Two potential 5' donor sites for splicing are located at +8 and +13 in transcripts

derived from the distal promoter. Drawing on the Drosophila analogy, a possible 3'

acceptor site in the region corresponding to the leader of the transcript derived from the

proximal promoter is located at +5. The utilization of these splicing sites would remove

redundant ATG triplets upstream of the coding region (+19 and +26) and would increase

the size of the mRNA by only 4 to 9 bases. Though this model for a dual promoter for

maize Adhl is simply conjecture due to the lack of any supporting evidence for the

utilization of the distal promoter in maize, it is interesting to note that an insertion of a 1.85

kb Mp-1 like element into the promoter (at -31) of Adhl reduced transcription levels of

this gene in seeds and anaerobically stressed seedlings to 6% of normal activity, whereas

ADH1 expression in pollen remained constant (Chen et al., 1987). The failure of Mp-1

to inactivate pollen expression suggests that in this case Adhl expression is regulated by a

promoter other than that associated with the proximal TATA. Again, a comparison

between this gene and Drosophila Adh may be made in that developmental control of

Adh expression is achieved by employing dual promoters (Benyajati et al., 1983). The

U,z x

cn cda

13 E cng
U .

I- .Lc 0


nC ui

1.-4 UA

crit ~ .0 1:U


Uo U ~q C.)

)U C
E 4 E 4|


E- C


C0 0

E- HE-

o (o

) CD)C ,D

0-+ +


E-c ) U

I E-c

0 E-I
U o

-4 *
E- 40


U-H -

< E-I
o CD
0 E-C

c 2) o-
u c



existence of dual promoters regulated by organ specificity cannot as yet be ruled out. It

will be interesting to pursue this question experimentally.

Even though splicing of the maize Adhl gene in sunflower tumors is faulty,

regulation of the transcriptional expression proved to be functional and closely mimicked

the response of this gene in its natural host. The following section focuses on localizing

and analyzing cis-acting regulatory elements of Adhl that are recognized in this

dicotyledonous plant system.



In order to identify and map regulatory elements of the maize Adhl promoter a

number of mutations were constructed by cloning and were analyzed in vivo using the

heterologous system of sunflower tumors. To facilitate in vitro manipulation of the Adhl

promoter the Adhl coding region and introns were removed and substituted by the coding

region of the 780 gene of T-right (pTil5955) (Karcher et al., 1984; Winter et al., 1984) as

shown in Fig. 3-1. This gene, so named because it produces an approximate 780 bp

transcript was enlisted as a reporter gene for the Adhl promoter because of the lack of

interfering restriction sites within the coding region and the absence of introns within the

resulting transcript.

In order to more accurately quantify transcription from the promoter mutants in

sunflower tumors, an internal standard of measurement was applied as utilized by

McKnight and Kingsbury (1982) and Bruce and Gurley (1987). This method of

quantification consists of incorporating two homologous genes into T-DNA which are

transferred in equal number into the genome of the recipient host. The DNA sequence

composition of one gene remained constant and is referred to as the reference gene. The

second gene, referred to as the test gene, can be mutated in vitro and incorporated into the

genome of sunflower along with the reference gene. Test and reference transcripts were

distinguished by the presence of a 22 base insert into the test gene 5' untranslated leader.

Thus, mRNA encoded by the test gene could be directly compared to the transcription

products of the reference gene.

Materials and Methods

Maize Adhl 5' and 3' deletions.

Construction of the test and reference genes required the construction of 3'

deletions with the end points within the untranslated leader sequence of Adhl. Plasmids
(10 gig) containing the entire promoter and the first 213 bp of the coding region were

linearized at a unique SstlI site (+106). The linearized fragment was suspended in 39 pi of

TE (10 mM Tris/HCl pH 8.0; ImM EDTA) and 10 il of a Bal 31 reaction buffer (100 mM

Tris/HCI pH 8.1, 60 mM MgC12, 60 mM CaCl2, and 1.5 M NaC1) was added. The
mixture was placed in a 300C water bath for approximately 5 minutes before adding 1 pl

(250 U/ml) of exonuclease Bal 31 (Bethesda Research Laboratories). Aliquots of 10 p.l

were subsequently transferred to separate "Microfuge" tubes containing 1 p of 40 mM

EGTA (giving a final EGTA concentration of 3.6 mM) at 20, 40, 60, 90, and 180 second

intervals in order to stop the exonuclease reaction. Each tube was immediately placed in a

650C water bath for 15 minutes to further arrest the reaction. The five aliquots were

pooled, extracted once with an equal volume of phenol/chloroform and once with an equal

volume of chloroform. The DNA was then ethanol-precipitated. Internal deletions of the

anaerobic response element (ARE) and the 5' promoter deletion mutations were similarly

produced. Removal of the ARE involved a 5' Bal 31 exonuclease digestion of the test

gene at a unique PstI restriction site (-140). Promoter mutated 5' deletions of the

Adhl/780 test gene chimeric construct were similarly generated. Test gene chimerics in a

pUC19 vector were opened at unique BamHI (-1094) or Xbal (-410) sites and deletions

obtained with the Bal 31 exonuclease reactions as previously described with the exception

that reaction intervals were 1, 2, 3, 4, and 5 minutes.

Any 5' overhangs generated by the exonuclease activity of the different Bal 31

digestions were subjected to fill in reactions and linker addition (Maniatis et al., 1982); the

5' promoter deletions with XhoI linkers (Bio Rad) and the ARE deletion with PstI linkers

(BRL). End points of the deletions were estimated by sizing on an 8% polyacrylamide gel

and compared to standard size markers. The end points of Adhl promoter 3' deletions for

construction of test and reference genes were also determined by sizing on an 8%

polyacrylamide gel. The actual Adhl/780 junctions and identification of ARE

constructions were subsequently determined by M13 sequence analysis.

M13 sequence analysis.

Localization of the test and reference gene Adhl promoter/780 coding junctions

and test gene promoter mutations were determined by M13 cloning/dideoxy sequencing

(Sanger et al., 1977). Chimeric test and reference genes were excised from p251 double

gene vectors used in tripartite matings. These genes were subcloned into M13 strain

mpl8 or mpl9 and sequenced to confirm the identity of mutations.

Transfer and incorporation of Adhl/80 chimeric test and/or reference genes into the Ti-


Tripartite matings (Comai et al., 1983; Fraley et al., 1983) were as previously

described in chapter 2 with the following modifications. E. coli strain LE 392 rec- was

used to harbor both the intermediate shuttle vectors and the helper plasmid pRK2013

(Ditta et al., 1980). The Agrobacterium tumefaciens 15955 mutant Ag5260 (strr;knr)

(Bruce and Gurley, 1987) was the recipient strain in mating events. This mutant strain

has the endogenous 780 gene deleted from T-right of the Ti-plasmid and replaced by a

kanamycin resistant gene from transposon Tn5 (Rothstein et al., 1980). Overnight

cultures of Ag5260 were grown in LB containing streptomycin (250 gg/ml) and

kanamycin (50 gg/ml). Two milliliter aliquots of each of the overnight cultures were

centrifuged at room temperature and the cells were washed twice with sterile 10 mM

MgSO4. The E. coli cells were resuspended in 50 mis of LB containing no antibiotic and

grown to log phase (2 1/2 hr) in a shaking 370C water bath at 250 rpm. The Ag5260 cells

were resuspended in 4 ml of LB with no antibiotic. Equal portions (200 pls) of the two

E. coli cultures harboring the shuttle vector and the helper plasmid, and Ag5260 were
mixed in a "Microfuge" tube. The cells were pelleted (45 seconds), resuspended in 100 pl

of LB, and spotted in the center of an LB agar plate for mating (48 hrs, 280C). Selection

of Agrobacterium with the incorporated chimeric genes were as previously described in
chapter 2 with the exception that the selection plates also contained kanamycin (20 p.g/ml).

Resulting colonies were transferred to LB containing the three antibiotics; streptomycin
(250 gg/ml), chloramphenicol (20pjg/ml), and kanamycin (20 .tg/ml). Transconjugants

were reselected by growth (2 days) in LB containing antibiotics (streptomycin 250 gg/ml,

kanamycin 20 g.g/ml, and chloramphenicol 20 gpg/ml).

Tumor formation and anaerobic stress.

Inoculation of sunflower (Helianthus annus, cv. Large Grey) seedlings with

Agrobacterium harboring the modified T-DNA of the Ti-plasmids were as previously

described in chapter 2 (materials and methods). Tumor tissue containing each construct

was either induced for six hr and frozen or were frozen immediately upon harvesting


Analysis of transgenic expression.

Poly(A)RNA was collected from the different tumor tissue as described in

materials and methods of chapter 2. S nuclease hybrid protection mapping was utilized

in analyzing relative transcription levels. The probe employed in this analysis was a 274

bp end labeled Hpall fragment of the test gene (see text) separated and isolated from an

8% polyacrylamide gel. The S nuclease reactions were as previously described with the

exception that 400 units/ml of enzyme (BRL) was used per reaction at room temperature.


Construction of test and reference genes.

The two chimeric constructs to be utilized as the test and the reference genes were

cloned as illustrated in figure 3-1. The initial step involved the excision of maize Adhl

genomic sequences from pB428 (Dennis et al., 1984) possessing 1,094 base pairs (bp) of

the 5' flanking region, the leader sequences, and the first 213 bp of the transcript coding

region with its subsequent insertion into the polylinker of pUC19. This second plasmid

was then opened within the region responsible for encoding the first exon of the transcript

at a unique SstlI restriction site located at +106 which is 6 bp downstream of the ATG

translation initiation codon (+100). This linearized DNA fragment was then "chewed

back" into the untranslated leader sequence (5' to ATG) by the exonucleic activity of Bal

31 with the subsequent addition of XhoI linkers. The promoter and remaining leader

sequences within the BamHI (-1094) to XhoI linkered fragments produced by this

manipulation were then ligated upstream of the T-DNA 780 gene coding sequences carried

within a pUC19 plasmid vector (pUC19:5' del) (Bruce and Gurley, 1987). The resulting

chimeric constructs were sequenced through the Adhl promoter/780 gene junction

whereupon two likely candidates to be utilized as the test and reference chimeric genes

were identified (Fig. 3-2).

The chimeric gene used as a test gene consisted of the Adhl promoter XhoI-

linkered within the leader sequence at +78 and positioned upstream of the 780 coding

region (Fig. 3-2C). A second chimeric gene used as the reference gene was comprised of

the Adhl promoter possessing sequences -1094 to +65 (65 base pairs into the leader)

situated directly upstream to the 780 coding sequences with no intervening linkers (Fig. 3-

2C). The reason this particular construct lacks the XhoI/SalI linkered junction is not

known, but this configuration of sequences achieved the desired objective of altering the

leader sequences compared to those of the test gene, and was advantageous to the desired

I Bun, HI
M1dilH KdiP Il


pUC19 HMdll
I* olhdoe fragmlot

(-1094) I 06)
Ban HI A0 !nOdlll


Xh i
loolato ffagment


B~nH1Sph I
BanHI pAdh780-T/R Spdil
4.8 kb


Fig. 3-1. Construction of test and reference genes. Shaded regions of plasmids
represent DNA fragments harboring either Adhl or the transcribed region of the T-DNA
780 gene. Plasmid pB428 is described in chapter 2 and pUC19:5' del. is a pUC19
derivative containing the 19 bp leader, coding sequences, and about 1.2 kb of 3' flanking
sequences of the 780 gene (Bruce and Gurley, 1987). Important restriction sites are as
labeled; the numbers in parenthesis designate position with reference to the Adhl
transcription start site and identify the 780 start site. Orientation of transcription of genes
are designated by arrows within plasmids. Legend: Apr, ampecillin resistance; S/X, union
of Sal and XhoI restriction sites by ligation.

DaI 31
rits I Unrkw






BS 040

* :C
e *A






40 T

Test gene


maize AdhI 780 coding
promoter linker sequences
+54 +65 +78 -1 +1 +10

maize Ad

Reference gene


22 bp

er 780 coding
54 +10 sequences

Fig. 3-2. Sequence identification of chimeric constructs to be used as the test gene
(A) and the reference gene (B). Panel C compares difference between the test and
reference genes at the Adhl promoter/780 coding region junction. Shaded region
highlights the 22 bp difference between the two constructs. Boxed sequences represent
ligation of Sall and XhoI linkers. Underlined sequences within box indicate
complementary bases of the two linkers.


outcome. The test gene has 22 base pairs at the Adhl/780 junction that the reference gene

does not possess.

Analysis of test and reference gene products.

The reference gene and the test gene were each cloned into the shuttle vector pW9

as described in chapter 2. Construction of the double gene reference vector required that

the reference gene become a permanent resident of the shuttle vector while maintaining

unique sites for the insertion of a test gene. This required that the pUC19-derived plasmid

pAdh780R harboring the reference gene be opened at the BamHI (-1094) restriction site

and changed to a Sall site by the addition of linkers. This construct was itself double-

digested with Sall and SphI restriction endonucleases whereupon the entire chimeric

reference gene was isolated and subcloned into like restriction sites in pW9 giving rise to

the reference gene shuttle vector pRef-W9 (Fig. 3-3). The reference gene was positioned

in this shuttle vector such that the 5' most region of the promoter was adjacent to

sequences derived from pACYC184 (Chang and Cohen, 1978) and the coding region is

upstream of the T-DNA portion. Similarly, the test gene harbored within pUC19

(pAdh780T) was isolated after restriction digestions at BamHI and HindIII and then

subcloned into pW9 giving rise to the test gene shuttle vector pTest-W9 (Fig. 3-4). This

second shuttle vector differed from pRef-W9 in the position and location at which the

Adhl chimeric construct was inserted. The promoter sequences were adjacent to the T-

DNA region of pW9 and the pACYC184 sequences were in turn downstream of the test

gene coding region. These two shuttle vectors were incorporated in separate experiments

into the T-DNA (T-left) of the Ag5260 Ti-plasmid and transferred to the genome of

sunflower seedlings from which resulting tumor tissue was harvested. Primary tumor

tissue possessing each of the constructs were either uninduced (no anaerobic stress) or

subjected to an anaerobic environment for 6 hr (induced). Transcription products from

each tumor tissue type were then subjected to Si hybridization analysis.

Sph I



Sph I
Isolate fragment

Sph I

Sph I


Reference C pACYC1 84
9.8 kb

Sph I 4 BamH1
(13220) 6b a (9062)

Fig. 3-3. Construction of the pRef-W9 shuttle vector harboring the reference
gene. Location and transcriptional orientation of represented genes are designated by
arrows within plasmids. Origin of DNA for each plasmid and important restriction sites
are as labeled. Numbers within parenthesis designate position of restriction sites within
octopine-type A. tumefaciens 15955 T-DNA (Barker et al., 1983). Legend: Cpr,
chloramphenicol resistance; S/X, Sall/XhoI junction; ocs, 4, 6a, and 6b are T-DNA genes
of A. tumefaciens 15955 (Willmitzer et al., 1981). Genes with dotted lines (gene 4 and
ocs) lack promoter sequences.

Bam H1

Bam HI
Hin dill
SIsolate fragment

Hin d II
i Barn Hi
41 (9062)

Bam H1

Sph I

Bam HI
Hin dill
*Isolate fragment

Sph I
Hin dill

Fig. 3-4. Construction of the pTest-W9 shuttle vector harboring the test gene.
Important regions and parameters are as previously described.

Sph I

4.2 kb

Sl hybrid protection mapping was employed to accurately quantify promoter

activity of mutated test genes. The probe employed in these experiments was a 283 base

pair Hpall DNA fragment isolated from the test gene and end labeled. It comprised 138

bases extending upstream of the recognized start of transcription of the Adhl promoter, 78

base pairs of the Adhl leader, 8 base pairs from the added Sall and XhoI linkers, a single

non-transcribed base (-1) and the first 59 bp of the transcribed region of the 780 gene

(Fig. 3-5). The S1 protection strategy required transcripts produced by the test gene in

sunflower tumors to hybridize and protect 146 bases of the end labeled DNA probe to the

Adhl start of transcription (Fig. 3-5A). In contrast, transcripts coded for by the reference

gene bound the same probe, but due to the lack of 22 complementary bases within the test

transcript, 22 bases of the probe looped out. This single stranded loop exposed these

non-complementary bases of the probe to S 1 nuclease attack producing two protected

fragments 65 and 59 bases in length with only the 59 base fragment being end labeled

(Fig. 3-5B).

The reference gene and the test gene that had been subcloned into separate shuttle

vectors (Figs. 3-3 and 3-4) were separately integrated into the T-DNA of a mutated form

of Agrobacterium tumefaciens 15955 (Ag5260) in which the endogenous 780 gene of T-

right had been excised and a kanamycin resistant gene substituted at this position (Bruce

and Gurley,1987). This lack of the endogenous 780 gene negated any interference in

analysis of the transcription products of the test and reference genes. Messenger RNA

collected from 6 hr anaerobically stressed sunflower tumors that had been infected by the

Agrobacterium separately harboring either of these two constructs were distinguishable

from each other in the predicted manner when subjected to S 1 nuclease mapping (Fig. 3-

6). The transcripts produced by the test gene protected 146 bases of the probe with no

band appearing at the location predicted for the reference gene (Fig. 3-6, lane 4). This

suggests that the chimeric test gene utilized the same transcription start site in sunflower

tumors as the endogenous Adhl gene in maize and tumor cells (see chapter 2). Messenger

aa3 -

0 C>


C U0 M

;.0 o
t ) 4 -0



u .- C
a .o

N -5 4)
5. S S

.- t C
0U C4
O @ 3 &
c It3 "f

(U< c"

10 2
C *

z z

< 4c
z z

+ IL





S" E
o Q0
0 .

< Tr CO

118- M

full length
-- probe

W test

4- reference

72 4

Fig. 3-6. S1 nuclease hybridization analysis of the reference gene and the test
gene. Hybridization of end labeled DNA probe previously described was with 1/10 (lanes
1,3, and 4) or 1/5 (lane 2) of poly(A)+RNA isolated from 25 to 40 grams of sunflower
tumor tissue. Poly(A)+RNA was collected from uninduced tumors harboring the
reference gene (lane 1) and induced tumors possessing the reference gene (lanes 2 and 3)
or the test gene (lane 4). Protected fragment are designated. Lane "0" contains end
labeled HaeII generated 4X174 markers with size designations.

RNA from tumors harboring the reference gene in turn protected 59 bases of the

end labeled probe (Fig. 3-6, lanes 2 and 3) indicating that the 22 base loop of the probe

was degraded by S1 nuclease activity. No transcript protection was observed for

reference gene products from uninduced tumor tissue (Fig. 3-6, lane 1). These

experiments indicate that not only was the natural transcription start site of the Adhl

chimeric constructs still recognized in sunflower cells, but the promoter was still

responsive to anaerobic stress. These preliminary experiments demonstrated that the

transcript products of each of these genes could be differentiated by hybrid protection

mapping using a single DNA probe.

Analysis of test gene promoter mutations.

As a first step in the identification of cis-elements within the Adhl promoter, a

number of 5' deletions of the test gene were constructed and assayed in sunflower. Two

of these mutant constructs utilized unique restriction sites within the promoter, XbaI

(-410) and PstI (-140). The initial deletion construct involved the digestion of the pUC19

vector harboring the test gene, pAdh780T, with XbaI and the subsequent addition of Sail

liners to this site. This construct was then double-digested with endonucleases Sail and

HindIII and subcloned into the polylinker of pUC19 giving rise to pAdh780T2 (Fig 3-

7A). These manipulations resulted in a test gene possessing 410 bp of the most proximal

upstream sequences to the start of transcription along with the acquisition of restriction

sites from the polylinker of pUCl9 immediately upstream to the gene. Particularly

important was the BamHI site of the polylinker that was necessary for subsequent cloning

into the shuttle vector. A similar procedure was used to produce a deletion to -140 by

subcloning a PstI (-140) to HindIII fragment of pAdh780T into pUC19 resulting in

pAdh780T3 (Fig. 3-7B). Additional 5' deletions of the test gene were constructed by Bal

31 exonuclease digestions initiated at the BamHl (-1094) or Xbal (-410) sites of the

promoter (pAdh780T) with the subsequent addition of Xhol linkers producing a series of

(-410) Pst I
iXbs 1 780

Blunt in dlli
Bam HI pAdh 780T HIndlil
(-1094) 4.8 kb


ZblunIIHin 4
l I liners slate frogmen
Pst I
(-410) Pst I

Bm Hlndilll I
(-1094) Hin ndll

L P 1 PUC19
Sal I
isolate fragment (-140)
Pst I S/X

Pst I
sa u o pSpo h


Pst I S/X
(-410) SaHll

pAdh 780T2 n Hindll
4.17 kb

\ pUC19

Fig. 3-7. Construction of 5' promoter deletion test genes pAdh780T2 to -410 (A)
and pAdh780T3 to -140 (B). Important sites and descriptions of plasmids are as
previously presented.

truncated test genes that were then subcloned into theSall and HindIII sites of pUC19
(Fig. 3-8). Location of the deletions were estimated by comparison to OX174 HaeIII

sizing markers on 8% polyacrylamide gels (data not shown). These Bal 31 generated

mutants resulted in test genes having 5' flanking sequences extending to approximately

-1000, -920, -850, -432, -324, and -288 relative to the transcript initiation site. Upon

insertion into pUC19 each of the exonuclease generated 5' deletion genes obtained

restriction sites (including BamHI) at the 5' end point from the polylinker (Fig. 3-8).

The eight 5' deletion test gene mutants, along with the original test gene that

retained the wild type promoter to -1094, were subcloned into the BamHI and HindIII

restriction sites of the pW9 shuttle vector derivative pRef-W9 (Fig. 3-9). This cloning

strategy resulted in a double gene shuttle vector (illustrated as pAdh-780TR-W9 in figure

3-9) that harbored the reference gene possessing the wild type Adhl promoter along with

a mutated test gene. Upon transfer into Agrobacterium 5260 and incorporation of the

shuttle vector into the Ti-plasmid, both genes along with T-DNA genes of the vector are

integrated into the T-DNA. Their position and orientation with respect to the naturally

occurring genes of T-left are presented in Fig. 3-10.

Tumor tissue that formed after infection of the sunflower seedlings with the

different derivatives of Ag5260 harboring the test and reference genes were anaerobically

stressed for 6 hr. Poly(A)+RNA isolated from the tumors was subjected to S1 protection

analysis as previously described. The poly(A)+RNA collected from anaerobically stressed

tumor tissue protected detectable levels of the end labeled probe representative of both the

test and reference genes (Fig. 3-11). Quantification employed the direct comparison of

each of the different test gene transcription products with transcripts produced by the

reference gene. The ratio of mRNA transcribed by the test gene possessing the wild type

Adhl promoter (-1094) versus the transcription products of the reference gene, also

possessing the Adhl wild type promoter, was arbitrarily assigned a relative transcription

level (RTL) value of 100. The average transcript level ratios and RTLs of the different

Bam H1 pA

Bal 31

Sph I
Hin dlll

Sph I

Sph I

Bam HI

Xho I
Hin dill
slate fragment
Pst I
Xho I

Sph I
pUC19 Hindll
Sail Hindlll

IXho I
Hin dill
isolatee fragment



Fig. 3-8. Construction of test gene 5' promoter deletion series. Important sites
and descriptions of plasmids are as previously presented.


Xho I

gene I

Sph I

Bam H1

Sph I
Hin dll

Fig. 3-9. Construction of the double gene shuttle vector pAdh780TR-W9.
Important sites and descriptions of plasmids are as previously presented.

40 c 0.E
C14 0 : 3 x 3
r 4- 0 0 0

0 0
c c ur

4 ) 4 > ) C
H: O '"E 0 4) g 0

C14 04 004)
4 2 O'= >

=.r-- 0

N Cd
0 ed 0

4.) 0 0 %
bG ~ ), 0 r >'.

4 0 0 0 0~ cu > 'I '
~= 4.~

r rA 00r
C sd w

0 b OO-o
z0 c.C
nsooz 00 u r

cd = 0 0 1

,a M S
18 "C d i

< '4:: 0 P4 C 4 -
4- 1-4 U U

cd = = u > 0 0 C
Su Ldo\ '

o brr
10 0 cd d r.(D0
044 r 4.. q
bi ed -04-40 r_

o ar. o E 06~
'd > 0 cq
G 0 R C7 4.)
~~Cd E 4-4 =











4- probe

M OW m 4

- -

m oft 4

4W M M

-- reference

Fig. 3-11. S1 hybridization assay of reference gene and 5' deletion test genes.
Each lane contains 1/10 of the poly(A)+RNA isolated from 25 to 40 grams of tumor tissue
that had been subjected to anaerobic stress for 6 hours. The sunflower cells were infected
by A. tumefaciens 5260 with incorporated double gene shuttle vector pAdh780TR-W9
(see Fig. 3-10). Lane designations represent 5' most point of test gene promoter with
respect to start of transcription. Protected fragments are designated.

-- test


Relative Transcription Levels (RTL) of 5' Deletion Mutants

5' border Number of Test/reference
of test gene experiments ratio avg. RTL

5'A-1094* 2 1.07 100
5'A-1000 1 0.90 84
5'A-920 2 1.08 101
5'A-850 2 1.34 125
5'A-432 2 0.69 65
5'A-410* 2 1.02 95
5'A-324 1 0.81 75
5'A-288 2 0.65 61
5'A-140* 2 0.77 72
*deletion end-point determined by DNA sequencing, others were
estimated by polyacrylamide gel analysis using restriction fragment
size markers

mutated test genes versus the constant reference gene were then calculated in relation to the

100 assigned the wild type. Table 3-1 summarizes the ratio and RTL values for each of

the 5' deletion mutants surveyed.

The most notable observation concerning the RTL values for the 5' deletion

mutations is that all of the RTLs remained relatively high. Initially, as 5' sequences are

deleted there was an actual rise in transcriptional activity. Removal of about 250 bp

(5'A-850) of the most distal 5' upstream sequences increased activity approximately 25%

above that of wild type. A more severe deletion to position -432 (5'A-432) lowered

transcriptional activity to an RTL value of 65, while an additional small deletion to -410

raised the RTL to that of wild type. The 5' deletions of the test gene promoter to -324,

-288, and -140 resulted in similar RTLs ranging from 61 to 72. These results suggest that

negative elements may exist within the upstream sequences of the Adhl promoter within

the small region between (approximately) -432 and -410. The 400 base pairs between

-850 and -432 appear to be required for maximum transcription when this putative

negative region is present, since a deletion to 5'A-432 markedly decreased the activity.

Upon deletion of the potential negative region, the transcript level was boosted to that of

wild type. A second possible explanation for this effect is that of a position effect of the

test gene within the shuttle vector. Unlike the reference gene utilized in these experiments

and the Adhl single gene analysis of chapter 2, the promoter of the test genes were

inserted into the shuttle vector in such a way as to be positioned immediately downstream

of the T-DNA (Fig. 3-9). Any activator-like sequences within the T-DNA region would

be in fairly close association with the test promoter and may have an effect upon the 5'

deletion analysis of the constructs. That is, as more sequences are deleted from the 5' end

of the promoter, the closer the remaining sequences and regulatory elements become

associated with the T-DNA.

Analysis of test genes transferred via shuttle vector p195.
The 5'A-140 deletion RTL of 72 closely corresponds to results observed by

Walker et al. (1987) when the maize Adhl promoter was fused to a CAT reporter gene

deleted to this same position and analyzed in maize protoplasts. However, these results

differ from lower transcription levels (<20% of wild type activity) observed for this

construct in maize protoplasts by Lee et al. (1987) and the preliminary results observed

concerning the maize Adhl mutants in sunflower tumors presented in this study (see

chapter 2). In an attempt to investigate any effect the T-DNA may have upon analysis of

the promoter deletions after incorporation of the shuttle vector into the Ti-plasmid,

sequences of the T-DNA situated upstream of the incorporated test genes were examined

more closely. As can be seen in figure 3-10 test genes assimilated into T-left are

downstream of the 3' coding region of the T-DNA gene 4 which codes for an enzyme

involved in cytokinin production, gene 6a which is involved in secretion of opines

(Messens et al., 1985), the 6b gene thought to control tumor size in some plant species

(Hookykaas et al., 1988), and the octopine synthase (ocs) gene (Hack and Kemp, 1980).

Upstream of the coding region of 6a is a region of 16 bp, ACGCAGCCGCTITCGA, that

is 67% homologous (11 of 16 bases) to the ocs enhancer-like element (Ellis et al., 1987b).

This sequence is located between -514 to -499 (9444 to 9459 in T-left, Barker et al.,

1983) relative to the transcription start site of 6a and within the 3' coding region of gene 4.

The ocs element-like sequence of genes 4 and 6a is 382 bp from the T-DNA/test

gene junction of the shuttle vector. Since these sequences are similar to the ocs enhancer-

like element, and due to its close proximity to the test genes, a region of the shuttle vector

was removed in an attempt to abolish any affect it may have had on transcription of the test

genes. An 858 bp PstI fragment within the T-DNA portion of the shuttle vector (pRef-

W9) was excised to give rise to pRef-W9-P (Fig. 3-12A). This second shuttle vector

lacked the promoter region for the 6a gene and specifically the 16 bp ocs-like element.

The reference gene and the T-DNA of pRef-W9-P were subsequently subcloned into the



Sph I



Fig. 3-12. Shuttle vectors lacking possible PstI fragment harboring potential ocs-
like element. (A) pACYC184 derived shuttle vector pRef-W9-P. (B) pSUP205 derived
shuttle vector p195. Important sites and descriptions of plasmids are as previously
presented with the exception that 6a/4 contains a 3' portion of the gene 4 coding region
immediately upstream of 3' coding sequences of 6a.

Bam HI and Sall restriction sites of pSUP205 (Simon et al., 1986) resulting in a third

shuttle vector, p195 (Fig. 3-12B). The 7.7 kb pSUP205 is a mobilizing plasmid

possessing the mobilization (Mob) site from the broad host range plasmid RP4 cloned into

the E. coli vector pBR325 (Bolivar et al., 1978). This Mob site was utilized during

tripartite matings to increase the transfer efficiency of the shuttle vector into the recipient


The test gene possessing the Adhl wild type promoter, 5'A-1094, and two 5'

promoter deletion mutants, 5'A-410 and 5'A-140, were inserted into the BamHI and

HindIII restriction sites of p195. This gave rise to double gene shuttle vectors p195-

5'A-1094, p195-5'A-410, and pl95-5'A-140, each possessing the standard reference

gene (Fig. 3-13). The test and reference genes were transferred into Ag5260 and

incorporated into the T-DNA (Fig. 3-14). The association of each of the chimeric genes

carried by the double gene vector p195 with surrounding DNA after incorporation into the

Ti-plasmid differed from those genes that had been transported via the pRef-W9 vector

(Fig. 3-14). The promoter of the test genes was again positioned adjacent to the shuttle

vector T-DNA but these sequences now lacked the 6a promoter and the 3' proximal most

sequences of the gene 4 coding region. Also the test and reference genes were separated

by about 7.3 kb of the pSUP205 sequences rather than the approximately 3.6 kb of

pACYC184 sequences derived from the pRef-W9 shuttle vector.

The Ag5260 transconjugants from the pSUP205 derived vectors harboring the

reference gene and the test gene promoter mutants in the T-DNA of the Ti-plasmid were

used to infect the hypocotyls of sunflower seedlings. The poly(A)+RNA isolated from the

tumors anaerobically induced for 6 hr was subjected to S 1 nuclease protection mapping

after hybridization to the Hpall end labeled probe isolated from the test gene as previously

described. The test and reference gene products were quantified and RTL derived for each

experiment. The average ratio of the wild type test gene products versus the reference

gene products was given the RTL value of 100. The average ratio of each of the deletion

Fig. 3-13. Double gene shuttle vector pl95-Test. Important sites and descriptions
of plasmid are as previously presented.


Cl cu


a 4
.-? a4-


U -!





0 -

0.. U









lo -


gene products versus the internal standard were compared to the wild type ratios and

assigned their own corresponding RTL value (Fig. 3-15; Table 3-2).

The test gene that had been deleted to the -410 position (195-5'A-410) did not

exhibit decreased activity, nor result in an RTL value similar to that the of wild type

promoter as was previously observed in experiments utilizing the pW9 derived shuttle

vector (Fig. 3-11; Table 3-1). Rather, the RTL was more than double the wild type.
Similarly, the RTL of the more drastic 5' deletion mutant to -140 (195-5'A-140) also

resulted in an RTL value 2-fold higher than wild type, but was slightly less than that of

195-5'A-410. This result may have been due to one of several possibilities. One is that at

a negative element exists within the region between -1094 and -410. Deletion of these

negative regulatory sequences may then have allowed for increased transcription levels of

the genes. A second explanation invokes the possibility of a position effect of the test

gene in relation to T-DNA of the shuttle vector, with a third possibility involving a

combination of both, deletion of a negative element region and position affect. Removal

of these same promoter sequences of the test genes transferred to Ag5260 by pRef-W9 did

not give rise to an increase in the RTL compared to the wild type RTL, but rather a slight

decrease (Table 3-1).

To further investigate the probability of position effect upon the test genes, the use

of filler DNA from the E. coli vector plasmid pBR322 (Bolivar et al., 1977) was

employed. A 1,125 bp BamHI to PstI fragment was subcloned upstream of the 5'A-140

test gene deletion mutant in pAdh780T3. This test gene along with its newly acquired

1,125 bp filler DNA was subsequently cloned into the BamHI and HindIII sites of the

double gene shuttle vector p195 (Fig. 3-16). This test gene is similar in content and

orientation within the p195 shuttle vector as 195:5'A-140, but is separated from the shuttle

vector T-DNA by the 1,125 bp insertion (195:5'A-140BR in Figs. 3-16 and 3-17). After

incorporation of 195:5'A-140BR into T-DNA, inoculation of seedlings, tumor formation,

and RNA extractions the mRNA from both induced (6 hr) and uninduced primary tumors





t- probe

4- test


72- -

- -*- reference


-410 -140










Fig. 3-15. S1 hybridization assay of reference gene and 5' deletion test genes
transferred into Ag5260 via p195. Poly(A)+RNA was isolated from tumor tissue that had
been subjected to anaerobic stress for 6 hours. Lane designations represent 5' most point
of test gene promoter with respect to start of transcription. Lanes 195:5'-1094,
195:5'-410, and 195:5'-140 contain 2.4 gg, 2.9 gg, and 18.8 ig of poly(A)+RNA,
respectively. Illustrations of each test gene in relation to T-DNA and the RTL value for
each are presented.





RTL Values Shuttle Vector p195
5' border Number of Test/reference
of test gene experiments ratio avg. RTL
5' A-1094 6 1.72 100 10.7
5'A-410 7 3.79 220 40.1
5'A-140 6 3.30 192 20.0
* standard deviation given as

pAdh 780T3 indll
3.9 kb


Bam HI
Isolate fragment

(-140) I
Bam H

pAdh 780T3 Hinal
3.9 kb


Eco Hin d

,k 3 kb

Hin dll
S. blunt



isolate fragment

Pall 4V 1125bp

Paet I

BOM t"I pAdh 78OT4


p195 I. isolate fragment

Fig. 3-16. Construction of test gene 5'A-140BR that has 1,125 bp of filler DNA
positioned upstream of the 5' deleted test gene to -140 and its incorporation into the p195

shuttle vector. pBR322 sequences are represented by lightly shaded region, 5'A-140 test
gene sequences by intermediate shaded region, and T-DNA by black region. Important
sites and gene designations are as previously presented.

Fig. 3-17. S1 hybridization assay of reference gene and 5' deletion test genes
transferred into Ag5260 via p195. Poly(A)+RNA was isolated from tumor tissue that had
been subjected to anaerobic stress for 6 hours. Lane designations represent 5' most point
of test gene promoter with respect to start of transcription and 1,125 bp of filler DNA
positioned upstream of the 5' deleted test gene to -140 designated by 195:5'-140BR.
Lanes 195:5'-1094, 195:5'-410, 195:5'-140, and 195:-140BR contain 2.4 pg, 2.9 gg,
18.8 gg, and 4.9 lg of poly(A)+RNA, respectively. Protected fragments are designated.
Illustrations of each test gene in relation to T-DNA and the RTL value for each are
presented with the pBR322 sequences represented by the hatched region.






279-* I

194. -

I, -


.- probe

-, '- test

118 .-


S-.- reference



-140 +1

| 780

* *////////////////_////////-/// I
T-DNA pBR322 1125 bp
pBR322 1125 bp






was subjected to S 1 hybrid protection analysis. While the test and reference gene

transcripts from the uninduced tissue was negligible (data not shown), detectable

transcript levels from induced tumors protected the end labeled probe (Fig. 3-17, lane 4).

Five separate experiments resulted in an average test/reference gene product ratio of 0.863

which converts to an RTL value of 50. This RTL is half that of the wild type test gene

and one fourth the level of the same 5' deletion lacking filler DNA (Fig. 3-17). Therefore,

by separating the -140 test gene from T-DNA by 1,125 bp of filler DNA the transcript

level dropped by almost 75%. These results indicate that T-DNA situated upstream

exerted a positive effect on the test gene transcription. The ocs element is most likely the

cause of this enhancement which was inversely proportional to distance.

Analysis of test genes transferred via shuttle vector p251.

The ocs element has not been shown to affect transcription in tobacco calli when

positioned 3' to a 5'-truncated ocs gene (Leisner and Gelvin, 1989), or 3' to an Adhl

promoter/CAT reporter gene in maize protoplasts (Ellis et al., 1987b). With this in mind,

in an attempt to develop a sunflower tumor assay system for the AdhlFl80 chimeric test

and reference genes free of any affect on transcription from the T-DNA, the shuttle vector

was re-designed such that the T-DNA portion would be 3' to the test gene as well as the

reference gene. The reference gene and T-DNA from pRef-W9 was subcloned into the

Sall and BamHI restriction sites of pSUP205. The resulting plasmid was opened at the

single HindIII site of the pSUP205 sequences, filled in, and blunt end ligated (Maniatis et

al., 1982) subsequently removing this HindII site. This plasmid was in turn opened at a

single Asp 718 (KpnI) restriction site within the T-DNA portion 776 bp from the T-

DNA/pSUP205 junction whereupon HindII linkers were inserted (Fig. 3-18). The

resulting shuttle vector, p251, possessed pSUP205 sequences that contributed a

chloramphenicol resistance gene and the mobilization site (Mob) for increased conjugation

efficiency, the reference gene, and the T-DNA necessary for recombination into T-left of

pSUP205 pRet-W9
7.7 kb 9.8 kb
W4 l ndilll
ndE Sphi IamH1
H1 (13220) 6b (9062)
T- DNA Aep 71
Sail Sall
SIsolate fragment Isolate fragment

Sa Cp

Spi 6b6a BamH1

T-DNA A7p 718


A.p 718

Aap 718
endig linkers

ocs aH
Sph I b BamH1

T-DNA Amndl1

Fig. 3-18. Construction of the shuttle vector p251. Important sites and
descriptions of plasmids and intermediates are as previously presented.
descriptions of plasmids and intermediates are as previously presented.

pTi-Ag5260. The newly acquired HindIII restriction site within the T-DNA allowed for

incorporation of test genes into the shuttle vector in the opposite orientation of any of the

previous double gene vectors (Fig. 3-19).

Test genes with the wild type promoter, 5' deletions to -410 and -140, and the

-140 5' deletion mutant with the pBR322 filler sequences were designated test genes 1, 2,

3, and 4 respectively (Fig. 3-21). Each of these test genes was subcloned into p251.

These double gene vectors were transferred into Ag5260 and incorporated into the T-DNA

(Fig. 3-20). The promoters of the test and reference genes were separated by the 7.3 kb

pSUP205 sequences and the T-DNA with the 6a, 6b, and ocs genes (and therefore the

ocs element) now situated 3' to the inserted test genes. Poly(A)+RNA was collected from

both uninduced and induced (6 hrs anaerobic stress) tumors formed by sunflower

seedlings infected with the Agrobacterium. Results of Sl hybridization protection

mapping of the transcripts hybridized to the test gene HpalI end labeled probe are shown

in figure 3-21. There were negligible levels of protected DNA with mRNA isolated from

each type of the uninduced tumor tissue. Test and reference gene transcripts from the

induced tumors protected probes which were quantified and test/reference ratios calculated

(Table 3-3).

The average wild type promoter test/reference gene ratio incorporated into pTi-

5260 utilizing the double gene vector p251 was assigned an RTL value of 100. Test gene-

2 that has 684 bp of the most distal sequences deleted to -410 had an RTL of 88 reflecting

an approximately 10% reduction from wild type promoter activity. The more extreme

deletions of test genes 3 and 4 to -140, without or with upstream spacer DNA from

pBR322, resulted in more drastic reduction of RTL values to 29 and 33 respectively (Fig.

3-21). These values markedly differed from previous results utilizing the double gene

vectors pRef-W9 and p195, each of which had higher RTL values for the 5'A-140

deletion. The test to reference gene product ratio of 5'A-140BR (5' deletion with 1125 bp

of spacer DNA) transferred by p195 (Fig. 3-16; 195:5'A-140BR) was directly compared

Bam H1

Bam Hi
Isolate fragment

Bam H1
Isolate fragment


Barn H1

Sph I
Hin dill


Fig. 3-19. Construction of the p251 double gene shuttle vector derivatives.
Important sites and descriptions of plasmids and intermediates are as previously presented.



to 0

c c

0 0
e- ~




r SR

o o-

4- 0

i~4 U