In vitro and functional analysis of the Arabidopsis Adh 5' flanking sequence

MISSING IMAGE

Material Information

Title:
In vitro and functional analysis of the Arabidopsis Adh 5' flanking sequence
Physical Description:
v, 168 leaves : ill., photos ; 29 cm.
Language:
English
Creator:
McKendree, William Lee, 1954-
Publication Date:

Subjects

Subjects / Keywords:
Arabidopsis   ( lcsh )
Cruciferae -- Genetics   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 159-167).
Statement of Responsibility:
by William Lee McKendree, Jr.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001747383
notis - AJG0206
oclc - 26371926
System ID:
AA00003730:00001

Full Text













IN VITRO AND FUNCTIONAL ANALYSIS OF THE
ARABIDOPSIS ADH 5' FLANKING SEQUENCE

















BY

WILLIAM LEE MCKENDREE, JR.


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

UNIVERSITY OF FLORIDA


1991
















ACKNOWLEDGMENTS


I thank my supervisory committee chairman, R. Ferl,

for making this work possible, and for continued encour-

agement and support, regardless of the weather. I also

thank my committee members, Dr. T. E. Humphreys, Dr. L.

C. Hannah, Dr. W. Gurley, and Dr. H. Nick, for their

thoughtful consideration and their time. Beth Laughner,

Joe Nairn, Anna-Lisa Paul, Tsako Hatori, John Baier,

Leonard Rosenkrantz, John Ingersoll, Kevin O'grady, Alice

Delisle, Dulce Barros, and Mohammed Ashraf have all

contributed significantly to this project, and to my

personal growth. I sincerely thank my wife, Lisa, and my

son, Mathew, for their continued love and faith through

good times and bad. I also thank my parents for their

love and concern. Finally, I thank my God and His Son

Jesus, who make all things possible.

















TABLE OF CONTENTS


ACKNOWLEDGMENTS................................ ii

ABSTRACT ........................................ iv

CHAPTER 1
INTRODUCTION ............. .... ................. 1

CHAPTER 2
CHARACTERIZATION OF PROTEIN INTERACTIONS WITH
ARABIDOPSIS ADH 5' FLANKING SEQUENCES ............29

CHAPTER 3
FUNCTIONAL ANALYSIS OF THE 5' FLANKING REGION
OF ARABIDOPSIS ADH................................52

CHAPTER
SUMMARY ........................................140

APPENDIX A......................................154

REFERENCES....................... ...............159

BIOGRAPHICAL SKETCH.............................. 168
























iii
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree Doctor of Philosophy


IN VITRO AND FUNCTIONAL ANALYSIS OF THE
ARABIDOPSIS ADH 5' FLANKING SEQUENCE

by

WILLIAM LEE MCKENDREE, JR.

December, 1991


Chairman: Robert J. Ferl
Major Department: Botany

The Adh (alcohol dehydrogenase) gene of Arabidopsis

is expressed constitutively in immature seedlings and

cells in suspension, and may be induced in roots of

mature plants. In vivo DMS footprinting of the 5' flank-

ing region of this gene has identified sites of protein-

DNA interaction which include the G-box element also

found in several light regulated genes (RbcS, Cab, and

Chs). The Adh G-box binds specifically to a protein

component of Arabidopsis extracts from cell culture and

mature leaves, and mutations of footprinted G-box bases

partially or completely disrupt binding. Corresponding G-

box mutations in the full length Adh promoter resulted in

>60% reduction in reporter gene activity when assayed in

Arabidopsis protoplasts. Adh promoter mediated expres-

sion was observed in both 4-day-old and 15-day old Arabi-










dopsis seedlings transformed by microprojectile bombard-

ment. Deletions of the Adh promoter were assayed in

Arabidopsis seedlings by particle bombardment and in

Arabidopsis protoplasts. Sequence domains which are

necessary for wild type gene activity were confined to

first 390 bp 5' to the transcription start site and

included all sites of in vivo protein-DNA interaction

previously identified by Ferl and Laughner (1989).

Deletion of 5' sequences from -390 to -289 decreased

activity by 50%, and further deletion to -177 resulted in

a decrease to less than 10% of full gene activity.
















CHAPTER 1
INTRODUCTION


Terrestrial green plants keep their photosynthetic

structures more or less fixed in space and are therefore

faced with the need to adapt to a variety of environmen-

tal stresses, as opposed to simply relocating to a more

favorable environment. Those plant species more able to

sense and respond to the stresses of their environment

are therefore selected through evolution. This unique

feature of green plants has provided a variety of model

systems for the inquiry of molecular bases for stress

adaptation.

Of the many hazards which befall the stationary organism,

flooding is one of the most life threatening. Molecular

oxygen concentrations in water preclude aerobic metabo-

lism for the organ of concern, and alternate biochemical

pathways which do not require oxygen are needed for the

flooding duration. Early work into the basis of this

flooding response in Zea mays (Hageman and Flesher, 1960)

led to the identification of the alcohol dehydrogenase

(ADH, E.C. 1.1.1.1.) enzymes (Schwartz and Endo, 1966;

Schwartz, 1966) and eventually to the cloning of genes

which encode them (Gerlach et al., 1982; Dennis et al.,

1984). ADH isozymes regenerate NAD+, thus allowing










glycolysis to continue in the absence of terminal oxida-

tion. Anaerobic ATP production is adequate to allow the

plant to survive limited periods of stress.

The means by which these stress induced genes are

regulated with regard to activity and cell type will be

the subject of this investigation. ADH is expressed in

cell types possessing elevated rates of metabolic activi-

ty in addition to oxygen deprived cells of the mature

root. ADH activity has been observed in a number of

higher plants including wheat, soybean, beet, rice, pea,

barley, tomato, tobacco, corn, pearl millet, and Arabi-

dopsis (from Dolferus and Jacobs, 1991). Expression

patterns of Adh vary slightly among those plant species

investigated, yet retain an overall similarity which

reflects Adh functionss.

The expression patterns of Adh in Zea mays and

Arabidopsis thaliana will be summarized briefly. This

will be followed by a review of 5' flanking sequence-

mediated mRNA synthesis and a comparison of Adh tran-

scriptional regulation among relevant genes and organ-

isms. Emphasis will be placed on protein/DNA interactions

within specific 5' flanking sequences and their apparent

ability to mediate cell-type specific, developmentally

regulated and inducible expression of Adh.











ADH Expression Characteristics


Zea mays

ADH activity in corn arises from a two-member gene

family, Adhl and Adh2. ADH enzymatic activity resulting

from one or both genes is found in the root, scutellum,

embryo, aleurone, endosperm, pollen, stem pith and nodes,

and the etiolated leaf (Bailey-Serres et al., 1987), as

well as cells in suspension (Paul and Ferl, 1991). The

level of ADH expression in roots, scutellum, immature

endosperm, and embryo increases under anaerobic stress.

Adhl and Adh2 differ in cell-type specific expression and

in their degree of inducibility (Bailey-Serres et al.,

1988; Paul and Ferl, 1991). In general, ADH is active in

tissues with high rates of metabolic activity or where

oxygen is in limited supply.

Adhl and Adh2 are found on chromosomes 1 and 4

respectively (Schwartz, 1971; Dlouhy, 1979). The two

genes are 82% similar in coding DNA and 87% similar in

amino acid sequence (Dennis et al., 1984; Dennis et al.,

1985). The Adh genes encode a 379 aa protein (Dennis et

al., 1985), with the functional ADH enzyme being a dimer

of one or both gene products (Freeling and Schwartz,

1973). Both 5' and 3' untranslated regions of the

maize Adh genes differ considerably in nucleotide se-

quence (Dennis, 1984; Dennis, 1985). However, some short

regions of 5' untranslated sequence are similar between











Adhl and Adh2 and will be considered in a following

section.

Anaerobically induced Adh activity arises from an

increase in Adh mRNA transcription (Gerlach et al., 1982;

Vayda and Freeling, 1986), eventually reaching 20 to 50

times that found in uninduced roots (Ferl et al., 1980;

Gerlach et al., 1982; Rowland and Strommer, 1986; Paul

and Ferl, 1990),. Anoxia is followed by the selective

translation of 20 proteins, including ADH (Sachs et al.,

1980). Once oxygen stress is relieved, Adh message is

rapidly depleted (Rowland and Strommer, 1986).

The two Adh genes differ slightly in their develop-

mental expression patterns in that Adh2 is not detectable

in pollen (Bailey-Serres et al., 1987), whereas Adhl is

abundant. They are, however, transcribed at roughly

equivalent levels in the cob, silk, leaf, node, stem, and

uninduced root of the mature plant (Paul and Ferl, 1990).

Adh mRNA levels in the root increase tenfold or more

following 8 hours of anaerobic induction, and Adh2 mRNA

is present at approximately twice the level of Adhl.

This pattern is also observed for cells in suspension,

where Adh2 mRNA increases to roughly three times that of

Adhl mRNA under similar induction conditions (Paul and

Ferl, 1990).

The rate of Adh mRNA synthesis may be regulated by

protein/DNA interactions within the 5' flanking sequences

of the Adh coding region (Ferl and Nick, 1987; Walker et










al., 1987; Chen et al., 1987; also, see review by DeLisle

and Ferl, 1990). Deletion or mutation of this region of

the Adhl gene results in reduced expression of a reporter

gene when assayed in protoplasts for transient expression

(Howard et al., 1987; Walker et al., 1987; Lee et al.,

1987) or in sunflower tumors for stable expression

(Ingersoll, 1990). A discussion of these and other

related studies will follow.

Regulation of tissue specific expression of Adh by

its 5' flanking sequence has not been conclusively demon-

strated in maize, and dicot species transformed with Adh

via Agrobacterium have given poor results (Ellis et al.,

1987). Indirect evidence supporting tissue specific

regulation by maize Adhl 5' flanking sequence does exist,

however, from an unstable, organ-specific Adhl mutant

isolated from a Robertson (Robertson, 1978) mutator line

(Chen et al., 1987). In this mutant, Adhl-3F1124, ADH1

is expressed at roughly 6% normal levels in seed and

anaerobically treated seedling, but at normal levels in

pollen. A Mu 1 transposable element insertion was found

31 bp 5' to the TATA element and creates a potential new

TATA element. 5' sequence mediated tissue specificity

for Adh expression is probable (but not proven) in that

other tissue-specific plant genes such as RbcS are known

to possess this property (Nagy et al., 1985).

Adh expression studies have been conducted in other

monocot plants, namely rice (Xie and Wu, 1989), barley










(Hanson and Brown, 1984; Hanson et al., 1984; Trick et

al., 1988), Wheat (Susseelan and Bathia, 1982), and pearl

millet (Bannuett-Bourrillon, 1982). In each of the above

mentioned species, Adh is encoded by two or three genes,

activity is induced with anoxia and expression is regu-

lated in a developmental and tissue specific manner.

One notable exception in Adh tissue specific expres-

sion occurs in rice. Adh activity may be induced in

mature roots, embryos, endosperm, etiolated and green

mature leaves (Xie and Wu, 1989). Although green leaves

show relatively low levels of Adh activity following

anaerobic induction, this and all other Adh activity

continues to increase for up to 96 hours. All other Adh

expression characteristics for rice are similar to corn,

including transcriptional regulation and induction of

activity by 2,4-D (Xie and Wu, 1989; Freeling, 1973).



Arabidopsis thaliana


Of the several dicot species in which Adh has been

characterized, Arabidopsis has proven to offer the widest

range of possibilities for analysis of expression both in

vitro and in the whole plant (Meyerowitz and Pruitt,

1985). Arabidopsis has been the system of choice for

gene expression studies due to its small habit and genome

size (70,000 kb)(Leutwiler et al, 1984; Meterowitz and

Pruitt, 1985). Plants may be grown by the thousands in a










single flat and will produce seed 4 to 6 weeks following

germination. Arabidopsis is susceptible to Agrobacterium

mediated transformation (Valvekens et al., 1988; Feldman,

1991) and has been used to characterize the expression of

a number of plant genes. ADH activity in Arabidopsis is

observed in germinating seeds, young seedling cotyledons

and roots (<10 DAP), the anaerobically induced mature

root (>10 DAP), in the stigma and in pollen grains

(Dolferus and Jacobs, 1985; Dolferus and Jacobs, 1990).

As in corn, ADH activity is not observed in green plant

parts except in the vicinity of vascular bundles in

stems, leaves and roots (Dolferus and Jacobs, 1990).

Specific ADH activity in young seedlings is highest 3

days following germination and decreases to undetectable

levels by day 10 (Dolferus and Jacobs, 1985).

ADH activity may be induced in the mature seedling

and callus tissue using either anoxia or the synthetic

auxin 2,4-dichlorophenoxyacetic acid (2,4-D). Anaerobic

induction of ADH activity in 10- to 14-day-old seedlings

results in a 10- to 15-fold increase in Adh mRNA (Dolfer-

us and Jacobs, 1985), versus the 20 to 50 fold increase

observed in corn. Poly (A+) mRNA isolated from induced

callus will direct the synthesis of ADH in an in vitro

translation system (Dolferus and Jacobs, 1985). Induc-

tion of ADH activity with 2,4-D has not been observed,

however, for cells in suspension (Dolferus and Jacobs,

1985).










Arabidopsis possesses only one Adh gene (Chang and Meye-

rowitz, 1986), as opposed to two for corn and three for

barley (Trick et al., 1988). This gene encodes a protein

identical in length and 81% similar in amino acid se-

quence to maize Adhl. Arabidopsis ADH forms an active

heterodimer with maize ADH when extracts from the two

species are mixed (Fisher and Schwartz, 1973). Further-

more, antibodies raised against maize ADH cross-react

with Arabidopsis ADH (Dolferus and Jacobs, 1991). The

Arabidopsis Adh gene is 73% similar in nucleotide coding

sequence to maize Adhl. Arabidopsis Adh possesses 6

introns, compared to the 9 found in maize Adh, and these

are found in the same positions as their counterparts in

maize. Introns IV, V, and VII of maize Adh are not

present in Arabidopsis Adh (Dolferus and Jacobs, 1990).

Comparison of 5' untranslated regions of the maize

and Arabidopsis Adh genes reveals a disappointing lack of

significant sequence homology. There are, however, small

elements of sequence homology which have been investigat-

ed for potential regulatory function in this and other

gene systems (Ferl, 1990) and whose investigation will be

continued in this study.


Regulation of mRNA Transcription

Several plant gene systems are now available for the

study of mRNA synthesis, as regulated by 5' flanking

sequences, and their subsequent developmental and/or

organ specific expression. This discussion will be










limited to those gene systems which share common putative

regulatory elements or characteristics with maize or

Arabidopsis Adh.



5' flanking DNA of maize Adh



Initial characterization of maize Adhl regulation fell

into two major categories: 1) Examination of protein/DNA

interactions within the 5' flanking sequences by in-vivo

DMS footprinting (Church and Gilbert, 1984; Ferl and

Nick, 1987) and DNase I or restriction endonuclease

hypersensitivity (Paul et al., 1987; Ferl, 1985) and 2)

Expression of reporter genes regulated by various domains

(deletions) of the Adhl 5' flanking region (Howard et

al., 1987; Lee et al., 1987; Walker et al., 1987). Virtu-

ally all current molecular studies in corn Adhl are

derived from this experimental foundation.

In vivo DMS footprinting of the maize Adhl promoter

(Ferl and Nick, 1987) reveals roughly five domains of

putative protein/DNA interaction. Cells in suspension

were either aerated normally or anaerobically induced

prior to DMS guanosine methylation and subsequent genomic

sequencing (Church and Gilbert, 1984). Interactions fell

into two general categories, those which are present

regardless of induction state (-110, -120, -130) and

those which appear in only anaerobically induced cell

suspensions (-95, -180, -130). As will be discussed










later, many of these regions have since been shown to

bind protein in vitro and play a role in regulated ex-

pression. In vivo footprinting of maize Adh2 (Paul and

Ferl,1990) reveals putative sites of protein/DNA interac-

tion as well. Sites of interaction appear in three

general regions. A footprint is observed, however, at

position -85 which is strikingly similar in pattern and

recognition sequence to that found at position -180 in

maize Adhl (Figure 1-1). The two footprints are labeled

"C' (Adhl) and "D" (Adh2) and are induced by anaerobic

stress. The possible involvement of these sites in

regulation has not as yet been tested functionally.

In situ digestion of DNA within intact nuclei using

either restriction endonucleases or DNase-I has been used

to further characterize protein/DNA complexes of Adhl

(Ferl, 1985, Paul et al., 1987). Using restriction

endonuclease Pst 1, a significant alteration in enzyme

access is observed at -147 following anaerobic induction.

Other regions examined were 5', at -417 and -1104, where

corresponding in vivo DMS footprinting data are not

available.

DNase-I hypersensitivity is an alternate method of

characterizing the accessibility of chromatin associated

with Adh transcription. Maize cells in suspension,

either aerobic or anaerobically induced, were used for

isolation of nuclei and in situ DNase-I digestion (Paul




































44


0


0


*1-



0






















*.4 a
44
a
-,
o
















I


S.
rzI
r-1
Q)
I *
2.^
&*x ;t
*^*or


OWtW




S0 0


0-




S0.00

N )
o4 -,- C



00 0
OC





,C0
UM i 0


0 (U

00
E aU P




C 0 0U


40 04 4

04




00 G
G 0 Q)





































uE-


Ut,
U

Ut,
u t

Ut

S Ut
uC
ut,.

u ,



CU
uu




u L
E-





u
UtU
-4
u00




u u






u u
It,
SU U
UtU
UO










OU
U
u ,





Ut, u
C U
u u
u u




4



CU
u



u u




'CE-
u) U
Ut


Ut,


u u
g4


u U
u u


Ut,
Ut,
4E-
u u
u u
CU
t,
< E-

u u



u u
< E-


Ut,
,UU


* Ut,
u u


u u
Su
uo


0- u uD

u u3
tU U
u u
u u 0
'-4
SU 0
t0
u u
u t
uL
LU

UtU
u
u
ut








u u

utU
ut,
.4



o u


Su


EC-
tU
u u




U U









tu





-C
E--
Ut,




4I
o f-

U.,

4.


U 0

u u
u u
uU
tU
UU

'4
Ut,


uU





L'C@
4u



UCO
'CE-



'g 4o










et al., 1987). Regions of hypersensitivity were then

classified constitutive (-160 to -700) or inducible (-35

to -150). These results agree well with those obtained

from the two previous methods with regard to anaerobic

induction, and may be used to delimit anaerobic respon-

siveness to within 200 bp of the transcription start.


Functional analyses; transient


Specific 5' flanking sequences of maize Adh have

been shown to direct the anaerobically induced expression

of a reporter gene in transformed maize protoplasts

(Howard et al., 1987). Functional analysis of 5' dele-

tions of the full length (1.1 kb) promoter shows that

full activity and anaerobic inducibility are retained

from -1094 to -140 (Walker et al., 1987). Inducible

expression is still observed following deletion to -124,

but this resulted in a significant (>50%) decrease in

activity. Linker scanning analysis through the -133 to

-72 region (Walker et al, 1987) identified two distinct

regions necessary for anaerobic inducibility at approxi-

mately -130 and -105. When either of these Anaerobic

Response Elements (AREs) are disrupted by DNA replace-

ment, there is a complete loss in anaerobic induction.

Lee et al. (1987) conducted a functional analysis of

the Adh 5' flanking region of maize as well, and ob-

tained results supporting Walker et al. (1987). Large

segments of the Adh promoter were removed by restriction










digest so that "deletions" corresponded to restriction

enzyme recognition sequence locations. The full length

promoter (1.1 kb) extended to the BamHl site and was

assigned 100% activity. Deletion to the Xbal site (-417)

resulted in a reduction to 65% of full activity, and

further deletion to Pstl (-147) reduces activity to 18%.

The question of anaerobic inducibility was not addressed

here.

The results obtained by Lee et al. (1987) are in

agreement with those from Walker et al. (1987) in that

147 bp of 5' flanking Adh sequence is sufficient for the

expression of a reporter gene. What remains in question

is the degree to which activity is reduced by the various

deletions. This matter is clarified by the work of

Ingersoll (1990) as described below.



Functional analyses; stable transformation



An alternative method for functional analysis of 5'

flanking sequences is the stable transformation of plant

tissue using Agrobacterium. Briefly, the sequence ele-

ment of interest is cloned into a plasmid carrying the

desired reporter gene, a reference gene if desired, and a

selectable marker. This plasmid is then introduced into

the Agrobacterium host which is then used to infect plant

tissue. If the appropriate sequence elements and viru-

lence genes are present, the desired DNA sequences will










be transferred to the plant and be incorporated into the

plant chromosome. The chromosomal site of incorporation

varies among transformation events, so several trans-

formed plants are needed for each construct to provide

statistical accuracy. This method has been successfully

applied to dicot species, but has had limited success in

monocots.

Regulation of tissue specific expression of a gene

is best analyzed in the transgenic plant. This method

has been employed to dissect the well characterized

cauliflower mosaic virus (CaMV) 35S promoter (Benfey and

Chua, 1990). This promoter is known to confer constitu-

tive expression on a reporter gene in both transient and

stable expression systems (Odell et al., 1985; Jefferson

et al., 1987). It was found that specific subdomains

acted in a modular fashion to confer expression of the

reporter gene on various organs (Benfey and Chua, 1990).

Furthermore, the manner in which the subdomains func-

tioned was dependent on the species in which they were

expressed. Their results emphasize the need for caution

when interpreting the tissue specificity of promoter

mediated expression within a heterologous genetic back-

ground.

Ingersoll (1990) has functionally analyzed the maize

5' flanking sequence in plant tumors derived from sun-

flower. This method has the advantage of allowing for the

analysis of larger amounts of transformed tissue than the










protoplast transformation affords, thus allowing for the

direct quantitation of test and reference RNA. This work

confirmed the location of the ARE(s) in maize Adhl to the

region between -140 and -100. A loss of 70% of full

activity resulted by deleting to -140, and this deletion

preserved anaerobic inducibility. These results are in

agreement with those obtained by Walker et al.(1987) and

Lee et al. (1987).

A convincing functional analysis of maize Adh in the

mature transgenic plant has not been accomplished. The

maize plant is refractory to Agrobacterium transforma-

tion, and maize Adh introns are not properly processed

in dicot species (Ingersoll, 1990). Ellis et al. (1987)

attempted to analyze maize Adhl 5' flanking DNA in trans-

genic tobacco. Although Adh 5' flanking sequences were

able to confer anaerobic inducibility, this required the

addition of non-Adh regulatory sequences (CAMV 35S) to

assist in overall transcription levels.

Many of the 5' flanking sequence elements of Adhl

which have been identified by one of the methods dis-

cussed above have been subjected to in-vitro analysis of

protein binding by the gel retardation assay (Fried and

Crothers, 1981). Gel retardation, or the mobility shift

assay, is performed by incubating a radioactive DNA

fragment with the protein extract of interest, which is

followed by fractionation by non-denaturing electrophore-

sis. This method of characterization allows for the










identification of the type(s) and amounts of protein

involved in a regulatory "domain." These proteins may

then be purified and used to obtain a clone of the gene

which encodes them. Such information is essential for

the elucidation of pathways governing development and

stress response.

In general, 5' flanking regions which produce an in

vivo DMS footprint appear to operate as functional units

when the results of deletion studies are used for com-

parison. Putative regulatory sequence elements in maize

Adhl are characterized by 1) the C-rich domain found at

-130 and 2) the GTGG motif at -98, -110, and -180 (Paul

and Ferl, 1990). This is also true for maize Adh2, with

a C-rich footprint at -210, and GTGG elements footprint-

ing at -160 and -80 (Paul and Ferl, 1990).

An oligonucleotide resembling the maize Adhl -130

region was constructed for analysis of in-vitro protein

binding (Ferl, 1990). A component of crude whole cell

extracts from maize suspension cultures was found to bind

specifically to this sequence. Further analysis of the

-130 binding protein (ARF-B2) shows that it is a multi-

component complex which binds the -130 element regardless

of the state of induction, supporting previous results

from in vivo DMS footprinting (Ferl and Nick, 1987).

Attempts to clone the gene encoding ARF-B2 and other

maize Adhl regulatory genes are in progress.










5' flanking DNA of Arabidopsis Adh


As with maize Adh, the in-vivo DMS footprint of

Arabidopsis Adh has revealed several sites of protein/DNA

interaction (Ferl and Laughner, 1989). Footprints are

observed at positions -145, -175, -190, -210, and -310

(Figure 1-2). The DNA sequence elements at these loca-

tions are similar to those found in maize in that they

are also either 1) C-rich (-145), or 2) have GTGG nearby

(-190, -210, and -310). The element which footprints at

position -210 is unique in that it possesses a perfect

GTGG dyad, and will be discussed in greater detail in a

following section. The element which footprints at

position -175 has no obvious similarity to footprinting

elements in maize Adh.

The dyad sequence element which footprints at posi-

tion -210 had been identified in a number of plant gene

5' flanking sequences, and has been termed the G-box

(Giuliano et al., 1988). The G-box element is present in

the 5' flanking sequence of ribulose-1,5-bisphosphate

carboxylase small subunit (RbcS) genes of several dicot

species, including Arabidopsis (figure 1-3). It is also

present in the promoters of several chalcone synthase

(CHS) genes, the patatin gene from Solanum, and the rolbc

gene of Nicotiana (Schulze-Lefert et al., 1989).

























'C-
0













r
04














*4

c
1-1








00


4 c
.a.


*p


0
,.C
t1-
't










(r ln
3r*


I I *
,MM >ia c

S4 o .






OU 0,-


, >1

0





SJ0 '
I .

-4 ~.-' to


ca .4 -
S0 0

4O -1 Co
4 J 0 A3



0 UO
( 3) 0 C.4.



EwCI .41
VL 03
o o ) 04



40 4c -kZ0
0) 4) 0v It
*P-P ^ IC (
(U U 1 lcc























4-
LU

4-

0 -
o 4 i








40 *u
IC 4

N a
ftJ




4
-C
::
U W
O Li


0. ul












em 1
UJ

4
4 -


4 -




































u a
-I




O Ui




U *
-U









0 UU
us
UL

** U o
@4 -
4-


C 4

UO
o C4








UU

O UL
U



UO








" Li




Lta









4-

WI U
4


( Li
v0
l U
U
U
-

4
<

4
4
Li
Li
<
I <
4
(1
0


<


<

f s
L

I0
u





0

o o
U
4












0 0

<
4




Li





4
4




Li
U
4




00







0U
4
4









Li

u
Li
4
4
Li
u
4
*Lu
o4

04
4
U


Li


[!



oLu
4-

4
L9
4
(U
4
-4
4
4
4
4
4
U-

4
4
Li
4k

Li
4

u
4>
4*
4

4


0

M *




0
0
<

O u
oe U
a t
4
Li


U
Li
U
4




Q
















o L




0I
.0
* 4
























4


0
Li
0 -
eg4





































If
Li
4









o 1-
*
4
U







a e-

* 4
U







OU


SU
o -
e







0 L
O h 4
4




















No

e.
O C-

a* 4
Li

4



U





* 4

Li
4

4
4


- U


U
LI

a U

U
4




a -U
*t 0
4r

















Tnm. RbcS-3A
Tnm. RbcS-1
Tob. Ntss-23
plumhb.RbcS-8B
Sonyhb. SRS-1
Soyb. SRS-4
Petunia
Petunia
Pea
Pea
pea
Pea
Pea
Ar;hbidopsis


SSU-301
SSU-611
RbcS-3.6
hbcS-8.0
RbcS-E9
RbcS-3A
RbcS-3C
RbcS-1A


TCATTCTGACACGTGGCACCCTTT
AT---- C ------ ---T-CA
*AGG---T----(G ---TCCA
GGTGCT-T----- ---TCCA
G****T-TC------T----TCCA
G*CC--CTC----------TTCCA
AAG-G--TC---------- TCCA
AGC-A---------- T--- TCCA
GGCA---T --------- TTA-CC
GGCA---T-T ---------TTA-CC
GGCA---T----------TTA--A
GGTAATATC---A------TG-CC
GTCA--ATC--------- T--CA
ATTA---TC---------TTA-C-


Figure 1-3. Comparison of RbcS G-box elements.

The sequence elements shown above are from 5' Flank-
ing sequence of selected RbcS genes. The areas of G-box
homology are shown by the circled G to the right, the
abbreviations shown are as follows. Tom.=tomato,
plumb.=tobacco, and soyb.=soybean. Numbers in parentheses
to the right are reference numbers in the original manu-
script. From Giuliano et al., (1989).


-278
-534
-275
-287
-240
-240
-269
-185
-217
-227
-232
-217
-172
-243


(24)
(25)
(26)
(26)
(1)
(1)
(20)
(20)
(6)
(2)
(2)
(21)


"~I'-------










RbcS 5' flanking sequence


Members of the RbcS gene family, like those of the

Adh gene family, are environmentally inducible and ex-

pressed in a cell-type specific manner (Tobin and Silver-

thorne, 1985). RbcS mRNA is induced in the presence of

light primarily in leaf and stem tissue and is not de-

tectable in roots (Sugita et al., 1987). RBCS activity

is, in part, regulated at the transcriptional level

(Tobin and Silverthorne, 1985; Kuhlemeir et al., 1987).

The RbcS gene family has been comparatively well

characterized at the molecular level. Comparison of 5'

flanking sequences from several RbcS genes has identified

three general elements, of sequence homology (Kuhlemeier

et al., 1987; Giuliano et al., 1988); 1) the G-box, C/A-

CACGTGGC (Giuliano et al., 1988), 2) the I box, which has

the consensus sequence 5'-GATAAG-3'(Giuliano et al,

1988), and 3) the GT box, which has a very broad consen-

sus sequence (Green et al, 1988; Kuhlemeir et al., 1988).

In order to demonstrate protein binding to these

putative regulatory elements, 5' flanking sequence frag-

ments from several RbcS genes were examined using the gel

retardation assay (Giuliano et al., 1988). Nuclear

extracts from tomato and Arabidopsis were found to con-

tain a protein component which binds specifically to

upstream RbcS sequences from tomato, pea, and Arabidop-

sis.










This result was further verified by performing DNase

I footprinting on the protein/DNA complex (Giuliano .op

et al, 1988). In this method, the DNA sequences of

interest are either bound to protein or naked, subjected

to partial digestion by DNasel, and fractionated by

denaturing electrophoresis. Regions which are "protect-

ed" from digestion show a "footprint" as compared to

naked DNA. These footprints correspond roughly to the

site of protein binding.

Footprinting of RbcS 5' flanking sequences, in

conjunction with gel retardation results, verified the

region of specific protein binding to be the G-box

(Giuliano et al, 1988). In contrast, DNasel hypersensi-

tivity was detected in the region of the I box. Foot-

printing of the GT box (Giuliano et al, 1988) failed to

reveal interactions with those sequence elements from

either tomato or Arabidopsis crude extract.

The GT box consensus and I box consensus for the Rbcs

gene family does not obviously align with regions of in-

vivo DMS footprinting in either maize or Arabidopsis Adh.

However, the GT box of a few Rbcs genes (Kuhlemeier et

al., 1987) resembles the GTGG observed in maize and

Arabidopsis Adh (-310, -190) and is homologous to the

SV40 core type II element (Zenke et al., 1986).

Functional analysis of the 5' flanking sequences of

Rbcs genes in transgenic plants has provided the bulk of










information available regarding sequence elements

necessary for gene activity and photoactivation. Initial

characterization delimited a 352 bp deletion mutant of

the pea rbcS-E9 gene which retained photoinducibility and

tissue specificity (Nagy et al., 1985). Deletion of

RbcS-E9 or RbcS-3A (Fluhr et al., 1986) confirmed the

results of Nagy et al. (1985) and further defined se-

quences necessary for organ specificity and photoinduci-

bility to within 240 bp 5' of transcription start.

More recently, Kuhlemeier et al. (1989) demonstrated

that a region from -50 to +15 of the RbcS-3A promoter

was able to confer light responsiveness, but not organ

specificity, on a reporter gene in transgenic plants.

Furthermore, it was demonstrated that this promoter

element interacts with an upstream element between -189

and -156 to allow high level expression. This region

contains no previously described element of homology. A

region from -150 to -230 does, however, produce an in-

vitro footprint (Green et al., 1987), and includes the G-

box (-217) and binds the G-box factor (Giuliano et al.,

1988). It is suggested that this domain (-189 to -156)

may contain a novel regulatory element. It should be

noted here that their constructs contained additional

enhancer elements (35S, soybean heat shock element)

foreign to the gene of interest in order to boost expres-

sion levels.










Further experiments designed to precisely define

thelocation of the light responsive elements) and the

function of the G-box were conducted on the Arabidopsis

rbcS-lA gene in transgenic tobacco (Donald and Cashmore,

1990; Donald et al., 1990). This work defined a 196 bp

region from -320 to -125 which was sufficient to confer

light regulated, tissue specific expression on an Adh

reporter gene. This region contains GT, G, and I boxes.

Mutations or deletions affecting GT box(es) had no sig-

nificant effect on expression. Disruption of the I box

resulted in a significant decrease in reporter gene

activity, and mutations or deletions which disrupted the

G-box abolished all reporter gene activity.

The presence of the G-box element in other light

regulated genes has led to an association of this element

with light responsiveness. In addition to functional

analyses of promoter function in the RbcS gene family,

the chalcone synthase (CHS) and chlorophyll a/b binding

protein (CAB) genes have also contributed to the under-

standing of the G-box and other putative 5' regulatory

elements.

CHS and CAB 5'flankinq sequence


The in vivo DMS footprint has not been successfully

conducted on an RbcS gene. Footprint data are available,

however, for the Chs gene of parsley (Schulze-Lefert et

al., 1989). CHS activity is inducible in parsley suspen-










sion cultures by UV light, and activation occurs at the

transcriptional level (Chappel and Hohlbrock, 1984).

Three sequence domains of the Chs promoter produce

an in-vivo DMS footprint upon UV irradiation of parsley

cells in suspension. These domains (labeled regions I,

II, and III) occur at positions -140, -165, and -235.

Region III resembles the GT box of RbcS genes, and, when

assayed for function in a transient expression, appears

to function as a transcriptional enhancer (Schulze-

Lefert et al., 1989). Region II is homologous to the G-

box of Rbcs and Arabidopsis genes, and footprints in a

manner nearly identical to that of the Arabidopsis G-box.

When region II is deleted or mutated for functional

analysis (Schulze-Lefert, 1989), there is a nearly com-

plete (>90%) loss of reporter gene activity, as well as a

loss in photoinducibility. It should be noted here,

however, that this deletion also included region I, and

region I deletion or mutation results in an equally

severe reduction in reporter gene activity.

In order to more carefully determine the structure

of putative cis-acting elements, site directed mutagene-

sis was employed to mutate and change spacing between

these elements (Block et al., 1990). These constructs

were then assayed for transient expression in parsley

protoplasts. Their results defined a 7 base core se-

quence critical for G- box activity. When the distance










between box I (as previously defined by Schulze-Lefert et

al.,( 1988)) and the G-box was increased by 4 bp, report-

er gene activity was essentially abolished.

However,deleting the space between the G-box and box III

had no significant effect. It is proposed that the G-box

functions in close association with at least one other

regulatory interaction.

This result is in agreement with those obtained for

5' flanking sequences of the RbcS gene family. Regulato-

ry interactions of 5' flanking sequences do not appear to

function with complete independence, in spite of their

modularity.



Summary



Plants may respond to environmental stress by modi-

fication of gene expression. The perception of changes

in atmospheric oxygen, light, temperature, and water

availability results in a modification of the type of

genes transcribed and translated. The modes of signal

transduction for these plant responses remain unknown.

In the case of plant Adh genes, mRNA transcription is

accelerated following anoxia, and rates of mRNA turnover

are low. Gene activity is therefore increased by induc-

tion of mRNA transcription.











Induction of gene activity is correlated with pro-

tein binding to 5' flanking sequences of that gene. The

in vivo footprint of the maize Adhl promoter changes

following anoxia. Sequences of the promoter which foot-

print in vivo, and bind protein in vitro, are contained

within regulatory promoter elements as determined by

deletion or mutation analysis.

One of these regulatory elements, the G-box, has

been found in genes of diverse function. Genes which

contain this element are expressed in a wide variety of

tissue types and are regulated by mechanisms which are

not obviously similar. Regulatory elements appear to

function with considerable modularity, and this G-box

element has apparently been recruited by many genes for

an as yet undetermined purpose.

Sites of protein-DNA interaction revealed by in vivo

footprinting of Arabidopsis Adh will be characterized for

in vitro protein binding. The Adh 5' flanking sequence

will then be subjected to functional analysis in trans-

formed Arabidopsis. Critical regulatory elements will be

identified and a model of Arabidopsis Adh transcriptional

activation will be proposed.
















CHAPTER 2
CHARACTERIZATION OF PROTEIN INTERACTIONS
WITH Arabidopsis Adh 5' FLANKING SEQUENCE

Introduction

Transcriptional regulation of gene expression in-

volves the interaction of DNA binding proteins with the

gene of interest (Dynan et al.,1985), and localization of

these sites of interaction has been facilitated by the

technique of in vivo dimeththylsulfate (DMS) footprint-

ing (Church and Gilbert,1984; Nick and Gilbert, 1985).

The in vivo DMS footprints of the maize Adhl and Arabi-

dopsis Adh genes have revealed several sites of putative

regulatory interaction which are similar in sequence

between these two genes, namely, the C rich box and a

moderately conserved 5'-GTGG-3' motif (Ferl and Nick,

1987; Ferl and Laughner, 1989). However, only in the

Arabidopsis Adh gene (at position -210) is this GTGG

motif present on both strands as a perfect dyad (Ferl and

Laughner, 1989). This symmetrical element (5'-CCACGTGG-

3') has also been identified in the 5' flanking sequence

of ribulose-1,5-bisphosphate carboxylase small subunit

(RbcS) genes of several species and is termed the G-box

(Giuliano et al., 1988).










Expression of the alcohol dehydrogenase (Adh) and

ribulose-1,5-bisphosphate (RbcS) genes of higher

plants is both cell-type specific and environmentally

inducible, yet the tissues in which they are expressed,

their modes of induction, and their protein functions are

quite distinct. RbcS mRNA is induced in the presence of

light primarily in leaf and stem tissue, and is not

detectable in roots (Sugita and Gruissem, 1987).

The G-box of tomato RbcS-3A has been shown to bind

specifically to protein from crude nuclear extracts from

tomato and Arabidopsis by the gel retardation assay and

G-box elements from other RbcS genes, including Arabidop-

sis RbcS-1A, compete specifically for G-box protein

binding (Giuliano et al., 1988).

In this chapter it will be demonstrated from the in

vitro binding competition assay (Fried and Crothers,

1981) and from both in vivo and in vitro DMS footprinting

(Church and Gilbert, 1984; Treisman, 1986; Ferl and Nick,

1987) that a protein component of Arabidopsis whole cell

extract from cultured cells and mature leaves binds the

Arabidopsis Adh G-box in a manner comparable to that

observed in cultured cells in vivo (Ferl and Laughner,

1989). Comparison of in vivo G-box binding in cultured

cells and leaves, however, reveals significant differ-

ences between these tissue types, and that in vivo Adh G-

box binding in leaves is very weak or does not occur.










Materials and Methods



Gel Retardation Assay


Whole cell extracts were prepared by variations of

Manley et al. (1980) and Wu (1984). Cell cultures or

leaves were frozen in liquid nitrogen before homogeniza-

tion at 4C in 15 mM Hepes, pH 7.6, 40 mM KC1, 5.0 mM

MgCl, 1.0 mM DTT, and 0.1 mM phenylmethanesulfonyl

fluoride (PMSF). One-tenth volume of 4.0 M ammonium sul

fate was added and the slurry centrifuged for 30 min at

19,000 g, 40C. To the supernatant, 0.3 g/ml ammonium

sulfate was added and mixed on ice for 60 min. Protein

was then precipitated at 15,000g for 20 min., 40C. This

pellet was resuspended (1.0 ml/10 g tissue) in 20 mM

Hepes, pH7.6, 40 mM KC1, 1.0 mM DTT, 0.5 mM PMSF, 0.1 mM

EDTA, and 10% glycerol), then dialyzed 4 hr in two

changes of 100 volumes of 20 mM Hepes, pH7.6, 40 mM KC1,

0.1 mM PMSF, 0.1 mM EDTA, 10% glycerol, and 5 mM -mer-

captoethanol. After dialysis the extract was frozen with

liquid nitrogen and stored at -80C until use. Binding

reactions were performed at room temperature in 13 ul

containing 10 ul (60 ug) of crude extract. All binding

reactions contained either 1 ul (50 ng) competitor DNA,

or extract buffer as the no competitor control, one ul of

end-labeled Adh G-box oligonucleotide (1.0 ng), and 1.0

ul of 1.0 M KCl (unless otherwise specified) for a final










KC1 concentration of 110 mM. A 5 minute pre-binding of

competitor DNA and protein was followed by addition of

labeled oligonucleotide probe, a 5 minute additional

incubation, and electrophoresis on a 5% non-denaturing 89

mM Tris-Cl, 89 mM Boric acid, and 2.6 mM EDTA polyacryl-

amide gel at 30 mA.



In Vitro DMS FootDrinting



In vitro DMS footprinting was carried out as de-

scribed (Treisman, 1986). Following electrophoretic

fractionation of binding reactions or probe alone, gel

slices were electroeluted, cleaved at methylated G resi-

dues with piperidine, and fractionated by electrophoresis

on a 10% polyacrylamide-urea gel. Approximately equal

counts were loaded per lane. The extent of methylation

protection varies among G residues within a given foot-

print and is highly reproducible.



UV Crosslinking



Modification of the cross-linking procedure of

Chodosh et al. (1986) was used for DNA probe synthesis

and binding reactions. Probe was synthesized by first

annealing a specific primer to the 3' end of the top

strand of the -210 oligonucleotide. The second strand










was synthesized using Klenow fragment in the presence of

25 ng of DNA template, 25 uM dGTP, 50 uM 5-bromo-2'deo-

xyuridine triphosphate (Sigma), and 50 uCi each of

-32P-dATP and -32P-dCTP. Binding reactions are essen-

tially identical to the gel retardation binding reactions

with the following exception that reactions contain

approximately Ing labeled oligonucleotide probe. Binding

is followed by 15 min exposure to UV light, nuclease

digestion and SDS gel electrophoresis (12.5 %). The gel

was fixed, dried, and autoradiographed for 48 hr.



RESULTS



The G-box binds a protein component of Arabidopsis crude

extract



Oligonucleotides corresponding to previously deter-

mined sites of in vivo protein-DNA interaction at posi-

tions -310, -210, and -140 of the Arabidopsis 5' flanking

region (Ferl and Laughner,1989) were synthesized. The

-210 oligonucleotide was utilized as the probe for all in

vitro gel retardation and DMS footprinting experiments.

Figure 2-1 shows the Arabidopsis Adh oligonucleotides

along with those corresponding to the G-box of Arabidop-

sis RbcS-1A (Krebbers et al., 1988), and the G-box like

sequence of the adenovirus major late promoter (MLP)
























Competitor DNA


-149 GCCCCTAGTATTCTGC


-134


-318 ACACCACGGCGTGACCAT -301

-222 GAATGCCACGTGGACTGCA -204


-264 ATCTTCCACGTGGCATTA


-63


GGCCACGTGACC


-247

-52


Figure 2-1. Oligonucleotides used for gel retardation
assay.
Sequences of the Arabidopsis Adh 5' flanking region
about which are observed in vivo DMS footprinting (Ferl
and Laughner, 1989) were used to design oligonucleotides
for the gel retardation assay. The -210 oligo (Adh G-
box) is used for all experiments as the probe. The RbcS
G-box oligonucleotide (RG) was synthesized from the -243
region of Arabidopsis RbcS-1A (Krebbers et al., 1989).
The MLP oligonucleotide corresponds to the -60 region of
the MLP of Adenovirus (Chodosh et al., 1989). Only the
top strand of the sequence is shown. All probe and
competitor DNA is in >90% double stranded form.


-140

-310

-210

RG

MLP


Sequence










enhancer (Chodosh et al., 1989). The MLP synthesized is

identical in length and sequence to that presented by

Chodosh et al. (1989) to be sufficient for specific

binding to a component of yeast crude extract, and was

chosen as a representative of G-box like mammalian

enhancer element for this experiment.

Subjecting crude whole cell extract (Manley et al,

1980) from Arabidopsis cell cultures to the gel retarda-

tion assay (Fried and Crothers, 1981) revealed two dis-

tinct types of G-box binding activity, Forms I and II

(Figure 2-2). A third less distinct form is observed

which migrates more rapidly than Forms I or II. Lane 1

shows bound and free forms resulting in the absence of

competitor DNA. Lanes 2 through 7 represent binding

reactions carried out in the presence of increasing

amounts of the heterologous competitor DNA, poly(dI-dC).

The addition of as little as 10 ng of poly(dI-dC) signif-

icantly increases the intensity of Form II, with a corre-

sponding reduction in intensity of Form I (lane 2).

Continued increase in poly(dI-dC) to 500 ng (lanes 3,4,5,

and 6) resulted in the reduction and eventual elimination

(lane 6) of Form I. Form II, however, increased in

intensity up to 500 ng. Further increase of poly(dI-dC)

to 1.0 ug prevented formation of Form I and significantly

reduced the intensity of Form II.













U

0 o 0 0o 0 O
o 0 0 0 0 0 0
a w-4 (M 1-4


- %w


.
a -'


II


w


1 2 3 4 5 6 7


Figure 2-2. Gel retardation assay of Arabidopsis crude
extract binding to the Adh B-box (-210).

Crude whole cell extract from Arabidopsis cells in
suspension is bound to end-labeled, double stranded -210
probe for gel retardation assay (Fried and Crothers,
1981). Binding reactions were carried out in the absence
of competitor DNA (lane 1), or in the presence of poly-
(dIdC) at 100 ng (lane 2), 200 ng (lane 3), 300 ng (lane
4), 400 ng (lane 5), 500 ng (lane 6), and 1.0 ug (lane
7).










In order to examine the influence of ionic strength

on binding affinity of Forms I and II, a retardation

assay was conducted with cell suspension crude extract in

the presence of increasing amounts of KCl (Figure 2-3,

A). Stringent binding conditions were maintained by the

addition of 1.0 ug poly(dI-dC) to each reaction. Lane 1

shows a binding reaction identical to that in lane 7 of

Figure 2-2, with KC1 at 40 mM. Increasing KC1 concentra-

tion to 100 mM results in a marked increase in binding

affinity of protein to the -210 probe. Continued in-

crease of KC1 to 200 mM greatly reduced the amount of

Form II, and binding in the presence of 300 mM essential-

ly precludes binding to the -210 probe.

To demonstrate that retarded electrophoretic species

are due to protein binding, a binding reaction was car-

ried out in the presence of proteinase k (Figure 2-3, B).

Lane one shows a typical Form I retarded complex using

1.0 ug poly(dIdC) competitor DNA. Form I is not present

when binding is carried out in the presence of proteinase

k (lane 2).

G-box protein binding specificity was determined by

competition analysis (Figure 2-4). Form I binding to

the Adh G-box appears to be nonspecific, as it is ob-

served in the absence of competitor DNA (lane 1), in the











A









0I -I


Figure 2-3. Influence of KC1 concentration on G-box
binding affinity.

A) Binding reactions of the Adh G-box (-210) probe
with Arabidopsis crude whole cell extract were conducted
in the presence of 1.0 ug poly-(dIdC), and KC1 at 40 mM
(lane 1), 100 mM (lane 2), 200 mM (lane 3), 300 mM (lane
4).
B) Binding reactions were carried out as in lane 1
above (lane 1), or in the presence of proetinase k (lane
2).


II -1


, Mwk
A'


1 2









presence of 50 ng of both heterologous and homologous

competitor oligonucleotides (lanes3 to 7), but not in the

presence of 1.0 ug of poly(dI-dC)(lane 2). Form II per-

sists in the presence of up to 1.0 ug poly(dI-dC)(lane 2)

as well as 50 ng of heterologous competitor DNA (lanes 3,

4, and 7), yet is obliterated by 50 ng of homologous

competitor Adh and RbcS G-boxes (lanes 5 and 6). There

fore, Form II appears specific for protein binding to

both the Adh and RbcS G-boxes. The G-box like MLP se-

quence is not effective in competition for G-box binding.

The results of competition analysis obtained with

leaf extract are similar to those from cell culture

extract (Figure 2-4, lanes 8-14) with two exceptions.

Form IV persists in the presence of all competitor DNA

(lanes 9 to 14), suggesting that it is nonspecific Adh G-

box binding. The specific binding activity from leaf

extract comparable with the Form II observed using cell

culture extract is represented by Form III, a protein-DNA

complex that was slightly reduced in migration rate

relative to Form II.

Competition experiments were also conducted in the

presence of 1.0 ug of poly(dI-dC) to demonstrate selec-

tive binding in excess heterologous carrier DNA. The

results of this experiment are shown in Figure 2-5.

Binding to the G-box -210 probe in cell culture and leaf

extract was conducted in the absence (lanes 1 and 6) or
















CELL


CULTURE


LEAF


I


II,-


ii


ill


- ill


milB I-"v


Figure 2-4. Adh G-box (-210) Binding Factor Is
Both Cell Culture and Leaves of Arabidopsis.


Found in


Crude whole cell extracts from cells in culture
(lanes 1 to 7) or mature leaf tissue (lanes 8 to 14) were
assayed for Adh G-box binding competition. Binding
reactions were carried out either in the absence of
competitor DNA (lanes 1 and 8), or in the presence of 1.0
ug of poly(dIdC)(lanes 2 and 9) or 50 ng of the following
oligonucleotides: -310 (lanes 3 and 10), -140 ( lanes 4
and 11), -210 Adh G-box (cold probe) (lanes 5 and 12), RG
(lanes 6 and 13), or MLP (lanes 7 and 14).










presence (lanes 2 to 5, and 7 to 10) of 1.0 ug of

poly(dI-dC) in addition to 50 ng of the following oligo-

nucleotide competitors: -310 (lanes 3 and 8), -210 (lanes

4 and 9), and RG (lanes 5 and 10).

Form I of the cell culture extract binding reaction

is present in lane 1 and absent in lanes 2 and 5. As was

seen in the previous experiment, poly(dI-dC) competitor

at 1.0 ug abolishes Form I [as will 1.0 ug of the -210 G-

box oligonucleotide (data not shown)]. Form II, however,

persists in the presence of either 1.0 ug poly(dI-dC)

alone (lane 2) or 1.0 ug poly(dI-dC) plus 50 ng of -310

(lane 3), yet is abolished in the presence of 1.0 ug

poly(dI-dC) plus 50 ng of -210 (lane 4) or 50 ng of RG

(lane 5). These results are entirely consistent with the

previous experiment (Figure 2-4), and are in support of

Form II being a protein-DNA complex that is specific for

G-box binding.

Results obtained using leaf extract in the presence

of high amounts of poly(dI-dC) (Figure 2-5, lanes 6 to

10) are also consistent with the previous experiment.

Bound Form III is observed in the absence of competitor

DNA (lane 6), in the presence of 1.0 ug of poly(dI-dC)

(lane 7), and in the presence of 1.0 ug poly(dI-dC) plus

50 ng of -310 (lane 8). Form III is abolished, however,

when 1.0 ug of poly(dI-dC) and 50 ng of homologous com-

petitor oligonucleotide are supplied to the binding














COMPETITOR
I A. enI o7 ,
1 a.
"" o-'Ji wQ l
o6O go


H


1-"11


1**UW ,vW-v


1 2 3 4 5 6 7 8 9 10 11 12 13 14


Figure 2-5. Binding specificity of Arabidopsis Crude
extract is preserved in the presence of poly(dIdC).
Crude whole cell extracts from cells in culture
(lanes 1 to 5) or mature leaf tissue (lanes 6 to 10) were
assayed as in Figure 2-4, with modification of competi-
tive binding conditions. Binding reactions were carried
out either in the absence of competitor DNA (lanes 1 and
6), or in the presence of 1.0 ug poly(dIdC) alone (lanes
2 and 7), or in the presence of 1.0 ug of poly(dIdC) plus
50 ng of one of the following oligonucleotide competitor
DNAs: -310 (lanes 3 and 8), -210 (lanes 4 and 9), and RG
(lanes 5 and 10).










reaction. Form IV, again, persists in the presence of

all competitor DNA. Therefore, it is concluded from

these results that bound Form III represents a

protein/DNA interaction of a specificity equivalent to

that observed for bound form II.



Protein-DNA crosslinking



UV crosslinking experiments were conducted for

determination of the molecular weight of protein(s)

binding to the G-box oligonucleotide. Binding reactions

are essentially identical to those performed for gel

retardation experiments, and included 1.0 ug of poly(dI-

dC) as heterologous competitor. No distinct bands are

observed in the no protein control (Figure 2-6, lane 1).

One predominant band of approximately 33 kd is observed

when protein is included in the binding reaction (lane

2).

Binding specificity was tested by the addition of

either -310 or -210 oligonucleotides. Competition with

the -310 oligonucleotide is not observed at 20 ng (lane

2) or 200 ng (lane 3), but is effective at 2.0 ug (lane

4) in eliminating both electrophoretic forms. The corre-

sponding analysis using -210 as competitor DNA is shown

in lanes 5 (20 ng), 6 (200 ng), and 7 (2.0 ug).















01

0 0
0 0


S92.5




69


---- 46


OOA- -o


021.5



00- 21.5


1 2 3 4 5 6 7 8


Figure 2-6. Protein-DNA Crosslinking of Arabidopsis
Crude Extract with the Adh G-box (-210) Probe.

The Adh G-box probe was synthesized incorporating 5-
Brdu for UV mediated protein-DNA crosslinking. Two
electrophoretic forms, a conspicuous band at 55 kd and a
faint band at 35 kd, are observed when binding is carried
out in the absence of competitor DNA (lane 1). Binding
reactions were also carried out in the presence of homol-
ogous competitor Adh G-box (cold -210) at 20 ng (lane 2),
200 ng (lane 3), or 2.0 ug (lane 4). Reactions were also
performed in the presence of heterologous -310 competitor
DNA at 20 ng (lane 5), 200 ng (lane 6), or 2.0 ug (lane










Again, competition is observed only in the presence of

2.0 ug of -210, and is equal for both bands. A faint

band is detectable in lane 5 at 55 kd, however, where one

is not observed in lane 8.


In Vitro DMS Footprintinc of the Adh G-box with Cell
Culture and Leaf Crude Extract

The in vitro competition analysis (Figure 2-4)

demonstrates equal specificity of protein factor G-box

binding from cell cultures and leaves, but it does not

indicate whether the protein-DNA interactions giving rise

to equivalent specificity are identical. An in vitro DMS

footprint (Treisman, 1986) of the Adh G-box complexes

from cell culture and leaf crude extract (Forms II and

III) is shown in figure 2-7.

Quantitative variation in the extent of methylation

protection in vitro may be observed among G residues

within either strand from both cell culture and leaf

extract (lanes 3,6,9,12). Moderate protection is detect-

ed for G residues at positions -210 and -211 (cell cul-

ture) of the top strand and -217 of the bottom strand,

whereas complete protection is evident for positions

-213 and -218 of the top strand and -216 and -214 of the

bottom strand. (The probe used for this assay was modi-

fied by the addition of PstI linkers at both the 5' and
















top strand
,Cal culture. ---Lt- _bottanm strnndl
Lo L loaf ,




-210 9
-211 -I

-213 -f4 l
-217 Wb.
-216


-21 -- -2164
7 10 II 12
I 2 4 I5








Figure 2-7. In vitro DMS footprinting of the Arabidopsis
Adh G-box from Cell Culture and Leaf Whole Cell Crude
Extract.
The -210 Adh G-box (Figure 2-1) double stranded
oligonucleotide containing PstI ends was 5'-end-labeled
on either the top strand only (left upper panel) or lower
strand only (right lower panel). After binding with cell
culture or leaf extract, the binding reaction was treated
with DMS and subjected to preparative electrophoresis on
gels identical to those used for the gel retardation
assay. The bands corresponding to the free, unbound
probe (lanes 2, 5, 8, and 11), bound form II (lanes 3 and
9), and bound Form III (lanes 6 and 12) were recovered by
electroelution. As an additional control, the probes
were treated with DMS in binding buffer without extract
(lanes 1, 4, 7, and 10). The asterisk indicates a gua-
nine residue outside the G-box, within the PstI end.
This G-box oligonucleotide with PstI ends has identical
bandshifting and competition qualities as the -210 oligo-
nucleotide without PstI ends.










3' ends. The G residue of the bottom strand, indicated

by a star in figure 3, is part of that linker sequence).

These results agree well with those obtained by in vivo

DMS footprinting of cell cultures (Figure 2-8).




Discussion


The G-box sequence is highly conserved among RbcS

genes (Giuliano et al., 1988) and some other higher plant

genes (Ferl and Nick, 1987; Ferl and Laughner, 1989;

Schulze-Lefert et al., 1989). in vivo DMS footprinting

of an RbcS gene is not yet available for comparison with

that of Adh, but in vivo studies conducted in parsley for

the chalcone synthase 5' flanking sequence have revealed

a G-box protein interaction virtually identical to that

observed in Arabidopsis Adh in cell culture (Schulze-

Lefert et al., 1989).

Although G-box sequences have been identified in

several RbcS and other plant genes, the only Adh gene

known to contains the conserved G-box element is that

from Arabidopsis. Maize Adhl and Adh2 do not have dyad

G-box elements (Gerlach, 1982; Dennis et al., 1985; Ferl

and Nick; 1987). Comparison of factor binding sites in

Arabidopsis and maize by in vivo DMS footprinting (Ferl

and Laughner, 1989) shows that the conserved 4C box and a

GTGG motif that is essentially half of a G-box are bound




















-220
*


-210
*


o o 00
o 0 00
AGAAATGCCACGTGGACGAATA
TCTTTACGGTGCACCTGCTTAT
00o
00 0


in vitro
in vivo


in
in


vivo
vitro


Figure 2-8. A Summary Comparison of in Vivo and in Vitro
DMS Footprinting for the Arabidopsis Adh G-box.

Protein-DNA interactions are designated by circles.
Open circles denote G residues protected from DMS methy-
lation and solid circles indicate enhanced methylation.










in the 5' flanking sequence of both genes. We have found

that these sequences do not compete for protein binding

to the G-box in vitro (data not shown), and therefore

suggest that these two element types are not bound by a

G-box factor in vivo. The G-box factor is, therefore,

separate and distinct from the factors) involved with

the 4-C box or GTGG motifs.

RbcS sequences containing the G-box confer both

organ specific and photoinducible expression in transgen-

ic plants (Fluhr et al.,1986; Kuhlemeier et al., 1987).

However, recent reports (Kuhlemeier et al., 1989; Ueda et

al., 1989) have shown that regions containing the G-box

are functioning as transcriptional enhancers, and that

the G-box itself is not essential for photoactivation. G-

box like sequences of animal promoters (Chodosh et al.,

1989) are known to function as enhancer elements, but our

binding competition results indicate that the MLP se-

quence either does not bind the same protein as the Adh

G-box, or does so with a greatly reduced affinity.

It is interesting to note that the slight variation

in electrophoretic mobility of the bound G-box between

cell culture and leaf extract of Arabidopsis (figure 2-

4) has also been observed between light grown and dark

adapted tomato leaf extract (Giuliano et al., 1988).

Extracts from dark adapted leaf tissue, in which RbcS

genes are not induced, exhibited G-box binding which was










slightly more rapid in migration rate than that from

light-grown plants, and this corresponds directly to what

we observe, with cell culture complex ex (Form II) mi-

grating more rapidly than that from leaves (Form III).

In contrast, binding specificity and in vitro DMS foot-

printing of cell culture and leaf extract are striking in

similarity, with the only notable difference being the

top strand footprint at nucleotide -211. In combination,

these results suggest that the G-box factor (GBF) may

undergo cell-type specific modification or interaction

with additional regulatory proteins. This is further

supported by the observation of ADH activity in dark

adapted leaves of maize (Bailey-Serres, 1987).

The differences we have observed in DMS modification

of G residues between the in vivo and in vitro foot-

prints, although possibly a result of the artificial

nature of the in vitro binding environment, may be a

result of examination of the G-box in the absence of

adjacent protein-DNA interactions (Ferl and Laughner).

This possibility may be addressed by increasing the

length of the probe used for the in vitro DMS footprint

to include other 5' flanking sequences which bind other

protein factors in vivo. This would more closely resemble

the in vivo binding conditions, and may assist in charac-

terizing other putative regulatory interactions which were

identified by the in vivo DMS footprint.










Mature Arabidopsis leaf tissue has also been sub-

jected to in vivo DMS footprinting (A-L. Paul, personal

communication), and is markedly dissimilar to that ob-

served for cells in suspensinsion. It is therefore

suggested that the in vivo difference between cell cul-

ture and leaf tissue in Adh G-box binding is not due to

G-box factor structure, and may be due to protein-protein

interactions or some cell-type variation in DNA modifica-

tion or chromatin structure. The ability of Adh and

RbcS-1A of Arabidopsis (and perhaps other genes with

related G-boxes such as chalcone synthase) to utilize

the same putative regulatory protein suggests a general

role for the G-box which varies with both cell type and

target gene in the induction of gene transcription in

higher plants.
















CHAPTER III
FUNCTIONAL ANALYSIS OF THE 5' FLANKING REGION OF
Arabidopsis Adh






In contrast to the extensive analysis which maize

Adh 1 5' flanking DNA has been subjected to, relatively

little is known about the function of potential regulato-

ry sequence elements of other Adh genes. Of the Adh

genes which have been cloned to date, only Arabidopsis

Adh has been characterized with regard to promoter func-

tion. Tobacco protoplasts were transformed (Dolferus and

Jacobs, 1990) with constructs containing either the full

960 bp of Arabidopsis Adh 5' flanking DNA or a fragment

which extended to -230, and subsequently assayed for

reporter gene expression. Anaerobic induction was ob-

served for full length promoter fused to both CAT and GUS

reporter genes, whereas the truncated promoter allowed

constitutive expression in the absence of oxygen. At

tempts to transform Arabidopsis protoplasts were unsuc-

cessful.

Dolferus and Jacobs (1990) also report a functional

analysis of the truncated (-230bp) Arabidopsis Adh pro-

moter fragment in transgenic Arabidopsis. CAT










activity was observed in root and shoot tissue, but not

in flowers. Full length promoter constructs were not

included in this study. These results are not in agree-

ment with those obtained using tobacco protoplasts.

Obviously, a genuine analysis of Adh regulation in Arabi-

dopsis must include transient (and stable) expression

characteristics of the Adh 5' flanking sequence in a

homologous genetic background (ie. Arabidopsis).

Several methods are now available for the analysis

of transient expression of reporter gene systems in

Arabidopsis tissue, yet none have been successfully

employed to date. Arabidopsis protoplast preparation and

transformation has been conducted using a PEG/Ca(NO3)2

method for the purpose of regenerating transgenic plants

from transformed protoplasts (Damm et al, 1988). Howev-

er, transformation of Arabidopsis protoplasts by this

method, or other methods such as electroporation, for

the analysis of transient gene expression has not been

reported.

Transformation by particle bombardment (Klein et

al., 1988a; Klein et al., 1988b;) has been successfully

employed in the transformation of cells in suspension and

intact plants, including Arabidopsis (Bruce, et al.,1989;

Seki et al, 1991). Particle bombardment requires far

less DNA (<10 ug) than the conventional methods of elec-

troporation or PEG incubation, and is not restricted to










any particular tissue or cell type.

Finally, plant cells or tissue may be transformed

by silicon carbide fiber ("whiskers") injection (Kaeppler

et al., 1990). Whiskers are suspended in water and mixed

with the tissue of interest in a buffered solution con-

taining DNA. The slurry is then vortexed briefly for

fiber injection. This method is being used routinely for

transformation of fly embryos (Cockburn, personal commu-

nication) and has been successful in transforming var-

ious tissues of the maize kernel (J. Baier, personal

communication).

In this chapter I report the transformation of

Arabidopsis for the functional analysis of the Arabidop-

sis Adh 5' flanking sequence. Deletions and mutations of

the Adh 5' flanking sequence are compared to the native

-950 bp fragment for regulation of transient expression

of the GUS reporter gene. Expression of native and

mutant Adh promoter regulated GUS constructs will be

assayed using both particle bombardment of Arabidopsis

seedlings and PEG-Ca(NO3)2 transformation of Arabidopsis

protoplasts.

Domains or sequence elements necessary for activa-

tion of Adh transcription are thereby identified, and a

comparison will be made between these results and those

obtained in the previous chapter.

Finally, a method for transformation of Arabidopsis










seedlings by silicon carbide fiber injection of plasmid

DNA will be described, and selected Adh constructs will

be examined by this method in order to evaluate its

usefulness as a tool for gene expression studies.


Materials and Methods

Plant materials



Arabidopsis cells in suspension were initiated from

seeds germinated on MS media (Sigma) containing 30%

sucrose, vitamins, myo-inositol, 0.8% agar, and 0.5 mg/L

2,4-d. Once established, callus was transferred to

liquid media described above. Cell suspensions were

maintained at room temperature and passage weekly.

Arabidopsis seeds were sterilized according to Valvekens

et al, (1989) and germinated in sterile MS plates without

2,4-D, or on potting soil in flats. Protoplasts were

prepared according to the method of Hauptmann et al.

(1987).



Gel Retardation assay

Conditions for gel retardation assay were identical

to those used in the previous chapter. Oligonucleotides

corresponding to the -210 probe used in the previous

study were synthesized to incorporate two contiguous base

pair substitutions at locations specified in the text,

and were used as competitor DNA for this assay..op












Construction of Expression Vectors



The GUS reporter gene system (Jefferson, 1987) was

selected based on the absence of detectable beta-glucuro-

nidase (GUS) in Arabidopsis. The GUS expression vector

pBI221 (Clonetech), shown in Figure 3-1 (Jefferson et

al., 1987) contains the GUS coding sequence and NOS

terminator cloned into the polylinker of pUC 19. A CAT

expression vector (obtained from L. C. Hannah), also

derived from pUC19, was used for the construction of Adh

5' chimeric plasmids in addition to pBI221.

Preparation of Arabidopsis Adh 5' flanking sequences

for cloning into the expression vector was begun by the

removal of a 2.2 kb EcoRl fragment from the Arabidopsis

Adh subclone, jAT3011 (obtained from E. Meyerowitz).

This fragment contained 1.0 kb of 5' flanking sequence in

addition to 1.2 kb of Adh coding sequence. This fragment

was then cloned into the EcoRl site of M13 for prepara-

tion of ssDNA. In vitro mutagenesis was performed by

the method of Kunkel et al.(1987). Oligonucleotides

previously synthesized for gel retardation analysis of G-

box mutations were utilized for the incorporation of

those base substitutions into the full length 5' flanking





























O-Glucuronidase (GUS)


i rr -
F r 5 i


Figure 3-1. GUS expression vector.

This vector was construct by cloning the bacterial
glucuronidase (GUS) coding region into the polylinker
region of plasmid pUC 19. From Jefferson et al., 1987.


T I


INOS-1 -










flanking sequence. Wild type or mutant (GM1 or GM2)

single stranded DNA from M13 was then used as a template

for PCR synthesis of fragments suitable for cloning into

the expression vectors (Figure 3-2).

Additional oligonucleotides were synthesized to be

used as PCR primers for preparation of Adh 5' cloning

fragments. The primers (17 to 19 bp long) were designed

to anneal either to the bottom strand at +30 relative to

transcription start, or the top strand at -940, -855,

-652, -482, -390, -289, -177, -177 and -77. Furthermore,

these primers were designed to incorporate either a BamH1

site at the 3' fragment end (+30 is 28 bp 5' of Adh

translation start), or an Sstl site at the various

"deletion" end points (Figure 3-3). Wild type or mutant

single stranded template was combined with the appropri-

ate primers for double stranded DNA synthesis and subse-

quent PCR amplification. PCR products obtained by this

method (shown in Figure 3-4) are practically homogeneous

and require only a minimum of preparation for the diges-

tion of 5' and 3' ends with restriction enzymes BamH1 and

Sstl.

The selection of 5' Sstl and 3' BamHl sites de-

scribed above was based on the CAT expression vector,

which contains the BamHl site closest to the CAT coding

region. Additional manipulations were required for

cloning into pBI221, and this is described in Appendix B.


















5'- AAATGCCa GGACGAA-3' --
3'- TAGTTGTCCTTTACGGTGCTGCTTATGATCGTTG5'


5'-


.3' -P


3'-


Figure 3-2. Mutagenic oligonucleotide binding to Adh
-210 region for in vitro mutagenesis

Oligonucleotides were designed to incorporate a
contiguous two base pair substitution at selected loca-
tions of the G-box was used for in vitro mutagenesis.
Arrows indicate the direction of second strand synthesis.
















+30


5 '- AAACTAACAA GAT CAAGCAAGTTTCTT -3 '
ATTGTT |CTA TTTCGTT -5'









-940


5' -TTGAGGAGC GTGAGAAAG---
3' CAAGAACTCCTCG JCACTCTTTCTAATTAAT -5'






Figure 3-3. PCR priming and mutagenesis for deletion
series construction.

The oligonucleotide PCR primers shown above (next to
arrows) were designed to anneal to either the Arabidopsis
Adh gene top strand at +30 (shown above) incorporating a
new BamHI site, or the bottom strand at various locations
on the 5' flanking sequence and incorporating an SstI
site. Arrows indicate direction of DNA synthesis.












S00 oh C0 un in
r- r4 N 0n n O




























1 2 3 4 5 6 7 8




Figure 3-4. PCR synthesis of Arabidopsis Adh deletion
fragments.

The products of PCR reactions using ssDNA template
and primers described in Figure 3-3 are shown above.
Fragment sizes are 100 bp (lane 1), 200 bp (lane 2), 330
bp (lane 3), 420 bp (lane 4), 510 bp (lane 5), 690 bp
(lane 6), and 880 bp (lane 7). Molecular weight markers
are shown in lane 8 (kb ladders).










Particle Bombardment



The method of gene transfer by high velocity micro-

projectiles (Klein et al., 1988a; Klein et al., 1988b)

was used for transformation of 4 day old Arabidopsis

seedlings. Five ug of supercoiled plasmid DNA was pre-

cipitated onto 1.2 um diameter tungsten particles for

subsequent bombardment. The bombardment device ("gun")

used for these experiments was designed and constructed

by this facility. Samples (still in the growth plate)

were inserted into the vacuum chamber of the gun (all

within a laminar flow hood) and a negative pressure of 26

mm Hg was obtained prior to bombardment. Samples were

then removed and plates were wrapped in parafilm. Fol-

lowing a 24 hr incubation at 240 C, 16 hr light/ 8 hr

dark, the samples were removed for histochemical staining

(Jefferson, 1987) at 370C for 24 hr.



Transformation by silicon carbide fiber injection



Arabidopsis was grown for 3 days on MS (2,4-D)

plates as previously described for particle bombardment.

Four day old seedlings are not amenable to this transfor-

mation method. On the 3rd day of growth, seedlings are

removed from plates and transferred to 1.5 ml eppendorf

tubes (approximately 50 seedlings per tube). Transforma-











tion was carried out as described by Kaeppler et

al.(1990) with the following modifications (Appel et al.,

1988) which are designed to ostensibly precipitate the

DNA onto the fiber; the transformation mixture contained

60 ug of supercoiled plasmid DNA, 125 mM CaC12, and 1.0

mM sodium phosphate, pH=7.0.

Samples are vortexed for one min, transferred to

plates, aspirated, and immersed in 400 ul of MS media.

Following 24 hr incubation, the samples are removed for

staining as previously described.



Protoplast transformation



Arabidopsis protoplasts were harvested in lots of

between 4.0 (107) and 8.0 (107), and allowed to incubate

for 30 min in 14 ml of MS/2,4-D media containing 0.4 M

mannitol. The protoplasts were then centrifuged for 10

min, 900 rpm, RT. The pellet was resuspended in MaMg

(Negrutiu et al., 1987) to a final protoplast concentra-

tion of 107/ml. Transformation was performed by the

method of Damm et al. (1989) using 50 ug each of test and

reference plasmid DNA (-950 Adh/CAT or 35S/luciferase,

obtained from H. Klee), 100 ug of salmon sperm DNA as

carrier, and a final concentration of 20 % PEG. Follow-

ing the 30 min PEG-CMS incubation, the transformation

mixture was transferred to 5.0 ml of MS (2,4-D) with 0.4










M mannitol in a 10 cm petri dish and wrapped in parafilm.

Samples were incubated 18 hr at RT in darkness. Proto-

plasts were recovered by centrifugation at 5000 g for 15

min. Following resuspension in 200 ul of the appropriate

extraction buffer, protein was extracted by grinding in a

1.5 ml eppendorf tube for 30 sec. Cell debris was re-

moved by a 2.0 min centrifugation at 40C, and supernatant

transferred to a fresh tube for assay of gene activity.


Assay of gene activity


GUS assays, histochemical or flourometric, were

performed according to Jefferson (1987). Biochemical

measurement of GUS for subsequent fluorimetric assay

included approximately 1.0 to 10 ug (100 ul) of total

protein (Bradford assay) in a final volume of 500 ul of

assay buffer. CAT activity was measured by the method of

Gorman (1982). Activity of the firefly luciferase refer-

ence gene was measured by a modification of the method of

Howell et al. (1989). (Protoplast transformations were

assayed separately for GUS and luciferase activity.) The

luciferase extraction buffer contained 0.1 M K phosphate

buffer, pH=7.8, 2.0 mM EDTA, 2.0 mM DTT, and 5% glycerol.

Luciferase assay reactions contained 10 ul of extract

(0.1-1.0 ug total protein), and 200 ul of reaction buffer

(25 mM tricine, pH=7.8, 15 mM MgC12, 5 mM ATP, and 0.5

mg/ml BSA.










Results

Gel retardation assay of G-box mutations


The 18 bp G-box oligonucleotide (-210) used for

characterization of G-box binding was used to evaluate

the influence of base substitutions on GBF binding affin-

ity. Those bases selected for substitution were highly

conserved G residues (Giuliano et al., 1988) known to

interact with protein from in vivo and in vitro DMS

footprinting (Ferl and Laughner, 1989; results of Chapter

2). The base sequences of native G-box and G-box muta-

tions 1 (GM1) and 2 (GM2) oligonucleotides are shown in

Figure 3-5.

Binding affinity assay of GM1 and GM2 is shown in

Figure 3-6. Crude whole cell extracts from cells in

suspension (lanes 1 to 4) or mature leaves (lanes 5 to 8)

are shown to bind the -210 oligonucleotide probe as

previously described in Chapter 2 (lanes 1 and 5).

Homologous -210 competitor DNA is effective in competing

for cell culture (lane 2) or leaf (lane 6) GBF binding at

50 ng. An equivalent amount of mutant G-box competitor

does not effectively compete for GBF binding to -210

probe (lanes 3, 4, 7, and 8). Partial competition is

observed for GM1 competitor using both cell culture (lane

3) and leaf (lane 7) extract. GM2 competitor, however,

is apparently unable to bind GBF and does not compete for

GBF binding to -210 probe.















-210

AAATGCCACGTGGACGAA
TTTACGGTGCACCTGCTT


AAATG
TTTAC


ACGTGGACGAA
TGCACCTGCTT


W ILD TYPE




GM1


AAATGCC4 GGACGAA
TTTACGG 1 CCTGCTT


Figure 3-5. Display of the Adh G-box Probe
Mutant Competitor Oligonucleotides.


and Adh G-box


Specific G-residues of the Adh G-box, which had been
shown previously (Figures 2-7, and 2-8) to contact pro-
tein, were selected for substitution in the synthesis of
mutant competitor oligonucleotides. Two locations of
mutations were selected, distal and central, with refer-
ence to the GTGG dyad axis of symmetry (shown above, GM1
and GM2 respectively). Contiguous two base substitutions
were selected with the intention of complete disruption
of protein binding.


GM2















.I
0
*r

04
0
U


0
.1.
O
0

0
0


rl
0 C
C I


k6S4i
p


U


1 2 3 4 5 6 7 8


Figure 3-6.
DNA binding.


Mutation of The Adh G-box disrupts protein-


The Adh G-box probe is bound by a protein component
of Arabidopsis crude whole cell extract from cell cul-
tures (lane 1) and mature leaves (lane 5) in the absence
of poly(dIdC). G-box/protein binding is not observed in
reactions which contain cold Adh G-box competitor DNA
(lanes 2 and 6). The mutant oligonucleotide competitor
DNAs, GM1 (lanes 3 and 7) and GM2 (lanes 4 and 8), do not
compete for Adh G-box protein binding.


J










Transformation of Arabidopsis Seedlings by Particle
Bombardment


The 1.0 kb 5' flanking region of Arabidopsis Adh, and

deletions or mutations thereof, were cloned into the GUS

expression vector pBI221 (Figure 3-7). These chimeric

constructs were used for the analysis of Adh regulated

transient expression in transformed seedlings and proto-

plasts.

Intact, four day old Arabidopsis seedlings were

subjected to transformation by particle bombardment. No

GUS activity was observed in seedlings which were bom-

barded with either no DNA or the CAT expression vector

(Figure 3-8,A). Seedlings bombarded with -940/GUS con-

tain numerous spots in all tissues (Figure 3-8, B), as do

those which were bombarded with -855/GUS, -652/GUS,

-482/GUS, -390/GUS, and -289/GUS (Figure 3-8; C, D, E, F,

and G). Each of these samples possess roughly equivalent

levels of GUS activity as estimated by the number of

spots and the intensity of blue color. These samples

also appear to have spots approximately equally distrib-

uted among tissues, with the exception of -390/GUS which

has an elevated number of spots in root tissue.

In contrast, very few blue spots are observed for

-177/GUS (Figure 3-8, H). GUS expression was not detect-

ed in the -77/GUS (Figure 3-8, I) construct.









G-BOX

4


I 1 II




I I II

I II



I I II
I I I





I II



GMI
I I I
GM2


-940/GUS

-855/GUS

-652/GUS

-482/GUS

-390/GUS

-289/GUS

-177/GUS

-77/GUS

-940.1/GUS

-940.2/GUS


Figure 3-7. Arabidopsis Adh 5' Deletion constructs and
G-box mutations.

The 5' flanking sequence of Arabidopsis Adh was
cloned into the GUS expression vector pBI221 (-940/GUS).
Deletions from the 5' end of the Adh flanking sequence
were prepared at approximately 100 bp intervals from
-940, and these were also cloned into pBI221 (-855/GUS,
-652/GUS, -482/GUS, -390/GUS, -289/GUS, -177/GUS, and
-77/GUS).
The Adh G-box mutations described in Figure 3-1 (GM1
and GM2) were also incorporated into the full length
promoter via. site directed mutagenesis, and these frag-
ments were cloned into pBI221 as well
These Arabidopsis chimeric constructs were used for
all transient expression experiments to follow.

























4-I
(0
04
>1





4-)


(0
C1
*H


.1-1
0







$4-
14 (




0)0
*u.



4-4
0
*Ia)












(0



0
14A
E3
C)
*Ha)












14J
10









0 .Q







ow
u
*^

I (0
r 0
>
c4,
3c
0)
*f -
c U


C -H r0 V
43-4 -
.1- *r ) 0 r-l -

3 0 0 U



S UH )
00 ,

:^ 6 ia r







0c^
S 0 rz















4H 0 4 M
4 (00 1





4 M 0) $4 ..
0


a'r ^



1 r r- i







14 M 3 1l
0 0 ( 0 i


0 H wa'lO






















L


C
'V


















a








I



4
H

Ali=,
Al










Each construct photograph is the combination of two

replicate samples from a single experiment, and repli-

cates were generally in close agreement.

Particle bombardment of 4 day old seedlings was also

conducted using the -940/GM1/GUS and -940/GM2/GUS con-

structs (Figure 3-9, A and B). GUS activity is observed

for both mutant constructs, with the number and intensity

of blue spots being only slightly reduced. A summary of

these results is provided in Table 3-1.

Mature Arabidopsis seedlings were also examined by

this method for Adh regulated expression (Figure 3-10).

Fourteen day old seedlings were transformed with -940/GUS

and subsequently allowed to incubate with or without

oxygen. GUS activity is observed in leaf tissue only,

and did not significantly vary with oxygen levels. The

relative absence of root activity (one spot in the -02

sample) most likely reflects the low frequency of trans-

formation inherent in this method. Expression of GUS in

4 day old seedlings was not significantly altered by

oxygen depletion (data not shown).

Particle bombardment was attempted for cells in

suspension using a variety of cell densities and cell

ages. GUS activity was not detected in these experiments

using either 35S/GUS or -940/GUS.

Biochemical measurement of GUS activities in seed-

lings was prevented by high background fluorescence.



























.C
4-)
3

'p

1-4
'o

(0




*4



0
r.








.,

S-




0

40

-4 U


0 .


0
C






S0




40 1
my






C I


*C4 En

13 -:
41.
C 0
a3

co


)O U
f 3

* 4
S.I-


4)C
0
-,4





O,
0al






0

Wa c
En










> '0
mUN
tOt



I.
ai '0

00



.- u
*W O
(U ^






76
,fl'!JS.,i i l;


. .,"'


4


r


p


~a~]

















Table 3.1

Summary of
results

Construct

-940/GUS
-855/GUS
-652/GUS
-482/GUS
-390/GUS
-289/GUS
-177/GUS
-77/GUS


-940.1/GUS
-940.2/GUS


Particle bombardment transformation


Activity

+++
++++
+++
++
+++
+++
+



++
++














f4e


,1

mnl


1
I '


' : 4 :
|k^-A


4'.


AM


Figure 3-10.
lings.


Transformation of mature Arabidopsis Seed-


Thirteen day old seedlings were transformed by
particle bombardment with -940/GUS. Fewer seedlings were
transformed and assayed due to seedling size. Root
tissue transformation was not frequent.


1

p: 1
;i- *










Transformation by Silicon Carbide Fiber "whiskers"
Infection


Silicon carbide fiber transformation was attempted

using seedlings at 1, 2, 3, and 4 days after germination.

Successful transformation was observed for 3 day old

seedlings only. In general, this method was far less

efficient than particle bombardment for transformation.

Transformed samples contained 1 to 5 spots, and spot

intensity varied considerably among replicates. When the

full length promoter construct (-940/GUS) was compared

with -289/GUS, and -177/GUS, activity was observed in

only the -940/GUS and -289/GUS treatments (Table 3-2).

An occasional blue staining was observed in the

shoot apex which appeared to have either spread exten-

sively from the point of origin, or was a result of

expression in more than one cell. This was observed in

three seedlings, and in each case the shoot apex or leaf

primordia were uniformly stained a dark blue or indigo.

One of those seedlings is shown in Figure 3-11. It

should be noted here that Arabidopsis seedlings are

growing rapidly by the third day after germination, and

continue to grow for the 24 hr following bombardment,

prior to staining.

Seedlings which had been transformed were examined

(G. Erdos) for fiber injection with scanning electron



























Table 3-2

Summary of Whiskers Transformation

Construct Activity

-940/GUS +++
-855/GUS +++
-390/GUS +++
-289/GUS +++
-177/GUS
-77/GUS














































Figure 3-11. Transformation of Arabidopsis by "whisker"
injection.

Three day old seedlings were transformed by the
"whiskers" method using -940/GUS. Seedlings remain
intact following one minute vortexing at high speed.
Blue stain appears in leaf primordia.


















































Figure 3-12. Electron Microscopy of "whisker" injection.

Seedlings subjected to transformation by silicon
carbide fiber injection were fixed and prepared for SEM.
This fiber is atypical in that it is bifurcated at one
end. Many fibers could be found in contact with the
plant surface without actual penetration.










microscopy (Figure 3-12). Many fibers were found in

contact with the plant surface, but very few had obvious-

ly penetrated the cell wall. Penetration was more fre-

quent in leaf tissue than roots.



Transformation of Arabidopsis protoplasts



Protoplasts were prepared on the sixth day of a

seven day passaging cycle. Protoplast yield from approxi-

mately 5.0 grams fresh weight of suspension culture cells

varied between 5 x 107 and 108. A minimum of 106 proto-

plasts/300 ul was necessary for efficient transformation.

Lower protoplast amounts resulted in non-linearity of GUS

activity. Cells in suspension appear yellow in color

when properly maintained, and unhealthy suspension cul-

tures (brown or pale yellow) generally yielded few proto-

plasts regardless of the amount of starting material.

Relative GUS activities of the full length, dele-

tion, and mutation constructs (Figure 3-7), using

-940/CAT as the reference construct, were examined on

6/21/91 by transformation of protoplasts, and the results

are shown in Table 3-3. Following cell harvest and

extraction, total protein was measured from 10 ul of

extract (column 1) and GUS assays were then performed by

measuring the production of 4-methylumbelliferone (4MU)

at various time points (columns 2 through 5). The re-

























Table 3-3
Total Protein

Total
Protein
(ua/ul)


0.563
0.589
0.592
0.601
0.518
0.484
0.398
0.495
0.402
0.452
0.459
0.472
0.474


and Fluorescence for Experiment 6/21/91

Fluorescence, (nM 4MU)


20'

0.343
1.650
0.350
9.023
3.487
4.775
7.921
2.330
2.220
7.830
3.140
0.520
0.437


40'

0.325
3.073
0.309
18.76
6.820
9.040
15.39
4.430
5.030
15.46
5.968
0.709
0.535


60'


0.349
4.336
0.294
28.76
10.19
13.12
33.89
6.320
6.360
24.70
8.199
0.831
0.539


80'


0.223
5.889
0.228
38.44
12.90

33.26
8.539
8.624
31.06
11.51
1.048
0.340

























Table 3-4

Protein per assay and Fluorescence per mg protein
for Experiment 6/21/91


protein per
assay
Imsg x 10
2.81
2.94
2.96
3.00
2.59
2.42
1.99
2.48
2.00
S2.26
S2.30
2.36
2.37


20'
1220
5600
1180
30060
13460
19720
39800
9380
11100
34640
13640
2200
1620


Corrected Fluoresce
(nmoles 4MU/mg protein)
40' 60' 80'
1160 1200 793
10440 14760 20032
1040 1020 770
62520 95880 128160
26320 39300 49841
37320 54180
77320 112980 167120
17840 25440 34431
25160 31800 43120
68440 109320 137440
25960 35640 50000
3000 3540 4440
2240 2280 1434










suits of Table 3-3 are normalized for protein amounts

(Table 3-4) and thenplotted as GUS activity (nM 4MU/mg

total protein) as a function of time (Figure 3-13, 3-14,

and 3-15).

The linearity of increasing 4MU fluorescence over

time is observed with all constructs except the no DNA

and 35S/CAT controls, and the -77/GUS construct. Corre-

lation coefficients range from 0.983 for -482/GUS to 1.0

for -940/GUS verifying enzyme activity to be in the

linear range. (An increase in activity of the -940/GUS

construct is occasionally observed by the addition of the

35S/CAT reference construct). The 35S/CAT was present in

all other deletion or mutation transformations.

The slopes of lines plotted in Figures 3-13, 3-14,

and 3-15 are then presented as GUS activities relative

to the full length construct (Figures 3-16 and 3-17).

Activities of the deletion constructs fell into three

general categories (Figure 3-16). Deletions to -855 or

-390 result in no significant reduction in GUS activity.

The -652, -482, and -289 deletion constructs produce less

than half of wild type activity, and further deletions to

-177 and to -77 essentially abolish gene activity.

The G-box mutation constructs 940.1/GUS and

940.2/GUS (Figure 3-17) were compared to -940/GUS, and

were found to significantly reduce the ability of the

full length promoter to activate GUS expression.



















200000








100000








0


0 20 40 60 80

time (min.)


Ia no DNA
2* -940/GUS*
3 35S CAT
4. -940/GUS


y = 1403.5 6.2050x RA2 = 0.631

y = 804.00 + 238.08x R^2 = 0.999
y = 1315.0 6.2500x R^2 = 0.895
y = 2760.0 + 1638.3x R^2 = 1.000


Figure 3-13. Time course of 4MU Fluorescence from exper-
iment 6/21/91.

Fluorescence values used here are from Table 3-4.
The slope of each line proceeds the x in each equation.
R^2 is correlation coefficient. The -940/GUS treatment
does not include the 35S/CAT reference construct. All
other treatments include 35s/CAT.
















200000

*2

5
0.





* 100000


0
E








0 20 40 60 80 100
time (min.)

I a -940/GUS y = 2760.0 + 1638.3x R^2 = 1.000
2 -855/GUS y = 5100.0 + 2088.1x R^2 = 0.990
3 M -652/GUS y = 1084.5 + 413.76x R^2 = 0.999
4 -482/GUS y = 2120.0 + 513.50x R^2 = 0.983
5 M -390/GUS y = 140.00 + 1746.4x RA2 = 0.995

6 0 -289/GUS y = 1620.0 + 593.80x R^2 = 0.995
7 A -177/GUS y = 1480.0 + 36.300x R^2 = 0.992
8 A -77/GUS y = 2023.0 2.5900x RA2 = 0.024








Figure 3-14. Time course of Fluorescence of Deletion
Constructs from experiment 6/21/91.

















140000


120000


100000


80000

60000


40000


20000


0


0 20 40 60 80


time (min.)


-940/GUS
-940.1 GUS
-940.2/GUS


y = 2760.0 + 1638.3x RA2 = 1.000
y = 1699.5 + 610.61x RA2 = 0.998
y = 2613.3 + 861.50x R^2 = 1.000


Figure 3-15. Time course of Fluorescence of G-box muta-
tion constructs from experiment 6/21/91.















200 -


100-


CAT -940 -855 -652 -482 -390 -289 -177 -77


construct









Figure 3-16. GUS activity of Deletion constructs rela-
tive to the full length promoter construct,-940/GUS.

Absolute GUS activity for -940/GUS varied between
1.5 and 3.0 umoles 4MU/min/mg total protein among experi-
ments. Relative activities are calculated as a percent-
age of -940/GUS activity in the same experiment.









































20


-940 -940.1 -940.2


construct








Figure 3-17. GUS activity of G-box mutant constructs
relative to the full length promoter construct.










Repeated attempts to obtain measurable CAT activity

uaing either 35S/CAT or -940/CAT were unsuccessful. The

-940/CAT construct was maintained in the transformation

mixture over the following 2 experiments in order to 1)

allow for comparisons between this and subsequent experi-

ments, and 2) to function as carrier DNA. This procedure

was continued until the procurement of the 35S/Luciferase

(Ow et al., 1986) construct.

A technical problem arises when attempting to meas-

ure relative activities of several constructs (limiting

protoplast yield), as well as time constraints at partic-

ular stages of transformation and assay. Initial experi-

ments were designed based on the assumption that activi-

ties (relative to full length) should agree from one days

experiment to the next. Each construct was transformed

(without replication), and the relative activities were

analyzed in combination with corresponding activities

obtained in subsequent experiments to derive an average

relative activity.

Therefore, the previous experiment was repeated on

6/27/91, as shown in Tables 3-5 and 3-6, and these re-

sults are plotted in Figures 3-18, 3-19. The resulting

GUS activities are displayed in Figures 3-20 and 3-21. A

comparison of the results of experiments 6/21/91 and

6/27/91 is presented in Figure 3-22. As opposed to the

first experiment, relative activities of the deletion

























Table 3-5

Total Protein and Fluorescence for Experiment
6/27/91

total Fluorescence (nM 4MU)
protein
(uq/ul) 20' 40 60 80
1 0.334 0.000 0.000 0.000 0.016
2 0.411 0.936 1.990 2.967 3.764
3 0.369 1.667 3.456 5.197 6.204
4 0.302 0.000 0.084 0.191 0.396
5 0.328 0.081 0.331 0.537 0.824
6 0.270 0.425 0.977 1.638 2.169
7 0.409 0.610 1.421 2.172 2.765
8 0.399 0.670 1.477 2.369 3.007
9 0.249 1.164 2.515 4.088 5.701
10 0.374 0.926 2.153 2.897 3.976
11 0.332 0.000 0.000 0.178 0.219
12 0.386 0.000 0.000 0.022 0.024
























Table 3-6

Protein per Assay and Fluorescence per mg Protein
for Experiment 6/27/91


protein per
assay
I(m x 1051
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25


20'
0000
37440
66680
0000
3240
17000
24400
26800
46560
37040
0000
0000


Corrected Fluorescence
(nmoles 4MU/mg protein)
40' 60'80'
0000 00006440
79600 118680 158088
138240 207880 260568
3360 7640 15840
13240 21480 32960
39080 65520 88929
56840 86880 113365
59080 94760 123287
100600 163520 239442
86120 125454 165401
0000 7120 8760
0000 880 960



































20 40 60 80
time (min.)


a -940/GUS
* -855/GUS
* -652/GUS
* -482/GUS
* -390/GUS
a -289/GUS
A -177/GUS
A -77/GUS


5516.0 + 3256.5x R^2 = 0.995
- 7924.5 + 1211.1x R*2 = 0.999
- 3862.5 + 1484.7x RA2 = 0.998
- 5303.5 + 1625.7x R^2 = 0.998
- 2.2861e+4 + 3207.8x R^2 = 0.994
- 2600.5 + 2122.1x R^2 = 0.997
- 4380.0 + 167.00x R^2 = 0.866
- 480.00 + 18.800x R^2 = 0.832


Figure 3-18. Time course of Fluorescence for Deletion
Constructs from Experiment 6/27/91.

Results are presented here as in Figures 3-13 to 3-
15. Slopes of these lines are used to determine absolute
GUS activity.


300000


200000





100000





0