ABSTRACT
Transcription of genes
DLEC2
and
PHS
[beta]
is specifically and coordinately regulated during maturation of
Phaseolus
embryos. Over-expression of the seed- specific factor PvALF in cotyledon cells results in transactivation
of either promoter. PvALF is related to the VP1 protein of maize, which
transactivates gene expression via G-boxes, Sph elements and AT-rich sequences. We used deletions and base substitutions in the
DLEC2
and
PHS
[beta]
promoters to demonstrate that several conserved RY-repeats were necessary for PvALF induction of both genes. A comprehensive
mutational and transactivation analysis was used to define functionally the
sequence of the
DLEC2
repeat RY3 as
G
/
C
CATGCxx
G
/
C
. We also found that an interaction between RY3 and the 3
'-flanking tetranucleotide CCAC increased PvALF transactivation. A preferred spacing and phasing requirement for the RY3 and CCAC motifs suggested the
possibility of interactions between cellular factors that recognize either
element. The high conservation of Sph-RY motifs in seed-specific promoters from monocots and dicots indicates that organ and
temporal specification by factors similar to VP1 and PvALF is common among seed
plants.
At the end of embryonic morphogenesis, higher plant embryos initiate a process
of maturation characterized by the accumulation of protein, lipid and
carbohydrate reserves, which is followed by the establishment of desiccation
tolerance late in embryogenesis. Viviparous mutations
vp1
and
abi3
in maize and
Arabidopsis
, respectively, disrupt normal induction of
Genes of VP1/ABI3-like factors have been cloned from maize (
VP1
,
9
), rice (
OsVP1
,
10
),
Arabidopsis thaliana
(
ABI3
,
11
) and
Phaseolus
(
PvALF
,
12
). VP1 was isolated first and it is considered as the prototype for the group.
Although VP1 and PvALF share only 38% overall identity at the amino acid level
(
12
) they are both transcription activators (
5
,
9
,
12
,
13
), and contain similar N-terminal, acidic segments that function as a transferable transcription
activation domains in yeast and plant cells (
9
,
12
). Because VP1 and PvALF lack typical DNA binding domains, they probably
activate transcription in combination with other factors, including DNA binding
proteins. Analyses of transactivation of
Em
and
C1
promoters by VP1 (
5
,
13
,
14
) have indicated a general requirement for three types of
cis
-acting DNA elements: G-boxes, originally described as recognition sequences for basic-leucine/zipper (bZIP) proteins and for certain types of bHLH
proteins (reviewed in
15
), an Sph element (
5
), and A/T-rich sequences (
14
). VP1 transactivates ABA-dependent
Em
expression via G-box complex I, and ABA-independent expression via different sequences which include the A/T-rich motifs and the Sph element (
14
). The latter is similar to the RY-repeats which were described first in genes
DLEC1
and
DLEC2
encoding the two subunits of seed phytohemagglutinin from
Phaseolus vulgaris
(
16
). Subsequently, identical or very similar motifs were found upstream of many
seed-specific genes from legumes and other dicots (
17
). Mutational analyses in tobacco demonstrated that the RY-repeats of legumin, glycinin and [beta]-conglycinin genes are necessary for positive regulation in
seed tissues (
18
-
21
), and negative regulation in vegetative organs (
18
,
22
). However, precise determinations of the nucleotides essential for positive or
negative transcriptional regulation, or the relations between the RY-repeat and other promoter DNA signals or cellular
trans
-acting factors remain largely unknown.
Previously, we reported that the seed-specific regulator PvALF transactivates the promoters of genes
DLEC2
and
PHS
[beta], which are coordinately induced at the beginning of seed maturation (
12
). As a step towards addressing the molecular interactions underlying PvALF transactivation, we have elucidated the structure of a PvALF response
complex. Our analysis revealed that some, although not all RY-repeats of genes
DLEC2
and
PHS
[beta] are necessary and sufficient for PvALF transactivation. Moreover, an
interaction between the RY-repeat and a neighboring CCAC-box in the
DLEC2
promoter enhances gene expression in response to PvALF. Preferred spacing and
phase (angular position around a B-DNA helix) requirements for this interaction suggest the presence of a
multiprotein complex associated with the RY/CCAC-box operator. These results constitute the first published demonstration
of a functional link between the RY-repeats of dicot storage protein genes and a cloned transcription factor,
and pave the way for further characterization of VP1/ABI3 regulatory pathways
in legumes and other dicots such as soybean, tobacco,
Brassica
and
Arabidopsis
.
Promoter fragments -295PHS, -120PHS, -230DLEC2, -133DLEC2 and MPHA were amplified by 30 cycles of PCR
(94oC, 1 min; 50oC, 1 min; 72oC, 1 min) from wild-type [beta]-phaseolin (
PHS
[beta]) or phytohemagglutinin L-subunit (
DLEC2
) promoters using synthetic oligonucleotide primers. Mutations near the 5'-ends of the MPHA and -133DLEC2 fragment were introduced into the 5' amplification primer. For internal mutations, two partially overlapping oligonucleotides
containing each mutation were prepared, and used as upstream and downstream
primers in separate PCR reactions with wild type 5'- and 3'-end primers. The products of PCR were separated from
unincorporated primers in a 2% agarose gel, eluted from a gel slice by overnight diffusion at
room temperature in TE buffer, and combined in a primerless PCR reaction for 5
cycles. The full length mutant promoter was then produced by adding 5'- and 3'-terminal primers followed by 20 cycles of
amplification. In all cases, the final promoter fragment was phenol and
chloroform extracted, EtOH precipitated, and digested overnight with
restriction enzymes; the restriction product was gel purified, eluted and
cloned into an expression vector containing the
uidA
gene encoding [beta]-glucuronidase. Restriction sites for
Eco
RI and
Xba
I were added to the ends of the MPHA fragments used in Figures
2
and
3
(wild type sequence is ggaatt
C
ACCATGCATGCTGCCACCTCAGCTCCCGCCTCTTCACCGTGTCTTTCTCTagagc, where lowercase represents added nucleotides) for cloning upstream of a -64 cauliflower mosaic virus 35S promoter fragment driving the
uidA
reporter gene; all other promoters were cloned into pBI221 (Clontech, San
Francisco, CA) between sites for
Hin
dIII and
Xba
I. All promoters were sequenced once cloned into the expression vector.
Leaves were collected from greenhouse-grown
Phaseolus vulgaris
cv `Tendergreen' plants, and surface sterilized in 20% bleach for 45 s. The
leaves were thoroughly rinsed in sterile water, and circular sections removed
with a sterilized 1' cork borer. Leaf disks were placed on Gamborg's B5 medium (BRL,
Bethesda, MD) pH 5.7, supplemented with mannitol (0.75 M), and 5 mM each
proline and glutamine, and solidified with 0.8% Bacto-agar (Difco, Detroit, MI). Leaf disks were incubated for >= 2 h prior to bombardment. Tungsten particles (average diameter ~1.3 micron; Bio-Rad, Hercules, CA) were washed three times with ethanol by
sonicating 30 s to suspend particles, and centrifuged 20 s, then washed twice with sterile water. A total of 3.75 [mu]g DNA in 37.5 [mu]l water was precipitated onto 3.75 mg prepared tungsten particles in 150 [mu]l water by the addition of 150 [mu]l 2.5 M CaCl
2
, 60 [mu]l 0.1 M spermidine with continuous vortexing. The mixture was vortexed 3 min and
the particles collected by centrifugation. The pellet was washed with 700 [mu]l 100% ethanol and resuspended in 60 [mu]l 100% ethanol, of which 10 [mu]l was used per bombardment. Bombardments were performed at 1550 p.s.i., 28 inches Hg vacuum in a Bio-Rad Biolistic particle delivery system. After bombardment,
leaf disks were incubated overnight at 26oC in darkness.
A -230
DLEC2
promoter that directs seed-specific expression of phytohemagglutinin in transgenic tobacco (
24
) is transactivated by PvALF in
Phaseolus
cotyledons (
12
). The importance of three RY-repeats shown in Figure
1
A for transactivation was investigated by the particle bombardment technique described previously (
23
). GUS activity was measured 24 h after bombardment with reporter DNA, plus
PvALF expression plasmid pALF (
12
), or a control plasmid lacking a PvALF coding sequence (pJIT), and the ratio of
GUS to luciferase activity (GUS/LUCm) was used as a relative measure of gene
expression. Throughout this article, GUS expression in the absence of the PvALF
effector gene is referred to as `endogenous expression', and is represented by
clear bars; `activated expression' designates the amount of GUS activity in
samples containing pALF, and is represented by shaded bars; `induction'
indicates the increase in GUS activity resulting from over-expression of PvALF, and is calculated as the ratio of activated versus
endogenous expression.
The canonical RY-repeat octamer CATGCATG (
16
) is a consensus deduced from natural DNA sequences (
17
) rather than the result of direct functional tests or protein-DNA binding assays. Like the experiments of the previous section,
mutational studies of RY-repeats have typically involved mutation (deletion, insertion or substitution) of several nucleotides at a time (
5
,
9
,
13
,
14
,
18
-
21
,
26
). Moreover, the different relative importance of individual RY-repeats for PvALF transactivation.
According to Figure
1
A and B, motif RY3 makes a large contribution to the transactivation of promoter
DLEC2
. In order to address future questions concerning protein contacts with specific
base pairs of this motif, and the role of its sequence context on PvALF
transactivation, a 50 bp fragment from the
DLEC2
promoter (nucleotides -115 to -65, fragment MPHA) was fused to a heterologous minimal promoter, -6435S. The resulting construct MPHA/35S is illustrated in Figure
3
A; Figure
3
B (bars labeled -6435S and MPHA/35S) shows that the level of PvALF induction increased
significantly with the inclusion of fragment MPHA. Subsequently, single-base mutations (purine-to-purine or pyrimidine-to-pyrimidine) were engineered at 20 positions indicated with a double-headed arrow in Figure
3
A, including and flanking motif RY3. The corresponding amounts of endogenous and
activated expression measured from each mutant promoter, and the level of
induction, are displayed in Figure
3
B. Individual nucleotides, their relative position and the mutations are shown
under the graph. Mutations within repeat RY3 at positions 3-7 and 10, decreased the level of induction 50% or more. Promoters
carrying mutations at the first two bases of repeat RY3 (positions 3 and 4)
also showed slight increases in the amounts of endogenous expression. These
results tentatively defined the pentanucleotide CATGC as the core of repeat
RY3. Additionally, mutations in the tetranucleotide CCAC located downstream of
repeat RY3 (positions 14-17), also reduced induction ~50%. In particular, the mutations at positions 15 and 16 caused significant
reductions in both endogenous and activated expression.
The sufficiency of the RY-repeat for PvALF transactivation was tested by fusing synthetic DNA
fragments containing one to three copies of the RY-repeat to the minimal -6435S promoter, as depicted in Figure
4
A. Plasmid MPHA/35S was used as a positive control, respectively. Figure
4
B shows that a single copy of the RY-repeat in construct 1RY/35S yielded a slight increase in the amounts of
endogenous and activated expression, and in the level of induction (4*) relative to the negative control -6435S promoter (1*-2*, data not shown). However, both activated
expression and the level of induction increased significantly with two and
three copies (2RY/35S and 3RY/35S), indicating that multimers of the RY-repeat can direct PvALF transactivation in the absence of a CCAC-box.
Due to the rigidity of double-stranded DNA helices over short spans of <60-100 bp, interactions among protein factors bound to DNA can be
greatly influenced by the position of their corresponding binding sites on the
surface of the DNA helix (
27
). A similar situation could apply to the RY3/CCAC-box complex where the cores of repeat RY3 and the CCAC-box are separated by one turn of B-DNA helix. An effort was made to test the range of distances
and phase angles that are compatible with PvALF transactivation from this
complex. Consequently, 2, 5, 7 or 10 bp were moved from a location 22 bp
upstream of motif RY3, and inserted between nucleotides T12 and G13 (Fig.
3
) that separate the RY3 repeat from the CCAC-box. A standard B-DNA value of 34.5o per base pair (
28
) yields calculated phase shifts of approximately 70, 170, 245, and 345 degrees,
respectively. The resulting mutant promoters RY+2 to RY+10 are depicted in
Figure
5
A. In these promoters, the distance and phase angles between the RY3 motif and
the TATA-box have also been changed. In order to control for those changes, and for
the effect of moving the spacer element to a different location within the
promoter, the same sequences were inserted downstream of the CCAC-box (constructs RY/CB+2 to RY/CB+10). Both series of mutant promoters were
tested in leaf transactivation assays and the corresponding results are
displayed in Figure
5
B and C. Figure
5
B shows that all the insertions between RY3 and the CCAC-box (RY+2 to RY+10) caused significant reductions in activated expression
and level of induction. In particular, promoters RY+2 and RY+7 yielded very low
levels of induction (2- and 3-fold, respectively), similar to the -6435S minimal promoter. Transactivation was partially
restored, however, in promoter RY+10 to 7-fold, corresponding to 70% of the amount seen with the intact WT promoter.
All the control promoters (RY/CB+2 to RY/CB+10) displayed comparable levels of
induction, which were also very similar to the wild-type (Fig.
4
B). Although these results resemble those from analogous `phasing' experiments
which have been performed with other systems, without testing a large number of
permutations of the spacer sequence(s) it is difficult to completely rule out
that changes in gene expression simply reflect the creation of new, transcriptionally active signals (either positive or negative). Therefore, while individual base
pairs between repeat RY3 and the CCAC-box, e.g. at nucleotides T12 and G13, can be changed without causing much
detriment to PvALF transactivation, altering the spacing and phase angle
between the two elements can seriously impair PvALF induction. However, the
activity of the operator is partially restored by introducing a full helical
turn (~10 bp) between repeat RY3 and the CCAC-box, suggesting that PvALF transactivation favors a configuration that has these two components of the complex on approximately the same side of DNA helix.
Analyzing the promoter elements involved in PvALF transactivation was a crucial step towards understanding the mode of action of this regulator.
Figure
6
depicts the main conclusions of this study. We provide the first experimental
evidence linking the conserved RY-repeats of dicot storage protein genes to a cloned transcription
activator, in this case PvALF, a member of a more general family of seed
transcription factors first identified in monocots. The relationship between
PvALF and the RY-repeats was suggested by the demonstration that VP1 transactivation of
monocot
C1
and
Em
promoters involved similar motifs, known as Sph elements (
5
,
9
,
10
,
13
,
14
). Although the nucleotide sequences of the
DLEC2
RY-repeats RY1, RY2 and RY3 are identical one to another, mutations in each
motif caused different changes in PvALF transactivation. We show that the
higher activity of repeat RY3 is due, at least in part, to its proximity and
spatial orientation relative to a CCAC-box. A CCAC tetranucleotide is also present within the Coupling Element 1
(CE1), a
cis
-acting element involved in abscisic acid (ABA) regulation of the
Hva22
gene from barley (
29
). Not surprisingly, the CCAC-box occurs frequently on many plant promoters active in seed and non-seed tissues (M. Bustos, unpublished). By contrast, RY-repeats have been found almost exclusively upstream of dicot
genes for storage proteins, lectins and oil-body proteins that are specifically expressed in the seed (
17
). It is important to note that the published sequence of the ABI3-regulated
At2S3
gene promoter of
Arabidopsis
(
8
,
30
) includes an RY/CB complex within 60 bp upstream of the TATA box, indicating
that ABI3 also regulates RY-containing genes in that species. Very recently, Kao
et al
.
(
31
) reported that an Sph element, which is very similar to the RY-repeats of dicot storage protein genes, is necessary for transactivation
of the C1 promoter of maize by VP1. Thus, the parallels between ABI3 and VP1
and their corresponding operators, i.e. RY and Sph elements, indicate a rather
high degree of conservation in the molecular mechanisms underlying the function
of VP1/ABI3 proteins in higher plants.
The authors are grateful to Dr Donald Helinski (UCSD, La Jolla) for the gift of
expression vector pJIT82, Dr Carol Greitner for expert maintenance of plant
materials. This work was supported by grants from the National Science
Foundation (MCB-9219203) and the US Department of Agriculture National Research Initiative
(No. 9303090) to M.M.B. M.-S.C. and A.J.B. were the recipients of Special Research Initiative Support/Graduate Research Assistantships from the University of Maryland Graduate School,
Baltimore.
*To whom correspondence should be addressed. Tel: +1 410 455 2769; Fax: +1 410
455 3875; Email: bustos@umbc.edu
+
Present address: Department of Biochemistry, School of Hygiene and Public
Health, Johns Hopkins University, Baltimore, MD 21205, USA
REFERENCES
Return