Investigations of ABA Insensitive Revertants of eral in Arabidopsis thaliana Sara Feriel Sarkar A thesis submitted in conforrnity with the requirements for the degree of Master of Science Graduate Department of Botany in the University of Toronto O Copyright by Sara Feriel Sarkar, 1999
88
Embed
Investigations of ABA Insensitive Revertants · insensitive, and have been designated _insemitive _reveflmts of gral. Genetic analysis reveaied that ire mutants suppress eral dominantly,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Investigations of ABA Insensitive Revertants of eral
in Arabidopsis thaliana
Sara Feriel Sarkar
A thesis submitted in conforrnity with the requirements for the degree of Master of Science
Graduate Department of Botany in the
University of Toronto
O Copyright by Sara Feriel Sarkar, 1999
National Library 1+1 of,,, Bibliotheque nationale du Canada
Acquisitions and Acquisitions et Bibliographie Services services bibliographiques
395 Wellington Street 395. nie Wellington ûîtawaON K l A W O(tawaON K l A W canada CaMda
The author has granted a non- exclusive licence allowing the National Libmy of Canada to reproduce, loan, distniute or seil copies of this thesis in microform, paper or electronic formats.
L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, disûi'buer ou vendre des copies de cette thèse sous la fonne de microfiche/film, de reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fkom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.
Investigations of ABA Insensitive Revertants of erul
Master of Science 1999
Sara Feriel Sarkar
Department of Botany
University of Toronto
Abstract A screen for suppressors of the ABA eral supersensitive mutant in Arabidopsis thaliana
had previously been perforrned in an effort to identifj. targets of the ERAl famesyl
transferase and to identiQ downstream components of the ERAl ABA-dependent
s i g n a h g pathway . Out of 124 lines that retested, eight suppressors of eral are ABA-
insensitive, and have been designated _insemitive _reveflmts of gral. Genetic analysis
reveaied that ire mutants suppress eral dominantly, but their insenkitive phenotype is
recessive. ire mutants represent three complementation groups, one of which is a new
mutation in the seed-specific AB13 transcription factor that has previously been shown to
mediate ABA responses during seed development In addition to the seed supersensitivity
of eral. ire mutants suppress adult phenotypes of eral. including.drought toIerance. This
is the first demonstration that AB13 has a genetic role outside of the seed.
Abbreviations ABA
AB1
o c
DNA
EDTA
EMS
F1
F2
FTase
g
GA
GGTase
kb
Ler
Ml
M2
MES
MC011 ++
RNA
SDS
S S C
TAE
clE
PM
UV
abscisic acid
abscisic acid insensitive
degrees Centigrade
deoxyribonucleic acid
ethylenediaminetetraacetic acid
ethylmechanesuifonic acid
first filial generation
second f ~ a l generation
farnes y 1 tram ferase
force due to gravity
gibberellic acid
gerany lgerany ltransferase
kilobase(s)
Landsberg erecta ecotype
EMS mutagenized seed, fmt generation
EMS mutagenized seed, second generation
rnorpholinoethanesuIfonic acid
Meyerow itz Columbia ecotype
ribonucleic acid
sodium dodecylsulfate
sodium chlonde/sodium citrate solution
Tridace tateEDTA solution
microeinstein
micromolar
ultraviolet light
Table of Contents .......... ......................................................................................................................... Abstract .. - i
Abbreviations ........................................................................................................................... M List of Figures .................................................................. ................................................... v List of Tables ........................................................................................................................... vi
Materials and Methods,.-.. .....- 1 3 Growth Conditions ................................................................................................................ 3
........................................................................................ Suppressor screen of era 1 1 4 ......................................................................... Determination of Hormone sensitivii 1 4
Results 2 3 ................................ ............................................................ Suppressor screen of era 7 .. 23 . . . .......................................................................................... ABA Insensitivity of ire mutants 23
ire mutants are Dominant for Suppression of era 1, but Recesive for ABA insensnivity ................................................................................................................................................ 30
.................................................................. ire mutants faIl into 3 complementation groups 36 .............................................................. .....*...................-....... 1 3-0404 is an allele of AB13 .- 36
Seed Fatty Acid Composition of iremutants ...................................................................... 41 .......................................................................................................... Expression of AtEm6 41
1 3-2202 Maps to Chromosome III .............................................. ... ............................... 46 ...................................................... ab& 1 and abi5 , but not abi4 are Epistatic to era 1-3 46
........................................................ ire mutants Suppress era 1 Vegetative Phenotypes 53
abi3. abi4 and abi5 seed epistasis with era 7 ................................................................... 64 era 7 affects al1 ire mutants. abi3 and abiQ ........................................................................... 66 ire mutants are Dominant Suppressors of era 1 .............................................................. 66 Maternaf Effects of ire mutants ............................................................................................ -69 ire suppression of era 1 adult phenotypes ......................................................................... 70
............................................................................................................... abi3 is an ire mutant 71
List of Figures
Figure
Schematic of the suppressor screen of eral
Dose response vs germination frequency of ire mutants
AtEm6 Northern analysis of ire mutants
Position of Iesion of di3-20
M o l e plant phenotypes of ire mutants
Branching patterns of ire mutants
Infiorescence architecture of ire mutants
Drought tolerance of 13-0404 (abi3-20)
List of Tables
Table
Primers used for sequencing of abi.3-20
Resistance of eral-2 and erul-3 suppressors to ABA
F2 segregation analysis of ire mutants crossed to the parent, erul-3
F2 segregation analysis of ire mutants crossed to wild-type, MCol
Cornplernentation analysis of ire mutants
Allelisrn tests of ire mutants with a.i.3, abi4 and d i 5
Fatty acid composition of ire mutants
Mapping of ire 13-2202
Epistatic analysis bet-ween eral-3 and abi3, abi4 and d i 5
Quantitation of branching of ire mutants
Introduction Hormones have been defmed as chernical controilers which are synthesised locally but act
at a distance and whose concentration defrnes the degree of command (Jacobs, 1979). Plant
hormones are therefore not true hormones, since they do not fit these two critena: they may act
over long distances but have also been known to act cell-autonomously and their concentration
does not always correlate to the degree of the effect king controlled. The only unifjhg concept of
plant hormones is that they c m affect physiological processes at extremely low concentrations.
Five classical plant hormones have been defmed by their eEects on plant development or
physiology, for example auxin is involved in apical dominance, mot growth, vascularization,
gravitropism and embryogenesis. This definition has also been misleadhg shce it is becoming
increasingly obvious that there are many more plant hormones than once believed (the cwrent tally
is 9) and that hormones do not entirely control a process but rather just aspects of a process
(Trewavas, 1991). Thus one hormone. in exquisite concert with other hormones, minerais,
metabolites and environmental factors such as Light, temperature and pH control a process. This
multiplicity of control factors is desirable for a plant since it makes for reliablity of a response
which may ofien be irreversible. Therein lies the challenge for plant biologists: how to dissect a
multi-faceted process, which is analogous to deciphering a spider's web.
The traditional approach has been to link hormone concentration with effect. This has been
effective in detexmining the possible sites of plant hormone action, but does not provide clear
answers as to their mode of action. Experiments involving the application of exogenous hormones
have k e n performed in the hopes of recapitulating a process thought to be controiied by the
hormone, but this has not provided any clarification, since there are concems of sequestration,
uptake and hormone sensitivity of different tissues (Bonetta and McCourt. 1998).
Another approach has k e n the genetic approach. This usually involves screening for
mutants which respond abnomally to exogenously applied hormone. This has yielded many
interesting mutants in some hormone signalling pathways, thereby allowing the systematic
dissection of a pathway. Perhaps the most successful example of this approach has been the
genetic dissection of the ethylene response in Arabidopsis. Mutant isolation has Ied to the
identification of numerous signahg components including the ethylene receptor, intermediaie
relays and ethylene responsive transcription factors and has thus provided a clear h e w o r k of the
ethylene response pathway (Woeste and Kieber, 1998). Screening for hormone insensitive mutants
has been successful in the case of ethyiene signailing, but has not been as trïumphant in the case of
other hormones. In these cases one possible problem has ken experimental design. These screens
are usualiy peiformed at concentrations of hormone much higher than endogenous levels, and
assume that signalling is occurring as it would at endogenous concentrations. At these saturating
concentrations the signalling system may very weli respond differently to the hormone. Sensitivity
is defined as the change in physiological response induced by a change in concentration.
Assuming that the response of the process can be saturated like any other dose response, the
maximal sensitivity will occur before saturation, in this case, at Iow concentrations of hormone.
Sensitivity takes into account all factors influencing the process, and is therefore a m e measure of
performance (Trewavas, 199 1)-
With these concerns in mind. new genetic screens need to be developed which more
accurately refelct the in vivo response system. Sensitizing screens for ABA action have
demonstrated that low exogenous hormone concentrations do enrich for different classes of
response mutants than insensitizing screens (Cutler et al-, 1996). This thesis provides a description
of mutants obtained from a screen performed at a low concentration of abscisic acid (ABA), close
to endogenous levels. in order to further elucidate the factors involved in ABA signalling.
Abscisic Acid Signalling ABA is a plant hormone implicated in promoting seed dormancy and development, as weil
as in transpiration and osrnotic stress responses. Genetic analyses have identified either ABA
biosynthetic or response genes and support a role of ABA in the above processes. ABA induced
responses have been categorized into fast (4 mins) and slow (>30 mins) responses, as
exempli fied by seed development and s tomat al closure respec tively . It has been hypo thesised that
there are therefore at least two separate ABA response pathways (Zeevaart and Creelman, 1988).
Physiological Investigations of ABA Action
Fast responses have mainly been studied at the physiological level, since these responses
are not thought to involve de novo gene expression. It is thought that either de novo ABA synthesis
or redistribution of ABA in guard cells leads to stomatai closure (Zeevaart and Creelman_ 1988)-
ABA may work in conjunction with a caZi-dependent pathway to effect stomaîal closure (Man et
ai., 1994).
Slow responses have b e n characterized by searching for genes that are upregulated by
increased concentrations of ABA. Many ABA-inducible genes have a consensus ABA-regulated
(ABRE) cis-acting promoter sequence. The EmBpl transcription factor of wheat binds to this
sequence upstream of the ABA-inducible wheat Eml gene, which is expressed during
embryogenesis (Guiltinan et al.. 1990) and this binding is aided by Vpl, another transcription
factor which mediates ABA-dependent and independent responses (Hill et al., 1996), and a 14-3-3
protein (Schultz et al., 1998). VP 1 was onginally identified in a mutant screen in maize for
embryos that showed a viviparous germination phenotype (Neill et al.. 1986; Robertson, 1955).
VP1 is not ABA-induced, and so the exact link of VPI and Eml to ABA has not been established.
Additionaliy, unlike Em 1. not al1 ABA-inducible genes contain ABRE boxes. suggesting that there
are other factors at play (Bray, 1993).
A large body of work has accumulated linking ABA with responses to environmental
stresses. Much of this work has k e n centred around the relationship of ABA to cold and drought
tolerance. Like Ernl, sorne genes that are drought-inducible are also induced by ABA (Bray, 1993;
Urao. 1996), one class of which are the Late Embryogenesis Abundaot (LEA) proteuis. Because
these are also expressed dunng seed development (Goldberg et al., 1989). one possible role of
these proteins is to protect cellular components during water-deficit conditions (Dure, 1993).
ABA Ieveis also increase during cold acclimation, leading to the hypothesis that ABA may
regulate this process (Zeevaart and Creelman, 1988). Arabidopsis plants treated with ABA are
tolerant of freezing, and many cold responsive genes are induced by ABA (Lang et al., 1989).
Thus, ABA is implicated in these processes, but no direct link between stress induction and ABA
signai transduction has been made.
Genetic Anaiysis of ABA Signalling
ABA-Deficient Mutants
AE3A deficient mutants have been used to understand the role of abscisic acid in several
plant species including tomato and Arabidopsis (Davies, 1995). These mutants have implicated
ABA in water stress relations and seed germination. Mutants defective in the accumulation of
AB A, abal mutants, were isolated as suppresson of a gibbereliin auxouophic mutant, gal
(Koomneef et al., 1982). The abal phenotypes supported roles for ABA in stomatal regulation
and seed dormancy, since mutants were wilty and non-dormant. Severe aileles of abal reduce
rneasurable ABA by 90% compared to wild-type. Recently, mutants in two other genes ABA2 and
ABA3 which were isolated by germination on GA inhibitors, have been shown to be
p henotypically sirniiar to abal mutants (Leon-Kloosterziel et al., 1 996). S tudies with aba2 have
shown this mutant to be defective in the dehydration-hduced accumulation of probe, which is
mediated by increased levels of ABA during drought stress (Nambara et al., 1998).
The abal mutant has k e n useful in detennining the spatial and temporal quirements for
ABA during seed dormancy induction. It was shown that in the developing seed, there are 2 peaks
of ABA concentration, one due to materna1 effects and another embryonic (Karssen et al., 1983).
The embryonic peak only was shown to be critical for dormancy, since neither matemal ABA nor
exogenous ABA could rescue the reduced dormancy phenotype of aba homozygous mutants. In
contrast, dormancy developed in wüd-type embryos even if the matemal tissue was homozygous
for the abal mutation. Matemal ABA may be required for normal seed development.
In tobacco, seed concentrations of ABA were dramatically reduced by the transgenic
expression of an ABA-specific antibody that sequesters free ABA (Philiïps et ai., 1997). These
mutants were phenotypicaiiy similar to abi.3-6 as well as to the relatedfics3 and lecl mutants.
These Iatter two genes appear to play essential roles in cotyledon identity in Arabidopsis. That
immunomodulation of endogenous ABA levels causes novel phenotypes in cornparison to ABA
auxotrophs suggests that aba biosynthetic mutants may be le*. In summary, ABA deficient
mutants have been useful in determining some roles of ABA but have not been able to provide any
clues as to how ABA functions.
ABA Insenstitive Mutants
Abil and Abi2 Mutants
These were isolated as mutants that could germinate on a concentdion of ABA that
inhibited wild-type germination (Koornneef et al., 1984). In addition to the reduced dormancy and
adult wiltiness seen in aba mutants, these dominant mutations cause seeds to be insensitive to
exogenous ABA. Their phenotypes suggested that they are involved in both fast and slow
responses to ABA in that seed domancy, stomatal regdation and ABA induced gene expression
are ail reduced in these mutants. Both of these genes have k e n cioned and shown to encode type
2C serine-threonine phosphatases (Leung et al., 1994; Leung et al.. 1997). Interestingly, ail ABA
insensitive mutant alleies of these two genes are due to the same base pair substitution suggesthg a
very limited range of mechanisms to conferring ABA insensitivity at these two loci. At present,
mutations in AB11 and Ai312 are thought to be dominant negative, since the mutant forms are not
gain-of-function mutations and have low phosphatase activity. AB11 and ABE may have paitialiy
overlapping functions, but are not redundant since ABU controls only a subset of ABIldependent
responses and ABA-dependent morphological and molecuiar responses to drought and cold are
impaired in Abil but not Abi2 (Gilrnour and Thornashow, 1991; Gosti et al., 1995; Vartanian et al.,
1994).
ab3 mutants
abi3-l was found as a mutant insensitive to exogenously appiied ABA at the level of seed
germination (Koornneef et al., 1984). Aside fiom reduced seed donnancy, this allele showed no
other obvious seed or vegetative phenotypes. Subsequently, more severe alleles have been found-
abi3-6, a deletion diele, and the most severe AB13 d e l e to date, was isolated by screening for
insensitivity to the gîbbereUin biosynthetic inhibitor, uniconazol (Nambara et al., 1994). This is an
example of isolahg hormone mutants without having to use large amounts of hormone outside of
the range of high sensitivity to hormone. Like Abil and Abi2 mutants, abi3 have reduced seed
donnancy, but thus far their effects have been limited to the seed.
Unlike Abil, Abi2 and aba mutants, severe abi3 mutants have other seed-specifïc defects
since seeds remain green, do not accumulate certain seed storage reserves, and are severely
dessication intolerant (Narnbara et al., 1994). The relative allelic strengths of ab3 mutants is
reflected b y their dessication intolerance in the series: WT> abi3-l>cr6i3-4>abi3-5>abi3-6
(Ooms et al., 1993). The AB13 pene was isolated by positional cloning, and found to encode a
transcriptionai regulator with homology to VP 1 in maize, which has been studied intensely. Some,
but not ali phenotypes of VP 1 and ABU are the same: both are highly non-dormant aithough obi.3-
6 is not viviparous, and some vpl mutants, unlike abcl-6 also lack anthocyanins in their seed coats
(Carson et al., 1997). There are 4 regions of homology between ABD and VP1, which are the
acidic N-terrninal A 1 domain, and three basic regions designated as B 1, B2 and B3 in order h m
the N-terminus. The B3 region has the highest homology between Vp 1 and AbU and other B3
domain proteins and is composed of 120 amino acids, of which 12 are invariant (Suzuki et
al., 1997). Vpl can activate uanscnption from 2 distinct types of ciselements: Sph-elements iike
that of the C 1 anthocyanin gene of rnaize, and G-boxes Like that of nce and wheat Eml genes. VP1
can also function as both a repressor and an activator, as evidenced by repression of a-amylase
gene in aleurone ceiis, and by the activation of Eml (Hoecker et al., 1995). Vpl is modular in
nature and different domains may be required for the activation of Sph and G-box coupled genes.
Work is stili under way to determine which regions of Vpl bind to which cis-elements, and which
regions of Vp 1 are responsible for repressor and activator fiinctions. Mutations in the B3 domain
block expression of the Sph-coupled Cl gene but do not prevent seed maturation or block the
repressor function of Vp 1 (Hoecker et al,, 1995). The B3 domain by itself can bind to the Sph-
element of C 1. The role of ABA in the action of VPl and AB13 is stiii unclear. AB13 has been
postulated to be a developmental factor which renders cells comptent to respond to ABA or it may
act as a shared component in both ABA signal transduction and seed maturation (Bonetta and
McCourt, 1998).
Epistatic Interactions Between ABA mutants
abal and abi3-I mutants by thernselves are not defective in seed morphology but the
double mutant is green and dessication intolerant like the severe abi3-6. Thus, a lack of ABA
magnifies a defect in ABD. Does ABA have a direct role in seed development?
fus3 was isolated as a non-dormant mutant which retained sensitivity to ABA, showing that
dormancy is mediated by ABA-independent as weli as ABA-dependent factors (Keith et al., 1994).
FUS3 mutants like AB13 mutants are dessication intolerant. However, the fus3 mutation results in
leaf-like CO tyledons, similar to the lecl mutant. FUS3 also encodes a VP 1/ABI3-like B3 -domain
transcription factor (Luerssen et al., 1998).
The genetic relationships between ABA insensitive mutants and FUS3 and LEC1 have k e n
investigated by constructing double mutants (Parcy et al.. 1997). In these studies, Abil fus3 is
more dessication intolerant than abi3 fus3 suggesting that ABII and FUS3 are additive and may
fûnction in different pathways, while AB13 and FUS3 may be involved in similar pathways. This is
supported by studies which argue that FUS3 and LEC1 act upstream of and activate the expression
of AB13 (Parcy et al., 19%).
Double mutants were made between different ABA-insensitive aileles to see whether they
could enhance each other. Enhancement would indicate an additive efflect, while non-enhancement
would indicate that components are in the same pathway (Finkelstein, 1994). These results are not
very concf usive, since such epistatic anaiysis should involve the use of nul1 aileles, and the
phenotypes assayed should be distinct. Interestingly, no combination of weak abi mutants resuIts
in a phenotype iike the severe abi3-6 or the immunomodulation of ABA.
Conflicting results as to whether abi3 and abil are in similar pathways were provided by
constitutively expressing AB13 (Parcy and Giraudat, 1997). These plants were able to accumulate
SSP proteins in response to ABA, but this effect was blocked by the abil mutation, suggesting that
ABIl acts genetically downstream of ABU. However, overexpressing ABU in an AbiI
background restores sensitivity of guard celis to ABA, suggesting that AB13 is downstream of
ABIl. Thus, the epistatic relationship of the two still remains to be clarified.
The eral Supersensitive Mutant
eral was isolated as a mutant that is supersensitive to ABA (Cutler et al., 1996). The
concentration of ABA used is at least tenfold lower than that used in insensitivïty xreens, and is
therefore closer to the sensitive range before saturation by exogenous ABA used in some of the
ABA-insensitive screens. This screen was intrinsicdy sensitive, since the signalling pathway was
selected for a better and not a worse response. The mutant eral seeds are hyperdomant, consistent
with the proposed role of ABA in promoting seed dormancy. They are supersensitive at 0.3 p M
ABA. Guard cells do not open fully and appear to show increased closure sensitivity to applied
AB A (Pei et al., 1998). This vegetative phenotype is also consistent with a proposed role for ABA
as suggested by ABA auxotrophic and abil and abi2 mutants. Unexpectedly with respect to ABA
regulated processes. eral mutants show defects in adult plant development. eral-3 siliques are
curved, apical dominance seerns to be increased, it bolts and grows slowly, it has premanirely
opened flower buds and sometimes fasciates @. Bonetta. pers. Comm). Moreover, cytological
examinations have shown eral mutants have a bigger meristem than wild-type. Some of these
phenotypes are enhanced in short day conditions. suggesting a role of light in these processes. The
ERAl wild-type gene has been shown to encode the B-subunit of a farnesyl transferase. This
enzyme. which has been studied in yeast and animal ceiis, has been shown that as a heterodimer
with the a subunit, ad& a lipophilic 15 carbon-chain to target proteins containing a C-terminal -
CAAX motif, where C is a cysteine, A is an aiiphatic residue, and X can be a variety of amino acids
(Schafer and Rine, 1992). Plant B subunits of farnesyltransferases have an acidic domain not
found in yeast or mammalian counterparts, and cannot substitute for yeast B subunits without the
plant a subunit. This implies that although the function is conserved, the manner in which it is
carried out may differ. Interestingly, in yeast the a subunit is shared with the B subunit of
geranylgeranyl transferases, which aiso transfer a lipophilic group to proteins with a sequence
sirnilar to the -CAAX. It has k e n shown that there may be some redundancy of these two
enzymes, with FTases recognizing GGTase targets, and vice versa. GGTases can also sometimes
transfer farnesyl groups as weU as geranylgeranyl lipid groups (Trueblood et al., 1993).
The ABA-sensitized background of eral mutants provides a useful genetic background to
develop new ABA sensitivity screens. The ABA sensitivity of the eral deletion allele could be
suppressed by second site mutations elsewhere in the genome. Such a screen may not ody
uncover new genes involved in ABA signalling, but may also identify targets of the ERAl faf~lesyl
transferase. This thesis involves the use of suppressor analysis of eral.
Pseudorevertant Analysis Once a mutant has k e n isolated, cloned and phenotypicaliy characte~ed, the next question
is how it interacts within a given developmental pathway. Recombinant DNA methods such as the
yeast 2-hybnd or phage display systems or biochernicd techniques such as affinity columns can
be used to detect protein interactions between a gene of interest and its potential target. The
inevitable problem is that one dways detects false positives with these systems, because they are
inherently arti ficial. The y do not accuratel y reflec t the intricate genetic, phy siologicd and
environment-sensitive balances of a Living plant.
Pseudoreversion is the reversion of a mutant phenotype back to the wild-type and this
genetic analysis is a powerful tool in uncovering a developmental pathway. Weli before the advent
of the aforementioned techniques for detecting interactions between gene products,
pseudoreversion analysis was used to dissect sequentiaiiy acting genes in developmental pathways
in bacteria (Jarvik and Botstein, 1973) and yeast (Moir and Botstein, 1982). Psuedorevertants are
also useful to idenrifv new mutations in known genes as weli as identifjring new genes.
Pseudoreversion readily occurs with many mutants and c m occur in four ways. The first is by
intragenic suppression, where a second mutation in a mutant gene can compensate for the original
mutation, and uius restore wiid-type function. If searching for genetic interactors, intragenic
suppression is undesireable, and so deletion mutants are suppressed in the hope of reversion
occurring in either of the two remaining ways. The second type of suppressors is bypass
reversions, which occur when there is a compensating mutation in a gene involved in a pathway
parailel to the one containhg the suppressed gene. The bypass mutation causes the wild-type
phenotype to be expressed by the activation of another developmental pathway distinct fiom the
original one k i n g studied. Possible molecular rasons for this include the suppressor gene king
highly homologous to the suppressed gene, or that it may be a gain-of- function mutation which
activates a new compensating pathway. Bypass mutations should by their very nature of activating
a parallel pathway, suppress al1 de les (except for dominant gain-of-function aileles) of the
pseudoreverted gene, and is therefore gene-specific and aliele non-specific. The third and most
usefiil type of pseudorevertants are interaction suppressors. These are mutations which occur in
genes coding for proteins that interact directly with the suppressed gene product. These mutations
compensate for the lack of functioning of the suppressed gene by causing conformational changes
causing activation of the devefopmental pathway in question. Thus, interaction mutations are gene-
specific and allele-specific. The fourth type of suppressor is a mutation in a gene downstream of
suppressed gene which has the opposite effect on the pathway as the suppressed gene. For
example, if the suppressed gene when wild-type, acts as a positive reguiator of the pathway, then
suppressing a loss of function allele of it would identify factors that act negatively when wild-type.
Once a suppressor mutant is isolated, it c m then be distinguished as one of these four types of
suppressors using genetic tests: mapping would determine if it were intragenic; if it suppressed ail
aileles of a given gene it would suggest that it was a bypass mutation, and if it did not then it would
be an interaction suppressor or a downstream suppressor.
Once a suppressor of a given gene has been found, the normal function of it needs to be
determined. However, since by defintion its phenotype is wild-type and always relative to the
suppressed gene, it is difficult to isolate it. The solution for this problem in bactena and yeast
(Jarvik and Botstein, 1973) has been to select for suppressors which in addition to the suppression
phenotype also have a phenotype of their own which exists independently of the suppressed gene.
Ln the study of P22 morphogenesis, suppressors of a cold sensitive (CS) mutation were selected,
and some of these were dso found to be temperature sensitive (Ts) (Jarvik and Botstein, 1973).
These Sup/Ts mutants were easily geneticaily analysed by virtue of the recessive Ts phenotype,
which could be used to select for homozygotes and for the isolated suppressor mutations and also
used in complementation tests. In a similar study of yeast ceii division cycle (cdc), suppressors of
a CS mutation were themselves Ts (Jarvik and Botstein, 1973). The Sup/Ts mutants were recessive
for temperature-sensitivity, but dominant for suppression.
Dominant interaction suppressors with a phenotype of their own are intriguing, because
these are probably very specific to the pathway of the gene king suppressed. Dominance of a
suppressor indicates that the flux of the entire pathway c m be changed simply by changing the
amount of the suppressor by 50%. By analogy to the activated Ras screen used by Rubin in
Drosophila (Karim et al., 1996), eral c m be considered a "sensitized" background. Rubin used
an allele of Ras whose expression was constitutive and restrîcted to the eye causing a rough eye
phenotype, which is non-lethal to the fly. Thus, the RadMapK pathway was constitutively active.
In order to isolate components of this pathway that may be essential and therefore Lethal when
homozygous, he reasoned that screening in this constitutively active and sensitized background
should detect critical components of the pathway in which a twofold reduction (ie mutation of one
copy) would alter the signalhg efficiency and thereby visibly modify the rough eye phenotype.
That is, dominant suppressors of a sensitized pathway should reptesent factors cnticd to that
pathw ay .
Therefore, the reasons for searching for pseudorevertants of eral are: 1) To idenw
possible targets of famesyl transferase; 2) To assess whether eral is specific to ABA signalling; if
it is. then suppressing its seed phenotype should also suppress its adult phenotypes; 3) To take
advantage of the Iow concentration of exogenous ABA needed for selection which is closer to
endogenous concentrations; this may select for a new spectnim of mutants involved in maximal
sensing of the hormone.
Materials and Methods
Growth Conditions AU seeds used were in a Meyerowia Columbia (MCol) background. MCol was used as the
wild-type. In experiments involving mutants from other ecotypes, the appropriate wild-type, usually
Landsberg erecta, was used. Seeds were surface sterilized by irnrnersing in 95% ethanol for 15-20
minutes, removing the ethanol and vacuum drying for 10 minutes to remove the ethanol.
Seeds were imbibed on Petri plates containing 0.8% agar supplemented with l.lg/L
Murashige and Skoog (MS) basal culture salts (Sigma Chemicals) buffered with 50mM
morpholinoethanesulphonic acid (MES) (Sigrna Chemicals) pH5.7 and chilied at 4OC for 4 days
to break dormancy. These were then moved to growth shelves at room temperature and illurninated
with approxirnately 200 pE m-2 s- lof iight. Germination was scored d e r 4 days. Abscisic acid
(ABA) was added to the agar at appropriate concentrations from a lOmM stock in methanol, which
was good for 2 weeks.
In cases where selection was not necessary, d e r sterilization seeds were dispersed and
chilied in 1 rnL of 0.2% agar (Sigma Chemicais) in Eppendorf tubes.
Seedlings from plates or seeds in 0.2% agar were transferred to a standard autoclaved soi1
medium containing equal parts of vermiculite. perlite and sphagnum and sanuated with 1gL of a
20-20-20 standard nutrient solution in water.
Plants were usually grown in continuous light. For branching measurements a long day
cycle (16 hour light) was implemented. In di cases illumlliation was at 200 pE m-2 s-1 at 220C
and 50% relative humidity.
Suppressor screen of eral This screen was performed by Dario Bonetta 15 000 erai-2 seeds were mutagenized with
0.25% EMS and 20 000 eral-3 seeds were mutagenized in 0.2% EMS for 16 hours and
immersed in distilled water over the course of 7 hours. Seeds were then chilled at 4 OC for 4 days,
and planted in pools of 30-350 seeds per pot for a total of 27 pools of eral-2 and 30 pools of
eral-3 (Figure 1). M 2 seed was hamested from each pool. 500 seeds per pool were screened on
0.3m ABA agar plates. 202 putative suppressors were picked as germinators afier 2 days and
transferred to soil. M3 seed was hanrested per M2 plant and retested on 0.3 pM ABA by chilling
for 4 days and scoring germination after 2 days. 197 iines retested- Two-thirds of these were
suppressors of eral-3, and the rest were suppressors of eral-2. That there were twofold more
suppressors of eral-3 than eral-2 was due to the higher and thus more toxic concentration of
EMS used to mutagenize eral-2 in addition to the lower number of eral-2 seeds mutagenized.
Determination of Hormone sensitivity ABA was used in Petri plates at appropriate concentrations ranging fkom 0-50pM. Seeds
were steriiized and plated and scored, imbibed and chilled at 4oC for 4 days at which time they
were pIaced on a growth shelf to dlow germination to occur- Germination was scored as the
number of seediings that had green, expanded cotyledons as compared to control plates.
Genetic Analysis In al1 F2 tests, 150-350 seeds were analysed.
Backcrosses to wild-type
Suppressor lines were backcrossed to MCol once. F2 seed was analysed to determine:
1) whether the suppressor mutations were single (standard mendelian ratios)
2) whether they suppressed eral in a dominant or recessive manner, as determhed by germination
ratios on 0.4w ABA. Dominant suppression would be indicated by a ratio of 15: 1 whüe
recessive suppression would be represented by a 13:3 ratio.
3) whether the ABA insensitivity observed in some h e s was dominant or recessive, as determineci
by germination ratios on 3pM ABA. Recessive insensitivity would be indicated by 1:3 ratio, while
dominant insensitivity would show up as a 3: 1 ratio.
Backcross to parent
Suppressor M4 Lines were crossed to eral-2 or era 1-3, depending on which was the parent,
to determine whether it suppressed eral in a dominant or recessive rnanner.
Ailelism Tests
Complemen ta tion
Suppressor M4 lines were reciprocaily crossed to each other to determine whether they were
aiielic to each other. Instead of the usuai testing of FI. lines were tested for alleiism at the F2 stage.
Non-complementation would then be indicated by 100% germination on 0.4pM and 3p.M ABA.
Cross to abi 3, abi4 and ab5
Suppressor M4 iines were crossed to abi3-l. abi4 and abi5 to test for aüelism. Again,
because of the few F1 seed avaiiable, F2 seed were analysed on MS and 3pM ABA plates.
Mapping Suppressor lines were crossed to Ler for SSLP mapping (Beii and Ecker. 1994). Lines 13-
2202 and 13-2903 were mapped by selecting for F2 seeds which germinated on 3pM ABA. DNA
was extracted from these seedlings and subsequently used in PCR reactions with SSLP rnapping
primers. Table1 shows the mapping pnmers used. PCR conditions were 94OC 3 mins for 1 cycle,
followed by 40 cycles at 9 4 ' ~ 30 secs, 54OC 30 secs, 70°C 30 secs and 1 cycle at 7 0 ' ~ for 3 mins.
Table 1
Primers used for sequencing of a6i3-20.
The abi.3-20 gene was divided into 2 regions, each of which were sequenced independently. F' indicates forward prïmers R' indicates reverse pnmers
f Primer Name 1 Direction 1 Seauence (5'.....,..... 3') 1
ire Mutants Fa11 into 3 Complementation Groups ire M5 lines were intercrossed and the F2 seed plated on MS and 3p.M ABA. 100%
germination on 3pM would indicate that the genes were allelic, since it was determined h m the
backcrosses that AB A insensitivity was inherited recessively. There appears to be at least 3 distinct
complementation groups, group 1: 13-0402, 13-0403 and 13-0404, groupII: 13-2202 and group
III: 13-2903 (Table 5). Because 13-0305 was not crossed in aü combinations, it was not assigneci
to a grouping.
13-0404 is an Allele of AB13 Crosses to abi3-1, d i 4 and abi5
From previous double mutant analysis of eralAbil, it was known that eral was epistatic to
Abil (Sarah Cooney, MSc thesis). and therefore Abil was not expected to be isolated as a
suppressor of eral. This was cofimed for the ire mutants by determinhg if any of the mutants
contained the Abil or Abi2 polymorphism. In aii cases, the AB11 and AB12 genes tested wild-type
(data not shown).
Partial alielism tests were performed with 3 known ABA insensitive mutants, abi3-1. abi4
and abi5 by crossing ire mutants with these mutants and observing the germination fkequency on
3p.M ABA. As in the complementation tests, 100% germination would indicate aiielism. Partial
results shown in Table 6 indicate that 13-0404 is allelic to abi3-2, and therefore from the
cornplementation tests, 13-0402, 1 3 - 0 3 and 13-0404 are also new alleles of abi3-l. 13-2202 is
not an allele of ABI3-1, AB14 or AB15 thus defining a new ABA insensitive gene, whilel3-2903 is
not allelic to abi3-l or abi4.
Table 5
Complementation Analysis of ire mutants
ire M4 lines were intercrossed and the F2 seed was assayed for the ability to germhate on 3pM ABA. A germination kquency of 100% indicated non-complementation. ire mutants faii into three distinct complementation groups.
Complementation Analysis
Table 6
Allelism tests of ire mutants with abi3, abi4 and abi.5.
d i 5
d d d d +
d d
ire
13-0305
13-0404 13-2202 13-2903
abi3-1
+ -
d d +
abi4
dd d d + +
Seed Fatty Acid Composition of ire mutants In Adidopsis seeds, storage reserves are in the form of proteins and triacylglycerols.and
synthesis of several of a number of genes involved in storage accumulation have been shown to be
ABA-inducible (Fuikeistein and Somerviile, 1990). Mutations in the ABI3 gene affect the
accumulation of eicosanoic acid (20: 1), the major storage form in seeds. This effect is not seen in
Abil or Ab2 mutant seeds, and therefore a defect in the accretion of 20: 1 is a usehl marker of
ABU action. ire mutant seed fatty acid compositions were determined and shown in Table 7. ab3
mutant levels of 20: 1 are reduced approximately fourfold compared to wild-type levels. A similar
outcorne is seen in 2 suppressor h e s , 13-0402 and 13-0404, which have approximately threefold
less 20: 1 than MCol. Interestingly, eral-3 but not eral-2 has twofold less 20: 1 than MCol. 13-
0403 also has twofold less of eicosanoic acid. This data is in agreement with the complemeatation
and alielism tests which indicate that 13-0402, 13-0403 and 13-0404 are new aüeles of ABI3.
Expression of AtEm6 As with the accumulation of eicosanoic acid, the ABA-induced expression of the late-
embryogenesis abundant protein, AtEm6 is also reduced in ab3 mutants (Fiîelstein, 1994).
However, in contrast to iipid accumulation, this LEA protein is also diminished in abi4 and ubS
mutants (Fielstein, 1994). The simïlar AtEm6 expression patterns in these three mutant
backgrounds has led to the suggestion that ABU, ABM and AB15 are in the same signalhg
pathway. In order to assess whether any ire mutants were involved in this pathway, a seed
Northem was performed using ArEM6 as a probe (Figure 3). Relative to MCol and eral-3 13-
0402, 13-0403, 13-0404 and 13-2903 al1 have less amounts of AtEm6, again supporting
complementation, allelism and fatty acid analyses which suggests that group 13-2903 may be a
mutated in either AB14 or ABIS. whereas l3-û3OS, 13-220 1 or 13-2202, because of their high level
of AtEm6 expression, are probably not defective in either of these genes.
ire mutants were probed with the seed specific marker, AtEM6. The lower panel shows the loading of RNA stained with ethidium brornide.
Sequencing of 13-0404
Fatty acid analysis, seed storage protein analysis and aiielism tests with abi3-l showed that
13-0404 is an ailele of the AB13 gene. In order to determine the exact nature of the mutation, the
abi3 gene of 13-0404 was sequenced. The 13-0404 mutation is due to a single GC to AT
transition at base pair position 3283 in exon 6, causing a missense mutation of an uncharged
glycine residue to the basic arginine. This residue is within the B3 domain (Giraudat et al., 1992),
which is thought to be involved in DNA binding as weii as dimerkation. Figure 4 shows
diagrarnmaticaily where the mutation has occurred in the protein, and shows the salient features of
the AB13 gene and protein. This new allele of AB13 has been titled "ubi3-20 ".
13-2202 Maps to Chromosome III 13-0305, 13-2202 and 13-2903 were crossed to Ler for F2 SSLP rnapping. 13-2903 Pl did
not germinate well, and is currently king propagated. The F2 of 13-2202 crossed to Ler was
rnapped using the markers indicated in Table 8. PrelimuIary results suggest that it is on
chromosome III possibty between g47 11 and AthGAPab. AB13 and AXR2 genes lie in this region,
but 13-2202 does not appear to be a mutation in either of these, as it complements abi3-l and has a
normal response to auxin as meaured by its root sensitivity to awcin as well as its gravitropic
response (results not shown). Further rnapping is necessary.
abi3-1 and abi5, but not abi4 are Epistatic to eraI-3 Crosses of eral-3 to abi.3-1, abi4 and abiJ
Since an allele of abi3 (abi3-20) had k e n found as a suppressor of eral-3, an important question
emerged - was the suppression of eral-3 by di3 aileie speciftc? This would give clues as to the
nature of the interaction between the two genes. The abi3-1 aliele used was originaily in a
Landsberg erecta background, but 1 crossed this allele into to MCol in order to reduce
Figure 4
Position of lesion of abi3-20 which was isolated as a suppressing mutation of eraA.3
DNA
lkb
Homdogyto f i z d basic - Wl
a d i c Nuclear Targeting signal - neuw Gln & Am rich region
Table 8
Mapping of ire 13-2202.
Lines tested were individual F2 progeny of 13-2202 crossed to Ler that were insensitive to 3 p M ABA SSLP mapping primes on chromosome 3 are indicated in italics. The numbers in brackets indicate the position of the primer from the top of chromosome 3 in centimorgans. C indicates a MCol aiiele. L indicates a Ler ailele.
III- NZTI (53 -8- 75.3) CC CC , CC CC CL a CC CC
a a CC a CC CL
1
CC CL CL LL CC CL
1
m - g4711 (53.8)
CC CC cc CC CC cc CC cc
cc CC cc CC a
, CC cc CC
CL a
LL
m- pmcabi3* (38 .O)
CC cc CC
cc
cc cc -IL CC LL CC cc CC CL CL CL
LL CL
m- AthGAPa b (62.7) CC CC cc CC CC cc cc cc a cc CC CC a CC CC CC a CL CL
III- NGAI62 (30.6)
CC CC CL CC CC CC CC
CC
CL LL CC CC CC a ‘IL 'CL - .
13- 2202xLer F2 abi h e #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #X #26 #27 #28 a9 #30 #3 1 #32 #33
III- nga172 (1.1)
CC CC CL
CC a CC
CC
CL LL CC CC CC a
l u #34 ICL #35
Table 9
Epistatic analysis between eral-3 and abi3-IC, ah4 and abi5
a) F2 progeny were plated on 0.4pM ABA on which ABA-supersensitive lines could not genninate,
b) F2 pmgeny were plated on 3pM ABA on which ouly ABA-insensitive Illies could genninate.
Mutant Ratio of Segregation Pattern Germinated:non germinated seeds
abi3-l 145:25 Recessive suppression ( 13:3)
abi4 52:35 Epistasis (9:7)
abS 102: 18 Recessive
Segregation Pattern
Recessive ABA insensitivity Recessive AB A insensitivity Recessive ABA insensi tivity
Mutant
abi3-l
abi4
abS
- Ratio of Germinated:non germinated seeds 30:60
15:111
55: 127
complications using different ecotypes in epistatic analyses with eral-3, and designated it a b 3 4 C.
As shown in Table 9a, eral-3 crossed to abi3-lC F2 seed give a 13:3 gennination:non-
germination ratio, which is characteristic for recessive suppression. The comparable cross with 13-
0404 gives a 15: 1 ratio, characteristic of dominant suppression. Therefore, regardless of which
AB13 allele is used, the results suggest that AB13 acts genetically at or downstream of ERAI.
The F2 of the cross of abS to eral-3 aiso gives a ratio characteristic of that for recessive
suppression, suggesting this gene also acts at or downstream of ERAL.
The F2 of the cross of eral-3 to abi4 indicates that there is an epistatic relationship
between these two mutations. However, unlike abi3-1C and abi5 mutants, abi4 does not
suppress eral -3, on the other hand, eral-3 appears to affect the phenotype of abi4 mutants as the
expected number of germinators for suppression is much reduced, with a concomitant increase in
the number of non-germinators.
To further study the genetic relationships of the above mutations, seeds fiom each F2 cross
were plated on 3pM ABA (Table 9b). Plating on this concentration of ABA assesses whether the
ABA insensitivity of the homozygous abi mutants is affkcted by the presence of erai-3. Only abi4
ABA insensitivity seems to be affécted byeral-3, which corroborates the epistasis result obtained
on 0.4pM ABA.
ire Mutants Suppress eral Vegetative Phenotypes
Morphology and Time to Flowering
Figure 5 shows plants germinated and grown in shon day conditions for 75 days. MCol
has just bolted and some panclades are beginning to elongate. The eral-3 mutant is lagging
behind in growth and has not bolted at this stageThe mutant leaves are flat and slightiy yeiiower
than MCol. AU ire mutants tested are suppressed for lagging growth and indeed are even more
advanced than MCol, as paraclades and rosette cofïorescences have elongated. AU
Figure 5
Whole plant phenotypes of ire mutants
Plants were grown in short day conditions for 75 days. The inset shows 3 eral-3 plants at 100 days old.
suppressor lines are very different in appearance from eral-3 at 100 days (inset, Figure 5), at
which time the wild-type architecture is deveioped enough to compare with the ire mutants. Lines
13-040 1, l 3 W 2 , 13-0403, 13-0404 and 13-2903 have many cauline and rosette leaves. 13-220 1
and 13-2202 are smaller and not very robust in these conditions.
Branching
As shown in Table10, the number of rosette coflorescences and the number of secondary
branches on the main stem of the eral mutunt are dramatically reduced compared to wild-type.
While there is no obvious pattern with the number of secondary branches, the number of rosette
coflorescences is decreased in eral-3 from 4 in MCol to 2 in eral-3. As with other vegetative
phenotypes, these branching defects of eral-3 are suppressed by aü ire mutants examined. Figure
6 represents this pictorially using silhouettes of plants at senescence that were grown in long day
conditions.
Inflorescence Architecture
eral-3 inflorescences have more buds compared to wild-type plants and charactensticaiiy,
many young buds (approximately stage 6- 1 1 according to (Smyth et al., 1990) are not tightly
closed. but instead are slightly open (Figure 7). This phenornenon is suppressed in 13-0305, 13-
0404. 13-2202 and 13-2903. The peneuance of this is variable in 13-0305 and 13-2903, and one
bud that is prematurely opened is seen in 13-2903.
Drought Tolerance
Because ire mutants can suppress many of the vegetative phenotypes of eral mutants, it was
of interest to determine if vegetative physiological responses of eral were also suppressed by ire
mutants. eral piants lose water more slowiy under conditions of drought In part, this reduced
water loss is due to the increased sensitivity of their stomates to closing induced by ABA (Pei et
al.. 1998). One ire mumt abi3-20, was tested for its ability to suppress the drought resistant
phenotype of eral-3. As shown in Fig 8. this mutant surprisingly c m suppress this specifically
vegetative phenotype of eral-3. and indeed seems to iose water siightiy faster than does wiid-type.
This provides more evidence for a rote of AB13 outside of the seed.
Table 10
Quantitation of branching of ire mutants.
Plants were grown in long day conditions and were 5 weeks otd. Results shown are an average of at least 5 plants. See figure 6 for designation of rosette and cauline paraclades.
Figure 6
Branching patterns of ire mutants.
Photocopied siihouettes of iremutants from above at 6 weeks old. RP indicates a rosette pamclade. CP indicates a cauline paraclade.
No. of rosette No. of cauline paraclades 1 1 paraclades
- - - - -
No. of nodes with -1
Figure 7
Inflorescence architecture of ire mutants.
Apical infiorescences were shown of plants gown in constant light.
Figure 8
Drought tolerance assay of 13-0404 (abi3-20)
The rate of water loss of 13-0404 is depicted relative to that of MCol and eral-3 as percentage of soi1 water content vs tbe len,oth of treatment without water.
Discussion
The goal of signal transduction dissection is to undentand the spatial and temporal
interaction of a cascade of gene products which lead to a defined state- A typical signai
transduction pathway consists of 3 components in addition to the Ligand: the receptor, the relays,
and the effectors. These various components allow for many control points and for reversibility of
the signal (McCourt, 1999). There are slight deviations of this such as in bacterial 2-compnent
systems or the marnmalian glucocorticoid receptor, where there are no intermediate relays (Bohen
et al., 1995; Pratt and Silhavy, 1995). Developmental pathways are not simple hear ones in plants
since the "defined state" is a convergence of many signals or ligands. Nonetheless, valuable
information about developmental pathways can be gleaned fiom analysis of the genetic pathways
and the mechanical pathways. The genetic parhway is defined by epistatic interactions and gives no
information on its own as to the molecular nature of the components involved. The mechanical
pathway is defined by interactions on a molecular level, and taken alone gives linle information as
to the placement relative to other components in the pathway. However, in combination these two
sets of information provides a usehl and true picture of a developmental pathway (McCourt,
1999).
abi3, abi4 and abi5 seed epistasis with eral
Epistasis is the masking of one phentoype by another. Strictly speaking, nuU aileles only
should be used. Furthemore, each phenotype should be distinct fiom the other (Avery and
Wasserman, 1992).
abi3-1 and abi.5 are epistatic to eral. This is suggests that they act downstream of erol,
which makes sense intuitively since ABU is a transcription factor. Thus famesyl transferase which
can be thought of as negative relay acts upstream of a VPl- like transcription factor, an effector.
AB15 also acts after ERA1, and it is possible that it acts at the same level as ABD, and is perhaps
dso a transcription factor, although it rnay act before or afier ABI3. These results add credibilty to
the double mutant result between abi3-l and abi5 (Fielstein, 1994), which implies that because
they do not enhance each other's sensitivity, they can be considered in the same pathway.
Therefore taken together, these results suggest that ERA 1, ABU and ABE al1 act in the same
pathway and that Al313 and AB15 act downstream of ERA1.
Surprisingly, eral is dorninandy epistatic to abi4. AB14 is therefore not fùnctioning at the
same level as AB13 and ABIS, as previously proposed (Finkelstein, 1994) but rather functions
upstream of ERA1. This result is intriguing for two reasons. Firstly, AB14 encodes an AP2- like
transcription factor (Finkeistein et al.. 1998). Thus, an effector seems to be acting More a relay,
atypicai of a Iinear signal transduction pathway. This result suggests a 2-tiered ABA response
pathway, the fust containing ABI4, which may transcribe genes necessary for M e r responcihg
to AB A, and the second containing ERA 1 . This picture can be enlarged by cons ide~g data
indicating that eral-3 is also epistatic to abil (Sarah Cooney, MSc Thesis; Pei et al., 1998). Thus,
AB14 and AB11 may act in the first tier. Secondly, the eral-3 effect becomes dominant in the
presence of abi4. If AB14 is indeed acting upstream of ERAI, then attenuating the primary signal
such as with the leaky abi4 alleIe changes the flux through the pathway such that it is more
sensitive to ERA 1 dosage.
Finkelstein (1994) attempts to dissect the AB1 genes into additive pathways based on
whether they enhance each others insensitivity, and suggests that ABD, AB14 and ABE ail act in
the same pathway, and AB11 and AB12 act in a separate pathway. These resufts are inconclusive
because nuII mutants were not used. However, in conjunction with epistatic analysis with eral, a
picture of a genetic pathway unfolds: AM4 and AB11 rnay act before ERAl followed by AB13 and
ABE. The above results are intriguing and must be further investigated using nul1 alleles of ABD,
AB14 and ABIS.
eral affects al1 ire mutants, abi3 and abi4 This is an attractive phenomenon, since it implies a level of interaction that has not been
previously considered. In most of these cases, eral does not completely mask the insensitive
phenotypes, but certainly impinges on their expression. There are numerous scenarios that can
account for this. One possibility is that there is another farnesyl tramferase or that a
geranylgeranyltransferase may be active on ERAl targets. Similar examples of reciprocal
suppression have k e n documented in the suppression of the yeast actin actl mutant (Adams and
Botstein, 1989). Reciprocal suppression is often an indication of interaction. This phenomenon
warrants m e r investigation.
ire mutants are Dominant Suppressors of eral Suppressors restore the phenotype of a mutant to that of wild-type. Moreover, if suppressors
have phenotypes separate from those of suppression then suppressors are epistatic only when their
own phenotype as determined afier they have been isolated frorn the suppressed mutation, ~ l l i t~ks
that of the suppressed phenotype (Botstein and Maurer, 1982).
Four ire mutants 13-0304, 13-0404, 13-2202 and 13-2903 ail dominantly suppressed eral-3
as evidenced from the 3: 1 ratios obtained on 0.3pM AB A of the F2 from crosses to eral-3.
Germination hquencies on 3 p M ABA indicate that these ire mutants are.recessive for ABA
insensitivity, although the deviation €rom a perfect 1:3 ratio implies that the ABA insensitivity of ire
mutants is in turn affected by eral. eral mutants respond to minute arnounts of exogenous ABA
and are 4 tirnes more sensitive to ABA than wild-type. That ire mutants are dominant suppressors
is significant because implies that they are key components in the signalling pathway that includes
eral and of which ABA is the inducer.
Dominant mutations are stereotyped as k i n g gain of function, but in the case of a sensitized
background screen, a dominant mutation could represent a loss of function mutation (Karim et al.,
1996). This is supported by the fact that when these mutations are homozygous, the seeds are
ABA insensitive. Homozygous recessive mutations in AB13 represent loss of function deles, and
it turns out that one ire is a mutation in ABU. This supports the hypothesis that dominant
suppressors of eral are loss of function mutations. In the case of 13-0402, which is semi-
dominant for insensitivity to AB A, and does not complement the abi3 ire (13-0404), chances that it
is a loss-of-function ailele is very small, as this aiiele would have to in addition to king a dominant
ire, also be a dominant for insensitivity. AU known alleles which are dominant for insensitivity to
ABA, Abi 1 and Abi2 are thought to be dominant negative (Fhkelstein, 1994). Therefore, 13-0402
may be a dominant negative allele of ABU. All other ire mutants are most probably 1 0 s of
function alleles, although a conclusive answer can only be provided by disceming the molecdar
nature of these mutations.
The insensitivity to ABA aiso provides more clues on the nahue of the pathways. Strongly
ABA insensitive suppressors c m definitely be placed in the main artery of the response pathway.
which includes ERA1, AB15 and AB13 since not only c m one mutant copy suppress the flux of
the pathway so it is no longer supersensitive, but two mutant copies can completely disable the
pathway. By this criterion of dosage, these factors are essential for transmitting the signal. No such
defuiitive hypotheses c m be drawn from the slightly insensitive or sensitive suppressor mutants
since if they are loss-of-function mutations, then dispensing with hem does not block ABA
signailing but removal of one copy dramatically affects signaihg effciency. These loci may
represent redundant functions or encode genes that affect the signalling efficiency at certain steps
but are not essential for transmitting the signal. Alternatively. they may be leaky mutations in key
factors.
That the ire mutations are suppressors of eral insinuates they affect positive regulaton of
the signalling pathway. Since eral-3 is a deletion de le , ire mutants cannot be intragenic
suppressors. There are three possible molecular scenarios. The fmt is that the ire is a mutation
which causes the deactivation of another distinct pathway, ie a bypass suppressor. This would
mean that there is more than one ABA signalling pathway in the seed. In this scenario, the signal
strength down this other pathway would be more substantial than that of the ERAl-dependent
pathway since an ire mutation in this pathway can not only bypass the sensitizing effect of eral,
but can completely override it. A dominant bypassing ire mutation can stU be envisaged as a Ioss-
of-function mutation in a key factor in the secondary pathway. A simple test of this possibility
would be the demonstration ttiat the bypass suppressor is ailele-non specific and therefore should
suppress al1 alleles of eral. One possible candidate for a bypass suppressor could be in the B-
subunit of a plant geranylgeranyl transferase (GGTase). In yeast, the O-subunits of GGTase and
FTase share a common a-subunit and the dimeric enzymes exhibit cross cross-speciflcity (Seabra
et al., 199 1 ; Trueblood et al., 1993). Therefore. an improved GGTase which c m recognize
fmesylation targets would suppress eral. The target sequences for these enzymes are very
similar: FTase recognizes -CAAX, while GGTase recognizes this or variations of it (Armstrong et
al., 1995). This would have to be a strong gain of function mutation since to cause ABA
insensitivity, it would have to necessarily decrease the signal by geranyigeranylating al l of the
targe t.
The second possibility is that an ire gene product directly interacts with farnesyl transferase.
Famesyl transferases in yeast, Drosophiia and mammalian systems are known to have targets such
as small GTPases of the Ras superfamily, G-protein y-sub units and yeast mating factors (Casey,
1995; Hancock et al., 1989; Leevers et al., 1994; Stokoe et al., 1994). The target could therefore be
similar to these exarnples and act as a positive regulator of ABA signaliing when not farnesylated.
The ire mutation could have caused a loss-of-hinction in this component Alternatively, a single
amino acid change in the target Ras, which is usualiy farnesylated, ailows it to be recognized and
altered by GGTase thereby rescuing a R a s e mutant (Trueblood et al., 1993). By this mechanism,
a target of ERAl may become a suppressor of eral. It is also possible that the suppressor
identifies a farnesylated component involved in the reception of ABA. To date, a large nurnber of
potential candidates for ERAl targets have been identified in siiico by s e a r c h g the Arobidopstr
thalima database for proteins which have a -CAAX box. Interestingly, one of these targets, DNAJ,
which has been shown to be farnesylated in plants, is also used to stabilize the glucocorticoid
receptor signalling pathway in animal ceiis (Fink, 1999; Zhu et al., 1993).
A third possibility is that IRE gene products act downstream of ERA1. Usuaüy suppressors
of a sensitized mutation identiQ factors geneticaiiy downstream of the original mutant protein. For
example, in the activated Ras suppressor screen, mutations in genes that function downstream of
Ras were obtained like Ra€, Mek and MapK (Karim et al., 1996).
Genetic tests are necessary to distinguish between these different pssibilities. Crossing to
different alleles of ERA1 wodd d e t e d n e the aüele specificity of the suppressors- If a suppressor
is allele-specific, then it is a bonafide signaiiing component, and possibly an interaçtor with ERAL.
This is unlikely since eral-3 is a deletion allele: this test is more pertinent to suppression of
missense mutations. Bypass suppressors are aiiele-non specific. Of special interest would be
crossing them to an ERA 1 gain-of-function allele since bonujide suppressors should have the
opposite effect, that is they should enhance this effect. Another test would be to cross to different
mutants that cause supersensitivity. This would pinpoint where in the pathway a suppressor is
acting as weU as defining whether ERA mutants Iie in the same pathway for exarnple, if an ire
mutation suppressed eral and era3, then these could be definitively placed in the same pathway.
Maternal Efiects of ire mutants Maternal effects have k e n documented for three ABA response mutants, Abil, Abi2 and
abi3 (FinkeIstein, 1994). In developing seeds there are 2 peaks of ABA synthesis, an early one at
14 days afier pollination that is matemally denved, and a later one at 16 days after pollination that
is embryonically synthesised and is responsible for the induction of dormancy (Karssen et al.,
1983). The purpose of the matemal peak of ABA remains unknown although it has been
dernonstrated to be necessary for AB13 function (Kmmneef et al., 1983). Putative matemal effects
of ire mutants may be related to perception of this maternai ABA, and reciprocal crosses with
ABA-deficient mutants must be made in order to more fully understand whether these abnormal
ratios are in fact due to matemal effects.
ire suppression of eral adult phenotypes Branches are initiated in leaf axils, and are clonally derived fiom the adaxial side of the subtending
leaf (McComell and Barton, 1998). Axillary buds are formed in a basipetal wave upon the
transition of the shoot apicai meristem (SAM) to an inflorescence rneristem 0 (Hempel and
Feldman, 1994). The initiation of these menstems is thought to be controlled by inhibitory signals
from the S A M , termed apical dominance, together with the distribution of growth substances and
cornpetition for mutrients (Schmitz and Theres, 1999). The ir;voIvement of hormones in lateral bud
formation is implicated by both physiologicai and genetic experiments. Classic experiments
involving the inhibition of secondary bud outgrowth due to the decapitation of the main stem by
the apical addition of auxin was a clear demonstration of the role of auxin as a potentiai negative
regulator of branch development (Thimann and Skoog, 1934). Moreover, mutants resistant to
auxin lack apical dominance (Estelie and Somerville, 1987; Maher and Martindale, 1980; Wilson et
al., 1990). Recently, other factors which rnay or may not be hormone dependent have been
identified in lateral branching, such as the lateral suppressor gene of tomato (Schumacher et al.,
1999) which is defective in a VHIID domain protein, sirnilar to GAI and RGA of Arabidopsis.
eral mutants are phenotypically similar to the revoluta mutant of Arabidopsis, which also
has a reduced number of rosette paraclades (Taiben et al., 1995). The mechanism by which this
reduction occurs is unknown. Because bud outgrowth involves the formation of new cells and is a
photosynthetic sink, possible mechanisms may entail the partitioning of nutrient or growth factors
or differential ce11 cycling due to developmentai cues. In Piswn safivum. he homolog of ERAl is
expressed in growing parts the plant, such as the junctions between stems and leaves, r w t tips and
shoot apices, and is repressed by light and sugar, arguing that expression of this FTase rnay have a
role in nutrient allocation (Zhou et al., 1997). A role for ABA in nutrient allocation is dso
suggested by studies with the prl 1 mutant of Arabidopsis which is hypersensitive not only to
sugars but also to several hormones including ABA (Nemeth et ai., 1998). That the branching
phenotype is suppressed by ire mutants means that they dong with ERAl are involved in bud
formation and outgrowth, and that ABA may be involved in this.
eral have flower buds which are prematurely opened but close again as the bud gets older.
This is due to a differential growth rate of sepals and the rest of the flower, which evenhiaiiy
catches up suggesting that coordinate growth of the flower structure is disrupted in eral mutants.
Further support for uncoordinated growth cornes from the edarged meristem defects seen in eral
grown in short days. Furthemore, the curved silique of eral mutants are due to a disorganized
growth of the normally file-ordered division pattern of the epidermis. Their curved nature means
that growth is occurring on one side differently fiom the other. In al1 of these cases, ire mutants
suppress the cell growth defects.
ire mutants suppress a range of eral phenotypes. Defects in eral aie many in number (D.
Bonetta, unpubl). As in many hormone mutants, it is diff~cult to pinpoint what molecular process is
causing the defect. eral farnesylates proteins, but what is the nature of the target and what
processes are they involved in? Evidence is accumulating that farnesylation is required for normal
cell cycling (Du et al., 1999). The common element of some of the vegetative phenotypes of eral
mutants is that they are defective in processes that involve cellcycling. ire mutants suppress al l of
these phenotypes suggesting that ABA may influence the coordination of cell cycling via ERA1.
That revertants suppress these varied phenotypes of eral argues that ERA1 is specific to ABA
signalling, and that AE3A plays an inhibitory role in celi-cyciing consistent with earlier studies on
ABA (Zeevaart and Creehan, 1988).
abi3 is an ire mutant That three ire mutations are in the AB13 transcription factor reveals two intriguing
phenomena. The fmt is that 2 different point mutations in the 8 3 domain can suppress eral in
different ways, one dominantly and one recessively.The second is that ABU which is supposedly
seed-specific, when mutated affects adult structures.
The obi3 -1 mutation causes a substitution of an aspanate residue for an asparagine at
amino acid position 580 (Giraudat et al., 1992). obi3-20 reported here is a nonsonservative
substitution of a glycine to an arginine at position 669. There is evidence that conservative
substitutions do not cause distortion of protein 3-dimensional structures, a postdate used as the
basis for alanine scanning mutagenesis in yeast (Wertman et ai., 1992). Therefore that abi3-20 can
suppress eral-3 dominantly may reflect the severity of a non-conservative amino acid substitution.
It suggests that specific amino acids and not just the entire region are important for functioning.
This is supported by the quantitative range of phenotypes of different aiieles of ABU: abi3-3, a
severe EMS allele, genninates more quickly on 2@l ABA than ubi.3-I (Nambara et al., 1992).
abi.3-3, abi3 -4 (G417-*stop), abi3-3 and &i3-6 (O.75kb deletion) are a i i more insensitive to ABA
than is abi3-I (Nambara et al., 1994; Ooms et al., 1993). Also, it is known that alleles of AB13
contribute to a pathway in a dosage sensitive mamer since abi3-6 is dorninantly insensitive to
uniconazol, an inhibitor of GA synthesis, but recessively insensitive to ABA (Nambara et al.,
1994). That abi3-20 suppresses eral -3 dorninantly suggests that it is a more severe d e l e than
abi3- 1.
The abi3-20 mutation is located in the B3 domain of the AB13 gene. which is thought to bind
cooperatively to Sph DNA elements (Suzuki et al., 1997). In VP1, defects in the B3 domain Iead to
AB A-independent abnormalities, sugges ting that other parts of this transcription factor are
necessary for mediating interactions with ABA responsive genes such as Eml, and implying that
the B3 domain mediates developmental processes (Carson et al-, 1997). This is consistent with
ABU king involved in mediating developmental States as an instructive factor (Bonetta and
McCourt, 1998).
Secondly, abi3-20 affects not only seed phenotypes but also vegetative phenotypes of eral.
This is surprising since no definitive rote of ABU outside of the seed has been shown (Giraudat et
al., 1992; Parcy et al., 1994). When ectopicaiiy expressed, AB13 c m affect vegetative phenotypes
(Parcy and Giraudat, 1997), but this misexpression was under artificial conditions and is most
likely not relevant to the in planta situation. However, this does indicate that ABU is sufficient for
ABA signalling. Two possible ways in which ABU can be interacting with ERAl are: 1) ABU is
expressed outside of the seed, but to very low levels that are difficult to detect; 2) Epigenetic effects
are occumng in the seed, which are "remembered" by the aduit plant.
If ABD is expressed outside of the seed then it is involved in branching, since it suppresses
the reduced paraclade phenotype of eral-3. Perhaps it affects the celi cycle or nutrient aliocation,
analogous to its role in the seed, where it is thought to change the cornpetence of ceils to respond
to ABA. Evidence that ABA is works by Iimiting the availablity of energy and nutrients is
accumulating (Garciarrubio et al., 1997) and AB13 may be involved in mediating the distribution of
resources.
The results imply that in eral-3, ABU rnay be functioning to d u c e branching. One way of
testing this hypothesis would be to check whether there is ectopic expression of AB13 in eral-3
plants. This would also imply that ERA1 rnay inhibit, aibeit indirectly, ABI3, and this rnay account
for the inability to detect AB13 since ERA L is expressed in adult tissue. .
Epigenetic effects in which signals occurring in the seed affect vegetative phenotypes have
been documented. Length of seed chilling affects the tirne to flowering of the adult plant (Sheldon
et al., 1999). The mechanism by which this occurs is not clear, but may involve DNA methylaîion
(Fîiegan et al., 1998). Protein modifications such as the acetylation of histones as weil as
chromatin silencing also confer epigenetic information (Gmnstein, 1998; Photta, 1998). More
recently, there have been reports that plants can "leam" by s t o ~ g information in signailing
pathways (Kudla et ai., 1999; Trewavas, 1999).
That abU-20 can suppress eral-3 drought avoidance is intriguing. eral-3 is drought
tolerant, as measured by its ability to survive weii past wild-type plants in drought conditions.
There are several possible strategies that c m be used by a plant to avoid drought stress. Faster and
longer guard ce11 closure, a "fast" response, is one possibility, and this is one strategy adopted by
eral-3 plants (Pei et ai., 1997). However, there are also slow responses to drought stress, which
involve de novo gene transcription, such as the production of protective osmolytes, the initiation of
root gmwth and the inhibition of drought-induced senescence. The drought tolerance of eral rnay
not simply be due to its guard ce11 phenosrpe, but is most probably also due to changes in these
other "slow" responses since e ra l affects the expression of senescence induced genes (SAG - McCourt unpubl. results) which are expressed in drought stressed plants. Therefore, abU-20 may
not necessarily be intedering with stomatal closure to effect suppression of the drought tolerant
phenotype of eral-3, but rnay aECect the implementation of the aforementioned alternative
strategies. Evidence supporting this hypothesis comes from work with LEA proteins which are
expressed during seed ernbryogenesis, but are simïlar to proteins expressed during drought stress
in whole plants, and are thought to act as dessication osmoprotectants (Dure, 1993). The
expression of severai LEA proteins is disrupted or abolished in AB13 mutants (Nambara et al.,
1995; Parcy et al., 1994). Additionaiiy, LEA proteins c m be found in leaves of plants ectopically
expressing ABU that have been ABA-treated, suggesting that expression of LEA proteins are
controlled by the presence of AB13 (Parcy et al.. 1994). Also, the ATMYB2 gene is drought
inducible, but the expression of MYB homologues are disrupted in abL3-4 (Kink et al., 1998; Urao
et al., 1996). Whatever the mechanism of suppression proves to be, this is the first clear
demonstration that AB13 has genetic effects on vegetative phenotypes.
The future of this work lies in deciphering the exact nature of the relationship between AB13
and ERAI. Many questions have been raised by this work: 1) 1s ABU truly seed-specific? in situ
hybridization studies wil3 help in answering this, since aithou@ ABU-GUS lines were observed,
AB13 may be expressed at very low levels undetectable by thîs rnethod or at very specific locations
not previously assayed (Parcy et al., 1994). ABZ3-GUS Lines that have k e n crossed into eral -3
are currently k i n g observed for blue staining, since it is hypothesised that in the eral mutant,
AB13 may be acting to inhibit paraclade formation in long day conditions.
2) If ABD is seed specific, how does it affect erd-3 adult phenotypes? This question is
intriguing. AB13 targets must be identified.
Additionaily, mapping and characterization of other ire mutants as weil as non-insensitive
revertants will provide more information on ABA signalling and the role of farnesylation. Indeed,
among this latter set, there are probably mutants in targets of farnesylation, which may now either
be gain of function mutants or mutants that c m now be efficientiy gerany lgeranylated. The recent
isolation of an eral mutant in the Ler background (Nocha Van Thielen, MSc) has already been
useful in mapping of these lines which have no obvious phenotype of their own but which can be
selected for on the basis of their eral suppression. More investigation into the maternai effects of
ire mutants and into the reciprocal effects between erul and ire mutants is warranted.
The work presented here represents a new and surprising perspective on ABU plant
development, since it was previously thought to act seed specincally. The relationship beween
ERAl and other ABA mutants in ABA signalling has been investigated and c l 6 e d . Fuaher
analysis of the collection of eral revenants is sure to provide additional information and may
ultimately lead to the elucidaiion of ABA signalling pathways-
References Cited
Adams, A.E.M. and Botstein, D. (1989). dominant suppressors of yeast actin mutations that are reciprocally suppressed. Genetics 121, 675-683.
Man, A. C., Fricker, M. DI, Ward, 3. L., Beale, M. H., and Trewavas, A- J. (1994). Two transduction pathways mediate râpid effects of abscisic acid in Commelina guard cells. Plant Cell 6, 1319-1328.
Armstrong, S. A., Hannah, V. C., Goldstein, J. L., and Brown, M. S. (1995). CAAX geranylgeranyl uansferase transfers faniesyl as eficiently as geranylgeranyl to RhoB. J Bi01 Chem 270,7864-8.
Avery, L., and Wasserman, S. (1992). Ordering gene function; hte interpretation ofepistasis in regdatory hierarchies. Trends Genet, 8 , 3 12-3 16.
Beli, C. I., and Ecker, J. R (1994). Assignment of 30 microsateliite loci to the linkage map of P~abidopsis. Genomics 19, 137- 144.
Bohen. S. P., Kraiii, A., and Yamamoto, K. R. ( 1995). Hold lem and fold 'em: chaperones and signal transduction, Science 268, 1303- 1304.
Bonetta, D., and McCoun, P. (1998). Genetic Analysis of ABA signal trasduction pathways. TIPS 3,23 1-235.
Botstein, D., and Maurer, R. (1982). Genetic approaches to the analysis of microbial development. h u Rev Genet 16,6 1-83.
Bray, E. A. (1993). Molecular responses to water deficit. Plant Physiol. 103, 1035-1040.
Carson, C. B., Hattori, T., Rosenkrans, L., Vasil, V., Vasil, 1. K, Petenon, P. A., and Mc-, D. R. (1997). The quiescent/colorless alleles of viviparous 1 show that the conserved B3 domain of VPl is not essential for ABA-regulated gene expression in the seed. Plant J 12, 123 1-40.
Casey, P. J. (1995). Mechanisms of protein prenylation and role in G protein function. Biochem Soc Tram 23, 16 1-6.
Cutler, S., Ghassemian, M., Bonetta, D., Cooney. S.. and McCourt, P. (1996). A protein farnesyl transferase involved in abscisic acid signal transduction in Arabidopsis. Science 273, 123941.
Davies, P. J. (1995). Plant Hormones: Their nature, occurrence and functions., P. J. Davies, ed. (Dordrecht: Kluwer).
Du, W., Lebowitz, P.F. and Prendergast, G.C. (1999). Cell growth inhibition by famesyl transferase inhibitors is mediated by a gain-of-function geranylgeranylated RhoB. Mol- Ceil. Biol. 19, 183 1-40
Dure, L. (1993). A repeating 1 1-mer amino acid motif and plant desiccation. Plant J 3,363-369.
Estelle, M., and Somerville, C. R. (1987). Auxin-resistant mutants of Arabidopsis thdiana with an altered morphology . Molec.Gen.Genet- 206,200-2%-
Fink, A. L. (1999). Chaperone-mediated protein folding. Physiol. Rev. 79,425429.
Fielstein, R. (1994). Matemal effects govern variable dominance of two abscisic acid response mutations in Arabidopsis thaiiana. Plant Physiol. 105, 1203- 1208.
Finkelstein, R. (1994). Mutations at two new Arabidopsis response loci are similar to the abi3 mutations. Plant J. 5,765-771
Finkelstein, R., and Somerville, C. R. (1990). Three classes of abscisic acid (ABA) insemitive mutations of Arabidopsis define gens that control overlapping subsets of ABA responses. Plant Physiol- 94, 1 172- 1 179.
Fielstein, R. R., Wang, M. L., Lynch, T. J., Rao, S., and Goodman, H. M. (1998). The Arabidopsis abscisic acid response locus AB14 encodes an APETALA 2 domain protein. Plant Ce11 10, 1043-54.
Fumegan, E. J., Genger, R. K., Kovac, K. Peacock, W. J., and Demis, E. S. (1998). DNA rnethylation and the promotion of flowering by vernalization. Proc. Natl Acad. Sci. U S A 95, 5824-5829.
Garcianubio, A., Legaria, J. P., and Covarnibias, A. A. (1997). Abscisic acid inhi'bits germination of manire Arabidopsis seeds by limiting the avaiiability of energy and nutrients. Planta 203, 182-7.
Gilmour, S. J., and Thornashow, M. F. ( 199 1). Cold acclimation and cold-regulated gene expression in ABA mutants of Arabidopsis thaliana. Plant Mol Bi01 17, 1233-40.
Giraudat, J., Hauge, B. M., Valon, C., Smalle, J., Parcy, F., and Goodman, H. M. (1992). Isolation of the Arabidopsis AB13 gene by positional cloning. Plant Ceil 4, 125 1-6 1.
Goldberg, R., Barker, S. J., and Perez-Grau, L. (1989). Regulation of gene expression during plant embryogenesis. Ce1 56, 149- 160.
Gosti, F., Bertauche, N., Vartanian, N., and Giraudat, J. (1995). Abscisic aciddependent and - independent regulation of gene expression by progressive drought in Arabidopsis thaliana Mol Gen Genet 246, 10-8.
Grunstein, M. (1998). Yeast heterochromatin: regulation of its assembly and inheritance by histones. Ce11 93, 325-328-
Guiltinan, M. J., Marcotte, W. R., Jr-. and Quatrano, R. S. (1990). A plant leucine zipper protein that recognizes an abscisic acid response element- Science 250,267-7 1.
Hancock, J. F., Magee. A. 1.. Childs, J. E., and Marshall, C. J. (1989). Al1 ras proteins are polyisoprenylated but only some are palmitoylated. Celi 57, 1167-77.
Hempel, F. D., and Feldman, L. J. (1994)- Bi-directional inflorescence deveio~ment in rabi id op sis thalima Acropetd initiation of flowers and basipetal initiation of paraclades. Planta 192,276-286.
Hili, A., Nantel, A., Rock, C. D., and Quatrano, R. S. (1996). A conserved domain of the viviparous-1 gene product enhances the DNA binding activity of the bZIP protein EmBP-1 and other transcription factors. J Bi01 Chem 271,3366-74.
Hoecker, U., Vasil, 1. K. and McCarty. D. R- (1995). Integrated control of seed maturation and germination programs by activator and repressor functions of Viviparous-1 of rnaize. Genes Dev 9, 2459-69.
Jacobs, W. P. (1979). Plant hormones and plant development (New York: Cambridge University Press).
Jarvik, J., and Botstein, D. (1973). A genetic method for detennining the order of events in a biological pathway. Proc Nati Acad Sci U S A 70,2046-50.
Karim, F. D., Chang, H. C., Themen, M.. Wassarman, D. A., Laverty, T., and Rubin, G. M. (1996). A screen for genes that hinction downstream of Ras 1 dunng Drosophila eye development. Genetics 143-3 15-29.
Karssen, C. M., Brinkhoerst-van der Swan. D. L. C., Breeklmd, A. E., and Koomneef, M. (1983). Induction of donnancy during seed deveiopment by abscisic acid: studies on abscisic acid deficient genotypes of Arabidopsis thaliana. Planta 157, 158- 165.
Keith, K, Kraxni, M., Dengler, N., and McCourt. P. (1994). fusca3: A heterochronic mutation affectïng late embryo development in Arabidopsis. Plant Cell6,589-600.
Kirik, V., Kolle, K., Wohlfarth. T., Misera, S.. and Baumiein, H. (1998). Ectopic expression of a novel MYB gene modifies the architecture of the Arabidopsis inflorescence. Plant J 13,729-42.
Koornneef, M., Jorna, M. L., Brinkhoerst-van der Swan, D. L- C., and Karssen, C. M. (1982). The isolation of abscisic acid (ABA) deficient mutants by selection of induced revrtants in non- germination gibberellin sensitive lines of Arabidopsis. Theor. Appl. Genet. 61,385-393.
Koomneef, M., Reuling, G., and Karssen, C. M. (1984). The isolation andcharacterization of abscisic acid insensitive mutants in Arabisopsis thaliana Physiol. Plant. 61,377-383.
Kudla, J., Xu, Q., Harter, K., Gntissern, W., and Luan, S. (1999). Genes for calcineurin B-like proteins in Arabidopsis are differentiaily regulated by stress signais. Proc Natl Acad Sci U S A 96, 47 1 8-4723 -
Lang, V., Heino, P., and Palva. E. T. ( 1989). Low temperature acchmation and treatment with exogenous abscisic acid induce cornmon polypeptides in Arabidopsis thaiiana (L.) Heynh. Theor. Appl. Genet. 77,729-734.
Laufs, P., Grandjean, O.. Jonak, C., Kieu, K., and Traas, J. (1998). Cellular parameters of the shoot apical meristem in Arabidopsis. Plant Ce11 IO, 1375- 1389.
Leevers, S. J., Paterson, H. F., and Marshall, C . 1. (1994). Requirement for Ras in Raf activation is overcome by targeûng Raf to the plasma membrane. Nature 369-4 1 1 4 14.
Leon-Kloosterziel, K. M., Gil, M. A.' Ruijs, G. I., Jacobsen. S. E., Olszewski, N. E., Schwartz, S. H., Zeevaart, J. A., and Koomneef, M. ( 1996). Isolation and characterization of abscisic acid- deficient Arabidopsis mutants at two new loci. Plant J 10,655-61.
Leung, J., Bouvier-Durand, M., Moms, P. C., Guemer, D., Chefdor, F., and Giraudat, J. (1994). Arabidopsis ABA response gene ABII: features of a calcium-modulated protein phosphatase. Science 264, 1448-52.
Leung. J., Merlot, S ., and Giraudat, J. (1 997). The Arabidopsis ABSCISIC ACID- INSENSI"ïrvE2 (ABD) and AB1 I genes encode homoIogous protein phosphatases 2C involved in abscisic acid signal transduction. Plant Cell9,759-7 1.
Lincoln, C., Bntton, J. H., and Estelle, M. (1990). Growth and development of the axrl mutants of Arabidopsis. Plant Ce11 2, 107 1 - 1080-
Luerssen, H., Kink, V., Herrmann, P., and Misera, S. (1998). FUSCA3 encodes a protein with a conserved VPl/AB 13-like B3 domain which is of functional importance for the regulation of seed maturation in Arabidopsis thaliana PImt J. 15,755-764.
Maher, E. P., and Martindale, S. J. B. (1980). Mutants of Arabidopsis thaliana with altered responses to auxins and gravity. Biochem. Genet. 18, 1041-1053.
McCo~eI1, J. R., and Barton, M. K. ( L998). Leaf polarity and meristem formation in Arabidopsis. Development 125,2935-2942.
McCourt, P. (1999). Genetic Anidysis of Hormone signaling. Annu. Rev:Ptant Physiol. Plant Mol. Biol. 50,219-243.
Moir, D., and Botstein, D. (1982). Determination of the order of gene function in the yeast nuclear division pathway using CS and ts mutants. Genetics 100,565-77.
Nambara, E., Kawaide, H., Kamiya, Y., and Naito, S. (1998). Characterization of an Arabidopsis thaliana mutant that has a defixt in ABA accumulation: ABA-ciependent and ABA-independent accumulation of fixe arnino acids during dehydration. Plant Cell Physiol39,853-8.
Namban, E., Keith, K., iMçCourt, P., and Naito, S. (1994). Isolation of an intemal deletion mutant of the Arabidopsis AB13 gene. Plant Celi Physiol. 35,509-5 13.
Nambara, E., Keith, K., McCourt, P., and Naito, S. (1995). A regdatory role for the AB13 gene in the establishment of embryo maturation in Arabidopsis thaliana. Development 121,629-636.
Nambara, E., Naito, S., and McCourt, P. (1992). A mutant of Arabidopsis which is defective in seed development and storage protein accumulation is a new abi3 aiiele. Plant J. 2,435-441.
Neill, S. J., Horgan, R., and Parry, A. D. (1986). The carotenoid and abscisic acid content of vivparous kemels and seedlings of Zea mays. Planta 169,87-96.
Nemeth, K., Salchert, K., Pu tnoky, P., Bhalerao, R., Koncz-Kalman, Z., S tankovic-S tangeland, B., Bako, L., Mathur, J., Okresz, L., Stabel, S., Geigenberger, P., Stitt, M., Redei, G. P., Schell, J., and Koncz, C. (1998). Pleiotropic control of glucose and hormone responses by PRL1, a nuclear WD protein, in Arabidopsis. Genes Dev 12,3059-73.
Ooms, J. J. J., Leon-Kloosterziel, K. M., Bartek, D., Koornneef, M., and Karssen, M. (1993). Acquisition of desiccation tolerance and longevity in seeds of Arabidopsis thdiana A comparative study using abi3 mutants. Plant Physioi. 102, 1185- 1 191.
Parcy , F., and Giraudat, J. (1997). Interactions between the AB1 1 and the ectopically expressed AB13 gene in controlling abscisic acid responses in Arabidopsis vegetative tissues. Plant J 11,693- 702.
Parcy, F., Valon, C., Kohara, A., Misera, S., and Giraudat, J. (1997). The ABSCISIC ACID- INSENSITIVE3. FUSCA3, and LEAFY COTYLEDON1 loci act in concert to control multiple aspects of Arabidopsis seed development, Plant Cell9, 1265-77.
Parcy, F., Valon, C., Raynai, M., Gaubier-Cornella, P., Delseny, M., and Giraudat, J. (1994). Regulation of gene expression programs dunng Arabidopsis seed development: roles of the AB13 locus and of endogenous abscisic acid, Plant Cell6, 1567-82.
Pei, 2. M., Ghassemian, M-, Kwak, C. M-, McCourt, P., and Schroeder, J. 1. (1998). Role of farnesyltransferase in ABA regulation of guard cell anion channels and plant water loss. Science 282,287-90.
Pei, Z. M., Kuchitsu, Ka, Ward, J. M., Schwarz, M., and Schroeder, J. 1. (1997). Differential abscisic acid regulation of guard ce11 slow anion channels in Arabidopsis wild-type and abil and abi2 mutants. Plant Cell9,409-23.
Phillips, J., Artsaenko, O., Fiedler, U., Horstmann, C., Mock, H- P., Muntz, KI, and Conrad, U. ( 1997). Seed-specific irnrnunornodulation of abscisic acid activity induces a developmentai switch. EMS0 J. 16,4489-4496.
Pirrotta, V. (1998). Polycombing the genome: PcG, trxG, and chromatin silencing. Cell93,333- 336.
Pratt, L. A., and Silhavy, T. J. (1995). Two-component signal transduction., J. A. Hoch and T. J. Silhavy, eds.: ASM Press).
Robertson, D. S. (1955). Genetic vivipary in rnaize. Genetics 40,745-760.
Roman, G., Lubarsky, B., Kieber, J. J., Rothenberg , M., and Ecker, J. R. (1995). Genetic analysis of ethylene signal transduction in Arabidopsis thaliana: five novel mutant loci integrated into a stress response pathway, Genetics 139, 1393-1409.
Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular cloning : a laboratory manual (Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory).
Schafer, W.R. and Rine, J. (1992). Protein prenylation: genes, enzymes, targets, and functions. Annu.Rev.Genet. 30,209-237,
Schmitz, G., and Theres, K. (1999). Genetic control of branching in Arabidopsis and tomato. Curr. Op. Plant Biol. 2,51-55,
Schultz, T. F., Medina, J., Hill, A., and Quatrano, R. S. (1998). 14-3-3 proteins are part of an abscisic a c i d - W A R O U S 1 (VP 1 ) response complex in the Em promoter and interact with VPI and EmBP 1. Plant Ce11 10,837-47.
Schurnacher, K., Schmitt, T.. Rossberg, M., Schmitz, G., and Theres, K. (1999). The lateral suppressor gene of tomato encodes a new member of the VHIID protein family. Proc. Natl. Acad. Sci. USA 96,290-295.
Seabra, M. C., Reiss, Y., Casey, P. J., Brown, M. S., and Goldstein, J. L. (1991). Protein farnesyltransferase and geranylgeranyltransfense share a common alpha subunit. Cell65,429-34.
Sheldon, C. C., Bum, J. E., Ferez, P. P., Metzger, J., Edwards, J- A., Peacock. W. J., and Dennis, E. S. (1999). The FLF MADS box geoe: a repressor of flowenng in Arabidopsis regulated by vemalization and methylation. Plant Cell 11.445-458.
Smyth, D. R., Bowman, J. L., and Meyerowitz, E. M. (1990). Early flower development in Arabidopsis. Plant Ce11 2,755-767.
Stokoe, D., Macdonald, S. G., Cadwdader, IC, Symons, M., and Hancock, J. F. (1994). Activation of Raf as a result of recruitment to the plasma membrane [see comments] [published erratum appears in Science 1994 Dec 16;266(5 192): 1792-31. Science 264, 1463-7.
Suzuki, M., Kao, C. Y., and McCarty, D. R. (1997)- The conserved B3 domain of VIVIPAROUS 1 has a cooperative DNA binding activity. Plant Celi 9,799-801.
Talbert, P. B., Adler, H. T., Paris, D. W., and Cornai, L. (1995). The REVOLUTA gene is necessary for apical meristem development and for limiting ceH divisions in the leaves and stems of Arabidopsis thdiana Developrnent 121,2723-2735.
Thirnann. K V., and Skoog, F. ( 1934). On the inhibition of bud development and other functions of growth substance in Vicia faba Proc. Roy. Soc- B- I14 ,3 17-339-
Trewavas, A. (199 1). Plant ceU signal transduction.Trends Genet- 7,356-61
Trewavas, A. (1999). How plants learn. Proc Nat1 Acad Sci U S A 96,4216-4218.
Trueblood, C. E., Ohya, Y.. and Rine, J. (1993). Genetic evidence for in vivo cross-specincity of the CaaX-box protein prenyltransferases famesyltransferase and geranylgeranyltransfera~e-I in Saccharomyces cerevisiae. Mol Cell Bi01 13,4260-75.
Urao, T., Yamaguchi-Shinozaki, K. Uno, S., and Shinozaki, K. (1 996). An Arabidopsis myb homolog is induced by dehydration stress and its gene product binds to the conserved MYB recognition sequence. Plant Cell 5. 1529-39.
Vartanian, N., Marcotte, L., and Giraudat, J. (1994). Drought rhizogenesis in Arabidopsis thaliana Differential responses of hormonal mutants. Plant physiology. 104,761-767-
Wertman, K F., Drubin, D. G., and Botsrein, D. (1992). Systematic mutational analysis of the yeast ACT1 gene. Genetics 132,337-50.
Wilson, A. K., Pickett, F. B., Turner, J. C., and Estelle, M. (1990). A dominant mutation in Arabidopsis confers resistance to auxin, ethylene and abscisic acid. Mol Gen Genet 222,377-83.
Woeste, K., and Kieber, J. J. (1998). The molecular b a i s of ethylene signalling in Arabidopsis. Philos Trans R Soc Lond B Bi01 Sci 353, 143 1-8.
Zeevaart, J. A. D., and Creelrnan, R. A. (1988). Metabolism and physiology of abscisic acid. AMU. Rev. Plant Physiol. Plant Mol. Biol. 45. 1 13- 14 1.
Zhou, D., Qian, D., Cramer, C. L., and Yang, 2. ( 1997). Developmentai and environmeoral regulation of tissue- and cell-specific expression for a pea protein famesylûansferase gene in transgenic plants. Plant J 12.92 1-30.
Zhu, J. K., Bressan, R- A., and Hasegawa, P. M. (1993). Isoprenvlation of the lan nt molecular chaperone ANJl facilitates membrane association and &don & high tempe&ture. Roc Nat1 Acad Sci U S A 90,8557-6 1.