-
The Arabidopsis Book ©2002 American Society of Plant
Biologists
As sessile organisms, plants are unable to move activelytowards
favorable or away from unfavorable environmentalconditions.
Therefore, through their evolution, plants haveadapted a high
degree of developmental plasticity to opti-mize their growth and
reproduction in response to theirambient environments. Light is one
of the major environ-mental signals that influence plant growth and
develop-ment. Not only is light the primary energy source for
plants,but it also provides them with positional information
tomodulate their developmental processes such as seedgermination,
seedling de-etiolation, gravitropism and pho-totropism, chloroplast
movement, shade avoidance, circa-dian rhythms, and flowering time.
Plants can detect almostall facets of light such as direction,
duration, quantity, andwavelength by using three major classes of
photorecep-tors: the red (R)/far-red (FR) light (600-750 nm)
absorbingphytochromes (phys), the blue (B)/UV-A (320-500
nm)absorbing cryptochromes (crys) and phototropins (phots),and the
UV-B (282-320 nm) sensing UV-B receptors(Kendrick and Kronenberg,
1994; Briggs and Olney, 2001;Briggs et al., 2001). These
photoreceptors perceive, inter-pret, and transduce light signals,
via distinct intracellularsignaling pathways, to photoresponsive
nuclear genes,which modulates plant growth and development.
The phenotypic changes associated with the
seedlingphotomorphogenic development are among the most dra-matic
events mediated by these photoreceptors. Seedlings
grown in the dark undergo skotomorphogenesis (etiolation)and are
characterized by long hypocotyls, closed cotyle-dons and apical
hooks, and development of the proplastidsinto etioplasts.
Light-grown seedlings undergo photomor-phogenesis (de-etiolation)
and are characterized by shorthypocotyls, open and expanded
cotyledons, and develop-ment of the proplastids into green mature
chloroplasts (thusa process considered “de-etiolation” of the
etioplasts,McNellis and Deng, 1995, Figure 1). The past decade
hasseen dramatic advances in our knowledge of plant pho-toreceptors
and in our understanding of their signal trans-duction pathways
that lead to various physiologicalresponses. Here, we briefly
review the most recentprogress that has provided new insights into
constructingan emerging integrated picture of phytochrome signaling
inArabidopsis, the model plant for molecular genetic
studies.However, results derived from other model organisms orplant
species which provide unique insights into phy-tochrome signaling
mechanism are also briefly discussed inthis review where deemed
appropriate. The interestedreaders are referred to the accompanying
reviews on othersubjects related to phytochrome signaling, such as
photo-morphogenesis (reviewed by Joanne Chory), blue light
sig-naling (reviewed by Winslow Briggs), circadian rhythms(reviewed
by C Robertson McClung and Steve Kay), pho-totropism (reviewed by
Mannie Liscum), flowering(reviewed by George Coupland, Caroline
Dean and Detlef
Phytochrome Signaling Mechanism
Haiyang Wang and Xing Wang Deng
Department of Molecular, Cellular & Developmental Biology,
Yale University, New Haven, CT, 06520-8104, USA
Key Words: Arabidopsis, phytochrome, light signaling,
photomorphogenesis
All correspondence should be addressed to:Professor Xing Wang
DengDepartment of Molecular Cellular and Developmental BiologyYale
University, P.O.Box 208104165 Prospect Street, OML 352ANew Haven,
Connecticut 06520-8104Phone 1-203-432-8908; Fax
1-203-432-5726e-mail: [email protected]
INTRODUCTION
-
The Arabidopsis Book 2 of 30
Weigel) and chapters of phytohormones (such as
auxin,brassinosteroids, and ethylene) and sugar sensing(reviewed by
Sue Gibson and Ian Graham).
THE DISCOVERY AND ACTION MODES OFPHYTOCHROMES
Early in the twentieth century it was shown that a
pigment,separable from the activity of photosynthesis, wasinvolved
in photoperiod detection and floral induction(Garner and Allard,
1920). The nature of this pigment wasnot discovered until 30 years
later. In the 1950’s, phy-tochromes were characterized as the
proteinous pigmentthat mediates the reversible control of
night-break of shortday flowering plants (such as tobacco and
soybean) andlettuce (c.v. Grand rapids) seed germination by red and
far-red light (Borthwick et al., 1952). In the lettuce seed
germi-nation case, red light stimulates seed germination, but
thisinduction can be inhibited by subsequent exposure to far-red
light. The seeds can be repeately treated by sequentialred or
far-red light, and the final germination response is
determined by the last light treatment. This
characteristicphotoreversibility of responses aided researchers to
purifyand characterize the responsible dichromic photoreceptorthat
was later termed phytochrome for “plant color”.Another
distinguishing feature of this response is its con-formity to the
Bunsen-Roscoe Reciprocity Law, whichstates that a response should
be dependent only on thetotal amount of photons received
irrespective of the dura-tion of the exposure. Thus, the
red/far-red reversibility andreciprocity constitute the hallmarks
of the classical phy-tochrome responses. This class of phytochrome
respons-es is now defined as the low fluence responses
(LFRs,Mancinelli, 1994).
In addition to the control of lettuce seed germination,LFRs also
induce other transient responses, such aschanges in ion flux, leaf
movement, and chloroplast rota-tion (Haupt and Hader, 1994; Roux,
1994). LFRs alsoinduce changes in gene expression during
de-etiolation,stem elongation, leaf expansion, and the transition
to flow-ering (Vince-Prue, 1994). Besides the R/FR reversibleLFRs,
two more modes of phytochrome action have sincebeen discovered: the
very-low-fluence responses (VLFRs)which are activated by extremely
low light intensities, suchas the expression of the LHCB gene, and
the high-irradi-ance responses (HIRs) which depend on prolonged
expo-sure to relatively high light intensities. HIRs are
prevalentprimarily in the control of the de-etiolation process
(e.g.inhibition of hypocotyl elongation) under all light
qualities(Mustilli and Bowler, 1997; Batschauer, 1998; Table
1).
THE PHYTOCHROME GENE FAMILY AND THECHROMOPHORE
Two Reversible Forms of Phytochromes
The realization that phytochrome is enriched in
dark-grownseedlings allowed its purification with relative ease due
tothe lack of photosynthetic pigments in dark-grownseedlings
(Butler et al., 1959, 1964). Based on variousphysiological
evidence, it was predicted that phy-tochromes exist in vivo in two
photoreversible forms. Thepurification of phytochrome confirmed
this view andshowed that in dark-grown plants, phytochrome is
presentin the Pr form. On exposure to red light, the Pr form is
con-verted to the Pfr form, which is considered as the
biologi-cally active form. The Pfr form is converted back to the
Prform on absorption of far-red light. This photoconversionof
phytochrome is correlated with changes in the absorp-tion maxima of
these two forms: the purified phytochrome
Figure 1. The contrasting phenotypes of dark- vs. light-grown
Arabidopsis seedlings.Dark-grown seedlings undergo a
skotomorphogenicdevelopment program (etiolation), which is
characterizedby elongated hypocotyls, closed cotyledons and
apicalhooks. Light-grown seedlings undergo photomorphogen-esis and
are characterized by short hypocotyls, open andexpanded green
cotyledons.
-
Phytochrome Signaling Mechanism 3 of 30
in the Pr form is blue in color and absorbs maximally at 666nm,
whereas the Pfr form is olive-green in color andabsorbs maximally
at 730 nm (Quail, 1997, Figure 2).
Classification of Phytochromes and The Gene Family
Purified phytochrome from etiolated seedlings was foundto be a
soluble homodimer, with each apoproteinmonomer bearing a covalently
attached linear tetrapyrrolechromophore. The molecular mass of the
phytochromeapoprotein is approximately 120 kDa. The chromophore
isattached via a thioether linkage to an invariant cysteine ina
well-conserved domain among all phytochromes. In the1980’s,
spectrophotometric studies indicated that thereare at least two
distinct pools of phytochromes, Type I(light labile) and type II
(light stable). The light-labile pooldegrades fairly rapidly upon
exposure to red or white light.In Arabidopsis, there are five
phytochromes (termed phyA-phyE) encoded by five distinct members of
the phy-tochrome gene family (Sharrock and Quail, 1989). phyA isa
type I phytochrome, and phyB-phyE are all type II phy-tochromes.
PHYB and PHYD polypeptides are about 80%identical and are somewhat
more related to PHYE thanthey are to either PHYA or PHYC (about 50%
identity). ThePHYB, PHYD and PHYE polypeptides are the most
recent-ly evolved members of the phytochrome family (Figure
3A;Table 2). Counterparts of PHYA, PHYB and other PHYgenes are
present in most, if not all, higher plants (Clack etal., 1994;
Sharrock and Quail, 1989; Mathews andSharrock, 1997).
All five phytochromes are expressed throughout theplant with
only minor differences in their expression pat-terns (Somer and
Quail 1995; Goosey et al., 1997), howev-er, their abundance and
stability differ dramatically. phyA ismost abundant in dark-grown
seedlings and its level dropup to 100 fold after exposure to light.
The degradation of
Figure 2. Absorption spectra of phytochromes and theirdual
physiological functions.(A) Absorption spectra of the two forms (Pr
and Pfr) ofphytochromes. The Pr form absorbs maximally at 660nm,
while the Pfr form absorbs maximally at 730 nm.(B) Dual roles
(sensory and regulatory) of the phy-tochrome molecules.
Phytochromes sense the light envi-ronment (shown on the left are
the principal parametersof the light signals), undergo a
photoconversion from theinactive Pr form to the active Pfr form,
and transduce thesignals through distinct signaling pathways, which
ulti-mately leads to regulated gene expression and appropri-ate
morphogenesis. Shown on the right are the majorfacets of plant
growth and development controlled byphytochrome actions (adapted
from Quail, 1997).
-
The Arabidopsis Book 4 of 30
phyA is light dependent and requires selective recognitionand
ubiquitination of Pfr (Clough et al., 1999; Hennig et al.,1999).
PHYA gene expression is also repressed at thetranscriptional level
by light treatment (Somers and Quail,1995; Canton and Quail, 1999).
Therefore, the regulation ofphyA protein level is the result of a
coordinated transcrip-tional and post-transcriptional regulation.
In light-grownplants, phyB is the most abundant phytochrome,
whereasphyC-phyE are the less abundant type II phytochromes(Clark
et al., 1994; Hirschfeld et al., 1998).
General Structure of Phytochromes
Studies with biochemically purified phyA holoproteins froma
number of plant species indicated that the phytochromemolecule
consists of two structural domains: a photosen-sory, globular
N-terminal chromophore-binding domainwhich is sufficient for light
absorption and photoreversibil-ity (~ 70 kDa), and a regulatory,
conformationally moreextended C-terminal domain which is important
for dimer-ization and downstream signaling (~ 55 kDa). These
twodomains are connected via a flexible hinge region. The
C-terminal domain contains several conserved subdo-mains/motifs
including the regulatory core sequence (Quailbox), the dimerization
motif, and the histidine kinase-relat-ed domain (HKRD). A pair of
Per-Arnt-Sim (PAS) motifoverlaps with the Quail box (Figure3B). PAS
domains havediverse function; they can be used either as
protein-proteininteraction platforms or as response modules to
small lig-ands or changes in light conditions, oxygen levels,
andredox potentials (Quail, 1997; Neff et al., 2000). Most
pointmutations in the PRD domain of both phyA and phyB donot affect
photoreversibility but eliminate the biologicalactivity (Quail et
al., 1995; Quail, 1997).
Chromophore Biosynthesis
As mentioned above, functional phytochrome holoproteinsrequire
the covalent attachment of a chromophore to eachphytochrome
apoprotein monomer. The structure of thephytochrome chromophore was
investigated initially bydegradation approaches and determined to
be a lineartetrapyrrole, phytochromobilin (PΦB). The ligation site
ofthe chromophore was investigated by proton NMR spec-troscopy of
phytochromopeptides isolated from sequentialpepsin-thermolysin
digestion of oat phytochrome in the Pr
form. PΦB was shown to ligate via the A-ring to a
cysteineresidue located in the N-terminal half apoprotein of
PHYA(Lagarias and Rapoport, 1980).
It is interesting to mention that each pyrrole ring of thelinear
tetrapyrrole chromophore may play a different role inchromophore
assembly and the photochromic propertiesof phytochromes. A recent
study found that the A-ring actsmainly as the anchor for ligation
to PHYB. The side chains
Figure 3. The Arabidopsis phytochrome family and thedomain
structure map of a generic phytochrome mole-cule.(A) The
phylogenetic distance tree of the five phy-tochrome species from
Arabidopsis thaliana. PHYB andPHYD share ~ 80% amino acid sequence
identity, andare the products of a recent gene duplication. They
arealso more closely related to PHYE (~55% identity) than toother
phytochromes, and these three genes are consid-ered to form a
subgroup of the Arabidopsis phytochromegene family. PHYA and PHYC
form the other branch ofthe family evolution tree (adapted from
Clack et al.,1994).(B) The domain structure map of a generic
phytochromemolecule. The coordinates indicate positions of the
con-sensus sequence derived from the alignment of
multiplefull-length phytochrome polypeptide sequences byMathews et
al. (1995). The N-terminal photosensorydomain (CBD, for chromophore
binding domain) and theC-terminal regulatory domain are joined by a
flexiblehinge region (H). The chromophore binding site (C374)
islocated in the N-terminal photosensory domain. The C-terminal
domain contains several conserved motifs,including the regulatory
core sequence (Quail box), twodimerization motifs (D1 and D2), two
PAS domains (P1and P2), and the histidine kinase-related domain
(HKRD).The positions for the junction site and these
individualmotifs are indicated (adapted from Quail, 1997).
-
Phytochrome Signaling Mechanism 5 of 30
of the B- and C-rings are crucial to position the chro-mophore
properly in the chromophore pocket of PHYBand for photoreversible
spectral changes. The side chainof the D-ring is required for the
reversible spectral changeof the adducts (Hanzawa et al.,
2001).
The synthesis of PΦB is directed by an enzymatic cas-cade in the
plastid that begins with 5-aminolevulinic acid.The early steps in
the PΦB pathway are shared with thoserequired to synthesize
chlorophyll and heme. The commit-ted step is the oxidative cleavage
of a portion of the hemepool by a ferredoxin-dependent heme
oxygenase (HO) to
form biliverdin IX (BV). BV is then converted into 3Z-PΦBby the
ferredoxin-dependent bilin reductase PΦB syn-thase. Both 3Z-PΦB and
its isomerized form 3E-PΦB canserve as functional precursors of the
phytochrome chro-mophore. PΦB is then exported to the cytoplasm
where itbinds to the newly synthesized apo-phys (Terry et al.,1997,
Figure 4 A and B). Absorption of red light triggers a“Z” to ‘E”
isomerization in the C-15 double bond betweenthe C and D rings of
the linear tetrapyrrole, resulting in thefar-red light absorbing
form Pfr. This Pr-to-Pfr transition isaccompanied by rearrangement
of the protein backbone.
-
The Arabidopsis Book 6 of 30
Pfr can be converted to Pr either by a slow non-photoin-duced
reaction (dark reversion) or much faster uponabsorption of far-red
light (Quail, 1997; Fankhauser, 2001;Figure 4C).
Arabidopsis photomorphogenic mutants defective in
thePΦB-synthetic pathway have been isolated. Thesemutants (hy1 and
hy2) have dramatically reduced levels ofPΦB and consequently
functional holo-phys, and thusexhibit severely impaired
photomorphogenesis (Parks andQuail, 1991). The Arabidopsis HY1
locus encodes a HO
(designated AtHO1) responsible for much of PΦB synthe-sis in
Arabidopsis (Davis et al., 1999; Muramoto et al.,1999; Table 2).
Three additional HO genes (AtHO2-4) existin the Arabidopsis genome,
and these additional HOs mayprovide alternative pathways for making
BV (Davis et al.,2001). The Arabidopsis HY2 locus, which appears to
be aunique gene in the Arabidopsis genome, encodes the
phy-tochromobilin synthase (Kohchi et al., 2001; Table 2).
It is generally assumed that all phys have the same
chro-mophore. Since type II phytochrome species are present in
Figure 4. Arabidopsis phytochrome chromophore.(A) The
biosynthesis pathway of Arabidopsis phytochrome chromophore
(adapted from Kohchi et al., 2001). (B) Chemical structures of
heme, BV, PF**B, PCB, and PEB. Heme oxygenase converts heme to BV
by an oxidativecleavage between rings A and D at the position
marked (arrow).(C) Red light (R) triggers a “Z” to “E”
isomerization in the C-15 double bond between the C and D rings of
the lineartetrapyrrole, which is accompanied by rearrangement of
the apoprotein backbone. This results in the photoconversion
ofphytochromes from the Pr form to the Pfr form. Far-red (FR) light
converts the Pfr form back to the Pr form.
-
Phytochrome Signaling Mechanism 7 of 30
very low abundance in plants, these have not been purifiedto
homogeneity using conventional biochemical techniques,except phyB.
The phyB purified from transgenicArabidopsis shows spectral
properties similar to phyA (Elichand Chory, 1997). It should be
pointed out that besidesPΦB, phycocyanobilin (PCB), the chromophore
of the light-harvesting pigment phycocyanin, also can bind
phy-tochrome resulting in Pr and Pfr spectra that are slightly
blueshifted compared with the PΦB adducts. PΦB differs fromPCB only
by substitution of the D-ring ethyl group with avinyl group
(Lagarias and Rapoport, 1980, Figure 4B). Thisfinding allowed the
constitution of photoreversible phy-tochromes by expressing
recombinant phytochrome pro-teins in yeast and assembling them in
vitro. Analysis ofreconstituted recombinant phyA, phyB, phyC and
phyErevealed that they have similar but not identical
spectralproperties. For example, the yeast-assembled phyC
has661/725 nm and phyE has 670/724 nm as the red/far-redabsorption
maxima (Kunkel et al., 1996; Remberg et al.,1998; Eichenberg et
al., 2000). Thus it is likely that all phy-tochromes possess
similar spectral properties.
Consistent with this notion, overexpression of the mam-malian
biliverdin reductase in Arabidopsis was found tocause the loss of
multiple phytochrome activities bydegrading phytochromobilin in
vivo and constituted a newclass of chromophore mutants which is
phenotypicallystronger than the hy1 or hy2 mutants. Many of the
trans-genic plants were highly chlorotic and did not survive,
sug-gesting an essential role for phytochromes in light-mediat-
ed plant growth and development (Lagarias et al., 1997).
ACTIONS AND INTERACTIONS OF PHYTOCHROMEFAMILY MEMBERS
Phytochromes regulate a variety of developmental process-es
throughout the life cycle of plants. In most instances, theroles of
individual phytochromes are studied in the contextof specific
responses and/or developmental stages.
Seed Germination and Seedlings De-etiolation
The isolation and construction of genetic mutants lackingone or
more of these photoreceptors as well as overex-pression studies of
individual phytochromes have nowmade it possible to assess the role
of each individual phy-tochrome in plant development. These studies
revealedthat different phytochromes play both distinct and
over-lapping roles within the spectrum of plant photomorpho-genesis
(Quail., 1998; Table 3). For example, analysis ofphyA and phyB
single and double mutants has shown thatthese two phytochromes
affect a number of identical
-
The Arabidopsis Book 8 of 30
processes in response to different fluences or wavelengthsof
light. Both phyA and phyB affect germination, however,phyA is
responsible for the photoirreversible VLFRresponses triggered by a
broad spectrum of irradiations(ultraviolet, visible and far-red
light), while phyB controlsthe R/FR photoreversible effects of low
fluence response(Reed et al., 1994; Botto et al., 1996; Shinomura
et al.,1996).
The regulation of hypocotyl elongation by light duringseedling
de-etiolation is another example of the complexinterplay among
these photoreceptors. Under far-red light,phyA is probably the only
active photoreceptor, as illus-trated by the quasi-complete lack of
de-etiolation of nullalleles of phyA (Nagatani et al., 1993;
Whitelam et al.,1993). In white or red light, phyB plays a major
role, buteven null mutants do not have a hypocotyl as long as
thatof dark-grown plants. The long hypocotyl and reducedcotyledon
expansion phenotype of phyB null mutants isenhanced in double
mutants with phyA or phyD, whichindicates that it is the co-action
of multiple photoreceptorsthat senses white light during
de-etiolation (Neff & VanVolkenburgh, 1994; Johnson et al.,
1994; Chory et al.,1996). Further, a recent high-resolution kinetic
analysis ofthe growth of Arabidopsis seedlings revealed that the
redlight inhibition of hypocotyl elongation is controlled by
asequential and coordinated action of phyA and phyB.
phyAcontributes to the initial hypocotyl growth inhibition (first
3hr of irradiation), while phyB functions in the later phase(Parks
and Spalding, 1999).
Vegetative Development
It should be emphasized that in reality it is the nature of
theincident light signal and the informational homeostasis
pro-vided by the action and interaction of multiple
signalingpathways of these photoreceptors that determines the
ulti-mate cellular responses. For example, by sensing the lowR/FR
ratio of light in their surroundings, plants initiate
theshade-avoidance response by increasing the elongationgrowth of
petioles and stems, the length-to-width ratio ofleaves, and
accelerating flowering (Smith and Whitelam,1997). This response is
adapted to an enrichment of far-redlight under a leaf canopy or to
reflected light from nearbyleaves, and is a mechanism for neighbor
detection.Accelerated flowering under shade, in which there is
morefar-red light, may allow plants to complete their life
cyclebefore the canopy of other plants becomes too dense.
Thismanifestation of light quality monitoring by phytochromescan be
phenocopied by end-of-day far-red (EOD-FR) treat-ments.
phyB-deficient plants, which have a constitutive
elongated-petiole and early-flowering phenotype, do notdisplay a
petiole elongation response to EOD-FR, but theydo respond to EOD-FR
by earlier flowering (Devlin et al.,1996), indicating that phyB
plays a major role in the percep-tion of low R/FR signals. In
contrast, phyA mutants show afairly normal shade avoidance response
but are impaired intheir perception of daylength (Bagnall et al.,
1995; Casal etal., 1998). However, an antagonism between the two
phy-tochromes can be detected upon overexpression of phyA
inlight-grown plants. The resulting transgenic lines no
longerdisplay the shade-avoidance response, which apparently isthe
manifestation of the opposing effects of phyA and phyBin response
to elevated levels of far-red light (Smith, 1995).This antagonism
is minimal in wild-type light-grown plantsowing to low phyA
levels.
The phyAphyB double mutants still respond to EOD-FRtreatments by
flowering early, suggesting the operation ofother phytochromes. The
recent identification of a natural-ly occurring mutation in the
PHYD gene in the ArabidopsisWassilewskija (WS) ecotype (Aukerman et
al., 1997) hasprovided evidence that phyD performs a similar role
ofphyB. The monogenic phyD mutant plants have no obvi-ous
phenotypic abnormality, whereas plants impaired inboth the PHYB and
the PHYD genes display significantlylonger hypocotyls under either
R or white light and flowerearlier than the phyB monogenic mutants
(Devlin et al.,1999). Moreover, the triple mutant phyAphyBphyD
stillretains the ability to respond to EOD-FR treatments
bydeveloping elongated rosette internodes and acceleratedflowering
responses (Smith and Whitelam, 1997; Whitelamand Devlin, 1997),
implicating the actions of phyC and/orphyE in these responses. The
isolation of the phyE mutantconfirmed this hypothesis. The phyE
mutants show nophenotypic alteration unless it is in the phyB
mutant back-ground and the phyBphyE double mutants flower
muchearlier than the phyB monogenic mutants (Devlin et al.,1998).
These studies show that phyB, phyD and phyE con-trol shade
avoidance responses and inhibit flowering in aconditional redundant
manner.
No mutation for phyC has been reported yet, but over-expression
studies suggest a role in primary leaf expan-sion (Qin et al.,
1997). ). Analysis of thephyAphyBphyDphyE quadruple mutant revealed
that phyCappears to play no role in response to low R/FR
ratio,although phyC does seem to be involved in the regulationof
gene expression (e.g. ATHB-2 is still induced by R,Whitelam, 2001).
The analysis of phyC null mutants will bevaluable in more precisely
assessing the function of phyC.
Recently, it was found that phyE also plays a role in thecontrol
of seed germination by FR light. Either phyE isdirectly involved in
the photoperception of FR for thisresponse or the action of phyA in
mediating seed germi-nation requires the presence of phyE. Other
VLFRs andHIRs for seedling de-etiolation are normal in phyE
mutants.
-
Phytochrome Signaling Mechanism 9 of 30
On the other hand, phyA mutants flower later than wild-type
plants under long days, whereas phyE mutants flowerslightly early.
Plants doubly null for both phyA and phyEflower at the same time as
phyE single mutant plants, indi-cating that the phyE mutation is
epistatic to phyA withrespect to flowering time (Whitelam,
2001).
Interestingly, the observed early flowering phenotype ofphyB
mutants grown under 210C is diminished whengrown under 160C.
Further, a substantial delay in floweringtime under 160C compared
with 210C was observed forevery genotype carrying the phyB mutation
(phyB,phyAphyB, phyBphyD, phyBphyE, phyAphyBphyD,phyAphyBphyE, and
phyAphyBphyDphyE), indicating thatat 210C phyB negatively regulates
flowering and at 160C,the effects of phyB are severely attenuated
(dampened).However, an early flowering phenotype is still observed
forthe phyE mutant at 160C, suggesting that phyE may con-trol
flowering time over a broader temperature range thanphyB (Halliday
et al, 2001).
Structural Basis for the Differential Functions
ofPhytochromes
The underlying mechanism for the observed functional
dif-ferences between different phytochromes and their actionmodes
has begun to be unraveled. Previous experiments bydeleting regions,
random mutagenesis of the full-lengthclone, or domain swapping
between phyA and phyB haveprovided some useful information
regarding the function ofvarious parts of phytochromes. For
example, the N-terminal70 amino acids (the 6-kD domain) of oat phyA
is required forcorrect chromophore/apoprotein interactions and
under-goes a substantial conformational change upon
photocon-version of Pr to Pfr. In this region, one domain (residues
13-62) is necessary for conformational stability and another
one(residues 6-12) is involved in attenuating phytochromeresponses
(Jordan et al., 1996, 1997). An oat phyA deletionmutant lacking
amino acids 7 through 69 is inactive in trans-genic tobacco
(Nicotiana tabacum, Cherry et al., 1992),whereas deletion of amino
acids 1-52 of oat phyA causes adominant negative interference in
Arabidopsis (Boylan et al.,1994). Clough et al (1999) showed that
both the N-terminaland C-terminal halves of phyA are essential for
Pfr degrada-tion. The N-terminal region provides important
selectiverecognition signals for ubiquitin conjugation, and an
intactC-terminal domain is essential for phyA breakdown. Further,a
domain swapping and deletion analysis suggests that theN terminus
of phytochrome is essential for its specific pho-tosensory
properties and that the C termini of phyA andphyB are
interchangeable (Wagner et al., 1996). It is impor-
tant to note, however, the domain swapping experimentwhere
fusion proteins between oat phyA and rice phyB wereectopically
expressed in wild-type Arabidopsis. It has beendocumented that the
dark reversion rates and the light labil-ity of monocot and dicot
phyAs are quite different (Neff etal., 2000). Therefore, these
results should be interpretedwith caution.
A recent study demonstrated that the continuous FRlight
treatment could be replaced by intermittent FR lightpulses to
induce the FR-HIR responses. Analysis of theseaction spectra
suggests that neither phyA in its Pr formsynthesized in the dark
nor in its photoconverted Pfr formis active in inducing the signal.
Instead the short-lived sig-nal was produced during
phototransformation from Pfr toPr (Shinomura et al., 2000). This is
in sharp contrast withthe case of phyB. Alternative irradiation
with R and FR lightphotoreversibly switches on or off the phyB
responses,indicating that Pfr is the active form from which the
R-induced signal is derived.
At the molecular level, it has been demonstrated thatexpression
of the nuclear photosynthetic gene LHCB inresponse to red light
depends on both phyA and phyB(Reed et al., 1994; Cerdan et al.,
1997). However, phyA andphyB respond to light of different
wavelengths and flu-ences (phyA is responsible for VLFRs, and phyB
is respon-sible for LFRs, Hamazato et al., 1997). Further, it has
beensuggested that phyA mediates the activity of the LHCBpromoter
in response to VLFRs and HIRs by targeting dis-tinct regions of the
same LHCB1*2 promoter (Cerdan et al.,2000), suggesting that these
different action modes ofphytochromes may entail distinct signaling
pathways. Thisnotion is consistent with the observations that VLF1
andVLF2 are distinct components of the VLFR pathway ofphyA response
(Yanovsky et al., 1997), whereas FHY3 pri-marily acts in the HIR
response pathway of phyA(Yanovsky et al., 2000).
Since phyB, phyD and phyE are most similar to eachother and
possess partially redundant functions, a recentstudy was carried
out to explore the functional determinantsof these photoreceptors.
Interestingly, neither PHYD norPHYE coding regions expressed under
the control of thePHYB promoter efficiently complements a phyB null
muta-tion, with phyE being partially active and phyD
completelyinactive. In agreement with this result, overexpression
ofphyD under the 35S promoter has a negligible effect on
thehypocotyl elongation response to red or white light, where-as
overexpression phyE can suppress the phyB mutantphenotype to
certain degree. These results indicate thatphyE can interact with
the phyB signaling components inthe cell, but not very efficiently,
and phyD appears to not sig-nal through the phyB pathway to any
significant extent.Further, chimeric coding sequences in which the
ends andthe central region of PHYB and PHYD were exchanged
andexpressed under the 35S promoter. It was found that most
-
The Arabidopsis Book 10 of 30
of the red hypersensitivity of the chimeras was attributableto
the central third of the molecule, with the amino-terminaland
carboxy-terminal regions being effectively interchange-able,
suggesting that the differential activities of phyB andphyD are
primarily determined by their central domains(Sharrock et al.,
2001).
PHYTOCHROME SIGNALING AND THE CIRCADIANCLOCK
Plants make use of an array of photoreceptors
(includingphytochromes) in gathering information about the
lightenvironment for setting the clock to oscillate with a periodof
about 24 hours (i.e. entrainment of the circadian clock).It has
been shown that mutations of photoreceptor genesPHYA and PHYB,
cause the circadian rhythm of CAB2 pro-moter activity to oscillate
at a pace slower (with a longerperiod length) than that of the wild
type under various lightconditions (Somers et al., 1998). This
study revealed thatin the regulation of the Arabidopsis circadian
clock, phyAacts in low intensities of red light and blue light,
while phyBfunctions in high-intensity red light. Recently it was
shownthat phyD and phyE also act as photoreceptors in red
lightinput to the clock, and that phyA and phyB act additivelyin
red light input to the clock (Devlin and Kay, 2000). On theother
hand, the action of the circadian clock governs, atany given time,
the effect of a photoreceptor (or a plant’sresponsiveness to the
light signal) on floral initiation, whichoften exhibits the
photoperiodic response rhythm. Such aregulation of the signal
transduction of photoreceptors bythe circadian clock has been
referred to as gating (Millarand Kay, 1996; Anderson et al., 1997).
One mechanism toachieve such a gating effect would be clock
regulation ofcomponents of the light input pathway. Indeed, the
mRNAabundance and transcriptional levels of PHYA, PHYB andPHYC have
been shown to be robustly rhythmic, whereasPHYD and PHYE expression
is, at most, weakly rhythmicin Arabidopsis (Bognar et al., 1999;
McClung, 2001).
Besides the photoreceptors themselves, severalArabidopsis
flowering-time genes have been recently iso-lated and shown to be
associated with the function of thecircadian clock. Interestingly,
these clock-related genesaffect both flowering time and hypocotyl
elongation(Dowson-Day and Miller, 1999; Somer, 1999), and
theyinclude ELF3, TOC1, CCA1, LHY, ADO1/ZTL/LKP1,ADO2/LKP2,
ADO3/FKF1, and GI (Hicks et al., 1996; Wangand Tobin, 1998;
Schaffer etal., 1998; Fowler et al., 1999;Park et al., 1999;
Nelson, et al., 2000; Somer et al., 2000;Strayer et al., 2000;
Table 2). It has been proposed thatELF3 and GI are most likely
components of the light input
pathway to the circadian clock, whereas TOC1, CCA1 andLHY are
likely to be components of the central oscillator.Like phyB
mutants, both the gi and the elf3 mutants dis-play elongated
hypocotyls in red light. However, the gimutants are late flowering,
which is in contrast with theearly flowering phenotype of the phyB
and elf3 mutants.This clearly suggests that ELF3 and GI play
different rolesor use different mechanisms in controlling hypocotyl
elon-gation and flowering responses. Both ELF3 and GI arenuclear
proteins and are most likely involved in regulatingthe expression
of flowering-time genes (Huq et al., 2000a;Liu et al., 2001).
Intriguingly, both ELF3 (a novel nuclearprotein) and ADO1/ZTL/LKP1
(a protein with an amino-ter-minus PAS domain, multiple kelch
repeats, and an F-box)are capable of directly interacting with phyB
(Liu et al.,2001; Jarillo et al., 2001). The implications of these
physi-cal interactions are not yet clear.
It has been shown that CCA1 binds in a circadian fash-ion to a
short element of the LHCB1*1 (CAB2) promoter,which is sufficient to
confer phytochrome responsivenessand circadian transcription (Wang
and Tobin, 1997).Interestingly, the expression of CCA1 and LHY
itself isunder the control of phytochrome signaling
(Martinez-Garcia et al., 2000), suggesting a mechanism for
phy-tochrome input to the clock.
The interactions and feedback regulations between phy-tochrome
signaling and the circadian clock make it difficultto determine
whether the abnormalities in flowering timeobserved in the
phytochrome mutants are the conse-quence of a malfunction of the
circadian clock, a manifes-tation of the direct action of the
respective photoreceptor,or both. Mutations in the PHYA and PHYB
genes affect thecircadian clock in a similar manner (they all cause
longerperiod length for the circadian expression of the
CAB2promoter), yet their effects on flowering time are
opposite(e.g. phyA mutants flower late but phyB mutants
flowerearly). These findings suggest that the observed alter-ations
in flowering time of the phyA and phyB mutants areunlikely to be
the direct consequence of a malfunction ofthe circadian clock.
Instead, these photoreceptors maydirectly affect the floral
initiation process, but the signaltransduction of photoreceptors
may be gated (rather thanexecuted) by the circadian clock. This
notion gained sup-port from a recent genetic study which showed
that ELF3requires phyB function in early morphogenesis but not
forthe regulation of flowering time, suggesting that ELF3 andphyB
control flowering via independent signal transductionpathways (Liu
et al., 2001).
PHYTOCHROMES AS LIGHT-REGULATED KINASES
-
Phytochrome Signaling Mechanism 11 of 30
It is clear now that phytochromes function as light-regulat-ed
switches for a number of developmental processes, buthow do
phytochromes initiate their signal transductionupon
photoconversion? A long-standing hypothesis is thatphytochromes act
as light-regulated kinases. This view issupported by the detected
kinase activity in purified phy-tochrome preparations by the
Lagarias group (Wong et al.,1986). However, this finding was
contested by otherresearchers who could not detect a kinase
activity in phy-tochrome preparations (Kim et al., 1989). Recently,
the dis-covery of phytochrome-like photoreceptors in
bacteria,collectively called bacteriophytochromes (BphPs),
hasdramatically expanded our understanding of the originsand modes
of action of phytochromes in plants. Thus, abrief summary of the
most relevant information on BphPshere is essential.
Bacterial Phytochromes
Bacteria constantly regulate their physiology and behaviorto
respond and adapt to their external environment using atypical
“two-component” system which consists of a sen-sor protein and a
response regulator protein. The sensorprotein detects a change in
the external environment andcommunicates this information to the
response regulatorprotein, which in turn either controls the
expression of spe-cific genes or initiates other appropriate
cellular functionsto respond to the environmental stimuli. The
communica-tion between these two components occurs via
phospho-rylation-dephosphorylation steps. It is well known
thatthese sensor proteins act as histidine kinases,
autophos-phorylate themselves and the phosphate group is
thentransferred to a regulator molecule, leading to a cascade
ofevents that modulates gene expression. Early physiologi-cal
studies revealed that cyanobacteria display photore-versible
effects analogous to those of plant phytochromesto optimize their
photosynthesis depending on light condi-tions (Vierstra and Davis,
2000).
The most intensively studied effect is complementarychromatic
adaptation (CCA), a process in which certaincyanobacteria can
regulate the composition of their lightharvesting complexes
(phycobilisome). In the cyanobac-terium Fremyella diplosiphon,
shifts in the ratio of red togreen light lead to transcriptional
changes and altered syn-thesis of several phycobilisome components.
These twocolors have opposite effects: red light activates
cpcB2A2,an operon encoding the biliprotein phycocyanin (PC),
andinactivates the cpeBA operon, which encodes phycoery-thrin (PE),
whereas green light activates PE synthesis andshuts down PC
synthesis. The effects of red and green
light on the transcription of these light-harvesting genesare
photoreversible. The study of CCA in the cyanobac-terium Fremyella
diplosiphon led to the identification ofRcaE, for response to
chromatic adapation E. Mutations inthis gene cause defects in CCA
and abolish the responsesto red or green light. A region in the
amino terminal portionof the protein (74 kDa) shows limited
homology to the CBDdomain of the plant phytochrome. Within this
region, RcaEalso contains a cysteine residue (C) at position 198,
and itwas recently demonstrated that RcaE can covalentlyattach a
tetrapyrrole chromophore in vivo and in vitro,dependent on the
presence of C198. The carboxy terminalportion of RcaE has the
signature (all four kinase motifsnecessary for catalysis H, N, D/F,
and G) of a typical bac-terial histidine kinase domain (HKD)
present in the two-component system, including the active-site
histidine(Kehoe, 1996). Both its proposed position in the CCA
sig-nal-transduction chain and its relation to phytochromessuggest
that RcaE controls CCA by acting as a photore-ceptor. In addition,
a response regulator (RcaF) for RcaEhas been identified, which is
located in the genome direct-ly downstream of RcaE (Kehoe,
1997).
Phytochrome-like sequences were also identified by theKazusa
sequencing project of the cyanobacteriumSynechocystis sp. PCC6803.
Four genes with varyingdegree of relatedness to RcaE and higher
plant phy-tochromes were uncovered (Hughes and Lamparter, 1999;Wu
and Lagarias, 2000). One of them, Cph1 for cyanobac-terial
phytochrome 1, was shown to be able to bind to theplant phytochrome
chromophore (phytochromobilin PΦBor phycocyanobilin PCB)
autocatalytically and to displayabsorption spectra with
photoreversible red (Pr) or far-red(Pfr) absorption maxima typical
of plant phytochromes.Further, Cph1 was shown to be a
light-regulated histidinekinase. Both autophosphorylation of Cph1
and transphos-phorylation of Rcp1 (the response regulator for Cph1)
areinhibited by red light and stimulated by far-red light (Yeh
etal., 1997; Lamparter et al., 1997). Although the in vivo
rel-evance of the protein kinase activity of Cph1 in light
sig-naling is still under investigation, these studies
clearlyestablish that Cph1 is a bona fide cyanobacterial
phy-tochrome. Indeed, the discovery of Cph1 created anexcitement in
the phytochrome research field becausewith the ability to exploit
bacterial genetics, these BphPsnow offer simple models to help
unravel the biochemicaland biophysical events that initiate phy
signal transmis-sion. In particular, the efficiency with which
highly solublerecombinant phytochrome can be prepared from E.
colioverexpressors offers fresh hope that the three-dimen-sional
structure of this class of photoreceptors could beresolved via NMR
and X-ray diffraction analysis of phy-tochrome crystals (Hughes and
Lamparter, 1999).
Later, phytochrome related photoreceptors were alsoidentified in
both the purple photosynthetic bacterium
-
The Arabidopsis Book 12 of 30
Rhodospirillum and non-photosynthetic bacteria such
asDeinococcus radiodurans, Pseudomononas putida andPseudomonas
aeruginosa (Jiang et al., 1999; Davis et al.,1999). Together, these
studies of phytochromes in prokary-otes have provided compelling
evidence that these phy-tochrome-like photoreceptors can perceive
the light sig-nals and relay the information via a His kinase
signaling
cascade, most likely through a response regulator, such asRcp1.
The organization and distribution of BphPs supportthe view that
phytochromes first appeared in eubacteria,were modified by
photosynthetic bacteria to better quanti-tatively mediate the shade
avoidance response, and thenwere transferred to plants from
cyanobacteria during theendosymbiotic formation of chloroplasts
(Fankhauser,2000; Vierstra and Davis, 2000).
It should be mentioned that the identities of the chro-mophores
used by these BphPs have not been explicitlydetermined. For
cyanobacteria, it is likely that PCB (and/or possibly PEB,
phycoerythrobilin, Figure 4B) is thechromophore since these species
make copious amountsof this bilin as an accessory pigment in
photosynthesis.The nature of the BphP chromophore(s) for the
non-pho-tosynthetic bacteria is even less clear since these
speciesare not known to produce linear bilins. Interestingly,
puta-tive heme oxygenase genes (HOs) in D. radiodurans,
P.aeruginosa and P. putida have been identified, indicatingthat
these species have the capacity to produce lineartetrapyrrole
biliverdin (BV). Remarkably, the D. radiodu-rans HO is encoded by a
single open reading frame (des-ignated Bph0) that is immediately
upstream (4 bp overlap)of the BphP in an apparent operon. In this
way, chro-mophore and apoprotein synthesis appears to be con-nected
transcriptionally. Further, this BphP can assemblewith BV to
produce a stable photochromic adduct, where-as neither higher-plant
phytochromes nor SynechocystisCph1 can bind BV. Thus it appears
that BV might be thechromophore specific to the D. radiodurans
BphP(Vierstra and Davis, 2000).
Higher Plants Phytochromes
The discovery of BphPs led to more careful sequenceanalysis of
higher plants phytochromes and similaritybetween the C-termini of
phytochromes and BphPs wasidentified (Schneider-Poetsch, 1992,
Figure 5A).However, plant phytochromes have two additionaldomains:
a serine-rich N-terminal extension domain(NTE) and a PAS
repeatdomain (PRD) located betweenthe CBD and the HKRD. In
addition, several criticalresidues required for activity in the
majority of bacterialsensor kinases are not conserved in all plant
phy-tochromes. Moreover, mutating some of the remainingcritical
residues for His kinase activity does not affect theactivity.
Therefore, it appears that plant phytochromesare not active His
kinases (Vierstra and Davis, 2000). Totackle this question, Clark
Lagarias’s group developed arecombinant system to express and
purify plant phy-
Figure 5. Arabidopsis phytochromes function as light-regulated
kinases.(A) Domain conservation between a cyanobacterial
phy-tochrome (Cph1) and Arabidopsis phytochromes. Theconserved
cysteine residue for chromophore binding isindicated (*). HKD:
histidine kinase domain; PRD: PASrelated domain; HKRD: histidine
kinase related domain.(H) highlights the conserved histidine on the
HKD domainof Cph1. The percent amino acid identities between theHKD
of Cph1 and both PRD and HKRD of Arabidopsisphytochromes are
indicated. Also note that Arabidopsisphytochromes have distinct
N-terminal extensions (NTE).(B) Proposed roles of Arabidopsis
phytochrome kinaseactivity. Light regulated kinase activity may
auto-phos-phorylate the phytochrome molecules themselves,
ortrans-phosphorylate their interacting partners (PIFs). Inturn,
these phosphorylation events could affect the stabil-ity of
photoreceptor in case of phyA, the subcellular local-ization of
phytochromes, their ability to interact with PIFs,and the
activities of other signaling intermediates. Theamino acid numbers
correspond to the oat phyA (adapt-ed from Neff et al., 2000).
-
Phytochrome Signaling Mechanism 13 of 30
tochromes in yeast. Purified oat phyA expressed in S.cerevisiae
and the green alga Mesotaenium caldariorumphytochrome expressed in
P. pastoris were assembledwith PCB in vitro. These phytochromes
display theexpected spectroscopic properties and protein
kinaseactivity in a light and chromophore-regulated manner.
Inaddition, the purified oat phyA can phosphorylate histoneH1 and
Rcp1, the response regulator substrate of Cph1.However, unlike
their cyanobacterial counterparts, theyauto-phosphorylate on
Ser/Thr rather than His/Asp.These studies provided strong
indication that the kinaseactivity is an intrinsic property of
plant phytochromes andnot an artifact due to co-purification of
another proteinkinase (Yeh and Lagarias, 1998).
Physiological Roles of Phytochrome Kinase activity
The claim that higher plant phytochromes function asprotein
kinases triggered many questƒions such as whatis the biological
role of the kinase activity? Which form ismore active? Pr or Pfr?
Answers to these questions havejust begun to be unraveled. For
example, recombinantoat phyA is a light and chromophore-modulated
proteinkinase with Pfr being more active than Pr (Yeh andLagarias,
1998). Two phosphorylation sites have beenmapped for oat phyA. The
Ser-7 is phosphorylated in vivoin both the Pr and Pfr forms, and
mutagenesis studiessuggest that this residue is implicated in
down-regulationof phyA signaling (Stockhaus et al., 1992). The
Ser-17 isphosphorylated by protein kinase A in vitro only in the
Prform. Mutation of the first 10 Ser of phyA to Ala (all con-tained
within the first 20 aa) or deletion of this regionresults in a
mutant that is hypersensitive to light(Stockhaus et al., 1992;
Jordan et al., 1996, 1997). Takentogether, these results suggest a
desensitization mecha-nism via these serines.
Another serine residue, Ser-598, is preferentially
phos-phorylated in the Pfr form in vivo. The importance of
thisresidue has been demonstrated in vitro because a S598Kmutant
loses light-regulated kinase activity (Fankhauser etal., 1999).
Consistent with this notion, when an oat PHYAcDNA with a Ser-598 to
Ala substitution was expressed inthe phyA mutants, the transgenic
plants exhibited hyper-sensitivity to far-red light, suggesting
that the Ser-598phosphorylation may serve as a desensitizing
mechanismof the Pfr activity by disrupting the interactions
betweenphytochromes with their downstream signaling partners(Park
et al., 2000, Figure 5B).
A complication on this issue arised from a recent studywhich
showed that although point mutations in the HKRD
region of phyB cause strong phenotypes in hypocotyllength and
flowering time, indicating that this domain isimportant for phyB
signaling. Quite surprisingly, deletion ofthis domain resulted in a
milder phenotype, suggestingthat this domain is dispensable (Krall
and Reed, 2000).
Do higher plants phytochromes initiate a two-compo-nent
phosphorelay cascade similar to that of RcaE? Thereis no
affirmative answer to this question, but a recent studyhints that
phytochromes could be the target of a two-com-ponent signaling
system operating in plants. A responseregulator from Arabidopsis
(ARR4) was identified as beingpredominantly expressed in response
to red light, sug-gesting that ARR4 may be involved in phyB
signaling.Further, ARR4 specifically interacts with the extreme
N-ter-minus of phyB both in vivo and in vitro. Interaction of
ARR4with phyB results in the stabilization of the active Pfr formof
the photoreceptor as determined by inhibition of Pfr—Prdark
reversion in vivo. Accordingly, transgenic Arabidopsisplants
overexpressing ARR4 display hypersensitivity to redlight with
respect hypocotyl and root growth as well asflowering time.
Further, the Asp residue involved in phos-photransfer in the
receiver domain of ARR4 was identified,and transgenic plants
overexpressing a mutated form ofARR4 in which the Asp was
substituted by a Asn residue,revealed a hyposensitive phenotype
regarding all phyB-dependent light responses. These data indicate
that phyBis the target of a novel two-component phosphorelay
sys-tem that modulates red-light-dependent signaling by
Figure 6. A summary of identified phytochrome-interact-ing
partners (PIFs).Some PIFs interact with both phyA and phyB,
whereasothers specifically interact with either phyA only or
phyBonly. Note that most of these physical interactions
weredetected with the yeast two-hybrid assay and/or in vitrobinding
assay.
-
The Arabidopsis Book 14 of 30
direct interaction of its response regulator (ARR4) with
thephotoreceptor (Sweere et al., 2001; Table 2). Further stud-ies
should reveal how a two-component system is func-tionally linked to
red light signaling and how other endoge-nous or environmental
signals modulate phyB activity. Forexample, the expression ARR4 has
been demonstrated tobe rapidly induced by cytokinin (Brandstatter
and Kieber,1998), suggesting a possible link between cytokinin
sig-naling and red light response mediated by phyB.
PHYTOCHROME-INTERACTING SIGNALINGPARTNERS
Phytochrome-interacting Partners (PIFs)
Protein-protein interactions are necessary for many
signaltransduction cascades. It is reasonable to expect that
phy-tochrome also interact with some partner protein(s) to
relayinformation about the light environment in the
cells.Identification of the molecular components responsible
forintracellular photosignal transduction is currently an area
ofintense research effort. Both general screenings for phy-tochrome
interacting partners and targeted protein-proteininteraction
studies have identified a number of phy-tochrome-interacting
factors (PIFs). Those include PIF3 (Niet al., 1998), PKS1
(Fankhauser et al., 1999), NDPK2 (Choiet al; 1999), cryptochromes
(both CRY1 and CRY2) and theAUX/IAA proteins (Ahmad et al., 1998;
Colón-Carmona etal., 2000; Mas et al., 2000; Reed, 2001; Figure 6;
Table 2).The physiological roles for some of these factors in
phy-tochrome signaling have been substantiated by recentmolecular
genetic studies. PIF3 is a nuclear localized basichelix-loop-helix
(bHLH) protein. Transgenic Arabidopsisseedlings with
antisense-imposed reductions in PIF3 levelsexhibited strongly
reduced responsiveness to light signalsperceived by phyB, and
partially reduced responsiveness tosignals perceived by phyA. These
data indicate that PIF3 isfunctionally active in both phyA and phyB
signaling path-ways in the plant cell, consistent with its binding
to bothphotoreceptors. Further, a T-DNA tagged pif3 mutant
(des-ignated poc1) also exhibits enhanced responsiveness to
redlight (Halliday et al., 1999). This exaggerated response ofthe
poc1 mutants to red light is caused by a T-DNA inser-tion into the
promoter of the PIF3 gene, and thus likely torepresent a
gain-of–function phenotype. The phyB mutationis epistatic to this
mutant, indicating that PIF3 is an authen-tic phyB signaling
component. PKS1 is a basic, soluble,cytoplasmic protein and has
been proven to be a substrate
for light-regulated phytochrome serine/threonine kinaseactivity,
indicating that protein phosphorylation is involvedin phytochrome
signaling, and which might modulate phy-tochrome kinase activity or
their subcellular localization.PKS1 overexpressing plants display
less sensitivity to redlight, suggesting that it acts as an
inhibitor of phyB signal-ing (Fankhauser et al., 1999). In
contrast, NDPK2 (nucleo-side diphosphate kinase 2) appears to be a
positive regula-tor of both phyA and phyB signaling. Although
hypocotylelongation is not obviously affected by this locus, its
loss offunction alleles have a small but significant reduction
incotyledon greening and opening of the hypocotyl/cotyledonhook
during de-etiolation (Choi et al; 1999).
Recently, using the C-terminal 300 amino acids ofArabidopsis
phyB as a bait, two additional phytochrome-interacting proteins,
PRP1 and PAB1 were isolated. PRP1(phytochrome related phosphatase
1) shares 30% identityto protein phosphatase 2Cs (PP2C) and has
been shownto possess the activity of a serine/threonine
phosphatase.Overexpression of the catalytic domain of PRP1
inArabidopsis plants leads to a reduced sensitivity to redlight. In
contrast, a knockout of PRP1 (prp1 mutant) ismore sensitive to red
light and less sensitive to far-red lightthan wild-type plants,
suggesting that PRP1 negativelyregulates phyB signaling and
positively regulates phyA sig-naling, possibly by dephosphorylating
phytochromes orother signaling components. PAB1 (phytochrome
actinbinding protein 1) contains six kelch repeats and binds
toactin filaments in vitro. The pab1 mutant is less sensitive tored
light and has reduced leaf expansion after a certaindevelopmental
stage, leading to smaller rosettes in matureplants, indicating that
PAB1 might play a fundamental rolein cell expansion. PAB1-GFP
localizes to a web-like pat-tern (possibly ER) in the cytoplasm of
hypocotyl cells(Chen et al., 2001).
Functional Domains and Intra-molecular Signaling
ofPhytochromes
It should be pointed out that these proteins interact
withdifferent structural motifs in the C-terminal domain of
phy-tochromes and that their interactions are differentially
reg-ulated by light. For example, NDPK2 binds to the Quail
boxpreferentially in the Pfr form in a GTP-dependent manner(Choi et
al., 1999; Park et al., 2000). PIF3 also binds to theQuail box
preferentially in the Pfr form. However, it is dif-ferent from
NDPK2 in that both the N-terminal and the C-terminal domains are
apparently required for full activity (Niet al., 1999; Zhu et al.,
2000). On the other hand, PKS1
-
Phytochrome Signaling Mechanism 15 of 30
binds to the Ser/Thr kinase motif equally well in both the Prand
Pfr forms, indicating that this motif is exposed in bothspectral
forms. However, PKS1 phosphorylation and phy-tochrome
autophosphorylation are stimulated by a factorof 2 to 2.5 in the
Pfr form (Fankhauser et al., 1999). Thus itis probable that
phosphorylation is an important regulato-ry event for the
phytochrome-PKS1 interaction. Thesestudies suggest that different
inter-domain crosstalks acti-vate a specific motif in the
C-terminal domain for recogni-tion by different factors and
contribute to the spectralintegrity of phytochromes. Direct
evidence for such anotion has come from chemical cross-linking
experimentswhich detected a R/FR-dependent interaction between
theN-terminal peptide and the distal C-terminal peptide (Parket
al., 2000). Consistent with this, a mutation in the CBDdomain of
phyB (phyB-401) causes a defect in the pho-toreversibility and
enhanced light sensitivity (Kretsch et al.,2000). Similarly,
another phyB missense mutation, phyB-101, is in the second PAS
repeats (Bradley et al., 1996).This mutation affects spectral
properties of the pigment,causing accelerated dark reversion from
Pfr to Pr, andalters the EOD-FR response in seedlings (Elich and
Chory,1997). These results also support the notion of inter-domain
cross-talk within the phytochrome molecules.
In vitro kinase assays have identified other substrates
ofphytochromes, including the blue light photoreceptors
CRY1 and IAA proteins (Ahmad et al., 1998; Colón-Carmona et al.,
2000). Although the phosphorylation ofCRY1 is not light-dependent
in an in vitro experiment, invivo analysis shows that cry1
phosphorylation is stimulat-ed by red light. Moreover, the
identification of the shy2mutant as a suppressor of phyB and of
phytochrome chro-mophore mutants and the gene encodes IAA3, one of
theearly auxin-inducible genes (Tian and Reed, 1999), sug-gesting
that the interactions between phytochromes andthe AUX/IAA proteins
and phosphorylation events are like-ly to be biologically
relevant.
GENETICALLY IDENTIFIED EARLY INTERMEDIATESOF PHYTOCHROME
SIGNALING
Genetic screening for Arabidopsis mutants potentiallydefective
in signaling intermediates either specific to phyAor phyB, or
shared by both pathways has identified a num-ber of candidate loci
(Figure 7).
Phytochrome A-specific Signaling Components
Mutants affected in phyA-specific signaling process werescreened
under a continuous far-red light (FRc) conditionand a number of
potential signaling components specificto this pathway have been
identified, including FHY1,FHY3 (Whitelam et al., 1993), FIN2 (Soh
et al., 1998), SPA1(Hoecker et al., 1998), FAR1 (Hudson et al.,
1999), FIN219(Hsieh et al., 2000), PAT1 (Bolle et al., 2000), EID1
(Bucheet al., 2000), HFR1/RSF1/REP1 (Fairchild et al.,
2000;Fankhauser and Chory, 2000; Soh et al., 2000) and
LAF1(Ballesteros et al. 2001). The fhy1, fhy3, fin2, fin219,
far1,laf1, laf6, and hfr1/rep1/rsf1 mutants show less sensitivityin
continuous far-red light, indicating that their respectivegenes
encode positive regulators of the phyA signalingpathway. On the
other hand, mutations in the SPA1 andEID1 genes cause increased
sensitivity to the FR light sig-nal, and it is most likely that
they act as negative regulatorsof the signaling cascade.
Among the positive regulators of phyA signaling identi-fied,
their loss-of-function mutants (mostly null mutationalleles) only
exhibit partial defects with different spectra andstrength in phyA
signaling, suggesting that phyA signalinginvolves multiple branches
or parallel pathways controllingoverlapping yet distinctive sets of
far-red light responses(hypocotyl growth, apical hook and cotyledon
opening,
Figure 7. A simplified genetic model for phytochrome-mediated
signaling pathways. phyA and phyB signalingentails specific as well
as shared components. These locipresumably act upstream of the
COP/DET/FUS genes,thus controlling HY5 and the degree of
photomor-phogenic development. Arrows indicate a positive
action,and the bars indicate a repressive effect.
-
The Arabidopsis Book 16 of 30
anthocyanin accumulation, far-red light pre-conditionedblocking
of greening, gravitropic response, light-regulatedgene expression
etc., Barnes et al., 1996; Hudson, 2000;Wang and Deng, 2001).
Moreover, the finding that the dou-ble mutants fhy3-1/far1-2,
fhy3-1/fhy1-1 and far1-2/fhy1-1all display an additive effect,
whereas the fhy3-1/spa1-3double mutant has an intermediate
hypocotyl-length phe-notype (Wang and Deng, 2001), also indicates
that there isno simple downstream/upstream relationship among
thesephyA signaling components. Further, complex genetic
rela-tionships such as non-allelic non-complementationbetween fin2
and fhy3-1 as well as between fin219 and fhy1have been reported
(Soh et al., 1998; Hsieh et al., 2000),suggesting that their gene
products may directly interact orengage in extensive
cross-talk.
Several phyA signaling intermediates have been charac-terized at
the molecular level (Table 2). LAF6 is a plastid-localized
ATP-binding-cassette protein involved in coordi-nating
intercompartmental communication between plas-tids and the nucleus
(MØller et al., 2000). PAT1 is a newmember of the GRAS family
(Bolle et al., 2000), whereasFIN219 is a GH3-like protein whose
expression is rapidlyinducible by auxin (Hsieh et al., 2000). Both
PAT1 andFIN219 are cytoplasmically localized proteins. FAR1,FHY3,
SPA1, HFR1, LAF1 and EID1 are all nuclear local-ized factors.
Interestingly, FAR1 and FHY3 encode twoclosely related proteins
that constitute one branch of alarge gene family (Hudson et al.,
1999; Wang and Deng,2001). HFR1 is an atypical bHLH transcription
factor close-ly related to PIF3 (Fairchild et al., 2000) and LAF1
is a MYBtype transcription activator (Ballesteros et al. 2001).
SPA1possesses a C-terminal WD-40 repeat domain that is mostclosely
related to that of COP1 (Hoecker et al., 1999),whereas EID1 is a
novel F-box protein most probablyinvolved in ubiquitin-dependent
proteolysis (Dieterle et al.,2001). The biochemical functions of
PAT1, FIN219, FAR1,FHY3 and SPA1 remain largely unknown.
The findings that fhy3 and far1 mutants display similaryet
distinct phenotypes and that FHY3 and FAR1 encodetwo homologous
proteins are particularly interesting. Morestrikingly,
overexpression of FAR1 or FHY3 can suppressthe phenotype of each
other’s loss-of-function mutations.It is also of interest to note
that overexpression of partialfragments of FHY3 in a wild-type
background causesreduced sensitivity to FRc in a dosage-dependent
manner.Especially, Arabidopsis seedlings homozygous for
thetransgene overexpressing the C-terminal portion of
FHY3(C473-839), which contains a Coil-coil domain, display
anapparent complete loss of FRc responses, remarkablysimilar to
phyA null mutants. This result indicates that theC-terminal
fragment of FHY3 may interact with other inter-mediates of phyA
signaling and that non-productive bind-ing of this truncated FHY3
protein with its interactive part-ners could shut down the entire
phyA signaling by a dom-
inant-negative interference. This interference is substan-tially
stronger than the effect of an fhy3 null mutation.Direct evidence
for this view is provided by the demon-stration that FHY3 and FAR1
directly interact with eachother in a yeast two-hybrid assay and an
in vivo co-immunoprecipitation assay (Wang and Deng, 2001). It
isconceivable that through interactions with FAR1 and
otherinteractive partners of FHY3 and FAR1, FHY3 could exertits
effect on a large number of phyA signaling intermedi-ates in
mediating FRc responses. Therefore, FHY3, togeth-er with FAR1,
could constitute a central regulatory knot inthe phyA signaling
network. Determining the protein-pro-tein interactions among these
phyA signaling intermedi-ates and identifying their novel
interactive partners shouldenhance our understanding of the phyA
signaling pathway.
Phytochrome B-specific Signaling Components
Putative phyB-specific signaling mutants have also
beenidentified, including red1, pef2 and pef3. They share anumber
of features with phyB mutants, such as a longhypocotyl phenotype
specifically under red light, earlyflowering in short days, and
elongated petioles, suggestingthat these loci positively regulate
phyB signaling (Reed etal., 1993; Ahmad and Cashmore, 1996; Wagner
et al.,1997). On the other hand, the srl1 mutants show
enhancedresponsiveness to red light, suggesting that SRL1 is a
neg-atively acting component specific to phyB signaling (Huqet al.,
2000b). The molecular identities of these genes arecurrently
unknown.
Signaling Components Shared by Both phyA andphyB
Additional mutants, pef1 and psi2, affect responses frommultiple
photoreceptors. The pef1 mutants show attenuat-ed red and far-red
responses, whereas the psi2 mutant ishypersensitive to red and
far-red light, and has necroticlesions in light-grown plants (Ahmad
and Cashmore 1996;Genoud et al., 1998), suggesting that these two
loci areshared by both the phyA and phyB signaling pathways.The
cloning of these genes and characterization of theirfunctions
should greatly enhance our understandingregarding how these two
pathways converge to regulatephotomorphogenesis.
-
Phytochrome Signaling Mechanism 17 of 30
G proteins, Calcium-binding Protein and Ion Flux inPhytochrome
Signaling
It should be pointed out that the multiplicity of
phy-tochrome-regulated responses signifies that this
diversityresults from operation of multiple signal transduction
path-ways triggered by the phytochrome. Previous microinjec-tion
and pharmacological studies have suggested theinvolvement of
G-proteins, cGMP and Ca2+/calmodulin inphytochrome signaling
(Bowler et al., 1994; Neuhaus et al.,1997). Particularly,
heterotrimeric G-proteins have beenimplicated in several processes
during plant growth anddevelopment, and they transduce
extracellular signals tothe cell. In general, heterotrimeric
G-proteins consist ofthree subunits: a, b, and g. Analysis of the
completegenome sequence of Arabidopsis indicated that
theArabidopsis genome contains only a single Ga gene, pre-viously
designated ATGPA1 (Ma et al., 1990), and a singleGb gene,
designated AGB1 (Weiss et al., 1994). Recentlythe Arabidopsis Gg
subunit was identified by a yeast two-hybrid screen using Gb as a
bait (Mason and Botella, 2001;Table 2) and it was shown that Gg
also is encoded by asingle copy gene in Arabidopsis.
Transgenic Arabidopsis plants conditionally overex-pressing the
Ga subunit of the heterotrimeric G-proteinunder the control of a
glucocorticoid-inducible promoterexhibited a light-dependent
hypersensitive response as aresult of reduced hypocotyl cell
elongation. Further, thisenhanced response in far-red and red light
requires func-tional phyA and phyB, respectively. Interestingly,
theresponse to far-red light depends on functional FHY1 butnot on
FIN219 and FHY3, suggesting that the ArabidopsisGa protein may act
only on a discrete branch of the phyAsignaling pathway (Okamoto et
al., 2001). However, a sep-arate study reported that
loss-of-function gpa1 mutantsdisplay partial de-etiolation in the
dark, with shorthypocotyls and open apical hooks typical of
light-irradiat-ed seedlings. The short hypocotyl of gpa1 seedlings
isreportedly due to a defect in cell division, not cell elonga-tion
(Ullah et al., 2001). The reason for the discrepancyregarding GPA1
in photomorphogenesis reported in thesetwo studies is currently
unknown. Regardless, these stud-ies provide genetic support for the
involvement of het-erotrimeric G proteins in light control of plant
photomor-phogenesis.
Genetic studies have also provided supporting evidencefor the
involvement of Ca2+ in phytochrome signaling. Thesub1 mutant
exhibits hypersensitive responses to both far-red light and blue
light. The SUB1 gene was found toencode a Ca2+-binding protein.
Genetic interaction stud-ies suggest that SUB1 is a component of a
cryptochromesignaling pathway and is a modulator of the phyA
signal-
ing pathway. Further, SUB1 negatively regulates HY5, abZIP
transcription factor and a positive regulator of
photo-morphogenesis (Guo et al., 2001; Table 2).
Some early studies also suggested that phytochromeexerts its
effects by first altering the permeability of theplasma membrane to
ions (Kendrick and Bossem, 1987).This is most likely to be true for
some light-regulatedresponses in certain plant species, such as the
bud induc-tion process of the moss Physcomitrella patens(Ermolayeva
et al., 1997), and the unrolling of the primaryleaf wrapped within
the oat coleoptiles (Viner et al., 1988).However, there have been
no reports of changes in cyto-plasmic ion (such as Ca2+)
concentration in Arabidopsishypocotyl cells in response to light,
neither is there anyelectrophysiological evidence that phytochrome
signalingin hypocotyls involves changes in ion fluxes (Parks
andSpalding, 1999; Spalding, 2000). Therefore, the role of ionflux
and membrane depolarization in phytochrome controlof Arabidopsis
photomorphogenesis remains elusive.
LIGHT-REGULATED CELLULAR LOCALIZATION OFPHYTOCHROMES
Phytochrome apoproteins are synthesized within thecytosol and
assembled autocatalytically with the plastid-derived chromophore.
For years, there has been a debateabout the intracellular
localization of phytochrome. Earlystudies using an
immunohistological approach and cellfractionation assay supported
the notion that phy-tochromes are predominantly localized outside
the nucle-us (Pratt, 1994). Recently, it was shown that upon
photo-conversion of Pr to Pfr, both phyA and phyB tagged withGUS or
green fluorescent protein (GFP) can translocatefrom cytoplasm into
the nucleus where they form intranu-clear spots (Sakamoto and
Nagatani, 1996; Yamaguchi etal. 1999; Kircher et al., 1999; Nagy et
al., 2000). In etiolat-ed seedlings and dark-adapted plants the
phyB:GFPfusion protein is localized in the cytosol. The
nucleartranslocation and spot formation of phyB:GFP is inducedby
continuous red light treatment, or multiple red lightpulses which
are reversible by subsequent far-red treat-ment, indicating that
the nuclear import of phyB is mediat-ed by a low-fluence response
(LFR) of phytochrome. It fol-lows that the nuclear import of phyB
is regulated by phyBitself, and/or by some other red-light
absorbing photore-ceptor(s). On the other hand, brief irradiation
with red, far-red, or blue light can induce rapid nuclear import
andintranuclear spot formation of phyA:GFP (one magnitudeof order
faster than that of phyB), preceded by an evenfaster cytosolic spot
formation of the fusion protein, a phe-
-
The Arabidopsis Book 18 of 30
nomenon reminiscent of SAP (sequestered area of phy-tochrome)
formation. Besides, nuclear translocation ofphyA is also induced by
continuous far-red but not by con-tinuous red light treatment.
Thus, the nuclear translocationof phyA is mediated by both the very
low fluence response(VLFR) and the far-red high irradiance response
(HIR), indi-cating that phyA regulates its own nuclear import
inde-pendent of any other types of phytochromes. It should
bepointed out that the transport seems to be not complete,i.e. a
significant portion of phyA:GFP remains cytosolic(Nagy et al.,
2000).
Full-length phyC-E have also been fused to GFP andexpressed in
transgenic Arabidopsis. Surprisingly, thenucleocytoplasmic
partitioning of these phytochromespecies seems not to be regulated
by light. They aredetected in the cytosol and the nucleus in
etiolatedseedlings. However, these fusion proteins are similar
tophyA or phyB:GFP in the formation of nuclear speckles ina
light-dependent manner (Nagy et al., 2000). Interestingly,the
nuclear import of all phytochrome species is regulatedby the
circadian clock and displays oscillations under con-stant
conditions, suggesting a new regulatory loop whichcould modulate
the gating of phytochrome signaling by thecircadian clock and
resetting of the circadian clock bythese photoreceptors (Nagy,
2001).
At the current stage, very little is known about themolecular
machinery and factors modulating the nucleo-cytoplasmic
partitioning of phytochromes. It has beenspeculated that the Pr
conformers of phytochromes areanchored/retained in the cytosol and
the Pfr conformersdo not interact with the anchoring proteins and
are thussubject to nuclear import. The autophosphorylation pat-tern
of phytochromes and phosphorylation of other pro-teins by the
kinase activity of phytochromes (such asPKS1) may play a role in
modulating the retention/releasestatus of phytochromes, thus
attributing to the control oftheir nuclear import (Nagy et al.,
2000). In addition, the var-ious phytochrome signaling
intermediates describedabove and other signaling cascades (such as
phytohor-mones) could also affect the intracellular distribution
ofphytochromes, thus modulating the amount of phy-tochromes
available for interaction with other componentsin the nucleus to
regulate light-responsive gene transcrip-tion. Further studies
aimed to map the possible nuclearlocalization signal (NLS) and
nuclear export signal (NES)and to determine the role of
phosphorylation and dephos-phorylation in regulating phytochrome
subcellular translo-cation may help to resolve these issues.
Particularly, thebiological significance of the nuclear
translocation eventsfor phytochromes remains to be
substantiated.
PHYTOCHROME SIGNALING AND THE
DOWNSTREAM COP/DET/FUS PROTEINS
The fact that distinct photoreceptor-triggered
signalingprocesses can lead to similar photomorphogenic
develop-ment implies that these signaling pathways converge
toregulate developmentally important genes through a set ofcommon
late signaling intermediates. Indeed, aside fromthese
phytochrome-specific signaling intermediates,genetic screens have
also identified eleven pleiotropicCOP/DET/FUS loci whose gene
products act as negativeregulators of photomorphogenesis and
function down-stream of multiple photoreceptors, including phyA
andphyB (Wei and Deng, 1996, Figure 8). Among theseCOP/DET/FUS
proteins, COP1 is a RING-finger proteinwith WD-40 repeats (Deng et
al., 1992). Studies with afunctional GUS-COP1 fusion protein
indicated that COP1acts within the nucleus to suppress
photomorphogenicdevelopment in darkness, and that inactivation of
COP1 bylight was accompanied by reduced COP1 abundance in
Figure 8. The phenotype of cop1 (constitutive photomor-phogenic)
mutants and the proposed roles of theCOP/DET/FUS proteins in
photomorphogenesis.(A) Dark-grown cop1 mutant seedlings
phenotypicallymimic light-grown wild-type seedlings. (B) A total of
eleven pleiotropic COP/DET/FUS loci func-tion as repressors of
photomorphogenesis. Light signalsperceived by multiple
photoreceptors are transduced toinactivate these COP/DET/FUS
proteins, and to turn onphotomorphogenic development.
-
Phytochrome Signaling Mechanism 19 of 30
the nucleus (von Arnim and Deng, 1994). In the dark,
COP1directly interacts with and targets HY5, a bZIP transcrip-tion
factor which acts to promote photomorphogenicdevelopment, for
degradation via a 26S proteasome-medi-ated process (Figure 9).
Thus, COP1 presumably functionsas a putative ubiquitin ligase
(Oyama et al., 1997; Ang etal., 1998; Osterlund et al., 2000).
Eight otherCOP/DET/FUS proteins are subunits (CSN1-CSN8) of
aprotein complex, the COP9 signalosome (Castle andMeinke, 1994;
Peng et al., 2001a, Peng et al., 2001b;Karniol et al., 1999; Kwok
et al., 1998; Serino et al., 1999;
Wei et al., 1994; Table 2). Its subunit-by-subunit similarityto
the lid subcomplex of the 26S proteasome and thedemonstrated
interaction between these protein complex-es suggest that the COP9
signalosome may be involved inregulated proteolysis (Kwok et al.,
1999; Peng et al.,2001c, Schwechheimer and Deng, 2000). This notion
isfurther supported by a recent demonstration that theCOP9
signalosome physically interacts with the SCFTIR1
E3 ubiquitin ligase and is required for efficient degradationof
a candidate substrate of SCFTIR1 (Schwechheimer etal., 2001). Based
on the pleiotropic nature of thecop/det/fus mutant phenotype, it is
likely that theCOP/DET/FUS proteins repress the activities of
multipletranscription factors or transcriptional
regulators.Consistent with this notion, multiple
COP1-interactingtranscriptional factors, besides HY5, have been
identified(Osterlund et al., 1999; Holm et al., 2001).
It is generally assumed that light-induced photomor-phogenic
development requires the inactivation of theseCOP/DET/FUS proteins.
However, little is known as to howthe light-activated
photoreceptors (including phy-tochromes) regulate the activities of
those downstreamCOP/DET/FUS proteins to bring about the
physiologicalresponses. At least one of the mechanisms through
whichphytochrome signaling regulates the downstreamCOP/DET/FUS
protein activities is achieved by triggeringthe nuclear depletion
of COP1 under their respective light-responsive regimes (FR and R
for phyA and phyB, respec-tively) (Osterlund and Deng, 1998). The
finding that thekinetics of GUS-COP1 nuclear depletion by FRc is
alsoaffected by other phyA-specific signaling mutations(including
phyA, fhy3, fhy1, far1, and fin219) suggests thatthose
phyA-specific signaling components functionupstream of COP1 and are
involved in regulating thenucleo-cytoplasmic partitioning of COP1.
As a result, theaccumulation of the bZIP transcription factor HY5
and, inturn, the degrees of photomorphogenic development ofthese
mutants are affected (Wang and Deng, 2001).However, it should be
pointed out that there is evidencethat these loci entail additional
signaling pathways besidessignaling through COP1 (Wang and Deng,
2001).
PHYTOCHROME REGULATION OF NUCLEAR GENEEXPRESSION
Transcriptional regulation of gene expression representsan
important step in the control of various processes ofplant growth
and development. The photoregulation ofgene expression in higher
plants has been extensivelystudied during the past two decades and
a number of pho-
Figure 9. The cellular basis and proposed biochemicalmode of
COP1 function.(A) Light-regulated nucleocytoplasmic partitioning
ofCOP1. In darkness, COP1 is enriched in the nucleus tosuppress
photomorphogenic development. Light signals(perceived by
phytochromes and other photoreceptors)trigger the nuclear depletion
of COP1, thus abrogatingthe repressive effect of COP1. Note that
the COP9 sig-nalosome may contribute to the control of nuclear
local-ization of COP1 or the stability of COP1 in the nucleus.The
roles of DET1 and COP10 in this cellular processhave not been
determined.(B) A putative role of COP1 as an E3 ubiquitin
ligase.COP1 mediates the ubiquitination of HY5 and its subse-quent
degradation via the proteasome by recruiting an E2and HY5 through
distinct interacting domains (RING-fin-ger and WD-40 repeat domain,
respectively, adaptedfrom Osterlund et al., 2000).
-
The Arabidopsis Book 20 of 30
toregulated genes have been identified from several
plantspecies. For example, the transcript level of both theLHCB
genes for light-harvesting chlorophyll a/b-bindingprotein of
photosystem II and the RBCS genes for ribu-lose-1, 5-biphosphate
carboxylase/oxygenase small sub-unit increases by light
illumination, and this light-induciblegene expression is mediated
by phytochromes.Consequently, the promoter region of LHCB and
RBCSgenes have been isolated and analyzed to identify
variouscis-regulatory elements responsible for light- or
phy-tochrome regulation, such as the GT-1 boxes of RBCS, G-box of
tomato RBCS-3A, GATA motif, 3AF binding sitesand AT-1 binding site.
In addition, trans-acting factorswhich bind to light-responsive
cis-regulatory elements(LREs) such as GT-1, GBF, GAF1, 3AF-1 and
AT-1 were iso-lated, and some of which have been shown to be
involvedin phytochrome-mediated light responsiveness (von Arnimand
Deng, 1996; Terzaghi and Cashmore, 1995; Menkenset al., 1995; Kuno
et al., 2000; Kuno and Furuya, 2000).Further, it was suggested that
the combinatorial interactionof multiple LREs is the key
determinant for mediating lightcontrol of promoter activity (Puente
et al., 1996). However,these traditional studies only revealed
limited informationon the role of individual phytochromes in the
photoregula-tion of gene expression.
Recently, the newly developed gene chip technology(Richmond and
Somerville, 2000) has been applied tostudy light regulation of gene
expression and to define theroles of some individual phytochromes
in gene regulationunder their respective light regimes (Ma et al.,
2001). It wasrevealed that a large number of genes, possibly over
onethird of the genome, are coordinatedly regulated by vari-ous
light signals, including the red and far-red lights thatare
primarily perceived by phytochromes. Utilization of thephyA and
phyB null mutants under specific light conditionsconfirmed the
roles of phyA and phyB in mediating the far-red and red light
regulation of genome expression. phyAseems to be the primary
photoreceptor for mediating far-red light reglation of gene
expression, whereas phyB isonly one of the phytochromes mediating
red light regula-tion of genome expression (Ma et al., 2001).
The genome profiling study revealed an interesting fea-ture of
light-regulated gene expression: many cellularmetabolic and
regulatory pathways are found to be coor-dinately regulated by
light. Some of them (such as all pho-tosynthetic genes, glycolysis
and the TCA cycle etc.) areactivated by light, whereas others (such
as cell wall-loos-ening enzymes and water channel protein
aquaporins) arerepressed by light (Ma et al., 2001). These results
substan-tiate the notion that light-regulated plant
developmentinvolves a coordinated regulation of different
pathways.Similar conclusions have also been drawn from studies
oncircadian clock control of gene expression (Harmer et al.,2000;
Schaffer et al., 2001).
AN EMERGING INTEGRATED PICTURE OFPHYTOCHROME SIGNALING
Among these fascinating achievements on phytochromeresearch, two
key discoveries have dramatically advancedour understanding of
phytochrome signaling. First, it hasbeen demonstrated that upon
photoconversion of Pr toPfr, phytochromes (both phyA and phyB) can
translocatefrom cytoplasm into the nucleus (Kircher et al., 1999;
Nagyet al., 2000). Second, the Pfr form of phytochromes hasbeen
demonstrated to interact with a light-responsive ele-ment (G
box)-bound bHLH protein, PIF3 (Ni et al., 1999;Martínez-Garcia et
al., 2000). These findings suggest anemerging model for phytochrome
signaling, whereby phy-tochromes perceive light, enter the nucleus,
interact withtranscriptional regulators, and thus regulate gene
tran-scription. It is conceivable that the direct targeting of
lightsignals to a promoter-element bound transcription factorwould
allow plants to continuously monitor their light envi-ronments and
to react to changes in light availability byconcomitant changes in
light-regulated gene expression.However, it should be pointed out
that the biochemicalbasis for the regulation of gene expression
brought aboutby the interaction of phytochrome with PIF3 is not
yetknown. By binding to the G-box-bound PIF3, phytochromecould
regulate transcription either: (a) directly, by function-ing as a
transcriptional coactivator or corepressor involvedin recruitment
or modulation of the pre-initiation complex,or (b) indirectly, by
biochemically or allosterically alteringthe intrinsic
transcriptional regulatory activity of PIF3(Quail, 2000).
Particularly, with the knowledge that theactivities of a number of
transcription factors are regulatedby phosphorylation (Hardtke et
al., 2000) and that phy-tochrome is a light-regulated kinase, it
will be interesting totest whether PIF3 is a substrate of
phytochrome kinaseactivity and how it may modulate the function of
PIF3.
A Class of bHLH Proteins Operate in PhytochromeSignaling
It should be noted that PIF3 could bind both phyA andphyB,
although it binds phyB with a higher apparent affin-ity (Zhu et
al., 2000). Also, transgenic Arabidopsis plantswith altered PIF3
levels have a markly greater effect onphyB- than on phyA-regulated
photoresponses (Ni et al.,1998; Halliday et al., 1999). These
results suggest thatPIF3 has a dominant role in phyB signaling, but
a moreminor role in phyA signaling. The finding that PIF3 is a
-
Phytochrome Signaling Mechanism 21 of 30
member of the bHLH family of transcription factors alsosuggests
that phytochromes might directly regulate adiverse set of genes
through an integrated transcriptionalnetwork orchestrated by an
array of interacting bHLH pro-teins. This is suggested by the
extensive size of this pro-tein family in Arabidopsis and the known
nature of bHLHproteins that are able to form homodimers and/or
het-erodimers. Heterodimers can have DNA-binding sitespecificity
different from that of their respective homod-imers, and they could
also function as convergence pointsfor integrating upstream signal
inputs specific to each ofbHLH proteins present in the heterodimers
(Quail, 2000,Figure 10A). Such a transcriptional network generated
bycombinatorial heterodimeric interactions of different mem-bers of
the bHLH family should explain the distinct yet
overlapping transcriptional profiles mediated by phyA andphyB,
respectively (Ma et al., 2001). Recent studies in Dr.Quail’s group
have provided new genetic and molecularevidence to support such a
model. They found two addi-tional bHLH proteins involved in
phytochrome signaling.One of them, PIF4, was isolated as a PIF3
interacting pro-tein and it also interacts specifically with the
Pfr form ofphyB, but not phyA. PIF4 could form homodimer with
itselfor heterodimer with PIF3, and both forms can bind to
thePIF3-binding promoter element. Missense mutants ofphyB that are
impaired in signaling show reduced bindingto PIF4, suggesting a
biologically relevant interaction.Overexpressing PIF4 in transgenic
Arabidopsis causes ahyposensitive phenotype specific to red light.
Conversely,antisense PIF4 plants are hypersensitive to red
light.
Figure 10. A molecular model depicting phytochromecontrol of
gene expression.(A) An illustration of the transcriptional
diversity generat-ed by combinatorial heterodimeric interactions of
thebHLH family members and their selective interactionswith
different phytochrome members. As shown, bHLH1interacts with phyA
only, whereas bHLH2 binds phyBonly. The homodimeric and
heterodimeric combinationsof the bHLH family members may target
different G-boxvariants in the promoters of a variety of
light-responsivegenes and regulate their expression (adapted from
Quail,2000).(B) An emerging integrated picture of phytochrome
sig-naling. Upon photoconversion to the Pfr form, phy-tochromes
translocate into the nucleus, where they inter-act with the
G-box-bound PIF3 and activate the expres-sion of the primary target
genes (such as the MYB classtranscription factors CCA1 and LHY).
The encoded pri-mary target gene products (many of them are
transcrip-tional regulators) in turn are responsible for
orchestratingthe expression of the secondary target gene
expression(such as the induction of LHCB expression by CCA1),thus
generating a transcriptional network controlling dif-ferent aspects
of phytochrome physiology. Other phy-tochrome-interacting partners
(PIFs) and their signalingintermediates may function as modifiers
to participate insuch a direct light signaling pathway. Moreover,
COP1and the COP9 signalosome could also interact (directly
orindirectly) with the transcriptional machinery, PIFs orother
signaling intermediates, and regulate their abun-dance through a
light-regulated proteolysis process. Notethat far-red light
converts the active Pfr form to the inac-tive Pr form, abolishing
its interaction with PIF3 and turn-ing off the transcriptional
cascade. R: red light; FR: far-red light; Pr: inactive, Pr
conformer of phytochromes; Pfr:active, Pfr conformer of
phytochromes; PIC: pre-initiationcomplex; TATA: TATA box.
-
The Arabidopsis Book 22 of 30
Further, a T-DNA knockout mutant of PIF4, srl2, was
inde-pendently identified from a genetic screen as a
hypersen-sitive mutant under red light. Together, these studies
pro-vided compelling evidence that PIF4 is an authentic
phyBsignaling component and that it acts early in the
signalingpathway. The mechanism for PIF4 attenuation of phyB
sig-naling is not yet elucidated. One possibility, however, isthat
the PIF3/PIF4 heterodimer might possess reducedtranscriptional
activity compared to the PIF3 homodimer.Alternatively, the PIF4
homodimer may compete with PIF3in binding the biologically active
Pfr form of phyB in thenucleoplasm, thus titrating out the
available phyB for inter-acting with PIF3 (Huq and Quail,
2001).
On the other hand, HFR1, another bHLH protein, isgenetically
identified as a positive regulator specific tophyA signaling, and
it is capable of forming a homodimeras well as heterodimerizing
with PIF