Integration of light and abscisic acid signaling duringseed germination and early seedling developmentHao Chen*, Jingyu Zhang†, Michael M. Neff†‡, Suk-Whan Hong§, Huiyong Zhang¶, Xing-Wang Deng¶,and Liming Xiong*�

*Donald Danforth Plant Science Center, St Louis, MO 63132; †Department of Biology, Washington University, St. Louis, MO 63130; ‡Department of Crop andSoil Sciences, Washington State University, Pullman, WA 99164; §Department of Applied Plant Science, Chonnam National University, Gwangju 500-757,South Korea; and ¶Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT 06520

Edited by Maarten Koornneef, Wageningen University and Research Centre, Wageningen, The Netherlands, and approved January 31, 2008(received for review November 13, 2007)

Seed germination is regulated by endogenous hormonal cues andexternal environmental stimuli such as water, low temperature,and light. After germination, the young seedling must rapidlyestablish its root system and the photoautotrophic capabilityappropriate to its surrounding environment. Light and the phyto-hormone abscisic acid (ABA) both regulate seed germination andseedling development, although how light and ABA signals areintegrated at the molecular level is not understood. Here, we foundthat the previously described light-signaling component HY5 alsomediates ABA response in seed germination, early seedlinggrowth, and root development in Arabidopsis. HY5 binds to thepromoter of the transcription factor ABI5 gene with high affinityand is required for the expression of ABI5 and ABI5-targeted lateembryogenesis-abundant genes in seeds. Chromatin immunopre-cipitation also indicated that the binding of HY5 to the ABI5promoter is significantly enhanced by ABA. Overexpression of ABI5restores ABA sensitivity in hy5 and results in enhanced lightresponses and shorter hypocotyls in the wild type. Our studiesidentified an unexpected mode of light and ABA signal integrationthat may help young seedlings better adapt to environmentalstresses.

light response � signal transduction

The plant hormone abscisic acid (ABA) plays essential rolesin several aspects of plant growth and development, includ-

ing seed maturation, seed dormancy, and adaptation to envi-ronmental stresses such as drought and high salinity (1–4).Genetic studies of ABA regulation of seed germination and geneexpression have identified a number of Arabidopsis mutants withaltered ABA sensitivities. One of the ABA insensitive mutants,abi5, was isolated for its ability to germinate in the presence ofhigher concentrations of exogenous ABA (5). ABI5 encodes abasic leucine zipper (bZIP) transcription factor whose accumu-lation inhibits seed germination and early seedling establishment(6–8). ABI5 regulates the expression of ABA induced, mostlyseed-specific, AtEM genes that encode class I late embryogen-esis-abundant (LEA) proteins important for seed maturation (6,9, 10). Interestingly, the ABI5 gene was also expressed in youngseedlings, suggesting that ABI5 may have additional functionsbeyond seed development and germination (11). It is known thatseed germination and early seedling development are alsoregulated by light (12). Yet little is known how light and ABAsignals are integrated in the regulation of this commondevelopmental transition from seeds to seedlings.

Whereas the role of ABA in seedling photomorphogenesis isunclear, ABA is well recognized for its critical roles in plantadaptation to drought stress. In addition to inducing stomatalclosure and the activation of stress-responsive genes, ABA wasrecently found to be involved in lateral root development(13–15). Genetic studies indicated that the inhibition of lateralroot elongation and potential promotion of primary root growthby drought stress and ABA are adaptive responses of roots to

drought stress that are linked to whole plant drought tolerance(15). To reveal the molecular mechanisms underlying this adap-tive response, we investigated genes potentially involved in thisprocess by studying ABA response in Arabidopsis mutants withaltered lateral root development. One of the mutants we exam-ined is hy5, a mutant originally isolated for its long hypocotylunder light (16) but also with more elongated lateral roots (17).The HY5 gene encodes a bZIP transcription factor and isresponsible for the activation of many light-regulated genes (18).We found that the hy5 mutant is tolerant to the inhibitory effectof ABA on the elongation of lateral roots, seedling growth, andseed germination. Furthermore, the steady-state transcript levelsof the ABI5 and its target genes were greatly reduced in hy5seeds. HY5 protein interacts with the ABI5 promoter, and the invivo interaction was enhanced by ABA. Ectopic expression ofABI5 in the hy5 mutant background restored its ABA sensitivityand overexpression of ABI5 in the wild-type background-enhanced light response. These results indicate that HY5 inte-grates both ABA and light signal transduction pathways partiallythrough direct activation of ABI5.

ResultsHY5 Mediates ABA Response in Roots and Is Required for OsmoticStress Tolerance in Seedlings. To determine whether HY5 isinvolved in ABA inhibition of lateral root development, wetested the response of a hy5 knockout mutant (SALK�096651;see Materials and Methods) to ABA. At either 0.5 or 1.0 �MABA, the growth of the wild-type seedlings was inhibited in thatthe primary root elongation and shoot growth were slower thanthat of the control plants, particularly at the 1.0-�M ABA level(Fig. 1 A and B). In contrast, the hy5 seedlings were much moretolerant to ABA in both primary root elongation and shootgrowth (Fig. 1 A and B). After 3-week growth, the shoot freshweights of the wild-type seedlings reduced by 30% and 60% with0.5- and 1.0-�M ABA treatment, respectively; whereas nostatistic difference in shoot fresh weights of hy5 seedlings wasfound among the treatments. As expected, ABA inhibits lateralroot growth in wild-type plants (Fig. 1 A and C) and the averagetotal length of lateral roots decreased 80% by even 0.5-�M ABAtreatment (Fig. 1C). However, the growth of lateral roots of hy5was much less inhibited by ABA (Fig. 1 A and C). Thus, hy5mutant seedlings are much less sensitive to ABA both in shootand root growth than the wild type.

Author contributions: H.C., M.M.N., and L.X. designed research; H.C., J.Z., and S.-W.H.performed research; H.Z. and X.-W.D. contributed new reagents/analytic tools; H.C., J.Z.,M.M.N., and L.X. analyzed data; and H.C. and L.X. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

�To whom correspondence should be addressed. E-mail: lxiong@danforthcenter.org.

This article contains supporting information online at www.pnas.org/cgi/content/full/0710778105/DC1.

© 2008 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0710778105 PNAS � March 18, 2008 � vol. 105 � no. 11 � 4495–4500

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

June

24,

202

0

Given that ABA mediates many aspects of abiotic stressresponse, we examined whether hy5 seedlings are compromisedin stress tolerance. We first checked the sensitivity of hy5 toosmotic stress. When transferred to media supplemented with 50or 100 mM mannitol, the shoot and root growth of hy5 was moreseverely retarded than that of the wild type (Fig. 1D). We thentested whether hy5 mutants are sensitive to ionic stress as well.We found that hy5 seedlings have slower growth rates of theprimary roots and also yellowish leaves under salt stress (Fig. 1Eand data not shown). Thus, hy5 mutants are more sensitive to theinhibition of ionic stress and osmotic stress. Interestingly,whereas NaCl inhibits the growth of lateral roots in the wild type,this inhibition is significantly reduced in hy5 seedlings (Fig. 1F).This observation implicates that ABA may play a more impor-tant role in NaCl inhibition of lateral root growth than in theinhibition of primary root growth.

Germination of hy5 Mutant Seeds Is More Tolerant to ABA andGlucose. On media supplemented with 0.5 or 1.0 �M ABA, it wasfound that hy5 seeds germinated more rapidly and at a higherrate than the wild type, although wild-type seeds eventuallygerminated at these ABA levels (data not shown). In the

presence of ABA, the germinated wild-type embryos werearrested, whereas those of hy5 were able to develop into normalgreen seedlings (Fig. 2B). This is not an indirect consequence ofearlier seed germination of hy5 than the wild type, because thegerminated wild-type embryos failed to develop into normalgreen seedlings even after prolonged incubation under the sameconditions.

Seed germination is also sensitive to salt stress and sugars.These stimuli inhibit germination partly through enhanced bio-synthesis of and responsiveness to ABA. Consistent with theirresistance to ABA during germination, hy5 seeds have signifi-cantly higher germination rates than the wild type at either 150mM NaCl or 333 mM glucose (Fig. 2 C and D).

HY5 Is Required for ABA-Inducible Gene Expression in Seeds andDuring Seed Germination. The impaired ABA responses of hy5mutants in roots and seed germination suggest that HY5 mayplay unexpected roles in ABA signal transduction. The fact thatHY5 is a DNA-binding transcription factor (18) prompted us toask whether HY5 regulates ABA-inducible gene expression.Although hy5 seedlings are more sensitive to osmotic stress (Fig.1D), no clear difference in gene expression was detected betweenyoung seedlings of hy5 and the wild type after ABA or osmoticstress treatments for eight known ABA up-regulated genesexamined. These genes include RD29A, RD22, DREB2A,DREB2B, CBF1, CBF2, CBF3, and ABF3 (data not shown). Anexception is RD29B, whose transcript level was lower in hy5 thanin the wild type under the ABA treatment (Fig. 3A). We thenwent on to investigate whether the expression of ABA-regulatedseed-specific genes is altered in hy5. The transcription factorsABI3 and ABI5 are known to regulate ABA sensitivity duringand immediately after seed germination. We thus checked thetranscript levels of ABI3 and ABI5 in dry seeds. Although verylittle difference in ABI3 transcript level was detected, ABI5

Fig. 1. hy5 seedlings are less sensitive to ABA inhibition of lateral rootgrowth and are more susceptible to stress. (A) Morphology of wild-type andhy5 seedlings on plates without or with 0.5 or 1.0 �M ABA. Five-day-oldseedlings grown on a half MS medium were transferred to half MS mediasupplemented with or without ABA. The pictures were taken 5 days after thetransfer. (B) Relative root length of wild-type and hy5 seedlings at 5 days afterbeing transferred to plates with the indicated concentrations of ABA. (C )Length of total lateral roots of two-week-old wild-type and hy5 seedlings asshown in A. (D) Morphology of wild-type and hy5 seedlings on plates withoutor with 50 or 100 mM mannitol. The pictures were taken 3 weeks afterseedlings were transferred to the shown plates. (E) Relative root length ofwild-type and hy5 seedlings at 8 days after being transferred to plates with theindicated concentrations of NaCl. (F) Length of lateral roots of the wild-typeand hy5 seedlings on media supplemented with the indicated concentrationsof NaCl. Data in B, C, E, and F represent means and SE (n � 6).

Fig. 2. hy5 seeds are less sensitive to the inhibition of ABA, salt, and sugaron germination and seedling growth. (A and B) hy5 is less sensitive to ABAinhibition of seed germination and seedling growth. Morphology of seedlingsin the absence (A) or presence (B) of 1.0 �M ABA. The pictures were taken 12days after the seeds were incubated at 22°C. (C) Germination rates of wild-typeand hy5 seeds on 150 mM NaCl. (D) Germination rates of wild-type and hy5seeds on 333 mM glucose. After sowing on the respective plates, seeds wereincubated at 4°C for 3 days before being placed at 22°C under constant whitelight for germination. Seeds with emerged radicals were counted as germi-nated. Numbers are averages of two independent experiments with �100seeds for each genotype.

4496 � www.pnas.org�cgi�doi�10.1073�pnas.0710778105 Chen et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

0

mRNA level was greatly reduced in hy5 seeds (Fig. 3B). Subse-quently, we monitored the transcript levels of three LEA genes,RAB18, AtEM1, and AtEM6, in hy5 seeds. AtEm1 and AtEm6 are

likely the direct target genes of ABI5, and RAB18 transcript levelis down-regulated in abi5 mutants (10, 19, 20). Consistent withthe reduction in ABI5 transcript level, the steady-state levels ofall of these ABI5-regulated genes are significantly down-regulated in hy5 seeds (Fig. 3B).

HY5 Protein Binds to the ABI5 Promoter with High Affinity andSpecificity. Because ABI5 and its target genes are down-regulatedin hy5 seeds (Fig. 3), and hy5 has reduced ABA sensitivityreminiscent of the abi5 mutant (Fig. 2), we speculated that HY5may directly regulate ABI5 transcription. HY5 is known to bindto the G-box CACGTG in light-regulated gene promoters (18).In searching the ABI5 promoter, four G-box core sequences werefound in the 1.7-kb fragment upstream of the translation startcodon (Fig. 3C). To test whether any of these G-boxes is involvedin HY5 binding, we did a gel retardation assay using Escherichiacoli expressed recombinant HY5 protein together with the fourABI5 promoter fragments. The P1 and P4 fragments each harborone G-box core, and the P2 contains two G-box cores, whereasP3 does not contain any consensus G-box motifs. RecombinantHY5 protein retarded the mobility of P1, P3, and P4, but not P2(Fig. 3 D and E). Among the HY5 retarded fragments, P1 andHY5 gave the strongest signal intensity (Fig. 3D). Competitionassays were performed with labeled P1 and unlabeled P1 orunlabeled P4. With increasing unlabeled P1, the P1–HY5 com-plex quickly diminished (Fig. 3F). With unlabeled P4, however,even at a concentration 300-fold more than the labeled P1, theshifted band was still present (Fig. 3F). Therefore, HY5 doesbind to ABI5 promoter with the highest affinity in the P1 regionfrom �1,754 to �1,294.

To confirm these in vitro binding results, we performed ChIPassays using HA-tagged HY5 transgenic seedlings (21). Fig. 3Gshows that consistent with the gel retardation data, anti-HAantibody specifically precipitated the ABI5 promoter P1 frag-ment, but not P2, P3, or P4. Interestingly, ABA treatmentsignificantly enhanced the binding of HY5 protein with the P1region of the ABI5 promoter. In contrast, ABA failed to enhancethe binding of HY5 to the promoters of two other light-regulatedHY5 target genes, CHS and RbcSA (Fig. 3G). This result wasfurther confirmed by quantitative real-time PCR. Comparedwith no enrichment of the CHS promoter, a 3.1-fold (P � 0.05)enrichment of the P1 fragment was detected in ABA-treatedrelative to ABA-untreated samples. Because ABA does notenhance HY5 expression nor increases HY5 protein stability(data not shown), the specific promotion of HY5 binding to theABI5 promoter may underlie the mechanism for ABA up-regulation of ABI5 expression (21).

To examine the effect of HY5 on ABI5 transcription, we dida transient transcription assay using tobacco BY-2 protoplasts.However, the ABI5 promoter-driven �-glucuronidase (GUS)activity is very limited in this system, and HY5 protein did notexhibit a significant effect on ABI5 transcription (data notshown). Although almost similar levels of HY5 transcript can bedetected in all of the tissues examined, including mature leaves(17), high expression of ABI5 is detected mainly during seedmaturation and at the early stage of seedling development(6, 11). Therefore additional seed or early seedling-specificfactors may be required to activate ABI5 transcription togetherwith HY5. Indeed, despite no evidence for direct regulation,ABI5 mRNA accumulation is altered in abi3, fus3, and lec1mutants (6, 22).

ABI5::GUS Activity in hy5 Is Reduced and Is Less Responsive to StressInduction. To address whether there is any tissue-specific regu-lation of ABI5 transcription by HY5, we made an ABI5::GUSconstruct consisting of a 1.7-kb fragment of the ABI5 promoter-driven GUS gene and transferred it into wild-type Col-0 and hy5.Sixteen independent transformants from each background were

Fig. 3. HY5 regulates the expression of a subset of ABA-inducible genes inyoung seedlings or dry seeds and binds to ABI5 regulatory sequences with highaffinity. (A) RD29B transcript levels in response to ABA treatment. Two-week-oldseedlings were sprayed with 100 �M ABA and incubated under light for 0.5 or 2 hbefore RNA extraction. Ten �g of the total RNA was loaded per lane. A �-tubulingene was used as a loading control. (B) Transcript levels of the indicated genes indry seeds of Col-0, hy5, abi5 (Ws background), and Ws. The 18s rRNA gene wasused as a loading control. (C–F ) Binding of HY5 to ABI5 regulatory sequences. (C )Schematic of the ABI5 promoter. The black filled box represents the promoterregion, gray filled box stands for intron, box with lines represents the 5� UTR, andemptyboxes standforG-boxes. (DandE)Gel-retardationassaysofABI5promoterP1 and P2 (D), P3 and P4 (E) with 150 �g GST (lanes 2 and 7) or increased amounts(50, 100, 150 �g) of GST-HY5 protein (lanes 3–5 and 8–10). Lanes 1 and 6 arelabeled probes alone to show the position of the free probes. The DNA–proteincomplexes are indicated by arrowheads. (F ) Competition gel-retardation assay ofP1 with cold P1 or P4 in the presence of 100 �g of GST-HY5. Lane 1 is labeled P1alone, and lane 2 is the labeled P1 with 100 �g of GST-HY5. Increased amounts(30�, 300�) of unlabeled P1 or P4 were added to lanes 3 and 4 and lanes 5 and6,respectively.TheDNA–proteincomplexesare indicatedbyarrowheads. (G)HY5binds to the P1 motif of ABI5 in vivo and ABA specifically enhances the binding.Shown are anti-HA ChIP assays with Col-0 and HA-HY5 transgenic plants with orwithout 0.5 �M ABA treatment. P1–P4 are fragments of the ABI5 promoter.RbcSA and CHS are light-regulated HY5 target genes. Input control: genomicDNA not subject to immunoprecipitation. Mock control: ChIP without HAantibody.

Chen et al. PNAS � March 18, 2008 � vol. 105 � no. 11 � 4497

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

June

24,

202

0

examined for GUS activity, and similar GUS expression wasfound among different lines in the same background. Data fromone representative line for each background are presented in Fig.4. Similar to what was previously reported (11), ABI5 promoteractivity was detected in cotyledons, hypocotyls, and roots ofwild-type young seedlings (Fig. 4A). On the contrary, theABI5::GUS activity was dramatically decreased in hy5 in all thesetissues (Fig. 4B), which is also similar to what was seen with otherpreviously characterized HY5 target genes in hy5 mutants (18).

As a critical regulator of photomorphogenesis, the HY5protein level is tightly controlled by the COP1 E3 ubiquitin ligasein the dark, such that the HY5 protein accumulates only in thelight (23). Therefore, if HY5 is essential for ABI5 transcription,dark-grown seedlings would have reduced ABI5::GUS activity.As expected, ABI5::GUS activity is much lower in wild-typeetiolated seedlings (Fig. 4 C and M). ABI5 transcripts could bedetected in floral tissues (24), which is coincident with the highaccumulation of HY5 protein in floral tissues (25). Consistently,ABI5::GUS activity is barely detectable in hy5 f loral tissues, incontrast to the strong GUS staining in the wild type (Fig. 4 D andE). Similarly, the GUS staining is weaker in hy5 siliques than inwild-type siliques (Fig. 4F).

The ABI5 promoter can be induced by ABA and salt stress(11). To investigate whether HY5 is also required for ABI5induction by ABA and stress, we treated the seedlings with either10 �M ABA or 100 mM NaCl for 1 day before staining orquantifying for GUS activity. Fig. 4 G–L and N show that withboth treatments wild-type seedlings display strong induction,whereas hy5 seedlings have only a limited augment of ABI5::GUSactivity. These results indicate that the induction of ABI5 by bothdevelopmental cues and abiotic stresses requires HY5, which isconsistent with the compromised stress tolerance of hy5 seed-lings (Fig. 1).

Overexpression of ABI5 Confers ABA Sensitivity to hy5 and InhibitsHypocotyl Elongation. If ABI5 is a direct target of HY5, one wouldexpect that overexpression of ABI5 may be able to rescue someof the hy5 mutant phenotypes such as its reduced sensitivity toABA. The ABI5 overexpression lines were generated in bothCol-0 and hy5 background (Fig. 5A). On medium without ABA,no obvious difference in growth between hy5 and hy5 trans-formed with 35S::ABI5 was noticed (Fig. 5B). With ABA sup-plement, however, seedlings of hy5 transformed with 35S::ABI5exhibited sensitive responses to ABA in that their seedlingdevelopment, seed germination, and primary root elongation aresignificantly inhibited relative to hy5 mutants (Fig. 5 C–E). Thus,expression of ABI5 is able to confer ABA sensitivity to hy5during seed germination and seedling growth. Nonetheless, theinsensitivity to ABA in lateral root growth was not rescued in hy5mutant overexpressing ABI5 (data not shown), suggesting thatthere are additional components controlling ABA response inlateral root development.

Because HY5 regulates the expression of ABI5, we askedwhether ABI5 plays any role in light signal transduction. Theabi5 mutant and 35S::ABI5 transgenic plants were hence inves-tigated for their response to light. Under different light regimessuch as blue light, red light, far-red light, or dark, no significantdifference between hypocotyl lengths of abi5 and the wild typewas observed (data not shown), suggesting that the absence ofABI5 alone does not compromise photomorphogenesis re-sponse. However, when ABI5 was overexpressed in the wild-typebackground, although there was no clear phenotypic alterationsin the dark, these transgenic plants exhibited a significantlyenhanced response to blue, red, and far-red light in repressinghypocotyl elongation (Fig. 5 F–H), which is consistent with theaccumulation of HY5 under a broad spectrum of light (26).Nonetheless, overexpression of ABI5 in hy5 was not sufficient torestore the light response of hy5, indicating that besides ABI5,

Fig. 4. Reduced ABI5::GUS activity in hy5 mutants. (A–C) Two-day-old seedlings of light-grown Col-0 (A), light-grown hy5 (B), and dark-grown Col-0 (C). (D andE) Floral tissues of Col-0 (D) and hy5 (E). (F ) Siliques of Col-0 (Left) and hy5 (Right) 14 days after pollination. (G–L) Four-day-old seedlings of Col-0 and hy5 without(G and J ) or with 100 mM NaCl (H and K) or 10 �M ABA (I and L) treatment. (M) GUS activity (arbitrary units) of light- or dark-grown Col-0 and hy5 seedlings.(N) GUS activity (arbitrary units) of Col-0 and hy5 without or with 10 �M ABA or 100 mM NaCl treatment. Data in M and N are means and standard errors of threeindependent experiments.

4498 � www.pnas.org�cgi�doi�10.1073�pnas.0710778105 Chen et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

0

other HY5 target genes are also required for the inhibition ofhypocotyl elongation during photomorphogenesis.

DiscussionBased on the insensitivity of hy5 mutants to light inhibition ofhypocotyl elongation, HY5 was defined as a positive regulatordownstream of diverse photoreceptors in the light signal trans-duction pathway (16, 27, 28). Interestingly, hy5 mutants also havemore lateral roots and these lateral roots grow more horizontally(17). Because drought stress and ABA inhibit lateral rootelongation (13–15), we examined whether HY5 is involved inABA-mediated inhibition of lateral root growth. Whereas ABAclearly represses lateral root growth in the wild type, it is muchless effective in hy5. Thus, wild-type HY5 is required for thisABA response (Fig. 1 A and C). Furthermore, the hy5 mutantseeds are less sensitive to the inhibitory effect of ABA, high salt,or glucose on seed germination (Fig. 2), indicating that HY5partly mediates these ABA responses as well. Perhaps as a result

of the impaired ABA response, hy5 seedlings were more sensitivein growth to salt and osmotic stress (Fig. 1 D and E). Nonethe-less, hy5 leaves did not exhibit any transpiration defects and theguard cell ABA responsiveness is unaltered (data not shown),which implies that either HY5 does not modulate ABA-inducedstomata movement or that HY5 is not involved in the ABAresponse of mature plants.

Besides its role in photomorphogenesis, recent research alsosuggests that HY5 is involved in the activation of a subset ofUV-B responsive genes. One of the down-regulated genes in hy5under UV-B stress is DREB2A (29), a well characterized ABA-and drought-inducible gene. Our results also demonstrate thathy5 mutants are defective in the accumulation of LEAs and ABI5mRNAs in mature seeds and ABA induction of RD29B in youngseedlings (Fig. 3 A and B). Given that HY5 and its homolog HYHredundantly regulate target genes’ transcription (30, 31), it is notunexpected that only a limited number of ABA-responsive geneswere altered in their expression in hy5 young seedlings (Fig. 3Aand data not shown). Indeed, transcript levels of several ABA-inducible genes including a dehydrin, a MYB-related gene,RD20, COR47, and LTI29 are greatly reduced in hy5-hyh doublemutant compared with single mutants where little reduction inthe transcript levels of these genes was detected (30). Thealteration of ABA responsive genes’ transcription in hy5 orhy5-hyh mutant strongly suggests that HY5 may play critical rolesin ABA signaling.

Intriguingly, genes with diminished accumulation in hy5 seedsand young seedlings appear to be either direct or indirect targetsof ABI5. In previous studies, ABI5 protein was shown to be ableto activate the promoters of AtEM1, AtEM6, and RD29B intransient transcription assays (6, 32). These genes may be thedirect targets of ABI5. On the other hand, although there was nodirect evidence for the activation of RAB18 transcription byABI5, RAB18 mRNA level is down-regulated in abi5 mutant(19). Coincident with reduced transcript levels of these ABI5target genes, the steady-state mRNA level of ABI5 was signifi-cantly reduced in hy5 seeds (Fig. 3B). Moreover, ABI5::GUSactivities were much lower in hy5 than in wild type under bothnormal or stress conditions (Fig. 4). Therefore, the decreasedtranscript levels of these ABA-responsive genes in hy5 are mostlikely caused by the down-regulated ABI5 expression in themutant. The fact that ABA insensitivity of hy5 mutants could berestored by ectopically expressing ABI5 (Fig. 5 B–E) supportsthis notion. Importantly, HY5 can bind the ABI5 promoter invitro and in vivo with high specificity (Fig. 3 D–G). This result isconsistent with a recent ChIP–Chip study that showed that theABI5 promoter was among the putative binding sites of HY5(21). ABA enhancement of HY5 binding to the ABI5 promoter(Fig. 3G) further suggests that HY5 is very likely a transcrip-tional activator of ABI5 expression in response to developmentaland environmental cues.

Recently, it was reported that seeds in shallow soil could senselight through phytochrome B to down-regulate ABI3 and pro-mote hypocotyl growth (33). Our results demonstrate thatoverexpression of ABI5 inhibits hypocotyl elongation under blue,red, or far-red light (Fig. 5 F–H). Interestingly, the rpn10 and kegmutants, which stabilize the ABI5 protein, exhibit a decreasedhypocotyl growth as well (34, 35). Thus, ABA may enhanceseedling light response and inhibit hypocotyl elongation, whichmay confer increased resistance to drought and other stresses. Infact, drought and ABA inhibition of hypocotyl elongation hasbeen documented, although the mechanism was previouslyunknown (36). Recent studies showed that light could affectABA metabolism in germinating seeds (37, 38). Our studydemonstrates that light also positively regulates ABA signalingthrough the HY5-ABI5 regulon that may facilitate seedlingestablishment under abiotic stress.

Fig. 5. Overexpression of ABI5 restores ABA sensitivity in hy5 and enhanceslight response in the wild type. (A) ABI5 transcript levels in the ABI5 overex-pression lines of Col-0 and hy5. (B and C) Morphology of 7-day-old seedlingsoverexpressing ABI5 in the wild-type or hy5 background without (B) or with(C) 1.0 �M ABA. (D) Germination rates of ABI5 overexpression lines in wild-type and hy5 backgrounds. Data are means from three independent experi-ments with �100 seeds for each genotype. (E) Primary root length of ABI5overexpression lines in the wild-type or hy5 background under the indicatedconcentrations of ABA (0, 0.5, 1.0, and 2.5 �M). Data are means and standarderrors (n � 10). (F–H) Hypocotyl lengths of seedlings of Col-0 and Col-0transformed with 35S::ABI5 under the indicated fluence rates of blue (F ),far-red (G), or red light (H). Data are means and standard errors (n � 30).

Chen et al. PNAS � March 18, 2008 � vol. 105 � no. 11 � 4499

PLA

NT

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

June

24,

202

0

Materials and MethodsPlant Materials, Plasmid Constructions, Plant Transformation, and GrowthConditions. Arabidopsis thaliana ecotype Columbia-0 was used in all experi-ments except for abi5–1 that is in the Wassilewskija (Ws) background (5). Theputative hy5 T-DNA insertion line, SALK�096651, was obtained from theArabidopsis Biological Resource Center (Columbus, OH). Homozygous hy5mutants, referred as hy5 in this study, were isolated by their long hypocotylsand confirmed by PCR-based genotyping. The full-length cDNA and the 1.7-kbpromoter of the ABI5 gene were amplified and the fragments were ligatedinto the pENTR-D-TOPO vector (Invitrogen). After sequence confirmation, theABI5 cDNA was cloned into the pMDC32 vector to make the cauliflower mosaicvirus 35S promoter-driven overexpression construct, and the ABI5 promoterwas cloned into the pMDC162 vector to generate the promoter GUS fusionconstruct. Agrobacterium tumefaciens GV3101 was transformed by electro-poration with these constructs and was used to vacuum-infiltrate 4-week-oldArabidopsis seedlings. Seed germination assays and stress and ABA treatmentsof seedlings were conducted as described (39). Root growth measurement(15), histochemical staining, and GUS quantification (11) were performed asdescribed.

RNA Blotting Analysis. Total RNA from dry seeds of Col-0, hy5, Ws, and abi5 wasextracted and analyzed as described (39). Total RNA from young seedlings ofCol-0 and hy5 was extracted by using TRI Reagent according to the manufac-turer’s manual (Molecular Research). For stress treatments, whole seedlings

were treated with 100 �M ABA for 0.5 or 2 h before extracting total RNA. Thegene-specific probes were prepared by PCR amplification with the respectiveprimers [supporting information (SI) Table 1].

Gel Retardation, ChIP, and Real-Time PCR Assays. The �1.7-kb region upstreamof the translational start codon of the ABI5 gene was divided into four overlap-ping fragments: P1, �1754 to �1294; P2, �1315 to �863; P3, �883 to �440; andP4, �460 to 0. These fragments were PCR-amplified with the respective primers(SI Table 1). Gel retardation was performed as described (18).

ChIP assays were performed as described (21) with 2-day-old seedlingsgrown on Murashige and Skoog medium (MS) or MS supplemented with 0.5�M ABA. The PCR primers for the P1–P4 fragments of the ABI5 promoter wereused to amplify the DNA fragments. Primers for real-time PCR of the ABI5 P1fragment were: 5�-TTTAGGTCGCTGGTTCGATTC and 5�-AACATGATTC-CGAACTTCCATTG.

Light Responses. Assays of seedling light response (hypocotyl length) wereconducted as described (40).

ACKNOWLEDGMENTS. We thank Dr. Ruth Finkelstein (University of Califor-nia, Santa Barbara) for an Arabidopsis ABI5::GUS line, Dr. Mark Running forcritical reading of the manuscript, and Dr. Chris Taylor (Donald Danforth PlantScience Center) for experimental materials. This study was supported bythe National Science Foundation Grants 0446359 (to L.X) and 0114726(to M.M.N.).

1. Finkelstein RR, Gampala SSL, Rock CD (2002) Abscisic acid signaling in seeds andseedlings. Plant Cell Suppl 14:15–45.

2. Schroeder JI, Kwak JM, Allen GJ (2001) Guard cell abscisic acid signaling and engineer-ing drought hardiness in plants. Nature 410:327–330.

3. Leung J, Giraudat J (1998) Abscisic acid signal transduction. Annu Rev Plant PhysiolPlant Mol Biol 49:199–222.

4. Xiong L, Schumaker KS, Zhu JK (2002) Cell signaling during cold, drought, and saltstress. Plant Cell Suppl 14:165–183.

5. Finkelstein RR (1994) Mutations at two new Arabidopsis ABA response loci are similarto the abi3 mutations. Plant J 5:765–771.

6. Finkelstein RR, Lynch TJ (2000) The Arabidopsis abscisic acid response gene ABI5encodes a basic leucine zipper transcription factor. Plant Cell 12:599–610.

7. Lopez-Molina L, Chua NH (2000) A null mutation in a bZIP factor confers ABA-insensitivity in Arabidopsis thaliana. Plant Cell Physiol 41:541–547.

8. Lopez-Molina L, Mongrand S, Chua N-H (2001) A postgermination developmentalarrest checkpoint is mediated by abscisic acid and requires the ABI5 transcription factorin Arabidopsis. Proc Natl Acad Sci 98:4782–4787.

9. Bensmihen S, et al. (2002) The homologous ABI5 and EEL transcription factors functionantagonistically to fine-tune gene expression during late embryogenesis. Plant Cell14:1391–1403.

10. Carles C, et al. (2002) Regulation of Arabidopsis thaliana Em genes: Role of ABI5. PlantJ 30:373–383.

11. Brocard IM, Lynch TJ, Finkelstein RR (2002) Regulation and role of the Arabidopsisabscisic acid-insensitive 5 gene in abscisic acid, sugar, and stress response. Plant Physiol129:1533–1543.

12. Koornneef M, Bentsink L, Hilhorst H (2002) Seed dormancy and germination. Curr OpinPlant Biol 5:33–36.

13. De Smet I, et al. (2003) An abscisic acid-sensitive checkpoint in lateral root developmentof Arabidopsis. Plant J 33:543–555.

14. Deak KI, Malamy J (2005) Osmotic regulation of root system architecture. Plant J43:17–28.

15. Xiong L, Wang RG, Mao G, Koczan JM (2006) Identification of drought tolerancedeterminants by genetic analysis of root response to drought stress and abscisic acid.Plant Physiol 142:1065–1074.

16. Koornneef M, Rolff E, Spruit CJP (1980) Genetic control of light-inhibited hypocotylelongation in Arabidopsis thaliana (L) Heynh. Z Pflanzenphysiol 100:147–160.

17. Oyama T, Shimura Y, Okada K (1997) The Arabidopsis HY5 gene encodes a bZIP proteinthat regulates stimulus-induced development of root and hypocotyl. Genes Dev11:2983–2995.

18. Chattopadhyay S, Ang LH, Puente P, Deng XW, Wei N (1998) Arabidopsis bZIP proteinHY5 directly interacts with light-responsive promoters in mediating light control ofgene expression. Plant Cell 10:673–683.

19. Finkelstein R, Gampala SS, Lynch TJ, Thomas TL, Rock CD (2005) Redundant and distinctfunctions of the ABA response loci ABA-INSENSITIVE(ABI)5 and ABRE-BINDING FAC-TOR (ABF)3. Plant Mol Biol 59:253–267.

20. Nakamura S, Lynch TJ, Finkelstein RR (2001) Physical interactions between ABA re-sponse loci of Arabidopsis. Plant J 26:627–635.

21. Lee J, et al. (2007) Analysis of transcription factor HY5 genomic binding sites revealedits hierarchical role in light regulation of development. Plant Cell 19:731–749.

22. Brocard-Gifford IM, Lynch TJ, Finkelstein RR (2003) Regulatory networks in seedsintegrating developmental, abscisic acid, sugar, and light signaling. Plant Physiol131:78–92.

23. Ang LH, et al. (1998) Molecular interaction between COP1 and HY5 defines a regula-tory switch for light control of Arabidopsis development. Mol Cell 1:213–222.

24. Kim SY, Ma J, Perret P, Li Z, Thomas TL (2002) Arabidopsis ABI5 subfamily members havedistinct DNA-binding and transcriptional activities. Plant Physiol 130:688–697.

25. Hardtke CS, et al. (2000) HY5 stability and activity in Arabidopsis is regulated byphosphorylation in its COP1 binding domain. EMBO J 19:4997–5006.

26. Osterlund MT, Wei N, Deng XW (2000) The roles of photoreceptor systems and theCOP1-targeted destabilization of HY5 in light control of Arabidopsis seedling devel-opment. Plant Physiol 124:1520–1524.

27. Ang LH, Deng XW (1994) Regulatory hierarchy of photomorphogenic loci: Allele-specific and light-dependent interaction between the HY5 and COP1 loci. Plant Cell6:613–628.

28. Pepper AE, Chory J (1997) Extragenic suppressors of the Arabidopsis det1 mutantidentify elements of flowering-time and light-response regulatory pathways. Genetics145:1125–1137.

29. Ulm R, et al. (2004) Genomewide analysis of gene expression reveals function of thebZIP transcription factor HY5 in the UV-B response of Arabidopsis. Proc Natl Acad SciUSA 101:1397–1402.

30. Holm M, Ma LG, Qu LJ, Deng XW (2002) Two interacting bZIP proteins are direct targetsof COP1-mediated control of light-dependent gene expression in Arabidopsis. GenesDev 16:1247–1259.

31. Sibout R, et al. (2006) Opposite root growth phenotypes of hy5 versus hy5 hyh mutantscorrelate with increased constitutive auxin signaling. PLoS Genet 2:e202.

32. Nakashima K, et al. (2006) Transcriptional regulation of ABI3- and ABA-responsivegenes including RD29B and RD29A in seeds, germinating embryos, and seedlings ofArabidopsis. Plant Mol Biol 60:51–68.

33. Mazzella MA, et al. (2005) Phytochrome control of the Arabidopsis transcriptomeanticipates seedling exposure to light. Plant Cell 17:2507–2516.

34. Stone SL, Williams LA, Farmer LM, Vierstra RD, Callis J (2006) KEEP ON GOING, a RINGE3 ligase essential for Arabidopsis growth and development, is involved in abscisic acidsignaling. Plant Cell 18:3415–3428.

35. Smalle J, et al. (2003) The pleiotropic role of the 26S proteasome subunit RPN10 inArabidopsis growth and development supports a substrate-specific function in abscisicacid signaling. Plant Cell 15:965–980.

36. Creelman RA, Mason HS, Bensen RJ, Boyer JS, Mullet JE (1990) Water deficit and abscisicacid cause differential inhibition of shoot versus root growth in soybean seedlings:Analysis of growth, sugar accumulation, and gene expression. Plant Physiol 92:205–214.

37. Seo M, et al. (2006) Regulation of hormone metabolism in Arabidopsis seeds: Phyto-chrome regulation of abscisic acid metabolism and abscisic acid regulation of gibberel-lin metabolism. Plant J 48:354–366.

38. Oh E, et al. (2007) PIL5, a phytochrome-interacting bHLH protein, regulates gibberellinresponsiveness by binding directly to the GAI and RGA promoters in Arabidopsis seeds.Plant Cell 19:1192–1208.

39. Xiong L, et al. (2001) Modulation of abscisic acid signal transduction and biosynthesisby an Sm-like protein in Arabidopsis. Dev Cell 1:771–781.

40. Ward JM, Cufr CA, Denzel MA, Neff MM (2005) The Dof transcription factor OBP3modulates phytochrome and cryptochrome signaling in Arabidopsis. Plant Cell 17:475–485.

4500 � www.pnas.org�cgi�doi�10.1073�pnas.0710778105 Chen et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

0