910Components of the Arabidopsis C-RepeatDehydration-Responsive

高考英语一轮总复习必修第一册提能训练：Unit 4 Natural DisastersⅠ.阅读理解A(2024·浙江1月高考题) On September 7, 1991, the costliest hailstorm (雹暴) in Canadian history hit Calgary's southern suburbs. As a result, since 1996 a group of insurance companies have spent about $2 million per year on the Alberta Hail Suppression Project. Airplanes seed threatening storm cells with a chemical to make small ice crystals fall as rain before they can grow into dangerous hailstones. But farmers in east-central Alberta—downwind of the hail project flights—worry that precious moisture (水分) is being stolen from their thirsty land by the cloud seeding.Norman Stienwand, who farms in that area, has been addressing public meetings on this issue for years. “Basically, the provincial government is letti ng the insurance companies protect the Calgary-Edmonton urban area from hail，” Mr. Stienwand says, “but they're increasing drought risk as far east as Saskatchewan.”The Alberta hail project is managed by Terry Krauss, a cloud physicist who works for Weath er Modification Inc. of Fargo, North Dakota. “We affect only a very small percentage of the total moisture in the air, so we cannot be causing drought，” Dr. Krauss says. “In fact, we may be helping increase the moisture downwind by creating wetter ground.”One doubter about the safety of cloud seeding is Chuck Doswell, a research scientist who just retired from the University of Oklahoma. “In 1999, I personally saw significant tornadoes (龙卷风) form from a seeded storm cell in Kansas，” Dr. Doswell says. “Doe s cloud seeding create killer storms or reduce moisture downwind? No one really knows, of course, but the seeding goes on.”Given the degree of doubt, Mr. Stienwand suggests, “it would be wise to stop cloud seeding.” In practice, doubt has had the opposite effect. Due to the lack of scientific proof concerning their impacts, no one has succeeded in winning a lawsuit against cloud-seeding companies. Hence, private climate engineering can proceed in relative legal safety.语篇导读：本文是一篇说明文。

Developmental CellArticleA MYB-Domain Protein EFMMediates Flowering Responsesto Environmental Cues in ArabidopsisYuanyuan Yan,1,2Lisha Shen,1,2Ying Chen,1Shengjie Bao,1Zhonghui Thong,1and Hao Yu1,*1Department of Biological Sciences and Temasek Life Sciences Laboratory,National University of Singapore,Singapore117543,Singapore 2Co-first author*Correspondence:dbsyuhao@.sg/10.1016/j.devcel.2014.07.004SUMMARYPlants adjust the timing of the transition tofloweringto ensure their reproductive success in changingenvironments.Temperature and light are major envi-ronmental signals that affectflowering time throughconverging on the transcriptional regulation ofFLOWERING LOCUS T(FT)encoding theflorigen in Arabidopsis.Here,we show that a MYB transcrip-tion factor EARLY FLOWERING MYB PROTEIN(EFM)plays an important role in directly repressingFT expression in the leaf vasculature.EFM mediates the effect of ambient temperature onflowering and is directly promoted by another major FT repressor, SHORT VEGETATIVE PHASE.EFM interacts with an H3K36me2demethylase JMJ30,which forms a nega-tive feedback regulatory loop with the light-respon-sive circadian clock,to specifically demethylate an active mark H3K36me2at FT.Our results suggest that EFM is an important convergence point that mediates plant responses to temperature and light to determine the timing of reproduction.INTRODUCTIONPlants respond to environmental signals to determine the timing of the transition toflowering,which is essential for their repro-ductive success.Temperature and light are two major related environmental cues that are intimately perceived by plants and contribute to precise regulation offlowering time in changing seasons.Plants have evolved various adaptive mechanisms to optimize their reproductive efforts in response to a wide range of tolerable temperatures.Low temperatures near the freezing point affect flowering in Arabidopsis in the vernalization pathway,mainly through a potent repressor FLOWERING LOCUS C(FLC) that directly suppresses the expression of twofloral pathway in-tegrators,FLOWERING LOCUS T(FT)and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1(SOC1)(Creville´n and Dean,2011;Li et al.,2008;Michaels and Amasino,1999;Searle et al.,2006;Sheldon et al.,2000).Warmer ambient growth temperatures also have a profound effect onflowering of Arabidopsis(Bla´zquez et al.,2003).This is mediated by several key regulators identified in the thermosensory pathway involving various molecular mechanisms.A bHLH transcription factor PHYTOCHROME INTERACTING FACTOR4(PIF4)binds and ac-tivates FT under relatively high temperatures in short days,which is facilitated by a decrease in a repressive H2A.Z nucleosome occupancy at the FT locus(Kumar et al.,2012;Kumar and Wigge,2010).In parallel to PIF4,two MADS-box transcription factors,SHORT VEGETATIVE PHASE(SVP)and FLOWERING LOCUS M(FLM),also play essential roles in mediating theflow-ering response to ambient temperatures(Balasubramanian et al.,2006;Lee et al.,2007,2013;Pose´et al.,2013).SVP inter-acts with FLM-b,a functional protein isoform encoded by a splice variant of FLM(Scortecci et al.,2001),to suppress down-streamflowering promoters,such as FT and SOC1,in a temper-ature-dependent manner(Lee et al.,2013;Pose´et al.,2013). Changes in ambient temperature not only regulate the abun-dance of SVP and FLM-b but also affect the formation of the re-sulting repressor complex and its ability to bind to targets.As a major downstream integrator regulated by various tem-peratures,FT is also well known as the major component of the long-soughtflorigen that move from leaves to the shoot api-cal meristem to induce the formation offloral meristems in response to light(Corbesier et al.,2007;Jaeger and Wigge, 2007;Liu et al.,2012;Mathieu et al.,2007).Under long-day con-ditions,FT messenger RNA(mRNA)expression is activated in vascular tissues of leaves in a circadian rhythmic manner by the transcriptional regulator CONSTANS(CO)whose expression and activity are controlled by light signaling pathways and the circadian clock(An et al.,2004;Imaizumi,2010;Samach et al., 2000;Song et al.,2013;Sua´rez-Lo´pez et al.,2001;Valverde et al.,2004;Wigge et al.,2005).Thus,transcriptional regulation of FT seems to be a key output resulting from integration of tem-perature cues with light signals,which is favorable to plant repro-ductive plasticity underfluctuating seasonal conditions.In this study,we show that a MYB transcription factor,EARLY FLOWERING MYB PROTEIN(EFM),plays an important role in directly repressing FT expression in the leaf vasculature in a dosage-dependent manner.EFM is directly promoted by SVP and mediates the effect of ambient temperature onflowering. EFM interacts with a Jumonji-C(JmjC)-domain-containing pro-tein,JMJ30,a dimethyl H3K36(H3K36me2)demethylase that forms a negative feedback regulatory loop with the light-respon-sive circadian clock,to specifically demethylate an active mark H3K36me2at the FT locus.Our results suggest that EFMmediates Arabidopsis responses to temperature and light to determine the timing of reproduction in response to changing environments.RESULTSEFM Represses Flowering in ArabidopsisTo elucidate SVP function during the floral transition,we per-formed chromatin immunoprecipitation (ChIP)-chip analysis as previously described (Tao et al.,2012)to identify enriched genomic regions for SVP in wild-type Arabidopsis plants over svp-41null mutants.One of the potential binding sites forSVPFigure 1.EFM Regulates Flowering Time in Arabidopsis in a Dosage-Dependent Manner(A)Schematic diagram shows the EFM coding re-gion,the transposon insertion site in efm-1(a Col near-isogenic line introgressed from N16895),and the target site of the AmiR in AmiR-efm .Exons and introns are indicated by black and white boxes,respectively.(B)EFM expression is undetectable in efm-1by semiquantitative PCR using the primers that amplify the full-length EFM transcript.TUB2was amplified as an internal control.(C)efm-1shows earlier flowering than a wild-type (WT)plant under long-day conditions.(D)Flowering time of efm-1under long-day and short-day conditions.Values were scored from at least 20plants of each genotype.Error bars indi-cate SD.Asterisks indicate significant differences in flowering time of efm-1compared with that of wild-type (WT)plants (Student’s t test,p <0.001).(E)Distribution of flowering time in T1transgenic plants carrying the EFM genomic fragment (gEFM )in efm-1background.(F and G)In (F),downregulation of EFM in inde-pendent AmiR-efm transgenic plants correlates to the degree of early flowering under long-day con-ditions.Expression of EFM in (F)and (G)was determined by quantitative real-time PCR in 9-day-old independent transgenic seedlings.Results were normalized against the expression of TUB2.The maximum expression is set as 100%.Error bars indicate SD.(G)Upregulation of EFM in in-dependent 35S:EFM transgenic plants correlates to the degree of late flowering under long-day conditions.See also Figure S1.was located at the genomic region of At2g03500on chromosome 2.This locus encodes an unknown MYB protein con-taining a single R3-type MYB domain (Du-bos et al.,2010)that is well conserved among various plant species (Figure S1A available online).To investigate the biolog-ical role of this gene,we isolated a corre-sponding Ds transposon insertion mutant (N16895)in the Nossen background from Nottingham Arabidopsis Stock Centre (NASC)(Figure 1A).This mutant exhibited an early-flowering phenotype under long-day conditions (Figure S1B).We further introgressed this mutant six times into the Columbia (Col)background to generate a Col near-isogenic line,and found that the Col mutants also flowered earlier than wild-type plants under both long-day and short-day conditions (Figures 1C and 1D).Thus,this MYB gene and the corresponding Col mutant were designated EARLY FLOWERING MYB PROTEIN (EFM )and efm-1,respectively.There was no detectable full-length EFM transcript in efm-1(Figure 1B).To verify that loss of EFM function is responsible for the early-flowering phenotype of efm-1,we transformed efm-1with a genomic construct (gEFM )harboring a 5.8-kb EFM genomic region including the 3.5-kb upstream sequenceDevelopmental CellEFM Mediates Flowering Responsesand the 2.3-kb full coding sequence plus introns.Most efm-1gEFM T1transformants exhibited comparable flowering time to wild-type plants (Figure 1E),suggesting that EFM acts as a flow-ering repressor.To confirm EFM function in the control of flower-ing time,we generated EFM knockdown transgenic plants by artificial microRNA (AmiR)interference.We created 16AmiR-efm independent lines that expressed an AmiR specifically tar-geting at the last exon of the EFM mRNA (Figure 1A),among which 12lines exhibited different degrees of early flowering under long-day conditions.In 8out of these lines,the degrees of early flowering were closely related to the levels of downregu-lation of EFM expression (Figure 1F).In contrast,floweringwasFigure 2.EFM Is Expressed in Vascular Tissues(A–F)Representative GUS staining of EFM:GUS transgenic plants displays EFM expression in a 5-day-old seedling (A),cotyledons (B),a root tip (C),a rosette leaf (D),a cauline leaf (E),and an open flower (F).Scale bars,1mm in (A),(B),(D),and (E);50m m in (C);and 200m m in (F).(G)Distribution of flowering time in T1transgenic plants carrying the KNAT1:EFM or SUC2:EFM construct in efm-1background.WT,wild-type.(H)Subcellular localization of EFM-GFP.GFP localization was observed in tobacco leaves infiltrated with 35S:EFM-GFP .GFP,GFP fluores-cence;DAPI,fluorescence of 40,6-diamino-2-phenylindole;Merge,merge of GFP,DAPI,and bright field images.Scale bar,10m m.See also Figure S2.delayed to varying degrees in transgenic plants overexpressing EFM ,which also correlated with increased levels of EFM in the transgenic plants examined (Fig-ure 1G).Taken together,these results demonstrate that EFM represses flower-ing in a dosage-dependent manner.EFM Is Specifically Expressed in Vascular TissuesWe examined EFM expression in various tissues of wild-type plants using quantita-tive real-time PCR and found its highest expression in rosette leaves (Figure S2A).To monitor the detailed expression pattern of EFM ,we generated an EFM:b -glucuronidase (GUS )reporter construct in which the EFM genomic fragment used for the gene complementation test (Fig-ure 1E)was fused to the GUS gene.Among 28independent EFM:GUS lines obtained,most lines consistently showed specific GUS expression in vascular tis-sues of various organs,including cotyle-dons,rosette leaves,and cauline leaves (Figures 2A–2F).There was no GUS stain-ing signal in the vegetative shoot apical meristem (Figure 2B),and in situ hybridi-zation further revealed barely detectable EFM expression in shoot apices during the floral transition (Figure S2B).To test the effect of spatial expression of EFM on flowering time,we transformed efm-1with the construct in which the EFM coding sequence was driven by the promoter of either SUCROSE TRANSPORTER 2(SUC2)or KNOTTED-LIKE FROM ARABIDOPSIS THALIANA 1(KNAT1),which is active specifically in phloem companion cells or shoot apical meristems,respec-tively.Most efm-1KNAT1:EFM T1transformants exhibited com-parable flowering time to efm-1,whereas all 40efm-1SUC2:EFM T1transformants flowered later than efm-1,among which 14lines displayed similar flowering time to wild-type plants (Figure 2G).Developmental CellEFM Mediates Flowering ResponsesThese results,together with the specific expression pattern of EFM in the leaf vasculature,indicate that EFM functions in leaf vascular tissues to repressflowering.We further examined the subcellular localization of EFM in to-bacco leaves using the greenfluorescent protein(GFP)fusion construct35S:EFM-GFP.EFM-GFP localized specifically in the nucleus(Figure2H),supporting that EFM functions as a tran-scription factor.EFM Expression Is Directly Promoted by SVPAs EFM wasfirst identified as a downstream target of SVP in our ChIP-chip analysis,we further confirmed the binding of SVP to EFM by ChIP assay using independent wild-type and svp-41seedlings grown under long-day conditions.We scanned the EFM genomic sequence for the CC(A/T)6GG (CArG)motif,a canonical binding site for MADS-domain pro-teins including SVP,with a maximum of one nucleotide mismatch and designed11primer pairs covering all the identi-fied motifs(Figure3A)to measure DNA enrichment in ChIP assay of SVP binding to EFM using anti-SVP antibody(Shen et al.,2011).Enrichment of SVP at the region between exons 2and3of EFM was consistently detected in three biological replicates(Figures3A and3B).To examine if physical binding of SVP on EFM directly affects EFM transcription,we detected EFM expression in svp-41and 35S:SVP seedlings harvested9days after germination when thefloral transition occurred in wild-type plants in long days un-der our growth conditions(at23 C).EFM expression decreased in svp-41but was elevated in35S:SVP(Figure3C),suggesting that SVP promotes EFM expression in young seedlings.Genetic analysis revealed that efm-1showed an early-flowering pheno-type similar to that of svp-41,while the svp-41efm-1double mutantflowered only slightly earlier than either of the single mutants(Figure3D),implying that SVP and EFM could act as flowering repressors mainly in the same regulatory pathway. In agreement with the role of EFM as aflowering repressor,its expression gradually decreased before thefloral transition and remained at low levels afterward(Figure3E).Thesefindings suggest that SVP directly promotes EFM expression to sup-pressflowering.Figure 3.SVP Directly Regulates EFM Expression(A)Schematic diagram shows the EFM genomic region.Exons are represented by black boxes, while other genomic regions are represented by white boxes.Arrowheads indicate the sites con-taining either a single mismatch or a perfect match to the consensus binding sequence(CArG box)of MADS-domain proteins.Eleven DNA fragments mostlyflanking these sites were designed for ChIP analysis of the SVP binding site as shown in(B). bp,base pairs.(B)ChIP analysis of SVP binding to the EFM genomic region.Nine-day-old seedlings grown under long-day conditions were harvested for ChIP analysis.Enrichment fold of each fragment was calculatedfirst by normalizing the amount of a target DNA fragment against a genomic fragment of TUB2as an internal control and then by normalizing the value for wild-type(WT)plants against that for svp-41.Error bars indicate SD of three biological replicates.A SOC1fragment (number6)that is highly associated with SVP(Shen et al.,2011)was included as a positive control. (C–E)In(C),EFM expression was determined by quantitative real-time PCR in wild-type(WT),svp-41and35S:SVP grown at23 C.Nine-day-old seedlings grown under long-day conditions were harvested for expression analysis.EFM expres-sion in(C)and(E)was normalized to TUB2 expression.Error bars indicate SD.Asterisks indicate significant differences in EFM expression levels between indicated genotypes and wild-type plants(Student’s t test,p<0.001).(D)Flowering time of efm-1,svp-41and svp-41efm-1grown under long-day conditions.Values were scored from at least20plants of each genotype.An asterisk indicates a significant difference inflow-ering time of svp-41efm-1compared with that of svp-41or efm-1(Student’s t test,p<0.05).Error bars indicate SD.(E)Temporal expression of EFM determined by quantitative real-time PCR in developing wild-type plants.Developmental Cell EFM Mediates Flowering ResponsesEFM Expression Is Regulated by Several PathwaysTo investigate how EFM regulates flowering in response to various flowering signals,we examined EFM expression in flow-ering mutants of various genetic pathways and also in different environmental conditions.EFM expression remained almost un-changed in the mutants of the photoperiod pathway (Figure 4A)and also did not exhibit a circadian oscillation under long-day conditions (Figure S2C).In addition,vernalization treatment did not affect EFM expression in both wild-type and FRI FLC plants (Figure 4B).These results indicate that EFM expression is not directly regulated by the photoperiod and vernalization pathways.EFM expression was consistently higher in GA-deficient ga1-3mutants than in wild-type plants grown under short-day condi-tions (Figure 4C),while gibberellic acid (GA)treatment for 2consecutive days significantly suppressed EFM in ga1-3grown under short-day conditions (Figure 4D),suggesting that the GA pathway represses EFM to promote flowering.EFM expression was consistently upregulated in loss-of-function mutants,such as fca-1and fve-1,of the autonomous pathway (Figure 4E),indi-cating a repressive effect on EFM by the autonomous pathway.As FCA and FVE also control flowering upstream of SVP in response to changes in ambient growth temperature in the ther-mosensory pathway (Bla´zquez et al.,2003;Lee et al.,2007),we further tested whether EFM expression was also modulated by ambient temperatures.EFM expression decreased with increasing temperature from 16 C to 27 C (Figure 4H)in the ambient temperature range (Wigge,2013).Moreover,EFM expression was consistently lower in svp-41but higher in 35S:SVP at 16 C,23 C,and 27 C (Figures 3C and 4G),suggest-ing that SVP promotes EFM expression at ambient tempera-tures.Loss of function of SVP ,a major player in the thermosen-sory pathway,showed a temperature-insensitive floweringphenotype as previously reported (Lee et al.,2007,2013;Pose´et al.,2013)(Figure 4H).In agreement with the finding on the promotion of EFM by SVP,efm-1also flowered in apartialFigure 4.EFM Expression Is Regulated by the Autonomous,GA,and Thermosensory Pathways(A–G)In (A),EFM expression in 9-day-old mutants of the photoperiod pathway is shown.EFM expression levels determined by quantitative real-time PCR in (A)through (G)were normalized to TUB2expression and shown as relative values to the maximal gene expression levels set at 100%.Error bars in (A)through (G)indicate SD.(B)Effect of vernalization treatment on EFM expression.Seeds were grown on Murashige and Skoog me-dium and vernalized at 4 C under low-light con-ditions for 8weeks.Nine-day-old seedlings grown under long-day conditions were harvested for expression analysis.WT,wild-type.(C)Compari-son of EFM expression in wild-type L er plants and ga1-3mutants (L er background).Four-day-old (D4)and 8-day-old (D8)seedlings grown under short-day conditions were harvested for expres-sion analysis.(D)Effect of GA treatment on EFM expression in ga1-3mutants grown under short-day conditions.Exogenous GA (100m M)or 0.1%ethanol (mock)was applied daily on to 3-week-old ga1-3seedlings for 2consecutive days.The seedlings before (0h)and 48hr after (48h)the first GA treatment were harvested for expression analysis.An asterisk indicates a significant differ-ence in EFM expression between GA-treated and mock-treated samples (Student’s t test,p <0.001).(E)EFM expression in 9-day-old mutants of the autonomous pathway.(F)EFM expression in 9-day-old wild-type seedlings grown at 16 C,23 C,and 27 C under long-day conditions.Asterisks indicate significant differences in EFM expression in seedlings grown at 16 C or 27 C compared with that at 23 C (Student’s t test,p <0.001).(G)EFM expression in wild-type,svp-41,and 35S:SVP grown at 16 C and 27 C.Nine-day-old seedlings grown under long-day conditions were harvested for expression analysis.Asterisksindicate significant differences in EFM expression levels between indicated genotypes and wild-type plants (Student’s t test,p <0.001).(H)Flowering time of svp-41and efm-1grown at 16 C,23 C,and 27 C under long-day conditions.The ratios of flowering time between 16 C and 23 C (16 C/23 C)and between 23 C and 27 C (23 C/27 C)for all the genotypes are indicated in the attached table.Values were scored from at least 15plants of each genotype.Error bars indicate SD.See also Figure S2.Developmental CellEFM Mediates Flowering Responsestemperature-insensitive pattern from 16 C to 23 C and in a more insensitive pattern from 23 C to 27 C (Figure 4H).These results suggest that EFM acts downstream of SVP to partially mediate the effect of the thermosensory pathway on flowering time.EFM Directly Represses FT Expression in LeavesSince SVP regulates flowering mainly through transcriptional regulation of SOC1and FT (Lee et al.,2007;Li et al.,2008),we proceeded to test whether EFM also affects the expressionofFigure 5.EFM Directly Represses FTExpression(A)Temporal expression of FT in developing efm-1and wild-type (WT)seedlings under long-day conditions.All samples were collected at the end of long-day conditions (ZT16).Gene expression levels determined by quantitative real-time PCR in (A)through (C)were normalized to TUB2expres-sion and shown as relative values to the maximal gene expression levels set at 100%.Error bars indicate SD.(B)Temporal expression of SOC1in developing efm-1,ft-10,ft-10efm-1,and wild-type seedlings under long-day conditions.(C)Diurnal oscillation of FT expression in 9-day-old efm-1and wild-type seedlings under long-day conditions.Samples were harvested at 4hr intervals over a 24hr period.Sampling time is expressed in hours as ZT,which is the number of hours after the onset of illumination.(D)Flowering time of various mutants grown under long-day conditions.Values were scored from at least 20plants of each genotype.Error bars indicate SD.(E)GUS expression in FT:GUS and efm-1FT:GUS .Representative GUS staining of 13-day-old FT:GUS and efm-1FT:GUS seedlings grown under long-day conditions is shown on the left,while quantitative comparison of GUS activity is shown on the right.Error bars indicate SD.(F)Schematic diagram shows the FT genomic re-gion.Exons are represented by black boxes,while other genomic regions are represented by white boxes.Gray boxes represent the DNA fragments amplified in ChIP analysis shown in (G).(G)ChIP analysis of EFM-GFP binding to the FT genomic region in 9-day-old seedlings.Enrich-ment fold of each fragment was calculated first by normalizing the amount of a target DNA fragment against a genomic fragment of TUB2as an internal control and then by normalizing the value for efm-1gEFM-GFP against that for efm-1.An ACTIN (ACT )fragment was amplified as a negative control.Er-ror bars indicate SD of three biological replicates.See also Figure S3.these two genes.FT expression was dramatically increased in developing efm-1seedlings (Figure 5A),but greatly suppressed in 35S:EFM (Figure S3A),at the vegetative phase and during the floral transition.Similar but less dramatic changing trends for SOC1expression were observed in efm-1and 35S:EFMseedlings (Figure 5B;Figure S3B).As FT positively regulates SOC1(Yoo et al.,2005),we then examined whether EFM affects SOC1expression through FT.Although SOC1was upregulated in efm-1,its expression was dramatically downregulated in ft-10efm-1to the levels comparable to those in ft-10(Figure 5B),indicating that EFM suppresses SOC1in an FT-dependent manner.Since FT is expressed in a diurnal pattern with a peak at the end of long days (Imaizumi and Kay,2006),we further measured its circadian expression in efm-1every 4hr over aDevelopmental CellEFM Mediates Flowering Responses24hr long-day cycle and found that rhythmic FT expression was constitutively suppressed by EFM (Figure 5C),demonstrating a strong negative effect of EFM on FT expression throughout the day.Genetic analyses revealed that compared with soc1-2,ft-10almost completely suppressed the early-flowering phenotype of efm-1(Figure 5D),substantiating that FT rather than SOC1is a downstream target of EFM.To understand the spatial effect of EFM on FT expression,we monitored the GUS staining patterns of FT:GUS (Takada and Goto,2003)in efm-1versus wild-type plants.The GUS signal was mainly restricted to the cotyledon vasculature and mi-nor veins toward the distal region of rosette leaves in wild-type plants,whereas the signal was obviously enhanced and expanded to almost all veins in efm-1(Figure 5E).Quantification of GUS activity consistently revealed greatly increased FT:GUS expression in efm-1(Figure 5E).In contrast,there were only mild changes in the GUS expression level and domain of SOC1:GUS (Li et al.,2008)in efm-1versus wild-type plants (Fig-ure S3C).These results,which are consistent with the negative effect of EFM expression in the leaf vasculature on flowering (Figure 2),support that EFM represses flowering through specif-ically suppressing FT in leaf vascular tissues.To examine whether EFM directly controls FT transcription,we first created efm-1gEFM-GFP transgenic lines containing the GFP -tagged EFM genomic fragment used for the gene comple-mentation assay (Figure 1E).Most efm-1gEFM-GFP T1trans-formants displayed similar flowering time to that of wild-type plants (Figure S3D),implying that the fusion protein of EFM-green fluorescent protein (GFP)is biologically functional.ChIP assays using a representative efm-1gEFM-GFP line showed that among 21fragments covering a 9-kb FT genomic region(Figure 5F)that includes sufficient cis elements required for rescuing the late-flowering phenotype of ft-10(Adrian et al.,2010),the number 5fragment was consistently enriched for EFM binding in three biological replicates (Figure 5G).These re-sults collectively suggest that EFM directly binds to the FT locus to suppress its expression.In addition to FT ,its closest homolog TWIN SISTER OF FT (TSF )(Yamaguchi et al.,2005)was also upregulated in devel-oping efm-1seedlings (Figure S3E),despite a less dramatic change than that occurred to FT ,implying that TSF could be another potential target of EFM.In contrast,expression of those known FT upstream regulators,such as AGAMOUS-LIKE 24,FLC ,SVP ,and several AP2-like members,was only slightly changed or remained unchanged in efm-1(Figures S3F–S3I).EFM Interacts with a JmjC-Domain-Containing Protein JMJ30To further elucidate how EFM affects FT transcription during the floral transition,we carried out yeast two-hybrid screening to identify interacting partners of EFM.The identified interactors included a JmjC-domain-containing protein JMJ30,which was previously reported to play a role in regulating the Arabidopsis circadian clock (Jones et al.,2010;Lu et al.,2011).Yeast two-hybrid assays confirmed the interaction between EFM and JMJ30(Figures 6A and 6B).To test the in vivo interaction be-tween EFM and JMJ30,we performed bimolecular fluorescence complementation (BiFC)assays and found the enhanced yellow fluorescent protein (EYFP)fluorescence signal only in the nuclei of tobacco epidermal cells (Figure 6C),implying a direct interac-tion between EFM and JMJ30in the nuclei of living tobaccocells.Figure 6.EFM Interacts with a Putative Histone Demethylase JMJ30(A)Yeast-two hybrid assay of the interaction be-tween EFM and JMJ30.Transformed yeast cells were grown on SD-Leu/-Trp medium (upper panel)and SD-Leu/-Trp/-His medium supplemented with 5mM 3-amino-1,2,4-triazole (3AT)(lower panel).(B)Quantification of the interaction between EFM and JMJ30in yeast by b -galactosidase assays.(C)BiFC analysis of the interaction between EFM and JMJ30.DAPI,fluorescence of 40,6-diamino-2-phenylindole;Merge,merge of EYFP,DAPI,and bright field images.(D)In vivo interaction between EFM and JMJ30shown by coimmunoprecipitation.Nu-clear extracts from 9-day-old efm-1gEFM-GFP and efm-1gEFM-GFP 35S:JMJ30-6HA seedlings were immunoprecipitated by anti-HA agarose.The input and coimmunoprecipi-tated proteins were detected by anti-GFP anti-body.(E)Coomassie blue staining of affinity-purified 6His-JMJ30and GST proteins that are overex-pressed in E.coli .(F)JMJ30demethylates H3K36me3and H3K36me2in vitro.Calf thymus free histones were incubated with various amount of JMJ30proteins and subjected to western blot analysis using antibodies that specifically recognize methylated H3histones.Purified GST was used as a reaction control,and the amount of H3served as a loading control.See also Figures S4and S5.Developmental CellEFM Mediates Flowering Responses。

Aerobic granular sludge–state of the artM.K.de Kreuk*,N.Kishida**and M.C.M.van Loosdrecht**Department of Biotechnology,Delft University of Technology,Julianalaan67,2628BC,Delft,The Netherlands(E-mail:m.dekreuk@tnw.tudelft.nl;m.c.m.vanloosdrecht@tnw.tudelft.nl)**Department of Chemical Engineering,Waseda University,3-4-1,Ohkubo,Shinjuku-ku,Tokyo,JapanAbstract In September2006,preliminary to the IWA biofilm conference,a second workshop about aerobic granular sludge was held in Delft,The Netherlands,of which a summary of the discussion outcomes is given in this paper.The definition of aerobic granular sludge was discussed and complemented with a few additional demands.Further topics were formation and morphology of aerobic granular sludge,modelling and use of the aerobic granular sludge in practice.Keywords Aerobic granular sludge;EPS;modelling;morphology;pilot plant;self-aggregationIntroductionAt the end of the1990s,research on biofilm structure and formation(van Loosdrecht et al.,1995,1997a)and on the role of storage polymers(van Loosdrecht et al.,1997b; Krishna and van Loosdrecht,1999)resulted in the idea of growing aerobic granules with-out carrier material on readily biodegradable substrates in a Sequencing Batch Reactor (SBR)(Morgenroth et al.,1997;Beun et al.,1999;Dangcong et al.,1999).The conver-sion of readily biodegradable COD into a substrate yielding a lower maximal growth rate facilitated granule formation.In these aerobic reactors,it was proven to be possible to grow stable granular sludge(Figure1)with integrated COD and nitrogen removal.In 1998,an international patent was submitted and granted(Heijnen and van Loosdrecht, 1998).An extension of thisfirst patent was submitted,including the description of anaerobic feeding(van Loosdrecht and De Kreuk,2004).From the year2000,aerobic granule formation has been excessively studied world-wide(Figure2).Many theories about aerobic granule formation made their way into different studies.Analysis of a cross section of literature published in the last decade shows that type of substrate,COD and N-load,superficial gas velocity or shear stress and oxygen concentration are important parameters.An important secondary parameter for the formation and maintenance of dense granules is the growth rate of the organisms that is influenced by cycle configuration or loading rates(De Kreuk et al.,2005a).An exten-sive overview of parameters that are important for anaerobic and aerobic granule for-mation has been given by Liu and Tay(2004).In the year2004,thefirst workshop on aerobic granular sludge was organised in Munich,Germany(Bathe et al.,2005).The conclusion that could be made from the aerobic granular sludge workshop2004was:“all participants agreed that up till then a lot of basic knowledge was gained on aerobic granule formation and that the pioneering stage wasfinished.Researchers should continue with specific research topics(e.g.factors influencing granule strength and formation,microbial diversity,conversion processes, pathogen removal,pilot-and demonstration-scale studies)and that the technology should be put into practice.Within a couple of years,this workshop should be repeated to see how research went from there;hopefully with new insights and with new applications.”Water Science & Technology Vol 55 No 8–9 pp 75–81 Q IWA Publishing 2007 75doi:10.2166/wst.2007.244Therefore,in September 2006,preliminary to the IWA biofilm conference,a second workshop about aerobic granular sludge was held in Delft,The Netherlands,of which a summary of the discussion outcomes is given below.Definition of aerobic granular sludge Despite the long-time application of anaerobic granular sludge in wastewater treatment,aerobic granular sludge is a new observation.Therefore a definition of aerobic granular sludge was made during the first aerobic granular sludge workshop,2004(De Kreuk et al.,2005b ):Granules making up aerobic granular activated sludge are to be understood as aggregates of microbial origin,which do not coagulate under reduced hydrodynamic shear,and which settle significantly faster than activated sludge flocs.The explanation of the different parts of the statement was discussed and analysed during the 2006workshop and delegates agreed on the statement,when the following explanation is used:(1)Aggregates of microbial origin:speaking of granular activated sludge in the state-ment implies that aerobic granules need to contain active microorganisms and cannot only consist of components of microbial origin (as proteins,EPS,etc.).The microbial population in aerobic granular sludge are to be expected more or less similar to the ones in activated sludge and or biofilms,thus there is no need to describe specific groups of microorganisms in the definition.Furthermore,this part implies that no carrier material is intentionally involved or added;the aggregate is formed without the dosage of such carriermaterial.Figure 1Activated sludge from a wastewater treatment plant (a)and aerobic granular sludge cultivated in a laboratory scale reactor (b)and in a pilot plant (c)# o f p u b l i c a t i o n s Year of publication Figure 2Number of publications about aerobic granular sludge per year (Source:Web of Science)M.K.de Kreuk et al.76(2)No coagulation under reduced hydrodynamic shear:this part describes the differ-ence in behaviour between activated sludge and aerobic granular sludge.Activated sludgeflocs tend to coagulate when they settle(when liquid-sludge mixture is not aerated or stirred),whilst granules do not coagulate and settle as separate units. (3)Which settle significantly faster than activated sludgeflocs:this means that SVI10(SVI after10min of settling)in combination with SVI30should be used forcharacterising the settleability of granular activated sludge as was suggested by Schwarzenbeck et al.(2004).The difference between the SVI10and SVI30value gives an excellent indication about the granule formation and indicates the extent of thickening after settling.(4)The minimum size of the granules should be as such that the biomass still fulfilspoint three.This minimum size was set to0.2mm,which was decided based on measurements in the past.This limit could be adjusted per case/granule type,as long as the other demands of the definition hold.(5)Sieving is considered a proper method to harvest granules from activated sludgetanks or from aerobic granule reactors,which also determines certain strength of the required biomass matrix.When an aggregate fulfils all characteristics as described above,it can be called aerobic granular sludge.This simplifies the interpretation of experimental results and clarifies when to speak about aerobic granular sludge,activated sludge or biofilms.Formation and morphology of aerobic granular sludgeMany factors are held responsible for the formation and stability of aerobic granular sludge,but it is undecided among scientists which factor is the dominant one.Discussions focussed on this topic and a few parameters were highlighted:a)use or appearance of specific self-aggregating cultures;b)selection by settling velocity;c)applied shear stress;d)growth rate of the organisms;e)substrate gradients inside the granules;f)formation of extra-cellular polymeric substances(EPS).Formation of aerobic granular sludge in laboratory experiments mostly occurs with a strict selection regime for well settling sludge by applying short settling times.The aggre-gate forming organisms will be maintained in the reactor,while other organisms are washed out with the effluent.To enhance the start-up of an aerobic granular sludge reactor, the use of specific self-aggregating cultures was suggested by the researchers of Nanyang University(Singapore).Non-pathogenic cultures with self-aggregating abilities were selected and added to a reactor during stat-up.This shortened the start-up time consider-ably(3days instead of9days without specific inoculum).The selection for specific organ-isms to enhance granule formation and stability has also been applied by other researchers, e.g.the use of phosphate or glycogen accumulating organisms(Dulekgurgen et al.,2003; De Kreuk and van Loosdrecht,2004)or nitrifying organisms(Liu et al.,2004a).However, the attending researchers agree that,based on published research,substrate gradients inside the granules are very important as well and that sharp decreasing substrate gradients inside the granules should be avoided by using the ability of converting readily biodegradable substrates into storage polymers and/or using organisms with low actual growth rates.EPS production by slow growing organisms enhance the granule formation (Liu et al.,2004b;McSwain et al.,2005,presented research results of National Taiwan University)and is seen as an important aspect by the attending researchers as well.Applied shear stress and settling time has been an important topic for studies in the past and still leads to discussions.At the aerobic granular sludge workshop2006,several researchers showed their results about this topic(Northwestern University,USA;Univer-sity of Beijing,China;Istanbul Technical University,Turkey).Shear stress is difficult to M.K. de Kreuk et al.77quantify and is often related to superficial gas velocity.However,factors such as stirring should not be underestimated and reported as well.There was general agreement on the positive effects of hydraulic selection pressure on granule formation and stability.When aerobic granular sludge is used with a membrane to filter the effluent,the selection pressure for aggregates by settling totally disappears.Even then,very dense and large flocs were obtained,having a positive effect on the membrane fouling.The high density flocs did only meet the definition of aerobic granular sludge during short periods of the total experiment,especially during the periods that autotrophic organisms were present (results presented by INSA,France).Modelling of aerobic granular sludge Mathematical modelling has proven very useful to study complex processes,such as the aerobic granular sludge systems (Beun et al.,2001;Lu ¨bken et al.,2005).Biological pro-cesses in the granules are determined by concentration gradients of oxygen and diverse substrates.The concentration profiles are the result of many factors,e.g.diffusion coeffi-cients,conversions rates,granule size,biomass spatial distribution and density.All of these factors tightly influence each other,thus the effect of separate factors cannot be stu-died experimentally.Moreover,experiments in granular sludge reactors take many weeks to reach steady state.A good computational model for the granular sludge process pro-vides significant insight in the most important factors that affect the nutrient removal rates and in the distribution of different microbial populations inside the granules.Further,models could also be used for process optimisation and for the scale-up and design (e.g.hydraulics)of a full-scale reactor.Aerobic granular sludge systems can be modelled in different ways,using different modelling tools depending on the fields of interest.When the overall reactor behaviour is described (substrate removal or sludge production),traditional biofilm modelling can be used,as in AQUASIM (Reichert,1998).Such models have already been developed by the groups in Munich and Delft.The effect of process parameters on the nutrient removal rates could be reliably evaluated with such models.Influence of oxygen concentration,temperature,granule diameter,sludge loading rate and cycle configuration have been analysed.Oxygen penetration depth in combination with the position of the autotrophic biomass played a crucial role in the conversion rates of the different components and thus on overall nutrient removal efficiencies (de Kreuk et al.,Accepted ).When a more detailed insight in microbial or EPS distribution inside the granule is desired or to study factors influencing granule shape (presence of filamentous outgrowth,biofilm structure modelling),individually based modelling can be used (Kreft et al.,2001).Such a model for aerobic granular sludge was presented by TU Delft and Universidade Nova de Lisboa,Portugal,to describe an aerobic granular sludge SBR (Xavier et al.,Submitted ).This multiscale model described the granular sludge reactor in detail,from the metabolism of microbial groups,through the spatial structure of granules to the dynamics of the whole reactor.However,simulations were computationally demanding and use will be restricted to a description of observed trends.With the model a preferential distribution of species along radial distances was shown,which were more heterogeneous than in strict layers.This heterogeneous structure and growth in microcolonies underlined the difficulty of representative micro-electrode measurements.Also,the accumulation of inert material in the cores of the granules was shown (Figure 3).Mainly the outer layers of the granule will be eroded,which contain less inert material.Therefore,aerobic granular sludge is expected to contain more inert material resulting from biomass decay than activated sludge.M.K.deKreuketal.78Waseda University (Japan)and TU Delft developed a similar multi-scale model with nitrifying granules.This model showed that within the nitrifying granules,EPS producing heterotrophic organisms will grow on products from cell-lyses.These heterotrophic organ-isms denitrify part of the produced nitrate and grow mainly inside of the granules.At this position,they excrete EPS,which strengthens the structure of the granule and enhances the growth of the total granule.The nitrifying organisms,growing mainly in the outer layers,will detach more which leads to a smaller sludge residence time for these species.Different effects of shear stress were not taken into account in the models as presented during the workshop.Dr.Ivanov,Nanyang University,Singapore pointed out that mechanical compaction of granules by shear stress caused by particle-particle collisions,air-bubbles or mechanical stirring might affect the distribution of microorgan-isms,detachment and porosity of the granular sludge.Changing porosity can affect the diffusion in the pores of the granule and detached parts from the granules can start new granules,both leading to a different microbial composition of the granule then described with current models.Aerobic granular sludge in practiceIn wastewater treatment,the activated sludge process is the dominant system.Biology of these systems has been optimised and the limits of the system have been reached.How-ever,sludge settleability and washout from clarifiers will remain a point of attention.Major concerns of water authorities are:minimising costs of wastewater treatment;meet-ing effluent requirements;being prepared for future developments in effluent demands;area availability;environmental aspects as smell and noise;energy consumption.Aerobic granular sludge could (partly)meet these concerns.Following successful pilot-plant studies at two sewage treatment plants in The Nether-lands by DHV,a first full-scale municipal wastewater treatment plant in The Netherlands,based on aerobic granular sludge is planned (presented by waterboard “Hollandse Delta”).As end-users,waterboards in The Netherlands see it as their (public)responsibil-ity to cooperate with consultancies and universities to develop this innovation and invest in full-scale applications.Several wastewaters from industrial and municipal origin were used to cultivate aerobic granules and to study the treatment process and the results were presented during the workshop;a pilot plant study (two reactors of 1.5m 3each)with the Nereda e system treating sewage at a Dutch wastewater treatment plant (presented by DHV Water,The Netherlands);a demonstration scale sequencing batch biofilter granular reactor (SBBGR,3.1m 3)treating sewage at a Italian wastewater treatment plant (presented byIRSA,(a)010203040506001234567X_NH X_NO X_PAO Inerts P A O ,i n e r t (gC OD /m 3)Ni t rifyers(g C OD/m 3)r (m)00.00020.0004(b) (c)Figure 3(a)Stained granule section (green ¼live cells;red ¼death cells);b:biomass distributioninside a matured granule according a simulation with an ASM-like aerobic granular sludge model (de Kreuk et al.,Accepted );c:biomass distribution inside a mature granule according a simulation with a 2-Dindividually based aerobic granular sludge model (red ¼PAO;green ¼NH 4þ-oxidisers;yellow ¼NO 22-oxidisers;grey ¼inert,Xavier et al.,Submitted ).Subscribers to the online version of Water Science and Technology can access the colour version of this figure from /wstM.K.de Kreuk et al.79Italy);laboratory scale experiments using livestock wastewater,containing high concentrations of nitrogen and phosphate (presented by Waseda University,Japan);using anaerobically pretreated abattoir effluent (presented by AWMC,Australia)and using low strength sewage (presented by Sumitomo Heavy Industries Ltd.,Japan).The main conclusion from all presentations was that the different aerobic granular sludge technologies were well applicable on tested wastewater;granules could be formed with good stability and conversion processes were satisfying.However,the more TSS in the wastewater and the more diluted the sewage,the less suitable tested systems were.It should be pointed out that in all experiments alternating anaerobic and aerobic periods were used,except for the SBBGR experiment and the experiment with diluted sewage,in which a pulse feed followed by aeration was used.Conclusions Research in aerobic granular sludge developed from mostly laboratory scale experimental research to modelling studies and studies at pilot and practical scale.The recommen-dation of the workshop 2004was that the pioneering stage of aerobic granular sludge research was finished.Researchers should continue with specific research topics (e.g.fac-tors influencing granule strength and formation,microbial diversity,conversion processes,pathogen removal,pilot-and demonstration scale studies)and that the technology should be put into practice.Partly this recommendation has been fulfilled;however,still many specific aspects about aerobic granular sludge formation,stability,diversity and process optimisation are unrevealed.One of the main concerns now for laboratory scale research is that a clear and standardised method for measuring EPS should be defined in the future,as only then the function of EPS in granule formation can be studied.For applications of aerobic granular sludge in sewage treatment or industry,the end-users need to make the first steps now towards this new technology and to allow research to obtain more specific insights in self aggregation of microorganisms and all related aspects.Acknowledgements We thank the participants from the aerobic granular sludge workshop 2006for their enthusiastic and active involvement and IWA for supporting this workshop.References Bathe,S.,De Kreuk,M.K.,Mc Swain,B.S.and Schwarzenbeck,N.(eds)(2005).Aerobic Granular Sludge IWA Publishing,LondonWater and Environmental Management Series (WEMS).Beun,J.J.,Hendriks,A.,van Loosdrecht,M.C.M.,Morgenroth,M.,Wilderer,P.A.and Heijnen,J.J.(1999).Aerobic granulation in a sequencing batch reactor.Wat.Res.,33(10),2283–2290.Beun,J.J.,van Loosdrecht,M.C.M.and Heijnen,J.J.(2001).N-removal in a granular sludge sequencing batch airlift reactor.Biotechnol.Bioeng.,75(1),82–92.Dangcong,P.,Bernet,N.,Delgenes,J.-P.and Moletta,R.(1999).Aerobic granular sludge–a case report.Wat.Res.,33(3),890–893.De Kreuk,M.K.and van Loosdrecht,M.C.M.(2004).Selection of slow growing organisms as a means for improving aerobic granular sludge stability.Wat.Sci.Technol.,49(11–12),9–19.De Kreuk,M.K.,De Bruin,L.M.M.and van Loosdrecht,M.C.M.(2005a).Aerobic granular sludge;from idea to pilot plant.In Aerobic Granular Sludge ,Bathe,S.,De Kreuk,M.K.,Mc Swain,B.S.and Schwarzenbeck,N.(eds),IWA,London,UK,pp.111–123.De Kreuk,M.K.,McSwain,B.S.,Bathe,S.,Tay,S.T.L.,Schwarzenbeck,N.and Wilderer,P.A.(2005b).Discussion outcomes.In Aerobic Granular Sludge ,Bathe,S.,De Kreuk,M.K.,Mc Swain,B.S.and Schwarzenbeck,N.(eds),IWA,London,UK,pp.153–169.M.K.de Kreuk et al.80De Kreuk,M.K.,Picioreanu,C.,Hosseini,M.,Xavier,J.B.and van Loosdrecht,M.C.M.(Accepted).Kinetic Model of a Granular Sludge SBR–Influences on Nutrient Removal.Biotechnol.Bioeng. Dulekgurgen,E.,Yesiladali,K.,Ovez,S.,Tamerler,C.,Artan,N.and Orhon,D.(2003).Conventional morphological and functional evaluation of the microbial populations in a sequencing batch reactor performing EBPR.J.Environ.Sci.Health Part a-Toxic/Hazardous Substances&EnvironmentalEngineering,38(8),1499–1515.Heijnen,J.J.and van Loosdrecht,M.C.M(1998)Method for acquiring grain-shaped growth of amicroorganism in a reactor,Patent Cooperation Treaty(PCT),pp.13.Technische Universiteit Delft,US, European patent.Kreft,J.U.,Picioreanu,C.,Wimpenny,J.W.T.and van Loosdrecht,M.C.M.(2001).Individual-based modelling of biofilms.Microbiology-Sgm.,147,2897–2912.Krishna,C.and van Loosdrecht,M.C.M.(1999).Effect of temperature on storage polymers and settleability of activated sludge.Wat.Res.,33(1),2374–2382.Liu,Y.and Tay,J.-H.(2004).State of the art of biogranulation technology for wastewater treatment.Biotechnol.Adv.,22(7),533–563.Liu,Y.,Yang,S.-F.and Tay,J.-H.(2004a).Improved stability of aerobic granules by selecting slow-growing nitrifying bacteria.J.Biotechnol.,108(2),161–169.Liu,Y.Q.,Liu,Y.and Tay,J.H.(2004b).The effects of extracellular polymeric substances on the formation and stability of biogranules.Appl.Microbiol.Biotechnol.,65(2),143–148.Lu¨bken,M.,Schwarzenbeck,N.,Wichern,M.and Wilderer,P.A.(2005).Modelling nutrient removal of an aerobic granular sludge lab-scale SBR using ASM3.In Aerobic Granular Sludge,Bathe,S.,De Kreuk, M.K.,Mc Swain,B.S.and Schwarzenbeck,N.(eds.),IWA,London,UK,pp.103–110.McSwain,B.S.,Irvine,R.L.,Hausner,M.and Wilderer,P.A.(2005).Composition and distribution of extracellular polymeric substances in aerobicflocs and granular sludge.Appl.Env.Microbiol.,71(2), 1051–1057.Morgenroth,E.,Sherden,T.,van Loosdrecht,M.C.M.,Heijnen,J.J.and Wilderer,P.A.(1997).Aerobic granular sludge in a sequencing batch reactor.Wat.Res.,31(12),3191–3194.AQUASIM2.0-Computerprogram for the Identification and Simulation of Aquatic Systems2.0.EAWAG, Dubendorf.Schwarzenbeck,N.,Erley,R.and Wilderer,P.A.(2004).Aerobic granular sludge in an SBR–system treating wastewater rich in particulate matter.Wat.Sci.Technol,49(11–12),41–46.van Loosdrecht,M.C.M.and De Kreuk,M.K.(2004)Method for the treatment of waste water with sludge granules,Bureau voor Industriele Eigendom.Technische Universiteit Delft,The Netherlands,International.van Loosdrecht,M.C.M.,Eikelboom,D.H.,Gjaltema,A.,Mulder,A.,Tijhuis,L.and Heijnen,J.J.(1995).Biofilm structures.Wat.Sci.Technol.,32(8),35–43.van Loosdrecht,M.C.M.,Picioreanu,C.and Heijnen,J.J.(1997a).A more unifying hypothesis for biofilm structures.FEMS Microbiol.Ecol.,24(2),181–183.van Loosdrecht,M.C.M.,Pot,M.A.and Heijnen,J.J.(1997b).Importance of bacterial storage polymers in bioprocesses.Wat.Sci.Technol.,35(1),41–47.Xavier,J.B.,de Kreuk,M.K.,Picioreanu,C.and van Loosdrecht,M.C.M.(Submitted).Microbial populations dynamics and nutrient conversions in aerobic granular sludge described by multi-scale modelling.Environ.Sci.Technol.M.K. de Kreuk et al.81。

REVIEWPlant tolerance to drought and salinity:stress regulating transcription factors and their functional significance in the cellular transcriptional networkDortje Golldack •Ines Lu¨king •Oksoon Yang Received:18February 2011/Revised:25March 2011/Accepted:25March 2011/Published online:8April 2011ÓSpringer-Verlag 2011Abstract Understanding the responses of plants to the major environmental stressors drought and salt is an important topic for the biotechnological application of functional mechanisms of stress adaptation.Here,we review recent discoveries on regulatory systems that link sensing and signaling of these environmental cues focusing on the integrative function of transcription activators.Key components that control and modulate stress adaptive pathways include transcription factors (TFs)ranging from bZIP,AP2/ERF,and MYB proteins to general TFs.Recent studies indicate that molecular dynamics as specific homodimerizations and heterodimerizations as well as modular flexibility and posttranslational modifications determine the functional specificity of TFs in environ-mental adaptation.Function of central regulators as NAC,WRKY,and zinc finger proteins may be modulated by mechanisms as small RNA (miRNA)-mediated posttran-scriptional silencing and reactive oxygen species signaling.In addition to the key function of hub factors of stress tolerance within hierarchical regulatory networks,epige-netic processes as DNA methylation and posttranslational modifications of histones highly influence the efficiency of stress-induced gene prehensive elucidation of dynamic coordination of drought and salt responsive TFs in interacting pathways and their specific integration in the cellular network of stress adaptation will provide newopportunities for the engineering of plant tolerance to these environmental stressors.Keywords Drought ÁEpigenetics ÁTranscription factor ÁRNAi ÁSalt tolerance ÁArabidopsisIntroductionA major challenge for current agricultural biotechnology is to satisfy an ever increasing demand in food production facing a constantly increasing world population that will reach more than 9billion in 2050(Godfray et al.2010;Tester and Langridge 2010).This growing demand for food is paralleled by dramatic losses of arable land due to increasing severity of soil destruction by abiotic environ-mental conditions.Thus,drought and salinity are the two major environmental factors that adversely affect plant growth and development and have a crucial impact on agricultural productivity and yields.Drought due to short-age of water is critical for crop production in large agro-nomic areas worldwide and it is usually coped with extensive irrigations.Although earth is rich in water,most water resources are highly salinized whereas high quality fresh water that is suitable for irrigation is extremely lim-ited.Accordingly,not only drought but also soil salinity becomes increasingly an agricultural problem due to extensive spreading of agricultural practices as irrigation (Flowers 2004)and it urgently requires the breeding of crops with increased water use efficiency and salt tolerance.Exposure of plants to excess salt causes ion imbalance and ion toxicity-induced imbalances in metabolism.Another component of salinity is hyperosmotic stress that results in water deficit in a comparable way to drought-induced water deficit.Plants basically counteract theCommunicated by R.Reski.D.Golldack (&)ÁI.Lu¨king ÁO.Yang Department of Biochemistry and Physiology of Plants,Faculty of Biology,Bielefeld University,33615Bielefeld,Germanye-mail:dortje.golldack@uni-bielefeld.dePlant Cell Rep (2011)30:1383–1391DOI 10.1007/s00299-011-1068-0negative effects of salinity and drought by activation of biochemical responses that include(1)the synthesis and accumulation of osmolytes,(2)maintaining the intracel-lular ion homeostasis,and(3)scavenging of reactive oxygen species(ROS)generated as a secondary effect of drought(Flowers2004;Ashraf and Akram2009).Plant engineering strategies for cellular and metabolic reprogramming to increase the efficiency of plant adaptive processes may either focus on(1)conferring stress toler-ance by directly re-programming ion transport processes and primary metabolism or(2)by modulating signaling and regulatory pathways of the adaptive mechanisms.The second approach seems to be more perspective because it is likely that signaling and regulatory factors orchestrate as key signaling components the transcriptional and transla-tional control of group(1)adaptive mechanisms(Die´dhiou et al.2008;Popova et al.2008).Accordingly,molecular re-programming to enhance stress tolerance of plants would probably require the genetic engineering of a single or a few master regulators of adaptation instead of modulating numerous metabolic and cellular adaptive mechanisms.However,although several plant stress signaling com-ponents have been dissected in detail the knowledge on integration of regulatory mechanisms in stress signaling cascades and on key regulators is still limited,although knowledge on the regulating key factors of stress adap-tation is highly necessary for biotechnological engineering of stress tolerance.In this review,we focus on recent advances in transcription factor(TF)-based engineering of increased drought and salt adaptation.Putative integra-tions and links of TFs in stress adaptive signaling net-works coordinating the endogenous programs of environmental adaptation will be highlighted.Accord-ingly,for this review TFs out of the wider range of all stress inducible TFs were selected so that we do not comprehensively cover all stress-related factors.Major criterion for this selection was a putative potential of the TFs in controlling sub-regulons of stress-adaptive cellular mechanisms within the hierarchical transcriptional net-work that will be discussed in the review.Arabidopsis and related model species:learningfrom species with different natural stress tolerance Traditional breeding attempts for sustainable agricultural use of dry and salinized soils have been clearly facilitated and stimulated by the wealth of knowledge of genomics and transcriptomics data available from the model species Arabidopsis(Arabidopsis thaliana)and rice(Oryza sativa). Linear general frameworks of plant drought and salt adaptation have been established that were mainly based on systematic and comprehensive mutant analyses.Thus,it is now accepted that changes in membrane integrity and modulation of lipid synthesis are key factors in the primary sensing of drought and salt(Kader and Lindberg2010).Secondary,osmotic stress-induced sig-naling involves changes in plasma membrane H?-ATPase and Ca2?-ATPase activities that trigger concerted changes of Ca2?influx,cytoplasmic pH,and apoplastic production of ROS(Beffagna et al.2005).In addition,osmotic stress-induced Ca2?fluxes are linked to abscisic acid(ABA),and calcium-responsive protein kinases act as key regulators in drought and salinity-induced signaling cascades(Die´dhiou et al.2008).As convergent down-stream elements of transcriptional activation,many genes that are responsive to drought and to salinity belong to the ABA-responsive element(ABRE)and dehydration-responsive element/C-repeat element(DRE/CRT)regulons(Yamaguchi-Shino-zaki and Shinozaki2005).Despite this knowledge derived from the model plants Arabidopsis and rice,the applicability of these data for biotechnological engineering of increased drought and salt tolerance is clearly prehensive comparisons of the salt inducible transcriptomes of the salt-sensitive spe-cies Arabidopsis and rice and,for example,transcriptional data of the closely related salt-tolerant model species Lobularia maritima(Brassicaceae)and Festuca rubra ssp. litoralis(Poaceae)show extensive differences in salt responsive expressional regulations(Popova et al.2008; Die´dhiou et al.2009b;Fig.1).In contrast to salt excluding and avoiding halophyte models as Thellungiella halophila with a very limited number of salt responsive transcripts, the salt-accumulating and-detoxifying halophytes L. maritima and F.rubra ssp.litoralis allowed identification of a wide range of transcripts with different salt responsive regulation in the salt-sensitive and salt-tolerant species (Volkov et al.2003;Taji et al.2004;Popova et al.2008; Die´dhiou et al.2009b).In addition,transgenic modulation of regulatory and signaling elements in Arabidopsis and rice according to the pattern in the halophytes L.maritima and F.rubra ssp.litoralis successfully activated stress adaptation in the sensitive model species(Die´dhiou et al. 2008;Yang et al.2009).Accordingly,understanding of stress-induced signaling complexity in stress-sensitive model species has to be complemented by comparisons with naturally tolerant species for a systematic identifica-tion of key regulators of stress tolerance with the potential of biotechnological application.bZIP TFs and their role in conferring stress tolerance to plantsResearch on salt and drought regulatory TFs has mainly focused on single factors and linear pathways.Emergingfindings increasingly suggest,however,integration of the TFs in dynamic network hubs as well as interaction and competition of pathways manifesting complexity of molecular links in stress adaptation.The emerging view of the salt-and drought-signaling network unequivocally supports a key and integrative function of members of the bZIP TFs in these regulatory networks(Fig.2)and the potential of these factors to confer enhanced stress tolerance has been demonstrated repeatedly.A key regulator of salt stress adaptation,the group F bZIP TF bZIP24,was identified by differential screening of salt-inducible transcripts in A.thaliana and a halophytic Arabidopsis-relative model species(Yang et al. 2009).Expressional regulation of bZIP24was different with induced transcription in the salt-sensitive and tran-scriptional repression in the halotolerant species,and RNAi-mediated repression of the factor conferred increased salt tolerance to Arabidopsis.The improved tolerance was mediated by stimulated transcription of a wide range of stress-inducible genes that are e.g.involved in cytoplasmic ion homeostasis,osmotic adjustment,as well as in plant growth and development demonstrating a central function of bZIP24in salt tolerance by regulatingmultiple mechanisms that are essential for stress adaptation (Yang et al.2009).Next to bZIP24and its function in salt adaptation,group A bZIP factors AREB1,AREB2,and ABF3have a key regulatory role in ABA signaling under drought stress.Thus,A.thaliana areb1areb2abf3triple knock out mutants had increased tolerance to ABA and reduced drought tolerance(Yoshida et al.2010).In addi-tion,in other species as rice and tomato transgenic modi-fication of group A bZIP TFs modified the tolerance of plants to water deficit and to salt stress(Amir Hossain et al. 2009;Hsieh et al.2010)strongly suggesting trans-species potential of these factors for increasing stress tolerance.From animal systems dynamic coordinations of numer-ous bZIP controlled signal transduction pathways by molecular re-organization and by posttranslational mech-anisms are well-known(Jindra et al.2004;Miller2009). Thus,specific homodimerizations and heterodimerizations within the class of bZIP TFs as well as modularflexibility of the interacting proteins and posttranslational modifica-tions might determine the functional specificity of bZIP factors in cellular transcription networks(Miller2009). Excitingly,evidences for involvement of homologous mechanisms in signaling hubs in plant systems are just now 20304050At Lm Os Fremerging.As an example,the three factors AREB1,AREB2,and ABF3can form homodimers and heterodi-mers as well as interact with a SnRK2protein kinase suggesting ABA-dependent phosphorylation of the proteins (Yoshida et al.2010).As another example for the function of bZIP factors in salt adaptation in A.thaliana ,salt stress induced proteolytic processing and translocation of the group B factor AtbZIP17to the nucleus followed by transcriptional up-regulation of salt-responsive transcripts (Liu et al.2007).The group F factor AtbZIP24shows salt-inducible subcellular re-targeting to the nucleus and for-mation of homodimers suggesting that molecular dynamics of bZIP factors could mediate new signaling connections within the complex cellular signaling network (Yang et al.2009).In contrast to the homodimerization of bZIP24,specific heterodimerization was shown for the salt-responsive group S AtbZIP1with group C bZIP TFs (Weltmeier et al.2009).In conclusion,it might be hypothesized that specific homodimerizations and hetero-dimerizations as well as posttranslational modifications (e.g.phosphorylations)might determine the functional specificity of bZIP factors in the cellular transcription networks of drought and salt adaptation.Interestingly,transgenic over-expression of rice SnRK2-type SAPK4in rice regulated ion and ROS homeostasis under salt stresssupporting the hypothesis of key functions of SnRK kina-ses in the intracellular signaling cascades of osmotic adaptation thus further supporting key modulatory function of posttranslational phosphorylations in diverse plant sys-tems that might,e.g.target bZIP factors (Die´dhiou et al.2008;Fig.2).Recently,it was recognized that general TFs might also have an important role in stress-responsive transcription.Thus,the TBP-associated factor AtTAF10has a specific and key function in plant salt and osmotic stress adaptation by regulating accumulation of Na ?and proline (Gao et al.2006).This functional overlap to bZIP24(Yang et al.2009)strongly suggests linked regulation and cofunctions of bZIP proteins and TAFs within the complex drought and salt signaling network—a hypothesis that awaits further clarification (Fig.2).The role of WRKY TFs and Cys2/His2zinc finger proteins in the regulation of adaptation to osmotic stressOur understanding of plant stress-inducible signaling has been greatly facilitated by research on TFs that regulate and control subsets of stress-responsive geneexpression.Fig.2Model of signaling pathways and regulatorytranscription factors involved in plant adaptation to drought and saltThus,WRKY proteins regulate diverse plant processes ranging from development to various biotic and abiotic stresses as well as hormone-mediated pathways(Rama-moorthy et al.2008).Involvement of WRKY factors in plant salt adaptation were shown for WRKY25and WRKY33that increased salt tolerance and ABA sensitivity independent of the SOS-pathway when over-expressed in A.thaliana(Jiang and Deyholos2009).In A.thaliana, wrky63knock out mutants showed decreased sensitivity to ABA and drought(Ren et al.2010).In these plants,the stomatal closure and the expression of the AREB1/ABF2 TF were affected indicating involvement of WRKY fac-tors in the ABA-dependent pathway of drought and salt adaptation(Ren et al.2010).Potential of WRKY-type TFs to confer increased salt tolerance by transgenic expression is further supported by the different salt-induced regula-tion of a WRKY protein in salt-sensitive rice and a hal-ophytic rice-relative model species(Die´dhiou et al.2009a, b).Interestingly,A.thaliana WRKY25and WRKY33are not only responsive to osmotic stresses but they are also regulated by oxidative stress(Miller et al.2008).In addition,down-stream regulated target genes of WRKY33 include transcripts with function in ROS detoxification as peroxidases and glutathione-S-transferases(Jiang and Deyholos2009)suggesting function of WRKY factors as key regulators in both osmotic and oxidative stress adaptation.Alternatively,it is tempting to hypothesize involvement of WRKY factors in the osmotic stress sig-naling via control of the intracellular stress-induced ROS levels(Fig.2).Interestingly,Zat proteins(TFIIIA-type Cys2/His2zinc finger proteins)have been suggested to control and regulate WRKY functions(Miller et al.2008).Thus,in soybean overexpression of GmWRKY54conferred increased salt and drought tolerance and regulation of the GmWRKY54 by Zat10/STZ was hypothesized(Zhou et al.2008).In addition,in rice stomatal closure is regulated by the Cys2/ His2zincfinger protein DST(drought and salt tolerance) via ABA-independent targeting of genes that are involved in ROS homeostasis(Huang et al.2009).Thesefindings further support involvement of zincfinger proteins and probably WRKY TFs in osmotic adaptation via ROS sig-naling(Fig.2).Interestingly,although both drought and salt stress might result in intracellular accumulation of toxic amounts of ROS,hydrogen peroxide(H2O2)and nitric oxide(NO)also function as signaling molecules in ABA-mediated stomatal responses(Miller et al.2010; Wilkinson and Davies2010).Mutation of a cellulose synthase-like protein induced accumulation of ROS, changed sensitivity to salt stress and to water deficit,and regulation of plant osmotic stress tolerance via control of intracellular stress-induced ROS levels has been suggested (Zhu et al.2010a).Stress adaptation and multi-transcriptional regulation: AP2/ERF,MYB,and bHLH TFsNext to TFs with possible upstream position in the hier-archical network of stress adaptation as the bZIP factors described above,integrative stress-adaptive functional roles of regulatory proteins from other diverse groups have been reported.These factors might be either integrated in the main pathways of environmental adaptation,likely under control of the key regulatory TFs,or they might have functions in regulating sub-networks of adaptation to drought and salt stress and in linking these stress adapta-tions to other stresses,developmental and hormonal responses.Thus,dual roles in both biotic and abiotic stress responses have been demonstrated for AP2/ERF proteins as soybean GmERF3and the ABA-responsive RAP2.6from A.thaliana(Zhang et al.2009;Zhu et al.2010b).Over-expression of Arabidopsis light and drought responsive RAP2.4led to defects in multiple developmental processes regulated by light and ethylene as well as drought tolerance (Lin et al.2008).Complementary to these observations, overexpression of AP2/ERF GmERF3in tobacco induced the expression of PR genes and of osmotin accompanied by enhanced accumulation of free proline and soluble carbo-hydrates(Zhang et al.2009).Members of the DREB/CBF subfamily of the AP2/ERF TFs have been recognized for a decade for their roles in stress tolerance via ABA-depen-dent and-independent pathways and for their regulation of a stress-response sub-transcriptome with more than hun-dred target genes inclusive regulatory factors as ZAT12 and RAP2.1(Shinozaki and Yamaguchi-Shinozaki2000). However,constitutive overexpression of the DREB/CBF pathway led to serious developmental defects of transgenic plants although accompanied by increased tolerance to drought,salt,and cold(Kasuga et al.1999).These data clearly demonstrate complexity of the stress adaptive net-work that requires major control points of the multiple transcriptional sub-regulons as well as cooperative and integrative function of the different stress sub-clusters to prevent impairing side effects.Nevertheless,members of the AP2/ERF TF family are integrated as a hub in signaling interconnections of complex biotic and abiotic environ-mental cues.Supporting the undeniable key function of AP2/ERF in terms of drought and salt tolerance the picture of integrative function of these factors in plant develop-mental processes as well as biotic and/or abiotic stress signaling in an interconnecting and linking way is,how-ever,only emerging.As another example for multi-functional regulations,the R2R3-MYB TF AtMYB41is transcriptionally induced in response to ABA,drought,salinity,and cold(Lippold et al. 2009).In addition,the factor influences cell expansion and cuticle deposition suggesting a linking function in abioticstress response and cell wall modifications(Cominelli et al. 2008).Interaction and competition of complex signaling pathways infine-tuning cellular responses is further illustrated by the A.thaliana basic-helix-loop helix TF bHLH92.The factor regulates only the expression of a subset of salt-and drought-responsive genes(Jiang et al. 2009).However,different peroxidases are down-stream targets of the factor and bHLH92might have a function in the control of ROS-mediated signaling thus linking salt and drought adaptation to ROS signaling(Fig.2). Here,more detailed work will be necessary to elucidate the precise integration of the diverse TFs in the cellular network of stress adaptation and to understand their potential in genetic engineering of improved stress tol-erance,probably via targeted engineering of defined subsets of stress adaptive mechanisms or sub-pathways of signaling to customize specific features of stress adaptation.NAC-triggered gene expression and miRNANAC type proteins are not only involved in diverse pro-cesses as developmental programs,defense,and biotic stress responses(Olsen et al.2005)but they also have a key function in abiotic stress tolerance inclusive drought and salinity.Thus,in rice ONAC5and ONAC6are transcrip-tionally induced by ABA,drought,and salt stress(Rabbani et al.2003;Takasaki et al.2010).ONAC5and ONAC6 transcriptionally activate stress-inducible genes as OsLEA3 by direct binding to the promoter and they interact in vitro suggesting functional dimerization of these TFs(Takasaki et al.2010).Interestingly,overexpression of the Arabid-opsis factors ANAC019,ANAC055,and ANAC072caused increased drought tolerance of transgenic plants but they only changed transcription of a limited number of non-particularly salt-and drought-responsive genes(Tran et al. 2004).These important results strongly suggest interaction or co-regulation of NAC factors with other regulatory pathways or subsets of stress-inducible molecular mecha-nisms for achieving the significant increased stress toler-ance that was observed(Tran et al.2004).Improved drought and salt tolerance could also be achieved by transgenic overexpression of diverse NAC factors in spe-cies ranging from A.thaliana and rice to chickpea,wheat, and tomato(Peng et al.2009;Yokotani et al.2009;Xia et al.2010;Yang et al.2011).Interestingly,in tomato two NAC TFs were salt-inducible in a salt-tolerant cultivar but showed different expression in salt-sensitive tomato plants (Yang et al.2011).These data indicate that differences in plant salt tolerance might be due to different and specific transcriptional activation of NAC-dependent regulatory pathways.As important examples for conferring increased stress tolerance underfield conditions,in rice transgenic over-expression of SNAC1enhanced salt and drought tolerance and OsNAC10improved drought tolerance and grain yield (Hu et al.2006;Jeong et al.2010).OsNAC10-regulated target genes mainly included protein kinases and TFs of AP2,WRKY,LRR,NAC,and Zn-finger types as well as the stress-responsive genes cytochrome P450and the potassium transporter HAK5(Jeong et al.2010).These results support the view that NAC type TFs might be part of the general frameworks of drought and salt adaptation by connecting or regulating subsets of linear adaptive pathways but the NAC factors themselves are likely to be controlled by global regulatory factors of the network of stress adaptive transcription and metabolism.Thus, important evidence for cooperative regulation of stress responses by members of different TF families was pro-vided by the study of Tran et al.(2007)that showed interaction and co-function of the drought,salt,and ABA inducible zincfinger protein ZFHD1and a NAC factor.As it was recognized recently,members of the CCAAT-HAP TF family also have a potential key function in conferring stress tolerance to crops.Transgenic maize plants with increased expression of the CCAAT-HAP-type factor ZmNF-YB2showed improved drought tolerance underfield conditions(Nelson et al.2007).This effect was achieved by mechanisms independent of ABA and DREB/ CBF pathways supporting the hypothesis of concerted action of different TF families within subsets of regulatory modules in the cellular stress-response network.Interestingly,members of the NAC TF family are potential regulatory targets of the small RNA(miRNA) posttranscriptional silencing machinery(Rhoades et al. 2002;Guo et al.2005).As an example,recently a NAC domain containing TF was identified as a target of miR164 in switchgrass(Matts et al.2010).Thus,regulation of NAC TFs by miRNA-mediated cleavage of mRNAs together with data showing differential regulation of NAC factors in response to drought and salt stress indicate that these TFs might participate in the regulation of environmental adaptation through miRNA pathways.Next to NAC pro-teins,TFs e.g.of SCL,MYB,and TCP types were iden-tified as targets of drought and salt inducible miRNAs as miR159,miR168,miR171,and miR396(Liu et al.2008). Accordingly,it might be hypothesized that the cellular networks of drought-and salt-stress tolerance are regulated by miRNA-mediated targeting of convergent and divergent adaptive pathways under control of different stress-specific TFs.Accordingly,relevance of modification of drought and salt stress-specific signaling pathways via the miRNA machinery in a biotechnological context might be a pow-erful approach for genetic engineering of improved toler-ance but remains to be discovered.Epigenetics:what is next in terms of biotechnological application?Next to transcriptional regulations of abiotic stress responses,epigenetic processes are becoming a new and current chapter in plant environmental adaptation.Effi-ciency of gene expression is highly influenced by chro-matin structure that might be modulated epigenetically by processes as DNA methylation and posttranslational mod-ifications of histones.The histone-mediated structure of nucleosomes in the chromatin might be posttranslationally modified at the N-terminal tails of the core histone com-plexes(H2A,H2B,H3,H4)and thus influence nucleosome density,binding efficiency of TFs,and transcriptional activity(Chinnusamy and Zhu2009;Kim et al.2010).In addition to methylations of histones,also acetylations and phosphorylations as well as other posttranslational modi-fications of histones as ubiquitination,biotinylation,and sumoylation might have a modulating impact on the reg-ulation of stress-specific gene expression(Chinnusamy et al.2008).Meanwhile,it is accepted knowledge that phenotypes within one species may transmit different epigenetic information based on covalent modifications of DNA or histones(Fazzari and Greally2004).Thus,plant popula-tions from stress exposed habitats may carry inherited memories of stress adaptation and transfer this epigeneti-cally to next generations.As an example,the desert shrub Zygophyllum dumosum was posttranslationally methylated at histone H3under wet but less under dry growth condi-tions indicating posttranslational regulation of gene expression activity(Granot et al.2009).As it was also reported recently,natural populations of mangroves were DNA hypomethylated when grown under saline conditions in contrast to populations from non-saline sites(Lira-Medeiros et al.2010).Based on these results,it seems obvious to think on simulation of inherited memories of stress adaptation in biotechnological applications to confer increased drought and salt tolerance to naturally sensitive species.However,in contrast to the detailed knowledge on influences of epigenetic mechanisms on developmental processes,information on epigenetic regulation of abiotic stress resistance is still rare.As a few examples,salinity-induced phosphorylation of histone H3and acetylation of histone H4in A.thaliana and tobacco have been reported(Sokol et al.2007).In addition, altered acetylation as well as trimethylation of histone H3 under drought stress in drought-responsive genes of A. thaliana have been observed(Kim et al.2008).In rice, expression of cytosine DNA methyltransferases was mod-ified by salt stress indicating functional importance of epigenetic modulation of genome activity also in monocot species(Sharma et al.2009).Detailed knowledge on the specific mechanisms that underlay epigenetic regulation under environmental expo-sure is,however,only slowly emerging.Thus,trans-gen-erational modifications of stress adaptations as salt stress include altered genomic DNA methylation as well as function of Dicer-like proteins suggesting involvement of small RNA pathways in epigenetic regulations(Boyko et al.2010).Interestingly,in barley expression of Poly-comb proteins with function in histone methylation was influenced by abscisic acid(ABA)suggesting involvement of ABA-mediated pathways in epigenetic modifications (Kapazoglou et al.2010).Thus,according to the current knowledge,an applica-tion of epigenetic processes to improve the stress-regulat-ing function of TFs will be a challenging and novel biotechnological approach for the engineering of plant tolerance to drought and salinity,however,many detailed information are still missing.Particularly,despite the importance of elucidating epigenetic mechanisms in model plants,it will be obligatory to extend investigations to systematic and comprehensive comparisons of stress rele-vant epigenetics in sensitive-and naturally tolerant species. Linking epigenetic processes to the key regulatory com-ponents of the general stress adaptive frameworks will be essential to further support the feasibility of epigenetics in the customized engineering of stress adaptation. Conclusion and perspectivesCellular effects of environmental stresses as drought and salinity are not only imbalances of ionic and osmotic homeostasis but also impaired photosynthesis,cellular energy depletion,and redox imbalances.Regulatory sys-tems inclusive TFs that link sensing and signaling of the environmental conditions and the cellular adaptive responses are emerging but are not well understood yet.As a next step,it will be important to identify master regula-tors and master pathways of stress adaptation in naturally stress-tolerant species as well as integration of the diverse regulatory factors in the network of intracellular stress adaptation pathways(Fig.2).Within this hierarchical net-work,cellular stress responses might befine tuned by interaction and competition of TFs that regulate sub-clus-ters of the stress transcriptome.Here,systematic and comprehensive data on the timing of all stress responsive TFs upon stress will be indispensable for detailed hierar-chical linking of all regulatory factors.In addition,more detailed understanding of shared and competing transcrip-tional regulation as well as modulated intramolecular interactions of different factors and epigenetic processes will be essential for targeted and efficient genetic engi-neering of improved drought and salt tolerance in plants.。

Arabidopsis EPSIN1Plays an Important Role in VacuolarTrafficking of Soluble Cargo Proteins in Plant Cells via Interactions with Clathrin,AP-1,VTI11,and VSR1WJinhee Song,Myoung Hui Lee,Gil-Je Lee,Cheol Min Yoo,and Inhwan Hwang1Division of Molecular and Life Sciences and Center for Plant Intracellular Trafficking,Pohang University of Scienceand Technology,Pohang790-784,KoreaEpsin and related proteins play important roles in various steps of protein trafficking in animal and yeast cells.Many epsin homologs have been identified in plant cells from analysis of genome sequences.However,their roles have not been elucidated.Here,we investigate the expression,localization,and biological role in protein trafficking of an epsin homolog, Arabidopsis thaliana EPSIN1,which is expressed in most tissues we examined.In the cell,one pool of EPSIN1is associated with actinfilaments,producing a network pattern,and a second pool localizes primarily to the Golgi complex with a minor portion to the prevacuolar compartment,producing a punctate staining pattern.Protein pull-down and coimmunoprecipitation experiments reveal that Arabidopsis EPSIN1interacts with clathrin,VTI11,g-adaptin-related protein(g-ADR),and vacuolar sorting receptor1(VSR1).In addition,EPSIN1colocalizes with clathrin and VTI11.The epsin1mutant,which has a T-DNA insertion in EPSIN1,displays a defect in the vacuolar trafficking of sporamin:greenfluorescent protein(GFP),but not in the secretion of invertase:GFP into the medium.Stably expressed HA:EPSIN1complements this trafficking defect.Based on these data,we propose that EPSIN1plays an important role in the vacuolar trafficking of soluble proteins at the trans-Golgi network via its interaction with g-ADR,VTI11,VSR1,and clathrin.INTRODUCTIONAfter translation in eukaryotic cells,a large number of proteins are transported to subcellular compartments by a variety of different mechanisms.Newly synthesized vacuolar proteins that are delivered to the endoplasmic reticulum(ER)by the cotrans-lational translocation mechanism are transported to the vacuole from the ER by a process called intracellular trafficking.Traffick-ing of a protein to the vacuole from the ER occurs through two organelles,the Golgi complex and the prevacuolar compartment (PVC)(Rothman,1994;Hawes et al.,1999;Bassham and Raikhel, 2000;Griffiths,2000).Transport of a protein from the ER to the Golgi complex is performed by coat protein complex II vesicles. Transport from the trans-Golgi network(TGN)to the PVC occurs via clathrin-coated vesicles(CCVs)(Robinson et al.,1998;Tang et al.,2005;Yang et al.,2005).Transport of a protein from the ER to the vacuole/lysosome requires a large number of proteins,including components of vesicles,factors involved in vesicle generation and fusion,reg-ulators of intracellular trafficking,adaptors for the cargo proteins, and other accessory proteins(Robinson and Kreis,1992;Bennett, 1995;Schekman and Orci,1996;da Silva Conceic¸a˜o et al.,1997;Kirchhausen,1999;Sever et al.,1999;Bassham and Raikhel, 2000;Griffiths,2000;Jin et al.,2001;Robinson and Bonifacino, 2001).Most of these proteins are found in all eukaryotic cells from yeast,animals,and plants,suggesting that protein traffick-ing mechanisms from the ER to the vacuole/lysosome may be highly conserved in all eukaryotic cells.Of the large number of proteins involved in intracellular traf-ficking,a group of proteins that have the highly conserved epsin N-terminal homology(ENTH)domain have been identified as playing a critical role at various trafficking steps in animal and yeast cells(Chen et al.,1998;De Camilli et al.,2002;Wendland, 2002;Overstreet et al.,2003;Legendre-Guillemin et al.,2004). The ENTH domain binds to phosphatidylinositols(PtdIns), although the lipid binding specificity differs with individual members of the epsin family.For example,epsin1binds to PtdIns(4,5)P2,whereas EpsinR and Ent3p bind to PtdIns(4)P and PdtIns(3,5)P2,respectively(Itoh et al.,2001).The ENTH domain is thought to be responsible for targeting these proteins to specific compartments and also for introducing curvature to the bound membranes to assist in the generation of CCVs(Legendre-Guillemin et al.,2004).However,the exact steps of intracellular trafficking in which ENTH-containing proteins play a role are complex.Epsin homologs can be divided into two groups based on the pathway in which they play a role.One group,which includes epsin1in animal cells and Ent1p and Ent2p in yeast cells,is involved in endocytosis from the plasma membrane (Chen et al.,1998;De Camilli et al.,2002;Wendland,2002).The other group,which includes EpsinR/clint/enthoprotin in animal cells and Ent3p and Ent4p in yeast cells,is involved in protein trafficking from the TGN to the lysosome/vacuole as well as1To whom correspondence should be addressed.E-mail ihhwang@postech.ac.kr;fax82-54-279-8159.The author responsible for distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors()is:Inhwan Hwang(ihhwang@postech.ac.kr).W Online version contains Web-only data./cgi/doi/10.1105/tpc.105.039123The Plant Cell,Vol.18,2258–2274,September2006,ª2006American Society of Plant Biologistsretrograde trafficking from the early endosomes to the TGN (Kalthoff et al.,2002;Wasiak et al.,2002;Hirst et al.,2003; Chidambaram et al.,2004;Eugster et al.,2004;Saint-Pol et al., 2004).Another common feature of epsin-related proteins is that they play a role in CCV-mediated protein trafficking at both the TGN and the plasma membrane.These proteins can bind directly to clathrin through their multiple clathrin binding motifs;thus,they may recruit clathrin to the plasma membrane or the TGN to generate CCVs(Rosenthal et al.,1999;Wendland et al.,1999; Drake et al.,2000).In addition,these proteins interact with many other proteins,such as heterotetrameric clathrin adaptor complexes(APs),monomeric adaptor Golgi-localized,g-ear–containing Arf binding proteins(GGAs),and soluble NSF attach-ment protein receptors(SNAREs).Epsin1interacts with AP-2, Epsin15,and intersectin(Chen et al.,1998;Legendre-Guillemin et al.,2004),whereas EpsinR/enthoprotin/clint and Ent3p interact with SNAREs such as vti1b and vti1p,respectively (Chidambaram et al.,2004)and with adaptor proteins such as GGAs and AP-1(Duncan et al.,2003;Mills et al.,2003).In addition,epsin homologs have ubiquitin-interacting motifs and are ubiquitinated(Oldham et al.,2002;Shih et al.,2002).Protein ubiquitination acts as a signal for endocytosis from the plasma membrane and trafficking from the TGN through the endosome/ PVC to the lysosome/vacuole(Polo et al.,2002;Horak,2003; Raiborg et al.,2003;Scott et al.,2004).The binding of epsin homologs to ubiquitin raises the possibility that epsin homologs may bind directly to cargo proteins that are destined for the vacuole/lysosome from either the plasma membrane or the TGN (Chen and De Camilli,2005;Sigismund et al.,2005).In plant cells,sequence analysis of the entire Arabidopsis thaliana genome reveals several proteins with the highly con-served ENTH domains(Holstein and Oliviusson,2005).However, their biological roles have not been addressed.In this study,we investigate the functional role of EPSIN1,an Arabidopsis epsin homolog,at the molecular level.In particular,we focus on its possible role in protein trafficking in plant cells.We demonstrate that EPSIN1interacts with clathrin,AP-1,VSR1,and VTI11and plays an important role in the vacuolar trafficking of a soluble protein from the Golgi complex to the central vacuole.RESULTSEPSIN1,a Member of the Epsin Family,Is Ubiquitously Expressed in ArabidopsisThe Arabidopsis genome encodes three highly similar epsin-related proteins,EPSIN1,EPSIN2,and EPSIN3(Holstein and Oliviusson,2005).In this study,we investigated the biological role of EPSIN1.EPSIN1has the highly conserved ENTH domain at the N terminus.However,the rest of the molecule is less similar to other epsin-related proteins,although it has motifs,such as LIDL and DPF,that may function as clathrin and AP-1binding motifs,respectively.To understand the biological role of EPSIN1,its expression in various plant tissues was examined.An antibody was raised against the middle domain of EPSIN1(amino acid residues153to 337).The antibody recognized a protein band at90kD,which was much larger than the expected size,60kD,of EPSIN1 (Figure1A).It was shown previously that epsin-related proteins migrate slower than expected in SDS-PAGE(Chen et al.,1998). The control serum did not recognize any protein bands.This re-sult suggested that the antibody specifically recognized EPSIN1. To confirm this,protoplasts were transformed with EPSIN1 tagged with HA at the N terminus(HA:EPSIN1)and protein extracts from the transformed protoplasts were analyzed by protein gel blotting using anti-HA and anti-EPSIN1antibodies. The anti-HA antibody specifically recognized a protein band from the transformed protoplasts,but not from the untransformed protoplasts,at90kD(Figure1B).In addition,the90-kD protein species was recognized by the anti-EPSIN1antibody,confirming that the90-kD band was EPSIN1.The expression of EPSIN1in various tissues was examined using the anti-EPSIN1antibody. Protein extracts were prepared from various tissues at different stages of plant growth and used for protein gel blot analysis. EPSIN1was expressed in all of the tissues examined,with the highest expression in cotyledons andflowers(Figure1C). EPSIN1Produces Both Network and PunctateStaining PatternsTo examine the subcellular distribution of EPSIN1,total protein extracts from leaf tissues were separated into soluble and membrane fractions and analyzed by protein gel blotting using anti-EPSIN1antibody.EPSIN1was detected in both membrane (pellet)and soluble fractions(Figure2A).As controls for the fractionation,Arabidopsis aleurain-like protease(AALP)and Arabi-dopsis vacuolar sorting receptor(VSR)were detected with anti-AALP and anti-VSR antibodies,respectively(Sohn et al.,2003). AALP is a soluble protein present in the vacuolar lumen,and VSR is a membrane protein that is localized primarily to the PVC with a minor portion to the Golgi complex(da Silva Conceic¸a˜o et al., 1997;Ahmed et al.,2000).As expected,AALP and VSR were detected in the supernatant and pellet fractions,respectively. These results indicated that EPSIN1localized to multiple loca-tions,consistent with the behavior of other epsin-related proteins (Legendre-Guillemin et al.,2004).Next,we defined the subcellular localization of EPSIN1.Our initial attempts to localize the endogenous EPSIN1with the anti-EPSIN1antibody failed.Thus,we determined the localization of EPSIN1protein transiently expressed in protoplasts.EPSIN1 was tagged with the HA epitope,greenfluorescent protein(GFP), or redfluorescent protein(RFP).The amount of total EPSIN1 protein was determined using various amounts of HA:EPSIN1 plasmid DNA by protein gel blot analysis with anti-EPSIN1an-tibody and was found to be proportional to the amount of plasmid used(Figure2B).For the localization,we used a minimal amount(5to10m g)of EPSIN1plasmid DNAs.Protoplasts were transformed with HA:EPSIN1,and localization of EPSIN1 was determined by immunostaining with anti-HA antibody.HA: EPSIN1produced primarily a punctate staining pattern(Figure 2Ca).In addition to punctate stains,we occasionally observed weakly stained strings that connected punctate stains(Figure 2Cc,arrowheads).By contrast,the nontransformed controls did not produce any patterns(Figure2Ce).In protoplasts trans-formed with EPSIN1:GFP and EPSIN1:RFP,both EPSIN1fusionEPSIN1in Vacuolar Trafficking2259proteins produced a network pattern with punctate stains (Fig-ures 2Cg and 2Ch),whereas GFP and RFP alone produced diffuse patterns (Figures 2Dh and 2Di),indicating that EPSIN1produces the network pattern with punctate stains.These results were further confirmed by cotransforming the protoplasts with either EPSIN1:GFP and HA:EPSIN1or EPSIN1:GFP and EPSIN1:RFP .The punctate staining pattern of EPSIN1:GFP closely over-lapped that of HA:EPSIN1(Figures 2Da to 2Dc).In addition,the network and punctate staining patterns of EPSIN1:GFP closely overlapped those of EPSIN1:RFP (Figures 2De to 2Dg).However,the fine networks revealed by EPSIN1:GFP in the live protoplasts were nearly absent in the fixed protoplasts.Thus,the differences in the staining patterns between fixed and live protoplasts may be attributable to the fact that the network pattern of live protoplasts are not well preserved under the fixing conditions used.In addi-tion,the strings occasionally observed in the fixed protoplasts may represent the remnants of the network pattern revealed by HA:EPSIN1.These results strongly suggest that EPSIN1is re-sponsible for the network pattern as well as the punctate stains.The network pattern was reminiscent of the ER or actin pattern in plant cells (Boevink et al.,1998;Jin et al.,2001;Kim et al.,2005),whereas the punctate staining pattern suggested that EPSIN1may localize to the Golgi complex or endosomes,as observed previously with epsin homologs in animal and yeast cells (Wasiak et al.,2002;Chidambaram et al.,2004;Saint-Pol et al.,2004).Therefore,protoplasts were cotransformed with EPSIN1:RFP and GFP:talin ,a marker for actin filaments consist-ing of GFP and the actin binding domain of mouse talin (Kost et al.,1998;Kim et al.,2005).As expected,GFP:talin produced the network pattern (Figure 3A)(Kost et al.,1998;Kim et al.,2005).Furthermore,the red fluorescent network pattern of EPSIN1:RFP closely overlapped the green fluorescent network pattern of GFP:talin (Figure 3A),raising the possibility that EPSIN1:GFP bound to the actin filaments rather than to the ER.To confirm this,the EPSIN1:RFP pattern was examined after treatment with latrunculin B (Lat B),a chemical agent known to disrupt actin filaments (Spector et al.,1983).Lat B–treated protoplasts produced the diffuse green fluorescent pattern of GFP:talin (Figure 3A),an indication of solubilized actin filaments,as observed previously (Kim et al.,2005).In addition,the Lat B–treated protoplasts displayed a diffuse red fluorescent pattern of EPSIN1:RFP (Figure 3A),indicating that EPSIN1is associated with actin filaments but not with the ER.Furthermore,the punc-tate staining pattern of EPSIN1:RFP also was not observed in the presence of Lat B,indicating that actin filaments played a role in yielding the punctate staining pattern of EPSIN1.In the same conditions,BiP:GFP,an ER marker (Lee et al.,2002),produced a network pattern,indicating that Lat B does not disrupt the ER network patterns (Figure 3Ai).To identify the organelle responsible for the punctate staining pattern of EPSIN1,its localization was compared with that of ST:GFP and PEP12p/SYP21.ST:GFP,a chimericproteinFigure 1.EPSIN1Is Expressed in Various Arabidopsis Tissues.(A)Generation of anti-EPSIN1antibody.The middle domain,corresponding to amino acid residues 153to 337,was expressed as the Hisx6-tagged form in E.coli and used to raise antibody in a rabbit.Control serum was obtained from the rabbit before immunization.Total protein extracts were obtained from leaf tissues and used to test the anti-EPSIN1antibody.(B)Specificity of the anti-EPSIN1antibody.Protein extracts were obtained from protoplasts expressing EPSIN1tagged with the HA epitope at the N terminus and used for protein gel blot analysis using anti-HA and anti-EPSIN1antibodies.(C)Expression of EPSIN1in various tissues.Total protein extracts from the indicated tissues were analyzed by protein gel blotting using anti-EPSIN1antibody.Leaf tissues were harvested 11and 20d after germination.Cotyledons were obtained from 5-d-old plants.The membranes were stained with Coomassie blue to control for protein loading.RbcL,large subunit of the ribulose-1,5-bis-phosphate carboxylase/oxygenase (Rubisco)complex.2260The Plant CellFigure 2.EPSIN1Produces Both Network and Punctate Staining Patterns.(A)Subcellular fractionation of EPSIN1.Total (T)protein extracts of leaf tissues were separated into soluble (S)and pellet (P)fractions and analyzed by protein gel blotting using anti-EPSIN1,anti-AALP,and anti-VSR antibodies.(B)Expression level of EPSIN1in transformed protoplasts.Protoplasts were transformed with various amounts of HA:EPSIN1DNA,and the level of EPSIN1was determined by protein gel blotting with anti-EPSIN1antibody.Protein extracts from untransformed protoplasts were used as a control.The membrane was also stained with Coomassie blue to control for loading.(C)Localization of EPSIN1.Protoplasts were transformed with the indicated constructs (5to 10m g),and the localization of EPSIN1was examined either by immunostaining with anti-HA antibody or by direct detection of the GFP or RFP signal.Untransformed protoplasts were immunostained with anti-HA antibody as a control.Bars ¼20m m.(D)Colocalization of EPSIN1proteins.The localization of EPSIN1protein was examined in protoplasts transformed with HA:EPSIN1and EPSIN1:GFP or with EPSIN1:GFP and EPSIN1:RFP .As controls,GFP and RFP alone were transformed into protoplasts.Bars ¼20m m.EPSIN1in Vacuolar Trafficking 2261亚细胞定位可以荧光观察也可以做western 检测Figure 3.Localization of EPSIN1in Protoplasts.2262The Plant Cellbetween rat sialyltransferase and GFP,localizes to the Golgi complex,and PEP12p,a t-SNARE,localizes to the PVC(da Silva Conceic¸a˜o et al.,1997;Boevink et al.,1998;Jin et al.,2001). Protoplasts were cotransformed with HA:EPSIN1and ST:GFP. The localization of these proteins was examined after staining with anti-HA antibody.ST:GFP was observed directly with the greenfluorescent signals.A major portion of the HA:EPSIN1-positive punctate stains closely overlapped with those of ST:GFP (Figures3Ba to3Bc).To further confirm the Golgi localization of HA:EPSIN1,protoplasts transformed with HA:EPSIN1were treated with brefeldin A(BFA),a chemical known to disrupt the Golgi complex(Driouich et al.,1993),and the localization of HA:EPSIN1was examined.In the presence of BFA,HA:EPSIN1 yielded a largely diffuse pattern with aggregates,but not the punctate staining pattern,indicating that BFA affects EPSIN1 localization(Figure3Be).In the same conditions,ST:GFP pro-duced a network pattern with large aggregates(Figure3Bg), confirming that the Golgi complex was disrupted.These results support the notion that EPSIN1localizes to the Golgi complex. Next,we examined the possibility of EPSIN1localizing to the PVC.Protoplasts were cotransformed with EPSIN1:GFP and PEP12p:HA.The localization of PEP12p:HA was examined after staining with anti-HA antibody.EPSIN1:GFP was observed di-rectly with the greenfluorescent signals.Only a minor portion of the EPSIN1:GFP-positive punctate stains overlapped with the PEP12p:HA-positive punctate stains(Figures3Bi to3Bk,ar-rows).These results indicated that EPSIN1localized primarily to the Golgi complex with a minor portion to the PVC.To obtain independent evidence for the localization,we ex-amined the colocalization of EPSIN1with VTI11,a v-SNARE that is distributed equally to both the TGN and the PVC(Zheng et al., 1999;Bassham et al.,2000;Kim et al.,2005).Protoplasts were cotransformed with EPSIN1:GFP and VTI11:HA,and the local-ization of these proteins was examined by immunostaining with anti-HA antibody.EPSIN1-positive punctate stains largely colo-calized with those of VTI11:HA(Figures3Bm to3Bo),confirming that EPSIN1localizes to both the Golgi complex and the PVC. EPSIN1Binds to and Colocalizes with ClathrinThe members of the epsin family have two clathrin binding motifs (Rosenthal et al.,1999;Wendland et al.,1999;Drake et al.,2000). Sequence analysis indicated that EPSIN1has a potential clathrin binding motif.To explore the possibility that EPSIN1binds to clathrin,glutathione S-transferase–fused EPSIN1(GST:EPSIN1) was constructed for a protein pull-down assay(Figure4A).GST: EPSIN1was expressed in Escherichia coli and purified from E. coli extracts(Figure4B).The purified GST:EPSIN1was mixed with protein extracts obtained from leaf tissues.Proteins pelleted with glutathione–agarose were analyzed by protein gel blotting using anti-clathrin antibody.GST:EPSIN1,but not GST alone, precipitated from the plant extracts a180-kD protein species that was recognized by anti-clathrin antibody(Figure4C),indi-cating that EPSIN1bound to clathrin.To further examine its binding to clathrin,EPSIN1was divided into two regions,the ENTH and the remainder of the molecule (EPSIN1D N)(Figure4A).These regions were expressed in E.coli as GST fusion proteins,GST:ENTH and GST:EPSIN1D N,re-spectively(Figure4B).Protein pull-down experiments using leaf cell extracts were performed with purified GST:ENTH and GST: EPSIN1D N.GST:EPSIN1D N,but not GST:ENTH,precipitated clathrin from the plant extracts(Figure4C).To identify the clathrin binding motif,the C-terminal region containing the putative clathrin binding motif,LIDL(Lafer,2002),as well as GST:RIDL, which contained an Arg substitution of thefirst Leu residue in the motif,were expressed as GST fusion proteins in E.coli(Figures 4A and4B).GST:LIDL,but not GST:RIDL,precipitated clathrin from protein extracts(Figure4C),indicating that the LIDL motif functioned as a clathrin binding motif.The in vitro binding of EPSIN1with clathrin strongly suggested that EPSIN1was likely to colocalize with clathrin.Therefore, immunohistochemistry for the localization of EPSIN1and clathrin was performed.Protoplasts were transformed with HA:EPSIN1, and the localization of HA:EPSIN1and clathrin was examined by staining with anti-HA and anti-clathrin antibodies,respectively. The anti-clathrin antibody produced a punctate staining pattern (Figure4D).A majority(60to70%)of the HA:EPSIN1-positive punctate stains closely overlapped with a pool(40to50%)of clathrin-positive punctate stains(Figure4D),consistent with an interaction between EPSIN1and clathrin.There was also a pool of clathrin-positive punctate stains that lacked the HA:EPSIN1 signal,suggesting that clathrin also was involved in an EPSIN1-independent process.To further characterize the interaction between EPSIN1and clathrin,we examined whether or not EPSIN1is permanently associated with CCVs.Protein extracts from leaf tissues were first separated into soluble and pellet fractions by ultracentrifu-gation.The pellet fraction was treated with Triton X-100and further fractionated by gelfiltration,and the fractions were ana-lyzed by protein gel blotting using anti-clathrin,anti-EPSIN,and anti-VSR antibodies.Clathrin was detected in a peak between 443and669kD(see Supplemental Figure1online).Interestingly, VSR,the vacuolar cargo receptor,was eluted at the same posi-tion with clathrin.By contrast,EPSIN1was eluted at90kD. These results suggest that EPSIN1is not permanently associ-ated with CCVs.Figure3.(continued).(A)Colocalization of EPSIN1with actinfilaments.Protoplasts were transformed with the indicated constructs,and the localization of these proteins was examined in the presence(þLat B)and absence(ÿLat B)of Lat B(10m M).Bars¼20m m.(B)Localization of EPSIN1to the Golgi complex and the PVC.Protoplasts were transformed with the indicated constructs,and localization of the proteins was examined after immunostaining with anti-HA.The GFP signals were observed directly in thefixed protoplasts.For BFA treatment,BFA(30 m g/mL)was added to the transformed protoplasts at24h after transformation and incubated for3h.Arrows indicate the overlap between EPSIN1:GFP and PEP12p:HA.Bars¼20m m.EPSIN1in Vacuolar Trafficking2263Figure 4.EPSIN1Binds to and Colocalizes with Clathrin.(A)Constructs.GST was fused to the N terminus.ENTH,the epsin N-terminal homology domain.DLF and DPF motifs are similar to AP-1and AP-3binding motifs,respectively.Q11indicates a stretch of 11Glu residues.The clathrin binding motif (LIDL)and the Leu-to-Arg substitution in the clathrin binding motif (RIDL)are shown in the C-terminal region.The numbers indicate amino acid positions.(B)Expression of GST-fused EPSIN1proteins.Constructs were introduced into E.coli ,and their expression was induced by isopropylthio-b -galactoside.GST fusion proteins were purified from E.coli extracts with glutathione–agarose beads.Purified proteins were stained with Coomassie blue.(C)Interaction of EPSIN1with clathrin.GST-fused EPSIN1proteins were mixed with protein extracts from leaf tissues.EPSIN1binding proteins were precipitated using glutathione–agarose beads and analyzed by protein gel blotting using anti-clathrin antibody.Supernatants also were included in the protein gel blot analysis.Subsequently,the membranes were stained with Coomassie blue.Bead,glutathione–agarose beads alone;P,pellet;S,supernatant (10%of total).(D)Colocalization of EPSIN1with clathrin.Protoplasts transformed with HA:EPSIN1were fixed with paraglutaraldehyde,and the localization of HA:EPSIN1and clathrin was examined by immunostaining with anti-HA and anti-clathrin antibodies,respectively.Bar ¼20m m.2264The Plant CellEPSIN1Interacts with VTI11Epsin-related proteins in animal and yeast cells are involved in either endocytosis or vacuolar/lysosomal protein trafficking(Chen et al.,1998;De Camilli et al.,2002;Wendland,2002;Overstreet et al.,2003;Legendre-Guillemin et al.,2004).To elucidate the pathway of EPSIN1involvement,binding partners of EPSIN1 were examined.In animal and yeast cells,epsin-like proteins have been shown to interact with SNAREs(Chen et al.,1998; Chidambaram et al.,2004).Because EPSIN1localized to the Golgi complex and the PVC,EPSIN1interactions with Arabidop-sis VTI11and VTI12(formerly At VTI1a and At VTI1b,respectively) were examined.VTI11is a v-SNARE that localizes to the TGN and travels to the PVC(Zheng et al.,1999;Bassham et al.,2000). VTI11and VTI12were tagged with HA at the C terminus and introduced into protoplasts.The expression of VTI11:HA and VTI12:HA in protoplasts was confirmed by protein gel blot analysis using anti-HA antibody.The anti-HA antibody detected protein bands at33and35kD(Figure5A),the expected positions of VTI11:HA and VTI12:HA,respectively.Purified GST:EPSIN1 from E.coli extracts was mixed with plant extracts from the VTI11:HA-or VTI12:HA-transformed protoplasts,and GST: EPSIN1-bound proteins were precipitated from the mixture using glutathione–agarose beads.The pellet fraction was analyzed by protein gel blotting using anti-HA antibody.VTI11:HA,but not VTI12:HA,was detected from the pellet(Figure5A).GST alone did not precipitate VTI11:HA from the plant extracts.These results indicated that although VTI11and VTI12are highly similar to each other,EPSIN1specifically binds to VTI11:HA.To further confirm this interaction,we performed a reciprocal protein pull-down experiment(i.e.,pull-down of EPSIN1with VTI11)using protein extracts obtained from protoplasts transformed with VTI11:HA and EPSIN1:GFP.VTI11:HA-bound proteins were immunoprecipitated with anti-HA antibody,and the immunopre-cipitates were analyzed by protein gel blotting using anti-HA, anti-GFP,and anti-calreticulin antibodies.Anti-calreticulin anti-body was used as a negative control.In addition to VTI11:HA, EPSIN1:GFP was detected in the immunoprecipitates(Figure 5B).However,calreticulin was not detected in the pellet.These results further confirm the interaction between VTI11and EPSIN1. To determine the VTI11binding domain of EPSIN1,proteinpull-down experiments were performed using GST:ENTH and GST:EPSIN1D N.GST:ENTH,but not GST:EPSIN1D N,precipi-tated VTI11:HA from the plant extracts(Figure5C),indicating that the ENTH domain contained the VTI11binding motif.Similarly,in animal and yeast cells,EpsinR and Ent3p have been shown to bind to vti1b and vti1p,respectively(Chidambaram et al.,2004). EPSIN1Binds to the Arabidopsis Homolog of g-Adaptinof AP-1Epsin homologs bind to adaptor proteins(APs)(Duncan et al., 2003;Mills et al.,2003).In animal cells,EPSIN1binds to the a-adaptin of AP-2via the D F F/W(where F indicates a hydro-phobic amino acid)and FXDXF motifs(Figure4A)(Brett et al., 2002).Arabidopsis EPSIN1has three DPF motifs to which a-adaptin of AP-2could bind.In addition,EPSIN1has two regions with motifs similar to the acidic Phe motif for binding AP-1and GGAs(Duncan et al.,2003).Therefore,the interactions of EPSIN1with AP complexes were examined.We isolated the Arabidopsis proteins g-adaptin related protein(g-ADR),a-ADR, and d-ADR,which were most closely related to g-adaptin, a-adaptin,and d-adaptin of AP-1,AP-2,and AP-3,respectively. These Arabidopsis proteins were tagged with GFP and ex-pressed transiently in protoplasts.Protein extracts from the transformed protoplasts were mixed with purified GST:EPSIN1, and the GST:EPSIN1-bound proteins were precipitated.The pellet was analyzed by protein gel blotting using anti-GFP antibody.GFP:g-ADR,but not a-ADR:GFP or d-ADR:GFP,was detected in the pellet(Figure6A).The control for the protein pull-down assay,GST alone,did not precipitate any of these proteins. These results strongly suggested that EPSIN1interacts with g-ADR specifically.To further confirm the interaction between EPSIN1and g-ADR,we performed a reciprocal protein pull-down experiment(i.e.,pull down of EPSIN1proteins with Figure5.EPSIN1Binds to VTI11.(A)Protein extracts were prepared from VTI11:HA-and VTI12:HA-transformed protoplasts and mixed with GST alone or GST:EPSIN1. EPSIN1-bound proteins were precipitated from the mixture with gluta-thione–agarose beads and analyzed by protein gel blotting using anti-HA antibody.(B)Coimmunoprecipitation of EPSIN1:GFP with VTI11:HA.Protein ex-tracts from protoplasts cotransformed with VTI11:HA and EPSIN1:GFP were used for immunoprecipitation with anti-HA antibody.The immuno-precipitates were analyzed by protein gel blotting with anti-HA,anti-GFP, and anti-calreticulin antibodies.P,immunoprecipitate;S,supernatant;T, total protein extracts(5%of the input).(C)For binding experiments,protein extracts from protoplasts trans-formed with VTI11:HA were mixed with GST alone,GST:ENTH,and GST:EPSIN1D N.Proteins were precipitated with glutathione-agarose beads and analyzed by protein gel blotting using anti-HA antibody.The amount of the input proteins is indicated.EPSIN1in Vacuolar Trafficking2265。

Article EIN2-Directed Translational Regulation of Ethylene Signaling in ArabidopsisGraphical AbstractHighlightsd Ectopic expression of EBF1/230UTR fragment leads toethylene insensitivityd30UTR mediates ethylene-induced translational repression in an EIN2-dependent wayd PolyU motifs within30UTR are critical for EIN2-directedtranslational inhibitiond EIN2targets EBF130UTR to cytoplasmic P-body viainteracting with EIN5and PABs AuthorsWenyang Li,Mengdi Ma,Ying Feng,..., Mingzhe Li,Fengying An,Hongwei GuoCorrespondencehongweig@In BriefThis study reports a novel translational repression mechanism during ethylene signaling in which30UTRs of mRNAs function as signaltransducers. Li et al.,2015,Cell163,670–683October22,2015ª2015Elsevier Inc./10.1016/j.cell.2015.09.037Article EIN2-Directed Translational Regulationof Ethylene Signaling in ArabidopsisWenyang Li,1,3Mengdi Ma,1,3Ying Feng,1Hongjiang Li,1,4Yichuan Wang,1Yutong Ma,1Mingzhe Li,1Fengying An,1,2 and Hongwei Guo1,2,*1The State Key Laboratory of Protein and Plant Gene Research,College of Life Sciences,Peking University,Beijing100871,China2Peking-Tsinghua Center for Life Sciences,Beijing100871,China3Co-first author4Present address:Institute of Science and Technology Austria,Am Campus1,3400Klosterneuburg,Austria*Correspondence:hongweig@/10.1016/j.cell.2015.09.037SUMMARYEthylene is a gaseous phytohormone that plays vital roles in plant growth and development.Previous studies uncovered EIN2as an essential signal trans-ducer linking ethylene perception on ER to transcrip-tional regulation in the nucleus through a‘‘cleave and shuttle’’model.In this study,we report another mechanism of EIN2-mediated ethylene signaling, whereby EIN2imposes the translational repression of EBF1and EBF2mRNA.Wefind that the EBF1/2 30UTRs mediate EIN2-directed translational repres-sion and identify multiple poly-uridylates(PolyU) motifs as functional cis elements of30UTRs.Further-more,we demonstrate that ethylene induces EIN2to associate with30UTRs and target EBF1/2mRNA to cytoplasmic processing-body(P-body)through in-teracting with multiple P-body factors,including EIN5and PABs.Our study illustrates translational regulation as a key step in ethylene signaling and presents mRNA30UTR functioning as a‘‘signal transducer’’to sense and relay cellular signaling in plants.INTRODUCTIONEthylene is a gaseous phytohormone produced by plants in response to various internal and environmental stimuli,which triggers a wide range of physiological and morphological re-sponses(Johnson and Ecker,1998).During the past decades, a relatively linear ethylene signaling pathway has been estab-lished through the application of molecular and genetic ap-proaches(Guo and Ecker,2004).In Arabidopsis,ethylene is perceived by a group of ER-located receptors(Chang and Stadler,2001).In the absence of ethylene signal,the hor-mone-free receptors activate a Raf-like protein kinase CONSTITUTIVE TRIPLE RESPONSE1(CTR1)(Gao et al., 2003;Kieber et al.,1993).Activated CTR1and the receptors cooperatively inhibit an ER-located membrane protein ETHYLENE INSENSITIVE2(EIN2)through physical interaction and protein phosphorylation(Alonso et al.,1999;Bisson and Groth,2011;Ju et al.,2012).EIN2is a key component in ethylene signaling pathway,evi-denced by completely ethylene-insensitive phenotypes of the ein2null mutants(Ji and Guo,2013).It is encoded by a single-copy gene in Arabidopsis,and is conserved from charophyte green algae to land plants(Ju et al.,2015).While the function of its N-terminal membrane-spanning domain is not clear,the C-ter-minal end of EIN2(CEND)is thought to participate in signaling output,as ectopic expression of this domain alone can partially activate ethylene responses(Alonso et al.,1999;Wen et al., 2012).Recent studies reported that CEND can be phosphory-lated by the receptors-activated CTR1in the absence of ethylene (Ju et al.,2012;Qiao et al.,2012).Upon ethylene application, inactivation of the receptors and CTR1abolishes the phosphory-lation state of CEND,leading to its proteolysis from the ER-teth-ered N terminus,followed by shuttling into the nucleus(Ju et al., 2012;Qiao et al.,2012;Wen et al.,2012).However,this‘‘cleave and shuttle’’mode might represent part of the EIN2actions,as induced nuclear localization of CEND only partially activates ethylene signaling(Ji and Guo,2013;Wen et al.,2012).Mean-while,ethylene also induces CEND to form discrete and promi-nent foci in the cytoplasm(Qiao et al.,2012;Wen et al.,2012), but the function of such cytoplasmic portion remains unexplored. In the nucleus,components working downstream of EIN2are two master transcription factors ETHYLENE INSENSITIVE3 (EIN3)and its homolog EIN3-LIKE1(EIL1),which regulate the vast majority of ethylene-directed gene expression(Chang et al.,2013;Chao et al.,1997).One of the key regulatory mech-anisms of ethylene signaling is the stabilization of EIN3/EIL1 proteins,wherein ethylene acts to repress the proteasomal degradation of EIN3/EIL1mediated by two F-box proteins, EIN3-BINDING F-BOX1(EBF1)and EBF2,in an EIN2-dependent manner(An et al.,2010;Guo and Ecker,2003;Potuschak et al., 2003).However,the molecular mechanism of how ethylene or EIN2represses the function of EBF1/2is still elusive. ETHYLENE INSENSITIVE5(EIN5),encoding a cytoplasmic 50-30exoribonuclease(AtXRN4),is another component positively modulating ethylene responses(Olmedo et al.,2006;Potuschak et al.,2006).Currently,little is known about how EIN5modulates ethylene signaling,except for the genetic evidence suggesting its participation in the regulation of EBF1/2function(Olmedo et al.,2006;Potuschak et al.,2006).Notably,small RNA frag-ments corresponding to EBF1and EBF2mRNA30UTR were pro-cessed and accumulated in ein5(Olmedo et al.,2006;Potuschak et al.,2006;Souret et al.,2004).Our recent work uncoveredthat670Cell163,670–683,October22,2015ª2015Elsevier Inc.EIN5,in combination with 30-50RNA decay pathway,is respon-sible for the removal of many defective coding transcripts as well as the cleavage fragments of miRNA targets,including 30UTRs,which are otherwise subjected to posttranscriptional gene silencing (Zhang et al.,2015).However,genetic evidence disfavored the possibility that 30UTR fragments of EBF1/2mRNA are processed and targeted to small RNA-mediated gene silencing pathway (Potuschak et al.,2006).Interestingly,ectopic expression of a 30UTR -truncated EBF2gene resulted in a stronger ethylene insensitive phenotype than that of the EBF2full-length gene (Konishi and Yanagisawa,2008),implying a negative role of 30UTR on the EBF2function.In this study,we sought to investigate the regulatory mecha-nisms of how ethylene signal is relayed from cytoplasm to nucleus,and how EIN2and EIN5participate in this signaling pro-cess.Strikingly,we found that ectopic expression of either EBF1or EBF230UTR fragments confers strong ethylene-insensitivity phenotypes through promoting the translation of endogenousEBF1/2mRNAs.Furthermore,we found that ethylene induces EIN2to target EBF130UTR to cytoplasmic processing-body (P-body)through interacting with EIN5and other P-body factors to repress EBF1/2translation.Our study uncovers another branch of ethylene signaling pathway mediated by cytoplasmic EIN2in translational control.RESULTSOverexpression of EBF130UTR Leads to Reduced Ethylene SensitivityPrevious studies revealed that the ein5mutant accumulated EBF1/2mRNA 30UTR fragments (Olmedo et al.,2006;Potu-schak et al.,2006;Souret et al.,2004).We thus speculated that the over-accumulated 30UTR fragments could contribute to the ethylene insensitivity of ein5.To test this speculation,we overexpressed the EBF130UTR region (1U )in wild-type Col-0plants (Figures 1A and 1B).The so-called ‘‘tripleresponse’’Figure 1.Overexpression of EBF130UTR Results in Reduced Ethylene Sensitivity(A)Schematic diagrams of the gene structure of EBF1and the 30-UTR -overexpressing construct.Full-length EBF130UTR (643bp after stop codon)plus a 66-bp flanking sequence was inserted into the multiple cloning site (MCS)prior to the NOS terminator in pDr vector.S in open circle,stop codon.(B)Quantification of 30UTR transcripts in etiolated seedlings of three independent transgenic lines grown on MS medium with (+)or without (–)ACC (an ethylene biosynthetic precursor).Vector means pDr-expressing transgenic plants while 30UTR means 30-UTR -overexpressing transgenic lines.Arrows denote the primers used for qRT-PCR to detect the levels of 30UTR .(C)The triple response phenotypes of seedlings corresponding to (B).(D)Quantification of hypocotyl lengths and root lengths of the seedlings in (C).**p <0.01.Mean ±SD,n >10.(E)Immunoblot assays showing EIN3protein levels of seedlings corresponding to (B).A nonspecific band served as a loading control.The numbers indicate the relative EIN3protein levels as calculated from three biological replicates.(F and H)Schematic maps of M1U (MYC-EBF130UTR )and G1U (GFP-EBF130UTR ),as well as two control transcripts MYC and GFP .A(n)represents the poly(A)tail.Of note,all these transcripts are driven by CaMV 35S promoters.(G and I)The triple response phenotypes of etiolated seedlings of wide-type Col-0as well as three independent lines of indicated transgenic plants.See also Figure S1.Cell 163,670–683,October 22,2015ª2015Elsevier Inc.671phenotype is commonly used as an ethylene-specific growth response in Arabidopsis,which refers to exaggerated apical hooks,shortened hypocotyls and roots of dark-grown seedlings exposed to ethylene or treated with ethylene precursor1-amino-cyclopropane-1-carboxylic acid(ACC)(Bleecker et al.,1988; Ecker,1995).Overexpression of1U conferred significant attenu-ation of triple response phenotypes to Col-0,resulting in elon-gated hypocotyls and roots compared with control seedlings (Figures1C and1D).Consistently,we found that the levels of EIN3protein were lower in1U transgenic plants than that in Col-0(Figure1E).Furthermore,we fused1U to the MYC tag and GFP coding sequence(referred to as M1U and G1U),respectively(Figures 1F and1H),and overexpressed these fusion genes in wild-type Col-0(Figures S1A–S1C and S1F–S1H).Similar to1U-overex-pressing seedlings,M1U-and G1U-overexpressing plants dis-played reduced ethylene sensitivity and impaired EIN3protein accumulation compared with control plants(Figures1G,1I, S1D,S1E,S1I,and S1J).Together,these results demonstrate that overexpression of1U,alone or in fusion with unrelated tran-scripts,reduces ethylene sensitivity.Overexpression of EBF130UTR Promotes the Translation of Endogenous EBF1/2mRNAs Interestingly,we found that ethylene hyposensitivity resulting from1U-overexpression was partially restored by a defect in either EBF1or EBF2(Figure2A).Due to the fatal effect of over-accumulated EIN3in ebf1ebf2double mutant,we next overexpressed M1U in b-estradiol-inducible EIN3-Flag/ein3 eil1ebf1ebf2(iEIN3/qm)(An et al.,2010),which was used as a substitution of the lethal ebf1ebf2double mutant(Fig-ure S2A).We found that M1U no longer affected the triple response phenotypes(Figure2B),and the abundance of EIN3protein was comparable between iEIN3/qm and M1U iEIN3/qm(Figure S2B).Together,these results demonstrate that the presence of EBF1/2is required for the1U-overexpres-sion-induced repression of ethylene responses,implying that exogenous30UTR expression modulates the function of EBF1/2.We found that the levels of both EBF1and EBF2tran-scripts were not evidently affected by1U overexpression (Figure S2C),excluding the modulation of EBF1/2at the level of transcription or RNA decay.We next examined whether the translation of EBF1/2mRNAs is under the regulation. Without good antibody against EBF1or EBF2available,two experiments were conducted for this ing poly-some profiling assays,we found that the translation of EBF1and EBF2mRNAs was repressed by ethylene,as the portion of high-density polysome-associated EBF1/2mRNAs was decreased upon ethylene application(Figure2C). Notably,1U overexpression recovered the drop of the portion of polysome-associated EBF1/2mRNAs(Figure2C).There-fore,1U overexpression augments the translation of endoge-nous EBF1/2mRNAs,which is subjected to repression by ethylene.Furthermore,we constructed transgenic plants harboring GFP-EBF1followed by1U or not(G1F and G1C,respectively) (Figure2D),and expressed an inducible1U(iEBF1U)in these plants to examine the effect of exogenous1U expression on G1F or G1C translation.We found that,while GFP-EBF1 mRNA levels were comparable,GFP-EBF1protein levels were downregulated by ethylene and upregulated by1U overexpres-sion in G1F plants(Figures2E and2F).By contrast,in G1C plants,the GFP-EBF1protein levels were virtually unchanged upon1U expression regardless of ethylene application(Figures 2E and2F).Collectively,these results suggest that the over-accumulation of1U transcripts boosts the function of EBF1/2 by enhancing their translation.Based on these observations,we propose a translational inter-ference model,in which ectopically expressed1U transcripts in-terferes with the endogenous EBF1/230UTRs that supposedly exert a repressive role on the translation of EBF1/2mRNAs. Such translational interference could arise from the competition and/or titration of translational repressors binding to the endog-enous30UTR regions(Figure2G).The30UTRs Impart Translational Inhibition to EBF1/2 mRNAs in Response to EthyleneWe next tested the translational interference model(Figure2G) by examining the effect of EBF130UTR on GFP mRNA trans-lation(Figures1H,1I,and S1F–S1J).We found that,with the comparable transcript levels(Figure S1G),seedlings express-ing G1U accumulated much lower GFPfluorescence or protein abundance than those expressing GFP alone,particularly when treated with ACC(Figures3A–3D).Ethylene caused over80% of decrease in the translational efficiency of G1U whereas had no effect on GFP alone(Figures3C and3D).The ACC-pro-moted reduction in GFP protein abundance was restored by the application of ethylene inhibitor silver ions(Ag+)(Figure3E). Taken together,these results indicate that EBF130UTR confers translational repression to its fusion mRNA in response to ethylene.Next,we determined the biological significance of the EBF1 mRNA30UTR–mediated translational repression in ethylene signal transduction.We constitutively expressed M1C(MYC-EBF1,MYC tag fused with the EBF1coding sequence)and M1F(MYC fused with the EBF1full-length transcript including coding sequence and30UTR)(Figure3F).Compared with control plants,M1F expression resulted in reduced ethylene sensitivity, whereas M1C expression conferred nearly complete ethylene insensitivity(Figure3G).In agreement with the triple response phenotype,the amount of MYC-EBF1protein was nearly con-stant in M1C but progressively decreased in M1F upon treatment with increasing doses of ACC(Figure3H).Given the comparable mRNA abundance between M1F and M1C(Figures S3A and S3B),we concluded that translational repression of EBF1 mRNA via its30UTR is critical for EBF1function in ethylene signaling.We further found that the overexpression of EBF230UTR(2U) also led to reduced ethylene sensitivity in GFP-EBF230UTR (G2U)transgenic plants(Figures S3C and S3D).Like EBF130 UTR,EBF230UTR also conferred translational repression to the GFP mRNA fused with it(Figure S3E).Thus,the30UTRs of both EBF1and EBF2act similarly to impose translational repres-sion to their respective mRNAs in response to the ethylene signal.672Cell163,670–683,October22,2015ª2015Elsevier Inc.Figure2.Overexpression of EBF130UTR Enhances the Translation of Endogenous EBF1/2mRNAs(A and B)Triple response phenotypes of etiolated transgenic seedlings expressing G1U(GFP-EBF130UTR)treated with ACC(A),and seedlings expressing M1U (MYC-EBF130UTR)treated with ACC in combination with DMSO or b-estradiol(B).iEIN3/qm is the b-estradiol-induced EIN3-Flag in the ein3eil1ebf1ebf2 quadruple mutant background,which was used to substitute for the lethal ebf1ebf2double mutant(An et al.,2010).(legend continued on next page)Cell163,670–683,October22,2015ª2015Elsevier Inc.673EIN2Is Essential for 30-UTR-Mediated Translational Repression of EBF1mRNAWe next investigated the role of key ethylene signaling com-ponents in 30-UTR-mediated translational regulation.The ethylene-induced repression of G1U mRNA translation,mani-fested by reduced GFP fluorescence,was similarly observed inCol-0and ein3eil1,but not in ein2and a receptor mutant etr1(Figures 4A and S4A),suggesting that the upstream signaling components including the receptors and EIN2are required for 30-UTR-mediated translational repression,whereas EIN3/EIL1are not.Expression of a b -estradiol-inducible version of EIN2was sufficient to restore such translation inhibition in ein2,and(C)Polysome profiling assays with sucrose density gradient accompanied by qRT-PCR to analyze translational status of EBF1/2mRNAs.A 254absorption was monitored together with fractionation (left).The fractions containing 40S,80S of ribosome,and polysomes are indicated.The abundance of EBF1and EBF2mRNA in each fraction was detected by qRT-PCR and quantified as a percentage relative to their total amount (right).UBQ5mRNA was used as a reference.(D)Structures of iEBF1U (b -estradiol-inducible EBF130UTR )transcript,G1F (GFP-EBF1full length containing CDS and 30UTR)and G1C (GFP-EBF1CDS).Arrows indicate the primer pair used to analyze the expression of iEBF1U .(E)Coexpression of G1F or G1C together with iEBF1U in etiolated seedlings treated with or without ethylene and b -estradiol for 4hr before RT-PCR and western blotting analysis.Protein loading was manifested by Coomassie brilliant blue (CBB)staining.(F)Quantitative measurements of GFP-EBF1proteins in (E)based on three biological repeats.*p <0.05;***p <0.001.(G)A translational interference model proposes that the exogenously overexpressed 30UTRs enhance the translation of endogenous EBF1/2mRNAs by competing with their inherent 30UTRs and thus titrating unknown repressor X bound to 30UTRs.See also Figure S2.Figure 3.EBF130UTR Confers Translational Repression to Its Fusion Transcripts in Response to Ethylene(A and B)GFP fluorescence in the roots of three independent transgenic seedlings expressing GFP or G1U (GFP-EBF130UTR )with (+)or without (–)ACC treatment (A)and the relative quantifications of GFP fluorescence (B).***p <0.001.Mean ±SD,n >20roots.(C)Immunoblot assays showing GFP protein abundance in whole etiolated seedlings with (+)or without (–)ACC treatment.(D)qRT-PCR analysis of GFP mRNAs and quantification of GFP proteins in (C).The ratio of protein to mRNA abundance was defined as the translation efficiency.***p <0.001;calculations based on three biological repeats.(E)Immunoblot assays showing GFP protein abundance in etiolated seedlings treated with (+)or without (–)ACC and/or silver ion.(F)Structures of MYC ,M1C (MYC-EBF1CDS),and M1F (MYC-EBF1full length containing CDS and 30UTR)transcripts.(G)Hypocotyl lengths of etiolated seedlings of three independent transgenic lines expressing MYC ,M1C ,and M1F .Mean ±SD,n >20.(H)Immunoblot assays indicating MYC-EBF1protein abundances (top)and their relative quantifications (bottom)in seedlings treated with increasing doses of ACC.Calculations were based on three biological repeats.See also Figure S3.674Cell 163,670–683,October 22,2015ª2015Elsevier Inc.Figure4.EIN2Is Required for EBF130-UTR-Mediated Translational Repression(A)GFPfluorescence in the roots of etiolated seedlings expressing G1U(GFP-EBF130UTR)in different genotype backgrounds(top).Immunoblot assays showing GFP protein abundance in whole seedlings(bottom).(B)Structure of the b-estradiol-inducible EIN2-HA gene(iEIN2-HA).(C)GFPfluorescence in the roots of etiolated seedlings transiently treated with or without ACC and b-estradiol for6hr.‘‘Removed,’’removal of both ACC and b-estradiol.(D)Profiles of polysome-associated EBF1,EBF2,and UBQ5mRNAs in Col-0and ein2-5.(E)Immunoblot assays showing MYC-EBF1protein abundance in etiolated seedlings of transgenic plants expressing M1F(MYC-EBF1CDS+30UTR)or M1C (MYC-EBF1CDS).(F)Immunoblot assays showing MYC-EBF1and EIN2-HA protein abundance in transgenic plants expressing iEIN2-HA together with M1F or M1C.Note that multiple processed C-terminal fragments of induced EIN2-HA(CEND-HA)were also shown.(G)Triple response phenotypes of etiolated seedlings corresponding to(F).(H)Quantitative measurements of hypocotyls(left)and roots(right)of etiolated seedlings in(G).*p<0.05;**p<0.01;***p<0.001.Mean±SD,n>20.See also Figure S4.Cell163,670–683,October22,2015ª2015Elsevier Inc.675the removal of b-estradiol led to the efficient translation of G1U again(Figures4B and4C).A similar scenario was observed with transiently expressed b-estradiol-inducible EIN2and G1U in tobacco(Figure S4B),supporting that EIN2is essential for EBF130UTR-directed translational repression.To gain further evidence for EIN2-regulated EBF1/2mRNA translation,we compared the polysome profiles of EBF1/2 mRNAs between Col-0and ein2(Figure4D).The polysome pro-files of EBF1/2mRNAs remained virtually unchanged in ein2 when treated with ethylene,in contrast with the apparent ethylene-induced polysome profile shifts observed in Col-0(Fig-ures2C and4D).Meanwhile,we found that the ethylene-evoked translational repression of M1F(MYC-EBF1full-length tran-script)was abolished in ein2(Figures4E and S4C),but exacer-bated by addition of EIN2function(Figure4F).By contrast,the translation of M1C(MYC-EBF1CDS)remained unaffected upon depletion or addition of EIN2(Figures4E,4F,and S4C). Furthermore,the partial ethylene-insensitivity phenotype of M1F transgenic plants was largely suppressed by the overex-pression of EIN2,whereas the strong ethylene insensitivity of M1C was hardly affected(Figures4G and4H).Taken together, these results indicate that30UTR is a critical ethylene-respon-sive element to repress EBF1translation,and EIN2is necessary and sufficient for directing such translational repression.EIN2-Directed Translational Repression Is Mediatedby PolyU Motifs of EBF1/230UTRsWe next dissected the functional cis elements within the EBF1/2 30UTRs by utilizing a dual-construct translation analysis system in tobacco leaves,in which a30UTR fragment of interest was fused with the GFP coding region,together with mCherry as the internal control in the same reporter construct(Figures5A and S5A).The GFP intensities relative to mCherry intensities were calculated to indicate the translation efficiency of GFP mRNA(Figure5B).Whereas the translation of GFP alone was not altered by introduction of EIN2and/or ACC application,the translation of G1U and G2U(GFP fused with EBF1/230UTR, respectively)was remarkably repressed by either expression of EIN2or ACC application(Figures S5B–S5D and S5K),and to a further extent when combining these two treatments(Fig-ure S5C).As a control,expression of EIN3protein had no effect on the translation of G1U(Figures S5E–S5H).These results confirmed the inhibitory effect of EBF1/230UTRs on translation in an EIN2-dependent manner.EBF130UTR was arbitrarily segmented intofive fragments ranging from98to150nt in length(Figure5C).Three fragments, including1Ua,1Ub and1Ud,were able to mediate EIN2-induced translational repression(Figure5D).Using the computation algo-rithm MEME and RNAfold,we identified a total of7poly-uridy-lates motifs in the predicted stem-loop structure within these three fragments(Figure5E).These sequences were designated as Ethylene Responsive RNA elements containing Poly-Uridy-lates(ERR-PolyU,or EPU for short)(Figure5E).Deletion of EPUs in each fragment or all seven EPUs in1U,which did not change their overall predicted secondary structures(Figure S5I), eliminated EIN2-directed translational repression(Figure5F). Similarly,five EPUs were found in EBF230UTR(Figure S5J), and they were all required for2U to mediate EIN2-induced trans-lational inhibition(Figure S5K).Sequence alignment of EBF30 UTRs from different plant species revealed that PolyU motifs are among the most conserved regions(Figures S5L and S5M),suggesting the30-UTR-mediated translational regulation as a well-preserved mechanism of ethylene signaling.To further investigate the role of EPUs in relaying ethylene signaling,we generated the transgenic plants expressing either the GFP-EBF1full-length transcript driven by its own promoter(pEBF1::G1F)or seven EPUs-depleted version (pEBF1::G1F D7U)in ebf1mutant background.While expression of pEBF1::G1F rescued ebf1to the wild-type level,the pEBF1::G1F D7U/ebf1seedlings exhibited nearly complete ethylene insensitivity,phenocopying pEBF1::G1C/ebf1plants (GFP-EBF1CDS driven by its own promoter)(Figure5G).Consis-tent with the ethylene-response phenotype,the levels of EIN3 protein were much lower in both pEBF1::G1F D7U/ebf1and pEBF1::G1C/ebf1than that in Col-0or pEBF1::G1F/ebf1, whereas the GFP-EBF1protein was more abundant in the former two lines,particularly under ethylene treatment(Figure S5N). These results suggest that EPU-mediated translational inhibition plays a key part in regulating EBF1protein abundance as well as ethylene signal transduction.From1Ud,we selected a region harboring two EPUs that is pre-dicted to form a hairpin structure(Figure5H),and repeated it three times to construct an artificial30UTR that possessed six EPUs(6x EPU)(Figure5H).Similar to G1U,the translation of GFP-6x EPU mRNA was highly reduced upon EIN2induction (Figure5I).Furthermore,transgenic overexpression of GFP-6x EPU but not GFP-1U D7U conferred ethylene insensitivity pheno-type(Figure S5O).Together,these results demonstrate that EPUs mediate the EIN2-directed translational repression of EBF1/2, which represents a crucial mechanism of ethylene signaling. We also examined the functional domain of EIN2in transla-tional repression.By taking advantage of the tobacco system, we narrowed down the C-terminal end of EIN2fragments (CEND)to amino acids(aa)654–1272that were required for translational repression(Figures5J,5K,and S5P).Within this re-gion,a predicted nuclear localization signal(NLS,aa1262–1269, LKRYKRRL)was previously identified to be required for the nu-clear translocation as well as the functionality of CEND(Ju et al.,2012;Qiao et al.,2012;Wen et al.,2012).We found that deletion or mutation of this NLS region also disrupted the func-tion of CEND in translational repression(Figures5J and5K). Interestingly,replacement of the NLS with a distinct K/R-rich NLS sequence(NLS’:KPKKKRKV)was able to relocate CEND into the nucleus but failed to restore its translational repression ability(Figures5K and S6G).Together,these results suggested that the short motif(aa1262–1269)was also critical for the trans-lational repression function of EIN2independent of its being a nuclear localization signal.Association and Co-localization of EIN2with EBF130 UTR in Cytoplasmic FociWe next investigated how EIN2imposes translational repression of1U/2U-containing mRNAs.Wefirst examined whether EIN2 associates with1U in vivo.RNA-immunoprecipitation assays (RNA-IP)in tobacco leaves indicated that EIN2preferentially associated with mRNAs containing1U(G1U,M1U),but not676Cell163,670–683,October22,2015ª2015Elsevier Inc.Figure5.PolyU Motifs in EBF130UTR Are Necessary and Sufficient for EIN2-Directed Translational Inhibition(A and B)Plasmids used in the dual-construct translation analysis system(A)as well as the workflow(B).The reporter plasmid harbors the reference gene mCherry and the reporter gene GFP-30UTR(GFP as control).The effector plasmid possesses EIN2-HA(HA as control)(A).ACC application was used to further activate the(legend continued on next page)Cell163,670–683,October22,2015ª2015Elsevier Inc.677。

∗Correspondingauthor:E-mail,m-takagi@aist.go.jp ;F ax,+81-29-861-3026.Plant Cell Physiol. 50(7): 1232–1248 (2009) doi:10.1093/pcp/pcp075, available online at © The Author 2009. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved.The online version of this article has been published under an open access model. Users are entitled to use, reproduce, disseminate, or display the open access version of this article for non-commercial purposes provided that: the original authorship is properly and fully attributed; the Journal and the Japanese Society of Plant Physiologists are attributed as the original place of publication with the correct citation details given; if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated. For commercial re-use, please contact journals.permissions@T ranscript ion fact ors (TFs) regulat e t he expression of genes at t he t ranscript ional level. Modifi cat ion of TFactivity dynamically alters the transcriptome, which leads t o met abolic and phenot ypic changes. Thus, funct ionalanalysis of TFs using ‘omics-based’ methodologies is one of the most important areas of the post-genome era. In this mini-review, we present an overview of Arabidopsis TFs and int roduce st rat egies for t he funct ional analysis ofplant TFs, which include bot h t radit ional and recent lydeveloped technologies. These strategies can be assigned t o fi ve ca t egories: bioinforma t ic analysis; analysis of molecular func t ion; expression analysis; pheno t ype analysis; and network analysis for the description of entire transcriptional regulatory networks.Keywords :Arabidopsis •Bioinformatics •Functional analysis •Methodology •Transcription factor .Abbreviat ions :AD ,activation domain ;CaMV ,caulifl ower mosaic virus ;CD,conserved domain ;ChIP ,chromatin immunoprecipitation ;CHX ,cycloheximide ;CRES-T ,chimeric repressor silencing technology ;DBD ,DNA -binding domain ;DEG ,differentially expressed gene ;DEX ,dexamethasone ;EAR ,ERF -associated amphiphilic repression ;ER ,estrogen receptor ;F DR ,false discovery rate ;F OX ,full-length cDNA overexpressor ;GF P ,green fluorescent protein ;GO ,gene ontology ;GR ,glucocorticoid receptor ;GUS ,β-glucuronidase ;LUC ,luciferase ;miRNA ,microRNA ;PPI ,protein –protein interaction ;RD ,repression domain ;RT–PCR ,reverse transcription–PCR ;SRDX ,modified EAR motif plant-specific repression domain showing strong repression activity ;ta-siRNA ,trans-acting small interfering RNA ;TF ,transcription factor ;Y1H ,yeast one-hybrid screening .I ntroduction I t is evident that transcriptional regulation plays a pivotal role in the control of gene expression in plants. Intensive studies of plant mutants have revealed that informative phenotypes are often caused by mutations in genes for tran-scription factors (TFs), and a number of TFs have been iden-tifi ed that act as key regulators of various plant functions. TFs, which regulate the fi rst step of gene expression, are usu-ally defi ned as proteins containing a DNA-binding domain (DBD) that recognize a specifi c DNA sequence. In addition, proteins without a DBD, which interact with a DNA-binding protein to form a transcriptional complex, are often catego-rized as TFs. Although some metabolic enzymes have been suggested to regulate gene expression directly in yeast ( H all et al. 2004 ), we do not focus on such multifunctional proteins in this review. In 2000, the entire genome sequence of A rabidopsis thaliana was determined and the genome was predicted to contain 25,498 protein-coding genes(Arabidopsis Genome Initiative 2000). Based on sequenceconservation with known DBDs, R iechmann et al. (2000)reported that around 1,500 of these genes encode TFs, andmore recent analyses have recognized > 2,000 TF genes in the Arabidopsis genome ( D avuluri et al. 2003 , G uo et al. 2005 , I ida et al. 2005 , R iano-Pachon et al. 2007 ). I n contrast to Arabidopsis, the number of TF genes found in D rosophila melanogaster and C aenorhabditis elegans ,which have similar sized genomes to that of Arabidopsis, is around 600, which is signifi cantly less than that in Arabidopsis ( R iechmann et al. 2000 ). The ratio of TF genes to the total number of genes in Arabidopsis is 5–10 % depending on databases, which is higher than that of D . melanogaster (4.7 % ) and of C . elegans (3.6% ) ( R iechmann et al. 2000 ), although it is comparable with that of human (6.0 %)(Venter et al. 2001 ). In addition to the larger number of TF genes in Arabidopsis, there is a greater variety of TFs, with a greaterdiversity of DNA binding specifi cities, compared with D. melanogaster or C . elegans (see later for more details).These characteristic features of Arabidopsis TFs suggest that transcriptional regulation plays more important roles in plants than in animals. Because transcriptional regulation isFunct ional Analysis of Transcript ion Fact ors in ArabidopsisNobutaka Mitsuda and Masaru Ohme-Takagi ∗R esearch Institute of Genome-Based Biofactory, National Institute of Advanced Industrial Science and Technology (AIST), Central 4,Higashi 1-1-1,Tsukuba,305-8562Japan Special Issue – Mini Reviewthe first step of gene expression and could affect various ‘omes’, namely the proteome, metabolome and phenome, the functional analysis of TFs is important and necessary for omics studies and for the elucidation of whole functional networks in plants. Although much effort has been made to identify the function of TFs, most of their functions remain to be clarifi ed. In this mini-review, we present an overview of Arabidopsis TF s and describe strategies for the functional analysis of plant TF s, which include both traditional and recently developed technologies.O verview of Arabidopsis transcription factorsA ccording to The Arabidopsis Information Resources (TAIR, h ttp://), there are 27,235 protein-coding genes in the Arabidopsis genome (ftp:///home/tair/Genes/TAIR8_genome_release/ README). Four independent reports have recently shown that approximately 2,000 genes encode TFs ( T able 1).These four representative databases of Arabidopsis TFs are: RARTF ( h ttp://rarge.gsc.riken.jp/rartf/) ( I ida et al. 2005 ), AGRIS ( h ttp:///AtTFDB/)(Davuluri et al. 2003 ), DATF ( h ttp:///) ( G uo et al. 2005 ) and PlnT DB ( h ttp://plntfdb.bio.uni-potsdam.de/ v2.0/index.php?sp_id=ATH) ( R iano-Pachon et al. 2007 ). Each database classifi ed TFs into families based on their own classifi cation criteria, and the number of loci in each family is different among the four databases. A total of 51, 51, 64 and 67 families (72 families in total) and 1,965, 1,837, 1,914 and 1,949 loci (2,620 loci in total), respectively, were identifi ed.A total of 1,318 loci are recognized by all four databases ( T able 1and Supplementary Table S1). These differences are mainly due to the different defi nition of a TF in each database. For example, the AGRIS database does not include AUX/IAA proteins as TF s as they do not directly bind to DNA but repress auxin-mediated gene transcription by interacting with ARF transcription factors ( O uellet et al. 2001 ), whereas the other three databases classifi es them as being TFs ( T able 1).A rabidopsis TFs are characterized by a large number of genes and by the variety of gene families when compared with those of D. melanogaster or C. elegans. F or example, zinc-fi nger TFs represent more than half of all TFs in D. mela-nogaster or C. elegans, whereas those in Arabidopsis repre-sent around 20 %( R iechmann et al. 2000 ). Around half of Arabidopsis TFs are plant specifi c and possess DBDs found only in plants ( R iechmann et al. 2000 and T able 1).AP2-ERF, NAC, Dof, YABBY, WRKY, GARP, TCP, SBP, ABI3-VP1 (B3), EIL and LF Y are plant-specifi c TF s. The three-dimensional structures of several plant-specifi c DBDs, i.e. NAC, WRKY, SBP, EIL, B3 and AP2-ERF, have been determined ( A llen et al. 1998 ,E rnst et al. 2004 ,Y amasaki et al. 2004a ,Y amasaki et al. 2004b ,Y amasaki et al. 2005a ,Y amasaki et al. 2005b ). Most Arabidopsis TFs form large families, which share similar DBD structures. For example, AP2-ERF and NAC domain families contain >100 loci each ( T able 1). MYB, MADS box, bHLH (basic helix–loop helix), bZIP and HB, which are not plant-specifi c families, also form large families. These families, such as the MADS box family, which includes a number of ABC fl oral homeotic genes ( R iechmann et al. 1996 ), play impor-tant roles in the control of plant growth and development.T F s act as transcriptional activators or repressors. In common with other eukaryotes, TF s containing domains rich in the acidic amino acids glutamine or proline, such as TOC1, DREBs, ARFs and GBF1, are transcriptional activators ( S chindler et al. 1992 ,U lmasov et al. 1999 ,S trayer et al. 2000 , S akuma et al. 2002 ). In addition, the AHA motif, which has a characteristic pattern of aromatic and large hydrophobic amino acid residues embedded in an acidic context, was shown to act as an activation domain (AD) in plant heat shock factors ( Döring et al. 2000 ).O n the other hand, transcriptional repressors in plants were not elucidated until the ERF-associated amphiphilic repression (EAR) motif was identifi ed in tobacco ETHYLENE RESPONSIVE ELEMENT BINDING F ACTOR 3 (EREBP3) ( O hta et al. 2000 ). Transcriptional repressors are roughly categorized into passive or active repressors. Passive repres-sors have neither an AD nor a repression domain (RD). Some repress transcription by binding to the promoter of the target gene, thereby competing with an activator that inter-acts with the same c is-element. Maize Dof2 is known to be a passive repressor ( Y anagisawa and Sheen 1998 ). Arabidopsis CAPRICE (CPC), TRIPTYCHON (TRY), ENHANCER OF TRY AND CPC1 (ETC1), ETC2 and ETC3, which are all small MYB proteins with a single R3-MYB domain, are negative regula-tors involved in the development of epidermal cells (reviewed in S imon et al. 2007 ) and are likely to act as passive repres-sors. They compete with other R2-R3 MYB proteins such as GLABRA1 (GL1) and WEREWOLF (WER) that positively reg-ulate epidermal cell development for interaction with bHLH proteins ( E sch et al. 2004 ,S imon et al. 2007 ,T ominaga et al. 2007 ). The active repressors possess distinct RDs that confer repressive activity to the TF. The EAR motif is a plant-specifi c repression domain. The minimum unit of the EAR-motif RD is only six amino acids, which comprise an amphiphilic fea-ture composed of leucine and acidic amino acids ( H iratsu et al. 2004 ). Because fusion of the EAR motif RD can convert a transcriptional activator into a strong repressor ( H iratsu et al. 2003 ), TFs that contain this motif are assumed to be transcriptional repressors, although experimental validation is required. Database analysis revealed that the EAR motif RD is found in 404 loci among 2,620 putative TFs (Supple-mentary Table S1). Interestingly, RDs are over-represented in the C2H2 zinc finger (68/136), AUX-IAA (28/29) and HB (32/93) families (Supplementary Table S1). Most RDs are conserved in various plants, including dicots and mono-cots, but are not obviously over-represented in TFs of otherFunctional analysis of transcription factorsN. Mitsuda and M. Ohme-TakagiT able 1 C omparison of plant TF databasesRARTF AGRIS DATF PlnTFDBFamily Loci Family Loci Family Loci Family Loci1.ABI3/VP151ABI3VP111ABI3-VP160ABI3VP156REM212.Alfi n-like47Alfi n-like7Alfi n7Alfi n-like73.AP2/EREBP93AP2-EREBP136AP2-EREBP146AP2-EREBP146ERF19Pti45Pti55Pti6184.ARF71ARF22ARF23ARF23RAV115.ARID6ARID7ARID10ARID106.AT-hook31––––––7.––––AS242––8.Aux/IAA21––AUX-IAA28AUX/IAA279.––BBR/BPC7BBR-BPC7BBR/BPC710.––BZR6BES18BES1811.bHLH157bHLH162bHLH127bHLH13412.––––––bHSH 113.bZIP56bZIP73bZIP72bZIP71TGA32714.C2C2(Zn)-CO-like51C2C2-CO-like30C2C2-CO-like37C2C2-CO-like17Pseudo ARR-B 515.C2C2(Zn)-Dof33C2C2-Dof36C2C2-Dof36C2C2-Dof3616.C2C2(Zn)-GATA37C2C2-Gata30C2C2-GATA26C2C2-GATA2917.C2C2(Zn)-YABBY5C2C2-YABBY6C2C2-YABBY5C2C2-YABBY 618.C2H2(Zn)177C2H2211C2H2134C2H29619.C3H-type 1(Zn)37C3H165C3H59C3H6720.––CAMTA6CAMTA6CAMTA 621.CBF52––––––AAT37CCAAT-DR12CCAAT-Dr12CCAAT43CCAAT-HAP210CCAAT-HAP210CCAAT-HAP310CCAAT-HAP311CCAAT-HAP513CCAAT-HAP51323.CPP(ZN)8CPP8CPP8CPP824.––––––CSD 425.––––––DBP 426.––––––DDT 427.E2F/DP8E2F-DP8E2F-DP8E2F-DP728.EIL6EIL6EIL6EIL 629.––––FHA16FHA1730.GARP51G2-like40GARP-G2-like42G2-like39ARR-B15GARP-ARR-B10ARR-B1331.––GeBP16GeBP21GeBP2032.––––GIF3––33.GRAS32GRAS31GRAS33GRAS3334.––GRF9GRF9GRF935.HB97Homeobox91HB87HB91PAIRED(w/o HB)236.HMG-box11––HMG11HMG11continuedFunctional analysis of transcription factorsTable 1 ContinuedRARTF AGRIS DATF PlnTFDBFamily Loci Family Loci Family Loci Family Loci37.––HRT3HRT-like2HRT238.HSF27HSF21HSF23HSF2339.C3H-type 2(Zn)10JUMONJI5JUMONJI17Jumonji17JUMONJI1340.LFY3LFY1LFY1LFY141.LIM-domain6––LIM13LIM642.––––LUG2LUG243.MADS106MADS109MADS102MADS10244.––––MBF13MBF1345.MYB superfamily189MYB130MYB149MYB145MYB-related67MYB-related49MYB-related6446.NAC106NAC94NAC105NAC10147.Nin-like14NLP9Nin-like14RWP-RK14AtRKD548.––––NZZ1NOZZLE149.PcG; E(z) class32PcG34SET33PcG; Esc class350.PHD-fi nger9PHD11PHD55PHD4351.––––PLATZ10PLATZ1152.––––––RB153.––––S1Fa-like3S1Fa-like354.––––SAP1SAP155.SBP17SBP16SBP16SBP1656.Sir22–––––-57.––––––Sigma70-like658.––––SRS10SRS1059.––––––SNF23860.SW136––––––61.Swi4/Swi61––––––62.––––TAZ9TAZ863.TCP24TCP26TCP23TCP2464.Trihelix31Trihelix29Trihelix26Trihelix2365.TUB11TUB10TLP11TUB1066.––––ULT2ULT267.––VOZ2VOZ2VOZ268.VIP31––––––69.––Whirly3Whirly2PBF-2-like370.WRKY(Zn)72WRKY72WRKY72WRKY7271.––ZF-HD15ZF-HD16zf-HD1772.––ZIM2ZIM18ZIM15Other81Other1Other69 Total1965Total1837Total1914Total1949 The number of loci in each database is shown. ‘–’ indicates that no corresponding TF family is defi ned in the database.organisms, such as yeast (N. Mitsuda et al. unpublished results). These suggest that the EAR motif RD and its mecha-nism of action is plant specific. Recently, novel RDs that could not be categorized according to the EAR motif ( H iratsu et al. 2004 ) were identifi ed in AtMYBL2 and B3 DBD TF s ( M atsui et al. 2008 ,I keda and Ohme-Takagi 2009 ). This sug-gests that unidentifi ed transcriptional repressors with novel RDs may be encoded in plant genomes. Activators and repressors act antagonistically to control the fi ne-tuning of gene expression. The molecular mechanism of transcrip-tional repression via the EAR motif RD remains to be clari-fi ed. Chromatin remodeling may be involved because the EAR motif interacts with TOPLESS (TPL), and mutations in H ISTONE ACETYLTRANSFERASE GNAT SUPERFAMILY 1sup-press the t pl-1phenotype ( L ong et al. 2006 ,S zemenyei et al. 2008 ). In animals, bifunctional TF s have been reported, which can act as transcriptional activators or repressors, depending on the environment or target genes ( A dkins et al. 2006 ). In plants, WRKY53 has been shown to act as either a transcriptional activator or repressor depending on the sequence surrounding the W-box ( M iao et al. 2004 ).T he activities of most TFs are controlled at the transcrip-tional level by other TF s, while several TF s are regulated post-transcriptionally, such as EIN3 ( Y anagisawa et al. 2003 ). Small RNAs that target TF genes are also important regula-tors of gene expression (see T able 3and Bioinformatic anal-ysis). TFs that are regulated at the post-transcriptional level may be regulators that act at early stages of transcriptional cascades.B ioinformatic analysisT he functional analysis of TFs using bioinformatic techniques has become an important and effective strategy. Databases concerned with the functional analysis of TFs are listed in T able 2. Initially, amino acid sequence analysis should be performed to fi nd evolutionarily conserved domains (CDs), including DBDs. Some TFs possess two DBDs. For example, RAV1 (At1g13260) group members have both AP2-ERF and B3 DBDs ( K agaya et al. 1999 ). Normally, a TF has a DBD and a transcriptional AD or RD. TFs having only a DBD are likely to be passive repressors, as reported for CPC and TRY, which interfere with the activity of the transcriptional activator (complex) ( S imon et al. 2007 ). CD searches against known motifs can be performed using many web-based programs. One of the most useful services is InterProScane provided by the European Bioinformatics Institute (EBI) ( h ttp://www.ebi. /Tools/InterProScan/) ( Q uevillon et al. 2005 ). This search is comprehensively performed against various CD databases and provides sophisticated graphical output. Finding known or unknown CDs among a set of proteins can be performed by MEME ( h ttp:///meme/ intro.html) ( B ailey et al. 2006 ). The SALAD database, which was developed specifi cally for plant proteins, also provides MEME-based CD searching with various other useful tools ( h ttp://salad.dna.affrc.go.jp/salad/en/ ).H omology searches, for example performed by BLAST, are also important for the bioinformatic study of TFs. Pro-teins which share high homology not only in their DBDs but also in other regions are likely to be functionally redundant, at least in tissues where they are co-expressed. However, it is frequently observed that proteins with high homology only in the DBD also function redundantly. For example, although CUP-SHAPED COTYLEDON1 (CUC1) and CUC3 are known to function redundantly, there is no significant sequence similarity outside the DBD ( V roemen et al. 2003 ,H ibara et al. 2006 ). BLAST searches against Arabidopsis can be performed through TAIR ( h ttp:///Blast/index.jsp). BLAST searches against multiple species, such as Arabidopsis and rice, can also be informative to assess functional redun-dancy. If three Arabidopsis proteins correspond to one rice protein, these three proteins might function redundantly. BLAST searches against a favorite combination of species can be performed through the NCBI website ( h ttp://blast. /Blast.cgi).A nalysis of the subcellular localization of putative TFs is important because TFs cannot function outside the nucleus. Some NAC domain TFs possess a transmembrane motif at the C-terminus and are liberated by proteolytic cleavage to move into the nucleus ( K im et al. 2006 ,K im et al. 2008 ). Sub-cellular localization of proteins can be predicted by com-puter programs such as SubLoc ( H ua and Sun 2001 ), TargetP ( E manuelsson et al. 2000 ) and WoLF PSORT ( H orton et al. 2007 ). Predictions of subcellular localization using 10 differ-ent computer programs and also from experimental evi-dence can be retrieved from the SUBAII database ( h ttp:// .au/suba2/) ( H eazlewood et al. 2007 ). This database also provides hydropathy plots of all Arabidopsis proteins.T he investigation of proteins that interact with TFs is also of great importance. Many TFs are known to form functional complexes. For example, some NAC TFs and MADS TFs form homo- or hetero-dimeric or tetrameric complexes ( H onma and Goto 2001 ,E rnst et al. 2004 ,H eazlewood et al. 2007 ). MYB TFs and bHLH TFs often form complexes ( Z immermann et al. 2004a ). A number of TFs are known to interact with kinases, resulting in TF phosphorylation ( H e et al. 2002 , F urihata et al. 2006 ,R obertson et al. 2008 ). Predicted or experimentally validated protein–protein interactions (PPIs) among Arabidopsis proteins can be retrieved from the TAIR, EBI and AtPID databases. The ‘Arabidopsis predicted interac-tome’, stored at TAIR, provides a set of >20,000 PPIs based on ortholog matching ( h ttp:///portals/ proteome/proteinInteract.jsp) (Geisler- L ee et al. 2007 ). EBI provides the IntAct database, which stores continu-ously updated PPI information of all organisms based on lit-erature curation ( h ttp:///intact/site/index.jsf )N. Mitsuda and M. Ohme-Takagi( K errien et al. 2007 ). The A rabidopsis thaliana Protein Inter-actome Database (AtPID) provides a search facility with graphical output against a predicted and literature-curated Arabidopsis PPI data set ( h ttp:///index. php) ( C ui et al. 2008 ).m icroRNA (miRNA) is also an important regulator of TF activity. According to the Arabidopsis small RNA Project (ASRP) ( h ttp:///db/)(Gustafson et al. 2005 ,B ackman et al. 2008 ), 200 genes are predicted to be targets of known miRNAs. Interestingly, 69 of these genesT able 2 L ist of useful databases for the functional analysis of TFsC ategory/database name URL CommentPlant ( A rabidopsis) transcription factorsRARTF h ttp://rarge.gsc.riken.jp/rartf/AGRIS h ttp:///AtTFDB/DATF h ttp:/// A part of a plant transcription factor database PlnTFDB h ttp://plntfdb.bio.uni-potsdam.de/v2.0/index.php?sp_id=ATH Data of other plants are also stored Conserved domain searchInterProScan h ttp:///Tools/InterProScan/For known motifsMEME h ttp:///meme/intro.html For discovering unknown motifsSALAD database h ttp://salad.dna.affrc.go.jp/salad/en/For known and unknown motifs Homology searchTAIR BLAST h ttp:///Blast/index.jsp For A rabidopsis onlyNCBI BLAST h ttp:///Blast.cgi For multispecies searchPrediction of subcellular localizationSUBAII h ttp://.au/suba2/Experimental data are also storedProtein–protein interactionA rabidopsis predictedinteractomeh ttp:///portals/proteome/proteinInteract.jspEBI IntAct h ttp:///intact/site/index.jsf For all organismsAtPID h ttp:///index.phpSmall RNAsASRP h ttp:///db/Includes data of miRNA, siRNA and ta-siRNA Repository of microarray dataNCBI GEO h ttp:///geo/EBI ArrayExpress h ttp:///microarray-as/ae/NASCArrays h ttp:///narrays/experimentbrowse.plBrowsing microarray data and co-expression analysisATTED-II h ttp://atted.jp/Genevestigator h ttps:///gv/index.jspBAR eFP browser h ttp://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgiFinding novel c is-elementsTAIR motif analysis h ttp:///tools/bulk/motiffinder/index.jspDatabase of known c is-elementsPLACE h ttp://www.dna.affrc.go.jp/PLACE/No longer updated after 2007AGRIS ATCISDB h ttp:///AtcisDB/GO categorizationTAIR GO annotation search h ttp:///tools/bulk/go/index.jspFunctional analysis of transcription factorsN. Mitsuda and M. Ohme-TakagiT able 3 L ist of TF genes targeted by miRNAm iRNA family miRNA locus Target family Target locusmiR156/miR157AT2G25095 (miR156A)SBP AT5G43270(SPL2)AT4G30972 (miR156B)AT2G33810(SPL3)AT4G31877 (miR156C)AT1G53160(SPL4)AT5G10945 (miR156D)AT3G15270(SPL5)AT5G11977 (miR156E)AT1G69170(SPL6)AT5G26147 (miR156F)AT2G42200(SPL9)AT2G19425 (miR156G)AT1G27370(SPL10)AT5G55835 (miR156H)AT1G27360(SPL11)AT1G66783 (miR157A)AT5G50570(SPL13)AT1G66795 (miR157B)AT3G57920(SPL15)AT3G18217 (miR157C)AT1G48742 (miR157D)miR159/miR319AT1G73687 (miR159A)MYB AT5G06100(ATMYB33)AT1G18075 (miR159B)AT3G11440(ATMYB65)AT2G46255 (miR159C)AT4G26930(ATMYB97)AT4G23713 (miR319A)AT2G32460(ATMYB101)AT5G41663 (miR319B)AT2G26950(ATMYB104)AT2G40805 (miR319C)AT5G55020(ATMYB120)AT3G60460(DUO1)TCP AT4G18390(TCP2)AT1G53230(TCP3)AT3G15030(TCP4)AT2G31070(TCP10)AT1G30210(TCP24)miR160AT2G39175 (miR160A)ARF AT2G28350(ARF10)AT4G17788 (miR160B)AT4G30080(ARF16)AT5G46845 (miR160C)AT1G77850(ARF17)miR164AT2G47585 (miR164A)NAC AT3G15170(CUC1)AT5G01747 (miR164B)AT5G53950(CUC2)AT5G27807 (miR164C)AT5G07680(ATNAC4)AT5G61430(ATNAC5)AT1G56010(NAC1)AT5G39610(ORE1)AT3G12977miR165/miR166AT1G01183 (miR165A)HB AT2G34710(PHB)AT4G00885 (miR165B)AT1G30490(PHV)AT2G46685 (miR166A)AT1G52150(CAN)AT3G61897 (miR166B)AT5G60690(REV)AT5G08712 (miR166C)AT4G32880(ATHB8)AT5G08717 (miR166D)AT5G41905 (miR166E)AT5G43603 (miR166F)AT5G63715 (miR166G)AT3G22886 (miR167A)ARF AT1G30330(ARF6)AT3G63375 (miR167B)AT1G37020(ARF8)AT3G04765 (miR167C)AT1G31173 (miR167D)miR169AT3G13405 (miR169A)CCAAT AT5G06510(NF-YA10)AT5G24825 (miR169B) AT1G72830(HAP2C)AT5G39635 (miR169C)AT1G17590(NF-YA8)AT1G53683 (miR169D)AT1G54160(NF-YA5)AT1G53687 (miR169E)AT5G12840(HAP2A)AT3G14385 (miR169F)AT3G20910(NF-YA9)AT4G21595 (miR169G)AT3G05690(HAP2B)continued(35 %) encode putative TF s ( T able 3and Supplementary Table S1), despite TFs representing only 5–10 %of all genes. For example, miRNAs that target TCP, NAC and SBP TFs are known to play very important roles in the control of plant growth and development ( P alatnik et al. 2003 ,W u and Poethig 2006 ,O ri et al. 2007 ,S chommer et al. 2008 ,K im et al. 2009 , L arue et al. 2009 ). ASRP also provides non-coding small RNA information (reviewed in R amachandran and Chen 2008 ), such as for small interfering RNAs (siRNAs) and trans-acting siRNAs (ta-siRNAs), in addition to that for miRNAs. Six TF genes are listed as targets of ta-siRNAs in ASRP.T he spatial and temporal expression profi le of a gene and the expression in response to varying conditions is funda-mental to its biological function. Development of microar-ray technologies and public data repositories, such as NCBI Gene Expression Omnibus (GEO) ( h ttp://www.ncbi.nlm.nih. gov/geo/) ( B arrett et al. 2007 ), EBI ArrayExpress ( h ttp:// /microarray-as/ae/) ( P arkinson et al. 2009 ) and NASCArrays ( h ttp:///nar-rays/experimentbrowse.pl) ( C raigon et al. 2004 ), enables us to access many kinds of microarray data easily. The expres-sion profi le of a gene of interest can also be easily accessed on many web sites, such as ATTED-II ( h ttp://atted.jp/) ( O bayashi et al. 2009 ), Genevestigator ( h ttps://www. /gv/index.jsp) ( Z immermann et al. 2004b , G rennan 2006 ) and BAR eF P browser ( h ttp://bbc.botany. utoronto.ca/efp/cgi-bin/efpWeb.cgi ) ( W inter et al. 2007 ). These web sites also provide information regarding co-ex-pression analysis. They provide lists of genes whose expres-sion profi les are positively or negatively correlated with theT able 3 Contuniedm iRNA family miRNA locus Target family Target locusAT1G19371 (miR169H)AT3G26812 (miR169I)AT3G26813 (miR169J)AT3G26815 (miR169K)AT3G26816 (miR169L)AT3G26818 (miR169M)AT3G26819 (miR169N)miR170/miR171AT5G66045 (miR170)GRAS AT2G45160AT3G51375 (miR171A)AT3G60630AT1G11735 (miR171B)AT4G00150(SCL6)AT1G62035 (miR171C)miR172AT2G28056 (miR172A)AP2-EREBP AT2G28550(TOE1)AT5G04275 (miR172B)AT5G60120(TOE2)AT3G11435 (miR172C)AT5G67180(TOE3)AT3G55512 (miR172D)AT4G36920(AP2)AT5G59505 (miR172E)AT2G39250(SNZ)AT3G54990(SMZ)miR393AT2G39885 (miR393A)bHLH AT3G23690(bHLH077) AT3G55734 (miR393B)miR396AT2G10606 (miR396A)GRF AT2G22840(ATGRF1) AT5G35407 (miR396B)AT4G37740(ATGRF2)AT2G36400(ATGRF3)AT3G52910(ATGRF4)AT5G53660(ATGRF7)AT4G24150(ATGRF8)AT2G45480(ATGRF9) miR778AT2G41616 (miR778A)SET AT2G35160(SGD9)AT2G22740(SDG23) miR824AT4G24415 (miR824A)MADS AT3G57230(AGL16) miR828AT4G27765 (miR828A)MYB AT1G66370(ATMYB113) miR858AT1G71002 (miR858A)MYB AT2G47460(ATMYB12)AT3G08500(ATMYB83)Functional analysis of transcription factorsquery gene ( Z immermann et al. 2004b ,T oufi ghi et al. 2005 , G rennan 2006 ,O bayashi et al. 2009 ). Co-expression analysis is particularly important for the functional analysis of TF s because co-expressed genes might encode proteins that are functionally related and/or are putative interacting proteins. They might also be downstream and/or upstream genes in the context of a transcriptional cascade. H irai et al. (2007) identifi ed MYB TF s as the key regulators of aliphatic glu-cosinolate biosynthesis by co-expression analysis. From pro-moter regions of co-expressed genes, short sequences that are statistically over-represented and may represent TF-binding sites can be identified using TAIR Motif Analysis (/tools/bulk/motiffinder/index. jsp). It is valuable to compare these sequences with known c is-elements stored in c is-element databases, such as PLACE ( h ttp://www.dna.affrc.go.jp/PLACE/) ( H igo et al. 1999 ) and AGRIS ATCISDB ( h ttp:/// AtcisDB/ ). Furthermore, functional characteristics of these genes can be analyzed using the TAIR Gene Ontology (GO) annotation search ( h ttp:///tools/bulk/ go/index.jsp). A set of favorite genes can be categorized based on a limited GO term ( A shburner et al. 2000 ) and shown as a graphical pie chart. In addition, we can compare these data with the results of functional categorization of all Arabidopsis proteins. These analyses help us to speculate on the biological processes in which the set of genes and the query TF are involved.M olecular analysisT he molecular analysis of TFs involves the characterization of their activation or repression activities. To this end, analy-sis using reporter and effector genes is often employed ( O hta et al. 2001 ;F ig. 1). A commonly used effector gene consists of a chimeric construct, in which a TF-coding sequence is fused to that of a heterogeneous DBD, such as the GAL4 DNA-binding domain (GAL4DB) from yeast, and which is driven by a strong promoter, such as caulifl ower mosaic virus (CaMV) 35S. The reporter gene is usually the fi refl yluciferase (LUC) or E scherichia coliβ-glucuronidase(GUS) gene, which is driven by a minimal promoter with upstream repeated c is-elements, such as the binding sequence of the GAL4DB. Another reporter construct, containing a constitu-tively expressed reporter, such as sea pansy L UC, is used as an internal control (reference). These reporter, effector and ref-erence constructs are transiently co-expressed by particle bombardment of leaf tissues or by polyethylene glycol-mediated transformation of leaf protoplasts or cultured cells. By assaying the activity of the reporter gene following co-expression of the effector gene, the activation activity of a TF can be examined. A transient expression assay using particle bombardment into leaf tissue is simple and repro-ducible and has several advantages for analyzing the molec-ular function of TFs ( U eki et al. 2009 ). Once a TF is identifi ed as a transcriptional activator, the AD can be determined by investigating the activities of truncated TF proteins. Activa-tion activity of TFs can also be assessed in a yeast system. However, ADs identifi ed in a plant system can sometimes differ from those identifi ed in a yeast system ( O hta et al. 2000).S ome Arabidopsis TFs are known to act as transcriptional repressors. To analyze whether the TF of interest is a repres-sor, the repressive activity of the effector construct, using a reporter gene containing a transcriptional enhancer in the promoter, such as that of the CaMV 35S promoter, is uti-lized. As in the case of an activator, by analyzing the repres-sive activities of truncated proteins, it is possible to identify the RD of the TF. This strategy for the molecular analysis of TFs is summarized in F ig. 1.Fig.1S chematic drawing summarizing the molecular analysis of TFs. The effector and reporter plasmids are co-introduced into Arabidopsis leaf by particle bombardment. Reporter activity is measured to examine whether a TF is an activator or a repressor.N. Mitsuda and M. Ohme-Takagi。

Chromatin regulation offlowering Yuehui HeDepartment of Biological Sciences,National University of Singapore,Temasek Life Sciences Laboratory,Singapore117604, Republic of SingaporeThe transition toflowering is a major developmental switch in the life cycle of plants.In Arabidopsis(Arabi-dopsis thaliana),chromatin mechanisms play critical roles inflowering-time regulation through the expres-sion control of keyflowering-regulatory genes.Various conserved chromatin modifiers,plant-specific factors, and long noncoding RNAs are involved in chromatin regulation of FLOWERING LOCUS C(FLC,a potentfloral repressor).The well-studied FLC regulation has provid-ed a paradigm for chromatin-based control of other developmental genes.In addition,chromatin modifica-tion plays an important role in the regulation of FLOW-ERING LOCUS T(FT,encodingflorigen),which is widely conserved in angiosperm species.The chromatin mech-anisms underlying FT regulation in Arabidopsis are likely involved in the regulation of FT relatives and,therefore,flowering-time control in other plants.Control of the transition tofloweringThe timing of the transition from a vegetative to reproduc-tive phase is critical for reproductive success in the angio-sperm life cycle.Many species have evolved multiple pathways responding to environmental cues and endoge-nous factors to regulateflowering time properly.In Arabi-dopsis thaliana,a facultative long-day plant,the vernalization,thermosensory,and photoperiod pathways, sensing cold winter,ambient temperature,and long days, respectively,together with other pathways responding to internal factors,such as age and the plant hormone gibber-ellins,form a regulatory network to control when toflower (Figure1).To date,keyflowering genes in this network have been uncovered and their expression regulation has been under intensive investigation.In this review,I focus on the regulation offlowering genes andflowering by chromatin-based mechanisms uncovered from Arabidopsis.In Arabidopsis,FLC functions as a centralfloral repres-sor,and its expression is under complex control[1,2]. FRIGIDA(FRI)activates or upregulates FLC expression to a higher level that inhibitsflowering,whereas vernali-zation,a prolonged period of cold exposure(a typical winter), overrides the function of FRI to silence FLC expression and so enableflowering[3,4].The vernalization-responsive winter annual versus rapid-cycling(early-flowering)growth habit is typically determined by the levels of FLC expression.Winter annuals have dominant alleles of FRI and FLC,whereas in rapid-cycling accessions that lack a functional FRI,FLC expression is repressed by the so-called ‘autonomous-pathway’genes or constitutive FLC repressors that repress FLC expression independently of environmental inputs[5,6].FT is another key player inflowering-time regulation,and the FT protein is aflorigen(flowering inducer)[7–10].FT expression is activated by CONSTANS(CO)in long days, whereas FLC directly represses FT expression[11–13].FT is expressed in the vasculature,typically in the phloem of leaves.The FT protein moves from the phloem to the shoot apical meristem,where it subsequently forms a complex with the bZIP transcriptional factor FD to activate the expression offloral-meristem identity genes LEAFY and APETALA1 (AP1),leading tofloral primordium formation[14,15].Chromatin modifications are involved in the regulation of developmental genes in plants.Such modifications,includ-ing nucleosome remodeling,DNA methylation,and various histone modifications,regulate chromatin structure and gene expression[16].In general,histone acetylation,histone H3lysine-4trimethylation(H3K4me3),H2B monoubiqui-tination(H2Bub1),and H3lysine-36di-and trimethylation (H3K36me2/me3)are linked with active gene expression, whereas histone deacetylation,H3lysine-9methylation,H3 lysine-27trimethylation(H3K27me3),and H2A monoubi-quitination(H2Aub1)are associated with gene repression [17].In Arabidopsis,chromatin-mediated FLC regulation has been well studied;various modifiers have been shown to mediate FLC regulation and soflowering[3,4,17].For in-stance,the ATX1H3K4methyltransferase and the EFS H3K36methyltransferase mediate H3K4and H3K36meth-ylation on FLC chromatin,respectively;both are also re-quired for FLC expression and so act to inhibit Arabidopsis flowering[18–20].By contrast,Polycomb repressive com-plex2(PRC2)-like complexes deposit H3K27me3at FLC to silence FLC repression[21,22].FLC regulation has become a paradigm for understanding the expression control of other developmental genes in plants.Recent advances in understanding FLC regulation have shown that FRI,a plant-specific scaffold protein,is part of a complex recruiting chromatin modifiers to the FLC locus, leading to FLC activation.In addition,it has recently been revealed that long noncoding RNAs(lncRNAs;with>100 nucleotides)trigger not only FLC repression in rapid-cycling accessions,but also vernalization-mediated FLC silencing in winter annuals.Here I review the current understanding of chromatin-mediated FLC regulation and discuss how FRI or lncRNAs‘engage’various chromatin modifiers at the FLC locus to regulate its expression and thusflowering.Further-more,as recent studies have found that FT expression is also partly regulated by chromatin-based mechanisms,I describe the current views on chromatin regulation of FT expression.Corresponding author:He,Y.(dbshy@.sg).Keywords:chromatin modification;FLC;FT;flowering time;lncRNAs.5561360-1385/$–see front matterß2012Elsevier Ltd.All rights reserved./10.1016/j.tplants.2012.05.001Trends in Plant Science,September2012,Vol.17,No.9FRI mediates chromatin modifications at the FLC locus to establish the winter-annual growth habitFRI,encoding a plant-specific scaffold protein,is a major determinant of natural variation in Arabidopsisflowering time[6],and many players involved in FRI-dependent FLC activation have been identified in genetic screens for mutants that suppress FLC activation in an FRI-containing line.These include both conserved chromatin modifiers and plant-specific components[3,5].Loss-of-function mutations in these components suppress FLC expression and so render a FRI-containing line earlyflowering.In addition,some of these players,such as chromatin modifiers,control multiple gene expression in the Arabidopsis genome.Here,I focus on their roles for FLC regulation.Thefirst identified conserved component for the function of FRI was the RNA Polymerase II(Pol II)Associated Factor 1complex(PAF1c)[23–25],which is highly conserved from yeast to plants and humans,and associates with Pol II during transcription.The Arabidopsis PAF1c is composed of six subunits,and its functional disruption leads to a reduction in H3K4me3,H3K36me2,and H3K36me3on FLC chromatin and suppresses FLC expression in the FRI background[19,23].In addition,PAF1c is also required for genome-wide H2Bub1[26].PAF1c itself does not possess histone-modifying activities,but is believed to serve as a platform for docking histone-modifying enzymes during transcriptional activation and elongation.COMPASS-like H3K4methyltransferase complexes deposit H3K4me3at FLC to activate its -PASS contains four conserved core subunits,namely a SET-domain H3K4methyltransferase plus three structural core components known as WDR5a,RBL,and ASH2R in Arabidopsis;these form a stable core subcomplex providing a structural platform for H3K4methylation[27,28].Two H3K4methyltransferases,known as ATX1and ATXR3 (or SDG2),and two putative ones called ATX2and ATXR7, are involved in H3K4trimethylation at FLC[18,29–32]. ATX1has been shown to associate with the WDR5a subcomplex[27];it is likely that ATX2,ATXR3,and ATXR7 also function in the context of COMPASS for H3K4me3on FLC chromatin.H3K4me3accumulates mainly in the region around the transcription start site(TSS)of FLC [23],and COMPASS components are directly required for this accumulation and for FLC activation[27,28].Further-more,overexpression of ASH2R leads to elevated H3K4 trimethylation at FLC and its activation[28],suggesting that elevated H3K4trimethylation is sufficient to activate FLC expression to inhibitflowering.In addition to H3K4me3,FLC activation in the FRI background also requires H3K36methylation and H2Bub1 catalyzed by EFS and the H2Bub1complex,known as HUB–UBC,respectively.EFS catalyzes H3K36me2and me3across FLC[19,20].The HUB–UBC complex compris-ing the E3ubiquitin ligases HUB1and HUB2,and an E2 ubiquitin-conjugating enzyme called UBC1or UBC2,cat-alyzes genome-wide H2Bub1,including the FLC locus [33–35].Using a reconstituted human chromatin-assembly system,H2Bub1has been shown to facilitate Pol II move-ment through nucleosomes in the body of a gene via histone chaperone FACT-mediated H2A-H2B replacement and nucleosome reassembly[36].H2Bub1in the gene body may function cooperatively with the conserved FACT to facilitate FLC transcription elongation because mutations in the FACT components SPT16and SSRP1cause FLC suppression[37].In addition,H2Bub1homeostasis at FLC is critical for its expression.H2B deubiquitination by the UBP26deubiquitinase on FLC chromatin is also required for FLC expression[26].This suggests that either appro-priately balanced H2Bub1levels or subsequent H2B deubiquitination following the HUB–UBC activity are critical for FLC transcription.FRI-meditated FLC activation also requires the deposi-tion of histone variant H2A.Z by the conserved SWR1 complex(SWR1c)at FLC[38,39].Functional disruption of SWR1c prevents H2A.Z deposition on FLC chromatin and suppresses FLC expression,leading to earlyflowering. SWR1c is an ATPase chromatin-remodeling complex that substitutes canonical H2A with H2A.Z near50ends of genes to promote transcriptional competence in Arabidop-sis[40].It has been reported that H2A.Z nucleosomes with further modifications(e.g.,acetylation)become labile and are evicted to facilitate active transcription[41,42].H2A.Z deposition around the FLC TSS may promote FLC tran-scription through this mechanism.In addition to the chromatin modifiers,FRI-dependent FLC activation requires two plant-specific factors called FRL1and FES1,and two components known as SUF4and FLX[43–47];these proteins,together with FRI(as a scaffold protein),form a putative transcription activator complex(FRIc)in which SUF4recognizes a cis-element in the FLC proximal promoter[48].Loss of function of each FRIc subunits suppresses FLC expression and results in earlyflowering without any other obvious phenotypes,Figure1.Flowering-time regulation in Arabidopsis.The vernalization pathwaysilences FLOWERING LOCUS C(FLC)expression,and the photoperiod andthermosensory pathways induce FLOWERING LOCUS T(FT)expression inresponse to long day and temperature rise,respectively;in addition,blue lightpromotes FT expression.FRIGIDA(FRI)activates FLC expression,whereas theautonomous or constitutive FLC repressors(e.g.,FCA,FVE,and FPA)repress FLCexpression independent of environmental cues.FLC protein directly represses theexpression of flowering-pathway integrators FT and SUPPRESSION OFOVEREXPRESSION OF COSTANS1(SOC1),which act to promote the expressionof the floral-meristem identity genes APETALA1(AP1)and LEAFY.In addition,developmental age and gibberellins also promote flowering.Unbroken lines witharrows indicate activation or upregulation of gene expression,whereas linesending in bars indicate repression.Trends in Plant Science September2012,Vol.17,No.9557suggesting that this complex is an FLC-specific activator. FRIc directly associates with the chromatin modifiers EFS and SWR1c[20,48].Furthermore,a functional FRI is required for WDR5a enrichment at FLC,suggesting that FRIc recruits the COMPASS H3K4methylation activity to FLC chromatin[27].Together,thesefindings suggest that FRIc recruits or enriches multiple chromatin modifiers at FLC to activate its expression.Indeed,multiple active chromatin modifications are required for,or associate with, FRI-dependent FLC activation,including H2A.Z deposi-tion,histone acetylation,H3K4me3,H2Bub1,and H3K36me2and me3.Recent studies show that there are functional dependences among these modifications at FLC; for instance,loss of EFS activity leads to a reduction in not only H3K36me3,but also H3K4me3in FRI-containing lines[20,49].In short,upon binding to the FLC proximal promoter,FRIc recruits or enriches multiple active chro-matin modifications at the FLC locus to establish a chro-matin environment conducive for the activation of FLC expression,leading to the establishment of the winter-annual growth habit.As described above,there is a functional dependence of FRIc on PAF1c at the FLC locus.The conserved Paf1c in yeast has been shown to associate with Pol II and multiple chromatin modifiers,including COMPASS,an H3K36 methyltransferase,and an H2Bub1complex[50].At the FLC locus,PAF1c may function cooperatively or directly associate with FRIc,resulting in the recruitment or en-richment of active chromatin-modification activities and so promotion of FLC transcription.Two models summarizing the role of FRIc in FLC transcriptional activation and the role of PAF1c in FLC transcription elongation are pre-sented in Figure2.Chromatin silencing of FLC in rapid-cycling accessions involves lncRNAsIn rapid-cycling accessions lacking a functional FRI,FLC is typically repressed by the autonomous or constitutive FLC repressors,among which are FPA,FCA,FY,FVE, FLOWERING LOCUS D(FLD),Polycomb(PcG)compo-nents,and HISTONE DEACETYLASE6(HDA6) [5,17,51,52].These components directly interact with the FLC locus and are involved in FLC chromatin silencing, leading toflowering promotion.It should be noted that,in addition to FLC,these genes also silence other loci.Here,Ifocus on their roles in FLC silencing.FLD,an H3K4demethylase,forms a corepressor-like complex with HDA6and the histone-binding proteins FVE or MSI5,an FVE homolog functioning redundantly with FVE,to repress FLC expression[51,53].Functional loss of each component of this complex leads to increased levels of acetylated core histones and H3K4methylation on FLC chromatin,resulting in FLC derepression[51,53–55]. Thus,the FLD–HDA6–FVE or MSI5complex deacetylates histone tails and demethylates H3K4at FLC to repress its expression and so promote thefloral transition.FLC silencing also requires PRC2-deposited H3K27me3, a repressive modification[56,57].PRC2,first identified in Drosophila,is conserved in higher eukaryotes.Arabidopsis PRC2-like components are required for H3K27me3at FLC and for its repression,which include two H3K27methyltransferases CURLY LEAF(CLF)and SWINGER (SWN),and structural subunits FERTILIZATION-INDE-PENDENT ENDOSPERM(FIE),VERNALIZATION2 (VRN2)and EMBRYONIC FLOWER2(EMF2;a VRN2 homolog)[56,58].These components are predicted to act in the context of a PRC2-like complex that deposits H3K27me3at FLC to repress its expression.I call this complex PRC2NV(NV for non-vernalization)to avoid confu-sion with another PRC2-like complex called PHD–PRC2 involved in vernalization-mediated FLC silencing(see de-scription below).PRC2NV may function in concert with the FLD complex to repress FLC expression.Indeed,there is functional interdependence of these two complexes at FLC: loss of FLD function causes a reduction in H3K27me3and, conversely,loss of PRC2NV activity leads to an increase in H3K4me3[56,59].This suggests that PRC2NV and the FLDAcSWR1cFRIcCOMPASSPol IIEFS(a) Chromatin-modifier recruitment at FLCPol IIK36me3 K36me3K4me3AcK36me3K4me3Ac K36me3EvictionH2A.Z nucleosomePol IIH2Bub1 H2Bub1K4me3EFSPoP l I I IIFSPA F1cFLC mRNA(b) Chromatin modifications and transcriptional activation(c) Transcriptional elongationK36me3K4me3 H2Bub1P AF1cH UBU BCH UBU BCTRENDS in Plant Science Figure2.Working models for FRIGIDA(FRI)-dependent FLOWERING LOCUS C (FLC)activation.(a)Model of FRI complex(FRIc)-mediated recruitment or enrichment of chromatin modifiers at the FLC locus around the transcription start site(TSS;indicated with the broken arrow).Upon binding to the proximal FLC promoter,probably via its subunit SUF4,FRIc increases the binding of SWR1c, COMPASS[for histone H3lysine-4trimethylation(H3K4me3)],and EFS[for histone H3lysine-36di-and trimethylation(H3K36me2/me3)]to the region around the TSS. In addition,RNA Polymerase II(Pol II)-Associated Factor1complex(PAF1c),in association with Pol II,may recruit the HUB–UBC complex and function cooperatively,or associate directly,with FRIc to enrich EFS and COMPASS at FLC.(b)Transcriptional activation of FLC expression.In the region around the FLC TSS,nucleosomes are modified with active marks,including acetylation(Ac), H3K4me3(K4me3),H3K36me3(K36me3),and H2B monoubiquitination(H2Bub1) (by HUB–UBC).H2A.Z is deposited by SWR1c into the nucleosome located at the TSS that is subsequently evicted to facilitate the transcription by Pol II.(c) Elongation of FLC transcription.The nucleosomes of the FLC gene body are modified with H3K36me3and H2Bub1.PAF1c travels with Pol II and may recruit EFS and HUB–UBC to catalyze H3K36me3and H2Bub1,respectively.Note that the FRI protein does not interact with the FLC gene body.Trends in Plant Science September2012,Vol.17,No.9558complex function coordinately to demethylate H3K4,dea-cetylate histones and deposit H3K27me3on FLC chromatin,leading to the establishment of a repressive chromatin environment to repress FLC expression.FLC silencing also involves lncRNAs,the RNA-binding proteins FCA and FPA,and the RNA 30-end processing factors FY,CstF64,and CstF77[52,60].At the FLC locus,sense and antisense transcripts are coexpressed,and there are two classes of antisense FLC transcript:Class I and Class II,due to alternative 30polyadenylation.Class I RNAs are polyadenylated at the proximal poly(A)site,whereas Class II results from polyadenylation at a distal site [52,60].The CstF 30-end processing complex containing CstF64and CstF77,is required for 30-end processing of antisense FLC transcripts;both FPA and FCA (in the context of FCA–FY association)promote CstF activity at the proximal poly(A)site,leading to the production of Class I lncRNAs.This co-transcriptional antisense RNA 30-end processing leads to FLC silencing [52],although the un-derlying mechanism for this silencing remains elusive.FCA-and FPA-mediated FLC repression requires the H3K4demethylase FLD [55,61].In addition,genetic anal-ysis of the fpa;fve double mutant has revealed that FPA and FVE act non-additively to repress FLC expression [62],suggesting that FVE ,similar to FLD ,functions in cooper-ation with FPA to mediate FLC chromatin silencing.Given the recent identification of the FLD–HDA6–FVE or MSI5complex [53],it is likely that this complex functions in concert with PRC2NV and the proximal 30-end processing of antisense FLC transcripts to repress FLC expression (Figure 3).As described above,FLC is repressed by chromatin silencing in many rapid-cycling accessions with a fri allele;introgression of a functional FRI into a rapid-cycling ac-cession leads to chromatin activation of FLC expression[48].Now the question is how FRI or FRIc overcomes FLC chromatin silencing exerted by the constitutive FLC repressors.It has been shown that,in FRI -containing lines,CLF (presumably PRC2NV )binding to FLC chroma-tin is substantially reduced [59].This suggests that FRIc or FRIc-mediated active chromatin modifications suppress PRC2NV binding to FLC chromatin,leading to a reduction of H3K27me3at FLC .Interestingly,in the FRI back-ground,FLD binding to FLC chromatin is not affected;the increased levels of H3K4me3at FLC are expected to result largely from elevated H3K4trimethylation by COM-PASS [27].Thus,FRIc acts to shift the equation of H3K4methylation and demethylation toward methylation.In conclusion,FRI or FRIc exploits at least two biochemical mechanisms to overcome the chromatin silencing of FLC by the constitutive FLC repressors,leading to chromatin activation of FLC .Epigenetic ‘Memory of Winter’involves lncRNAs and Polycomb componentsWinter annuals require vernalization to acquire the com-petence to flower.Vernalization silences FLC expression,leading to acceleration of the floral transition upon the return of plants to a warm condition (typically approxi-mately 20–258C).This silencing involves lncRNAs and PcG components,is maintained after the plants resume growth in warm conditions,and so is mitotically stable,conferring the cold-experienced plants the ‘memories of winter’[4,63,64].Previous genetic screens for mutants with a defect in vernalization response have identified a few PcG or PcG-related components required for FLC silencing,known as VERNALIZATION 1(VRN1),VRN2,VERNALIZATION INSENSITIVE 3(VIN3),VRN5(or VIL1),and LIKE HET-EROCHROMATIN 1(LHP1)[4,21].VIN3,encoding a plant homeodomain (PHD)protein,is induced by pro-longed cold exposure and expressed only in the cold [65].VIN3,VRN5,VRN2,CLF (or SWN),FIE,and MULTI-COPY SUPPRESSOR OF IRA 1(MSI1)form the PHD–PRC2complex during vernalization [21,22].This complex deposits H3K27me3in the region around the first exon of FLC ,the nucleation region,to initiate FLC silencing [64].Upon return to warm conditions,the PHD–PRC2(with-out VIN3)continues to catalyze H3K27me3,which is further spread from the nucleation region to across the entire FLC locus,including the 50promoter,gene body,and 30end;in addition,the levels of H3K27me3are strongly enhanced [21,64,66].This spreading and enhancement requires mitotic activity (i.e.,DNA replication)in Arabi-dopsis [66].The H3K27me3mark is recognized and bound by LHP1[67],a component of the Arabidopsis Polycomb repressive complex 1(PRC1)[68],that acts with PRC2for PcG-mediated FLC silencing.VRN1,a putative PRC1component [68,69],may function with LHP1to maintain FLC silencing after the cold.In summary,the spreading and enhancement of H3K27me3at the FLC locus and subsequent reading of this mark by LHP1(presumably PRC1)function to lock in FLC chromatin at a silenced state,which is stably maintained during cell division after the plant resumes growth in the warm,leading to flowering competence.H3K4 demethylationDeacetylation FLD FVE/MSI5HDA65ʹ IncRNAs(A)nCstFFCA–FYFPAProximal polyadenylationPRC2NVPRC1Antisense transcription5ʹSense transcription5ʹCrosstalkH3K27 trimethylationH2Aub1TRENDS in Plant ScienceFigure 3.Model of chromatin silencing of FLOWERING LOCUS C (FLC ).At the FLC locus,sense and antisense transcripts are coexpressed.The CstF-mediated 30-end processing of FLC antisense transcripts at the proximal poly(A)site,promoted by FPA and FCA–FY,leads to FLC silencing.This processing functions in concert with the FLD–HDA6–FVE or MSI5complex,A Polycomb repressive complex 2called PRC2NV and a PRC1-like complex containing LIKE HETEROCHROMATIN 1(LHP1)and EMBRYONIC FLOWER 1(EMF1)that deposits histone H2A monoubiquitination (H2Aub1),to establish a repressive chromatin environment for FLC silencing.Unbroken black lines with arrows indicate promotion,broken black arrows indicate crosstalk,and black lines ending with bars indicate repression.Abbreviation:HDA6,HISTONE DEACETYLASE 6.Trends in Plant Science September 2012,Vol.17,No.9559The length of cold exposure exerts a quantitative effect on FLC silencing in the warm:that is,quantitative epige-netic memory[4,64].Upon vernalization,FLC chromatin has a bistable epigenetic state:active or silenced[64]. Recent studies have revealed that the quantitative FLC silencing in vernalization-responsive tissues in the warm is determined by the percentage of the cells in which FLC chromatin has been switched to a stable silenced state[64].A longer period of cold leads to an increased fraction of cells with a silenced FLC and,consequently,a lower level of FLC transcripts in the warm at a tissue level.Why cells in a vernalization-responsive tissue(e.g.,root tip)need different periods of cold exposure to switch FLC chromatin from an active to silenced state is unknown.This might be due to variability in the time that a cell takes to perceive cold,its mitotic state and/or the time it takes for a chro-matin modifier(e.g.,VIN3)to be induced in a cell.PcG silencing activity is recruited to FLC chromatin in the cold by lncRNAs.Cold exposure transiently induces the expression of two types of lncRNA transcribed from the FLC locus:COOLAIR and COLDAIR[70,71].COOLAIR com-prises the Class I and II polyadenylated antisense FLC transcripts,and has been proposed to trigger vernaliza-tion-mediated FLC silencing by a co-transcriptional mecha-nism[70].A recent study using FLC mutants with insertional T-DNAs that can block the full-length transcrip-tion and separate the sense and antisense transcription,has revealed that elimination of COOLAIR transcription does not impair FLC silencing by vernalization[72].Thus,COOLAIR is not essential for FLC silencing.Of note,it is likely that COOLAIR might participate in vernalization-mediated FLC silencing via a co-transcriptional mechanism involving an overlapping transcription of sense and anti-sense transcripts.However,the precise role of COOLAIR in FLC silencing remains to be determined.Beside COOLAIR,the unpolyadenylated COLDAIR lncRNAs transcribed from thefirst intron of FLC are also transiently induced by the cold[71].The CLF subunit of PHD–PRC2associates with COLDAIR,leading to the enrichment of this complex on FLC chromatin in the cold [71].A partial deletion of the COLDAIR promoter or knockdown of COLDAIR expression compromises FLC silencing[67,71].Thus,COLDAIR is required for FLC silencing by vernalization.In summary,the currentfind-ings suggest that cold exposure transiently induces the expression of lncRNAs that recruit PcG activity to FLC chromatin,leading to FLC silencing in the cold and upon return to warm conditions.Chromatin-mediated regulation of FT expressionFT expression is induced in the vasculature by a long-day photoperiod and ambient temperature rise.Recent studies have shown that various chromatin modifiers are involved in the regulation of FT expression,including SWR1c, PRC2,LHP1,the REF6H3K27demethylase,and the PKDM7B(also known as AtJMJ4or JMJ14)H3K4 demethylase(Figure4).FT expression is repressed by PcG activity in both long and short days.It has been shown that CLF binds to FT chromatin and is required for H3K27me3deposition on FT chromatin and for FT repression[56].In addition,other PRC2components,including SWN,EMF2,and FIE are also required for FT repression[56,58],suggesting that a PRC2-like complex deposits H3K27me3at FT to repress its expression in the vasculature.Trimethyl H3K27is dynami-cally removed by H3K27demethylases.A recent study has revealed that REF6,a JmjC-domain H3K27demethylase,is involved in H3K27demethylation at FT and is required for its expression[73].Hence,the levels of H3K27me3at FT are dynamically controlled by the PRC2and REF6.As described above,the H3K27me3mark is read by LHP1.Indeed,LHP1 binds directly to FT chromatin to repress FT expression in the vasculature[69].In addition,another putative PRC1-like component,called EMF1,is also required for FT repres-sion[74,75].Together,thesefindings suggest that a PRC1-like complex acts in cooperation with the PRC2to repress FT expression and so inhibitflowering.FT chromatin has a bivalent structure,simultaneously carrying the active H3K4me3and repressive H3K27me3 marks[56].PRC2-dependent H3K27me3suppresses,but does not eliminate,H3K4me3on FT chromatin,and vice versa.Loss of PRC2function not only eliminates H3K27me3, but also causes an increase in H3K4me3at FT[56].Recent studies have revealed that the di-and trimethyl H3K4 demethylase PKDM7B binds to FT chromatin and mediates H3K4demethylation at FT to repress its expression[76–78]. Loss of PKDM7B activity leads to a increase in H3K4me3 and a reduction in H3K27me3[76,77].Thus,H3K4and H3K27trimethylation act antagonistically at the FT locus, and the relative levels of H3K4me3versus H3K27me3play a critical role in the regulation of FT expression.FT expression is induced by ambient temperature rise through the thermosensory pathway,and H2A.Z-containingPhotoperiodCO H3K4me3PKDM7BH3K27me3Temperature riseH2A.Z evictionAP2-likeREF6PRC2FLCSVPTEM1 & TEM2P R C1FTTRENDS in Plant ScienceFigure4.Chromatin-mediated regulation of FLOWERING LOCUS T(FT)expression in the leaf vasculature(phloem).Long-day photoperiod induces the accumulation of CONSTANS(CO)protein that activates FT expression.CO might promote histone H3 lysine-4trimethylation(H3K4me3),whereas PKDM7B(also known as AtJMJ4or JMJ14)demethylates H3K4to repress FT expression.A Polycomb repressive complex2(PRC2)deposits H3lysine-27trimethylation(H3K27me3),whereas REF6 demethylates H3K27,to maintain a proper level of H3K27me3at FT in the vasculature.A PRC1complex reads the H3K27me3mark and acts to repress FT expression.A rise in temperature causes the eviction of the H2A.Z nucleosome located around the transcriptional start site(TSS)of FT to facilitate transcription.FT expression is directly repressed by three types of transcriptional factor:SHORT VEGETATIVE PHASE(SVP)and FLOWERING LOCUS C(FLC)(MADS box), TEMPRANILLO1and2[TEM1and TEM2(B3domain)],and APETALA2(AP2)-like. Unbroken blue lines with arrows indicate promotion,broken blue lines indicate potential promotion,and unbroken blue lines ending with bars indicate repression.Trends in Plant Science September2012,Vol.17,No.9560。

Antisense suppression of the Arabidopsis PIF3gene does not a¡ectcircadian rhythms but causes early £owering andincreases FT expressionAtsushi Oda a ,Sumire Fujiwara a ,Hiroshi Kamada a ,George Coupland b ,Tsuyoshi Mizoguchi a ;ÃaInstitute of Biological Sciences,University of Tsukuba,Tsukuba,Ibaraki 305-8572,Japan bMax Plank Institute for Plant Breeding,Carl von Linne Weg 10,D-50829Cologne,GermanyReceived 13November 2003;revised 8December 2003;accepted 8December 2003First published online 29December 2003Edited by Takashi GojoboriAbstract Photoperiodic control of £owering is regulated by light and a circadian clock.Feedback regulation of the tran-scription of clock components is one of the most common and important mechanisms that control clock functions in animals,fungi,and plants.The Arabidopsis circadian clock is believed to involve two myb-related proteins,LHY (late elongated hypoco-tyl)and CCA1(circadian clock associated 1),which negatively regulate TOC1(timing of cab expression 1)gene expression through direct binding to the TOC1promoter.PIF3(phyto-chrome-interacting factor 3),a bHLH transcription factor binds promoter regions of the LHY and CCA1genes,a¡ecting the light induction of these genes,and interacts with TOC1protein.Although the positive feedback regulation of clock components in plants has been predicted,and PIF3has been assumed to be involved,the molecular nature of this process has not been elu-cidated.Here we demonstrate that the antisense suppression of the PIF3gene causes higher levels of mRNA of £oral activator genes CO (constans )and FT (£owering locus T )and results in early £owering under long days (LD).Neither the circadian rhythms of the clock-controlled genes (CCGs)under constant conditions nor the diurnal rhythms of the CCGs under LD con-ditions are a¡ected by the reduction in PIF3gene expression.These results suggest that PIF3may play an important role in the control of £owering through clock-independent regulation of CO and FT gene expression in Arabidopsis .ß2003Federation of European Biochemical Societies.Pub-lished by Elsevier B.V.All rights reserved.Key words:Circadian clock;Feedback regulation;Transcriptional regulation;Flowering time;Photoperiod;Arabidopsis1.IntroductionThe molecular genetic dissection of £owering time in Ara-bidopsis has identi¢ed several of the clock components,photo-receptors,and light signaling proteins that are involved in the photoperiodic control of £owering time [1,2].Loss-of-function of one of these genes,constans (CO ),causes late £owering under inductive long days (LD)conditions [3],whereas the gain of CO function results in early £owering even under non-inductive short days (SD)conditions [4].CO is a tran-scriptional activator that accelerates £owering time under LDconditions by,at least in part,increasing the expression of the gene £owering locus T (FT )[1,5,6].Changes in the levels of CO and FT mRNAs are responsible for the alterations in £owering time observed in two late £owering mutants,gigan-tea (gi )and late elongated hypocotyl-1(lhy-1),and in three early £owering mutants,early £owering 3(elf3)[5],early £ow-ering 4(elf4)[7],and timing of cab expression 1(toc1)[1,6],all of which have circadian defects [1,2].The regulation of £ower-ing time by day length is thought to require the integration of temporal and environmental light information at the molec-ular level [1,2].It has been proposed that this integration takes place at the level of CO [1,5].Recent models suggest that an external coincidence mechanism,based on the circa-dian control of CO mRNA levels,and the modulation of CO function by light may constitute the molecular basis for the regulation of £owering time by day length in Arabidopsis [1,5,6].Circadian clocks represent widespread endogenous mecha-nisms that allow organisms to time biological processes ap-propriately throughout the day^night cycle.At least three genes are putative central oscillators of a circadian clock in Arabidopsis [1,2].Two of these,LHY and circadian clock as-sociated 1(CCA1),encode closely related transcription factors of the MYB family [2].The third,TOC1,encodes a protein with a sequence related to the receiver domain of two-compo-nent signaling [1].The TOC1gene has also been described as Arabidopsis pseudo-response regulator 1(APRR1)and Arabi-dopsis has four additional homologs (APRR3,5,7and 9)[8].The reciprocal regulation of clock-associated genes is central to the function of all circadian oscillators [9].Molecular bio-chemical analyses and molecular genetics have shown that LHY and CCA1negatively regulate TOC1expression in a direct manner [10,11].TOC1has been assumed to function reciprocally as a positive e¡ector of LHY and CCA1expres-sion;however,because TOC1lacks DNA-binding domain motifs and because there is no experimental evidence that it directly binds to DNA,TOC1may require protein partners in order to regulate LHY and CCA1.Recently two mutations in the Arabidopsis pseudo-response regulator 7(PRR7)gene have been reported [12].The PRR7gene has been shown to be required for the negative regulation of LHY and CCA1in etiolated seedlings in response to light pulses.prr7also showed a clear defect in the sustained circadian expression pattern of LHY and CCA1[12].It has been suggested that the basic helix-loop-helix (bHLH)transcription factor phytochrome-interacting factor0014-5793/03/$30.00ß2003Federation of European Biochemical Societies.Published by Elsevier B.V.All rights reserved.doi:10.1016/S0014-5793(03)01470-4*Corresponding author.Fax:(81)-298-537723.E-mail address:mizoguchi@gene.tsukuba.ac.jp (T.Mizoguchi).FEBS 27998FEBS Letters 557(2004)259^2643(PIF3)acts as a positive element in LHY and CCA1expres-sion,at least in etiolated seedlings in response to light pulses [13].and a protein^protein interaction between TOC1and PIF3has been demonstrated[14].Therefore,PIF3appears to be a good candidate for a positive regulator of LHY and CCA1expression.Although the regulation of hypocotyl elon-gation by PIF3under various light conditions and the expres-sion of light-inducible genes by PIF3in etiolated seedlings have been fully studied,and detailed biochemical analyses of the interactions between PIF3and phytochromes have been conducted[15,16],the possible roles of PIF3in the control of processes such as£owering and circadian rhythmicity have not been reported to our knowledge.Here,we describe the¢rst characterization of the antisense suppression of the PIF3gene with respect to circadian rhythms,£owering times,and the expression of the£oral ac-tivator genes CO and FT in light/dark(L/D)cycles.2.Materials and methods2.1.Plant materials and growth conditionsThe No-0ecotype of Arabidopsis thaliana was used unless otherwise indicated.The PIF3antisense line A22was described previously[15]. Plants used for the reverse transcription-polymerase chain reaction (RT-PCR)and Northern blot analysis were grown on agar plates in controlled environment rooms under LD(16h light/8h dark)con-ditions for12days.For continuous light(LL)experiments,the LD-grown plants were transferred to LL conditions.For the measurement of£owering times,plants were grown on soil under LD(16h light/8h dark)and SD(10h light/14h dark)conditions.2.2.Northern blot analysis and analysis of period lengthRNA(10W g)was separated on1.2%agarose/formaldehyde dena-turing gels and transferred to Biodyne B membranes(Nippon Genet-ics,Tokyo,Japan).Hybridization was done in0.3M sodium phos-phate bu¡er(pH7.0),7%sodium dodecyl sulfate(SDS),1mM ethylenediamine tetraacetic acid(EDTA),and1%bovine serum albu-min overnight at65‡C.The blot was washed with0.2U standard sodium citrate(SSC)and0.1%SDS for30min at65‡C.Probes were full-length LHY,CCA1,TOC1,and GI cDNAs[11].Images were visualized using a BioImaging Analyzer(BAS5000;Fuji Photo Film,Tokyo,Japan);signal intensity was quanti¢ed with Science Lab 98Image Gauge software(version3.1;Fuji Photo Film,Tokyo,Ja-pan).Fourier transforms and period estimates were obtained using the fast Fourier transform-non-linear least squares program(FFT-NLLS) as described[17].The relative amplitude error,RAE,is the value of the amplitude error estimate divided by the value of the most prob-able amplitude estimate.RAE can range from a value of0for an in¢nitely well-determined rhythmic component(zero error)to a value of1,theoretically,for a minimally determined rhythmic component (error in the amplitude equals the amplitude value itself).2.3.RT-PCR analysisRT-PCR was performed with1W g of total RNA using a Super-Script1¢rst-strand synthesis system for RT-PCR(Invitrogen,Carls-bad,CA,USA).CO[5],FT[18],and TUB2[19]primers have been described.The products were separated on1.5%agarose gels and analyzed as described above.2.4.Measurement of£owering timeFlowering time was measured by scoring the number of rosette and cauline leaves on the main stem.Data are presented as meanþS.E.M. Measurement of£owering time was done twice in LD and SD with similar results.All di¡erences in£owering times were con¢rmed as statistically signi¢cant using Student’s t-test(P60.00005).3.Results3.1.Reduction of PIF3gene expression causes early£oweringunder LD conditionsPIF3has been predicted to function as one of the positive regulators of LHY and CCA1gene expression in the Arabi-dopsis circadian clock[13].LHY and CCA1have partially redundant roles in maintaining circadian rhythms.A reduc-tion in the total amount of LHY and CCA1protein by the loss-of-function of either the LHY or CCA1gene causes a short-period phenotype under LL and continuousdark Fig.1.Early£owering phenotype of the PIF3antisense line.No-0(left)and pif3as line(right)plants were grown for3weeks under LD condi-tions(A and B).Means are shownþstandard deviation.Open and¢lled boxes represent the numbers of rosette leaves and cauline leaves,re-spectively.(DD)conditions[11].If PIF3is involved in the positive reg-ulation of LHY and CCA1gene expression and if Arabidopsis does not have genes with functions redundant with those of PIF3,then circadian rhythms should be a¡ected in a pif3as line.Numerous Arabidopsis circadian clock mutants display not only an altered light sensitivity during seedling emergence [1,2],but also a reduction or even an absence of sensitivity to day length[20].The latter phenotypes are often associated with changes in£owering time[21].First we tested pif3as for£owering time phenotypes.We used the pif3as line A22 (Section2)in which a dramatic reduction of PIF3mRNA levels were reported previously.Under SD conditions,the pif3as line had a subtle early£owering phenotype,but under LD conditions,pif3as plants£owered much earlier than did the wild-type in our experiments(Fig.1);Student’s t-test con-¢rmed that these di¡erences were statistically signi¢cant(Sec-tion2).3.2.Antisense suppression of PIF3gene does not a¡ectrhythmicity of expression of clock-controlled genes(CCGs)The loss-of-function of either LHY or CCA1causes a short-period phenotype associated with early£owering under light/ dark cycles,especially under SD conditions[11].Similar cor-relations between period lengths and£owering times have been reported in toc1and zeitlupe(ztl);toc1shortens free running rhythms(FRRs)of CCGs under LL conditions and causes early£owering under SD conditions[22],whereas ztl lengthens FRRs and causes late£owering[23].Therefore,we next tested whether the expression of CCGs was altered in the pif3as line under LL conditions.In wild-type plants,LHYand Fig.2.Antisense suppression of PIF3does not a¡ect the expression of CCGs under LL conditions.Shown are the Northern blot analysis of LHY(A),CCA1(B),TOC1(C),and GI(D)and the abundance of rRNA(E)in No-0(open circles)and pif3as lines(¢lled circles).Plants were entrained under LD(16h light/8h dark)conditions for12days and then released into LL conditions.The analysis is shown from the time24h after transferring to LL.Open and¢lled boxes indicate subjective day and night periods,respectively.Each experiment was per-formed at least twice with similar results.Quanti¢cation was performed with Science Lab98Image Gauge software as described in Section2.CCA1gene expression peaked around subjective dawn at Zeitgeber time (ZT)24,ZT 48,ZT 72,and ZT 96(Fig.2A and B ),as reported previously [11].Surprisingly,the antisense suppression of the PIF3did not a¡ect FRRs or the amplitude of expression of the LHY or CCA1gene expression (Fig.2A and B ).Similar results were obtained for the CCG,GI and TOC1,which normally peak in expression in the evening (Fig.2C and D ).There was no statistical di¡erence in the rhyth-micity of the expression of CCGs between pif3as and wild-type plants (Table 1).To con¢rm the strength of rhythms,the RAE (see Section 2)were calculated (Table 1).All of the rhythms were statistically signi¢cant (RAE 61).3.3.The early £owering phenotype of pif3as correlates with adramatic increase in the FT mRNA level under LD conditionsConsistent with the early £owering phenotype (Fig.1),the expression level of the £oral marker gene FT increased dra-Table 1Free running period estimates of the expression of CCGs in No-0and the PIF3antisense lines under constant white lightExperiment 1Experiment 2Period (h)S.D.RAE Period (h)S.D.RAE LHY WT 22.70þ1.640.5822.77þ1.170.36pif3as 23.07þ0.700.3622.56þ1.880.64CCA1WT 23.01þ1.150.7722.71þ1.380.42pif3as 23.35þ0.880.3223.94þ1.960.48GI WT 22.16þ1.260.5022.25þ0.970.31pif3as 22.59þ0.960.3722.39þ3.430.69TOC1WT 22.87þ1.200.3421.58þ1.480.45pif3as22.73þ1.580.5721.98þ1.420.41The gene expression data in Fig.2(Experiment 1)and an independent experiment (Experiment 2)were Fourier transformed,and period esti-mates were derived using FFT-NLLS [17].The periods are given as the variance-weighted means (period)of the estimate with variance-weighted standard deviations (S.D.).The RAE is shown in Section 2.WT iswild-type.Fig.3.Reduction of PIF3increases mRNA levels of the two £oral regulator genes FT and CO but does not alter those of CCGs,LHY ,CCA1,TOC1and GI under LD conditions.Shown are the RT-PCR analyses of FT (A),CO (B),and TUB2(C)expression and the Northern blot analyses of LHY (D),CCA1(E),GI (F),and TOC1(G)expression and the abundance of rRNA (H)in No-0and pif3as lines under LD conditions.Tissue was harvested from 12-day-old seedlings entrained in a LD cycle.Expression levels were normalized against TUB2(A and B)and rRNA (D^G).The periods of light and dark are indicated as open and ¢lled boxes,respectively.ZT 0is the time point just before lights on.Each experiment was performed at least twice with similar results.Quanti¢cation was performed as described in Fig.2.matically in pif3as under LD conditions(Fig.3A).CO mRNA levels were also signi¢cantly higher in pif3as at all time points (Fig.3B).Under LD conditions,the expression patterns of LHY and CCA1were not a¡ected(Fig.3D and E),and TOC1mRNA levels were similar in pif3as and the wild-type (Fig.3G).These results indicated that the early£owering defect of pif3as was not due to a generalized alteration in the expression of clock-regulated,£owering time genes.In contrast,we noted a substantially higher level of GI expres-sion in pif3as than in the wild-type twice in our triplicate experiments under LD conditions(Fig.3F;Section2),sug-gesting that GI might function in the regulation of FT[24]. Under LL conditions,almost no di¡erence was detected in the maximum levels of GI expression between the wild-type and pif3as(Fig.2C).4.DiscussionRecently,PIL1(PIF3-like1)was identi¢ed as a TOC1/ APRR1-interacting protein[14].The PIL1gene encodes a putative bHLH transcription factor with an amino acid se-quence that is highly similar to that of PIF3[14].It has also been reported recently that,as seen with PIF3(Fig.2and Table1),the loss-of-function of PIL1(pil1-1;T-DNA inser-tion mutant of PIL1)does not a¡ect the rhythmicity of LHY or CCA1gene expression under DD or light/dark cycles[25]. In fact,Arabidopsis has at least¢ve additional genes(PIL2, PIL5,PIL6,PIF4/SRL2,and HFR1)encoding bHLH pro-teins that are highly similar to both PIF3and PIL1,and all PIF3family members,except HFR1,interact with TOC1/ APRR1in yeast two-hybrid analyses[25].Because closely related genes often have redundant functions,as has been demonstrated for the two closely related genes LHY and CCA1,it is possible that the loss-of-function of one redundant gene would not cause a severe phenotype,but would result in only a subtle defect.It may be that the defects in pif3are completely or partially compensated for by PIF3-related genes,even though PIF3might play a vital role in clock functions.PIF4and PIL6have recently been shown to be clock-controlled genes and have been proposed to play impor-tant roles in the control of circadian rhythms[25],supporting this explanation.Alternatively,PIF3might not be involved in maintaining a circadian clock.The antisense suppression of PIF3accelerated£owering (Fig.1)without a¡ecting the expression of CCGs under LD conditions(Fig.2),therefore we investigated the role(s)of PIF3in the regulation of key genes that a¡ect£owering time.FT integrates several£owering time pathways [1,19,26],and its expression is lower in phyA(phytochrome A)and cry2(cryptochrome2)mutants under conditions in which these mutants are late£owering[1].The early£owering phenotype of pif3as was seen under LD conditions(Fig.1) and was not as severe under SD conditions in our experiments (data not shown).Under LD conditions,we observed a large increase in FT mRNA levels in pif3as(10-and20-fold wild-type levels;Fig.3).PIF3interacts with Pfr forms of phyB [16].The similarities in phenotypes in terms of(i)long hypo-cotyls under red light,(ii)a reduction in gene expression in-duced by red light(e.g.CAB and RBCS),(iii)lower chloro-phyll content,and(iv)long petiole length strongly suggest that PIF3plays a role in phyB signaling[13,15].Consistent with this idea,the loss of phyB function accelerates£owering under both LD and SD conditions[27],and signi¢cant in-creases in both CO and FT expression similar to those ob-served in pif3as were seen in phyB under LD conditions(Fig. 3A^C)[27].The loss-of-function of TOC1(toc1-2and TOC1RNAi lines)results in a reduced sensitivity to red light and far-red light in the control of hypocotyl elongation[20];this pheno-type is similar to that of pif3as[15].The short-period muta-tion toc1causes early£owering under SD conditions.This is associated with a phase advance of CO expression,which leads to relatively high levels of CO mRNA during the illumi-nated part of the day at dusk and the upregulation of the £oral activator gene FT[1].Although£owering time is not a¡ected,substantial increases in FT and CO expression in toc1 are seen during the day time under LD conditions[1,6].The increase in FT expression in toc1is less than that in pif3as under LD conditions(approximately2-fold vs.10-to20-fold; Fig.3A)which might explain in part the early£owering in pif3as but not in toc1under LD conditions.We think it is likely that the early£owering of pif3as is caused by the up-regulation of the£oral activator gene FT through a substan-tial increase in CO expression.One alternative possibility is that the increase in FT expression might be caused by a post-translational modi¢cation of the CO protein induced by light signaling[1,5].Light might activate or stabilize the CO pro-tein directly or indirectly to increase FT expression. Although several phenotypes of pif3as and toc1are com-mon,functional cooperation in vivo between PIF3and TOC1 is still unclear.Investigations using pif3/toc1double mutants and multiple mutants of the PIF3family members would pro-vide new insights into the positive feedback loop of the Ara-bidopsis clock and the connections between circadian rhythms and light inputs in the photoperiodic control of£owering. Acknowledgements:The pif3as line(A22)was kindly provided by Dr. Peter Quail.This work was supported in part by a grant from the PROBRAIN(to T.M.)and a Grant-in-Aid for Scienti¢c Research from the Ministry of Education,Science,Sports and Culture of Japan (No.15770021to T.M.).The authors are grateful to Ms.Midori Moro-oka for her technical assistance.References[1]Yanovsky,M.J.and Kay,S.A.(2003)Nat.Rev.Mol.Cell Biol.4,265^275.[2]Hayama,R.and Coupland,G.(2003)Curr.Opin.Plant Biol.6,13^19.[3]Putterill,J.,Robson,F.,Lee,K.,Simon,R.and Coupland,G.(1995)Cell80,847^857.[4]Onouchi,H.,Igen‹o,M.I.,Pe¤rilleux,C.,Graves,K.and Coup-land,G.(2000)Plant Cell12,885^900.[5]Suarez-Lopez,P.,Wheatley,K.,Robson,F.,Onouchi,H.,Val-verde,F.and Coupland,G.(2001)Nature410,1116^1120. [6]Bla¤zquez,M.A.,Tre¤nor,M.and Weigel,D.(2002)Plant Physiol.130,1770^1775.[7]Doyle,M.R.,Davis,S.J.,Bastow,R.M.,McWatters,H.G.,Koz-ma-Bognar,L.,Nagy,F.,Millar,A.J.and Amasino,R.M.(2002) Nature419,74^77.[8]Makino,S.,Kiba,T.,Imamura,A.,Hanaki,N.,Nakamura,A.,Suzuki,T.,Taniguchi,M.,Ueguchi,C.,Sugiyama,T.and Mi-zuno,T.(2000)Plant Cell Physiol.41,791^803.[9]Dunlap,J.C.(1999)Cell96,271^290.[10]Alabad|¤,D.,Oyama,T.,Yanovsky,M.J.,Harmon,F.G.,Ma¤s,P.and Kay,S.A.(2001)Science293,880^883.[11]Mizoguchi,T.,Wheatley,K.,Hanzawa,Y.,Wright,L.,Mizogu-chi,M.,Song,H.R.,Carre¤,I.A.and Coupland,G.(2002)Dev.Cell2,629^641.[12]Kaczorowski,K.A.and Quail,P.H.(2003)Plant Cell15,2654^2665.[13]Mart|¤nez-Garc|¤a,J.F.,Huq,E.and Quail,P.H.(2000)Science288,859^863.[14]Makino,S.,Matsushika,A.,Kojima,M.,Yamashino,T.andMizuno,T.(2002)Plant Cell Physiol.43,58^69.[15]Ni,M.,Tepperman,J.M.and Quail,P.H.(1998)Cell95,657^667.[16]Zhu,Y.X.,Tepperman,J.M.,Fairchild,C.D.and Quail,P.H.(2000)A97,13419^13424.[17]Plautz,J.D.,Straume,M.,Stanewsky,R.,Jamison, C.F.,Brandes, C.,Dowse,H.B.,Hall,J.C.and Kay,S.A.(1997) J.Biol.Rhythms12,204^217.[18]Bla¤zquez,M.A.and Weigel,D.(1999)Plant Physiol.124,1025^1032.[19]Kobayashi,Y.,Kaya,H.,Goto,K.,Iwabuchi,M.and Araki,T.(1999)Science286,1960^1962.[20]Ma¤s,P.,Alabad|¤,D.,Yanovsky,M.J.,Oyama,T.and Kay,S.A.(2003)Plant Cell15,223^236.[21]Mizoguchi,T.and Coupland,G.(2000)Trends Plant Sci.5,409^411.[22]Somers,D.E.,Webb,A.A.,Pearson,M.and Kay,S.A.(1998)Development125,485^494.[23]Somers, D.E.,Schultz,T.F.,Milnamow,M.and Kay,S.A.(2000)Cell101,319^329.[24]Hayama,R.,Yokoi,S.,Tamaki,S.,Yano,M.and Shimamoto,K.(2003)Nature422,719^722.[25]Yamashino,T.,Matsushika,A.,Fujimori,T.,Sato,S.,Kato,T.,Tabata,S.and Mizuno,T.(2003)Plant Cell Physiol.44,619^ 629.[26]Kardailsky,I.,Shukla,V.K.,Ahn,J.H.,Dagenais,N.,Christen-sen,S.K.,Nguyen,J.T.,Chory,J.,Harrison,M.J.and Weigel,D.(1999)Science286,1962^1965.[27]Cerda¤n,P.D.and Chory,J.(2003)Nature423,881^885.。