ORIGINAL RESEARCH published: 04 July 2016 doi: 10.3389/fpls.2016.00962 Genomic Analysis Reveals Contrasting PIFq Contribution to Diurnal Rhythmic Gene Expression in PIF-Induced and -Repressed Genes Guiomar Martín, Judit Soy and Elena Monte * Center for Research in Agricultural Genomics, CSIC-IRTA-UAB-UB, Barcelona, Spain Members of the PIF quartet (PIFq; PIF1, PIF3, PIF4, and PIF5) collectively contribute to induce growth in Arabidopsis seedlings under short day (SD) conditions, specifically promoting elongation at dawn. Their action involves the direct regulation of growth- related and hormone-associated genes. However, a comprehensive definition of the PIFq-regulated transcriptome under SD is still lacking. We have recently shown that SD and free-running (LL) conditions correspond to “growth” and “no growth” conditions, respectively, correlating with greater abundance of PIF protein in SD. Here, we present a genomic analysis whereby we first define SD-regulated genes at dawn compared Edited by: to LL in the wild type, followed by identification of those SD-regulated genes whose Rob McClung, expression depends on the presence of PIFq. By using this sequential strategy, we have Dartmouth College, USA identified 349 PIF/SD-regulated genes, approximately 55% induced and 42% repressed Reviewed by: Biswapriya Biswavas Misra, by both SD and PIFq. Comparison with available databases indicates that PIF/SD- University of Florida, USA induced and PIF/SD-repressed sets are differently phased at dawn and mid-morning, Zhaojun Ding, Shandong University, China respectively. In addition, we found that whereas rhythmicity of the PIF/SD-induced gene Andreas Hiltbrunner, set is lost in LL, most PIF/SD-repressed genes keep their rhythmicity in LL, suggesting University of Freiburg, Germany differential regulation of both gene sets by the circadian clock. Moreover, we also *Correspondence: uncovered distinct overrepresented functions in the induced and repressed gene sets, Elena Monte elena.monte@cragenomica.es in accord with previous studies in other examined PIF-regulated processes. Interestingly, promoter analyses showed that, whereas PIF/SD-induced genes are enriched in direct Specialty section: PIF targets, PIF/SD-repressed genes are mostly indirectly regulated by the PIFs and This article was submitted to Plant Cell Biology, might be more enriched in ABA-regulated genes. a section of the journal Keywords: phytochrome-interacting factors PIFs, PIF-regulated transcriptional network, diurnal growth, Frontiers in Plant Science Arabidopsis, short day, circadian clock Received: 24 April 2016 Accepted: 15 June 2016 Published: 04 July 2016 INTRODUCTION Citation: Martín G, Soy J and Monte E (2016) Environmental light conditions change within each day and season, and plants have adapted to Genomic Analysis Reveals these oscillations in solar energy (and the related changes in temperature and water availability) Contrasting PIFq Contribution by timing their physiological responses to specific times of the day and/or the year. Of particular to Diurnal Rhythmic Gene Expression in PIF-Induced and -Repressed importance in this respect is the diurnal photoperiod (or more accurately, the duration of the Genes. Front. Plant Sci. 7:962. daily dark period), primarily sensed by the phytochrome (phy) family of photoreceptors (phyA doi: 10.3389/fpls.2016.00962 to phyE in Arabidopsis; Rockwell et al., 2006; Galvao and Fankhauser, 2015). An important aspect Frontiers in Plant Science | www.frontiersin.org 1 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression of phy function resides in the ability to bind in its photoactivated et al., 2008a; Nozue et al., 2011). Growth-related genes regulated state to the PIF (Phy-Interacting Factor) family of basic helix- by PIF4 and PIF5 under photoperiodic conditions were defined loop-helix (bHLH) transcriptional regulators (Ni et al., 1999; Bae by Nozue et al. (2011). Many of these genes are direct targets of and Choi, 2008; Leivar and Monte, 2014; Xu et al., 2015). This both PIF and TOC1 and their expression is limited to the pre- phy-PIF interaction induces the rapid degradation of the PIF dawn hours (Soy et al., 2016). However, a comprehensive analysis quartet (PIFq) PIF1, PIF3, PIF4, and PIF5 (Bauer et al., 2004; Al- of the PIFq-regulated transcriptome under SD conditions is still Sady et al., 2006; Lorrain et al., 2008; Shen et al., 2008; Ni et al., lacking. 2013), altering the transcription of the target gene network within Here, using microarray analysis, we identify genes specifically minutes (Jiao et al., 2007; Pfeiffer et al., 2014). regulated at the end of the night during the growth conditions After seed germination and emergence into sunlight, of SD compared to LL, where growth is marginal, and define Arabidopsis seedlings are subjected to alternating light/dark their dependence on PIFq for regulation. Our analysis provides cycles. Hypocotyl elongation under these conditions depends a comprehensive description of the SD- and PIFq-regulated on the duration of the dark period in a non-linear fashion, transcriptome at dawn, which identifies ∼60% and ∼40% and it is accelerated specifically in the long nights of short-day of genes that are PIF-induced and –repressed, respectively. (SD) photoperiods (Niwa et al., 2009). In SD, seedlings display Additionally, we show that the pattern of rhythmic expression diurnal rhythmic growth with maximal growth rates at dawn, of the PIF/SD-gene set and the contribution of PIFq is which are rapidly reduced during the first hours of light (Nozue markedly different in PIF/SD-induced and PIF/SD-repressed et al., 2007). In comparison, growth is greatly reduced under genes. Finally, we describe a functional and promoter sequence SD-entrained seedlings released into constant light (LL; Soy dichotomy between the PIF/SD-induced and -repressed gene et al., 2014), which indicates that the dark period is necessary sets. for the induction of elongation. The promotion of growth at the end of the night involves the combined actions of PIF1, PIF3, PIF4, and PIF5 (Nozue et al., 2007; Niwa et al., 2009; Soy MATERIALS AND METHODS et al., 2012, 2014). Precise regulation of their accumulation and time of action under diurnal conditions has been proposed to Plant Materials and Growth Conditions involve different parallel mechanisms. First, at the transcriptional Three biological replicates of Arabidopsis thaliana wild type (WT) level, PIF4 and PIF5 genes are directly targeted by several clock Col-0 and pifq mutant seeds (Leivar et al., 2008) were plated on components that impose an internal rhythm of expression with GM medium without sucrose at room temperature as described increased transcript levels at the end of the night (Nozue et al., (Monte et al., 2003). Seedlings were then stratified for 4 days at 2007; Nusinow et al., 2011; Huang et al., 2012). In contrast, 4◦ C in darkness, and then placed in SD conditions [8 h white PIF1 and PIF3 transcript levels are maintained fairly constant light (85 µmol m−2 s−1 ) + 16 h dark] at the beginning of the during the diurnal cycle (Soy et al., 2012, 2014). Second, at the light period, considered as day 1. On day 3, seedlings were either post-transcriptional level, phy-induced degradation imposes kept under SD or transferred to continuous light (LL; Soy et al., oscillation of PIF3 and likely PIF1 proteins to peak at dawn (Soy 2014). Samples were harvested at the end of the dark period et al., 2012, 2014), and fine-tunes the timing of PIF4 and PIF5 (ZT24, SD samples) or the subjective night (CT24, LL samples; accumulation (Nozue et al., 2007; Yamashino et al., 2013). In Figure 1A). addition, PIF4 has been proposed to be a target of the kinase BIN2 to mark it for proteasome regulation (Bernardo-García Microarray-Based Expression Profiling: et al., 2014). Third, DELLAs and ELF3 have been shown to Sample Preparation and Data Analysis interfere with DNA binding of the PIFs (de Lucas et al., 2008; Total RNA was prepared using the RNeasy Plant Mini Kit Feng et al., 2008; Arana et al., 2011; Nieto et al., 2015). Finally, (Qiagen). cRNA synthesis, and microarray hybridizations and we have recently shown that the transcriptional activity of washes were performed as described by Affymetrix in the PIF3 (and possibly of the other PIFq members) is inhibited Genomics Facility at CRAG. Arabidopsis Affymetrix Gene 1.0 ST by the core component of the circadian clock TOC1 to gate arrays were used for gene expression detection. hypocotyl elongation to the end of the night (Soy et al., Data analysis was performed using the Rosetta Resolver Gene 2016). Expression Analysis System, version 7.0 (Rosetta Biosoftware) Previous work has established that PIFq stimulation as previously described (Sentandreu et al., 2011). Briefly, two of hypocotyl elongation under SD involves the direct separate statistical analyses were carried out to define (a) a regulation of growth-related genes such as PIL1 “SD-regulated” gene list of transcripts whose expression is (PHYTOCHROME-INTERACTING FACTOR-3 LIKE significantly altered by in the WT upon exposure to long 1), HFR1 (LONG HYPOCOTYL IN FAR-RED 1), XTR7 night (SD) compared to subjective night (LL); and (b) a (XYLOGLUCAN ENDOTRANSGLYCOSYLASE 7), GA2OX6, “PIFq-regulated” gene list of transcripts whose expression is and PAR1 (Soy et al., 2012, 2016), which are up-regulated in significantly altered in pifq compared to WT under SD. These conditions where hypocotyl elongation is induced (Salter et al., gene lists were calculated by performing a two-group, two-way, 2003; Lorrain et al., 2008; Hornitschek et al., 2009; Leivar et al., error-weighted, Benjamini-Hochberg false discovery rate error- 2009; Nozue et al., 2011), and the regulation of auxin-related corrected analysis of variance with a p-value cutoff of 0.05. genes that oscillate in phase with hypocotyl growth (Michael This statistical significance test was combined with experimental Frontiers in Plant Science | www.frontiersin.org 2 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression FIGURE 1 | Overlap between SD and PIF-regulated genes at dawn defines 349 PIF/SD-regulated genes. (A) Schematic representation of seedling growth used for microarray-based transcriptome analysis. Samples of wild type (WT) and pifq seedlings grown for 3 days under short days (3dSD), or 2 days in SD plus 1 day in free running conditions (2dSD + 1dLL) were harvested at the end of the night at ZT24 or CT24, respectively. (B) Venn diagram shows pairwise comparison between SS1.5F differentially expressed genes comparing WT SD and pifq SD (PIF-regulated genes), and WT SD and WT LL (SD-regulated genes). The number of genes that are differentially expressed in each set is indicated. The list of genes in each class is provided in Dataset S1. SS1.5F: genes whose expression changed statistically significantly and by at least 1.5 fold. (C) Scatterplot of log2 fold change (FC) values shows for each gene a quantitative measure of the correlation in responsiveness between the PIF- and SD-regulation described in B: Black dots represent genes that are shared between the two combinations, whereas orange (SD-regulated) and purple (PIF-regulated) dots represent genes that are specifically present in one of the combinations but not in the other (p-value = 2.2e-16, one-sided binomial test between quadrants). consistency by further reducing the statistically significantly (SS) Statistics of Gene Expression Analysis transcript list to only those transcripts exhibiting an absolute FC Binomial test was performed in Figure 1C to asses the of greater than 1.5-fold (SS1.5F genes). PIF/SD-regulated genes association between the PIFq and SD regulation of PIF/SD- were defined as SS1.5F genes in both SD- and PIFq-regulated regulated genes. Differential gene expression shown in Figure 2A gene lists. and Supplementary Figures S2 and S6 was determined by Willcoxon test. SS differences are indicated (∗ p-value < 0.05; ∗∗ p-value < 0.01; ∗∗∗ p-value < 0.001). Promoter Analysis for DNA Binding Motifs Promoter analysis was performed within 3 kb upstream of the Functional Categorization translation initiation site using the Patmach and Motif analysis PIF/SD-regulated genes were functionally categorized according tools available at the TAIR website1 . For enrichment analysis to their established or predicted subcellular localization. performed for each of the investigated motifs, the total number of A single subcellular localization was assigned to each locus genes on the Arabidopsis Affymetrix Gene 1.0 ST array containing and designations were determined using the annotation of at least one motif of interest, was compared with the total the Arabidopsis genome (TAIR10). To identify enriched gene number of genes within our PIF/SD-regulated subset of genes ontology (GO) biological processes, analyses were performed that contain the motif. SS (p-value ≤ 0.05) enrichment of these using the functional annotation classification system DAVID3 motifs in the various gene subsets was then calculated from the (Huang et al., 2007). hypergeometric distribution. Promoter analysis using the SCOPE motif finder2 was performed within 2.5 kb upstream of the Data Analyses translation initiation site. Phases of expression were analyzed using the publicly available gene phase analysis tool PHASER4 (Mockler et al., 2007) using 1 3 http://www.Arabidopsis.org/ http://david.abcc.ncifcrf.gov 2 4 http://genie.dartmouth.edu/scope/ http://phaser.mocklerlab.org Frontiers in Plant Science | www.frontiersin.org 3 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression FIGURE 2 | Contribution of PIFq to the expression of PIF/SD-regulated genes. (A) Vioplot of the log2 expression values of the 191 PIF/SD-induced (top) and the 145 PIF/SD-repressed genes (bottom) in the WT and pifq SD and LL microarray samples. Statistically significantly (SS) differences from WT SD or pifq SD by Willcoxon test are indicated in the upper and lower part, respectively (∗ p-value < 0.05; ∗∗ p-value < 0.01; ∗∗∗ p-value < 0.001. n.s., non-significant). (B) Bar graph shows the mean ± SE of log2 FC values for PIF/SD-induced (top) and repressed (bottom) genes relative to the expression value of WT LL sample. The mean FC value for each sample is indicated. (C) Bar graph of microarray data showing the FC in gene expression relative to the WT SD. Relative expression of two representative PIF/SD-induced genes (PIL1 and IAA19; top) and repressed genes (PSY and CYP97A3; bottom). Data correspond to biological triplicates, and bars indicate SE. a cutoff of 0.7. Transcript abundance of diurnal and circadian PIF abundance is greatest (Soy et al., 2012, 2014; Yamashino et al., photocycles was analyzed using the publicly available genome- 2013) and at subjective dawn in LL (CT24; Figure 1A). wide expression data in DIURNAL5 (Mockler et al., 2007) for the Microarray analysis shows that the expression of 2,184 following conditions: SD (Col-0_SD), 12:12 (Col_LDHH), and genes changes statistically significantly (SS) and by 1.5 fold Free Running (LL23_LDHH). (SS1.5F) during growth in SD relative to LL (WT-SD vs. WT- LL; Figure 1B and Supplementary Figure S1; see Dataset S1 online for the gene list). By comparison, pifq seedlings in SD RESULTS display SS1.5F alterations in the expression of 538 genes relative to the SD-grown wild type (WT-SD vs. pifq-SD; Figure 1B Night-Induced Transcriptomic Changes and Supplementary Figure S1; see Dataset S1 online for the under SD Underlying PIFq-Regulated gene list). Of these 538 genes, 349 (64.8%) are identical to those altered in expression in the WT by exposure to the long Growth night of SD (Figure 1B and Supplementary Figure S1). Of To delineate the transcriptomic changes underlying PIFq these common 349 SS1.5F genes, 191 (54.7%) are SD- and PIF- induction of growth under short days (SDs), we performed induced (i.e., are up-regulated in the WT in SD compared to microarray analysis to define gene expression in 3 day-old LL, and down-regulated in pifq compared to the WT in SD), WT and pifq seedlings grown under SD for 2 days and then and 145 (41.5%) are SD- and PIF-repressed (i.e., are down- kept in SD (3dSD) or released into continuous white light regulated in the WT in SD compared to LL, and up-regulated (LL) for an additional day (2dSD+1dLL; Figure 1A). This in pifq compared to the WT in SD; Supplementary Figure experimental configuration was designed based on our previous S1), and were designated PIF/SD-induced and PIF/SD-repressed, results showing that the hypocotyl of seedlings in SD elongate respectively (Dataset S1). The remaining 13 genes (0.03%) during the night period, whereas seedlings transferred to LL displayed opposite direction of their response between night and do not grow during the subjective night (Soy et al., 2014). PIFq, and were designated as ambiguous (Supplementary Figure Thus, SD and LL correspond to “night-induced growth” and “no S1, Dataset S1). The scatterplot shown in Figure 1C illustrates night-induced growth” conditions, respectively, correlating with the significant strong qualitative and quantitative correlation presence (SD) or absence (LL) of PIF accumulation (Soy et al., between the expression elicited by the long night and by PIFq 2012, 2014). Samples were collected at dawn in SD (ZT24) when in SD compared with LL for the 349 regulated genes (black 5 ftp://diurnal.mocklerlab.org symbols; p-value = 2.2e-16). In addition, this plot reveals that Frontiers in Plant Science | www.frontiersin.org 4 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression genes that are SD-regulated only (orange symbols) responded we ranked the percent-contribution values for the induced and similarly or with only slight differences in expression in the pifq repressed genes separately (Supplementary Figure S3). All of compared to the WT (Figure 1C and Supplementary Figure the PIF/SD-induced genes showed a contribution larger than S2A), indicating that they are only marginally regulated by 50%, with 64% of the genes having a contribution that was the PIFs. However, a considerable additional number of the 100% or greater. In contrast, of the PIF/SD-repressed gene SS1.5F genes that respond to SD in the pifq mutant only (PIF- set, 35% of the genes contributed less than 50%, and only regulated only, purple symbols) also respond in the WT in SD 14% of the genes had a contribution that was 100% or higher compared to LL, albeit to a lesser extent compared to the fold (Supplementary Figure S3). Together, the data suggest that change (FC) difference between WT and pifq (Figure 1C and gene induction and repression under SD might be differently Supplementary Figure S2B). Together, these data indicate that regulated by the PIFs: whereas the PIFq appears to be necessary there are extensive similarities in the direction and extent of the and sufficient to explain the night induction of PIF/SD-induced gene expression changes induced by genetic removal of the PIFq genes, other proteins might be involved in the night repression proteins in SD and the changes elicited by suppression of the of PIF/SD-repressed genes. As examples of the contribution night period in SD-entrained seedlings released in continuous of PIFq for specific PIF/SD-regulated genes, Figure 2C shows light, and suggest that the reciprocal regulation by light and the expression of representative PIF/SD-induced genes PIL1 PIFs previously described to take place during deetiolation (AT2G46970) and IAA19 (AT3G15540), and PIF/SD-repressed (Leivar et al., 2009) is also reproduced under SD conditions. The genes PSY (AT5G17230) and CYP97A3 (AT1G31800), in WT and degree of this reciprocal regulation was most significant for the pifq seedlings grown under SD or LL. identified PIF/SD-regulated subset of 349 genes (Dataset S1). These PIF/SD-regulated genes were therefore considered as good Phase Enrichment and Diurnal candidates to underlie the PIFq-regulated changes in seedling Expression of PIF/SD-Regulated Genes growth and development under SD conditions, and were selected Time-of-day-expression enrichment-analysis of the PIF/SD- for further analysis. regulated genes was performed using the available data at the PHASER website6 (Mockler et al., 2007). The data showed that, under SD photocycles, PIF/SD-induced genes displayed PIFq Contribution to the Expression of an overrepresented phase of expression at the end of the dark PIF/SD-Regulated Genes period (Figure 3A), with 49% of these genes phased at pre- To estimate the contribution of the PIFq proteins to the night- dawn between 18 and 23 h (Dataset S1), when PIF abundance induced expression of the PIF/SD-regulated genes, we calculated is maximum. By contrast, the phase-overrepresentation pattern the average contribution of PIFq to the night-induced changes was remarkably different for the PIF/SD-repressed genes, which in expression elicited in the WT in SD compared to LL. The displayed an overrepresented phase of expression in the morning expression levels of the 191 PIF/SD-induced gene-set in pifq (Figure 3A), with 26% of these genes phased between 2 and seedlings grown in SD resembled those in WT in LL (and also 3 h (Dataset S1). Consistent with this phase enrichment, in pifq in LL), indicating that PIFq proteins are necessary to the expression of PIF/SD-regulated genes under SD shows an induce their gene expression during the night period under SD oscillatory pattern (Figure 3B, top). The expression of PIF/SD- (Figure 2A). Examination of the contribution of PIFq to the induced gene set is kept low during the day, progressively expression of the PIF/SD-induced genes shows that, compared increases during the night and displays a maximum at the end to the normalized log2 expression of genes in WT LL conditions of the dark period (Figure 3B, top). In contrast, the PIF/SD- set at 0, the mean log2 of the FC of WT SD is 1.2, whereas in repressed gene set presents a similar increase during the night pifq SD and LL is −0.13 and −0.14, respectively (Figure 2B). but the peak of maximum expression occurs during the day Similarly, for the 145 PIF/SD-repressed gene set, expression (Figure 3B, top). Interestingly, these oscillatory patterns are not levels in pifq seedlings grown in SD were more similar to maintained under free-running conditions (Figure 3B, bottom). those in LL-grown WT (and also pifq), than to SD-grown Indeed, expression for PIF/SD-induced genes is kept at low levels WT, although there were significant differences among them across the day and the subjective night under LL (Figure 3B, (Figure 2A). In accord, the mean log2 FC of WT SD was (−1.3) bottom left). Strikingly, for PIF/SD-repressed genes, whereas, the compared to the WT LL, whereas the mean FC in pifq SD expression in LL displays a comparable night increase compared was (−0.4; Figure 2B). The mean log2 of the FC expression in to SD, there is a significant phase shift from mid morning pifq LL was 0.05, similar to WT LL (Figure 2B). When these to the end of night (Figure 3B, bottom right), also apparent differences between WT and pifq are expressed as percentage when compared to the corresponding entrainment conditions contribution of PIFq to the change induced in the WT in SD in 12:12 that also display a mid-morning peak (Supplementary compared to LL, these results indicate that, on average, PIFq Figure S4). Accordingly, whereas most PIF/SD-regulated genes contributes to 100% of the upregulation of PIF/SD-induced are rhythmic under SD conditions, the percentage of rhythmic genes, whereas their contribution to the repression of gene genes under free running is reduced to 60% in the PIF/SD- expression in WT SD compared to WT LL is ∼69%, suggesting induced gene set, and to 90% in the PIF/SD-repressed gene set that other factors must be responsible for the remaining 30% (Figure 3C). Together, these data suggest that PIF accumulation change in expression. To visualize the array of the contribution of PIFq to the expression of the individual genes under SD, 6 http://phaser.mocklerlab.org Frontiers in Plant Science | www.frontiersin.org 5 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression FIGURE 3 | Expression dynamics of PIF/SD-induced and -repressed genes under SD compared to LL photoperiods. (A) Expression phases in SD of the 191 PIF/SD-induced (left) and the 145 PIF/SD-repressed (right) genes. Phases, defined by PHASER (http://phaser.mocklerlab.org), are indicated on the circumference, and FC phase enrichment of genes (count/expected) on the radius. Day: yellow; Night: gray. (B) Median values (thick line) and upper and lower quartiles (thin lines) of log2 expression for all PIF/SD-induced (left) and PIF/SD-repressed (right) under SD (top) compared to free running conditions in seedlings entrained in 12:12 (bottom). See Supplementary Figure S4 for values in 12:12. Data were obtained from http://diurnal.mocklerlab.org. Day: yellow; Night (top): gray; Subjective night (bottom): light gray (C) Percentage of rhythmic genes in SD or LL in the PIF/SD-induced (left) and the PIF/SD-repressed (right) gene sets. Data are from http://phaser.mocklerlab.org. Frontiers in Plant Science | www.frontiersin.org 6 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression during the night underlies the phase and amplitude of the described conditions where PIFs are regulating development: rhythmicity of PIF/SD-induced and -repressed genes, to peak at shade (Leivar et al., 2012), deetiolation (Leivar et al., 2009), pre-dawn and in the morning, respectively. and photoperiod (Nozue et al., 2011). Next, we compared it to the subset of PIF/SD-regulated genes described here filtered for Comparison of PIFq-Regulated Genes in presence in the ATH1 microarray platform used in the three abovementioned genomic analyses (166 PIF/SD-induced and SD with PIF-Regulated Genes During 133 PIF/SD-repressed genes). We found that the percentage of Deetiolation, Shade, or Photoperiodic overlap was similar for PIF/SD-induced and PIF/SD-repressed, Growth with 54 and 53% of the genes having been previously described Previously, we defined a common set of PIF-regulated “central- as PIF-induced or PIF-repressed, respectively (Figure 4A). Of the class” genes that were associated to the light/PIF-regulated 22 PIF-induced “central class” genes (Leivar and Monte, 2014), 16 responses of deetiolation, shade, and diurnal conditions (Leivar were PIF-induced in all experiments including SD (Dataset S1). and Monte, 2014). These genes might represent common AT3G15850, the only “central class” PIF-repressed gene (Leivar effectors in different light environments to implement PIF- and Monte, 2014), was also a PIF/SD-repressed gene. regulated seedling growth and development. To establish whether To further investigate the light and PIF-mediated regulation of our PIF/SD-regulated set of genes might also be in common genes not defined as common in all classes, we next examined the with genes associated to other PIF-regulated developmental behavior of the subsets of genes that (1) were previously defined processes, we decided to examine how this gene set compares as PIF-regulated genes but were not in common with our defined to the previously defined PIF-regulated gene sets. To this PIF/SD-gene set under the diurnal SD conditions assayed here end, we first compiled a list of PIF-induced (591) and PIF- (Figure 4A), and (2) were defined here as PIF/SD-genes but were repressed (711) genes defined in at least one of previously not previously identified as PIF-regulated genes (Supplementary FIGURE 4 | Comparison of PIF-regulated gene expression in SD with the expression of previously defined PIF-regulated gene sets during deetiolation, shade, and photoperiodic conditions. (A) Venn diagram shows the overlap between the PIF/SD-induced (left) or PIF/SD-repressed (right) genes, and the composite of 591 PIF-induced and 711 PIF-repressed genes previously described during deetiolation (Leivar et al., 2009), photoperiod (Nozue et al., 2011), and shade (Leivar et al., 2012). (B) Bar graph of WT and pifq microarray data (SD and LL) showing the log2 FC expression relative to the WT LL of the PIF-induced (green) or PIF-repressed (yellow) genes defined during deetiolation (left), photoperiod (middle) and shade (right), that are not PIF/SD-regulated in our experimental conditions. Frontiers in Plant Science | www.frontiersin.org 7 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression Figure S5). For the previously defined PIF-regulated genes not assigned subcellular localization, the data show that genes coding in common with our defined PIF/SD-gene set (501 PIF-induced proteins that localize to the nucleus are the most abundant (36%; and 641 PIF-repressed genes; Figure 4A), the data show that Figure 5A). This result is similar to the previously reported the average FC expression of these genes followed a pattern abundance of transcription factors among PIF-regulated and similar to the PIF/SD-regulated genes (Figure 4B, compare with PIF-target genes during deetiolation and shade (Monte et al., Figure 2B). Indeed, the PIF-induced gene sets in shade (117 2004; Leivar et al., 2009; Hornitschek et al., 2012; Zhang et al., genes; Leivar et al., 2012), photoperiod conditions (43 genes; 2013; Pfeiffer et al., 2014) and it is consistent with the PIFs Nozue et al., 2011), and during deetiolation (388 genes; Leivar acting as master regulators of a transcriptional network in the et al., 2009) that were not in our PIF/SD-regulated gene set, adaptation of seedling morphology to alternating day and night exhibited induced expression under SD compared to WT LL, an conditions. Genes in this category include the transcriptional induction that was either absent (deetiolation and photoperiod) regulators PIL1, FHL, HFR1, and HAT2 (described to be PIF- or much reduced (shade) in pifq SD (Figure 4B). Similarly, induced specifically at the end of the night under SD; Nomoto the PIF-repressed gene sets in shade (82 genes; Leivar et al., et al., 2012; Soy et al., 2016). Other well-represented subcellular 2012), in photoperiod conditions (32 genes; Nozue et al., 2011), categories are cell membrane (14%) and extracellular (16%) and during deetiolation (567 genes; Leivar et al., 2009) that related genes (Figure 5A). Among the overrepresented biological were not in our PIF/SD-regulated gene set, exhibited repressed processes, PIF/SD-induced genes are significantly enriched in expression under SD compared to WT LL, a repression that was genes involved in responses to light and hormone stimuli, partially reduced in pifq SD (Figure 4B). These data indicate transcription, and cell growth processes (Figure 5B). Among the that previously defined PIF-regulated genes that are not in our response to hormone stimulus, auxin-related genes are dominant PIF/SD-regulated gene set were not selected here because they (namely we found four SAUR genes and the IAA genes IAA2, did not meet the SS1.5 FC requirement in the comparisons of IAA19, and IAA29) and thus “response to auxin stimulus” is WT SD vs. WT LL and/or pifq SD vs. WT SD (Figure 4B). strongly enriched (Figure 5B). There is also enrichment in However, these PIF-regulated genes follow a similar regulatory the response to gibberellins (GAs) and brassinosteroid (BR) pattern under SD compared to the selected PIF/SD-regulated stimuli. The cell growth category includes genes that are gene set (Figure 2B), indicating that they are also regulated in involved in cell wall loosening and modification such as the the long nights of SD by the PIFs, albeit to a lesser degree. XYLOGLUCAN ENDOTRANSGLUCOSYLASES (XTH) XTH8, Notably, PIF-regulated genes during deetiolation have the lowest XTH15, XTH30, and XTH33, the CELLULOSE-SYNTHASE- average FC (WT SD compared to WT LL) under SD conditions LIKE C4 (CSLC04), the BARWIN-LIKE ENDOGLUCANASES (Figure 4B). This result suggests that the response of dark-grown SUPERFAMILY PROTEIN EXP3, the EXPANSIN11 (EXPA11), seedlings to light might be stronger compared to the response and the EXPANSIN-LIKE A2 (EXLA2). This analysis, together of SD-grown seedlings to the long nights of SD compared with previous studies (Michael et al., 2008a; Leivar et al., 2009; to LL, or the response of light-grown seedlings to shade. For Nozue et al., 2011; Hornitschek et al., 2012; Pfeiffer et al., the second set of genes that was selected here as PIF/SD- 2014) suggests that PIFq might induce elongation during the regulated but had not been described before as PIF-regulated long nights of SD by upregulating the expression of hormonal during deetiolation, shade or photoperiod (76 PIF-induced and pathways such as auxin, GA and BR, and genes involved in cell 63 PIF-repressed genes; Figure 4A), we asked whether it might wall modification to support cell growth. This PIFq-mediated represent genes with a specific role under SD conditions, or induction of growth-related gene expression is suppressed in the whether they are genes that were also regulated in the previous morning upon light exposure and consequent reduction in PIF deetiolation, photoperiod, or shade studies, but did not meet protein levels (Figure 3; Nozue et al., 2007; Soy et al., 2012, 2014). the FC or SS requirements applied. Supplementary Figure S5 Similarly, under LL, PIFs do not accumulate (Soy et al., 2014) and shows that expression of both induced and repressed PIF/SD- this correlates with reduced expression of growth-related genes regulated genes followed a similar PIF-dependent expression and growth arrest under the subjective night of LL compared to pattern during deetiolation, photoperiod or shade, albeit with the long night of SD (Figure 3). a FC that was reduced compared to SD (Figure 2B). We Of the PIF/SD-repressed genes, the data show that genes conclude that, to a great extent, the genetic network imposed coding for proteins predicted or established to localize to by the PIFs under different conditions involves the same genes. the chloroplast are the most abundant (52%), followed by However, differences in the expression FC comparing dark or the genes assigned to code for proteins localized to the shade with light, and/or the contribution of each gene to the nucleus (16%; Figure 5A). Other prominent categories are the overall morphological, cellular, and subcellular phenotype might mitochondrion (11%) and extracellular (7%; Figure 5A). Among be specific under each condition. the overrepresented biological processes, PIF/SD-repressed genes are significantly enriched in genes involved in photosynthesis and in response to light and radiation stimuli (Figure 5B). The Functional Categorization of photosynthesis/chloroplast categories include genes involved in PIF/SD-Regulated Genes photosynthesis (e.g., the photosystem I PGR5-LIKE A, and the The PIF/SD-regulated genes were classified according to their photosystem II PsbP and PSB28), genes coding for structural established or predicted subcellular localization and GO terms components of the chloroplast [e.g., acyl carrier protein 4 (Figure 5; Dataset S1). Of the PIF/SD-induced genes with an (ACP4)], genes necessary for chlorophyll biosynthesis (CHLI1, Frontiers in Plant Science | www.frontiersin.org 8 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression FIGURE 5 | Subcellular localization and functional analysis of the PIF/SD-regulated genes. (A) Subcellular localization of PIF/SD-regulated genes based on the gene ontology (GO) annotations available at TAIR (http://www.Arabidopsis.org). Each class is represented as percentage of the total annotated genes excluding the genes annotated as having unknown localization. The list of genes belonging to each class is provided in Dataset S1. (B) Enriched GO categories (biological process) of PIF/SD-induced (top) and -repressed (bottom) gene sets. DAVID p-value indicates significance (Fisher’s exact test; p-value < 0.05). Frontiers in Plant Science | www.frontiersin.org 9 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression HEMA1), and genes involve and carbon fixation, such as the small results suggest that a high % of the PIF/SD-regulated genes might chain of the ribulose bisphosphate carboxylase. Interestingly, it be direct targets of the PIFs. Indeed, comparison with previously also includes the master regulators of chloroplast development defined PIF-bound genes in deetiolation and shade (Hornitschek GLK2 and SIGMA5 (Fitter et al., 2002; Waters et al., 2009; et al., 2012; Pfeiffer et al., 2014) showed that 57.5% of the PIF/SD- Noordally et al., 2013). It is worth noting that the expression induced genes defined here were identified as PIF-bound under of these genes, even though they are PIF-repressed genes in SD these conditions (110 genes; Figure 6A). Based on our previous conditions compared to LL, is actually progressively induced data showing that PIF-bound genes in etiolated and shade-grown during the long night of SD as shown above (Figure 3B, seedlings can also be bound by the PIFs under SD (Soy et al., top). Together, these results suggest that PIFq prevents over- 2012, 2016), we propose that these SD/PIF- induced genes that expression of chloroplast-related genes during the night. Once are PIF-bound are likely direct targets of the PIFs under diurnal the seedling is exposed to light, PIFs are degraded, the repression conditions. is lifted and these genes reach maximum expression during the Interestingly, these genes were also enriched (p-value: 7.75E– morning hours (Figure 3B, top). 11) in the extended G-box/PBE element (cacrtggg, where r can be a “g” or “a”; Figure 6A and Table 1) that we have described Promoter Analysis of PIF/SD-Regulated recently (Soy et al., 2016). This motif was found to be enriched Genes in the flanking regions of the PIF binding sites in genes bound PIF proteins have been experimentally shown to bind sequence by both PIFs and TOC1 with an expression phase at the end specifically to the G-box motif CACGTG and the PIF-binding of the night (Soy et al., 2016). This result suggests that this element (PBE)/Hormone-up at dawn box (HUD) CACATG PIF/SD-induced gene set might be enriched in genes that are (Hornitschek et al., 2012; Oh et al., 2012; Zhang et al., 2013; directly targeted by PIF and TOC1. Out of the 110 PIF-bound Pfeiffer et al., 2014). We therefore examined the PIF/SD- genes in the PIF/SD-induced set (Pfeiffer et al., 2014), 11% regulated genes identified in our analysis for the presence of of the comparable PIF-bound genes (84 genes, as defined in these motifs within 3 kb upstream of the transcription start site. Soy et al., 2016) are predicted to be bound as well by TOC1 This analysis is presented in Figure 6A and Table 1. We found (Huang et al., 2012). Conversely, expression analysis of the that genes in the induced set displayed SS enrichment of both PIF and TOC1 bound genes that were defined as “predawn- elements. The G-box and PBE were present in 54% (p-value: specific PIF-TOC1” (Soy et al., 2016) showed that most are 5.04E–10) and 90% (p-value: 3.79E–07) of the genes, respectively. induced at ZT24 compared to LL in the conditions described Similar results were obtained using SCOPE (Table 1). These here, and that this induction is dependent on the presence of PIFq (Supplementary Figure S6). As a control, genes defined as PIF and TOC1 bound but not phased at predawn (Soy et al., 2016) do not show this regulation under SD conditions at ZT24 (Supplementary Figure S6). Together, these results suggest that a subset of the PIF/SD-induced genes might be directly targeted by PIF and TOC1 under SD, consistent with our recent description of TOC1 repressive action on PIF transcriptional activity as a mechanism to gate growth at dawn (Soy et al., 2016). In agreement, growth-related genes targeted by TOC1 under SD conditions (like PIL1, HFR1, CKX5, FHL, or HAT2) that were shown to have early expression in a toc1 mutant under SD conditions (Soy et al., 2016), are included in our PIF/SD-induced gene set (Dataset S1). In contrast to the induced set, we found that genes in the repressed set were only marginally enriched in the G-box and the PBE element, and were not significantly enriched in the combination of G-box+PBE (p-value: 0.43; Figure 6A and Table 1). Accordingly, only 10% of them were previously identified as PIF-bound genes (Hornitschek et al., 2012; Pfeiffer et al., 2014; Figure 6A). These results suggest that the majority of the PIF/SD-repressed genes are not direct targets of the PIFs, in agreement with Leivar and Monte (2014) and Martín et al. (2016). Interestingly, analysis of DNA binding motifs revealed that the PIF/SD-repressed gene set is significantly enriched in the FIGURE 6 | Promoter analysis of PIF/SD-regulated genes. (A) Venn ABA-responsive (ABRE) ACGTGGC element (p-value: 0.00046, diagram shows the overlap between the PIF/SD-induced (top) or present in 28% of the genes). Similar results were obtained PIF/SD-repressed (bottom) genes and the previously defined PIF-bound genes using SCOPE (Table 1). This finding suggests that changes in (Hornitschek et al., 2012; Pfeiffer et al., 2014). (B) Enriched DNA elements in the promoters of PIF/SD-induced (top) or PIF/SD-repressed (bottom) genes. ABA levels or response might underlie the PIF-mediated down- regulation of these genes at the end of the night. Frontiers in Plant Science | www.frontiersin.org 10 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression TABLE 1 | DNA motifs associated with PIF/SD-induced and PIF/SD-repressed genes. Motif Sig value Coverage p-value Sig value Coverage p-value G-BOX cacgtg 49.1 50.80% 5.04E–10 −1.5 35.30% 0.040 PBE/HUD cacatg 61.7 85.10% 3.79E–07 −1.2 67.60% 0.801 G-BOX + PBE cacrtg 103.9 95.60% 5.15E–11 1.8 77.70% 0.434 T1ME caca 14.2 100% 0.980 34.3 100% 0.985 ME ccacac –3.9 51.90% 0.390 1 51.10% 0.578 PBX atgggcc 1.8 23.20% 0.487 − 1.5 28.10% 0.086 EE aaaatatct –6.4 17.70% 0.098 −10.6 15.10% 0.396 EE-like aatatct –2.7 63.50% 0.039 −7.8 56.80% 0.511 GATA ggata –6 97.80% 0.779 −0.9 98.60% 0.085 SBX aagccc –6.5 38.70% 0.997 −0.9 49.60% 0.280 TBX aaaccct –2.6 39.80% 0.947 −8.3 51.10% 0.363 Extended G-BOX cacgtggg 40 11% 9.32E–08 −8.1 3.60% 0.110 Extended PBE cacatggg 13.6 11.60% 9.38E–05 −7.9 1.40% 0.912 Extended G-BOX +PBE cacrtggg 53.2 22.10% 7.75E–11 −10.1 5.00% 0.487 ABRE acgtggc –4.1 19.30% 0.029 15.3 28.80% 4.67E–04 Presence of previously described diurnal and circadian DNA motifs (Michael et al., 2008b; Soy et al., 2016) in the promoters of PIF/SD-induced (left, in blue) and PIF/SD-repressed genes (right, in red). Analyses were performed using the SCOPE motif finder within 2.5 kb upstream of the translation initiation site (TSS; http://genie.dartmouth.edu/scope/) and the Patmach and Motif analysis tools available at the TAIR website (http://www.Arabidopsis.org/cgi-bin/patmatch/nph-pat match.pl) within 3 kb upstream of the TSS. Base pair “r” refers to either “g” or “a”. Sig value (a measure of how likely the motif in question is to have been this over- represented. Values greater than zero are meaningful, and the higher the number, the better) and coverage (indicating the percentage of the genes in the query that contain at least one instance of the sequence being examined) are as defined by SCOPE. DISCUSSION directed to specifically determine the genes bound by TOC1 and PIFs under SD are needed to establish a comprehensive list of In Arabidopsis seedlings grown under SD conditions, PIF activity PIF and TOC1 target genes under the SD conditions assayed peaks at dawn coinciding with maximum PIF protein levels and here. declining abundance of their repressor TOC1 (Soy et al., 2016). In contrast to the PIF/SD-induced set, our observation The genomic analysis presented here defines the PIFq-regulated that the repressed gene set is still largely rhythmic under LL transcriptome at the end of the night in SD, and establishes (Figure 3) suggests that these genes are likely clock-output that approximately 60% of the SD- and PIFq-regulated genes genes. This establishes a mechanistic difference in how induced correspond to induced genes, whereas 40% are SD- and PIFq- and repressed genes are regulated under SD. Interestingly, in repressed. The data identify a contrasting PIFq contribution addition, the expression dynamics of the repressed gene set is to the expression, rhythmicity, and function in the induced not the mirror image of the induced set, i.e., they are not genes and repressed gene sets, and suggest different PIF-regulatory that are progressively repressed during the night as a result mechanisms to control growth and other aspects of seedling of PIF accumulation. Instead, the PIF/SD-repressed genes are development under SD. also progressively induced during the night (Figure 3). The Our observation that PIF-induced genes oscillate in SD but mechanism underlying this pattern is currently not understood, are not rhythmic under LL expands our recent observation and could involve induction by yet an unknown mechanism. for a smaller subset of PIF-induced genes (Soy et al., 2016). In the WT, these genes peak in the morning, whereas genetic This result provides additional supporting evidence that PIF- removal of the PIFs induces higher accumulation at dawn. These induced genes at dawn do not cycle in conditions where PIFs observations indicate that the PIFs are suppressing an early do not accumulate (Soy et al., 2014, 2016), and therefore cannot transcription induction during the last part of the night, and be considered classical clock outputs. Interestingly, our results suggest that the PIFs are actively preventing these genes to peak showing that 11% of the PIF/SD-induced genes are directly during the night. In the morning, when PIF levels are low due to targeted by PIF and TOC1 is in accordance with our recent phytochrome-induced degradation, these genes are induced and description of the TOC1/PIF antagonistic interplay to gate display their maximum peak of expression. growth at the end of the night (Soy et al., 2016). Additionally, Our findings that the promoters of PIF/SD-induced genes are because both the TOC1 and PIFq ChIP-seq experiments referred enriched in the G-box and in the PBE/HUD elements, and that to in here were performed under conditions different to SD (12:12 the PIF/SD-induced set is enriched in putative PIF-bound genes and deetiolation, respectively; Huang et al., 2012; Pfeiffer et al., (Figure 6), agree with previous studies (Leivar et al., 2009; Zhang 2014), the actual % of PIF-TOC1 targets among the PIF/SD- et al., 2013; Leivar and Monte, 2014; Pfeiffer et al., 2014) and are in induced genes might be larger. For example, the PIF direct target line with the PIFs acting as transcriptional activators (Huq et al., PIL1 was not identified as TOC1 target by Huang et al. (2012), 2004; Al-Sady et al., 2008; de Lucas et al., 2008; Shen et al., 2008; whereas we have experimentally shown that it is bound by TOC1 Hornitschek et al., 2009). The PBE/HUD element was found to under SD conditions (Soy et al., 2016). Further experiments be enriched in the promoters of hormone-related genes induced Frontiers in Plant Science | www.frontiersin.org 11 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression at dawn (Michael et al., 2008a). Accordingly, we found AUTHOR CONTRIBUTIONS enrichment in growth- and hormone-related genes in the PIF/SD-induced set (Figure 5). Interestingly, the induced set was JS and EM designed and performed the research. GM and EM also enriched in the extended G-box/PBE element that we have analyzed the data and wrote the manuscript. recently described in the promoters of predawn-specific PIF and TOC1 cotarget genes (Soy et al., 2016), which agrees with the PIF/SD-induced set being co-targeted antagonistically by both FUNDING factors to time growth. Importantly, although we found that the PIF/SD-repressed This work was supported by grants from the Spanish “Ministerio gene set is not enriched in putative PIF-bound genes, some of de Ciencia e Innovación” BIO2009-07675, “Ministerio de the repressed genes are also likely targeted directly by the PIFs Economía” BIO2012-31672, and from the Generalitat de (10%, 15 genes; Figure 6; Pfeiffer et al., 2014), suggesting that Catalunya 2014-SGR-1406 to EM. the PIFs might also act as transcriptional repressors under SD, in accord with previous results during the deetiolation response ACKNOWLEDGMENT (Toledo-Ortiz et al., 2010; Chen et al., 2013; Liu et al., 2013; Zhang et al., 2013; Leivar and Monte, 2014; Pfeiffer et al., We thank Pablo Leivar for helpful discussions. 2014). Notably, the master regulators of chloroplast development SIGMA5 and GLK2 (Waters et al., 2009; Noordally et al., 2013; Martín et al., 2016) are among the PIF/SD-repressed and PIF- SUPPLEMENTARY MATERIAL bound genes (Oh et al., 2012; Pfeiffer et al., 2014; Song et al., 2014). This finding suggests that they could act as intermediaries The Supplementary Material for this article can be found online between the PIFs and the chloroplast-related PIF/SD-repressed at: http://journal.frontiersin.org/article/10.3389/fpls.2016.00962 genes that are not directly bound by the PIFs, delivering the FIGURE S1 | Definition of 191 PIF/SD-induced and 145 PIF/SD-repressed PIF signal to the downstream targets indirectly repressed by genes. (A) Venn diagram showing the four-way comparison between SS 1.5F PIFq. The mechanism of how PIFs repress gene expression is PIF-regulated (WT SD vs. pifq SD, induced and repressed) and SD-regulated (WT still largely unknown, but a recent paper showing that PIF3 SD vs. WT LL, induced and repressed) gene sets. The number of differentially interacts with the deacetylase HDA15 to repress chlorophyll expressed genes in each set is indicated. The list of genes in each class is provided in Dataset S1. SS1.5F: genes whose expression changed statistically biosynthesis and photosynthetic genes (Liu et al., 2013) supports significantly and by at least 1.5 fold. (B) Two-dimensional-cluster diagram the notion that PIFs might associate with repressive histone- depicting the identified PIF- regulated SS1.5F genes in 3-day-old SD- and modification factors to function as transcriptional repressors. LL-grown WT and pifq seedlings. A total of 207 genes are upregulated Additionally, our finding that PIFq cannot explain 100% of (PIF-repressed) in the absence of PIFq, whereas 331 correspond to genes that are the repression (in contrast to the PIF/SD-induced genes set, downregulated (PIF-induced), as defined in (A). apparently regulated solely by PIFq) agrees with the possibility FIGURE S2 | Gene expression of SS1.5F PIF-specific and SD-specific that part of the repression mechanism might involve chromatin- regulated genes. (A) Vioplots of log2 expression values in WT and pifq (SD and remodeling events independent of PIFq. Alternatively, factors LL) of the 881 SD-induced (left) and the 954 SD-repressed (right) genes that are not regulated by PIFq, as defined in Supplementary Figure S1A. (B) Vioplots of other than PIFq that accumulate or are activated in the dark log2 expression values in WT and pifq (SD and LL) of the 130 PIF-induced (left) could be contributing to the repression of these genes in and the 59 PIF-repressed (right) genes that are not regulated by SD, as defined in SD compared to LL. Finally, our finding that the PIF/SD- Supplementary Figure S1A. (A,B) Statistically significant differences from WT regulated genes (and more significantly the repressed set) SD or WT LL by Willcoxon test are indicated in the upper and lower part, are enriched in genes containing the ABA-responsive element respectively (∗ p-value < 0.05; ∗∗ p-value < 0.01; ∗∗∗ p-value < 0.001. n.s., non-significant). The mean log2 FC value relative to the expression value of WT LL ABRE opens the intriguing possibility that ABA mediates some for each sample is indicated. aspects of PIF/SD-regulated development. To our knowledge, ABRE has not been recognized before as an overrepresented FIGURE S3 | PIF/SD-regulated genes ranked based on the percentage contribution of PIFq to their expression in SD. PIF/SD-induced (left) and DNA binding element for PIF-regulated genes, which might PIF/SD-repressed genes (right) were arranged in descending percentage based indicate a role for ABA in PIF-regulated development specifically on the contribution of PIFq to their expression in SD. Percentage of genes with a under SD. The hormone ABA has been described to regulate contribution of <50% or >100% is indicated for each gene set. growth, stomata opening and stress responses (Finkelstein, FIGURE S4 | Diurnal pattern of PIF/SD-regulated genes in 12:12. Median 2013). Further studies are necessary to explore the possible values (thick line) and upper and lower quartiles (thin lines) of log2 expression for involvement and role of ABA under the SD conditions examined all PIF/SD-induced (left) and PIF/SD-repressed (right) under 12:12. Data were here. obtained from http://diurnal.mocklerlab.org. Day: yellow; Night: gray. FIGURE S5 | Expression of the SD-specific PIF/SD-regulated genes under the previously described shade, deetiolation and photoperiodic growth ADDITIONAL INFORMATION conditions. Bar graph of microarray data [deetiolation (Leivar et al., 2009), shade (Leivar et al., 2012), and photoperiodic conditions (Nozue et al., 2011)] showing the log2 FC expression relative to the WT Light in deetiolation and shade, and Microarray data reported in this study have been deposited in the relative to pif4pif5 in growth phase, of the 76 PIF/SD-induced (top) and 63 Gene Expression Omnibus database under the accession number PIF/SD-repressed (bottom) genes not previously defined as PIF-regulated in any of GSE81813. the other experimental conditions. SS, statistically significantly; FRed: Far red. Frontiers in Plant Science | www.frontiersin.org 12 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression FIGURE S6 | Expression of the previously defined “dawn-specific Willcoxon test are indicated (∗ p-value < 0.05; ∗∗∗ p-value < 0.001. n.s., PIF-TOC1” and “no dawn-specific PIF-TOC1” in SD conditions. Box plots non-significant). of microarray data showing the log2 FC expression in WT and pifq in SD and LL, relative to the WT LL, of the “dawn-specific PIF-TOC1” (green) and “no DATASET S1 | Expression data and characterization (PIF- and dawn-specific PIF-TOC1” (purple) genes recently defined by Soy et al. (2016). SD-regulation, gene class, phase, and subcellular localization) of the 2,373 Statistically significant differences among FC values relative to WT SD by SD- and/or PIF-regulated genes identified. REFERENCES Jiao, Y., Lau, O. S., and Deng, X. W. (2007). Light-regulated transcriptional networks in higher plants. Nat. Rev. Genet. 8, 217–230. doi: 10.1038/nrg2049 Al-Sady, B., Kikis, E. A., Monte, E., and Quail, P. H. (2008). Mechanistic duality Leivar, P., and Monte, E. (2014). PIFs: systems integrators in plant development. of transcription factor function in phytochrome signaling. Proc. Natl. Acad. Sci. Plant Cell 26, 56–78. doi: 10.1105/tpc.113.120857 U.S.A. 105, 2232–2237. doi: 10.1073/pnas.0711675105 Leivar, P., Monte, E., Oka, Y., Liu, T., Carle, C., Castillon, A., et al. Al-Sady, B., Ni, W., Kircher, S., Schäfer, E., and Quail, P. H. (2006). Photoactivated (2008). Multiple phytochrome-interacting bHLH transcription factors repress phytochrome induces rapid PIF3 phosphorylation prior to proteasome- premature seedling photomorphogenesis in darkness. Curr. Biol. 18, 1815– mediated degradation. Mol. Cell. 23, 439–446. doi: 10.1016/j.molcel.2006.06.011 1823. doi: 10.1016/j.cub.2008.10.058 Arana, M. V., Marin-de la Rosa, N., Maloof, J. N., Blazquez, M. A., and Alabadi, D. Leivar, P., Tepperman, J. M., Cohn, M. M., Monte, E., Al-Sady, B., Erickson, E., (2011). Circadian oscillation of gibberellin signaling in Arabidopsis. Proc. Natl. et al. (2012). Dynamic antagonism between phytochromes and PIF family basic Acad. Sci. U.S.A. 108, 9292–9297. doi: 10.1073/pnas.1101050108 helix-loop-helix factors induces selective reciprocal responses to light and shade Bae, G., and Choi, G. (2008). Decoding of light signals by plant phytochromes in a rapidly responsive transcriptional network in Arabidopsis. Plant Cell 24, and their interacting proteins. Annu. Rev. Plant Biol. 59, 281–311. doi: 1398–1419. doi: 10.1105/tpc.112.095711 10.1146/annurev.arplant.59.032607.092859 Leivar, P., Tepperman, J. M., Monte, E., Calderon, R. H., Liu, T. L., and Quail, Bauer, D., Viczián, A., Kircher, S., Nobis, T., Nitschke, R., Kunkel, T., et al. P. H. (2009). Definition of early transcriptional circuitry involved in light- (2004). Constitutive photomorphogenesis 1 and multiple photoreceptors induced reversal of PIF-imposed repression of photomorphogenesis in young control degradation of phytochrome interacting factor 3, a transcription factor Arabidopsis seedlings. Plant Cell 21, 3535–3553. doi: 10.1105/tpc.109.070672 required for light signaling in Arabidopsis. Plant Cell 16, 1433–1445. doi: Liu, X., Chen, C. Y., Wang, K. C., Luo, M., Tai, R., Yuan, L., et al. 10.1105/tpc.021568 (2013). PHYTOCHROME INTERACTING FACTOR3 associates with the Bernardo-García, S., de Lucas, M., Martínez, C., Espinosa-Ruiz, A., Davière, histone deacetylase HDA15 in repression of chlorophyll biosynthesis and J. M., and Prat, S. (2014). BR-dependent phosphorylation modulates PIF4 photosynthesis in etiolated Arabidopsis seedlings. Plant Cell 25, 1258–1273. doi: transcriptional activity and shapes diurnal hypocotyl growth. Genes Dev. 28, 10.1105/tpc.113.109710 1681–1694. doi: 10.1101/gad.243675.114 Lorrain, S., Allen, T., Duek, P. D., Whitelam, G. C., and Fankhauser, C. (2008). Chen, D., Xu, G., Tang, W., Jing, Y., Ji, Q., Fei, Z., et al. (2013). Antagonistic basic Phytochrome-mediated inhibition of shade avoidance involves degradation helix-loop-helix/bZIP transcription factors form transcriptional modules that of growth-promoting bHLH transcription factors. Plant J. 53, 312–323. doi: integrate light and reactive oxygen species signaling in Arabidopsis. Plant Cell 10.1111/j.1365-313X.2007.03341.x 25, 1657–1673. doi: 10.1105/tpc.112.104869 Martín, G., Leivar, P., Ludevid, D., Tepperman, J. M., Quail, P. H., and de Lucas, M., Daviere, J. M., Rodriguez-Falcon, M., Pontin, M., Iglesias- Monte, E. (2016). Phytochrome and retrograde signalling pathways converge to Pedraz, J. M., Lorrain, S., et al. (2008). A molecular framework for antagonistically regulate a light-induced transcriptional network. Nat. Comm. light and gibberellin control of cell elongation. Nature 451, 480–484. doi: 7:11431. doi: 10.1038/ncomms11431 10.1038/nature06520 Michael, T. P., Breton, G., Hazen, S. P., Priest, H., Mockler, T. C., Kay, Feng, S., Martinez, C., Gusmaroli, G., Wang, Y., Zhou, J., Wang, F., et al. (2008). S. A., et al. (2008a). A morning-specific phytohormone gene expression Coordinated regulation of Arabidopsis thaliana development by light and program underlying rhythmic plant growth. PLoS Biol. 6:e225. doi: 10.1371/ gibberellins. Nature 451, 475–479. doi: 10.1038/nature06448 journal.pbio.0060225 Finkelstein, R. (2013). Abscisic Acid synthesis and response. Arabidopsis Book Nov Michael, T. P., Mockler, T. C., Breton, G., McEntee, C., Byer, A., Trout, J. D., et al. 1:e0166. doi: 10.1199/tab.0166 (2008b). Network discovery pipeline elucidates conserved time-of-day-specific Fitter, D. W., Martin, D. J., Copley, M. J., Scotland, R. W., and Langdale, J. A. cis-regulatory modules. PLoS Genet. 4:e14. doi: 10.1371/journal.pgen.0040014 (2002). GLK gene pairs regulate chloroplast development in diverse plant Mockler, T. C., Michael, T. P., Priest, H. D., Shen, R., Sullivan, C. M., Givan, species. Plant J. 31, 713–727. doi: 10.1046/j.1365-313X.2002.01390.x S. A., et al. (2007). The DIURNAL project: DIURNAL and circadian expression Galvao, V. C., and Fankhauser, C. (2015). Sensing the light environment in plants: profiling, model-based pattern matching, and promoter analysis. Cold Spring photoreceptors and early signaling steps. Curr. Opin. Neurobiol. 34, 46–53. doi: Harb. Symp. Quant. Biol. 72, 353–363. doi: 10.1101/sqb.2007.72.006 10.1016/j.conb.2015.01.013 Monte, E., Alonso, J. M., Ecker, J. R., Zhang, Y., Li, X., Young, J., et al. (2003). Hornitschek, P., Kohnen, M. V., Lorrain, S., Rougemont, J., Ljung, K., Lopez- Isolation and characterization of phyC mutants in Arabidopsis reveals complex Vidriero, I., et al. (2012). Phytochrome interacting factors 4 and 5 control crosstalk between phytochrome signaling pathways. Plant Cell 15, 1962–1980. seedling growth in changing light conditions by directly controlling auxin doi: 10.1105/tpc.012971 signaling. Plant J. 71, 699–711. doi: 10.1111/j.1365-313X.2012.05033.x Monte, E., Tepperman, J. M., Al-Sady, B., Kaczorowski, K. A., Alonso, J. M., Ecker, Hornitschek, P., Lorrain, S., Zoete, V., Michielin, O., and Fankhauser, C. (2009). J. R., et al. (2004). The phytochrome-interacting transcription factor, PIF3, acts Inhibition of the shade avoidance response by formation of non-DNA binding early, selectively, and positively in light-induced chloroplast development. Proc. bHLH heterodimers. EMBO J. 28, 3893–3902. doi: 10.1038/emboj.2009.306 Natl. Acad. Sci. U.S.A. 101, 16091–16098. doi: 10.1073/pnas.0407107101 Huang, D. W., Sherman, B. T., Tan, Q., Collins, J. R., Alvord, W. G., Roayaei, J., Ni, M., Tepperman, J. M., and Quail, P. H. (1999). Binding of phytochrome B to et al. (2007). The DAVID Gene Functional Classification Tool: a novel biological its nuclear signalling partner PIF3 is reversibly induced by light. Nature 400, module-centric algorithm to functionally analyze large gene lists. Genome Biol. 781–784. doi: 10.1038/23500 8:R183. doi: 10.1186/gb-2007-8-9-r183 Ni, W., Xu, S. L., Chalkley, R. J., Pham, T. N., Guan, S., Maltby, D. A., et al. Huang, W., Perez-Garcia, P., Pokhilko, A., Millar, A. J., Antoshechkin, I., (2013). Multisite light-induced phosphorylation of the transcription factor Riechmann, J. L., et al. (2012). Mapping the core of the Arabidopsis circadian PIF3 is necessary for both its rapid degradation and concomitant negative clock defines the network structure of the oscillator. Science 336, 75–79. doi: feedback modulation of photoreceptor phyB levels in Arabidopsis. Plant Cell 10.1126/science.1219075 25, 2679–2698. doi: 10.1105/tpc.113.112342 Huq, E., Al-Sady, B., Hudson, M., Kim, C., Apel, K., and Quail, P. H. (2004). Nieto, C., Lopez-Salmeron, V., Daviere, J. M., and Prat, S. (2015). ELF3-PIF4 Phytochrome-interacting factor 1 is a critical bHLH regulator of chlorophyll interaction regulates plant growth independently of the evening complex. Curr. biosynthesis. Science 305, 1937–1941. doi: 10.1126/science.1099728 Biol. 25, 187–193. doi: 10.1016/j.cub.2014.10.070 Frontiers in Plant Science | www.frontiersin.org 13 July 2016 | Volume 7 | Article 962 Martín et al. PIFq Contribution to Diurnal Gene Expression Niwa, Y., Yamashino, T., and Mizuno, T. (2009). The circadian clock regulates factors PIF3, 4, and 5. Mol. Plant 7, 1776–1787. doi: 10.1093/mp/ the photoperiodic response of hypocotyl elongation through a coincidence ssu109 mechanism in Arabidopsis thaliana. Plant Cell Physiol. 50, 838–854. doi: Soy, J., Leivar, P., Gonzalez-Schain, N., Martín, G., Diaz, C., Sentandreu, M., 10.1093/pcp/pcp028 et al. (2016). Molecular convergence of clock and photosensory pathways Nomoto, Y., Kubozono, S., Yamashino, T., Nakamichi, N., and Mizuno, T. (2012). through PIF3-TOC1 interaction and co-occupancy of target promoters. Circadian clock- and PIF4-controlled plant growth: a coincidence mechanism Proc. Natl. Acad. Sci. U.S.A. 113, 4870–4875. doi: 10.1073/pnas.16037 directly integrates a hormone signaling network into the photoperiodic control 45113 of plant architectures in Arabidopsis thaliana. Plant Cell Physiol. 53, 1950–1964. Soy, J., Leivar, P., Gonzalez-Schain, N., Sentandreu, M., Prat, S., Quail, P. H., et al. doi: 10.1093/pcp/pcs137 (2012). Phytochrome-imposed oscillations in PIF3 protein abundance regulate Noordally, Z. B., Ishii, K., Atkins, K. A., Wetherill, S. J., Kusakina, J., Walton, hypocotyl growth under diurnal light/dark conditions in Arabidopsis. Plant J. E. J., et al. (2013). Circadian control of chloroplast transcription by a 71, 390–401. doi: 10.1111/j.1365-313X.2012.04992.x nuclear-encoded timing signal. Science 339, 1316–1319. doi: 10.1126/science. Soy, J., Leivar, P., and Monte, E. (2014). PIF1 promotes phytochrome-regulated 1230397 growth under photoperiodic conditions in Arabidopsis together with PIF3, Nozue, K., Covington, M. F., Duek, P. D., Lorrain, S., Fankhauser, C., Harmer, S. L., PIF4, and PIF5. J. Exp. Bot. 65, 2925–2936. doi: 10.1093/jxb/ert465 et al. (2007). Rhythmic growth explained by coincidence between internal and Toledo-Ortiz, G., Huq, E., and Rodríguez-Concepción, M. (2010). Direct external cues. Nature 448, 358–361. doi: 10.1038/nature05946 regulation of phytoene synthase gene expression and carotenoid biosynthesis Nozue, K., Harmer, S. L., and Maloof, J. N. (2011). Genomic analysis of circadian by phytochrome-interacting factors. Proc. Natl. Acad. Sci. U.S.A. 107, 11626– clock-, light-, and growth-correlated genes reveals PHYTOCHROME- 11631. doi: 10.1073/pnas.0914428107 INTERACTING FACTOR5 as a modulator of auxin signaling in Arabidopsis. Waters, M. T., Wang, P., Korkaric, M., Capper, R. G., Saunders, N. J., and Plant Physiol. 156, 357–372. doi: 10.1104/pp.111.172684 Langdale, J. A. (2009). GLK transcription factors coordinate expression of Nusinow, D. A., Helfer, A., Hamilton, E. E., King, J. J., Imaizumi, T., Schultz, T. F., the photosynthetic apparatus in Arabidopsis. Plant Cell 21, 1109–1128. doi: et al. (2011). The ELF4-ELF3-LUX complex links the circadian clock to diurnal 10.1105/tpc.108.065250 control of hypocotyl growth. Nature 475, 398–402. doi: 10.1038/nature10182 Xu, X., Paik, I., Zhu, L., and Huq, E. (2015). Illuminating progress in phytochrome- Oh, E., Zhu, J. Y., and Wang, Z. Y. (2012). Interaction between BZR1 and PIF4 mediated light signaling pathways. Trends Plant Sci. 20, 641–650. doi: integrates brassinosteroid and environmental responses. Nat. Cell Biol. 14, 10.1016/j.tplants.2015.06.010 802–809. doi: 10.1038/ncb2545 Yamashino, T., Nomoto, Y., Lorrain, S., Miyachi, M., Ito, S., Nakamichi, N., Pfeiffer, A., Shi, H., Tepperman, J. M., Zhang, Y., and Quail, P. H. (2014). et al. (2013). Verification at the protein level of the PIF4-mediated external Combinatorial complexity in a transcriptionally centered signaling hub in coincidence model for the temperature-adaptive photoperiodic control of Arabidopsis. Mol Plant 7, 1598–1618. doi: 10.1093/mp/ssu087 plant growth in Arabidopsis thaliana. Plant Signal. Behav. 8:e23390. doi: Rockwell, N. C., Su, Y. S., and Lagarias, J. C. (2006). Phytochrome structure 10.4161/psb.23390 and signaling mechanisms. Annu. Rev. Plant Biol. 57, 837–858. doi: 10.1146/ Zhang, Y., Mayba, O., Pfeiffer, A., Shi, H., Tepperman, J. M., Speed, T. P., et al. annurev.arplant.56.032604.144208 (2013). A quartet of PIF bHLH factors provides a transcriptionally centered Salter, M. G., Franklin, K. A., and Whitelam, G. C. (2003). Gating of the rapid signaling hub that regulates seedling morphogenesis through differential shade-avoidance response by the circadian clock in plants. Nature 426, 680–683. expression-patterning of shared target genes in Arabidopsis. PLoS Genet. doi: 10.1038/nature02174 9:e1003244. doi: 10.1371/journal.pgen.1003244 Sentandreu, M., Martín, G., González-Schain, N., Leivar, P., Soy, J., Tepperman, J. M., et al. (2011). Functional profiling identifies genes involved in organ- Conflict of Interest Statement: The authors declare that the research was specific branches of the PIF3 regulatory network in Arabidopsis. Plant Cell 23, conducted in the absence of any commercial or financial relationships that could 3974–3991. doi: 10.1105/tpc.111.088161 be construed as a potential conflict of interest. Shen, H., Zhu, L., Castillon, A., Majee, M., Downie, B., and Huq, E. (2008). Light-induced phosphorylation and degradation of the negative regulator Copyright © 2016 Martín, Soy and Monte. This is an open-access article distributed PHYTOCHROME-INTERACTING FACTOR1 from Arabidopsis depend upon under the terms of the Creative Commons Attribution License (CC BY). The use, its direct physical interactions with photoactivated phytochromes. Plant Cell 20, distribution or reproduction in other forums is permitted, provided the original 1586–1602. doi: 10.1105/tpc.108.060020 author(s) or licensor are credited and that the original publication in this journal Song, Y., Yang, C., Gao, S., Zhang, W., Li, L., and Kuai, B. (2014). Age- is cited, in accordance with accepted academic practice. No use, distribution or triggered and dark-induced leaf senescence require the bHLH transcription reproduction is permitted which does not comply with these terms. Frontiers in Plant Science | www.frontiersin.org 14 July 2016 | Volume 7 | Article 962