Arabidopsis noncoding RNA mediates control of photomorphogenesis ...

Download PDF
2 downloads 44 Views 1MB Size Report
Comment
Jul 15, 2014 - the entire plant life cycle (1). One of the ... are an important class of. ncRNAs that recently have been shown to possess a diverse set of.
Arabidopsis noncoding RNA mediates control of photomorphogenesis by red light Yuqiu Wanga, Xiuduo Fana, Fang Lina, Guangming Hea, William Terzaghib,c, Danmeng Zhua,1, and Xing Wang Denga,c,1 a State Key Laboratory of Protein and Plant Gene Research, Peking-Tsinghua Center for Life Sciences, College of Life Sciences, Peking University, Beijing 100871, China; bDepartment of Biology, Wilkes University, Wilkes-Barre, PA 18766; and cDepartment of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT 06520

Contributed by Xing Wang Deng, May 23, 2014 (sent for review April 23, 2014)

light signaling

| transcriptional regulation

L

ight is one of the most important environmental cues influencing the growth and development of plants throughout the entire plant life cycle (1). One of the best-characterized lightcontrolled developmental processes is seedling morphogenesis: photomorphogenesis (de-etiolation) in light and skotomorphogenesis (etiolation) in darkness. Photomorphogenesis is characterized by the development of short hypocotyls, opened cotyledons, and chlorophyll synthesis, whereas skotomorphogenesis is characterized by the development of long hypocotyls, closed cotyledons with apical hooks, and undifferentiated plastids (2). The switch from skotomorphogenesis to photomorphogenesis is critical for seedling survival and is dependent on the precise control of geneexpression patterns by genetic and epigenetic pathways (3–6). Plants have evolved multiple photoreceptors that are capable of perceiving and propagating a variety of light signals. For example, five phytochromes (phyA–phyE) that perceive far-red and red light, two cryptochromes (CRY1 and CRY2) and two phototropins (PHOT1 and PHOT2) that sense blue/UV-A light, and a UV-B photoreceptor (UVR8) have been identified in the model plant Arabidopsis thaliana (7, 8). Traditional genetic and molecular analyses combined with recent genomic studies have identified a number of protein-coding genes that function as positive or negative regulators of seedling photomorphogenesis under different light conditions (1, 2, 9). Among these genes, a family of basic helix– loop–helix (bHLH) transcription factors, designated “phytochromeinteracting factors” (PIFs), has been shown to repress seedling photomorphogenesis in the dark. PIF1, PIF3 (the founding member), PIF4, and PIF5 are the most extensively characterized members of this family. Specifically, pif3, pif4, and pif5 mutants have been shown to exhibit hyper-photomorphogenic phenotypes in response to continuous red (cR) light, whereas the quadruple mutant

www.pnas.org/cgi/doi/10.1073/pnas.1409457111

(pifq) has been shown to display a constitutive photomorphogenic phenotype in darkness (10–14). Recent studies have revealed that these PIFs are targeted for rapid degradation via the ubiquitin– proteasome pathway by photo-activated phytochromes in light (15). Genome-wide transcriptomic and ChIP-sequence analyses have identified numerous genes regulated by PIFs. Many targets of PIFs encode transcription factors, suggesting that PIFs act early in and define a central hub in the phytochrome-mediated light-signaling pathways controlling seedling photomorphogenesis (3, 16, 17). However, the way in which these PIF genes are regulated at the transcriptional level is still the subject of intense investigation. Recent genome-wide studies have shown that noncoding RNAs (ncRNAs) comprise a significant portion of the transcriptome in animals and plants. Despite their lack of protein-coding potential, many ncRNAs have been recognized as essential regulators of gene expression (18, 19). Long ncRNAs (lncRNAs), which vary in length from 200 nt to dozens of kilobases, are an important class of ncRNAs that recently have been shown to possess a diverse set of functions in eukaryotes (20). Although thousands of lncRNAs have been systematically identified or predicted in silico in Arabidopsis, wheat, and maize (21–23), very few have been characterized functionally (24–29). Specifically, no report to date has outlined the function of lncRNAs in photomorphogenesis. Building on our recent global annotation of Arabidopsis 50- to 300-nt ncRNAs and our large-scale reverse genetic analysis (30), here we report the identification and characterization of an evolutionarily conserved ncRNA of 236 nt in land plants, HID1 (HIDDEN TREASURE 1), that modulates red-light–mediated seedling photomorphogenesis in Arabidopsis. Knocking down HID1 led to increased levels of PIF3 mRNA, which in turn correlated directly with the elongated hypocotyl phenotype Significance The dynamic regulation of gene-expression programs is both critical to and regulated precisely in the light-mediated seedling photomorphogenesis of higher plants. Our work adds HIDDEN TREASURE 1 (HID1), a noncoding RNA that acts as a positive regulator of photomorphogenesis, to the current group of pivotal genetic factors known to control photomorphogenesis. Specifically, our data obtained by numerous approaches reveal that HID1 modulates red light-mediated photomorphogenesis by directly repressing PHYTOCHROME-INTERACTING FACTOR 3, which encodes a key transcription factor that inhibits red light responses. HID1 appears to be highly conserved among higher plants. Author contributions: Y.W., D.Z., and X.W.D. designed research; Y.W., X.F., F.L., and D.Z. performed research; Y.W., G.H., W.T., D.Z., and X.W.D. analyzed data; and Y.W., W.T., D.Z., and X.W.D. wrote the paper. The authors declare no conflict of interest. The data reported in this paper have been deposited in GenBank database (accession no. KM044009) and in the Gene Expression Omnibus (GEO) database, www.ncbi.nlm. nih.gov/geo (accession no. GSE57806). 1

To whom correspondence may be addressed. E-mail: [email protected], [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1409457111/-/DCSupplemental.

PNAS | July 15, 2014 | vol. 111 | no. 28 | 10359–10364

PLANT BIOLOGY

Seedling photomorphogenesis is a sophisticated developmental process that is controlled by both the transcriptional and posttranscriptional regulation of gene expression. Here, we identify an Arabidopsis noncoding RNA, designated HIDDEN TREASURE 1 (HID1), as a factor promoting photomorphogenesis in continuous red light (cR). We show that HID1 acts through PHYTOCHROMEINTERACTING FACTOR 3 (PIF3), which encodes a basic helix–loop– helix transcription factor known to be a key repressor of photomorphogenesis. Knockdown of HID1 in hid1 mutants leads to a significant increase in the expression of PIF3, which in turn drives the development of elongated hypocotyls in cR. We identified two major stem-loops in HID1 that are essential for its modulation of hypocotyl growth in cR-grown seedlings. Furthermore, our data reveal that HID1 is assembled into large nuclear protein–RNA complex(es) and that it associates with the chromatin of the first intron of PIF3 to repress its transcription. Strikingly, phylogenetic analysis reveals that many land plants have conserved homologs of HID1 and that its rice homolog can rescue the mutant phenotype when expressed in Arabidopsis hid1 mutants. We thus concluded that HID1 is a previously uncharacterized noncoding RNA whose function represents another layer of regulation in the precise control of seedling photomorphogenesis.

Fig. 1. The ncRNA mutant hid1 exhibits hyposensitivity to cR. (A) Phenotypes of 5-d-old seedlings grown under various light qualities. (Scale bar: 1 mm.) (B–E) Hypocotyl lengths of seedlings grown in cFr (B), cR (C), or cB (D) over a range of fluence rates or in darkness (E). Data are mean ± SD (n ≥ 20). (F and G) Cotyledon opening angles (in degrees) (F) and total chlorophyll contents (in milligrams per gram) (G) of 5-d-old seedlings grown in cR. *P < 0.05, **P < 0.01 (t test).

observed in the hid1 mutant seedlings grown under cR. We further demonstrated that HID1 is a chromatin-bound RNA and functions as a direct repressor of PIF3 transcription in cR. Thus, to our knowledge, HID1 appears to be the first lncRNA identified as being involved in the precise control of light-mediated seedling development. Results hid1 Exhibits a Hypo-Photomorphogenic Phenotype Under cR. To

identify RNA molecules that function in the light-signaling pathways, we screened a collection of Agrobacterium transferred DNA (T-DNA) insertion mutants that were compromised in either the expression or structure of specific ncRNAs for photomorphogenic phenotypes (30). One such mutant, hid1, was identified and found to express defectively a polycistronic cluster of four ncRNAs (Fig. S1). Upon comparing the hypocotyl lengths of hid1 and WT seedlings grown in continuous far-red (cFr), cR, or continuous blue (cB) light over a range of fluence rates, we observed a significant reduction in the inhibition of hypocotyl growth among the hid1 mutants specifically under cR (Fig. 1A). Moreover, the difference in hypocotyl length observed between the hid1 and WT seedlings was more apparent under higher cR fluence rates (Fig. 1C). In other light conditions and in darkness, hid1 mutants were indistinguishable from WT plants (Fig. 1 B–E). The hid1 mutant cotyledons also were less open than their WT counterparts over the tested range of cR fluence rates (Fig. 1F), and, hid1 seedlings grown under cR contained less chlorophyll than their WT counterparts (Fig. 1G). Overall, our results demonstrate that hid1 mutants exhibit a decreased responsiveness to cR. hid1 is Defective in the ncRNA HID1. In the hid1 mutant, the T-DNA

inserted into a polycistronic cluster of four noncoding genes, causing a dramatic reduction in the expression of all four ncRNAs as verified by Northern blot analysis (Fig. S1B). To determine whether the lack of these ncRNAs was the direct 10360 | www.pnas.org/cgi/doi/10.1073/pnas.1409457111

cause of the elongated hypocotyl phenotype observed in the hid1 seedlings grown under cR, we transformed a DNA fragment encoding all four ncRNAs driven by the CaMV 35S promoter into the hid1 mutant background (Fig. 2A). The resulting transgenic plants (35S:A/hid1) expressed all four ncRNAs at levels slightly higher than WT and completely rescued the hid1 phenotype in cR (Fig. 2 B–D), indicating that the decreased expression of these noncoding genes was responsible for the observed hid1 phenotype in cR. Next, to determine which of these ncRNAs was responsible for the hid1 phenotype, we made several constructs expressing subsets of these ncRNAs under the control of the 35S promoter and transformed them into the hid1 mutant background. We found that the transgenic line (35S:B/hid1) harboring the construct expressing nc3019, nc3018, and nc3017 at WT levels still exhibited the hid1 phenotype in cR (Fig. 2 B–D), thus indicating that these three ncRNAs are not essential for cR-mediated photomorphogenesis. Therefore we reasoned that the reduced expression of nc3020 might be responsible for the hid1 phenotype. We tested this hypothesis by expressing nc3020 from the nc3020 promoter in the hid1 mutant background and found that the hypocotyls of 5-d-old transgenic seedlings grown in cR were the same length as those of WT seedlings (Fig. 2 B–D). Therefore, our data demonstrated that nc3020, hereafter referred to as “HID1,” is a negative regulator of hypocotyl elongation and is necessary for cR-mediated seedling photomorphogenesis. Given that HID1 is a 236-nt ncRNA that had been identified and verified in our previous genomic annotation (30), we next attempted to determine independently whether it acts directly or via a translational product. HID1 has a potential ORF encoding a 44-aa peptide. However, no homolog of this peptide was found in the annotated peptides or proteins of Arabidopsis. To investigate whether this predicted peptide was needed for HID1 function, we prepared two constructs and transformed them into hid1: M1, which contained two mutations changing ATG to ATA in the first predicted ORF, and M2, which changed ATG to AG. This latter mutation broke the predicted ORF but kept the predicted RNA secondary structure intact (Fig. 3A). The transgenic lines harboring either M1 or M2 expressed mutated HID1 at a level comparable to that observed in the WT plants and also fully rescued the reduced inhibition of hypocotyl elongation observed in hid1 under cR (Fig. 3 B–D). Therefore we concluded that HID1 most likely is a bona fide regulatory ncRNA in Arabidopsis that does not need a translational product to exert its function.

Fig. 2. HID1 is the predominant player promoting cR-mediated seedling photomorphogenesis in the hid1 mutant. (A) Schematic illustration of constructs containing different members of the ncRNA gene cluster. (B) Phenotypes of 5-d-old WT, hid1, 35S:A/hid1, 35S:B/hid1, and pHID1:HID1/hid1 seedlings grown in cR. (Scale bar: 1 mm.) (C) Northern blot analysis showing the expression levels of nc3020 (HID1), nc3019, nc3018, and nc3017 in WT, hid1, and indicated transgenic lines, with 5S rRNA as the loading control. (D) Hypocotyl lengths of seedlings grown under the conditions in B. Data are mean ± SD (n ≥ 20). **P < 0.01 (t test).

Wang et al.

HID1 in eight different organs/developmental stages by Northern blot analysis. We observed that HID1 was comparably abundant in all the organs and developmental stages analyzed, suggesting that HID1 was expressed ubiquitously throughout the Arabidopsis development processes (Fig. S2A). In addition, we observed that the expression of HID1 did not vary between seedlings grown in darkness and those grown under the various light conditions included in this study. Similarly, HID1 expression was not altered in mutants lacking the photoreceptor that detects the tested wavelength (e.g., phyA mutant grown in cFr) (Fig. S2B) under any of our light conditions. This result indicates that the abundance of HID1 is not regulated by light. We next examined the role of secondary structure in HID1 function. HID1 was predicted to fold into four stem-loops by the mfold Web Server (http://mfold.rna.albany.edu/?q=mfold/RNAFolding-Form) (Fig. 3A). To test the functional importance of each of these motifs, we generated four constructs containing the endogenous promoter driving HID1 mutants in which one motif, i.e., one of the major stem-loops 1, 2, 3, or 4 (SL1–SL4), was deleted. Each of these deletions removed only an individual motif and thus did not change the secondary structures of the remaining motifs predicted by mfold (Fig. 3E). After each construct was introduced into the hid1 background, the expression of each mutant was confirmed by Northern blot analysis in at least two independent transgenic lines. Then the hypocotyl growth of these two transgenic lines was examined under cR. As shown in Fig. 3 F and G, cR-grown mutants lacking either SL1 or SL3 had hypocotyls similar in length to those in WT plants. In contrast, mutants lacking SL2 or SL4 had elongated hypocotyls similar to cR-grown hid1 seedlings, suggesting that both SL2 and SL4 are essential for HID1-regulated hypocotyl elongation under cR. Thus, our results suggest that HID1’s function in response to cR is not limited to a particular submotif. Instead, it requires a complex structural organization of the HID1 molecule.

Fig. 3. HID1 contains two essential stem-loops. (A) The predicted secondary structure of HID1 showing four major stem-loops, SL1–4. M1 and M2 represent two mutated HID1 derivatives. SL1 is shaded pink, SL2 is blue, SL3 is light green, and SL4 is yellow. (B) Phenotypes of 5-d-old cR-grown WT, pHID1:M1/ hid1 (M1/hid1), and pHID1:M2/hid1 (M2/hid1) seedlings. (Scale bar: 1 mm.) (C and D) The expression levels of HID1 examined by Northern blot (C) and hypocotyl lengths (D) determined in WT, M1/hid1, and M2/hid1 seedlings grown under the conditions in B. Data are mean ± SD (n ≥ 20). (E) Secondary structures of Arabidopsis HID1-deletion derivatives predicted by the mfold web server (Upper) and Northern blot analysis showing their expression in the corresponding transgenic lines as indicated by the black arrowheads (Lower). 5S rRNA was used as the loading control. (F and G) Phenotypes (F) and hypocotyl lengths (G) of 5-d-old cR-grown seedlings of WT, hid1, and transgenic lines expressing the deletion derivatives of HID1. (Scale bar: 1 mm.) **P < 0.01 (t test).

Wang et al.

HID1 Acts Through PIF3 to Modulate Hypocotyl Elongation. To investigate the mode of HID1 action, we analyzed the expression of genes located at either end of the HID1-coding sequence using quantitative RT-PCR (qRT-PCR). Our results showed that the expression of the tested neighboring genes in hid1 or 35S:A/hid1 mutants was comparable to the levels observed in WT plants, or was not changed by defective HID1 (Fig. S3). In addition, we did not find any known light-signaling mediators within 10 kb of either end of the sequence encoding HID1. Thus, our data did not support the hypothesis that the elongated hypocotyl growth observed in cR-grown hid1 seedlings was the consequence of the in-cis regulation of expression of neighboring genes by HID1. To elucidate further the function of HID1 under cR, we compared global changes in gene expression in 5-d-old cR-grown WT and hid1 seedlings by RNA-seq. We found ∼635 genes that displayed statistically significant twofold changes in expression in hid1 mutants as compared with WT plants. Gene Ontology analysis (http://bioinfo.cau.edu.cn/agriGO/) showed that these genes primarily were enriched in Gene Ontology terms related to “response to stimulus” (P = 4.19e−16) (Fig. S4A). Among those genes, a small set of red light–responsive genes that were up-regulated more than twofold in hid1 mutants compared with WT plants particularly caught our attention (Fig. S4B). Notably, PIF3, a bHLH transcription factor that functions antagonistically in photomorphogenesis under prolonged cR and is one of the known positive regulators of hypocotyl elongation, was included in this group. To confirm further the role of HID1 in regulating PIF3 in vivo, we performed qRT-PCR to validate the mRNA levels of PIF3 and other PIFs in WT and hid1 seedlings grown under cR. Our data showed that the mRNA levels of PIF3, but not of the other PIFs, were increased notably in hid1 mutants (Fig. 4A). The increase in PIF3 expression observed in hid1 mutants was confirmed under cR over a wide range of fluence rates (Fig. 4 B and C) and was correlated with the longer hypocotyls observed in the cR-grown hid1 seedlings. In addition, our data showed that in cR-grown 35S:A/hid1 and pHID1:HID1/hid1 seedlings in which the expression of HID1 was equal to or greater than the WT levels, the PIF3 transcript levels declined to WT levels (Fig. 4D), supporting the negative correlation between HID1 and PIF3 mRNA levels. We therefore examined the genetic interaction between HID1 and PIF3 by generating hid1pif3 double mutants. As shown in Fig. 4 E and F, the hypocotyl lengths of the hid1pif3 seedlings closely resembled those of the pif3 single-mutant seedlings, indicating that PIF3 acts downstream of HID1 genetically. HID1 Negatively Regulates PIF3 Gene Expression. To determine whether the increased expression of PIF3 in hid1 mutants is regulated at the transcriptional level, we first treated WT and hid1 seedlings with a transcriptional inhibitor. As shown in Fig. 4G, the rate of PIF3 mRNA decay after the arrest of transcription by actinomycin D did not differ between 5-d-old cR-grown WT and hid1 seedlings, suggesting that PIF3 mRNA stability was not affected by the hid1 mutation. Next, we used qRT-PCR to analyze the transcript levels of PIF3 in WT and hid1 seedlings subjected to cR-to-dark treatments. Intriguingly, PIF3 expression continued to be about twofold greater in hid1 mutants than in WT seedlings during the transfer from 5-d cR to dark for 3–12 h, confirming our hypothesis that HID1 played a role in downregulating PIF3 transcript levels (Fig. 4H). We then examined PIF3 protein levels over the same cR-to-dark time course. Our immunoblot analysis showed that PIF3 proteins accumulated rapidly in hid1 mutants 3 h after the plants were transferred from cR to dark and continued to increase over the course of the subsequent dark treatment (Fig. 4I). In contrast, only a mild increase in PIF3 protein was observed in WT seedlings over the course of the same period; but this increase could be detected clearly until 12 h of dark treatment. Consistent with the induction of PIF3 protein levels, three representative genes known to be direct targets of PNAS | July 15, 2014 | vol. 111 | no. 28 | 10361

PLANT BIOLOGY

HID1 Is Constitutively Expressed and Contains Two Essential StemLoops. We next examined the steady-state expression levels of

PIF3—SNRK2.5, XTR7, and AT4G35720 (16)—also showed significant induction in hid1 compared with WT plants during our dark treatment (Fig. 4 J–L). Thus, our data suggest that HID1 mediates the transcriptional control of PIF3 under cR. HID1 Is Part of Large Nuclear Protein–RNA Complex(es) and Is Capable of Associating with Chromatin. To gain further insights into the

HID1 regulatory mechanism, we determined its cellular localization by biochemical fractionation. Notably, the majority of HID1 was detected in the nuclear fraction (Fig. 5A). Similarly, through the use of an alternative cellular fractionation method (31), we found that ∼90% of HID1 (the P1 fraction) was extracted by mild detergent treatment, suggesting that it is soluble in the nucleoplasm (Fig. 5B). Furthermore, most of the remaining HID1 (the S2 fraction) was associated with the chromatin fraction, suggesting that HID1 may regulate gene

Fig. 4. HID1 represses PIF3 expression to modulate hypocotyl elongation in cR. (A) qRT-PCR analysis showing the transcript levels of PIFs in 5-d-old cRgrown WT and hid1 seedlings. Expression of PIF1 in WT seedlings was set as 1. Error bars represent SD of triplicate biological replicates. **P < 0.01 (t test). (B and C) qRT-PCR analysis showing the transcript levels of PIF3 in 5-d-old WT and hid1 seedlings grown in various light conditions and darkness (B) and under a range of R light intensities (C). (D) PIF3 transcript levels in 5-d-old cR-grown WT, hid1, 35S:A/hid1, and pHID1:HID1 seedlings. (E and F) Phenotypes (E) and hypocotyl lengths (F) of 5-d-old WT, hid1, pif3, and hid1pif3 seedlings grown in cR. (Scale bar: 1 mm.) (G) Transcript abundances of PIF3 following actinomycin D treatment in WT and hid1 seedlings. (H) qRT-PCR analysis showing the expression of PIF3 in WT and hid1 seedlings grown in cR for 5 d and then transferred to dark for the indicated times. (I) Immunoblots showing PIF3 protein levels in WT and hid1 seedlings collected from the time-course in H. (J–L) qRT-PCR analysis showing the transcript levels of SNRK2.5 (J), XTR7 (K), and AT4G35720 (L) in WT and hid1 seedlings collected from the time-course in H. CSN6 was used as the loading control. *P < 0.05, **P < 0.01 (t test).

10362 | www.pnas.org/cgi/doi/10.1073/pnas.1409457111

expression directly. Because ncRNAs usually function in the form of RNA–protein complexes, we performed a gel filtration analysis on 5-d-old cR-grown WT and hid1 seedlings. We detected most HID1 in the fractions around 500 kDa in size, suggesting that it is assembled into large complex(es) (Fig. 5C). HID1 Associates with the PIF3 Promoter and Modulates Its Transcriptional Activity. Our detection of the majority of soluble HID1 on chromatin

suggested that HID1 could be closely associated with genomic DNA. Therefore we investigated whether HID1 is associated with the promoter of PIF3. To this end, we incorporated the S1 streptavidin-binding RNA aptamer tag into the SL3 stem-loop of HID1 (Fig. 5D). To examine the in vivo activity of this S1-tagged HID1, we transformed S1-HID1 driven by the HID1 promoter into the hid1 mutant background. This construct rescued the hid1 mutant phenotype in cR (Fig. 5 E–G), suggesting that S1-HID1 is biologically functional. Moreover, we were able to enrich S1-HID1 by immunoprecipitation with streptavidin-conjugated Sepharose beads (Fig. 5H). We then used this stable transgenic line to perform ChIPqPCR analysis to determine if and where HID1 binds to the PIF3 locus. Intriguingly, compared with ACT7 controls, we observed a significant enrichment in the first intron of the PIF3 5′UTR (amplicon 3 and 4) in pHID1:S1-HID1/hid1 seedlings relative to hid1 seedlings, suggesting that HID1 associates with the proximal promoter of PIF3 (Fig. 5 I and J). Next, to test the functional significance of this association, we examined the activity of the full-length PIF3 promoter sequence (WT-p) and the same sequence lacking the HID1-binding region (dI-p) using firefly luciferase as reporter in the WT and hid1 mutant plants. Firefly luciferase activity driven by the intact promoter was found to be two times greater in hid1 than in WT plants kept under cR (Fig. 5K). In contrast, no difference was observed in hid1 and WT seedlings kept under cFr or cB (Fig. 5K). Likewise, no difference in firefly luciferase activity was detected between WT and hid1 plants treated with the dI-p construct, suggesting that HID1 binding in the PIF3 promoter is required for the transcriptional regulation of PIF3 in response to cR (Fig. 5L). HID1 Is Evolutionarily Conserved in Land Plants. Interestingly, when we used the entire primary sequence of HID1 for BLASTN searches of the National Center for Biotechnology Information (www.ncbi.nih.gov) and Phytozome databases (www.phytozome. net), orthologs of Arabidopsis HID1 (AtHID1) were found only in the plant kingdom, ranging from moss to Arabidopsis (E < 10−5). ClustalW alignment of these sequences in five representative plant species from disparate taxa identified a highly conserved region (∼74 nt) near the 3′ end of AtHID1. The resulting secondary structures predicted by RNAalifold (http://rna.tbi.univie. ac.at/cgi-bin/RNAalifold.cgi) also were well conserved (Fig. 6A). The full-length transcripts of AtHID1 and of HID1 in rice (OsHID1) had been experimentally identified previously (32), and their expression was confirmed by Northern blot analysis (Fig. 6C). Moreover, the entire predicted secondary structures of OsHID1 and AtHID1 are notably similar. Specifically, SL4 was highly conserved, whereas SL2 was relatively conserved, further suggesting the functional significance of HID1 (Fig. 6B). Therefore we tested whether OsHID1 was functional in Arabidopsis by expressing OsHID1 from its own promoter in the hid1 mutant background. Interestingly, OsHID1 was highly expressed from its own promoter in Arabidopsis and rescued the hid1 elongated hypocotyl phenotype (Fig. 6 D–F). This finding suggests that the function of this structurally conserved ncRNA may be conserved in monocots and dicots.

Discussion Photomorphogenic development in seedlings has been studied extensively for decades, and the underlying mechanism has been established as primarily regulated proteolysis by the ubiquitin– proteasome pathway. In this study we introduced a lncRNA, HID1, as another layer of regulator and demonstrated that it functions in cR by regulating the transcription of key light intermediates in light Wang et al.

signaling exemplified by PIF3. To our knowledge, this report is the first example of an ncRNA functioning in light-mediated plant development. Mutant hid1 seedlings were observed to be hyposensitive to cR, exhibiting elongated hypocotyls, reduced cotyledon angles, and reduced chlorophyll accumulation (Fig. 1). Several lines of evidence indicate that HID1 acts as an in trans transcriptional repressor that associates with the PIF3 promoter and negatively regulates PIF3 transcription in Arabidopsis. This process, in turn, may explain how the red light–dependent function of HID1 is achieved. Disruption of HID1 results in a mild PIF3 overexpression phenotype, as shown by the elongated hypocotyls of the seedlings grown in cR (Fig. 4). The transcriptional regulation of PIF3 by HID1 represents both an elaborate control of the expression of a key transcription factor and an additional mechanistic layer of the light-signaling pathway. This function may be a way to coordinate transcription with the translation of PIF3 to maintain proper abundance of a key signaling mediator under cR. Wang et al.

Fig. 6. HID1 is a plant-specific ncRNA with conserved structure and function in monocots and dicots. (A) Sequence and structural alignment of AtHID1 in the five representative organisms prepared using RNAalifold showing a conserved region near its 3′ end. (B) Predicted secondary structures of AtHID1 and OsHID1. Conserved sequences are indicated in red. (C) Northern blot analysis of HID1 showing its expression in seedlings and leaves of Arabidopsis and rice respectively, with 5S rRNA as the loading control. (D) Northern blot analysis showing the expression of OsHID1 in pOsHID1:OsHID1/hid1 (OsHID1/hid1) transgenic plants. (E and F ) Phenotypes (E ) and hypocotyl lengths (F) of 5-d-old cR-grown WT, hid1, and pOsHID1:OsHID1/hid1 seedlings. (Scale bar: 2 mm.) Data are mean ± SD (n ≥ 20). **P < 0.01 (t test).

PNAS | July 15, 2014 | vol. 111 | no. 28 | 10363

PLANT BIOLOGY

Fig. 5. Nuclear-localized HID1 associates with the PIF3 promoter to repress PIF3 expression. (A) Northern blot analysis showing HID1 expression levels in purified nuclear (N) and nuclear-depleted (C) fractions extracted from 5-dold cR-grown seedlings. T, total extract. tRNA0119 was used as the marker for the cytoplasmic fraction. (B) Northern blot analysis showing the chromatin-bound fractions of HID1 (P1 or S2). tRNA0119 served as a marker for RNAs unbound to chromatin. (C) Gel filtration profiles of HID1 in cR-grown WT (Upper) and hid1 seedlings (Lower). The fraction numbers and molecular weights are indicated. T, total soluble extracts used for gel filtration. (D) Schematic illustration of HID1 tagged with the minimal S1 aptamer. SA. streptavidin Sepharose beads. (E and F) Phenotypes (E) and hypocotyl lengths (F) of 5-d-old cR-grown WT, hid1, and indicated transgenic lines. (Scale bar: 1 mm.) Data are mean ± SD (n ≥ 20). (G) Northern blot analysis showing the expression levels of S1-HID1 in two independent transgenic lines, with 5.8S rRNA as the loading control. (H) Northern blot analysis showing the expression levels of S1-HID1 in input and immunoprecipitation (IP) samples. (I) Diagram of the PIF3 amplicons used in the ChIP assays. (J) ChIP-qPCR analysis of S1-HID1 at the PIF3 locus. Data represent means of four biological replicates ± SD. *P < 0.05 (t test). (K) Transient dual-luciferase (Dual-LUC) transcription assay showing PIF3 promoter activity in WT and hid1 seedlings grown under the indicated light conditions. (L) Transient Dual-LUC transcription assay showing the differences between WT and truncated PIF3 promoter activity in WT and hid1 seedlings grown under cR. Luciferase/Renilla ratios in WT were normalized to 1. Data are means of triplicates ± SD. *P < 0.05, **P < 0.01 (t test).

Knowledge of lncRNAs’ functional motifs is essential to uncover their regulatory mechanisms. Genomic analyses of a variety of animal species have revealed that lncRNAs are significantly less conserved than protein-coding genes in animals. Furthermore, some, but not all, have clearly conserved regions that are both under selective pressure and functionally important (33, 34). To date, however, we have only a limited understanding of the structure and function of the lncRNAs found in plants. Phylogenetic analyses showed that HID1 occurs only in plants, indicating that it evolved after plants diverged from other eukaryotes and thus functions in processes unique to plants. Two pieces of evidence suggest that HID1 functions as part of a complex structure rather than as a precursor of a small RNA. First, our research has demonstrated that two unique modules of HID1, SL2 and SL4, are both essential for its function in cR. This finding, in turn, suggests that more than one structural feature of HID1 is required for its function (Fig. 3). Second, our BLAST searches of the available Arabidopsis small RNA databases did not reveal the existence of any small RNA that could be derived from SL2 or SL4, suggesting that HID1 is unlikely to function as a precursor of a small RNA. Thus, it appears that HID1 may be a good place to begin an investigation of the structures and functions of lncRNAs in plants. Further study of the structure of HID1 should yield new insights into the mechanisms by which lncRNAs function. Case studies of lncRNAs in animals have shown that these RNAs regulate transcription by binding specific sites in trans in the nucleus. Many lncRNAs that bind chromatin to regulate the chromatin functional state have been identified in animals, including Drosophila roX1 and roX2 RNA (35), human HOTAIR (36), XIST (37), and mammalian pRNA (38). However, our identification of Arabidopsis HID1 is, to our knowledge, the first

example of an lncRNA that associates with chromatin at a distant gene locus in land plants. Given both the abundant expression of HID1 detected in numerous tissues during plant development and the pleiotropic phenotypes observed in hid1 mutants, we postulate that HID1 may have other targets in addition to PIF3. In fact, it seems likely that HID1 may associate with multiple chromatin targets. Identification of HID1’s chromatin occupancy and that of its partners in the RNA–protein nuclear complex(es) undoubtedly will provide new insights into the details of this specific lncRNA’s function. More broadly, this identification may uncover evolutionarily divergent regulation of lncRNAs in plants and animals. In addition, we anticipate that further studies connecting the RNA-based gene-regulatory network to the known protein-based mechanisms of gene regulation should lead to a better understanding of both RNA biology and the regulation of gene expression in general. 1. Jiao Y, Lau OS, Deng XW (2007) Light-regulated transcriptional networks in higher plants. Nat Rev Genet 8(3):217–230. 2. Chen M, Chory J (2011) Phytochrome signaling mechanisms and the control of plant development. Trends Cell Biol 21(11):664–671. 3. Leivar P, et al. (2009) Definition of early transcriptional circuitry involved in light-induced reversal of PIF-imposed repression of photomorphogenesis in young Arabidopsis seedlings. Plant Cell 21(11):3535–3553. 4. Hu W, Su YS, Lagarias JC (2009) A light-independent allele of phytochrome B faithfully recapitulates photomorphogenic transcriptional networks. Mol Plant 2(1): 166–182. 5. Tepperman JM, Hwang YS, Quail PH (2006) phyA dominates in transduction of red-light signals to rapidly responding genes at the initiation of Arabidopsis seedling de-etiolation. Plant J 48(5):728–742. 6. Ma L, et al. (2001) Light control of Arabidopsis development entails coordinated regulation of genome expression and cellular pathways. Plant Cell 13(12):2589–2607. 7. Chen M, Chory J, Fankhauser C (2004) Light signal transduction in higher plants. Annu Rev Genet 38:87–117. 8. Heijde M, Ulm R (2012) UV-B photoreceptor-mediated signalling in plants. Trends Plant Sci 17(4):230–237. 9. Li J, Terzaghi W, Deng XW (2012) Genomic basis for light control of plant development. Protein Cell 3(2):106–116. 10. Huq E, Quail PH (2002) PIF4, a phytochrome-interacting bHLH factor, functions as a negative regulator of phytochrome B signaling in Arabidopsis. EMBO J 21(10): 2441–2450. 11. Kim J, et al. (2003) Functional characterization of phytochrome interacting factor 3 in phytochrome-mediated light signal transduction. Plant Cell 15(10):2399–2407. 12. Khanna R, et al. (2007) The basic helix-loop-helix transcription factor PIF5 acts on ethylene biosynthesis and phytochrome signaling by distinct mechanisms. Plant Cell 19(12):3915–3929. 13. Leivar P, et al. (2008) Multiple phytochrome-interacting bHLH transcription factors repress premature seedling photomorphogenesis in darkness. Curr Biol 18(23): 1815–1823. 14. Shin J, et al. (2009) Phytochromes promote seedling light responses by inhibiting four negatively-acting phytochrome-interacting factors. Proc Natl Acad Sci USA 106(18): 7660–7665. 15. Leivar P, Quail PH (2011) PIFs: Pivotal components in a cellular signaling hub. Trends Plant Sci 16(1):19–28. 16. Zhang Y, et al. (2013) A quartet of PIF bHLH factors provides a transcriptionally centered signaling hub that regulates seedling morphogenesis through differential expression-patterning of shared target genes in Arabidopsis. PLoS Genet 9(1): e1003244. 17. Bai MY, et al. (2012) Brassinosteroid, gibberellin and phytochrome impinge on a common transcription module in Arabidopsis. Nat Cell Biol 14(8):810–817. 18. Wang KC, Chang HY (2011) Molecular mechanisms of long noncoding RNAs. Mol Cell 43(6):904–914.

10364 | www.pnas.org/cgi/doi/10.1073/pnas.1409457111

Materials and Methods Plant materials and growth conditions, phenotype analyses, plasmid construction, and the generation of transgenic plants are described in SI Materials and Methods. The detailed procedures of Northern blot analysis, qRT-PCR, RNA-Seq, mRNA decay assay, Western blot analysis, biochemical fractionation assay, gel filtration chromatography, RNA immunoprecipitation assay, ChIP assay, and transient transcription dual-luciferase (Dual-LUC) assay are provided in SI Materials and Methods. The sequences used for probing ncRNAs in this study are listed in Table S1. ACKNOWLEDGMENTS. We thank Dr. Haiyang Wang, Dr. Yijun Qi, and Dr. Ligeng Ma for their helpful discussions and comments on this project; Wei Chen and Xuncheng Wang for bioinformatics assistance; Huikun Duan and Junjie Ling for technical assistance; and Abigail Coplin for critical reading of the manuscript. This work was supported by National Natural Science Foundation of China Grant 31171156, National Basic Research Program of China (973 Program) Grant 2012CB910900, and in part by the State Key Laboratory of Protein and Plant Gene Research at Peking University and the PekingTsinghua Center for Life Sciences.

19. Guttman M, Rinn JL (2012) Modular regulatory principles of large non-coding RNAs. Nature 482(7385):339–346. 20. Geisler S, Coller J (2013) RNA in unexpected places: Long non-coding RNA functions in diverse cellular contexts. Nat Rev Mol Cell Biol 14(11):699–712. 21. Kim ED, Sung S (2012) Long noncoding RNA: Unveiling hidden layer of gene regulatory networks. Trends Plant Sci 17(1):16–21. 22. De Lucia F, Dean C (2011) Long non-coding RNAs and chromatin regulation. Curr Opin Plant Biol 14(2):168–173. 23. Zhang YC, Chen YQ (2013) Long noncoding RNAs: New regulators in plant development. Biochem Biophys Res Commun 436(2):111–114. 24. Franco-Zorrilla JM, et al. (2007) Target mimicry provides a new mechanism for regulation of microRNA activity. Nat Genet 39(8):1033–1037. 25. Heo JB, Sung S (2011) Vernalization-mediated epigenetic silencing by a long intronic noncoding RNA. Science 331(6013):76–79. 26. Swiezewski S, Liu F, Magusin A, Dean C (2009) Cold-induced silencing by long antisense transcripts of an Arabidopsis Polycomb target. Nature 462(7274):799–802. 27. Cifuentes-Rojas C, Kannan K, Tseng L, Shippen DE (2011) Two RNA subunits and POT1a are components of Arabidopsis telomerase. Proc Natl Acad Sci USA 108(1): 73–78. 28. Ding J, et al. (2012) A long noncoding RNA regulates photoperiod-sensitive male sterility, an essential component of hybrid rice. Proc Natl Acad Sci USA 109(7): 2654–2659. 29. Zhou H, et al. (2012) Photoperiod- and thermo-sensitive genic male sterility in rice are caused by a point mutation in a novel noncoding RNA that produces a small RNA. Cell Res 22(4):649–660. 30. Wang Y, et al. (2014) Genomic features and regulatory roles of intermediate-sized non-coding RNAs in Arabidopsis. Mol Plant 7(3):514–527. 31. Ji X, et al. (2013) SR proteins collaborate with 7SK and promoter-associated nascent RNA to release paused polymerase. Cell 153(4):855–868. 32. Liu TT, et al. (2013) A global identification and analysis of small nucleolar RNAs and possible intermediate-sized non-coding RNAs in Oryza sativa. Mol Plant 6(3):830–846. 33. Chodroff RA, et al. (2010) Long noncoding RNA genes: Conservation of sequence and brain expression among diverse amniotes. Genome Biol 11(7):R72. 34. Ulitsky I, Shkumatava A, Jan CH, Sive H, Bartel DP (2011) Conserved function of lincRNAs in vertebrate embryonic development despite rapid sequence evolution. Cell 147(7):1537–1550. 35. Franke A, Baker BS (1999) The rox1 and rox2 RNAs are essential components of the compensasome, which mediates dosage compensation in Drosophila. Mol Cell 4(1): 117–122. 36. Gupta RA, et al. (2010) Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 464(7291):1071–1076. 37. Simon MD, et al. (2013) High-resolution Xist binding maps reveal two-step spreading during X-chromosome inactivation. Nature 504(7480):465–469. 38. Schmitz KM, Mayer C, Postepska A, Grummt I (2010) Interaction of noncoding RNA with the rDNA promoter mediates recruitment of DNMT3b and silencing of rRNA genes. Genes Dev 24(20):2264–2269.

Wang et al.