ABSTRACT
Plant long noncoding RNAs (IncRNAs) have emerged as important regulators of chromatin dynamics, impacting on transcriptional programs leading to different developmental outputs. The lncRNA AUXIN REGULATED PROMOTER LOOP (APOLO) directly recognizes multiple independent loci across the Arabidopsis genome and modulates their three-dimensional chromatin conformation, leading to transcriptional shifts. Here, we show that APOLO recognizes the locus encoding the root hair (RH) master regulator ROOT HAIR DEFECTIVE 6 (RHD6) and controls RHD6 transcriptional activity leading to cold-enhanced RH elongation, in association with the Polycomb Repressive Complexes (PRC) 1 and 2. Additionally, we demonstrate that APOLO interacts with the transcription factor WRKY42 and modulates its binding to the RHD6 promoter. WRKY42 is required for the activation of RHD6 by low temperatures and WRKY42 deregulation impairs cold-induced RH expansion. Collectively, our results indicate that a novel ribonucleoprotein complex involving APOLO and WRKY42 forms a regulatory hub which activates RHD6 by shaping its epigenetic environment and integrates different signals governing RH growth and development.
INTRODUCTION
Root hairs (RHs) are single cell projections developed from specialized epidermal trichoblast cells able to increase their size several hundred times in a polar manner to reach and promote the uptake of water-soluble nutrients, interact with microorganisms and anchor the plant to the soil. The specification of epidermal cells into RHs involves a cell fate determination process whose underlying mechanisms are only partially understood. In Arabidopsis thaliana, RH cell fate is controlled by a developmental program involving a complex of transcription factors (TFs) promoting the expression of the homeodomain protein GLABRA 2 (GL2)(1–4). GL2 blocks RH development by inhibiting the transcription of the master regulator ROOT HAIR DEFECTIVE 6 (RHD6)(5). In the trichoblast cells that differentiate into RHs, a second TF complex suppresses GL2 expression (3), forcing the cells to enter the RH cell fate program via the concomitant activation of RHD6 along with downstream TFs (6,7). Briefly, RHD6 together with its homolog RHD6-LIKE 1 (RSL1) induce the expression of other TFs from the bHLH family, including RSL2 and RSL4, ultimately triggering the differentiation of the RHs and their subsequent polarized tip-growth (8–10). In addition, it was proposed that RSL4 controls the expression of a small subset of nearly 125 genes (9, 11–13), including several cell wall extensins (EXTs) (14, 15) sufficient to promote RH growth (16).
RH expansion is regulated both by cell-intrinsic factors (e.g. endogenous phytohormones such as auxin) and external environmental signals (e.g. phosphate (Pi) availability in the soil) (17, 18). Pi starvation is one of the key environmental factors promoting rapid RH growth (9, 12, 13). In Arabidopsis, it triggers RSL4 expression via an enhanced auxin production, activating downstream effector genes mediating cell growth (9, 12, 17–19). Accordingly, several auxin-related TFs have been implicated in Pistarvation signaling in roots, including WRKY proteins that control the expression of the Pi transporter families Pi-permease PHO1 and PHOSPHATE TRANSPORTER (PHT) (20–23). Under Pi-sufficient conditions, WRKY6 and WRKY42 bind to W-boxes of the PHO1 promoter and suppress its expression. During Pi starvation, WRKY42 is degraded by the 26S proteasome pathway, resulting in the activation of PHO1 transcription (21, 23). In addition, WRKY42 functions as a positive regulator of PHT1;1, by binding to its promoter under Pi-sufficient condition (23). Overall, WRKY42 is part of the components activating root early-responses to Pi starvation, although its role in controlling RH growth remains unexplored.
In recent years, plant long noncoding RNAs (lncRNAs) have emerged as important regulators of gene expression, and several among them have been functionally linked to Pi homeostasis. For instance, the lncRNA INDUCED BY PHOSPHATE STARVATION 1 (IPS1) can sequester the Pi starvation-induced microRNA miR-399, attenuating miR-399-mediated repression of PHO2, a gene involved in Pi uptake (24). In addition, the cis-natural antisense (cis-NAT) transcript PHO1;2, induced under Pi deficiency, was shown to promote the translation of the PHO1;2 mRNA involved in Pi loading into the xylem. The expression of this cis-NAT is associated with the transport of the sense–antisense RNA pair toward the polysomes (25). More recently, it was shown that the lncRNA AUXIN REGULATED PROMOTER LOOP (APOLO) recognizes multiple spatially independent genes by sequence complementarity and DNA-RNA duplex formation, known as R-loops. Upon recognition, APOLO shapes the three-dimensional (3D) conformation of its target regions by decoying the Polycomb Repressive Complex 1 (PRC1) component LIKE HETEROCHROMATIN PROTEIN 1 (LHP1), thereby regulating their transcription (26, 27).
Here, we show that the lncRNA APOLO directly regulates a subset of genes involved in RH development, including the master regulator of RH initiation RHD6. APOLO activates RHD6 transcription by modulating the formation of a local chromatin loop encompassing its promoter region, an epigenetic regulatory mechanism likely involving PRC1 and PRC2 components. Furthermore, we found that APOLO interacts with the TF WRKY42, forming a new hub that regulates RHD6 to induce RH growth in response to low temperatures.
RESULTS
APOLO regulates root hair cell elongation in response to low temperatures
The lncRNA APOLO was previously reported to recognize a subset of independent loci enriched in categories related to cell wall composition and organization (27). A closer look at APOLO bona fide targets allowed us to identify seventeen genes involved in RH growth and expansion (Supplementary Table 1), a process dependent on cell wall remodeling molecules, including EXTs and EXT-related proteins (14, 15, 28–30). Interestingly, according to single-cell RNA-seq datasets (31) APOLO transcripts are enriched in RH cells (Supplementary Figure 1). Notably, sixteen APOLO direct targets were upregulated and one downregulated (EXT18) upon APOLO over-expression (Supplementary Table 2). Furthermore, 52 additional RH-related genes were upregulated in 35S:APOLO seedlings, albeit not identified as APOLO direct targets (Supplementary Table 2) (27). Among them, the RH central TF genes RHD6 (as a direct target), RSL2 and RSL4 (as indirectly regulated) were induced upon APOLO over-expression.
It was reported that the APOLO locus is targeted by the RNA-polymerase Pol V and silenced by RNA-directed DNA Methylation (RdDM, (26)). A search in a small RNA-Seq performed in WT roots subjected to different temperature treatments (32) revealed that RdDM-related 24nt siRNA accumulation over the APOLO locus is less abundant at low temperatures (15°C; Figure 1A), suggesting that APOLO transcription is regulated by cold. Accordingly, we found that APOLO transcriptional accumulation increases in roots after 24h at 10°C (Figure 1B). An analysis of the promoter activity of the 5.2kb region upstream APOLO (27) directing the expression of a GFP reporter gene, additionally revealed a higher transcriptional activity at low temperatures in the RHs (Figure 1C). Strikingly, we observed that two RNAi-APOLO and two 35S:APOLO independent lines (26, 27) exhibit a basal increase of RH length at 22°C, and uncovered a strong induction of RH elongation in WT and RNAi-APOLO at 10°C, in contrast to 35S:APOLO lines (Figure 1D). Accordingly, RHD6 is induced in response to cold in WT roots and RNAi-APOLO roots display higher RHD6 basal levels than the WT (Figure 1E). Collectively, our findings suggest that APOLO participates in the induction of cold-mediated RH elongation and that a deregulation of APOLO transcript levels can impact RH growth.
Previous studies pointed out a key role of RHD6 (together with RSL1) in RH development, which is mediated by RSL4 and RSL2 as downstream regulators of RH cell elongation (6, 7). Considering that RHD6, RSL2 and RSL4 transcript levels were upregulated in 35S:APOLO seedlings (Supplementary Table 2; (27)), we assessed if these TFs were also controlling the promotion of RH growth by low temperatures. To this end, we tested how rhd6/rsl1/rsl4 and rsl2, rsl4 and double mutant plants rsl2/rsl4 respond to low temperatures in comparison with control conditions. The rsl2 mutant was highly responsive to low temperatures in a similar manner to WT while rsl4 was impaired in the response to cold. The double mutant rsl2/rsl4 and the triple mutant rhd6/rsl1/rsl4 did not develop RHs in either of the two conditions (Supplementary Figure 2A). In addition, constitutive expression of RSL4 (35S:RSL4) as well as its expression under the control of the RH specific EXPANSIN7 promoter (EXP7p:RSL4) boosted basal RH growth without further enhancement in response to cold (Supplementary Figure 2B). These results demonstrate that RSL4 is largely required for RHD6-dependent activation of RH growth at low temperatures, and RSL2 might participate to a lower extent.
APOLO directly modulates the three-dimensional chromatin conformation of the root hair specific locus RHD6
Among APOLO targets involved in RH development, we found the master regulator of RH initiation RHD6 (6, 7). The epigenetic profile of the RHD6 locus corresponds to typical APOLO targets (Figure 2A, (27)), including H3K27me3 deposition (track 1), LHP1 recognition (track 2, chromatin immunoprecipitation (ChIP)-Seq, (33)), and APOLO binding regions (tracks 3 to 5, chromatin isolation by RNA purification (ChIRP)-Seq, (27)). A GAAGAA box, shown to be important for APOLO target recognition (27) is located in the RHD6 locus and coincides with APOLO binding site. In addition, a small peak indicates the presence of an R-loop coinciding with APOLO recognition sites over RHD6 (tracks 6 to 9, DNA-RNA hybrid immunoprecipitation (DRIP)-Seq (34)). The possibility that RHD6 recognition by APOLO occurs specifically in RHs may explain the dilution of the signal coming from entire seedlings (34).
Remarkably, APOLO recognition and R-loop formation are also detectable over RHD6 neighbor gene, located 3.2 kb upstream RHD6 transcription start site (Figure 2A). According to DpnII Hi-C datasets from Arabidopsis seedlings (35), a chromatin loop encompassing the intergenic region upstream RHD6 was detected (Figure 2B), exhibiting APOLO binding at both sides of the loop base (Figure 2A, ChIRP-Seq). By performing a ChIRP-qPCR with two independent sets of biotinylated probes to purify APOLO (ODD and EVEN; (27)) and one additional set used as a negative control (LacZ), we confirmed that APOLO RNA-RHD6 DNA interaction occurs in wild-type (WT) and is lost in APOLO knockdown (RNAi) seedlings ((26); Figure 2C). In addition, the quantification of relative RHD6 loop formation in RNAi-APOLO and APOLO over-expressing (35S:APOLO; (27)) seedlings, revealed impaired loop formation in both lines (Figure 2D), hinting at a stoichiometric requirement of APOLO for RHD6 chromatin loop formation. Accordingly, an RNA-Seq dataset of 35S:APOLO seedlings vs. WT (27) indicates that RHD6 transcript levels are increased upon APOLO over-expression (Figure 2E, Supplementary Table 1), suggesting that the chromatin loop including RHD6 promoter region precludes transcription. Altogether, our results indicate that APOLO lncRNA directly regulates RHD6 transcriptional activity by fine-tuning local chromatin 3D conformation.
It was previously reported that PRC2 actively participates in the regulation of RH growth (36) and that the RHD6 locus exhibits H3K27me3 deposition and LHP1 recognition (Figure 2A; (33)). Thus, we decided to explore the role of PRC1 and 2 in APOLO-mediated RHD6 activation at low temperatures. At 22°C, RHD6 suffers a reduction of H3K27me3 in the PRC2 mutant curly leaf (clf), in contrast to the PRC1 mutant lhp1 (Supplementary Figure 3A; (33)). Interestingly, we observed that H3K27me3 deposition and LHP1 binding diminish in WT roots treated for 24h at 10°C compared to 22°C (Supplementary Figure 3B), in agreement with the induction of RHD6 in response to cold (Figure 2E). Moreover, lhp1 and clf mutants exhibit a basal decrease of RH length together with a slight decrease of cold-induced RH elongation in lhp1, and a strong decrease of cold-induced RH elongation in clf (Supplementary Figure 3D). Consistently, RHD6 transcriptional activation by cold is abolished in the clf mutant (Supplementary Figure 3C), hinting at an important role of chromatin rearrangement for RHD6 activation in response to cold.
APOLO interacts with the transcription factor WRKY42 to coordinate the activation of RHD6
With the aim of identifying novel actors involved in cold-induced transcriptional regulation of RH growth, we analyzed the sequence of the promoter regions of RH-related APOLO targets. Notably, 13 out of the 17 APOLO targets contained between 1 and 4 canonical WRKY TF binding sites (W-box) in their promoters, including RHD6 (Supplementary Table 1). According to the Arabidopsis eFP Browser (37), the TF-encoding gene WRKY42 is induced in roots when seedlings are subjected to 4°C for 24h (Figure 3A). Remarkably, WRKY42 is involved in the response to Pi starvation (23), an environmental condition that promotes RH cell expansion (18) in a similar manner to low temperatures. At 10°C, we observed that WRKY42 transcriptional accumulation augments significantly in roots (Figure 3B). By using a 35S:WRKY42:GFP line, we demonstrated that WRKY42 can directly bind to the promoter region of RHD6 (Figure 3C). Accordingly, the over-expression of WRKY42 (35S:WRKY42:GFP line) led to a basal increase of RHD6 levels (Figure 3D) and RH elongation (Figure 3E) at ambient temperature, mimicking the effect of cold. On the contrary, cold-mediated induction of RHD6 is abolished in the wrky42 mutant (23; Figure 3D), which consistently exhibits shorter RHs at 22°C and partially impaired RH elongation at low temperatures (Figure 3E). Taken together, our results suggest that WRKY42 is an important regulator of RHD6- mediated RH growth in response to cold.
We thus wondered to what extent WRKY42 regulates the epigenetic landscape of the RHD6 locus. We first observed that H3K27me3 deposition over RHD6 is significantly augmented in the wrky42 mutant background, in contrast to AZG2, an APOLO target non-related to WRKY42 (Figure 4A), consistent with reduced RHD6 basal levels reported in wrky42 (Figure 3D). Considering that other TFs were shown to directly interact with lncRNAs in animals (38–40), we wondered if WRKY42 can recognize APOLO in planta. Thus, APOLO-WRKY42 interaction was assessed and validated by RNA immuno-precipitation (RIP-qPCR) in tobacco leaves and in Arabidopsis plants transitory or stably transformed with 35S:WRKY42:GFP, respectively (Figure 4B). Therefore, we evaluated the mutual contribution of APOLO and WRKY42 to their respective recognition of the RHD6 locus. APOLO ChIRP-qPCR in the WT and wrky42 mutant (Figure 4C) revealed similar binding to RHD6, indicating that WRKY42 does not participates in APOLO-target recognition. Reciprocally, we assessed the control of APOLO over WRKY42 recognition of the RHD6 locus. To this end, chromatin was extracted from 35S:WRKY42:GFP seedlings and increasing amounts of in vitro transcribed APOLO were added before cross-link and regular WRKY42 ChIP over RHD6. Strikingly, increasing concentrations of APOLO gradually decoy WRKY42 away from the RHD6 locus (Figure 4D), hinting at a stoichiometric regulation of APOLO over the activity of its partner TF.
Collectively, our results indicate that the regulation of RHD6 expression in response to cold depends on Polycomb-dependent H3K27me3 dynamic deposition. The WRKY42-APOLO complex modulates the epigenetic environment of RHD6, activating its transcription and promoting RH growth at low temperatures. RHD6 activation further triggers the expression of RSL2 and RSL4 that control the transcriptional RH program inducing cell expansion in response to cold (Figure 4E).
DISCUSSION
Cell fate determination in the epidermis has been extensively studied (1–4). Once trichoblast cells differentiate in the root epidermis, RHs develop as fast polar growing protuberances in response to endogenous and environmental signals (6, 7, 9, 19). RHs are one of the main entry points in the roots for water-soluble macronutrients, such as Pi and nitrates. Pi is an essential element for plant growth and development, and the availability of this macronutrient is a factor limiting plant productivity. Low Pi in the soil triggers auxin synthesis and transport in Arabidopsis roots, enhancing RH elongation to promote Pi uptake (18). Thus, auxin mediates low Pi-induced promotion of RH cell expansion. Under low soil Pi, auxin synthesis is enhanced specifically in the root cap (41) and transported (mostly by AUX1, PIN2, and PGP4) from the apex to the differentiation zone, specifically leading to an increase of auxin levels in trichoblasts (18, 42, 43). In response to the high-auxin microenvironment, RHs protrude from the root epidermis controlled by RHD6 and RSL1 (6, 7). High levels of auxin in trichoblasts trigger a signaling cas-cade mediated by TIR1-ARF19 (and possibly also ARF7) that directly induce the expression of RSL4 (and likely of RSL2) and promote RH elongation (9, 17, 18, 44). ARF7 and ARF19 also activate other RH genes independently of RSL4 (45). Interestingly, our results indicate that the lncRNA APOLO participates in the response to low temperatures. APOLO is directly activated by ARF7 and regulates the transcriptional activity of its neighboring gene PINOID (PID) by shaping local 3D chromatin conformation (26, 27). PID encodes a kinase responsible for accurate auxin polar transport by localizing PIN2 in the root cell membrane (46). More recently, it was shown that APOLO can recognize a subset of distant genes across the Arabidopsis genome, most of them being related to auxin synthesis and signaling (27). In this work, we demonstrate that a group of RH related genes are directly regulated by APOLO in response to cold, including the RH master regulator RHD6. Collectively, our results uncover a lncRNA-mediated epigenetic link between environmental signals and auxin homeostasis modulating RH growth.
As mentioned above, nutrient availability is known to activate RH expansion through a transcriptional reprogramming governed by RHD6 and downstream TFs. The quantification of RH growth of WT plants in response to increasing concentrations of nutrients (0.5X to 2.0X MS (Murashige and Skoog) medium) indicates that high concentrations of nutrients completely abolish RH growth triggered by low temperatures (Supplementary Figure 4A). In a similar way, by increasing agar concentration in the MS medium (from 0.8% to 2.5%) restraining nutrient mobility, cold-induced RH is clearly blocked (Supplementary Figure 4B). Altogether, these observations suggest that low temperatures may restrict nutrient mobility and availability in the culture medium, leading to the promotion of polar RH growth. Further research will be needed to determine what is the limiting nutrient mediating the effect of cold on RH growth.
Although substantial progress has been achieved in the identification of the molecular actors that control RH development, the impact of chromatin conformation in the transcriptional regulation of central TFs remains poorly understood. In this study, we have revealed a new mechanism of gene regulation in RHs by which the lncRNA APOLO integrates chromatin-associated ribonucleoprotein complexes together with the TF WYRK42, participating in the transcriptional activation of RHD6 and the down-stream RH gene network (Figure 4D). APOLO directly regulates the chromatin 3D conformation of the genomic region encompassing the RHD6 locus and stoichiometrically recruits WYRK42, previously linked to Pi-starvation (21, 23). Our results suggest that an APOLO-WRKY42 hub regulates RH cell elongation through the master regulator RHD6, although the APOLO-WYRKY42 hub potentially targets several additional cell wall related genes (Supplementary Table 1) at the end of the pathway controlled by RHD6 and the RHD6-downstream TFs RSL2/RSL4 (17, 47).
Participation of epigenetic factors in root cell identity determination strongly suggests that the default pattern for epidermal cell fate can be overridden by environmental stimuli (48). Interestingly, it was reported that the expression of GLABRA2 (GL2), a gene encoding a TF repressing RHD6 in atrichoblasts, is tightly regulated at the epigenetic level. By using 3D fluorescence in situ hybridization, it was shown that alternative states of chromatin organization around the GL2 locus are required to control position-dependent cell-type specification in the root epidermis (49). Furthermore, GL2 epigenetic regulation was proposed to be responsive to salt stress (50). In addition, a comprehensive characterization of alternative mutant lines uncovered the role of PRC2 in the regulation of RH development (36). Loss-of-function mutants in different PRC2 subunits develop unicellular RHs but fail to retain the differentiated state, generating a disorganized cell mass from each single RH. It was shown that the resulting RHs are able to undergo a normal endoreduplication program, increasing their nuclear ploidy, although they subsequently reinitiate mitotic division and successive DNA replication. It was proposed that aberrant RH development in PRC2-related mutants is due to the epigenetic deregulation of key genes such as WOUND INDUCED DEDIFFERENTIATION 3 (WIND3) and LEAFY COTYLEDON 2 (LEC2) (36). Here, we showed that the single mutants clf (PRC2) and lhp1 (PRC1) are affected in RH growth. Moreover, we showed that H3K27me3 deposition throughout the RHD6 locus is partially impaired in the clf back-ground, which is affected in RH elongation promoted by cold. Altogether, our results suggest that Polycomb proteins participate in the control of RH-related genes transcriptional reprogramming at low temperatures.
Notably, CLF and LHP1 were shown to interact with a subset of lncRNAs in Arabidopsis, modulating the activity of PRC target genes (51). Interestingly, several lncRNAs have been linked to the control of transcription in response to cold. FLOWERING LOCUS C is regulated by at least three lncRNAs. First, the alternative splicing of a set of antisense transcripts, collectively named as COOLAIR, depends on the prolonged exposure to cold, epigenetically repressing FLC (52). The use of the COOLAIR proximal poly(A) site results in down-regulation of FLC expression in a process involving FLOWERING LOCUS D (FLD), an H3K4me2 demethylase (53). A second lncRNA called COLD ASSISTED INTRONIC NONCODING RNA (COLDAIR) is fully encoded in the first intron of FLC. Similar to COOLAIR, its accumulation oscillates in response to low temperatures. It was proposed that COLDAIR recruits the PRC2 component CLF to target FLC for H3K27me3 deposition (54). More recently, a third lncRNA modulating FLC transcription was identified (55). The cold-responsive lncRNA COLDWRAP is derived from the FLC proximal promoter and it also interacts with PRC2. It was suggested that COLDWRAP functions in cooperation with the lncRNA COLDAIR to retain Polycomb at the FLC promoter through the formation of a repressive intragenic chromatin loop (55). Another lncRNA named SVALKA was shown to mediate the response to low temperatures (56). Interestingly, the activation of SVALKA by cold triggers the transcription of a cryptic down-stream lncRNA, which overlaps the antisense locus of the C-repeat/dehydration-responsive element Binding Factor 1 (CBF1), involved in the early response to cold in Arabidopsis. Antisense transcription causes Pol II head-to-head collision modulating transcriptional termination of CBF1 (56). Here, we showed that the auxin-responsive lncRNA APOLO is also transcriptionally modulated by cold. The differential accumulation of 24nt siRNAs across the APOLO locus at low temperatures indicates that this activation is related to a decrease in RdDM. Moreover, we showed here that the intergenic region between PID and APOLO acting as a divergent promoter is also activated at low temperatures in RHs, as revealed by the reporter gene GFP. Thus, the lncRNA APOLO integrates external signals into auxin-dependent developmental outputs in Arabidopsis.
In the last decade, lncRNAs have emerged as regulators of gene expression at different levels, ranging from epigenetics to protein modifications and stability (57). Notably, it has been shown in animals that noncoding transcripts can be recognized by TFs. In humans, it was proposed that the interaction with the lncRNA SMALL NUCLEOLAR RNA HOST GENE 15 (SNHG15) stabilizes the TF Slug in colon cancer cells. It was shown that SNHG15 is recognized by the zinc finger domain of Slug preventing its ubiquitination and degradation in living cells (38). Also, the transcriptional activity of the human gene DIHYDROFOLATE REDUCTASE (DHFR) is regulated by a lncRNA encoded in its proximal promoter. It was proposed that the nascent noncoding transcript forms a hybrid with its parent DNA and decoys the regulatory TF IIB away from the DHFR promoter, dissociating the transcriptional pre-initiation complex in quiescent cells (40). The lncRNA P21 ASSOCIATED ncRNA DNA DAMAGE ACTIVATED (PANDA) was identified in human cancer and it was activated in response to DNA damage (39). PANDA is transcribed from the promoter region of the CDKN1A gene and interacts with the TF NF-YA to limit the expression of pro-apoptotic genes. The activity of PANDA has been linked to the progression of different tumors (58, 59). Interestingly, it was shown that in addition to NF-YA, PANDA interacts with the scaffold-attachment-factor A (SAFA) as well as PRC1 and PRC2 to modulate cell senescence. In proliferating cells, SAFA and PANDA recruit Polycomb components to repress the transcription of senescence-promoting genes. Conversely, the loss of SAFA–PANDA–PRC interactions allows expression of the senescence program (60). In this work, we showed that the PRC1-interacting lncRNA APOLO can also be recognized by the TF WRKY42, hinting at general lncRNA-mediated mechanisms linking Polycomb complexes with the transcriptional machinery across kingdoms. Furthermore, our observations indicate that the deregulation of WRKY42 affects the epigenetic environment of RHD6. It was previously shown that the addition of in vitro transcribed APOLO to RNAi-APOLO chromatin extracts was able to partially restore R-loop formation over APOLO target genes, and that high levels of APOLO may titer LHP1 away from chromatin (27). Here we showed that the relative accumulation of the lncRNA APOLO can modulate the binding activity of its partner TF to common target genes. Collectively, our results strongly support that environmentally controlled cell fate in Arabidopsis relies on a transcriptional reprogramming governed by a network of epigenetic regulatory complexes, lncRNAs, TFs and effector proteins.
MATERIALS AND METHODS
Plant Material and Growth Conditions
All the Arabidopsis thaliana lines used were in the Columbia-0 (Col-0) background. WRKY42 over expression transgenic plants were generated through Agrobacterium tumefaciens (strain EHA105)-mediated transformation (61). 35S:WRKY42:GFP transformant lines were selected on MS/2 medium supplemented with kanamycin (40μg/mL) and WRKY42 expression levels were measured by RT-qPCR (primers used are listed in Supplementary Table 3). The wrky42 mutant line belongs to the SALK collection (SALK_121674C), as the one previously characterized (23). Homozygous plants were obtained in our laboratory and genotyped using the oligonucleotides indicated in Supplementary Table 3. Seeds were surface sterilized and stratified at 4°C for 2d before being grown under long day conditions (16h light, 140μE.m-2.sec-1/8h dark), on ½-strength Murashige and Skoog media (1/2 MS) (Duchefa, Netherlands) with 0.8% plant agar (Duchefa, Netherlands).
Cloning procedure
The coding region of WRKY42 (AT4G04450) excluding the STOP codon was amplified by PCR, cloned into the Gateway entry vector pENTR/D-TOPO (Invitrogen) and recombined by Gateway technology (LR reaction) into the pK7FWG2,0 vector containing a p35S-GFP cassette (http://www.psb.ugent.be/gateway/index.php).
Root hair phenotype characterization
For quantitative analyses of RH phenotypes, 100 fully elongated RHs were measured (using the ImageJ software) from 10-20 roots grown on vertical agar plates for 5 days at 22° and 3 days at 10°C. Measurements were made after 8 days. Images were captured with an Olympus SZX7 Zoom Stereo microscope (Olympus, Japan) equipped with a Q-Colors digital camera and software.
Confocal microscopy analysis of root hairs
Confocal laser scanning microscopy was performed using Zeiss LSM5 Pascal (Zeiss, Germany) and a 40x water-immersion objective, N/A=1.2. Fluorescence was analyzed by using 488 nm laser for GFP excitation (Laser Intensity: 70%, Detector Gain:550, Amplifier Offset:0.1, Amplifier Gain:1), and emitted fluorescence was recorded between 490 and 525nm for GFP tag. Z stacks were done with an optical slice of 1μm, and fluorescence intensity was measured in 15μm ROI (Region Of Interest) at the RH tip and summed for quantification of fluorescence using ImageJ. Five replicates for each of ten roots and 15 hairs per root were observed. Col-0 wild type root hairs were used as a negative control, to check autofluorescence signal occurrence and no signal were detected in the wavelengths range stated above.
RNA extraction and RT-qPCR
Total RNA was extracted using TRIZol (Invitrogen) and 2 μg were subjected to DNase treatment according to the manufacturer’s protocol (Thermo Scientific). One μg of DNase-free RNA was reverse transcribed using Maxima H Minus Reverse Transcriptase (Thermo Scientific). RT-qPCR were performed using the LightCycler 480 SYBR Green I Master Kit on a LightCycler480 apparatus (Roche) using standard protocols (40 cycles, 60°C annealing). PP2A (AT1G13320; primers are listed in Supplementary Table 3) was used as reference.
RNA Immunoprecipitation
RNA immunoprecipitation (RIP) assays were performed on transiently transformed N. benthamiana leaves as described in (62), or in 10-day-old A. thaliana 35S:WRKY42:GFP seedlings as described in (63), using anti GFP (Abcam ab290) and anti-IgG (Abcam ab6702). RIP was performed using Invitrogen Protein A Dynabeads. Precipitated RNAs were prepared using TRI Reagent (Sigma-Aldrich), treated with DNase (Fermentas) and subjected to RT-qPCR (High Capacity cDNA Reverse Transcription Kit (Thermo); primers used are listed in Supplementary Table 3). Total RNAs were processed in parallel and considered the input sample.
Chromatin Immunoprecipitation
Chromatin immunoprecipitation (ChIP) assays were performed on 10-day-old WT seedlings treated or not during 24h at 10 °C, using anti H3K27me3 (Diagenode pAb-195-050), anti LHP1 (Covalab pab0923-P) and anti-IgG (Abcam ab6702), as described in (27). Crosslinked chromatin was sonicated using a water bath Bioruptor Pico (Diagenode; 30sec ON/30sec OFF pulses; 10 cycles; high intensity). ChIP was performed using Invitrogen Protein A Dynabeads. Precipitated DNA was recovered using Phenol:Chloroform:Isoamilic Acid (25:24:1; Sigma) and subjected to RT-qPCR (primers used are listed in Supplementary Table 3). Untreated sonicated chromatin was processed in parallel and considered the input sample. For in vitro competition assays, APOLO was transcribed using the T7 transcription kit (Promega; (27)). After regular chromatin isolation from 10-day-old 35S:WRKY42:GFP seedlings, the sample was split in 4 independent tubes and diluted to 1ml in Nuclei Lysis Buffer without SDS. 0 μg, 0.1 μg, 1 μg and 10 μg of APOLO were added to each sample respectively, and incubated in rotation at 4 °C for 3h. Then, cross-linking was performed with 1% formaldehyde for 5 min at 4 °C, followed by 5 min with a final concentration of 50 mM glycine. SDS was added to a final concentration of 0.1% prior to sonication and the subsequent steps of a regular ChIP protocol.
Chromatin Isolation by RNA Purification followed by qPCR
A method adapted from the ChIRP protocol (64) was developed to allow the identification of plant DNA associated to specific lncRNAs, as described in (26, 27). Briefly, plants were in vivo crosslinked and cell nuclei were purified and extracted through sonication. The resulting supernatant was hybridized against biotinylated complementary oligonucleotides that tile the lncRNA of interest and lncRNA-associated chromatin was isolated using magnetic streptavidin beads. One hundred pmol of probes against APOLO (ODD and EVEN set of probes (26, 27)) and the corresponding negative set against LacZ were used for independent purification. Co-purified ribonucleoprotein complexes were eluted and used to purify RNA or DNA, which were later subjected to downstream assays for quantification as previously described (27).
Chromatin Conformation Capture
Chromosome conformation capture (3C) was performed basically as previously described in (27) starting with 2g of seedlings. Digestions were performed overnight at 37°C with 400U DpnII (NEB). DNA was ligated by incubation at 16°C for 5h in 4 ml volume using 100U of T4 DNA ligase (NEB). After reverse crosslinking and Proteinase K treatment (Invitrogen), DNA was recovered by Phenol:Chloroform:Isoamilic Acid (25:24:1; Sigma) extraction and ethanol precipitation. Relative interaction frequency was calculated by qPCR (primers used are listed in Supplementary Table 3). A region free of DpnII was used to normalize the amount of DNA.
Author contributions
MM, JMP, LL, CFF, JRM, AC, MB and FA performed the experiments. JB, MB, FI, MC, JE and FA analyzed the data. JE and FA conceived the project. FA, JE and CFF wrote the manuscript with the contribution of all authors.
LEGENDS TO SUPPLEMENTARY TABLES
Acknowledgments
We thank Chang Liu for providing the Hi-C plot of Figure 2B. This work was supported by grants from ANPCyT (PICT2016-0132 and PICT2017-0066) and Instituto Milenio iBio – Iniciativa Científica Milenio, MINECON to JME; ANPCyT (PICT2016-0007 and PICT2016-0289) to FA; UNL (CAI+D 2016) to LL; Laboratoire d’Excellence (LABEX) Saclay Plant Sciences (SPS; ANR-10-LABX-40) to MC; and CNRS (Laboratoire International Associé NOCOSYM) to MC and FA. LL, FI, JME and FA are researchers of CONICET; MM, JMP, CFF and JRM are fellows of the same institution.