DNA Methylation and Transcriptomic Changes in Response to Different Lights and Stresses in 7B-1 Male-Sterile Tomato

We reported earlier that 7B-1 mutant in tomato (Solanum lycopersicum L., cv. Rutgers), an ABA overproducer, is defective in blue light (B) signaling leading to B-specific resistance to abiotic and biotic stresses. Using a methylation-sensitive amplified polymorphism (MSAP) assay, a number of genes were identified, which were differentially methylated between 7B-1 and its wild type (WT) seedlings in white (W), blue (B), red (R) lights and dark (D) or in response to exogenous ABA and mannitol-induced stresses. The genomic methylation level was almost similar in different lights between 7B-1 and WT seedlings, while significant differences were observed in response to stresses in D, but not B. Using a cDNA-AFLP assay, several transcripts were identified, which were differentially regulated between 7B-1 and WT by B or D or in response to stresses. Blue light receptors cryptochrome 1 and 2 (CRY1 and CRY2) and phototropin 1 and 2 (PHOT1 and PHOT2) were not affected by the 7B-1 mutation at the transcriptional level, instead the mutation had likely affected downstream components of the light signaling pathway. 5-azacytidine (5-azaC) induced DNA hypomethylation, inhibited stem elongation and differentially regulated the expression of a number of genes in 7B-1. In addition, it was shown that mir167 and mir390 were tightly linked to auxin signaling pathway in 5-azaC-treated 7B-1 seedlings via the regulation of auxin-response factor (ARF) transcripts. Our data showed that DNA methylation remodeling is an active epigenetic response to different lights and stresses in 7B-1 and WT, and highlighted the differences in epigenetic and transcriptional regulation of light and stress responses between 7B-1 and WT. Furthermore, it shed lights on the crosstalk between DNA hypomethylation and miRNA regulation of ARFs expression. This information could also be used as a benchmark for future studies of male-sterility in other crops.


Introduction
The 7B-1 mutant in tomato (Solanum lycopersicum L., cv. Rutgers) is a genic photoperioddependent male-sterile in long days with stamens that are shrunken and produce non-viable DNA methylation can be detected via bisulfite conversion, methylation-sensitive restriction enzymes, methyl-binding proteins and anti-methylcytosine antibodies. Methylation-sensitive amplified polymorphism (MSAP) technique has been successfully applied to study DNA methylation variations in many plants, including tomato [20,21,28]. The main objective of this study was to investigate the differences in DNA methylation and transcriptional regulation in response to different lights and stresses between 7B-1 and WT, which were associated with the 7B-1 mutation and male-sterility in 7B-1. DNA methylation localization across the genome was profiled in 7B-1 and WT seedlings in different light conditions and in response to exogenous ABA and mannitol using the MSAP approach. Global changes of the genomic methylation level were also measured in 7B-1 and WT seedlings in the above mentioned conditions using anti-methylcytosine antibodies. Transcriptomic changes in blue light, dark, and in response to ABA, mannitol, and also 5-azaC (DNA demethylating agent) were studied in 7B-1 and WT seedlings using a cDNA-AFLP assay.

Plant material and stress treatments
Tomato seedlings were grown either on a basal MS medium [29] or in soil. Seedlings were grown under continuous W, B, and R lights or in D in temperature-controlled growth chambers set at 23°C (Microclima 1000E, Snijders Scientific B.V., The Netherlands) equipped with different light sources. W was provided by white cool fluorescent tubes (Philips TLD-36W/54, Phillips, USA). B and R were provided by blue (Philips TLD-36W/18-Blue) and red (Philips TLD-36W/15-Red) fluorescent tubes with a maximum irradiance at 460 and 660 nm, respectively. The total photon fluence rate was calibrated to 10 μmol m -2 s -1 by the Department of Biophysics at Palacky University in Olomouc using a portable spectroradiometer (model LI-1800, LI-COR, Lincoln, Nebr.). Stress treatments were carried out as described by Fellner and Sawhney [4] by growing the seedlings on MS mediums containing 10 μM ABA or 140 mM mannitol or 10 μM fluridone (an inhibitor of ABA biosynthesis). 5-Azacytidine treatment was carried out by placing drops of 40% ethanol solution containing 50 uM 5-azaC directly to the shoot apical meristem of 2-month old seedlings grown in long days (16 h light) once a day for 30 days. Control seedlings were treated similarly, but using ethanol drops instead of 5-azaC. Three biological replicates of WT and 7B-1 were included for all the treatments.

MSAP analysis
Genomic DNA was extracted from the biological replicates using the DNeasy Plant Mini kit (Qiagen), and pooled in equimolar ratio. MSAP analysis was performed as described by Portis et al. [30]. List of adapter sequences and primers used for pre-and selective PCR amplifications is provided in S1 Table. In brief, 200 ng of genomic DNA was digested using 20 U of either HpaII or MspI and 10 U of EcoRI (New England Biolabs) in a final volume of 20 μl for 4 h at 37°C. Digested DNAs were then ligated to EcoRI and HpaII/MspI adapters and preamplified using primers complementary to the adapter sequence plus a single base extension (Eco+A and H/M+T). Primary PCR products were diluted 10 x and used as templates for selective PCRs using 24 primer combinations. MSAP products were separated on 6% sequencing acrylamide gels and visualized by silver staining.

Measuring global genomic methylation level
Global genomic methylation level was measured using the MethylFlash Methylated DNA Quantification kit (Epigentek). In brief, DNA was first bound to high affinity strip wells, and then the methylated fraction of DNA was detected using capture/detection antibodies and subsequently quantified in an ELISA-based assay by reading the absorbance at 450 nm. The amount of methylated DNA is proportional to the absorbance reading. Level of 5-methylcytosine was quantified as the percentage of methylated cytosine in total genomic DNA using the formulas described in the kit manual.

cDNA-AFLP analysis
Total RNA was extracted from the biological replicates using the RNeasy Plant Mini kit (Qiagen), and pooled in equimolar ratio. First-strand cDNAs were synthesized using the Prime-Script First Strand cDNA Synthesis kit (TAKARA). Second strand cDNAs were synthesized using the NEBNext mRNA Second Strand Synthesis kit (New England Biolabs), and used for cDNA-AFLP analysis following the protocol described by Bachem et al. [31]. List of adapter sequences and primers used for pre-and selective PCR amplifications is provided in S2 Table. In brief, 100 ng of double stranded cDNAs were digested with EcoRI and MseI. Digested products were then ligated to EcoRI and MseI adapters. Pre-amplification was carried out using EcoRI and MseI adapter-specific primers. Pre-amplified reactions were diluted 10 x and used as templates in selective PCRs using 24 primer combinations. AFLP products were separated on 6% sequencing acrylamide gels and visualized by silver staining.

Cloning and characterization of differentially amplified fragments
Differentially amplified fragments were isolated from the gel following the procedure described by Wang et al. [24] and reamplified using same sets of primers used for the selective PCR. Amplified fragment were subsequently cloned into the pGEM-T Easy Vector (Promega) and subjected to sequence analysis. Sequences were annotated using blast search against the SOL Genomics Network (SGN; http://solgenomics.net/) and NCBI databases and gene ontologies were assigned from SGN, EBI (http://www.ebi.ac.uk/interpro/) or UniProt Knowledgebase (UniProtKB; http://www.uniprot.org/uniprot/).

Quantitative PCR
QPCR validations were carried out using the SensiFAST SYBR Lo-ROX kit (Bioline) and firststrand cDNAs as templates. Gene accessions and gene-specific primers are listed in S3 Table. Housekeeping α-tubulin and CAC genes were used as reference genes for data normalization (data were shown only for α-tubulin). PCR conditions were set at 95°C for 2 min, followed by 40 cycles of 95°C for 5 s, and annealing/extension at 60°C for 20 s. Expressions of miRNAs and tasiRNAs were validated using the Mir-X miRNA First-Strand Synthesis and SYBR qRT-PCR kit (Clontech). In a single reaction, sRNA molecules were polyadenylated and reverse transcribed using poly(A) polymerase and SMART MMLV Reverse Transcriptase provided by the kit. List of miRNA and tasiRNA forward primers is provided in S3 Table. U6 small nuclear RNA was used as a reference for data normalization. QPCR conditions were set at 95°C for 10 s, followed by 40 cycles of 95°C for 5 s, and annealing/extension at 60°C for 20 s. Changes of expressions were calculated as normalized fold ratios of three replicates using the ΔΔCT method [32].

Experimental design and statistical analysis
The experiments were arranged in a completely randomized design with three replications. Data were analyzed using the analysis of variance (ANOVA). Duncan new multiple range test (DNMRT p = 0.05) was used for comparison of the means.

Results
Methylation profile in 7B-1 seedlings in response to different lights Isoschizomers HpaII and MspI display differential sensitivity to DNA methylation as presented in Table 1. MSAP bands could be divided into four types based on their restriction patterns as illustrated in Fig 1. Type I bands, present in the gel for both enzyme combinations, which indicates presence of unmethylated C; type II bands, present only for EcoRI/HpaII (hemi-methylation); type III bands, present only for EcoRI/MspI (internal full methylation); and type IV bands, absent from both enzyme combinations (external full methylation). Methylation profiles in 7B-1 and WT seedlings were analyzed in W, B, R, and D using a MSAP assay. Twenty four primer combinations were used in our study (S1 Table) and a total of 20 polymorphic bands were identified from the gels and sequenced ( Table 2). Several fragments encoded proteins with known regulatory functions, including a phosphatase/tensin-like protein, ribonuclease 3 (RTL3), GTPase-activating protein (GAP), cytochrome b6, anthocyanidin synthase, 14-3-3 protein, histone-lysine N-methyltransferase (HMT), ABC transporter, serine/threonine protein kinase (CIPK), S-adenosylmethionine decarboxylase proenzyme (SAMDC), pentatricopeptide repeatcontaining (PPR) protein, and poly(A) RNA polymerase. These proteins regulate biological processes, including protein dephosphorylation, RNA processing and polyadenylation, regulation of G proteins, electron and metabolite transport, anthocyanidin and polyamine biosynthesis, protein-binding, histone methylation, and ABA signaling pathway ( Table 2). Other sequenced fragments were originated either from uncharacterized genomic regions or encoded uncharacterized putative proteins. The fragments were differentially methylated either in different lights (including D) or in between 7B-1 and WT seedlings (S4 Table). Phosphatase/tensin, HMT and 14-3-3 were ummethylated in WT and fully methylated (type III) in 7B-1 in all lights. RTL3 and cytochrome b6 were both unmethylated in WT in all lights and in 7B-1 only in R, but fully methylated (type III) in 7B-1 in other lights. GAP, anthocyanidin synthase, and poly(A) RNA polymerase were fully methylated (type IV) in WT, but unmethylated in 7B-1 in all lights. ABC transporter was Methylation profile in 7B-1 seedlings in response to different stresses Methylation profiles in response to exogenous ABA and mannitol-induced stresses in B and D were analyzed in 7B-1 and WT seedlings. A total of 19 polymorphic bands were identified from the gels and sequenced (Table 3). Among those, several sequences encoded stress or defenserelated proteins, which have been previously characterized in plants, including SNF4, UDPglucose flavonoid 3-O-glucosyltransferase (UF3GT), galactinol synthase (GolS), Pti4, Nresistance protein, and MYB transcription factor (AIM1). These fragments were differentially methylated in different light/stress conditions either in WT or 7B-1 or in between (S5 Table). SNF4 was hemi-methylated in D/MS in both WT and 7B-1, but fully methylated (IV) in WT and 7B-1 in other light/stress conditions. UF3GT was fully methylated (type IV) in WT in B and D by mannitol, but hemi-methylated in both WT and 7B-1 in other light/stress conditions. GolS was fully methylated (type IV) in WT and 7B-1 in B/MS and D/MS, while hemi-methylated in WT and 7B-1 in other light/stress conditions. Pti4 was fully methylated (type IV) in WT in B by mannitol, while hemi-methylated in other conditions. In 7B-1, it was fully methylated (type IV) in all conditions. N gene was fully methylated (type IV) in WT in D by mannitol, but unmethylated in other conditions. In 7B-1, it was unmethylated in B/MS and D/MS, but fully methylated (type III) in other conditions. AIM1 was fully methylated (type IV) in WT in all conditions. In 7B-1, it was fully methylated (type IV) in B/MS and D/MS, but unmethylated in other conditions. Similar to experiments with the lights, these results also indicated an active DNA methylation reprogramming in response to stresses in B and D with distinctive methylation patterns between 7B-1 and WT.
DNA methylation dynamic in 7B-1 in response to different lights and stresses Even though there were minor variations of methylation level in different lights and a higher methylation level was observed in D in both 7B-1 and WT and in R in 7B-1 seedlings, yet genomic DNAs did not undergo an extreme methylation changes. Nevertheless, these observation suggested that there might be a dedicated epigenomic mechanism for fine-tune regulation of gene expression in D and R distinctive from those in W and B, as implied by the higher methylation levels in D and R. Global changes of genomic methylation level were also measured in 7B-1 and WT seedlings in response to exogenous ABA, mannitol, and fluridone in B and D (Fig 3). In WT, methylation level was not changed by ABA and mannitol in B, but increased by fluridone. Methylation level in D was higher than B in untreated WT and significantly increased in response to ABA, mannitol and fluridone. In 7B-1, methylation level was decreased in response to ABA and fluridone in B, but slightly increased by mannitol. Methylation level in D was higher than in B in untreated 7B-1 and decreased in response to ABA and mannitol, but increased by fluridone. In B/stresses conditions, methylation changes were not much significant between 7B-1 and WT, except for the fluridone, however these changes were strikingly prominent in D/stresses conditions, as 7B-1 had a much lower methylation level in general compared to WT. These results indicated extensive changes of genomic methylation landmarks in response to stresses in both light (as the case of B) and dark, which were also distinctively different between 7B-1 and WT.
Transcriptomic profiling in 7B-1 seedlings in blue light and in response to stresses Transcriptomic changes between 7B-1 and WT in B and D and also in response to exogenous ABA and mannitol were analyzed using a cDNA-AFLP assay. A total of 19 differentially regulated bands were isolated from the gel and sequenced (Table 4). Several sequences encoded proteins with known regulatory functions, including pectate lyase, transmembrane protein, phosphoserine phosphatase, lipase, RNA helicase, protein phosphatase 2C (PP2C), protein kinase EXS, H + -ATPase, and WRKY transcription factor. These proteins are involved in regulation of processes, including cell wall modification, membrane transport, amino acid and lipid biosynthesis, stress response, microsporogenesis, ATP biosynthesis, and regulation of transcription (Table 4). These fragments were differentially regulated between WT and 7B-1 in B or D (in no stress condition) or in response to stresses (S6 Table; comparisons were made based on the intensity of the bands in WT in no stress condition as a reference). In no stress condition, pectate lyase, lipase, kinase, and H + -ATPase were down regulated in 7B-1 in B. Phosphoserine phosphatase and RNA helicase genes were also down regulated in 7B-1 but in D. Several other uncharacterized sequences were also differentially regulated between 7B-1 and WT in B and D (S6 Table). In response to stresses, pectate lyase and the gene encoded transmembrane protein were both down regulated in 7B-1 in B by mannitol, but not affected by ABA. In D, expression of both genes remained unchanged. Phosphoserine phosphatase was not affected by any stresses in B, but down regulated in both 7B-1 and WT in D by ABA, but not mannitol. Lipase was down regulated in 7B-1 in B by ABA and mannitol. In D, it was down regulated in both 7B-1 and WT by mannitol. RNA helicase was up regulated in 7B-1 in B, but down regulated in D by both ABA and mannitol. PP2C was up regulated in 7B-1 in B by ABA and mannitol, while it was down regulated in both 7B-1 and WT in D by mannitol. Kinase EXS was up regulated in 7B-1 in B by ABA and mannitol, but not in D. H + -ATPase was down regulated in WT and 7B-1 in B and D by ABA and mannitol. WRKY was up regulated in 7B-1 in B and D by ABA and mannitol, but not in WT.
Expressions of lipase, RNA helicase, PP2C, H + -ATPase, and WRKY genes were further validated using qRT-PCR (Fig 4), and data were in good agreements with the observation from the gel. In addition, expression of a gene encoding 14-3-3 protein was also analyzed using qRT-PCR. 14-3-3 was down regulated in 7B-1 in B in no stress condition, but not in D. In response to ABA and mannitol, it was down regulated in 7B-1 and WT in B as well as D. Transcriptomic analysis in our study revealed that light and stress-induced regulation of gene expression in 7B-1 and WT are different at least to some extents as implied by differential regulation of the above subset of genes.
DNA methylation and transcriptomic profiling in 7B-1 seedlings in response to 5-azaC 5-azaC induces aberrant DNA demethylation and genome-wide transcriptional reactivation of silenced genes [28,33]. Global changes of the genomic methylation level in 7B-1 and WT seedlings (2-month old seedlings grown in LD) were measured in response to 5-azaC (Fig 5). Methylation level was relatively similar in untreated 7B-1 and WT seedlings; however in response to 5-azaC, methylation level was reduced (hypomethylation) in 7B-1 and WT of about 1.5 and 1.4 folds, respectively. Some morphological changes were also associated with the treatment, including inhibition of stem elongation in WT and more predominantly in 7B-1 (S1 Fig), and promotion of the lateral shoot formation in both 7B-1 and WT. Structure of the flowers was not affected by 5-azaC and male fertility was not restored in 7B-1. To investigate the effects of aberrant DNA demethylation on transcriptional regulation of gene expression, transcriptomic changes in 5-azaC-treated 7B-1 and WT seedlings were profiled using a cDNA-AFLP assay. Total of 14 differentially regulated bands were isolated from the gels and sequenced (Table 5). Among those, six transcripts encoded proteins, which have been previously characterized from plants, including subtilisin-like protease (SDD1), CCCH-Type zinc finger, Ycf4, NPH3-type protein, auxin response factor 8 (ARF8), and ABA responsive transcription factor (ABF4). These proteins regulate plant processes, including stomatal morphogenesis, photosynthesis, response to light and hormones ( Table 5). Expressions of these genes were validated using qRT-PCR (Fig 6). SDD1 was up regulated in both 7B-1 and WT seedlings in response to 5-azaC. Zinc finger was down regulated in 7B-1, but slightly up regulated in WT by 5-azaC. Expression of Ycf4 was not affected in 7B-1, but increased in WT by 5-azaC. NPH3 was strongly up regulated in 7B-1 by 5-azaC, but not in WT. ARF8 expression was higher in untreated 7B-1 seedlings as compared to WT; however it was down regulated strongly in both 7B-1 and WT by 5-azaC. ABF4 transcripts were found more abundantly in untreated 7B-1 compared to WT. In response to 5-azaC, ABF4 was strongly up regulated in both 7B-1 and WT. These results indicated that changes of the gene expression in response to 5-azaC were not similar between 7B-1 and WT as indicated by differential regulation of the above subset genes. In addition to the above genes, expressions of CRY1/2, PHOT1/2, and elongated hypocotyl 5 (HY5) were also analyzed in 5-azaC-treated 7B-1 and WT seedlings (Fig 7). Expressions of CRY1/2, and PHOT1/2 were similar between 7B-1 and WT seedlings, but HY5 expression was higher in 7B-1. Despite some minor changes (less than 2-fold changes), expressions of CRY1/2, and PHOT1/2 were not significantly affected by 5-azaC in 7B-1 and WT, while HY5 strongly down regulated in 7B-1, but not WT. These results indicated that blue light receptors were not affected by the induced DNA hypomethylation and 5-azaC-induced inhibition of hypocotyl elongation in 7B-1 is independent and not regulated by CRY1/2 and HY5. Mir167 and mir390 regulate auxin response factors in response to 5-azaC Mir167 cleaves ARF8 transcripts and mir390 triggers the production of tasiRNAs from TAS3 mRNA, which in turns cleave and down regulate ARF2, 3, and 4 [34]. To investigate if these miRNAs were linked to auxin signaling pathway in response to 5-azaC, expressions of mir167, mir390, TAS3-derived tasiRNAs (D7 and D8) and their target ARF transcripts (ARF2/3/4/8) were analyzed in 5-azaC-treated 7B-1 seedlings using qRT-PCR (Fig 8). Mir167, mir390, D7 tasiRNA were all strongly up regulated in response to 5-azaC, while D8 tasiRNA expression remained unaffected. Primary transcripts of ARF2, 3, 4, and 8 were found more abundantly in untreated 7B-1 seedling compared to WT (S2 Fig). Interestingly, ARF2, 3, and 8 were all strongly down regulated in response to 5-azaC, while ARF4 was slightly up regulated (Fig 8). Expression of sRNAs and/or ARFs could have been independently modulated by either DNA hypomethylation or each other in a feedback response or as a result of likely changes of the auxin gradient due to 5-azaC.

Discussion
The main goal of our study was to understand and highlight the differences in DNA methylation dynamic and transcriptional regulation in response to different lights, stresses and 5-azaC-induced DNA hypomethylation between 7B-1 and WT. In light experiments, "0" means no change of expression in response to 5-azaC compared to the reference, "+" means up regulation and "-"means down regulation.
Expression profiles are based on the intensities of the bands in 5-azaC-treated tissues as compared to their corresponding untreated tissues. methylation changes were mainly displayed in different lights in 7B-1 or WT or in between. Genes with different regulatory functions, including protein dephosphorylation, RNA processing and polyadenylation, regulation of G proteins, electron/metabolite transport, anthocyanidin and polyamine biosynthesis, histone methylation, ABA signaling, and defense response were differentially methylated by lights; however, we did not find any of the key receptors/regulators of light response in our analysis. Among the sequences noteworthy here were RTL3, poly  (A) RNA polymerease, and histone methyltransferase, which the first two encode RNA processing enzymes and third one catalyzes histone methylation. Histone methylation and alternative RNA polyadenylation have been shown to be strongly correlated with light regulation of the gene expression in plants [35][36][37][38]. However, it's not clear if these enzymes themselves could also be regulated by DNA methylation. These findings suggested that DNA methylation represented a layer of epigenetic control over the light-induced regulation of gene expression. However, it is unclear if the differences of DNA methylation patterns between 7B-1 and WT are solely due to different lights or could be the primary effect of the 7B-1 mutation, introducing new methylation marks independent of lights. As MSAP assay does not have the potential to provide a global overview of the methylation dynamic, global methylation changes in different lights were measured in 7B-1 and WT seedlings. Despite some notable variations, genomic DNA did not go through extreme methylation changes in different lights. The data suggested that epigenetic response to different lights in 7B-1 does not encompass extreme methylation reprogramming; however, there could be specialized epigenetic responses for fine-tune regulation of the gene expression in different lights as higher methylation levels was observed in D and R compared those in W and B. In response to ABA and mannitol, several stress-related genes, including SNF4, Pti4, N resistance gene, and MYB were identified, which were differentially methylated between 7B-1 and WT. Tobacco plants infected with tobacco mosaic virus showed strong CG hypomethylation at LRR region of the N gene [39]. Our results indicated that epigenetic regulation of stress response in 7B-1 and WT requires different DNA methylation reprogramming as indicated by differential methylation of the identified genes. Identification of diverse category of genes with altered DNA methylation patterns in our study provided a clear evidence that firstly, DNA methylation remodeling plays a critical role in plant adaptation to environmental cues, such as light and stresses, and secondly these interactions were different at least party between 7B-1 and WT. Different DNA methylation dynamics were observed in response to ABA, mannitol, and fluridone between 7B-1 and WT in B and D. Changes of DNA methylation level were not so significant between 7B-1 and WT in B in stress conditions, except for the fluridone. On the other hand, these differences were quite remarkable in D, suggesting that DNA hypermethylation in D is an active epigenetic response to stresses in WT. The results indicated that in addition to the crosstalk between light and hormonal signaling pathways, the WT response to stresses in B and D could partially be regulated through an extensive DNA methylation reprogramming. In contrary, 7B-1 response to stresses in B and D does not impose an extreme DNA methylation remodeling, but it's very likely to be different from those of WT in D. Hypermethylation in D is in good agreement with the slow rate of global transcription in D in general, in contrast to the faster rate of light-induced transcription. Stress-induced increase of methylation level in WT in D could also be explained in a way that hypermethylation further did reduce the global transcription rate in favor of slowing down the energy consumption, while expression of a subset of hypomethylated stress-related genes could likely be expanded to help the cell to cope with the stress. However, it remains unclear if the sudden drop of DNA methylation level in 7B-1 in response to ABA and mannitol in D is due to the direct effects of stresses or as a result of autonomous interplay between 7B-1 mutation and affected hormonal sensitivities in 7B-1.
Fluridone prevents ABA biosynthesis, but it was not clear if it affected ABA level in 7B-1 and WT equally. Different methylation response of 7B-1 to fluridone in B compared to WT could have been regulated by elevated sensitivity or higher endogenous level of the ABA in 7B-1 seedlings. Not taking the light-ABA interaction into account in D, fluridone increased methylation level similarly in 7B-1 and WT. Interestingly, ABA treatment or inhibition of ABA synthesis did not influence the methylation level in WT in D, but they reduced and induced the methylation level in 7B-1, respectively.
Using a cDNA-AFLP assay, several transcripts were identified, which were differentially regulated between 7B-1 and WT in B or D (no stress condition). These transcripts encoded regulatory proteins involved in process, such as cell wall modification, lipid and amino acid biosynthesis, microsporogenesis, and ATP biosynthesis. These findings highlighted some of the differences of blue light and dark regulation of gene expression between 7B-1 and WT. H + -ATPase and 14-3-3 genes were both down regulated in 7B-1 in B. PHOT1/2 mediate B-dependent activation of H + -ATPase in guard cells via phosphorylation and subsequent binding of the 14-3-3 proteins [40]. In contrary, ABA could suppress stomatal opening by inhibiting phosphorylation of H + -ATPase [41]. Hlavinka et al. [42] reported that B-induced stomatal opening is impaired in 7B-1, but unlikely to be affected by H + -ATPase. Our data suggested that B has positively regulated H + -ATPase expression in WT, as a defect in B perception in 7B-1 resulted in lower expression of H + -ATPase transcripts. In addition, B-insensitivity and elevated level of ABA in 7B-1 could have possibly led to partial inactivation of H + -ATPase. On the other hand, down regulation of 14-3-3 expression in B could slow down H + -ATPase phosphorylation rate, which in turn reduces the B-induced stomatal opening.
In response to ABA and mannitol, several stress-related genes, including lipase, RNA helicase, PP2C, and WRKY were differentially regulated between 7B-1 and WT in B or D. Several studies have shown differential expression of lipases in response to abiotic stresses [43,44], but the actual function of these genes in plant adaptation to stress is largely unknown. Understating how down regulation of this gene in 7B-1 is connected to stress response requires further functional studies. RNA helicases have been implicated in every step of RNA metabolism [45], while their involvement in response to abiotic stress is only beginning to emerge. There are few reports displaying up regulation of the DEAD-box family of RNA helicases in response to abiotic stress, suggesting that they might play important roles in stabilizing plant growth under stress conditions by regulating stress-induced pathways [46][47][48]. In pea, two DEAD-box helicases, PDH45 and PDH47, were induced by a variety of abiotic stresses [48], while in Arabidopsis, two DEAD-box RNA helicases, STRS1 and STRS2, negatively regulated the stress response [49]. Up regulation of RNA helicase in 7B-1 by ABA and mannitol could be associated with B-specific stress-tolerance of 7B-1, however down regulation of this gene in D in no stress condition and by ABA and mannitol, suggests an alternative regulatory function for this gene rather than a stress-related one. PP2Cs act redundantly in ABA signaling and stress response in plants [50][51][52][53]. PP2Cs have been reported to be differentially expressed in response to different abiotic stresses [51][52][53]. Inactivation of PP2Cs by ABA receptors, such as OsPYL5 and AtPYL enhanced drought tolerance in rice and Arabidopsis, respectively [51,52]. Expression of OsSIPP2C1 was up regulated in rice in response to ABA, drought and salinity [53]. Up regulation of PP2C in 7B-1 by ABA and mannitol in B could in part contribute to 7B-1 stress tolerance, but down regulation of this gene in WT and 7B-1 by mannitol in D, suggests a different mode of action for this gene in D as compared to B, which has yet to be understood. Several WRKYs have been identified, which positively regulate plant response to a range of abiotic stresses [54][55][56]. Up regulation of WRKY in 7B-1 in response to ABA and mannitol in B and D, could in part mediate the higher tolerance of 7B-1 to these stresses. Response to stress in light and dark is coordinated by a synchronized action of many stress-inducible genes, some being positively regulated while others down regulated [57]. Our findings indicated that blue light and dark regulation of stress response in 7B-1 is modulated at least partly through a different transcriptional reprogramming in comparison to those in WT. These differences could be due to the indirect effect of 7B-1 mutation, which has impaired B signaling and hormonal balance in 7B-1, altering the downstream gene expression as a result. In addition to the above mentioned protein-coding mRNAs, several non-coding transcripts were also identified in our cDNA-AFLP experiment, which have been differentially expressed in different lights and stresses. These fragments are likely the amplified long non-coding RNAs (lncRNAs), which could have been transcribed from intronic, intergenic, intragenic, promoters, and untranslated regions. Several lncRNAs have been identified from Arabidopsis [58], rice [59] and wheat [60], which were suggested to be related to abiotic stress response based on their expression profile. However, lncRNAs are still largely unknown in plants, which makes their analysis rather challenging, therefore we turned our attention to those protein-coding transcripts.
Phenotypic and developmental changes could be induced in plants by aberrant demethylation of DNA [28,33,61]. 5-azaC-induced hypomethylation in 7B-1 (1.5-fold decrease of genomic methylation level) inhibited stem elongation and promoted shoot branching, but did not affect the flower structures and male sterility. Transcriptomic changes were mainly profiled in the stem tissue as induced morphological changes were manifested primarily in this tissue. Among those differentially regulated transcripts, several transcripts encoded regulatory proteins involved in processes, such as stomatal morphogenesis, photosynthesis, and response to light and hormones. These transcripts were differentially regulated in response to 5-azaC in either 7B-1 or WT or in between. These results indicated that induced hypomethylation has affected the transcriptional regulatory network or at least a subset of genes unequally between 7B-1 and WT.
NPH3 (non-phototropic hypocotyl 3) is a phototropin-interacting protein, which is essential for PHOT1-dependent phototropic response [62]. Blue light together with PHOT1 stimulate NPH3 activity via dephosphorylation, but it not clear if B induces the NPH3 expression [63]. NPH3 had a lower expression in untreated 7B-1 seedlings compared to WT, and it was up regulated in response to 5-azaC in 7B-1, but not in WT. Earlier experiments in our lab showed that the phototropism was impaired in 7B-1 seedlings (Bergougnoux and Fellner, unpublished data). This could also be addressed by lower level of NPH3 transcripts in 7B-1 seedlings, which itself could be due to a tradeoff mechanism between inactive form of the NPH3 protein (phosphorylated) and NPH3 transcripts level. Up regulation of this gene by 5-azaC in 7B-1 and not WT could possibly be explained by i) the indiscriminate removal of epigenetic marks from NPH3 exclusively in 7B-1, which could have induced the NPH3 expression, or ii) partial restoration of B sensitivity in 7B-1 in response to 5-azaC, which could have balanced off the NPH3 transcripts. This brings into question if phototropic response in 7B-1 is restored by 5-azaC, which has to be yet verified. PHOT1/2 expressions were similar between 7B-1 and WT, which indicated that the reduced rate of phototropism in 7B-1 was independent and not mediated through its receptors at the transcriptional level. Similar expression of CRY1/2 between 7B-1 and WT also indicated that that the defect in blue light signaling as well as the reduced deetiolation in 7B-1 were rather independent and not regulated by CRY1/2, at least at the transcriptional level, suggesting that the 7B-1 mutation had probably affected the downstream components of the light signaling pathway, and not the receptors. HY5 plays a key role in promoting photomorphogenesis by regulating the transcription of a wide range of genes, and its abundance is negatively correlated with hypocotyl elongation [64]. In our study, HY5 had a higher expression in 7B-1 compared to WT seedlings. Keeping in mind that 7B-1 had a longer hypocotyl than WT, brings the observation to the contrary with the general function of HY5 in inhibition of hypocotyl elongation, which higher abundance of HY5 protein is associated with shorter length of the hypocotyl. However, it should be noted that higher level of HY5 transcript does not necessarily imply to a higher protein accumulation. Even though, 5-azaC inhibited hypocotyl elongation in 7B-1, yet down regulation of HY5 in 7B-1 in response to 5-azaC could not be agreeably tailored to its central role in light-inhibition of hypocotyl elongation. Increasing evidences suggest that hormones, such as ethylene, gibberellins and cytokinins also regulate the accumulation of HY5 [65], therefore primary expression of HY5 in 7B-1 and in response to 5-azaC could perhaps be in part mediated by hormonal imbalance and altered sensitivity in 7B-1.
ARFs modulate auxin signaling by regulating the expression of auxin-response genes [66]. Mir167 and mir390 are tightly connected to auxin signaling pathway via down regulation of ARF8 and ARF2/3/4 transcripts, respectively [34]. Effect of DNA hypomethylation on the crosstalk between miRNAs, ARFs, and auxin signaling was investigated in our study. Mir167, mir390, D7 tasiRNA were all up regulated in 5-azaC-treated 7B-1 seedlings, which expectedly down regulated the expression of their target ARF transcripts except for ARF4; ARF4 was slightly up regulated. These results pointed to a combinatory interplay between DNA hypomethylation and miRNA modulation of auxin signaling via spatial regulation of ARFs. Furthermore, our findings underlined the vital regulatory functions of miRNA in balancing off the gene expression in response to uncommon epigenetic marks, such as induced DNA hypomehtyation. However, it is unclear if the changes of miRNA expressions were due to the direct effect of 5-azaC treatment or modulated through a feedback regulation mediated by auxin gradient in 7B-1 seedlings. We did not measure the auxin level in 5-azaC-treated 7B-1 seedlings, but as auxin transport is required for hypocotyl elongation in light [67], one possible explanation for inhibition of stem elongation in 7B-1 could be that 5-azaC has reduced the rate of auxin synthesis in shoot apical meristem, slowing down its transport to other tissues, which in turn could have negatively regulated the stem elongation. This notion could be coupled by the observations that ARFs were down regulated by 5-azaC, but also to mention that the primary transcripts of ARFs were found more abundantly in untreated 7B-1 seedlings compared to WT, which could imply a more active auxin signaling/responsive gene expression in 7B-1. ARFs seem to control both auxin sensitivity and homeostasis as reported in Arabidopsis [66,68]. Several reports indicated involvement of ARFs, such as ARF2, 6, and 8 in regulation of the hypocotyl growth in an auxin-dependent fashion [68][69][70]. Target genes of ARF2, 3, and 4 remain largely unknown, thus the mechanisms linking these ARFs to context-specific cellular responses are poorly understood. Tian et al. [70] suggested that ARF8 could control the free level of IAA in a negative feedback by regulating the expression of GH3 genes. Based on our findings, it could be concluded the defect in blue light signaling in 7B-1 was not mediated by its photoreceptors and the mutation had probably affected downstream components of the light signaling pathway. Differences in the regulation of gene expression in response to lights and stresses between 7B-1 and WT clearly showed that 7B-1 mutation has extensively affected the light-and stress-responsive gene expression in 7B-1. The mutation also has changed the methylation dynamic in 7B-1 and brought along a whole new set of methylation marks in response to different light and stress scenarios, which could be either the direct effect of the mutation or as an adaptive response to changes of the gene expression portfolio. This information adds on to the fact that the 7B-1 is a complex mutation with its primary effect yet unknown. A probable explanation could be that male-sterility in 7B-1 is governed by interplays between blue light signaling defect and altered level of endogenous hormones, which in turn could have adversely affected the expression of genes involved in anther and pollen development. Comparative studies on light-and stress-responsive epigenomes and transcriptomes will enhance our understanding of plant adaptation to lights and stresses. Overall, our study highlighted some of the differences in epigenetic and transcriptional regulation of light and stress responses between 7B-1 and WT, and shed lights on the crosstalk between DNA hypomethylation and miRNA regulation of ARFs.