CHR729 Is a CHD3 Protein That Controls Seedling Development in Rice

CHD3 is one of the chromatin-remodeling factors that contribute to controlling the expression of genes associated with plant development. Loss-of-function mutants display morphological and growth defects. However, the molecular mechanisms underlying CHD3 regulation of plant development remain unclear. In this study, a rice CHD3 protein, CHR729, was identified. The corresponding mutant line (t483) exhibited late seed germination, low germination rate, dwarfism, low tiller number, root growth inhibition, adaxial albino leaves, and short and narrow leaves. CHR729 encoded a nuclear protein and was expressed in almost all organs. RNA-sequencing analysis showed that several plant hormone-related genes were up- or down-regulated in t483 compared to wild type. In particular, expression of the gibberellin synthetase gibberellin 20 oxidase 4 gene was elevated in the mutant. Endogenous gibberellin assays demonstrated that the content of bioactive GA3 was reduced in t483 compared to wild type. Moreover, the seedling dwarfism, late seed germination, and short root length phenotypes of t483 were partially rescued by treatment with exogenous GA3. These results suggest that the rice CHD3 protein CHR729 plays an important role in many aspects of seedling development and controls this development via the gibberellin pathway.


Introduction
Chromatin is a macromolecular complex that consists of DNA, protein, and RNA. At least three processes control chromatin assembly and regulation: DNA methylation, histone modification, and ATP-dependent chromatin remodeling [1]. Chromatin is highly dynamic and undergoes structural alterations in response to developmental changes and changes in the cellular environment [2,3]. Dynamic regulation of chromatin structure is important for transcriptional regulation and other functions in higher eukaryotes, and many coordinating and gene expression analysis. All materials for crossing and analysis were grown in the experimental field at the Chinese Academy of Agricultural Sciences, Beijing and Sanya.
For germination analysis, rice seeds were soaked in tap-water at 37°C in the dark for three days. The soaked seeds were incubated at 28°C with 12 h of light and 12 h of darkness for eight days, then seed germination rate was measured by counting only those seeds with shoots longer than 2 cm. For gibberellin treatment, mutant and WT seeds were surface sterilized in 2.5% NaClO, soaked with tap-water in sterile Petri dishes at 37°C in the dark for one day. Seeds were then soaked in different concentrations of gibberellin acid solution and incubated at 28°C with 12 h of light and 12 h of darkness.

Chlorophyll content measurement and transmission electron microscopy analysis
Chlorophyll contents were measured using a spectrophotometer according to the method of Arnon [18] with minor modifications. Briefly, equal weights of freshly collected second top leaves from two-week-old seedlings were marinated in 95% ethanol for 48 h in darkness. For thorough chlorophyll extraction tubes were periodically inverted five times. Residual plant debris was removed by centrifugation. The supernatants were used to measure chlorophyll content by a DU 800 UV/Vis spectrophotometer (Beckman Coulter) at 665, 649 and 470 nm.
For transmission electron microscopy analysis, leaf samples from two-week-old plants growing in paddy field were first fixed in 2% glutaraldehyde solution and then transferred into 1% OsO 4 for two days. After fixation, samples were stained with uranyl acetate and dehydrated in an ethanol series, and then embedded in Spurr's medium before ultrathin sectioning. Samples were stained with uranyl acetate again and observed with a transmission electron microscope (Hitachi H-7650, Japan).

Map-based cloning of the mutated gene in t483
To map the mutated gene, t483 was crossed with cv. 93-11 (indica). The plants with the mutant phenotype in F 2 population were selected for a genetic linkage analysis. Molecular markers distributed throughout the rice genome were chosen for preliminary mapping [19,20]. Additional Insertion-deletion (IN) markers for fine mapping were developed according to the DNA sequence difference between japonica and indica. PCR procedure was as following: 95°C for 5 min, followed by 35 cycles of 95°C for 30 s, annealing for 30 s, extension 72°C for 30 s, and a final extension at 72°C for 5 min.

Generation of transgenic rice plants
Because the genomic sequence of CHR729 was very large, containing 13,848 bp (not including promoter sequence), it was difficult to construct a complementary transformation plasmid. We therefore used a pCUbi1390-ΔFAD 2 RNAi vector to generate a CHR729-RNAi construct [21]. The specific region of CHR729 used for the RNAi construct was identified by alignment with the basic local alignment search tool (http://www.gramene.org). A 346 bp specific fragment of the CHR729 gene was amplified with primer pairs RNAi-SF/ RNAi-SR and RNAi-AF/ RNAi-AF, then cloned into the pCUbi1390-ΔFAD 2 vector as described by Mao et al. [22]. The RNAi construct was introduced into wild type Nipponbare by Agrobacterium tumefaciens-mediated transformation as previous report and empty pCUbi1390 vector was also introduced as a control [23].

Quantitative real-time PCR analysis
Total RNA was extracted from various tissues using RNA Prep Pure Plant Kit (Tiangen Co., Beijing), and was reverse transcribed using a SuperScript II Kit (TaKaRa), according to the user's manual. qRT-PCR (quantitative real-time PCR) analyses were performed using the 7900 HT Fast Real-Time PCR System (ABI). UBIQUITIN gene (Os03g0234200) was chosen as a reference gene. Reactions containing SYBR premix (TaKaRa) were carried out in final volumes of 20 μL with 2 pmol of the appropriate primers. The 2 -ΔΔCT method was used to calculate relative levels of gene expression [24].

Subcellular localization of CHR729
To determine the subcellular localization, the CHR729 cDNA fragment was amplified (primer pair GFP-F/GFP-R) and ligated into pA7-GFP vector (p35S-CHR729-GFP). As a control, the cDNA of a previously characterized nuclear protein, OsMADS3, was fused to the mCherry gene (p35S-OsMADS3-mCherry) [25]. Protoplasts were isolated from rice seedlings, and cotransfected with the p35S-CHR729-GFP and p35S-OsMADS3-mCherry vectors. The transformed protoplasts were incubated at 28°C in darkness for 16 h before detection. Fluorescence of GFP was observed using a confocal laser scanning microscope (Leica TCS SP5).

RNA-sequencing
Two-week-old whole WT and t483 seedlings were immediately frozen in liquid nitrogen and stored at -80°C. Material from five different plots was pooled together. Total RNA was extracted using a RNA Prep Pure Plant Kit (Tiangen Co., Beijing), and treated with RNase-free DNase I (NEB, Ipswich, MA, USA) to remove any genomic DNA contamination. For each sample, at least 10 μg of total RNA was used for illumina HiSeq2000 sequencing conducted at Beijing Novo Co. (Beijing). After sequencing, the raw reads were first purified by trimming adapter sequences and removing low-quality sequencing date. The clean reads were mapped to the reference genome of japonica cv. Nipponbare using SOAP2 software [26]. Genes differentially expressed (DEGs) between t483 and WT were identified using DEGseq R package (1.12.0; TNLIST, Beijing). P-values were adjusted using the Benjamini and Hochberg method [27]. Corrected p-values of 0.001 and log2 (fold change) of ±1 were set as the threshold to determine significant differential expression.
Gene ontology (GO) analysis was performed using the open-source MAS3 database (http:// bioinfo.capitalbio.com/mas3/). A threshold of a two-fold change in gene expression levels and a FDR of <0.05 were used to identify DEGs. The p-values and FDRs of DEGs were calculated as previously reported [28].
Determination of endogenous GAs levels 500 mg of the plant material powder was extracted with 5 mL of 90% aqueous methanol (MeOH). Simultaneously 2 ng of each D-labelled GA compound was added to the extracting solvents as internal standards for GAs content measurement. MAX and MCX Cartridges (6 mL, 500 mg) were purchased from Waters Corporation (Milford, MA, USA). The MAX cartridge was activated and equilibrated with 10 mL MeOH, water, 5% NH4OH, 90% MeOH in turn, while MCX with 10 mL MeOH, water and 90% MeOH. After the two columns were connected with an adapter, the crude extracts were subjected to the tandem cartridges. Then the MAX cartridge was disconnected and washed with 5% NH 4 OH in 5% MeOH, MeOH in sequence. At last GAs were eluted with 2% FA in 90% MeOH. After dried with N 2 stream, the eluent was reconstructed with 200 μL 80% MeOH and subjected to UPLC-MS/MS analysis.

Phenotypic characterization of the rice t483 mutant
When compared to WT rice plants, t483 plants exhibited abnormal growth at every stage of development. The germination rate was significantly lower in the t483 mutant than in WT, with a clear difference first becoming apparent three days after germination ( Fig 1A). Eight days post-germination, shoots of t483 were shorter than those of WT and roots in the mutant were shorter and lower in number (Fig 1B-1D). Furthermore, approximately 30% of the t483 mutant seeds arrested at the germination stage ( Fig 1B, 1E and 1F). These results indicate that t483 is defective in seed germination.
The t483 mutant was also defective at other development stages. At the tillering stage, stature and tiller numbers were reduced compared to WT (Fig 2A). At the mature stage, mutant plants were reduced in height and had smaller and fewer panicles than WT plants (Fig 2B-2D, Table 1). In addition, leaf blade lengths and widths were reduced in the t483 mutant compared   to WT. Mutant leaves exhibited pale coloration on the adaxial side ( Fig 2E, Table 1) and normal coloration on the abaxial side (data not shown). Spectrophotometric analysis showed that chlorophyll levels were 27% lower in t483 than in WT (Fig 2F). Consistent with the alteration in leaf color, the numbers of thylakoid lamellar and stacked grana in chloroplasts of adaxial mesophyll cells were lower in t483 than in WT, as determined using transmission electron microscopy (Fig 2G-2J). The chloroplast structure and thylakoid lamellar were, however, developed normally in abaxial mesophyll cells of the mutant (data not shown). In summary, the t483 mutant had morphological and growth defects that affected seed germination, plant height, tillering, panicle, leaf color, and leaf and root dimensions.
Map-based cloning of the mutated gene in t483 Segregation analysis indicated that the t483 phenotype was controlled by a single recessive gene (55:174, χ 2 1:3 = 0.118, P>0.70). The mutated gene was isolated by positional cloning. An F 2 population with 1,350 recessive mutant plants was developed from a cross between t483 and indica cv. 93-11. The mutated gene was located in a 27 kb region between markers IN27 and IN35 ( Fig 3A). Three predicted open reading frames (ORFs) were predicted in this region, and the genomic sequences of each were obtained from WT and t483. A single nucleotide substitution of T for A in ORF3 in t483 was predicted to encode an early stop codon ( Fig 3A). This nucleotide change was confirmed in the cDNA sequence.
Just after we analyzed the mutated gene in t483, two similar mutants, named chr729 and oschr4, were reported elsewhere [17,29]. The location and sequence of these mutants suggested that CHR729/OsCHR4 was a candidate for t483. To investigate this further, we generated several transgenic lines by introducing the specific RNA interference construct into Nipponbare plants. In a typical homozygous T 2 line, RNAi-3, CHR729 transcript levels were significantly lower than in control, as determined by real-time PCR analysis (Fig 3F). Phenotypic evaluation of RNAi-3 revealed that plant height (Fig 3B), seed germination (Fig 3C, 3D and 3G), chlorophyll content (Fig 3E and 3H), panicle number and length, and the length and width of the flag leaf were also significantly reduced in the RNAi-3 line compared to control ( Table 2). The phenotype of the silenced plants mimicked that of the t483 mutant. Together, the sequence and phenotype data suggested that the defect in the t483 line was caused by a mutation in CHR729/ OsCHR4. Henceforth, t483 is used to denote the rice mutant line and CHR729 is used for the gene name.
Sequence comparison between genomic DNA and cDNA indicated that CHR729 contained ten exons and nine introns, and encoded a 2259 amino acid protein. CHR729 belonged to the CHD3 protein subfamily and contained a PHD zinc finger domain, two chromo domains, a SNF2-related helicase domain, and a DNA-binding domain.

CHR729 is ubiquitously expressed and encodes a nuclear protein
To examine the spatial expression pattern of CHR729, qRT-PCR analysis of endogenous CHR729 transcripts was performed using total RNA isolated from different WT tissues. As shown in Fig 4A, CHR729 was expressed in all the examined organs, including culms, leaf blades and sheaths, young inflorescences, and seedling-stage roots. CHR729 was also expressed in flower organs, including anthers, pistils, lemmas, and paleas. CHR729 therefore appeared to be constitutively expressed, which was consistent with the morphological defects observed in t483 at many developmental stages.
To determine the subcellular localization of CHR729, a GFP fusion construct was transiently expressed in rice leaf protoplasts. The resulting GFP signal co-localized with the nuclear marker OsMADS3-mCherry [30] (Fig 4B-4D), suggesting that CHR729 was located in the nucleus.

Expression of genes putatively related to phytohormone function is altered in the t483 mutant
To understand the molecular mechanism of CHR729 in the regulation of plant development, gene expression changes were analyzed in young t483 seedlings using RNA sequencing (RNAseq). Transcriptional analysis showed that 346 and 154 genes were up-and down-regulated, respectively, in t483 relative to WT (P<0.05; S1 Table). Many of the differentially expressed genes (DEGs) were identified as transcription factor genes (S2 Table). Gene ontology (GO) enrichment analysis found that five GO categories were enriched in the group of DEGs Data are presented as means ± SE. Significance of differences between control and RNAi-3 was detected using Student's t-test (n = 5).
doi:10.1371/journal.pone.0138934.t002  Table). The most enriched category was "protein phosphorylation", and this group included genes encoding protein kinases. Of the DEGs, several were predicted to involve in phytohormones such as auxin, abscisic acid (ABA), cytokinins, gibberellin, and ethylene (S4 Table). Only three of the total 19 phytohormone genes were downregulated, and more than half were involved in ethylene signaling. The RNA-seq results were confirmed using qRT-PCR (Fig 5B). One of the upregulated genes, EIN3 (Ethylene-insensitive 3), encodes a protein that, after activation by EIN2, activates downstream transcription factors to induce various ethylene responses [31][32][33][34]. In addition to EIN3, expression of several other genes encoding ethylene- responsive transcription factors was elevated in t483 compared to WT (S4 Table, Fig 5B). These results suggest that altered phytohormone signaling might contribute to the morphological and growth defects observed in the t483 line.

Responses of the t483 mutant to gibberellins
The t483 mutant exhibited gibberellic acid (GA) deficiency phenotypes such as short roots, dwarfism, and late seed germination (Fig 1). In addition, the gene encoding gibberellin 20 oxidase 4 (OsGA20ox4), an active GA synthetase, was upregulated in t483 compared to WT (S4 Table). To evaluate if GA metabolism was altered in the mutant, endogenous GA content was measured in the t483 mutant and WT. Two weeks after germination, the levels of 13-OH GAs were significantly lower in the t483 mutant than in WT. In particular, GA 20 and active GA 3 levels were reduced by 50% compared to WT (Fig 6B). These results indicate that the GA synthesis pathway is defective in t483.
To further test the effects of GA in t483, WT and t483 mutant plants were treated with exogenous GA 3 . Seed germination vigor was improved in t483 by the addition of GA 3 (Fig 7A), but there was no change in the overall seed germination rate (data not shown). Seedlings were also treated to investigate the ongoing response of the t483 mutant to GA 3 . Root lengths of WT seedlings were unaffected by GA 3 treatment, but root lengths in t483 seedlings increased significantly to lengths similar to those in WT (Fig 7B and S1B Fig). Consistent with the morphological responses, expression of OsGA20ox4 was more responsive to GA 3 treatment in t483 roots than in WT roots (Fig 7C). Shoot length was enhanced equally in WT and t483 seedlings upon GA 3 treatment (Fig 7B and S1C Fig). In untreated shoots, OsGA20ox4 mRNA accumulated to a higher level in t483 than in WT, but OsGA20ox4 mRNA levels were decreased to similar CHR729 and Rice Seedling Development levels in t483 and WT when treated with GA 3 . In untreated roots, OsGA20ox4 mRNA accumulated to higher levels in t483 than in WT. On treatment with GA 3 , there was no change in accumulation of OsGA20ox4 mRNA in WT, but levels increased in t483 (Fig 7C). CHR729 expression also responded to GA 3 treatment (S1D Fig). These results suggest that t483 responds to GA and that exogenous GA 3 can overcome the mutant seed germination and root and shoot development defects, indicating that CHR729 regulates plant development through the GA pathway.

Discussion
A rice mutant was identified with pleiotropic phenotypic traits including dwarfism, lower tiller number, small panicle size, root growth inhibition, short and narrow leaves, and reduced chlorophyll content. The mutant gene responsible for the mutant phenotype was shown to be CHR729, which encoded a CHD3 protein. Our data and former reports demonstrated that mutation of this gene affected expression of a substantial number of downstream genes [16,17,29,35]. The large number of DEGs identified and the pleiotropic phenotype exhibited by the mutant suggested that CHR729 might be a hierarchical regulator of transcriptional cascades in plant growth and development. RNA-seq analysis identified 500 genes that were differentially regulated in t483 compared to WT, approximately 70% of which were upregulated in the mutant (S1 Table). About 10% of DEGs were transcription factor genes, and most of these were also upregulated in t483 (S2 Table). The mutation in the t483 line was predicted to lead to premature termination of CHR729. CHD3 proteins are generally considered to be transcriptional repressors, and the depression of transcription as a result of lower levels of active CHR729 is consistent with the upregulation of the majority of DEGs.
Hormones are signaling molecules that are vital to the life-cycles of many multicellular organisms [36]. There are eight major classes of plant hormones: auxin, ABA, gibberellin, ethylene, cytokinin, brassinosteroid, salicylic acid, and strigolactone. While each class performs specific functions, hormones regulate development in conjunction with one or more hormones of other classes [37,38]. Of the DEGs identified in this study, 19 were phytohormone-related genes up-or down-regulated by CHR729 (S4 Table). These genes encoded proteins that were involved in auxin, ABA, cytokinin, gibberellin, and ethylene functions. These results suggest that CHR729 affects different plant tissues throughout development through its influence on several different phytohormone pathways.
In general, seed dormancy and germination are regulated by the antagonistic actions of ABA and GA. ABA induces and maintains the dormant seed state [39,40], whereas GA negatively regulates dormancy by releasing coat-mediated seed dormancy and promoting or permitting seed germination [41]. Ethylene acts in concert with GA to promote seed germination, and can counteract the effects of ABA in seed germination [42]. Seed germination is thus a highly complex process that requires concerted interactions between ABA, ethylene, GA, and possibly other hormones. More research is needed to further elucidate the multiple roles of hormones in seed germination. The t483 mutant exhibited late germination and a low rate of germination. The expression of genes related to ABA, GA, and ethylene biosynthesis was altered in the mutant, as was signaling gene expression (S4 Table). Endogenous GA levels and levels of bioactive GA 3 were lower in the t483 mutant than in WT (Fig 6). Addition of exogenous GA 3 in the t483 line rescued the late seed germination phenotype (Fig 7A), but a similar effect was not seen with added ethylene (data not shown). This suggests that the t483 mutant is defective in GA biosynthesis and that late seed germination may be caused by the reduction in endogenous GA. As GA can regulate ethylene biosynthesis and the response to ethylene, it is possible that the differential expression of ethylene-related genes observed between WT and t483 might also be due to low endogenous GA levels [43]. The t483 mutant line thus constitutes a valuable genetic resource as not only can it be used to investigate the mechanisms underlying control of seed germination by phytohormones, but it can also be used to examine CHD3-mediated control of seed germination via the phytohormone pathway.
GA plays a key role in the regulation of plant growth and development, and many GA biosynthetic, catabolic, and signaling proteins have been identified [44]. While GAs are primarily known as growth-promoting hormones, they can also function as negative regulators. For example, bioactive GA in rice positively regulates seed germination and root elongation [41], but negatively regulates tillering through the Homeobox 1 and Teosinte branched1 pathway [45][46][47]. GA also negatively controls adventitious root development [48]. The t483 mutant displayed GA deficiency phenotypes such as short roots, dwarfism, and late seed germination (Figs 1 and 2). Endogenous GA assays in WT and t483 lines showed most of the 13-OH GAs to be present in both WT and t483. GA levels were lower in t483 compared to WT, but the bioactive GA 1 was not detected in either WT or t483. Although this differed from a previous report, in which GA 1 was detected in rice seedlings in the vegetative stage [49], the conflicting results might be a consequence of different sample timings. Here, treatment with exogenous GA 3 rescued most of the phenotypic effects seen in the mutant (Fig 7 and S1 Fig). GA deficiency generally causes thick culms; however, tiller number was reduced in t483. It is possible that the CHR729-controlled genes that are involved in tillering control do not act antagonistically to the GA pathway.
GA20ox is a key enzyme that catalyzes the penultimate steps in bioactive GA synthesis from GA 53 /GA 12 to GA 20 /GA 9 . Rice contains four GA20ox-like genes. GA20ox2, also named SD-1 (semi-dwarfing gene), is known as the rice "green revolution gene" due to the dramatic yield increases (and semi-dwarfism) seen in sd-1 lines [50]. A second gene, GA20ox1, also affects plant stature, and its overexpression results in tall plants and a GA-overproduction phenotype [51]. GA20ox3 contributes to disease resistance as well as to plant stature [52]. The function of GA20ox4 remained unclear until now. RNA-seq and qRT-PCR analysis performed in this study showed that GA20ox4 was upregulated in t483 compared to WT. This result was contradictory when considered in the context of the functions of the other three GA20ox genes. Specifically, higher GA20ox levels would be expected to produce a GA-overproduction phenotype, but the t483 mutant contained low GA levels and exhibited features characteristic of GA deficiency. This suggests that GA20ox4 plays a different role to the other GA20ox genes in GA metabolism.
CHR729 appeared to negatively control the OsGA20ox4 gene. Compared with WT, CHR729 expression decreased and OsGA20ox4 increased in the t483 mutant. GA 3 supplementation data were consistent with these observations (Fig 7 and S1 Fig). Previous research showed that the CHD3 protein regulated target gene expression via H3K27me3 [17]. It is possible that control of OsGA20ox4 by CHR729 may be regulated in a similar manner. Further studies will enhance our understanding of CHR729 and its regulatory mechanisms.