Flowering-Related RING Protein 1 (FRRP1) Regulates Flowering Time and Yield Potential by Affecting Histone H2B Monoubiquitination in Rice (Oryza Sativa)

Flowering time is a critical trait for crops cultivated under various temperature/photoperiod conditions around the world. To understand better the flowering time of rice, we used the vector pTCK303 to produce several lines of RNAi knockdown transgenic rice and investigated their flowering times and other agronomic traits. Among them, the heading date of FRRP1-RNAi knockdown transgenic rice was 23–26 days earlier than that of wild-type plants. FRRP1 is a novel rice gene that encodes a C3HC4-type Really Interesting Novel Gene (RING) finger domain protein. In addition to the early flowering time, FRRP1-RNAi knockdown transgenic rice caused changes on an array of agronomic traits, including plant height, panicle length and grain length. We analyzed the expression of some key genes associated with the flowering time and other agronomic traits in the FRRP1-RNAi knockdown lines and compared with that in wild-type lines. The expression of Hd3a increased significantly, which was the key factor in the early flowering time. Further experiments showed that the level of histone H2B monoubiquitination (H2Bub1) was noticeably reduced in the FRRP1-RNAi knockdown transgenic rice lines compared with wild-type plants and MBP-FRRP1-F1 was capable of self-ubiquitination. The results indicate that Flowering Related RING Protein 1 (FRRP1) is involved in histone H2B monoubiquitination and suggest that FRRP1 functions as an E3 ligase in vivo and in vitro. In conclusion, FRRP1 probably regulates flowering time and yield potential in rice by affecting histone H2B monoubiquitination, which leads to changes in gene expression in multiple processes.

Introduction residue is in the fifth of the eight Zn coordinating sites [28][29][30]. Most RING finger proteins are potential ubiquitin ligases.
Protein ubiquitination is achieved through the action of the E1, E2 and E3 ubiquitin ligases [31][32][33][34]. There are two main types of protein ubiquitination: multiubiquitination and monoubiquitination. Most of multiubiquitination mark proteins to be degraded by the 26S proteasome, whereas monoubiquitination mediates the localization or activation of the ubiquitinated proteins [35,36].
Ubiquitination plays an important role in regulating cellular activity by coordinating it with other posttranslational modifications. Numerous proteins are labeled ubiquitin; one of them, histone ubiquitination, is mostly associated with histone H2B and H2A. Ubiquitination of histones is mainly involved in regulating eukaryotic gene expression by altering chromatin structures [37] and controlling the transcription machinery [38,39].
In plants, several RING finger proteins such as E3 ubiquitin ligases have been predicted or known functions in the control of flowering, cell cycle, hormone signaling and stress responses [40][41][42][43][44].
In this study, we cloned a gene, Flowering Related RING Protein 1 (FRRP1) (Os10g0565600), that encodes a RING finger protein from rice. Down-regulation transcript of FRRP1 by RNAi knockdown in transgenic rice plants produced an early flowering phenotype. Interestingly, we also found changes in an array of agronomic traits. Therefore, further researches were performed on the FRRP1-RNAi knockdown transgenic rice. The results indicated that the changes in expression of several key genes probably were the reasons for the phenotypes observed.

Plant Materials and Growth Conditions
We used Nipponbare (Oryza sativa L. japonica. cv. Nipponbare) as the WT plant in our experiments. Plants for experimental analysis were grown in normal rice growing seasons under natural field conditions in the Shangzhuang Experimental Station of China Agricultural University, Beijing, China. Rice seeds were sown in a seed bed in early May and then transplanted to the field in early June. Field management, including irrigation, fertilizer application and pest control followed usual agricultural practice.
The seeds of Arabidopsis used in this study were ecotype Columbia-0 (Col-0). For the mutant complementation assay, the seeds of the hub2 mutant were used, corresponding to SALK_071289 [45]. For planting, the seeds were kept at 4°C for 3 days, then germinated on Murashige and Skoog (MS) medium. After 6-7 days of growth under continuous white light, the plants were transplanted to soil and grown under conditions with daily cycles of 16 h of light at 22°C and 8 h of dark at 18°C.

Cloning FRRP1 and Sequence Analysis
Rice variety Nipponbare were field grown under normal conditions in the greenhouse of China Agriculture University. RNA samples were prepared from leaves. The complementary DNA (cDNA) samples were synthesized using M-MLV reverse transcriptase (Promega, USA). The protocols followed a previous method [46]. The PCR amplification conditions were 95°C for 5 min; 35 cycles of 95°C for 1 min; 55°C for 1 min; and 72°C for 3 min 30 s; with a final extension at 72°C for 10 min, with primers FRRP1-F and FRRP1-R (S1 Table). The amplified 2535 bp coding sequence (CDS) fragment was sequenced and then inserted into pMD18-T Simple Vector (TaKaRa, Japan) and sequenced. Sequence analysis was performed with DNAMAN (version 6). The phylogenetic tree was constructed in MEGA (version 5.1) using the neighborjoining method, and bootstrap analysis was performed with 1000 replications.

Plasmid Construction and Rice Transformation
We had known the HUB2 affected the Arabidopsis flowering, so we found FRRP1 in rice database by blast with HUB2 sequence. Then the mutant population were generated by construct RNAi plasmid. To construct the FRRP1-RNAi plasmid, a 500-bp fragment of FRRP1 CDS, only identity with the FRRP1 CDS in the rice genome, was amplified using primers FRRP1-303-F1/FRRP1-303-R1 and FRRP1-303-F2/FRRP1-303-R2 (S1 Table) and inserted into the pTCK303 vector. The final vector pTCK303-FRRP1 was introduced into Nipponbare via Agrobacterium tumefaciens mediated transformation to produce RNAi knockdown transgenic lines. Details of the protocols were as described previously by Wang et al. [47] with a few modifications.

PCR and Histochemical Analysis of GUS Activity
To determine the integration of the pTCK303-FRRP1 in the genome, PCR was performed using specific primer pairs (FRRP1-Rcheck-F and FRRP1-Rcheck-R) (S1 Table) to amplify the specific fragment region from T 0 and T 3 generation transgenic rice plants. Genomic DNA (gDNA) was extracted and purified from young leaves using the cetyl trimethyl ammonium bromide (CTAB) method [48]. For histochemical GUS analysis, the roots of T 0 and T 3 generation transgenic rice samples were analyzed according to the protocol described by Jefferson et al [49].

Phenotypic Data Collection
The heading date was the day when the first panicle of the plant emerged to a height of 2 cm. The panicle length was the average length of three main panicles measured from the bottom neck to their tips for each plant. Plant height was measured from the ground to the tip of the tallest tiller of the plant. We measured one main panicle three times, from which 10 randomly selected grains were aligned one by one and the average length was recorded as the grain length. A Student's t-test was used in all statistical analyses (P0.05).

Semiquantitative Reverse Transcription PCR and Quantitative Real-Time PCR
RNA samples were prepared from leaves about 1 month before WT plants flowering (expression analysis of key genes for flowering time). The cDNA samples were synthesized by use of M-MLV Reverse Transcriptase (Promega, USA). The protocols followed a previous method [46]. qPCR analysis was performed with a CFX-96 real-time system (Bio-Rad) using SYBR Premix Ex Taq Mix (TaKaRa, Japan). LA Taq polymerase (TaKaRa, Japan) was used for semiquantitative RT-PCR with the primers qFRRP1-F and qFRRP1-R. Several primers from previous papers are listed in S1 Table [45,50,51].

Genetic Complementation in Arabidopsis
For complementation of Arabidopsis hub2 mutants (SALK_071289), the CDS of FRRP1 was cloned into the pCAMBIA vector 1305.1 with primers FRRP1-1305.1-F and FRRP1-1305.1-R (S1 Table), downstream of the CaMV35S promoter. The resultant construct was introduced into hub2 mutant plants by the floral dip method with Agrobacterium tumefaciens (strain EHA105) [45].

In Vivo H2B Monoubiquitination Assay
For the detection of overall H2Bub1 levels between both WT plants and the FRRP1-RNAi transgenic rice, total proteins were extracted from 70-day-old plants with plant extraction buffer (0.25 M NaCl, 1% sodium dodecyl sulfate, 1% β-mercaptoethanol, 0.05 M phosphatebuffered saline, pH 7.4, 1 mM phenylmethylsulfonyl fluoride). The proteins were then separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes for blotting with an antibody against ubiquityl-histone H2B (H2Bub1) monoclonal (Cell Signaling). Histone H3 was used as a loading control and was detected with an anti-histone H3 antibody (Cell Signaling). The protocols followed a previous method [52].
In Vitro Ubiquitination Assay

Knockdown of FRRP1 Resulted in Pleiotropic Phenotypes, Especially Early Flowering in Rice
To obtain further knowledge of floral induction in rice, we used RNAi to produce transgenic plants and measured flowering time in each line. We found that one of the transgenic plants showed significantly early flowering time compared with wild type (Nipponbare). We cloned the full-length cDNA of the gene disrupted in this RNAi line and found it comprised of 2873 base pairs (bp). Its open reading frame (ORF) was 2535 bp, corresponding to an 844-aminoacid polypeptide with a C3HC4-type RING in the C-terminal. It had a predicted molecular mass of 96.5 kilodaltons (kDa) and calculated isoelectric point (pI) of 4.90 ( Fig 1A). We named the protein Flowering Related RING Protein 1 (FRRP1) (Os10g0565600).
Thirteen T 0 transgenic FRRP1-RNAi rice lines were produced after several rounds of selection by polymerase chain reaction (PCR) and β-glucuronidase (GUS) staining (Fig 1B and 1C). Three independent lines randomly selected from the T 3 populations derived from these lines were labeled RL1, RL2 and RL3 and used for phenotypic analysis. Further quantitative reverse transcription (qRT)-PCR analysis of FRRP1-RNAi transgenic plants confirmed that expression of FRRP1 in the transgenic lines RL1, RL2 and RL3 were reduced to 40%, 48% and 46% of the wild-type (WT) level, respectively (Fig 1D), establishing that the knockdown of FRRP1 were implemented successfully in the selected transgenic lines.
To analyze the total days to heading, seeds from the T 3 generations were planted in natural long-day (NLD) conditions (day length >14 h). Nipponbare was the wild type used as a control. Compared with the 118.5 days to heading for WT plants, the three transgenic knockdown lines flowered 23-26 days earlier under the same growth conditions ( Fig 1E, Table 1). This suggested that knockdown of FRRP1 significantly accelerated flowering in rice under NLD condition. In addition, FRRP1-RNAi knockdown transgenic plants exhibited other phenotypic traits: their plant height were 10-20 cm higher than that of WT plants, and the grain length were 0.06-0.08 cm longer compared with that of WT plants. Also, panicle length showed a change to some extent in NLD conditions, and two lines were 2cm longer than that of WT plants (Fig 1F and 1G, Table 2)

FRRP1 Complements the Arabidopsis hub2 Mutant
Most RING finger proteins were E3 ligases [53], and FRRP1 was 50% identical with AtHUB2 (At1g55250), an E3 ligase in Arabidopsis regulating flowering time and involved in H2Bub1 (On the histone H2B plus a ubiquitin protein, such as transcription regulation process, not the degradation of protein) [45]. To determine how FRRP1 regulated flowering way, a genetic complementation assay was performed by transforming FRRP1 under control of a constitutive CaMV 35S promoter into Arabidopsis hub2 mutant plants [45]. We obtained nine 35S:FRRP1/ hub2 plants lines, three of which were complemented. The T 3 generation homozygotes of transgenic hub2 harboring FRRP1 were used for observation and analysis. The flowering time of the transgenic hub2 mutant was complemented by rice FRRP1 (Fig 3A and 3B, Table 3). Furthermore, qPCR analysis of the FLC-clade genes demonstrated that expression of FLC and MAF5 in the transgenic hub2 plants can be restored to the level in WT plants (Fig 3C). To verify whether FRRP1 affects gene expression through its E3 ligase activity, we extracted total proteins from the FRRP1-RNAi transgenic lines and used the monoclonal antibody against ubiquity1-histone H2B(H2Bub1) to detect the monoubiquitinated histone H2B in the transgenic rice plants, using histone H3 as a loading control. The H2Bub1 level was significantly reduced in the FRRP1-RNAi knockdown transgenic lines compared with that of WT plants (Fig 4). This suggested that the FRRP1 functioned as an E3 ligase and monoubiquitinated histone H2B in vivo.

FRRP1-F1 Have the Ability of Self-Ubiquitination In Vitro
To get more evidences that FRRP1 is an E3 ligase, a ubiquitination test was performed in vitro.
Because we couldn't obtain the full length and soluble protein of FRRP1, we created a truncated protein, named FRRP1-F1. It was the C-terminal region of FRRP1 (amino acid residues 786 to 844). In order to find out the key amino acid involved in the function of E3 ligase activity in the RING domain, Cys 830 was substituted with Ser 830 in the FRRP1-F1, named FRRP1-F1 C830S . Then FRRP1-F1 and FRRP1-F1 C830S were fused with a maltose-binding protein (MBP-FRRP1-F1 and MBP-FRRP1-F1 C830S ), expressed in E.coli TB1, and purified using amylose resin. The ubiquitinated protein MBP-FRRP1-F1 of high molecular weight was detected at the presence of E1 (ubiquitin-activating enzyme), E2 (ubiquitin-conjugating enzymes), ubiquitin and ATP, suggesting that MBP-FRRP1-F1 was capable of self-ubiquitination (Fig 5A and  5B, lane 2). But the mutant MBP-FRRP1-F1 C830S did not have the ability of self-ubiquitination  Fig 5A and 5B lane 7). The in vitro ubiquitination tests identified FRRP1-F1(FRRP1) as E3 ligase. However, FRRP1-F1 C830S had lost the E3 ligase activity in vitro. And the results of ubiquitination reaction assay also showed that the amino acid Cys 830 in the RING domain was required for the activity of FRRP1-F1(FRRP1).

The Expression Profiling of Key Genes for Flowering Time and Other Traits in the Knockdown Transgenic Rice Lines
To explain the molecular mechanism affecting the phenotypic alteration in the transgenic rice, expression of several key genes was evaluated, such as OsMADS50 [60], Hd1, Ehd1, Hd3a and RFT1. The FRRP1-RNAi knockdown transgenic lines and the WT plants were grown under natural field conditions(NLD condition) (day length >14h). About 30 days before rice flowering, we collected samples for analysis. The expression of OsMADS50, Hd1 and Ehd1 was reduced and RFT1 was barely detectable in the transgenic rice compared to that in WT plants.
By contrast, the transcript of Hd3a was greatly increased in the transgenic lines, whereas it was almost undetectable in the WT plants (Fig 6). These results showed that it was the final result that early flowering in the FRRP1-RNAi knockdown transgenic lines was due to the altered expression of a combination of genes regulating flowering. Upregulation of Hd3a corresponded with earlier flowering in the transgenic lines. We also examined expression of a cluster of genes (RID1/OsID1/Ehd2, Ghd7, DTH8/Ghd8 and HGW) that controlled both flowering time and other agronomic traits [3]. The transgenic lines all showed reductions in mRNA levels of these genes, except HGW, whose expression increased (Fig 7). Possibly the changes of the expression of the aforementioned genes varied the phenotypes of the transgenic plants.
Brachypodium distachyon, a RING domain protein (XP_008668811.1) from Zea mays, a RING domain protein (GenBank accession number XP_002530869.1) from Ricinus communis, histone monoubiquitination 2 (AtHUB2) (At1g55250) from Arabidopsis thaliana and a RING domain protein (GenBank accession number XP_009801022.1) in Nicotiana sylvestris. Eight amino acid residues conserved in and characteristic of the RING domain proteins are framed in red, and these amino acid residues were identical in all six proteins. doi:10.1371/journal.pone.0150458.g002    After examining the known flowering-related genes, we further investigated whether FRRP1 acts through GIF1, GW2, GW5 and GS3 to regulate grain size and weight. The results showed the expression of GW2 and GS3 was up-regulated, whereas expression of GW5 and GIF1 was down-regulated (Fig 8), suggesting that FRRP1 also affected the expression of grain-related genes in transgenic rice.
From all of the results, we concluded that the most important and direct reason for the significant early flowering of transgenic rice plants was the marked increase in the expression of Hd3a. The trait changes in transgenic rice may be regulated collectively by a series of genes involved in controlling flowering time and yield potential, as all of them were possibly affected via FRRP1.

Basic Function of the RING Proteins such as FRRP1 Is Conserved between Dicotyledonous and Monocotyledonous Plants
Here we performed characterization and analysis of FRRP1, a RING finger protein from rice. Sequence alignment showed that FRRP1 was 50%-80% identical with some other RING finger proteins in different plant species (Fig 2B), and the RING domains featured a more significant sequence identity (~90%-97%). The high level of identity of these RING finger proteins from different plant species suggested that they were highly conserved across both dicotyledonous and monocotyledonous plants. A detailed analysis obtained from the phylogenetic tree is shown in Fig 2A. AtHUB2 is a RING finger protein and an E3 ubiquitin ligase. A defect in H2Bub1 in the hub2 mutant of Arabidopsis inhibited expression of genes in the FLC-clade and thus accelerated the transition to flowering in the hub2 mutant of Arabidopsis [41,44,45]. H2Bub1 levels were reduced in the FRRP1-RNAi knockdown rice lines used in our experiments (Fig 4). At the same time, transformants of Arabidopsis hub2 plants harboring overexpressing FRRP1 flowered at the same time as WT plants (Fig 3A). Furthermore, there was no difference in the transcript levels of FLC and MAF5 between the transformants and WT plants, although the expression of MAF4 was inhibited in the transformants (Fig 3C). Such an inconsistency of these three genes in 35S:FRRP1/hub2 plants and WT Arabidopsis may reflect functional divergence between dicotyledonous and monocotyledonous orthologs. In support of this claim, the Arabidopsis hub2 mutant showed dwarfism [45], but FRRP1-RNAi transgenic rice lines were taller than the WT (Fig 1E). These observations suggested a combination of conserved and divergent functions of FRRP1 compared with AtHUB2.
Increased Hd3a Transcript Level Accounts for Early Flowering of FRRP1

Knockdown Transgenic Rice under NLD Conditions
Knowledge obtained from the considerable efforts made to date in studies of flowering-time regulation has facilitated an understanding of molecular mechanisms controlling heading date in rice. Hd3a (a major floral activator under SD conditions) and RFT1 (a major floral activator under LD conditions) were the most important floral activators, as mobile flowering signals acting to promote the floral transition [8,12,15]. Short days activate the pathway leading to flowering in rice. Hd3a is activated by Hd1, and also by Ehd1, which encodes a B-type response regulator and functions independently of Hd1 under SD conditions [10,14,16]. Furthermore, RFT1 was activated later during the development of SD flowering in Hd3a RNAi plants [15]. However, Under LD conditions, flowering time is regulated by both activation and suppression pathways interacting in a network. Hd1 suppresses Hd3a expression under LD conditions [14,15], and RFT1, with its both positive (OsMADS50 and Ehd1) and negative (Hd1 and Ghd7) regulators form a gene network to control LD flowering in rice [8].
We constructed FRRP1-RNAi knockdown transgenic rice, and found the transgenic rice plants displayed a significant phenotype of early flowering. We investigated the expression of the genes related to flowering time in transgenic rice and the WT plants grown under the same natural field in NLD conditions (Figs 6 and 7). Our qRT-PCR results revealed that the mRNA expression of OsMADS50, Ghd7, DTH8, RID1/OsID1/Ehd2, Ehd1, Hd1 and RFT1 was reduced in the FRRP1-RNAi knockdown transgenic rice lines compared with the expression of these genes in WT plants, but transcription of Hd3a increased significantly. According to previous studies, under LD conditions, Hd3a and RFT1 were both suppressed by Hd1 and Ghd7, but RFT1 was promoted by OsMADS50 and Ehd1. Furthermore other studies indicated Ehd1 was up-regulated by OsMADS50 and RID1/OsID1/Ehd2 but suppressed by Hd1, Ghd7 and DTH8 under LD conditions. Hence all of the regulation mechanisms involved with Ehd1 may be independent of each other, and indicated that Ehd1 integrated multiple pathways in the regulation of flowering time [1,7,8]. We conjectured that the genes worked concertedly on Ehd1, which resulted in a significant decrease of the Ehd1 transcript and in the reduced expression of OsMADS50, which together decreased the expression of RFT1. But the lessened expression of Hd1 and Ghd7 imparted a significant increase to the Hd3a expression as shown in our results (Fig 6). Finally, an early flowering time in our knockdown transgenic rice was observed ( Fig  1E, Table 1). The results indicated that there was a net genetic regulation promoted NLD flowering via Hd3a but not RFT1 in control of multiple genes.

Pleiotropic Effects of FRRP1 May Be a Result of a Reduction in H2Bub1 Level
A number of genes involved in controlling heading date, grain size or plant height have been reported recently [1,3,7,61]. Ghd7 has been proven to have pleiotropic effects on flowering time, plant height and on rice cultivars adapting to cold climate regions [1]. DTH8 not only delayed rice flowering but also played an important role in regulating plant height and grain number [7]. Besides early flowering time, FRRP1-RNAi knockdown transgenic rice displayed an array of phenotypes, such as higher plant height and longer grain length. These implied FRRP1 was important for rice development. Therefore the FRRP1 and FRRP1-RNAi knockdown transgenic rice could be used to breed superior rice through genetic engineering and/or conventional breeding.
Up to now, posttranslational modifications of histone, such as acetylation, methylation, phosphorylation and monoubiquitination have been associated with control of several complex processes including plant development and adaptation to environmental conditions [62][63][64]. Histone H2B modification by monoubiquitination is mainly involved in transcriptional activation and plays a key role in the plant life cycle, such as seed germination and initiation of flowering [65]. For instance, the monoubiquitination of histone H2B was proven to be essential for the enhancement of H3K4 and H3K36 hypermethylation and FLC/MAF transcriptional activation in Arabidopsis, the deficiency of which leads to earlier transition to flowering time in the hub2 mutant [43,44,45].
As changes in transcripts of a number of genes were associated with displayed phenotypes of flowering time and yield potential in the FRRP1 knockdown transgenic rice (Figs 6-8; Tables 1 and 2), the underlying reason was probably the reduction in H2Bub1 level. This may trigger alteration of expression in certain master genes, then cause a series of changes in their regulated pathways and finally produce discernible pleiotropic effects of the phenotypes. In addition, the MBP-FRRP1-F1 had the ability of self-ubiquitination in vitro( Fig 5). Thus, we argue that FRRP1 regulated flowering time, plant height and yield potential by monoubiquitinating histone H2B; however, more evidence is required to confirm this hypothesis.

Conclusion
We obtained several FRRP1-RNAi knockdown transgenic rice lines and the rice showed an early flowering and high plant, long grain phenotypes. The transgenic lines flowered much earlier, 23-26 days before the WT rice plants. The phenotype of flowering time was excellent in transgenic rice. To understand the reason and determine the function of FRRP1, we performed qRT-PCR and verified lower levels of histone H2B monoubiquitination of the transgenic rice. And in vitro we confirmed the MBP-FRRP1-F1(FRRP1) had the ability of self-ubiquitination. In addition, we performed some tests in Arabidopsis. All of the results showed the increase in levels of Hd3a was the main reason for the rice early flowering, and many other genes such as Ehd1, Hd1, Ghd7 formed a network regulating flowering time and other agronomic traits. H2Bub1 decreased in the transgenic rice and the ability of self-ubiquitination with MBP-FRRP1-F1(FRRP1) in vitro, which suggests that FRRP1 was an E3 ligase in vivo and in vitro, regulating flowering time, plant height and grain length.
Supporting Information S1