| Edited by Masao Tasaka. Yukimoto Iwasaki: Corresponding author. E-mail: iwasaki@fpu.ac.jp |
As the world population continues to grow, food demand is rising, requiring a continuous increase in food supplies. Cereals are the most important source of food, therefore higher production in crop plants may prove to be necessary to satisfy the increasing demand in food. Rice yield potential is determined by several factors including seed size (or weight), number of panicles per plant and number of seeds per panicle (Takeda and Matsuoka, 2008; Song and Ashikari, 2008). Among these factors, we have been focusing on the study of seed size and continue to identify regulatory genes (Fujisawa et al., 1999; Tanabe et al., 2005; Kitagawa et al., 2010), in order to understand the molecular mechanisms that underline seed size regulation. Knowledge of these genes could be useful for breeding purpose.
Causal genes of small (or short) seed phenotype have been identified by us and other research groups through map-based cloning, namely D1 (also named RGA1) encoding the heterotrimeric G protein α subunit (Ashikari et al., 1999; Fujisawa et al., 1999), D11 encoding a cytochrome P450 involved in brassinosteroid (BR) biosynthesis (Tanabe et al., 2005), BRD1 and D2 encoding different types of cytochrome P450 that are also involved in BR synthesis (Hong et al., 2002, 2003), BRD2 (also named rice Dim/dwf1) encoding an enzyme associated with BR biosynthesis (Hong et al., 2005), D61 (also named OsBRI1) encoding a BR receptor (Yamamuro et al., 2000), and finally SRS3 encoding the kinesin 13 protein (Kitagawa et al., 2010). These studies have shown that both the heterotrimeric G protein signaling pathway and the BR signaling pathway play a major role in the regulation of seed size. During seed formation in rice, it was shown that D1 regulates the cell number (Izawa et al., 2010) and SRS3 regulates the cell length (Kitagawa et al., 2010), whereas BR-related genes seem to regulate both cell number and cell elongation (Yamamuro et al., 2000; Hong et al., 2002; Nakamura et al., 2006). Therefore, there are different types of functional genes that regulate seed size through the control of either cell length, cell number, or both.
There have been studies in which these regulatory genes were used to improve the seed size in rice. One study showed the use of a chimeric gene, encoding a constitutive active form of the heterotrimeric G protein α subunit (QL). When QL was introduced into the rice mutant d1, which is defective for the α subunit gene, the seed length and weight were substantially increased in the transformants (Oki et al., 2005). Similarly, another study showed that, when a chimeric gene of a sterol C-22 hydroxylase, involved in maize BR biosynthesis (Zm-CYP) was expressed in rice plants, seeds were heavier in the transformants than in the WT (Wu et al., 2008). These studies suggested that the enhancement of G protein signaling could lead to an increase in seed size and by controlling the BR levels through the manipulation of BR biosynthesis related genes, it is possible to increase grain yield in rice. Further progress in our understanding of the regulation of seed formation through molecular and genetic studies may prove to be useful for the improvement of rice yield.
A number of genes that regulate seed size have been identified in rice (Nagato and Yoshimura, 1998). Previously we have reported the rough map position on chromosome 7 of a new gene responsible for the small and round seed phenotype. The gene determining seed size by regulating the size of lemma and palea was named the SMALL AND ROUND SEED1 (SRS1) gene (Tanabe et al., 2007). Here we report that the SRS1 gene encodes a novel protein that has no known functional domains.
Oryza sativa L. cv. Taichung 65 (T65) (japonica) and Fujiminori (japonica) were used as recurrent parents in this study. These recurrent parents were abbreviated as WT in the paper. Three mutants, 20KY2036, TCM1201 and 00T69 were obtained from the T65-library mutagenized by N-methyl-N-nitrosourea treatment. A mutant, Fu-1 was obtained from the Fujiminori-library mutagenized by γ-irradiation. The recurrent parent for a mutant, Kaihenaikoku d1 was unknown. 20KY2036 was allelic to TCM1201. We roughly mapped the causal gene of 20KY2036 (srs1) at 81–94 cM on chromosome 7 (Tanabe et al., 2006). As described this work, we found that TCM1201, 00T69, Fu-1 and Kaihenaikoku d1 have a mutation in the SRS1 gene. In this paper, 20KY2036 (srs1), TCM1201, 00T69, Fu-1 and Kaihenaikoku d1 were renamed srs1-1, srs1-2, srs1-3, srs1-4 and srs1-5, respectively. Rice plants were grown in a greenhouse at 30°C (day) and 25°C (night) or in the field.
For SRS1 mapping, we used Kasalath (indica) as a parental line to develop F2 population. The SRS1 locus has previously been mapped to the long arm of chromosome 5 (Tanabe et al., 2007). To increase mapping resolution, we analyzed the genomic DNA from leaves of 1500 F2 plants bearing small and round seeds (F3 seeds). The genomic DNA was extracted by the cetyltrimethylammonium bromide method (Murray and Thompson, 1980). The genetic linkage between the SRS1 locus and molecular markers was determined by PCR using the sequence tagged site (STS) markers reported by the Rice Genome Research Program (http://rgp.dna.affrc.go.jp/J/index.html) and microsatellite markers (McCouch et al., 2002).
In addition to these markers, eight primers for the two STS markers (25249 and 25363) and two SNPs markers (25285As and 25340As) on rice chromosome 7 were used: 25249F (5’-GGAAAGCATCGAGTACAATA-3’) and 25249R (5’-AAACCTGAAACCTGATGGC-3’), 25363F (5’-CGGTTTCATGGACTTTGCAC-3’) and 25363R (5’-GTGTCAAGATCGACGACAGC-3’), 25285AsF (5’-TCTTGGAACAGCAACAGCCA-3’) and 25285AsR (5’-CAACTGCCTGGTTTGTGTGTG-3’), 25340AsF (5’-TCTATACGCGGTTCCGGTAC-3’) and 25340AsR (5’-GGAGTTCTGTATGAGATAC-3’C). The positions of markers 25249, 25285As, 25340As, 25363 were 25,963K, 26,000K, 26,054K and 26,077K on rice chromosome 7 by the rice annotation project database, build 5 (http://rapdb.dna.affrc.go.jp/). The PCR conditions were 94°C for 2 min, 94°C for 30 s, 55°C for 30 s, 72°C for 1 min in 30 cycles.
For the identification of the causal gene in srs1, more than 96 sets of PCR primers were designed in the candidate region (55 kb) between marker 25285As and 25340As. The amplified DNA fragments were directly sequenced without cloning, using the same primers as for amplification.
For the identification of the mutation site in the srs1-1 mutants, 15 sets of PCR primers in which were covered 10 exons in the SRS1 gene were designated. The amplified DNA fragments using genomic DNA were directly sequenced without cloning, using the same primers as for amplification.
In srs1-2 and srs1-5, mutations were detected in the intron 3. In order to confirm the possibility of the abnormal splicing in the srs1-2 and srs1-5 mutants, SRS1-2 and SRS1-5 cDNAs were prepared. Briefly, total RNA was extracted from the mutants using the RNeasy Mini Kit (QIAGEN, Hilden, Germany) and cDNAs were synthesized from total RNA and T-tailed primer using the SuperScript III system (Invitrogen, Carlsbad, CA, USA). For the analysis of SRS1-2 and SRS1-5, the cDNA fragment covering from exon 3 to exon 7 was amplified using the following primers (SRS1-E3-F, 5’-AGAGGTTCGTGCGGTTCGTG-3’) and (SRS1-E7-R, 5’-TGATGACACCGAGTAATGAG-3’) and the KOD-FX (TOYOBO Ltd., Japan). The PCR conditions were 94°C for 2 min, 98°C for 10 s and 68°C for 3 min in 30 cycles. The amplified cDNA fragments were directly sequenced without cloning on an ABI PRISM 3100 Genetic Analyzer, using SRS1-E3-F primers.
For the quantification of SRS1 mRNA, real-time RT-PCR was carried out with the SYBR Premix Ex TaqTM II (Takara-Bio., Tokyo, Japan). Two primers, (SRS1-RT-PCR-F; 5’-CCAGTTGAGCGTTTCCTCTG-3’ / SRS1-RT-PCR-R; 5‘-GGCTCATGTTGGCAAGATTG-3’) were used for amplify 227-bp cDNA fragment of the SRS1 cDNA. Two primers, (forward; 5’-CTGTTCTAGGGTTCACAAGTCTGC-3’ / reverse; 5’-GGACACAATGATTAGGGATC-3’) were used for amplify OsUbiquitin1 gene (Accession number, Os060681400). PCR products were quantified by real-time RT-PCR using a Thermal Cycler Dice Real Time System (Takara-Bio., Tokyo, Japan).
A cDNA fragment encoding the amino acid sequence from 844 to 1221 of SRS1 protein was amplified by PCR using pENTER-SRS1 as a template. The fragment was digested with KpnI and SacI (TOYOBO, Osaka, Japan) and ligated into the pET51b vector (Novagen, Merck, Darmstadt, Germany) digested with KpnI and SacI (TOYOBO, Osaka, Japan) to obtain pET51b-SRS1. The vector was transformed in BL21(DE3) pLys cells (Novagen, Merck Ltd., Japan). The SRS1-His-tagged protein was induced and purified according to a previously described method (Iwasaki et al., 1997).
Polyclonal antibody against the SRS1 protein was raised in rabbits by immunization with purified His-tagged SRS1 protein.
Seedlings grown in a growth chamber under 14 hr light at 30°C and 10 hr dark at 25°C conditions for 7 days were used for the preparation of the intracellular membrane fractions, according to previously described method (Iwasaki et al., 1997). Briefly, the aerial parts were homogenized with three volume of the grinding buffer (0.5 M sucrose, 75 mM 3-N-morpholino propansulfonic acid (MOPS), 5 mM ethylenediaminetetraacetic acid (EDTA), 5 mM ethyleneglycol bis(2-aminoethylether)tetraacetic acid (EGTA) and 10 mM KF pH7.6). The homogenate was centrifuged at 1,000 g for 10 min and the pellet was designated as 1K. The supernatant was centrifuged at 10,000 g for 10 min and the pellet was designated as 10K. The supernatant was centrifuged at 100,000 g for 60 min. The pellet was designated as 100K and the supernatant was designated as the soluble fraction (Sup).
Proteins were separated by SDS-PAGE on a 5–20% gradient gel (Bio Craft, Tokyo, Japan) and blotted onto PVDF membrane (Immobilon-P; Millipore, Bedford, USA). Western blotting was carried out as described previously (Kato et al., 2004). The amount of SRS1 proteins was quantified using an imaging device, Multi Gauge Ver 2.2 (LAS-1000plus, Fujifilm, Tokyo, Japan).
Total RNAs and total proteins were prepared from the same sample. First, total RNAs were extracted from 100 mg of tissues, using an RNeasy Mini Kit (QIAGEN, Hilden, Germany). Total RNAs were used for real time RT-PCR analysis. Total proteins were prepared from the flow-through fractions of the RNeasy Mini Kit, as previously described (Morse et al., 2006). Proteins obtained from 100 mg of tissues were dissolved in 400 μl of a dissociation buffer (50 mM Tris-HCl pH6.8, 5% glycerol, 2% SDS and 4% 2-mercaptoethanol) for SDS-PAGE and western blot analysis (Iwasaki et al., 1997).
We previously isolated the small and round seed1 mutant (srs1) and roughly mapped the chromosomal location of SRS1, the causal gene, at 81–94 cM on chromosome 7 (Tanabe et al., 2007). We also collected more than 50 mutants with small and round seed phenotypes from various genetic backgrounds and found that four mutants roughly mapped to a region near srs1. As described in this work, we found that all four mutants have a mutation in the SRS1 gene. In this paper, 20KY2036 (srs1), TCM1201, 00T69, Fu-1 and Kaihenaikoku d1 were renamed srs1-1, srs1-2, srs1-3, srs1-4 and srs1-5, respectively.
The phenotype of the srs1 mutants was compared with that of each corresponding WT from which the mutants were generated (Fig. 1). Overall, the srs1 mutants showed a shorter plant phenotype, smaller grains indicating seed with hulls and smaller seeds (Fig. 1). To get a better understanding of the altered phenotype, measurements of internodes and grains at the developed stage were carried out on the srs1-1, srs1-2 and srs1-3 mutants as well as those of WT from which the mutants were generated (Table 1 and Table 2). The overall length of internodes was reduced in all the srs1 mutants, although the degree of reduction of each internode was different among the srs1 mutants. The length of grains was also reduced in all the srs1 mutants, although the degree of reduction of each grain was different among the various srs1 mutants. The grains were slightly wider and thicker in srs1-1 and srs1-2 than in WT, but no significant difference was found between srs1-3 and WT in this regard. Grain weight was also reduced in general in the srs1 mutants, which correlates with their reduced length. The weight of seeds (brown rice) in WT, srs1-1, srs1-2 and srs1-3 were 25.7 ± 1.1 (100%), 21.0 ± 0.3 (81.8%), 22.1 ± 1.2 (86.1%), 14.7 ± 1.1 (57.2%), respectively. This result showed that the reduction of seed weight was proportional to that of grain weight in the srs1 mutants. Grain number per panicle was not statistically different between the srs1 mutants and WT, but the number of panicles per plant was reduced in srs1-2 and srs1-3. Among all mutants, the srs1-3 mutant showed the most severe phenotype, including the leaf erection at mature stage. As shown by these results, it becomes apparent that the SRS1 gene is involved in the regulation of both longitudinal and lateral elongation of the grains. Some morphological differences among the srs1 mutants may be linked to various degree of alteration of the SRS1 function by the different mutations in the gene.
 View Details | Fig. 1 Plant type and seed morphology of WT, srs1-1, srs1-2, srs1-3, srs1-4 and srs1-5. (A) Whole plant (upper panel), grain (middle panel) and seed (lower panel) morphology of T65 and the srs1 mutants, srs1-1, srs1-2 and srs1-3. T65, which is the recurrent parent of the srs1-1, srs1-2 and srs1-3 mutants, was used as the wild-type plant (WT). (B) Whole plant (upper panel), grain (middle panel) and seed (lower panel) morphology of Fujiminori and the srs1-4 mutant. Fujiminori, which is the recurrent parent of the srs1-4 mutant, was used as the wild-type plant (WT) (C) Whole plant (upper panel), grain (middle panel) and seed (lower panel) morphology of the srs1-5 mutant. The recurrent parent of the srs1-5 mutant was unknown. All plants were grown in the field to the heading stages. Bars = 20 cm (upper panel) and 0.5 cm (middle and lower panels). |
 View Details | Table 1 Morphological measurements of internodes in WT and srs1 mutants |
 View Details | Table 2 Morphological measurements of grains in WT and srs1 mutants |
Shortening of the second internode has been observed in the Gα deficient mutant, d1 and the BR-related mutants, d2, d11 and d61. In order to investigate whether SRS1 gene is involved in the G protein signaling and/or the BR signaling pathways, the length of the second internode of WT, srs1-1, srs1-2, srs1-3, d1-1, d2-2 and d61-2 were compared. Fig. 2 shows that the length of the second internode was similar between the srs1 mutants and WT, but severely reduced in d1-1, d2-2 and d61-2 mutants. This suggests that the SRS1 gene may not be involved in a regulatory pathway that is dependent of the heterotrimeric G protein signaling and BR-signaling pathway in regard to the elongation inhibition of the second internode.
 View Details | Fig. 2 Length of internodes relative to the total length of the culm. Schematic representation of the elongation patterns of internodes in WT (T65), the recurrent parent of all mutants: the srs1 mutants (srs1-1, srs1-2, srs1-3), the Gα-deficient mutant (d1-1), the BR-deficient mutants (d2-2) and the BR-insensitive mutant (d61-2). The relative length of internodes of srs1 mutants was similar to that of WT. The second internodes in d1-1, d2-2 and d61-2 were remarkably shortened, compared with that in WT. |
The photomorphogenic phenotype of the srs1 mutants was compared to that of the d1, d2-2 and d61-2 mutants (Fig. 3). When WT plants and the srs1 mutants were grown in the dark, the internodes were elongated in both WT and the srs1 mutants, whereas no elongation could be observed in this region in the d1-1, d2-2 and d61-2. Apparently, the skotomorphogenesis occurs normally in the srs1 mutants and WT, but not in the d1-1, d2-2 and d61-2 mutants.
 View Details | Fig. 3 Skotomorphogenesis of the srs1 mutants. Gross morphology of WT (T65), the srs1 mutants (srs1-1, srs1-2, srs1-3), the Gα-deficient mutant (d1-1) and the BR-deficient mutants (d2-2) and the BR-insensitive mutant (d61-2) grown in darkness for 2 weeks on agar media. Arrows indicate the position of nodes. The internodes (the part between two arrows) of WT and the srs1 mutants elongated, indicating that WT and the srs1 mutants had a skotomorphogenic growth phenotype. The internode of the Gα-deficient mutant (d1-1) and the BR-related mutants (d2-2 and d61-2) did not elongate under the same condition. Thus, Gα-deficient mutant and the BR-related mutants shows a de-etiolated phenotype. Bar = 3 cm. |
In order to identify the cause of the small and round seed phenotype in the srs1-1 mutant, the length and width of the cells in the inner epidermal tissues of the lemma in WT and srs1-1 at the heading stage were analyzed by scanning electron microscopy (SEM). The distribution of cell length in the mutant and in the WT is shown in Fig. 4. Measurements of the lemna, the cell length and the estimated cell number of the lemma in srs1-1 are summarized in Table 3. The cell number in the longitudinal and the lateral direction of the lemma was estimated by dividing the length and the width of the organs by the length and the width of the cell, respectively. In the longitudinal direction, the organ length, the averaged cell length and the deduced cell number in srs1-1 was 85.5%, 91.4%, and 93.5% of those in WT respectively. These results suggest that the reduced organ length was due to the reduction of both the cell length and the cell number, thus, the SRS1 gene appears to regulate both the cell number and the cell length in the longitudinal direction. In the horizontal direction, the organ length, the averaged cell length and the deduced cell number in srs1-1 was 111.5%, 114.2%, and 97.5% of those in WT respectively, indicating that the increased width in srs1-1 was due to the increased cell length. These results suggest that loss of SRS1 function enhanced cell elongation in the lateral direction of the lemma.
 View Details | Fig. 4 Analysis of cell length in WT and srs1-1. (A and B) Inner epidermal cells of lemma from WT (T65) and srs1-1 observed by SEM. Bar = 1 mm. (C) Cell length of WT (white bar) and srs1-1 (black bar). Average cell lengths of WT (open triangle) and srs1-1 (closed triangle). The significant differences for cell length were detected between WT and srs1-1 (0.01 > p). (D) Cell width of WT (white bar) and srs1-1 (black bar). Average cell width of WT (open triangle) and srs1-1 (closed triangle). The significant differences for cell length were detected between WT and srs1-1 (0.001 > p). |
 View Details | Table 3 Cell length and cell number in lemma of WT and the srs1-1 mutant at the heading stage |
Linkage mapping using the srs1-1 mutant placed the SRS1 locus on chromosome 7 (Tanabe et al., 2007). Identification of the SRS1 gene was performed by map-based cloning using F2 plants from a cross between srs1-1, a japonica strain and Kasalath, an indica strain. The F2 plants segregated into two groups in a 3:1 ratio, that former group showing the normal seed phenotype and the latter showing the small and round seeds phenotype. The 1500 F2 plants bearing small and round seeds were used for a high-resolution mapping of the SRS1 locus. The candidate genomic region of srs1-1 was delimited to a 55 kb interval between markers 25285As and 25340As with two recombinant plants (Fig. 5A).
 View Details | Fig. 5 Physical map of the srs1 locus and mutation sites in the srs1 mutants. (A) High-resolution linkage and physical map of the srs1 locus. Vertical lines represent the positions of molecular markers (25249, 25285As, 25340As, 25363) and the numbers of recombinants are indicated below the linkage map. The nine candidate genes for the srs1 mutation were present between marker 25285As and 25340As. (B) Structure of SRS1 and positions of the mutations in the srs1 mutants, srs1-1, srs1-2, srs1-3, srs1-4 and srs1-5. Black boxes indicate exons of the SRS1 gene. The SRS1 gene consists of 10 exons and 9 introns and codes for 1365 amino acids. (C) Mutation sites in srs1-1, srs1-2, srs1-3, srs1-4 and srs1-5. Mutated nucleotides are underlined. The 38-base deletion in exon 7 in srs1-1, one base substitution (G to T) in a conserved intron splicing site in srs1-2, one base substitution (A to G) resulting change from the stop codon to Trp in srs1-3, the 31-base deletion in exon 6 in srs1-4 and one base substitution (G to A) in a conserved intron splicing site in srs1-5. Amino acid residues generated by abnormal splicing, resulting the frame shift in srs1-5 were also shown in the bottom. |
In the candidate region, nine genes are annotated in the Rice Annotation Project Database (RAP-DB, http://rapdb.dna.affrc.go.jp/). To identify the mutation in srs1-1, we have used over 96 sets of PCR primers to cover these open reading frames and sequenced the fragments amplified by PCR. A sequence polymorphism was found in Os07g0616000 suggesting that this gene is the causal gene for the small seed phenotype of the srs1-1 mutant. We named the gene, SRS1. The SRS1 cDNA contains an open reading frame of 1366 codons. The SRS1 gene consists of 10 exons and 9 introns. The mutation in the srs1-1 mutant is a 38-bp deletion in the seventh exon (Fig. 5B and C). The SRS1 gene was found to be identical to the DENSE AND ERECT PANICL 2 gene (DEP2) reported previously (Li et al., 2010).
By rough map-based analysis, the causal gene of a small and round seed phenotype from 00T69, TCM1201, Fu-1 and Kaihenaikoku d1 was localized on the same chromosome 7 (data not shown). Because the candidate for the causal gene of srs1-1 is also localized on chromosome 7, we hypothesized that the mutations in these mutants also occurred in the SRS1 gene. Therefore, sequence of the SRS1 gene in these mutants was analyzed (Fig. 5C). As expected, we found that 00T69 has one base substitution (G to T) in the third intron of SRS1, which changed the conserved sequence of the splicing site GTGAG (Irimia and Roy, 2008) of the WT gene to TTGAG. Based on this result, it was hypothesized that such a mutation led to abnormal splicing of SRS1 in 00T69. 00T69 was renamed srs1-2. To investigate the occurrence of abnormal splicing in the srs1-2 mutant, mRNAs were prepared from the mutant and cDNA was synthesized. Sequence analysis of the cDNA from SRS1-2 revealed the 120-bp intron 3 was not spliced out, indicating that splicing is affected by the mutation. Because the sequence of intron 3 does not contain stop codons, theoretically, the mutated SRS1-2 protein could have an additional sequence of 40 amino acids, which could potentially impair the function of the protein. The mutant TCM1201, renamed srs1-3, also has one base substitution (A to G) located in the stop codon, resulting in a change to Trp (TGG). A new stop codon could be found at 183 base downstream from the mutation site. The mutated SRS1-3 protein could be 61 amino acids longer than the wild-type SRS1 protein which could potentially impair the function of the protein. The mutant Fu-1, renamed srs1-4, has a deletion of 31-bp in exon 6 which is the same as that in the dep2-1 mutant. Kaihenaikoku d1, renamed srs1-5, had one base substitution (G to A) in the third intron of SRS1, which changed the AG dinucleotide of the conserved splicing site (Irimia and Roy, 2008) into AA. Such a change is expected to lead to abnormal splicing. To investigate the occurrence of abnormal splicing, the mRNAs from srs1-5 was prepared and cDNA was synthesized and analyzed. We confirmed that splicing has occurred between the GT of intron 3 and the AG of intron 4, leading to the elimination of exon 4 from the srs1-5 transcript. As a result, a frame shift occurs at the splicing junction between the exon 3 and exon 5 in srs1-5. A new stop codon is generated 60 bp downstream of the exon 3 of SRS1 in srs1-5.
BLAST searches revealed that the deduced amino acid sequence of the SRS1 protein shared some similarity with proteins from Arabidopsis: 32% identity and 73% similarity with At3g14172, 25% identity and 71% similarity with At1g72410 (NP_177385), and 23% identity and 73% similarity with At1g17360 (NP_173179). Functional domains could not be found in the SRS1 protein and it’s orthologs from Arabidopsis, At3g14172, At1g72410 and At1g17360 using the SMART algorithm (http://smart.embl-heidelberg.de/smart/set_mode.cgi?NORMAL=1).
About 180 amino acid residues of the N-terminus of the SRS1 protein are highly similar to that of COP1-interacting protein 7 (CIP7) (Yamamoto et al., 1998). However, this region is not a functional domain in CIP7. Therefore, the function of SRS1 remains to be elucidated.
Western blot was performed to determine the intracellular localization of the SRS1 protein. Four fractions, namely pellets of the 1,000 g, 10,000 g and 100,000 g and supernatants, were analyzed by western blotting, using the anti-SRS1 antibodies raised against the C-terminus of the SRS1 protein (Fig. 6). The analysis revealed a cross-reaction with a band of about 160 kDa from the pellet fraction of 100,000 g.
 View Details | Fig. 6 Intracellular localization of SRS1 protein. (A) Predicted size of the SRS 1 proteins from WT and srs1-1. The number of amino acids of SRS1 is indicated in parenthesis. The region of the recombinant protein used for antibody production is indicated by an arrow. (B) Proteins from WT (10 μg) and the srs1-1 mutant (10 μg) were separated by SDS-PAGE on a 5–20% gradient gel and analyzed by western blotting (WB) using anti-SRS1 antibody. Lanes 1, 1,000 g pellet; lane 2, 10,000 g pellet; lane 3, 100,000 g pellet and lane 4, 100,000 g supernatant. The arrow indicates the candidate band for the cross-reaction of the antibodies with the SRS1 protein. |
To confirm that the 160 kDa protein as indeed the SRS1 protein, we probed protein extracts of srs1-1 with the anti-SRS1 antibodies. The predicted open reading frame of the mutated SRS1 proteins in srs1-1 is 819 amino acid-long, which does not contain the C-terminus region of the WT protein used for generating antibodies. As expected, no cross-reaction was observed with protein extracts from srs1-1 (Fig. 6). From these results, we concluded that the 160 kDa polypeptide localized in the crude microsomal fractions, that cross-reacted with the antibodies was the SRS1 protein.
Accumulation of SRS1 transcripts and proteins during rice plant development was investigated by real-time RT-PCR and western blot analyses. First, high levels of SRS1 proteins were found in internodes and in panicles both before and after heading stage (Fig. 7A, left panel). Then, we examined the expression profile of SRS1 proteins in leaves at different plant developmental stages by choosing the L4 stage, where the first (L1), second (L2) and third (L3) leaves have reached full development, while the fourth leaf is still developing (Fig. 7A, middle panel). At this stage, we found the expression levels of SRS1 proteins to be significantly higher in the fourth leaf (L4) than in other leaves. This shows that the SRS1 protein is highly expressed in developing organs and the expression diminishes as the organs reach final development. The levels of SRS1 proteins accumulation during the development of panicles were also examined (Fig. 7A, right panel). Results show that the amounts of SRS1 proteins are higher in the developing panicles than in the fully developed ones, which corroborates the results observed with leaves. Fig. 7B shows the relative levels of SRS1 proteins in the above-analyzed organs, taking the expression level in the 2.5 cm-long panicles as a reference. It shows that the overall levels of SRS1 proteins in panicles are higher than in other organs. In parallel with the protein analysis, we also carried out real-time quantitative PCR to examine the levels of the SRS1 transcripts in the corresponding organs. As shown in Fig. 7C, the levels of SRS1 transcripts correlated fairly well with those of the SRS1 protein overall.
 View Details | Fig. 7 Expression of the SRS1 protein. (A) Total proteins prepared from 0.5 mg of tissue from WT were analyzed by SDS-PAGE and western blot (WB) using the anti-SRS1 antibody. Lane 1, aerial parts containing all leaves from 12 days-old-plants; lane 2, roots from 12 days-old-plants; lane 3, internodes; lane 4, panicles (> 15 cm) before heading (BH); lane 5, panicles after heading (AH); lane 6, first leaves (L1); lane 7, second leaves (L2); lane 8, third leaves (L3); lane 9, fourth leaves (L4); lane 10, panicles with 2.5 cm length; lane 11, panicles with 5 cm length; lane 12, panicles with 7 cm length; lane 13, panicles with 10 cm length. SRS1 protein is shown by the arrow. (B) Amount of SRS1 protein in various organs is expressed relative to the amount of SRS1 protein in 2.5 cm-long panicles (Fig. 7A, lane 10). Quantification of signal on WB was done using a Multi Gauge ver 2.2 imaging device. (C) Real-time RT-PCR analysis of SRS1 mRNA. Total RNAs were isolated from various organs (same as in panel A). The amount of SRS1 mRNA was expressed relative to that of OsUbiquitin1 (an ubiquitin gene). |
We report the identification of the causal gene of a new small and round seed mutant 1 (srs1) and its morphological analysis. The SMALL AND ROUND SEED 1 (SRS1) gene was found to be identical to the DENSE AND ERECT PANICL 2 (DEP2) gene, which was shown to be involved in the panicle architecture (Li et al., 2010).
Analysis of the plant type, grain and seed morphology revealed that the srs1 mutations exert pleiotropic effects on the plant development, including small and round seeds, shortened internodes, and erect leaves, in addition to the dense panicle phenotype reported earlier (Li et al., 2010). Among these characteristics, reduction in the longitudinal direction of grains and seeds, internodes and panicles was commonly observed among the srs1, d1 and BR-related mutants. However, the srs1 mutants exhibited certain specific phenotypes. Unlike d1 and BR-related mutants, srs1 mutants showed similar elongation pattern of the second internode and normal skotomorphogenesis as WT. Therefore, SRS1 proteins appear to be involved in a regulatory pathway that is independent of the heterotrimeric G protein signaling and BR-signaling pathways in regard to the inhibition of the elongation of the second internode and skotomorphogenesis.
In the srs1-1 mutant, the grain length was reduced and the grain width was increased (Table 2). In the dep2 mutants (alleles of srs1), the panicle length was reduced and panicle diameter was increased (Li et al., 2010). These results suggest that the SRS1/DEP2 gene reduced the development in the longitudinal direction but enhanced the development in the lateral direction of plural organs in rice.
By analyzing the length and width of cells in the inner epidermal tissues of the lemma of WT and the srs1-1 mutant at heading stage, we found that the small and round seed phenotype in the srs1-1 mutant was due to the reduction in cell length and cell number in the longitudinal direction and the increased cell length in the horizontal direction of the lemma (Table 3). In the case of the srs1-1 mutant, the size of lemma and palea was the limiting factor to the size of grain. BRI1, a BR-related gene, is known to regulate cell length (Yamamuro et al., 2000; Oki et al., 2009) and cell number through fine-tuning of the direction and rate of cell division (Nakamura et al., 2006) in the leaf sheathes and the internodes. Thus, the reason of shortened seed phenotypes in BR-related mutants may also be the reduction of the cell number via the impaired cell elongation, but the morphological study of grains in the BR-related mutants remains to be investigated. By analogy, the SRS1/DEP2 gene may have similar function as the BR-related genes in regards to the regulation of the cell elongation and the cell number, though the function of the SRS1/DEP2 gene was different from that of BR-related genes in the elongation pattern of the second internode and skotomorphogenesis. Among small and round seed mutants, the cause of the small seed phenotype in the d1-5 mutant is mainly due to the reduction of cell number in the lemma (Izawa et al., 2010). The cause of the small seed phenotype in the srs3 mutant is mainly due to the reduction of cell length in the lemma (Kitagawa et al., 2010). The SRS3 gene encodes a kinesin 13 protein (Kitagawa et al., 2010). The biochemical function of the SRS3 protein remains to be determined and it remains to be elucidated whether SRS3 functions in the G protein signaling or the BR signaling, or both. As results, these data suggested that the function of SRS1/DEP2 gene may be similar to that of BR-related genes, but not D1 and SRS3, in regards to the regulation of the cell number and cell length of the lemma.
We have shown that the causal gene for a novel small and round seed mutant (srs1), SRS1/DEP2 codes for a novel protein without any functional domains (Fig. 5). We also found four more mutants having mutations in the SRS1/DEP2 gene. The srs1-1 and srs1-4 mutants showed similar abnormal phenotypes, namely the shorten grain, seed and internode and the erect leaf. However, the recurrent parent of the srs1-1 and srs1-4 mutant is Taichung 65 and Fujiminori, respectively. These results show that the loss of SRS1/DEP2 function correlates with these abnormal phenotypes, regardless of the genetic background. In srs1-2 and srs1-3, the presence of an aberrant SRS1 protein is expected. Some morphological differences among srs1-1, srs1-2 and srs1-3 may depend on the level of impairment in the SRS1 proteins. It will be important to study the function of the aberrant SRS1 proteins, in addition to the SRS1 protein.
The size of the SRS1 protein was estimated at 160 kDa by SDS-PAGE analysis and the protein is mainly located in the crude microsomal fraction, indicating that the SRS1 protein is a membrane protein. It will be important to identify the target membrane in which the SRS1 protein is located. Previously, the DEP2-GFP fusion protein was detected in the cytoplasm, plasma membrane and nucleus in onion epidermal cells and N. benthamiana leaves by transient assays through particle bombardment (Li et al., 2010).
The SRS1 mRNAs and proteins were shown to accumulate in the developing organs, namely young leaves and panicles, suggesting that the SRS1 protein may regulate cell elongation and/or the cell division. The amount of the SRS1 protein per total proteins was highest in the young panicles. The result showed that tissues with high amount of the SRS1 proteins closely corresponded to the tissues that show abnormality in the srs1 mutant.
The mechanisms regulating grain size are likely to be complex. First, there are many different genes regulating development of the lemma/palea, namely the D1 gene regulating cell number, the SRS3 gene regulating cell length and the SRS1/DEP2 and the BR-related genes regulating cell number and length. Second, it is proposed that the lemma/palea and the caryopsis (seeds after maturation of ovary) may be regulated by different pathways during grain formation (Takeda and Takahashi, 1970; Takeda, 1982). This suggests that there may be genes regulating seed size independently from genes regulating the size of lemma/palea. In other words, the seed size may not be determined only by lemma/palea size. In order to elucidate these mechanisms, it would be useful to identify new grain and/or seed size mutants, to identify the causal genes, and to carry out genetic cross studies using these mutants.
This work was supported by the Ministry of Education, Culture, Sports, Science and Technology, Japan [Grant-in aid for Scientific Research on Priority Areas (No. 19060003)] and the Ministry of Agriculture, Forestry and Fisheries of Japan [Genomics for Agricultural Innovation IPG-0002].
We thank Drs. Yasuo Nagato and Hikaru Sato for the gift of rice mutants and Ms. Hiromi Matsuoka for her help with plasmid construction.
|
Index