Introduction

The male gametophyte, pollen, develops within the anther compartment of the stamen and requires cooperative functional interactions between gametophytic and sporophytic tissues. Pollen development consists of several distinct stages. A microspore mother cell undergoes meiosis to give rise to a tetrad of four microspores. After release upon the dissolution of the callose wall, each uninucleate microspore undergoes an asymmetric mitotic division to give rise to two cells with distinct fates: the vegetative cell and the generative cell. The larger vegetative cell is transcriptionally active (Bedinger 1992; McCormick 2004). Recently, a number of anther-specific cDNA clones have been identified and isolated using DNA microarray and macroarray techniques from several model plants, Lotus japonicus, Arabidopsis thaliana, and Oryza sativa (Endo et al. 2002; Amagai et al. 2003; Honys and Twell 2003; Endo et al. 2004). Although pollen-specific genes have been studied extensively in most cases, it remains unclear what roles these genes play.

The genes encoding small cysteine-rich proteins form a gene family in several plant species. The small cysteine-rich proteins are highly differentiated in their biological function and were related to several physiological phenomena, e.g. storage, protection, lipid transfer, cell–cell communication, etc. (Jose-Estanyol et al. 2004). In the case of cell–cell communication, SP11/SCR, encoding a male S determinant in Brassica self-incompatibility (SI), is a small cysteine-rich protein, and functioned as a ligand molecule for female S determinant, SRK, S receptor kinase (reviewed in Watanabe et al. 2003; Takayama and Isogai 2005). In another case, lipid transfer protein, LTP, also encoding a different type of small cysteine-rich protein, has an important role in transferring lipids in several tissues (Kader 1996). The pollen-specific gene, LAT52, also encodes a small cysteine-rich protein, which could interact with pollen-specific receptor protein kinase, LePRK2 (Twell et al. 1989; Tang et al. 2002). LAT52 is required for the pollen to germinate in vitro and achieve fertilization in vivo (Muschietti et al. 1994).

In the previous experiments, we identified over 100 anther-specific genes in rice (Endo et al. 2004). In this current study, we identified three novel pollen-specific clones, OsSCP1, OsSCP2, and OsSCP3. The deduced amino acid sequences of OsSCPs contained an open reading frame (ORF) encoding a small cysteine-rich protein. In order to determine whether the promoter regions of OsSCPs were functional in pollen or not, transgenic tobacco plants having the β-glucuronidase (GUS) gene regulated by OsSCPs promoter were made. A blue GUS signal was specifically observed in mature pollen grains and pollen tubes. Furthermore, OsSCP1 and OsSCP3 formed a multigene family in the rice genome. We discuss the function of the pollen-specific OsSCP genes.

Materials and methods

Plant material and mRNA isolation

Plants of O. sativa cv. Koshihikari were grown in a green house. Anther and pistil tissues in various stages (A1: anther of uninucleate stage; A2: anther of binucleate stage; A3: anther of trinucleate stage; P1: pistil of uninucleate stage; P2: pistil of binucleate stage; and P3: pistil of trinucleate stage) were collected for isolation mRNA as described previously (Endo et al. 2004). Leaf sheath (Ls) and leaf blade (Lb) were harvested at the heading stage. Furthermore, root (R) and shoot (S) of seedling were also collected for RNA experiment. Lemma and palea (L/P) were also harvested for mRNA isolation. Isolation of poly (A)+ RNA from each tissue was performed using a FastTrack 2.0 mRNA isolation kit (Invitrogen, San Diego, CA) as described in Watanabe et al. (2000).

RT-PCR

RT-PCR was performed according to Endo et al. (2002). Briefly, poly (A)+ RNA was reverse-transcribed to synthesize the first strand cDNA by using the First-Strand cDNA synthesis kit (Amersham-Pharmacia, Uppsala, Sweden). Then, cDNA was used as a template for PCR amplification with a set consisting of primers specific to each gene: OsSCP1RT-F(5′-ATGGCCCAGAACAAGACCAT-3′),OsSCP1RT-R(5′-GCAACTGTCTACGCACTTCTTGTTGGTG-3′),OSCP2 RT-F(5′-ATGGCCCAAAACAAGACCAT-3′),OsSCP2 RT-R(5′-TCCTTGGCTTCCTTAGACATG-3′),OsSCP3R T-F(5′-CAGAACAAGACCATTGCGGT-3′), and OsSCP3 RT-R(5′-CGTTGTTTGAATATAAGCGACGAC-3′). PCR was performed with ExTaq DNA polymerase (TaKaRa Shuzo, Shiga, Japan) for 25 or 35 cycles of denaturation for 30 s at 94°C, annealing for 30 s at 58°C and extension for 1 min at 72°C, followed by a final extension for 5 min. Actin primers as a control were 5′-TCCATCTTGGCATCTCTCAG-3′ (forward) and 5′-GTACCCGCATCAGGCATCTG-3′ (reverse).

5′-RACE and sequence analysis

Full-length cDNA clones of OsSCPs from the cultivar, Koshihikari, were obtained by the 5′-RACE procedure according to Watanabe et al. (1999). cDNA inserts were sequenced by the dideoxy chain-terminator method using a model 310 DNA sequencer (Applied Biosystems, Foster City, CA). The DNA sequence was analyzed with Genetyx software (Software Development, Tokyo, Japan). A homology search was performed using BLAST (Altschul et al. 1997). The motif search was performed with the TAIR web site program (http://www.arabidopsis.org/cgi-bin/patmatch/nph-patmatch.pl). The presence of a signal peptide was predicted using the PSORT program (http://psort.nibb.ac.jp).

Construction of promoter–GUS fusion

The promoter regions of OsSCP1, OsSCP2, and OsSCP3 were identified from a DNA database of genomic sequences of rice. These promoter regions were amplified by PCR using each of the following primers: OsSCP1pro-F (5′-ACTAGTTCTGTTGTTCTACCCAATATTGTTAC-3′), OsSCP1pro-R (5′-GGATCCTAGCACCGTTTCTATTGA GGGTAACTGGTGTGGAT-3′), OsSCP2pro-F (5′-ACTA GTACTGGGTCGACTTGGGCCGGGGAGGAGAGAGA- AA-3′), OsSCP2pro-R (5′-GGATCCTAGCACCGTTTAT GCTGAGCAGAATTGGT-3′), OsSCP3pro-F (5′-ACT AGTCTGCCACTGTCTACAGCGGGAAGGTTAACAAT-3′), and OsSCP3pro-R (5′-GGATCCTAGCGCTGTTTT TCTTGAGTGTAATTGGTGTGGA-3′). All PCR fragments were subcloned into the pCR2.1 plasmid vector (Invitrogen). The nucleotide sequence of these promoter fragments was confirmed by sequencing of the plasmid using the DNA sequencer (ABI 310; PE Biosystems, Foster City, CA). The binary vector, pBI–GSH, was used. This vector was derived from pBI–BG vector (Okada et al. 2000) by replacement of the Bra r 1 promoter region to multi-cloning sites. The promoter region of each OsSCP was taken as a SpeI–BamHI fragment from each plasmid and inserted into the SpeI–BamHI site of pBI–GSH. As a result, each promoter region was placed in front of the reporter gene, GUS. The constructs, each formed by promoter–GUS fusion, were named OsSCP1–GUS, OsSCP2–GUS, and OsSCP3–GUS, respectively.

Transformation of tobacco plants

The constructed OsSCP1–GUS, OsSCP2–GUS, and OsSCP3–GUS were transferred to Agrobacterium tumefaciens strain EHA101 using the freeze-thaw method (An et al. 1988). Leaf disks of Nicotiana tabacum cv. Petit Havana SR1 were transformed according to the procedure of Horsch et al. (1988). Transgenic tobacco plants were selected on the basis of hygromycin resistance (100 μg/ml). These plants were maintained under greenhouse conditions until maturity. The vegetative and floral tissues were used in GUS assay.

Histochemical GUS assay

Vegetative tissues, and anther and pollen of different developmental stages, were obtained from primary transgenic plants and wild-type plants as a control. GUS staining was performed according to Tsuchiya et al. (1994), except for addition of 20% methanol in the staining solution. Briefly, for staining, several tissues were incubated in 200 mM NaPO4, pH 7.0, 20% methanol, 0.3% Triton X-100, 12.5 mM K3Fe(CN)6, 12.5 mM K4Fe(CN)6, and 38.3 mM X-Gluc (5-bromo-4-chloro-3-indoyl-β-d-glucuronide). Plant tissues were vacuum infiltrated briefly, then incubated at 37°C overnight. After staining, chlorophyll was cleared from the sample by 75% ethanol treatment. Photos were taken using a Nikon E800 microscope system. The developmental stages of microspores and pollen grains were determined by staining nuclei with DAPI (Watanabe et al. 1991). Pollen germination was performed according to Lush et al. (1998).

Results

Identification and characterization of OsSCP genes

In our previous study, we isolated 89 mature anther-specific genes (Cluster RA3 genes) in rice by using cDNA microarray analysis (Endo et al. 2004). However, the function of 60% of the Cluster RA3 genes (53 genes) was unknown at the time we published the array data. Because a large number of ESTs (expressed sequence tags) and full-length cDNA clones were deposited in the public database after publishing our data (Yazaki et al. 2004), we reanalyzed these unknown function genes and their encoding proteins from the viewpoint of the specific amino acid motifs. Within 53 genes, we found that three genes, Os-26, Os-32, and Os-169 (renamed as OsSCP1, OsSCP2, and OsSCP3), had a small ORF, which was constructed from 127, 129, 129 amino acid residues, respectively, and encoded specific cysteine-rich motif (Cys–X3–Cys–X13–Cys–X3–Cys; Fig. 1). Based on the homology search of OsSCPs against the rice full-length cDNA database with the BLAST program, OsSCP1, OsSCP2, and OsSCP3 were found to be identical to full-length cDNA clones, AK064756 (chromosome 2), AK071781 (chromosome 8), and AK120999 (chromosome 2), respectively. Among three OsSCPs, the sequence identity was 35–47% at the amino acid level, and eight cysteine residues were completely conserved. Interestingly, the nucleotide difference between our cDNA clone and the full-length cDNA clone in each gene was not observed in both ORF and UTR (5′ and 3′) regions in spite of the difference of material cultivar, Koshihikari (our cDNA clone) and Nipponbare (full-length cDNA clone). Based on the comparison between cDNA and the genomic clone, the OsSCPs had no intron. By using the full-length cDNA information of OsSCPs, a BLASTP search of the OsSCPs was performed against the public database. However, OsSCPs showed no significant similarity to other known proteins. As described above, a number of small cysteine-rich proteins have been identified and characterized (Jose-Estanyol et al. 2004). When the motif search against OsSCPs was performed to find proteins, which have the same cysteine-rich motif as the OsSCPs on the web site, eight clones encoding OsSCP-like proteins having a specific cysteine motif were found only in the rice genome, not in other plant genomes (data not shown). Furthermore, we performed a signal peptide search with PSORT program. All three OsSCP proteins had N-terminal signal peptide, like other pollen-specific cysteine-rich proteins, SLR1-BP, PEC-3, and PCP-A1 (Stanchev et al. 1996; Toriyama et al. 1998; Takayama et al. 2000a). When compared the nucleotide sequences of these eight cDNA clones each other, eight cDNA clones were divided into three groups. In the first groups, the nucleotide sequence of three cDNA clones was identical to that of OsSCP1. In the second group, the nucleotide sequence of two cDNA clones was identical to that of OsSCP2. In the third group, the nucleotide sequence of three cDNA clones was identical to that of OsSCP3. Furthermore, the OsSCPs formed a small gene family in the rice genome. From these results, OsSCPs were found to be a novel type of cysteine-rich protein (Fig. 2). Thus, we carefully analyzed the OsSCPs.

Fig. 1
figure 1

Multiple alignment of amino acid sequences of OsSCP1, OsSCP2, and OsSCP3. Gaps were introduced to optimize the alignment. Conserved amino acid residues are indicated by boxes. The conserved eight cysteine residues in the mature proteins are shown as C1–C8

Full size image
Fig. 2
figure 2

Multiple alignment of the cysteine motifs of OsSCPs with the cysteine motifs of other proteins. Os, Oryza sativa; Br, Brassica rapa; Bo, Brassica oleracea; Ms, Medicago sativa; Mt, Medicago truncatula; Zm, Zea mays; Le, Lycopersicon esculentum

Full size image

In order to confirm the expression pattern of OsSCPs, RT-PCR was performed with several rice tissues. In each gene, specific amplification was observed only in the anther at the trinucleate stage containing mature pollen grain. In other anther at the developmental stages and other tissues, no band was detected (Fig. 3). This expression pattern was coincident to that observed by cDNA microarray analysis (Endo et al. 2004).

Fig. 3
figure 3

RT-PCR analysis of anther-specific cDNA clones, OsSCP1, OsSCP2, and OsSCP3. The mRNA was isolated from root of seedling (R), shoot of seedling (S), leaf sheath at the heading stage (Ls), leaf blade at the heading stage (Lb), lemma and palea (L/P), pistil of uninucleate stage (P1), pistil of binucleate stage (P2), pistil of trinucleate stage (P3), anther of uninucleate stage (A1), anther of binucleate stage (A2), and anther of trinucleate stage (A3) in rice. The gene encoding actin was used as a positive control

Full size image

In order to determine the promoter region in each OsSCP, we compared the nucleotide sequence between full-length cDNA clone and genomic DNA clone of each OsSCP gene. We defined the promoter region as the nucleotide sequence between OsSCP and the nearest upstream ORF (intergenic region). The length of the putative promoter region of each OsSCP was 2033 bp, 952 bp, and 1845 bp, respectively (Fig. 4). The putative TATA box sequence was found at −74 for all OsSCP promoter sequences when the start of initiation codon (ATG) was marked as +1. Furthermore, in order to determine the spatial and temporal expression pattern of the OsSCPs, the 2033 bp promoter region (nucleotides −2034 to −1) of OsSCP1 (OsSCP1-b, see below), the 952 bp promoter region (nucleotides −953 to −1) of OsSCP2, and the 1845 bp promoter region (nucleotides −1846 to −1) of OsSCP3 (OsSCP3-a, see below) were fused with β-glucuronidase (GUS) genes (Fig. 4). For histochemical GUS staining, two to three independent tobacco transgenic plants were used in each OsSCP gene. All the transgenic lines of OsSCP1 (OsSCP1-b)–GUS, OsSCP2–GUS, and OsSCP3 (OsSCP3-a)–GUS exhibited GUS activity in pollens at a very late developmental stage, as described below (Fig. 5). No GUS activity was detected in other floral organs and vegetative organs (data not shown).

Fig. 4
figure 4

Physical map of the promoter regions of OsSCP1, OsSCP2, and OsSCP3. Nucleotides are numbered with the first nucleotide of the translation initiation codon marked as +1

Full size image
Fig. 5
figure 5

Histochemical analysis of GUS activity conferred by OsSCP1–GUS and wild-type in transgenic tobacco. Developmental stage of microspore and pollen was determined by DAPI staining (A, B, C). Representative photographs of uninucleate microspore (D), early binucleate pollen (E), late binucleate pollen (F), and germinated pollen and pollen tube (G) in transgenic plant are shown. Blue staining indicates GUS activity. In wild-type plant, GUS activity was not observed in late binucleate pollen (H) and germinated pollen and pollen tube (I). WT, wild type; VN, vegetative nucleus; GC, generative cell; M, microspore; P, pollen; GPP, GUS-positive pollen; GNP, GUS-negative pollen; GPPt, GUS-positive pollen tube; Pt, pollen tube

Full size image

For the next step, we examined the GUS activity in anthers of transgenic plants according to different developmental stages. Developmental stage of microspore and pollen was determined by staining of nuclei with DAPI (Fig. 5A–C). Because transgenic plants having each OsSCP construct showed the similar GUS expression pattern, we presented the OsSCP1 construct as a representative. GUS activity was specifically observed in late binucleate pollen grain (Fig. 5F). In other words, GUS activity was not detected in uninucleate microspores or in early binucleate pollen grains (Fig. 5D and E). After growing the pollen tube in the pollen germination medium in vitro, GUS expression was also detected from the pollen tube of all transgenic plants (Fig. 5G). Segregation of GUS-positive and GUS-negative pollen grains was observed, as shown in Fig. 5C. The ratio of GUS-positive and GUS-negative pollen grains was 504:464 in transgenic plant having OsSCP1–GUS construct, indicating a segregation pattern of 1:1 (significant at the 5% level). In the non-transgenic wild-type plants, no GUS expression was detected in the pollen grain and the pollen tube (Fig. 5H and I).

Identification of pollen-specific cis-elements of the OsSCP genes

We surveyed the potential pollen-specific cis-elements, which have been already identified in other plant species, in each promoter region. The sequence motifs of the LAT52/LAT56 (TGTGG and TGTGA) enhancer element, LAT52 quantitative element (TGGTTA), LAT52 pollen-specific activation element (AGAAA), g10 late pollen gene element (GTGA), which were required for the specific expression of mature pollen (Twell et al. 1990; Eyal et al. 1995; Bate and Twell 1998; Rogers et al. 2001), were observed. In the case of LAT52/LAT56, the nucleotide sequences of the enhancer elements (TGTGG and TGTGA) were found at −1850, −1285, and −1130 for the OsSCP1 (OsSCP1-b) promoter, at −345 for the OsSCP2 promoter, and at −617 and −345 for the OsSCP3 (OsSCP3-a) promoter (Fig. 4). The sequence motifs of the LAT52 quantitative element were present at −1059 and −201 for the OsSCP1 (OsSCP1-b) promoter, at −208 for the OsSCP2 promoter, and at −201 for the OsSCP3 (OsSCP3-a) promoter (Fig. 4). In the case of LAT52, the nucleotide sequences of the pollen-specific activation element (AGAAA) were found at −1322, −694, and −326 for the OsSCP1 (OsSCP1-b) promoter, at −923, −790, −620, −548, −524, and −344 for the OsSCP2 promoter, and at −577 and −326 for the OsSCP3 (OsSCP3-a) promoter (Fig. 4). The sequence motifs of the g10 late pollen gene element (GTGA) were present at –1657, −842, and −288 for the OsSCP1 (OsSCP1-b) promoter, at −294 for the OsSCP2 promoter, and at −756 for the OsSCP3 (OsSCP3-a) promoter (Fig. 4). Furthermore, the repeated sequence (TAAGGAA) was present at position −195 and −53 for the OsSCP1 (OsSCP1-b), and −202 and −53 for the OsSCP2, though the function of this repeated sequence was unknown (Fig. 4).

Genomic organization of the OsSCP genes

As a next step, we surveyed the genomic organization of each OsSCP gene. In the case of OsSCP2 gene, we could not find the gene duplication in the rice genome, indicating that this OsSCP2 gene was single copy in rice genome. As previously described, we found two cDNA clones of OsSCP2 in the public database. Combining these two data, these two cDNA clones were derived from single OsSCP2 gene.

In contrast, in the case of OsSCP1 and OsSCP3 genes, we identified gene duplication in the rice genome. In OsSCP1, seven OsSCP1 homologous genomic clones (OsSCP1-a to OsSCP1-g) were tandemly duplicated within the 100-kb genomic region in chromosome 2 of the rice genome. In our experiment, we used popular type of the promoter of the genomic clone, OsSCP1-b, for promoter analysis, as previously described. From the Huge plot analysis, three long duplications were also detected (Fig. 6). Interestingly, the direction of the transcription of seven duplicated genomic clones was the same. Furthermore, when we compared the promoter regions (−2033 to +1) of seven homologous sequences, they were divided into four groups (Fig. 7A). The first group was OsSCP1-a. The second group contained three genomic clones, OsSCP-b, -e, -f. In these three genomic clones, the nucleotide sequence difference up to −2033 bp was not observed. The third group contained two genomic clones, OsSCP-c, -d. In these two genomic clones, the nucleotide sequence of promoter region up to −2033 bp was completely identical. The fourth group was OsSCP1-g. In the case of the OsSCP1-a promoter sequence, the region of the high sequence similarity to other groups (99%) was restricted up to −483 bp (Fig. 7A). Furthermore, several cis-elements, which were required for pollen-specific expression as described above, were also found in each genomic clones, though the location of the elements was different among the four groups (data not shown).

Fig. 6
figure 6

Duplication in the OsSCP1 region. Sequence similarities within 100 kb containing seven OsSCP1 homologous sequences are shown by Huge plot, which was constructed with Genetyx software. Within the plot, each dot represents a match of successive six nucleotides

Full size image
Fig. 7
figure 7

Schematic structure of promoter regions in OsSCP1 homologs (A) and OsSCP3 homologs (B)

Full size image

In the case of OsSCP3, one duplicated homologous clone (OsSCP3-b) was located at 3.7-kb downstream of OsSCP3 (OsSCP3-a), tandemly. Although a low sequence similarity in the promoter region (−1845 to +1) was observed between OsSCP3-a and OsSCP3-b, cis-elements were found in both OsSCP3-a and OsSCP3-b (Fig. 7B).

Discussion

In this study, we identified and characterized novel pollen-specific genes encoding small cysteine-rich proteins (OsSCPs). Small cysteine-rich proteins having eight conserved cysteine residues are widely distributed inside and outside the plant kingdom (Jose-Estanyol et al. 2004). These types of proteins contain the male S determinant of Brassica self-incompatibility (SP11; Takayama et al. 2000b; Watanabe et al. 2000), pollen coat protein, which can interact with SLR1 stigma-specific protein (PCP-A1, Stanchev et al. 1996; SLR1-BP, Takayama et al. 2000a), defensin peptides, which have anti-bacterial activity (Broekaert et al. 1995; Kanzaki et al. 2002; Park et al. 2002), nodule-specific cysteine-rich protein, whose functions are related to the nodule development and defense system (NCR; Mergaert et al. 2003), lipid transfer protein, which functions to transfer several different phospholipids (Zachowski et al. 1998), and pollen-specific ligand molecule, LAT52, which can interact with pollen-specific receptor kinase, LePRK2 (Twell et al. 1989; Tang et al. 2002). When aligned with these proteins at the amino acid level, the sequence and cysteine motif are highly diverted among these cysteine-rich proteins (Fig. 2), indicating that conserved cysteine motif would be important for their specific functions in each protein.

To date, several genes encoding the pollen-expressed cysteine-rich proteins (SP11, SLR1-BP, PCP-A1, PEC-3, and LAT52) have been characterized (Takayama et al. 2000a, b; Watanabe et al. 2000; Stanchev et al. 1996; Toriyama et al. 1998; Twell et al. 1989). All of these proteins had N-terminal signal peptide, indicating that these proteins should be extracellular proteins in pollen surface. As discussed above, a part of functions has been determined, and was important for pollen development and/or pollen–stigma interaction. From these data, OsSCPs would be important for reproductive process, like other pollen-expressed cysteine-rich proteins.

Gene duplication was observed in several genes in plants (Blanc et al. 2000). In the case of the plant disease resistance gene (R gene), genetically linked multigene families were clustered into a single locus. This R gene cluster should have contributed to the generation of the new R gene, because the nucleotide sequence diversity was observed among the duplicated homologous genes (Song et al. 1997; Dixon et al. 1998). The Arabidopsis genome contained gene families homologous to the pollen coat proteins (LCRs) and male S determinant of Brassica SI (SCRLs). Both gene families encoded different types of small cysteine-rich protein, and a part of the gene family formed a gene cluster (Vanoosthuyse et al. 2001). In the gene cluster, each clone was highly diverted. Although there was no experimental data showing that these clustered genes were inherited together as a unit (haplotype), these genes might have had different functions, respectively. In the Brassica SI, the duplication of the SP11 was also observed. In this duplication, the nucleotide sequence of two SP11 genes was completely identical. However, the deletion of the promoter region was found in one SP11 gene, indicating that this clone should have been a pseudogene (Shiba et al. 2004). In this study, in OsSCP1 and OsSCP3, gene duplication was observed (Fig. 6). Interestingly, the nucleotide sequences of ORF in the duplicated clones were completely conserved except for one clone in OsSCP1. Furthermore, pollen-specific cis-elements were also observed in each clone, indicating that these six duplicated clones of OsSCP1 would have been redundant, though the precise expression pattern was determined in only the representative clone, OsSCP1-b. From the RT-PCR analysis, OsSCPs were specifically expressed in the mature anther containing trinucleate pollen grains (Fig. 3). In addition to these data, the nucleotide sequence of OsSCPs in two different varieties, Koshihikari and Nipponbare was completely conserved, though several nucleotide variations in other genomic regions were observed in these two varieties (Shirasawa et al. 2004). Combining the results of cis-elements and RT-PCR of OsSCPs, OsSCPs, which were identified and characterized in this experiment, would have been mature anther-specific genes. This high redundancy and high sequence conservation of OsSCPs indicate the importance of this gene in pollen maturation, pollen germination, and pollen tube growth.

In our previous study, 111 mature anther-specific genes were identified in model legume, Lotus japonicus (Endo et al. 2002). Out of 111 mature anther-specific genes, one interesting gene encoding receptor-like protein kinase was found. On the basis of our preliminary data, this gene was specifically expressed in mature pollen grain in L. japonicus (H. Masuko, M. Endo, and M. Watanabe, unpublished data). To date, two types of the pollen-specific receptor kinases (PRK1 and LePRK2) have been isolated and characterized (Mu et al. 1994; Muschietti et al. 1998). The receptor domain of LePRK2 could interact with pollen-specific small cysteine-rich protein, LAT52 (Tang et al. 2002). This pollen-specific ligand and receptor should be important for pollen germination and pollen tube elongation (Tang et al. 2002). In the rice genome, 2210 candidate genes encoding receptor-like kinases have been identified (Shiu et al. 2004). However, the functions of the receptor-like kinases have been discovered in only a few genes (Song et al. 1995; Nonomura et al. 2003; Suzaki et al. 2004). If the OsSCPs which were identified in this study were functioned as a ligand, the discovery of the orphan receptor would be important for understanding the pollination and fertilization in future.