Introduction

Prunus mume, which originated in China, is a flowering tree and an important commercial fruiting tree due to its flavourful fruit (Hou et al. 2011). P. mume belongs to the Amygdaloideae subfamily and exhibits gametophytic self-incompatibility (GSI) that is controlled by a single S locus with multiple alleles (de Nettancourt 2001). The specificity of the SI response is determined by the haplotypes of the polymorphic S locus, which contains at least two genes, i.e., one for the pistil determinant and one for the pollen determinant. Pollen is rejected when its S haplotype is the same as either of the S haplotypes in the diploid pistil.

The determinants of S-specificity in the pistil are ribonucleases (S-RNases) (Lee et al. 1994; Murfett et al. 1994). Based on the comparison of the cDNAs of several S-RNases in the Japanese pear (Norioka et al. 1996; Ishimizu et al. 1998; Castillo et al. 2002; Takasaki et al. 2004) and apple (Broothaerts et al. 1995; Li et al. 2011), the structural features of rosaceous S-RNases have been characterised. All S-RNases share five conserved domains (C1, C2, C3, RC4, and C5), and one hypervariable (HV) region is involved in allelic specificity in GSI reactions (Ushijima et al. 1998).

The SLF (S locus F-box) gene that controls specificity in the pollen determinant SLF genes were first reported in the Plantaginaceae and are located approximately 9 kb downstream from the S-RNase (Lai et al. 2002). In the Prunus species, the putative pollen S gene encodes an F-box protein, which is termed the S haplotype-specific F-box protein gene (SFB) in sweet and sour cherries (Yamane et al. 2003a, b; Ikeda et al. 2004, 2005; Wunsch et al. 2010), almonds (P. dulcis) (Ushijima et al. 2003), Japanese plums (Zhang et al. 2007), peaches (Tao et al. 2007), and Japanese apricots (P. mume) (Entani et al. 2003; Yamane et al. 2003b). The primary structural features of Prunus SFBs, based on a comparison of 13 alleles, include the presence of one F-box motif, two variable regions (V1 and V2), and two hypervariable regions (HVa and HVb) in the amino acid sequence (Ikeda et al. 2004). The physical distance between SFB and S-RNase varied from 380 bp to 30 kb depending on the S haplotypes (Entani et al. 2003; Ushijima et al. 2003; Yamane et al. 2003a).

P. mume exhibits S-RNase-based gametophytic self-incompatibility similar to that of most other Prunus species. Based on the analyses of PCR and genomic DNA blots, a common S-RNase gene, which is designated as S f-RNase, has been found in Japanese apricots (Tao et al. 2000, 2002). Three partial cDNA fragments (MSRN 1–3) (Yaegaki et al. 2001) and seven S-RNase genes (PMSRNA 1–7) have been identified using PCR methods. Subsequently, two S-RNase genes have been isolated and are designated as S 8 -RNase (Tao et al. 2002) and S 9 -RNase (Entani et al. 2003). The PmS 10 –S 16 -RNases have been identified in cultivars of P. mume (Heng et al. 2008). Based on the sequences of S-RNase, the corresponding SFBs and the physical distances between S-RNase and SFB in the S haplotypes of trans-specific were identified in the current study. Our results may help reveal the characteristics of SFBs and the patterns of trans-specific evolution in Prunus.

Materials and Methods

Plant Materials and Genomic DNA Extraction

The P. mume cultivars ‘Xiaoyezhugan’ (S 10 S 14 ), ‘Fubantiaozhi’ (S 12 S 13 ), ‘Duoezhusha’ (S 3 S 4 ), ‘Dantaofen’ (S 4 S 12 ), ‘Longyanmei’ (S 13 S 14 ), ‘Hongding’ (S 15 S 16 ), ‘Xiyeqing’ (S 3 S 15 ) and ‘Musashino’ (S 5 S 11 ) (Heng et al. 2008) as well as the P. armeniaca cultivars ‘Jiguang’ (S 8 S 9 ) and ‘Erhong’ (S 9 S 11 ) were used. Young leaves were collected in the spring and used for total DNA extraction (Bourguiba et al. 2010; Campoy et al. 2010).

PCR Amplification for PmSFBs

Consensus primers were designed from the conserved regions SFB-F1 and SFB-R1 (Ikeda et al. 2004) for amplifying PmSFB 10 , PmSFB 12 , PmSFB 13 , PmSFB 14 , PmSFB 15 , PmSFB 16 , and ParSFB 9 , respectively. The PmS 11 -RNase-specific primer designed between RC4 and C5 and the PmSFB-F1 primer were used to obtain the intergenic sequence that contains PmSFB 11 . The PCR amplification was performed using ExTaq polymerase (Takara, Kyoto, Japan) using a temperature profile of a cycle of 95°C for 3 min; 10 cycles of 94°C for 30 s, 58°C for 60 s, and 72°C for 1 min 30 s; 25 cycles of 95°C for 30 s, 55°C for 60 s, and 72°C for 1 min 30 s; and a final extension cycle at 72°C for 10 min in a Thermal Cycler (BioRad, PTC-200). The PCR products were purified from agarose gels using the QIAEX® kit (Qiagen, Hilden, Germany) and cloned into the pGEM-T vector (Promega, Madison, WI, USA) according to the manufacturer's instructions. Approximately 20 plasmid clones were analysed using the restriction enzymes Csp6 I and EcoR I. Three clones with the same restriction patterns were sequenced to obtain a consensus sequence. If the same sequence was isolated from two cultivars possessing a common S-allele, the sequence was considered to represent the shared allele. Subsequent additional sequences were attributed to the remaining allele.

Analysis of the Physical Distance Between S-RNase and SFB in P. mume

To determine the physical distance between the S-RNase and SFB in the S haplotype, the S-RNase gene-specific primer and the SFB-F2 primer, which were designed from the variable region of S-RNase and the conserved region of the 3′ SFBs in Prunus sequences, respectively, were used (Table 1). The PCR reactions, purification, cloning and sequencing were performed as described above. The DNA sequence data were analysed using the BioEdit ver. 7.0 software and the ClustalX program (Zhang et al. 2011).

Table 1 The primers for amplification of SFB genes in P. mume and P. armeniaca
Full size table

Assessment of SFB Intron Polymorphism in P. mume

The SFB-5′A primer and the SFB-specific primer (Table 1), which were designed from the conserved region of the 5′UTR and the variable region of SFB, respectively, were used to amplify the intron that was located in the 5′UTR of each isolated SFB allele. The amplification products were purified, cloned, and sequenced as described above.

Results

PCR Amplification for PmSFBs

Partial sequences of PmSFB 10 , PmSFB 12 , PmSFB 13 , PmSFB 14 , PmSFB 15 , PmSFB 16 , and ParSFB 9 were isolated successfully from the various cultivars as follows: PmSFB 10 from ‘Xiaoyezhugan’, PmSFB 12 from ‘Dantaofen’ and ‘Fubantiaozhi’, PmSFB 13 from ‘Fubantiaozhi’ and ‘Longyanmei’, PmSFB 14 from ‘Xiaoyezhugan’ and ‘Longyanmei’, PmSFB 15 from ‘Hongding’ and ‘Xiyeqing’, PmSFB 16 from ‘Hongding’, and ParSFB 9 from ‘Jiguang’ and ‘Erhong’. However, PmSFB 11 was not confirmed because no other cultivar shared the same PmSFB. The PmS 11 -RNase-specific primer that was designed between RC4 and C5 and the PmSFB-F1 primer were used to obtain the intergenic distance in this haplotype. Based on the DNA sequence analysis, open reading frames were identified that corresponded to PmS 11 -RNase (partial) and PmSFB 11 (partial). In addition, the intergenic distance between PmS 11 -RNase and PmSFB 11 was determined.

In an alignment of the SFBs of P. mume, 170 out of 377 sites were conserved, and an additional 62 sites had only conservative replacements, which were scattered throughout the SFB. Many residues were conserved at the N-terminal region of the F-box motif and at most of the variable sites, which were located at the C-terminal region of SFB, which also contained two hypervariable regions, HVa and HVb (Fig. 1) as previously reported (Ushijima et al. 2003; Ikeda et al. 2004). The identities amongst these PmSFBs ranged from 73.2 to 90.9% at the amino acid level.

Fig. 1
figure 1

Alignment of the deduced amino acid sequences of the P. mume SFB, S 1 , S 7 and S 10 –S 16 by the ClustalX method using BioEdit software. The F-box motif and the hypervariable regions V1, V2, HVa and HVb are underlined. Asterisks indicate identical residues, colons indicate conserved substitutions, dots indicate semi-conserved substitutions and dashes within the sequences correspond to gaps

Full size image

Physical Distance Between SFB and S-RNase

S-RNase gene-specific primers and the SFB-F2 primer were used to investigate the PmS 10 –PmS 16 haplotypes. However, only four out of seven haplotypes were confirmed. The intergenic distances for PmS 10 , PmS 11 , PmS 12 and PmS 13 haplotypes were 339 bp, 974 bp, 1,710 bp and 623 bp, respectively; were lowly polymorphic with similarities that ranged from 31.9% to 41.9%; and displayed C+G content ranging from 32.91% to 40.29% at the nucleotide level. The intergenic region between the S-RNase and the SFB genes in ParS 9 haplotypes was also successfully obtained using ParS 9 -RNase-F, which was a forward S-RNase-specific primer, and the SFB2 primer, which was a reverse primer that was designed from the 3′ region of Prunus SFB (Table 1). The size of the intergenic region was 619 bp, suggesting that the two genes were in opposite transcriptional orientations. SFB was located downstream of the S-RNase and in the reverse transcriptional orientation. The relative order and transcriptional orientation of S-RNase and SFB genes were conserved amongst the four P. mume S haplotypes.

Discrimination of the Introns of New SFB Alleles

The putative introns in PmSFBs were observed in the 5′UTRs showing small length polymorphisms. The intron lengths of PmSFB 10 , PmSFB 11 , PmSFB 13 , PmSFB 14 , PmSFB 16 and PmSFB f varied from 89 bp to 114 bp. The sequence polymorphism of introns varied from 69.3% (PmSFB 10 with PmSFB 11 ) to 94.4% (PmSFB 14 with PmSFB 16 ), and GC content ranging from 28.1% to 30.7% at the nucleotide level. The introns of SFBs in P. mume exhibited high sequence polymorphisms with those of other species in Prunus, which similarity scores ranged from 59.1% (PmSFB 10 with PaSFB 5 ) to 98.9% (PmSFB 13 with ParSFB 9 ) at the nucleotide level (Fig. 2). Similar to reports regarding the sweet cherry, the conserved motif at the 3′-intron border ‘TDCAG’ (except for PaSFB 16 ) and the conserved motif at the 5′-intron border ‘TAAGT’ were noted in Prunus SFBs (Yamane et al. 2003a).

Fig. 2
figure 2

Comparison of SFB introns amongst the species in Prunus

Full size image

Comparison of PmS 11 with PspS 3-1 Haplotype and PmS 13 , PspS 8 with ParS 9 Haplotype

The identity between PmSFB 11 and PspSFB 3-1 was exceptionally high (Nunes et al. 2006). The alignment of these two SFBs showed that only 8 bp were different in the coding sequences with 98.8% similarity, and four residues differed from each other yielding 99.1% similarity at the amino acid level (Fig. 3a). Another alignment showed that the identity between the PmS 11 -RNase (Heng et al. 2008) and the PspS 3-1 -RNase was 99.2% at the nucleotide level with differences in 5 bp, and 99% at the amino acid level with differences in two residues (Fig. 3b). Moreover, sequence comparisons between the two introns in S-RNase indicated that the identity at the nucleotide level was 99.2% between the first intron sequences and 96.4% between the second intron sequences. The distance between the PspS 3-1 -RNase and PspSFB 3-1 was 1,343 bp greater than that in the PmS 11 haplotype. Surprisingly, the identity between the intergenic sequences of these two haplotypes was 94.6%.

Fig. 3
figure 3

a Comparison of the amino acid sequences of PmSFB 11 and PspSFB 3-1 . The hypervariable regions V2, HVa and HVb are underlined. Asterisks indicate identical residues, colons indicate conserved substitution, dots indicated semi-conserved substitution and dashes correspond to gaps. b The amino acid sequences were compared between the PmS 11 -RNase and the PspS 3-1 -RNase. The conserved regions (C1, C2, C3, RC4 and C5) and the hypervariable region (RHV) are underlined. Asterisks indicate identical residues, colons indicate conserved substitutions, dots indicated semi-conserved substitution and dashes correspond to gaps

Full size image

The identities amongst PmSFB 13 , PspSFB 8 and ParSFB 9 were also exceptionally high. The alignment showed that seven residues were different between PmSFB 13 and PspSFB 8 and five residues were different between PmSFB 13 and ParSFB 9 (Fig. 4a). PmS 13 -RNase (Heng et al. 2008) was similar to PspS 8 -RNase with 99.1% identity at the protein level and 98.3% at the nucleotide level, and it was similar to ParS 9 -RNase with 99.1% identity at the protein level and 99.3% at the nucleotide level (Fig. 4b). Sequence comparison showed that the identity of the two intron sequences of PmS 13 -RNase was high compared with PspS 8 -RNase or the ParS 9 -RNase. The identities between PmS 13 -RNase and PspS 8 -RNase were 96.9% at the first intron sequences and 97.4% at the second intron sequences. The identities between PmS 13 -RNase and ParS 9 -RNase were 98.8% at the first intron sequences and 98.3% at the second intron sequences. Moreover, the intergenic sequence comparison showed that the similarity of PmS 13 was 96.1% to the PspS 8 haplotype and was 99.0% to the ParS 9 haplotype at the nucleotide level.

Fig. 4
figure 4

a Comparison of the amino acid sequences of PmSFB 13 , PspSFB 8 and ParSFB 9 . The F-box motif and the hypervariable regions V1, V2, HVa, and HVb are underlined. Asterisks indicate identical residues, colons indicate conserved substitutions, dots indicate semi-conserved substitutions and dashes within the sequences correspond to gaps. b The amino acid sequences were compared amongst the PmS 13 -RNase, the PspS 8 -RNase and the ParS 9 -RNase. Five conserved regions (C1, C2, C3, RC4 and C5) and the hypervariable (RHV) region are underlined. Asterisks indicate identical residues, colons indicate conserved substitutions and dashes correspond to gaps

Full size image

Discussion

Sequence Analysis and Comparison of SFBs

Similar to other Prunus SFBs, PmSFBs were found to have one F-box domain, V1, V2, HVa and HVb, which are involved in the allelic specificity of the GSI reaction (Ushijima et al. 2003). The identities amongst the PmSFBs ranged from 73.2 to 90.9% at the amino acid level. Interestingly, the high identities amongst the SFBs occurred frequently across Prunus species. For example, the identity between PmSFB 11 and PspSFB 3-1 was 94.6%, and the identity of PmSFB 13 was 98.7% with PspSFB 8 and 99.3% with ParSFB 9 at the amino acid level. These high identities indicated that the Prunus S haplotypes might have a common ancestor.

The sizes of the introns varied from 89 to 114 bp, and the similarity amongst them ranged from 69.3% to 94.4%. Compared with the first and second introns of the S-RNases, the introns of SFBs exhibited lower length polymorphism, which may be constrained due to a specific regulatory role for the 5′UTR intron in gene expression (Chung et al. 2006). In addition, 50 bp was the minimum requirement for effective splicing (Deutsch and Long 1999). Moreover, the introns of three pairs, PmSFB 13 with ParSFB 9 , PmSFB 14 with PdSFB-c, and PaSFB 6 with ParSFB 1 , showed high similarities, whereas the coding sequences of the SFBs, with the exception of PmSFB 13 with ParSFB 9 , did not. These results indicate that the evolution of SFB coding and non-coding sequences may not occur simultaneously.

SFBs are Tightly Linked with S-RNases

The tight linkage between the SFB and S-RNase alleles is one of the mechanisms that are used to suppress recombination between these two genes (Zhang et al. 2007). In the current study, four intergenic sequences at the S locus were obtained from P. mume cultivars using allele-specific primers and SFB-F2. We demonstrated that the sizes of these sequences ranged from 339 to 1,710 bp and that the two intergenic sequences at the S locus were obtained from P. armeniaca cultivars. The intergenic regions in the Pm 14 –Pm 16 S haplotypes were not obtained successfully due to their great length as was also found for the P. avium S 4 haplotype (Ikeda et al. 2005). Moreover, the sequence order at the two ends of the intergenic region showed that SFB was located downstream of the S-RNase and in the reverse transcriptional orientation.

Encoding the Sequences of S-RNase and SFB Gene

SFBs that were recovered from the pairs of PmS 11 /PspS 3-1 and PmS 13 /ParS 9 /PspS 8 showed a low degree of variation. However, this variation was greater than that amongst the S-RNases. There were 1–7 amino acid substitutions in the corresponding regions. Variant residues did not cluster in or near the four HV regions, which included the hypervariable residues that were identified by Ikeda et al. (2004) but were distributed uniformly throughout the polypeptide chain. Likewise, the alignment of the amino acid sequences of two pairs indicated differences that were distributed throughout the sequence, most of which were non-synonymous substitutions. In the pairs PmS 11 /PspS 3-1 and PmS 13 /ParS 9 , there were two residues from C1 to the stop codon region. Interestingly, the pairs PmS 11 /PspS 3-1 and PmS 13 /ParS 9 /PspS 8 shared identical RHV regions, which may encode specificity, and changes at these sites are more likely to generate new specificities (Ushijima et al. 1998). Given that S-RNases and SFB polypeptides interact in a specific way, we expect that the identity and polypeptide structure are highly conserved.

S-allele Evolution

Prunus S-RNase alleles have one small, variable-length intron immediately upstream of the C1 region and the other introns, which are located within the RHV region, have notable characteristics such as their lengths and sequence polymorphisms (Tao et al. 1999; Igic and Kohn 2001). The presence or absence of introns acts as a taxonomic marker between closely related species. The absence of introns from Maloideae S-RNases agrees with the marked divergence between the Maloideae and Prunoideae S-RNases, which has been noted by Ushijima et al. (1998). As non-coding regions, the introns of the S-RNase alleles have accumulated random mutations and show significant differences amongst the S haplotypes. The similar scores of the second introns, which ranged from 96.4% to 98.3%, between the PmS 11 -RNase and the PspS 3-1 -RNase as well as the PmS 13 -RNase and ParS 9 -RNase or the PspS 8 -RNase may indicate that a relatively recent divergence of these alleles or a recent introgression event occurred after the divergence of the species but prior to their geographical spread.

Šurbanovski et al. (2007) have found a reproductive Prunus species, which had a different SFB allele that was coupled with the S-RNase allele. In addition, Sutherland et al. (2008) has reported six pairs of S-loci alleles that exhibited exceptionally high identities in Prunus species. In the current study, we found two pairs of S haplotypes that exhibited exceptionally high identities. All of the pairs of these S haplotypes may arise from introgressive hybridisation with each other or from a common ancestor that existed before the separation of these species. After species divergence, the S haplotypes may have retained the same protein sequences in different lineages. Moreover, two pairs of intergenic sequences at the S locus were exceptionally high. These haplotypes may be formed by trans-specific pairs (Sutherland et al. 2008). Minor divergences are evident in the Prunus S-RNase and the SFB sequences of the trans-specific pairs, and it is still unclear whether these divergences are sufficient to produce a new S-allele specificity. Ideally, the function and specificity of the trans-specific S haplotypes in Prunus that are reported here should be examined in controlled crosses.