Diversification and Alteration of Recognition Specificity of the Pollen Ligand SP11/SCR in Self-Incompatibility of Brassica and Raphanus

Abstract

The recognition specificity of the pollen ligand of self-incompatibility (SP11/SCR) was investigated using Brassica rapa transgenic plants expressing SP11 transgenes, and SP11 of Raphanus sativus S-21 was found to have the same recognition specificity as that of B. rapa S-9. In a set of three S haplotypes, whose sequence identities of SP11 and SRK are fairly high, R. sativus S-6 showed the same recognition specificity as Brassica oleracea S-18 and a slightly different specificity from B. rapa S-52. B. oleracea S-18, however, showed a different specificity from B. rapa S-52. Using these similar S haplotypes, chimeric SP11 proteins were produced by domain swapping. Bioassay using the chimeric SP11 proteins revealed that the incompatibility response induction activity was altered by the replacement of Region III and Region V. Pollen grains of Brassica transgenic plants expressing chimeric SP11 of the B. oleracea SP11-18 sequence with Region III and Region V from B. rapa SP11-52 (chimeric BoSP11-18[52]) were partially incompatible with the B. rapa S-52 stigmas, and those expressing the R. sativus SP11-6 sequence with Region III and Region V from B. rapa SP11-52 (chimeric RsSP11-6[52]) were completely incompatible with the stigmas having B. rapa S-52. However, the transgenic plant expressing chimeric RsSP11-6(52) also showed incompatibility with B. oleracea S-18 stigmas. These results suggest that Regions III and Region V of SP11 are important for determining the recognition specificity, but not the sole determinant. A possible process of the generation of a new S haplotype is herein discussed.

INTRODUCTION

The molecular mechanism of pollen-pistil recognition in self-incompatibility has been intensively studied in Brassica oleracea and Brassica rapa. The pollen-pistil recognition takes place by the interaction between membrane-spanning Ser-Thr protein kinase in the stigma, i.e., S-locus receptor kinase (Stein et al., 1991; Takasaki et al., 2000), and low–molecular weight Cys-rich protein in pollen, i.e., S-locus protein 11(SP11)/S-locus Cys-rich protein (SCR) (Schopfer et al., 1999; Suzuki et al., 1999). The genes encoding these molecules, i.e., SRK and SP11/SCR (referred to as SP11 hereafter), are located in the S locus, and another molecule encoded in the S locus is S-locus glycoprotein (SLG), the function of which in self-incompatibility is still controversial (Gaude et al., 1995; Okazaki et al., 1999; Dixit et al., 2000; Suzuki et al., 2000; Takasaki et al., 2000; Silva et al., 2001). The alleles of SRK, SP11, and SLG are transmitted to the progeny as one set named S haplotype. On the basis of the nucleotide sequence similarity of SLG and SRK, S haplotypes fall into two classes, class I and class II (Nasrallah et al., 1991). The identities of the deduced amino acid sequences of SLG and the S domain of SRK are >72% in both classes and <70% between the two classes (Chen and Nasrallah, 1990; Nishio and Kusaba, 2000). The class-II S haplotypes are generally recessive to class-I S haplotypes in pollen and express weaker self-incompatibility than the class-I S haplotypes (Nasrallah et al., 1991).

The numbers of S haplotypes in B. oleracea and B. rapa have been estimated to be 50 and >100, respectively (Nou et al., 1993; Ockendon, 2000). Nucleotide sequence analyses of SRK, SP11, and SLG of many S haplotypes have revealed that B. oleracea and B. rapa have pairs of S haplotypes in common, the amino acid identities of SRK and SP11 between the pairs being >90% (Sato et al., 2002, 2003). These interspecific pairs have been proved to possess the same recognition specificity (Kimura et al., 2002; Sato et al., 2003). Intergeneric pairs have also been identified between Brassica and Raphanus, the identity of amino acid sequences being lower than that of the interspecific pairs (Okamoto et al., 2004). The recognition specificity of the intergeneric pairs has not been investigated.

The recognition mechanism between SP11 and SRK has been fairly well elucidated (Kachroo et al., 2001; Takayama et al., 2001), but the regions of these recognition molecules important for recognition specificity have not been identified. Sato et al. (2003) have assigned six regions to SP11, Region I to Region VI, on the basis of conserved Cys residues, and considered Regions III, V, and VI to be more important for recognition specificity than Regions I, II, and IV, because amino acid sequences in Regions III, V, and VI are conserved between the interspecific pairs. Determining the solution structure of the SP11 protein of B. rapa S-8, Mishima et al. (2003) have identified a hypervariable region, which they considered to be important for recognition specificity. This hypervariable region corresponds to Region IV, which has been considered to be unimportant by Sato et al. (2003). Recently, Chookajorn et al. (2004) have shown that the specificity of SP11 can be altered by the substitution of four continuous amino acid residues in Region V.

In this study, we investigated the recognition specificity of an intergeneric pair between Brassica and Raphanus using transgenic plants. We also compared the recognition specificity of a set of three S haplotypes, the shared amino acid identities of which are fairly high, in Raphanus sativus, B. oleracea, and B. rapa. Chimeric SP11 genes were produced by swapping the sequences between a set of S haplotypes, and the recognition specificity of chimeric SP11 proteins was investigated using a bioassay with recombinant SP11 proteins and a pollination test with transgenic plants expressing the chimeric SP11 genes. The regions of the SP11 protein important for the recognition specificity and the process of generation of a new S haplotype are herein discussed.

RESULTS

The Same Recognition Specificity between R. sativus S-21 and B. rapa S-9

We investigated the recognition specificity of an intergeneric pair between Raphanus and Brassica, in which amino acid identities of SRK and SP11 are lower than those of the interspecific pairs between B. oleracea and B. rapa (Sato et al., 2003). The most similar pair between Raphanus and Brassica was a pair of S-21 in R. sativus (RsS-21) and S-9 in B. rapa (BrS-9), the amino acid identity of SP11 being 89.5% and that of the S domain of SRK being 91.0% (Table 1) (Okamoto et al., 2004). Different amino acid residues between SP11 of RsS-21 (RsSP11-21) and that of BrS-9 (BrSP11-9) are located in Regions II, IV, and VI (Figure 1A). Our trial to investigate the recognition specificity of this intergeneric pair by a bioassay using recombinant SP11 proteins failed because of the low yield of the recombinant RsSP11-21 protein synthesized by Escherichia coli. The yield of the RsSP11-21 protein was about one-tenth of that of the BrSP11-9 protein. Our preliminary experiment using several chimeric SP11 genes produced by DNA shuffling (Zhao and Arnold, 1997) between RsSP-21 and BrSP11-9 suggested that the difference of the efficiency of producing recombinant SP11 protein was because of the amino acid residues in Region VI. A bioassay of BrSP11-9 using R. sativus stigmas was also unsuccessful, because of the difficulty of the treatments of the recombinant proteins on the small stigmas of R. sativus and unstable results of control experiments.

Table 1.

Amino Acid Identity of SP11 and S Domain of SRK between S Haplotypes Used in This Study

	Amino Acid Identity
Pairs of S Haplotypes	SP11	SRK
RsS-21 versus BrS-9	89.5	91.0
RsS-6 versus BoS-18	70.9	88.3
RsS-6 versus BrS-52	69.1	86.6
BoS-18 versus BrS-52	58.2	87.6
The highest in intraspeciesa	57.1	87.3

^aThe highest amino acid identity in B. oleracea.

Figure 1.

B. rapa Transgenic Plants Expressing RsSP11-21 and Chimeric RsSP11-21(9).

(A) Deduced amino acid sequences of RsSP11-21 and BrSP11-9 (shaded background). Boxes indicate Cys residues, and asterisks show the same amino acid residues between RsSP11-21 and BrSP11-9. The sequence of chimeric RsSP11-21(9) used for the plant transformation is also shown.

(B) The T region of the vector used for B. rapa transformation. The coding region of SP11 is indicated by a shaded box. SP11 pro, BrSP11-46 promoter; NOS T, terminator of nopaline synthase gene; 35S pro, Cauliflower mosaic virus 35S promoter; NPTII, neomycin phosphotransferase gene; OCS T, terminator of octopine synthetase gene; RB, right border; LB, left border.

(C) DNA gel blot analysis of the transgenes. After electrophoresis and transfer to a membrane, DNA digested with BamHI and HindIII was hybridized with a probe of a mixture of the coding regions of RsSP11-21 and chimeric RsSP11-21(9). The arrow indicates the 1.5-kb band of the transgenes. DNA size markers in kb are shown to the left. M, molecular marker; WT, nontransgenic BrS-52/S-60 plant. T21, transgenic plants carrying wild-type RsSP11-21; T21(9), those carrying chimeric RsSP11-21(9).

(D) RNA gel blot analysis of the transgenes. The same probe as in the DNA gel blot analysis was used. RNA size markers in kb are shown to the left. BrS-9, a BrS-9 homozygote.

Because of the difficulties of the bioassay using the recombinant SP11 proteins, we constructed genes of a 1-kb sequence of BrSP11-46 promoter, cDNA of SP11, and NOS terminator, in which the SP11 cDNA was wild-type RsSP11-21 or chimeric RsSP11-21(9) composed of Region I to Region V of RsSP11-21 and Region VI of BrSP11-9, and introduced them into B. rapa cv Osome (a BrS-52/S-60 heterozygote) (Figures 1A and 1B). Five independent transgenic plants carrying wild-type RsSP11-21 (T21-A, T21-B, T21-C, T21-D, and T21-E) and three independent transgenic plants carrying chimeric RsSP11-21(9) [T21(9)-A, T21(9)-C, and T21(9)-D] were obtained. DNA gel blot analysis after digestion with a mixture of BamHI and HindIII showed a 1.5-kb band in all the transgenic plants (Figure 1C). Expression of wild-type RsSP11-21 and chimeric RsSP11-21(9) was detected in all the transgenic plants by RNA gel blot analysis, but the expression levels in T21(9)-C and T21(9)-D were low (Figure 1D). The pollen grains of the transgenic plants expressing wild-type RsSP11-21 or chimeric RsSP11-21(9) were pollinated onto the stigmas of homozygotes of BrS-9 and BrS-46 (Table 2, Figure 2). The pollen grains of all the transgenic plants were incompatible with the BrS-9 stigmas, but were compatible with the BrS-46 stigmas. This result indicates that RsSP11-21 has the same recognition specificity as BrSP11-9.

Table 2.

Incompatibility of Pollen Grains of B. rapa Transgenic Plants Expressing RsSP11-21 or Chimeric RsSP11-21 with BrS-9 Stigmas

			Stigmas
Pollen Donors	S Haplotypes	Transgenes	BrS-9	BrS-46
T21-A	BrS-52/S-60	RsSP11-21	−	++
T21-B	BrS-52/S-60	RsSP11-21	−	++
T21-C	BrS-52/S-60	RsSP11-21	−−	+
T21-D	BrS-52/S-60	RsSP11-21	−	++
T21-E	BrS-52/S-60	RsSP11-21	−−	+
T21(9)-A	BrS-52/S-60	Chimeric RsSP11-21(9)	−	+
T21(9)-C	BrS-52/S-60	Chimeric RsSP11-21(9)	+−	++
T21(9)-D	BrS-52/S-60	Chimeric RsSP11-21(9)	−	+
Wild type	BrS-52/S-60		+	++
BrS-9			−	++

The indices are as follows: −−, completely incompatible (no or few germinating pollen grains on a stigma); −, incompatible (>30 pollen grains on a stigma and <5 pollen tubes penetrating papilla cells); +−, partially incompatible (5 to 30 pollen tubes penetrating papilla cells); +, compatible (30 to 100 pollen tubes penetrating papilla cells); ++, completely compatible (>100 pollen tubes penetrating papilla cells).

Figure 2.

Pollen Tube Growth from the Pollen Grains of the B. rapa Transgenic Plants Carrying RsSP11-21 and Chimeric RsSP11-21(9) in the Stigma of a BrS-9 Homozygote.

The BrS-9 homozygote and the wild-type plant of BrS-52/S-60 heterozygote (WT) were crossed as the controls. The pollen grains of the transgenic plants [T21-B and T21(9)-A] were incompatible with the stigmas of BrS-9.

Recognition Specificity of a Set of S Haplotypes in Raphanus and in Two Species of Brassica

Comparison of the nucleotide sequences of SRK and SP11 in many S haplotypes of B. oleracea, B. rapa, and R. sativus revealed a set of three fairly similar S haplotypes, RsS-6, BoS-18, and BrS-52 (Table 1). Interspecific hybrid plants, which have been successfully used for the test of recognition specificity of S haplotypes between B. oleracea and B. rapa (Kimura et al., 2002; Sato et al., 2003), were produced to investigate the recognition specificity of BoS-18 and BrS-52. BoS-18 and BrS-52 homozygotes were crossed with BrS-60 and BoS-15 homozygotes, respectively, in which BoS-15 and BrS-60 are class-II S haplotypes. A hybrid between BoS-18 and BrS-60 was an amphidiploid plant (BoS-18/BrS-60), whereas that between BrS-52 and BoS-15 was a diploid plant (BrS-52/BoS-15). The latter was sterile like diploid interspecific hybrids between these species reported so far, the sterility of which is caused by inability of chromosome pairing in meiosis. Therefore, only the stigmas of the hybrids were used for the pollination test.

Pollen grains of BoS-6, BoS-8, BoS-18, BrS-8, BrS-9, BrS-36, BrS-45, and BrS-52 homozygotes were pollinated onto the stigmas of the interspecific hybrid plants (Table 3). When pollinated onto the stigmas of BoS-18/BrS-60, only pollen grains of BoS-18 were incompatible. The pollen grains of BrS-52 were completely compatible with the stigmas of BoS-18/BrS-60. When pollinated onto the stigmas of BrS-52/BoS-15, only pollen grains of BrS-52 were incompatible, whereas those of BoS-18 were completely compatible with the stigmas of BrS-52/BoS-15. These results indicate that BoS-18 and BrS-52 have different recognition specificity. The inability of the BrS-52 stigmas to recognize BoSP11-18 as self was also shown by a pollination test using B. rapa transgenic plants carrying a BoSP11-18 transgene as the control experiment for the test of a chimeric gene (data are shown later in Table 5).

Table 3.

Pollination Tests of BoS-18 and BrS-52 Using Interspecific Hybrids Having BoS-18/BrS-60 and BrS-52/BoS-15

	Stigmas
Pollen Donors	BoS-18/BrS-60	BrS-52/BoS-15
BoS-6	++	++
BoS-8	++	++
BoS-18	−−	++
BrS-8	++	++
BrS-9	++	+
BrS-36	++	++
BrS-45	++	++
BrS-52	++	−

The indices are as follows: −−, completely incompatible (no or few germinating pollen grains on a stigma); −, incompatible (>30 pollen grains on a stigma and <5 pollen tubes penetrating papilla cells); +, compatible (30 to 100 pollen tubes penetrating papilla cells); ++, completely compatible (>100 pollen tubes penetrating papilla cells).

Table 5.

Incompatibility between B. rapa S-52 and the Transgenic B. rapa Plants Expressing Chimeric BoSP11-18(52) or BoSP11-18

			Stigmas
Pollen Donors	S Haplotype	Transgenes	BrS-52	BrS-46
T18(52)-A	BrS-60	Chimeric BoSP11-18(52)	+−	++
T18(52)-G	BrS-60	Chimeric BoSP11-18(52)	+−	++
T18(52)-I	BrS-60	Chimeric BoSP11-18(52)	+−	++
T18-A	BrS-60	BoSP11-18	+	++
T18-C	BrS-60	BoSP11-18	+	++
T18-E	BrS-60	BoSP11-18	+	++
T18-F	BrS-60	BoSP11-18	+	++
Wild type	BrS-60		+	++
BrS-52			−	++

The indices are as follows: −, incompatible (>30 pollen grains on a stigma and <5 pollen tubes penetrating papilla cells); +−, partially incompatible (5 to 30 pollen tubes penetrating papilla cells); +, compatible (30 to 100 pollen tubes penetrating papilla cells); ++, completely compatible (>100 pollen tubes penetrating papilla cells).

The recognition specificity of RsS-6 was investigated using transgenic plants of B. oleracea carrying an RsSP11-6 transgene. The RsSP11-6 gene under the BrSP11-46 promoter was introduced to B. oleracea cv Green Comet, heterozygotes of BoS-39/S-15 or BoS-2/S-15 (Figure 3A), the pollen phenotype of which is S-39 or S-2, respectively, because S-2 of B. oleracea is dominant to S-15 in pollen (Thompson and Taylor, 1966) and S-39 is also dominant to S-15 (data not shown). Three independent transgenic plants (T6-A, T6-B, and T6-C) were obtained. DNA gel blot analysis detected a 2.1-kb band in the transgenic plants after double-digestion with EcoRI and HindIII (Figure 3B). The expression of the transgene was observed in all the transgenic plants using RNA gel blot analysis (Figure 3C). Pollen grains of the three transgenic plants expressing RsSP11-6 were incompatible with the stigmas of the BoS-18 homozygote and the BoS-18/BrS-60 interspecific hybrid (Figure 4, Table 4). For the pollination test of BrS-52, the BrS-52/BoS-15 interspecific hybrid was used to avoid the interspecific barrier. The pollen grains of all the transgenic plants were partially incompatible with those of the BrS-52/BoS-15 interspecific hybrid. These results indicate that RsSP11-6 is recognized as self by the BoS-18 stigmas and as a little different from BrSP11-52 by the stigmas having BrS-52.

Figure 3.

B. oleracea Transgenic Plants Expressing RsSP11-6, Chimeric RsSP11-6(18), and Chimeric RsSP11-6(52).

(A) The T region of the vector used for B. oleracea transformation. The coding region of SP11 is indicated by a gray box. SP11 pro, BrSP11-46 promoter; NOS T, nopaline synthase gene terminator; 35S pro, Cauliflower mosaic virus 35S promoter; NPTII, neomycin phosphotransferase gene; HPT, hygromycin phosphotransferase gene; RB, right border; LB, left border.

(B) DNA gel blot analysis of the transgenes. DNA was digested with EcoRI and HindIII, and probed with a mixture of the coding regions of RsSP11-6 and BrSP11-52. The arrow indicates the 2.1-kb band of the transgenes. DNA size markers in kb are shown to the left. M, molecular marker; WT, nontransgenic BoS-39/S-15 heterozygote (39), BoS-2/S-15 heterozygote (2); T6, transgenic plants carrying wild-type RsSP11-6; T6(18), those carrying chimeric RsSP11-6(18); T6(52), those carrying chimeric RsSP11-6(52).

(C) RNA gel blot analysis of the transgenes. The same probe was used as in the DNA gel blot analysis. RNA size markers in kb are shown to the left. BrS-52, a BrS-52 homozygote.

Figure 4.

Pollen Tube Growth from the Pollen Grains of the Transgenic Plants Carrying RsSP11-6, Chimeric RsSP11-6(18), and Chimeric RsSP11-6(52) in the Stigmas of the BoS-18 Homozygote (BoS-18) and the BrS-52/BoS-15 Interspecific Hybrid (BrS-52*).

Pollen grains of the transgenic plants carrying RsSP11-6 (T6-B) were incompatible with the BoS-18 stigmas, but only partially incompatible with the stigmas having BrS-52. The transgenic plants carrying chimeric RsSP11-6(18) [T6(18)-A] were incompatible with BoS-18, but compatible with BrS-52*. Those carrying chimeric RsSP11-6(52) [T6(52)-A] were incompatible with the stigmas of both BoS-18 and BrS-52*.

Table 4.

Incompatibility of Pollen Grains of B. oleracea Transgenic Plants Expressing RsSP11-6 with the Stigmas of BoS-18 and BrS-52

			Stigmas
Pollen Donors	S Haplotypes	Transgene	BoS-18	BoS-18/BrS-60	BrS-52/BoS-15	BoS-8
T6-A	S-39/S-15	RsSP11-6	−−	−−	+−	++
T6-B	S-2/S-15	RsSP11-6	−−	−−	+−	++
T6-C	S-2/S-15	RsSP11-6	−	−	+−	++
WT1	S-39/S-15		++	++	++	++
WT2	S-2/S-15		++	++	++	++

The indices are as follows: −−, completely incompatible (no or few germinating pollen grains on a stigma); −, incompatible (>30 pollen grains on a stigma and <5 pollen tubes penetrating papilla cells); +−, partially incompatible (5 to 30 pollen tubes penetrating papilla cells); ++, completely compatible (>100 pollen tubes penetrating papilla cells).

Analysis of the Regions Important for Recognition Specificity of BrSP11-52 by Bioassay Using Recombinant SP11 Proteins

Using BoS-18 and BrS-52, which are similar but have different recognition specificity, chimeric SP11 proteins were produced by domain swapping to investigate the regions important for the recognition specificity. Seven chimeric SP11s [52(II-1), 52(II-2), 52(III), 52(IV-1), 52(IV-2), 52(V-1), and 52(V-2)] were produced to have the sequence of BrSP11-52 with short segments of Region II, III, IV, and V from BoSP11-18 (Figure 5A). Regions I and Regions VI were not replaced because of the high similarity between BoSP11-18 and BrSP11-52. Recombinant proteins of the seven chimeric SP11s, wild-type BrSP11-52, and wild-type BoSP11-18 were applied onto the surface of the stigmas of S-52 homozygotes in B. rapa and compatible pollen grains were pollinated. The levels of incompatibility induced by the recombinant proteins were rated with indices, i.e., 1 (incompatible) to 4 (compatible), based on the number of pollen tubes penetrating the stigma. The results of this bioassay were not as stable as those of the pollination test using transgenic plants; therefore, the bioassay was performed for 23 d using 55 to 75 flowers for each treatment in total. Averaged indices of BrSP11-52 and BoSP11-18 were 1.7 and 2.5, respectively (Figure 5B). Replacements of Region III and Region V reduced significantly the incompatibility response induction activity in the BrS-52 stigmas, whereas those of Region II and Region IV did not. The recombinant proteins having Region II or Region IV of BoSP11-18 induced significantly higher incompatibility response than BoSP11-18. These results suggest that Region III and Region V of BrSP11-52 are more important for the recognition specificity than Region II and Region IV.

Figure 5.

Sequences of Chimeric SP11s and Analysis of Recognition Specificity of the Chimeric SP11 Proteins by Bioassay.

(A) The chimeric SP11-52s [52(II-1), 52(II-2), 52(III), 52(IV-1), 52(IV-2), 52(V-1), and 52(V-2)] were produced by replacing the BrSP11-52 sequence with small segments of the BoSP11-18 sequence. A chimeric SP11-18 [18(III,V)] was produced by replacing Region III and Region V of the BoSP11-18 sequence with the sequence of BrSP11-52. The white and gray boxes indicate the conserved Cys residues and the amino acid sequences that were replaced, respectively.

(B) The bioassay of incompatibility response inducing ability of recombinant proteins of chimeric SP11s using the stigmas of S-52 homozygotes in B. rapa. The level of incompatibility was rated with the indices as follows: 1, the pollen tubes penetrating into the stigma being <10; 2, those being 10 to 30; 3, those being from 30 to 100; 4, those being >100. The values are the means ± se of 54 to 75 stigmas. ** and * represent significant differences from an incompatible control, BrSP11-52, at 1 and 5% levels, respectively. ^†† and ^† represent significant differences from a compatible control, BoSP11-18, at 1 and 5% levels, respectively.

The BoSP11-18 sequence was replaced with the sequences of BrSP11-52 in Region III and Region V [18(III,V), Figure 5A], and the incompatibility response induction activity of 18(III,V) was tested by the bioassay using 54 stigmas of the BrS-52 homozygotes. Significantly higher activity inducing the incompatibility response than BoSP11-18 was observed (Figure 5B).

Alteration of the Recognition Specificity of SP11 by Swapping Region III and Region V

The bioassay revealed Region III and Region V of SP11 to be important for the recognition specificity. To confirm this finding, we investigated the recognition specificity of chimeric SP11s using the plant transformation and pollination tests. Chimeric BoSP11-18(52), in which Region III and Region V of BoSP11-18 were replaced by those of BrSP11-52, was produced (Figure 6), and chimeric BoSP11-18(52) and wild-type BoSP11-18 were introduced into B. rapa S-60 homozygotes. Five independent transgenic plants carrying chimeric BoSP11-18(52) [T18(52)-A, T18(52)-B, T18(52)-F, T18(52)-G, and T18(52)-I] and seven independent transgenic plants carrying wild-type BoSP11-18 (T18-A, T18-B, T18-C, T18-D, T18-E, T18-F, and T18-G) were obtained. DNA gel blot analysis after double-digestion with BamHI and HindIII detected a 1.5-kb band in all the transgenic plants, except for 18-G, in which a 3.0-kb band was detected (Figure 7A). The expression of the transgenes in the transgenic plants was confirmed by RNA gel blot analysis except for T18(52)-B, T18-D, and T18-G (Figure 7B). The expression of the transgene was not detected in T18(52)-B, and the levels of the expression in T18-D and T18-G were low. T18(52)-F was sterile and DNA gel blot analysis showed T18-B to have an additional band. Therefore, the transgenic plants of T18(52)-A, T18(52)-G, T18(52)-I, T18-A, T18-C, T18-E, and T18-F were used for pollination tests. When the pollen grains of T18(52)-A, T18(52)-G, T18(52)-I were pollinated onto the stigmas of BrS-52, they showed partial incompatibility (Table 5). The pollen grains of T18-A, T18-C, T18-E, and T18-F were compatible with the stigmas of BrS-52.

Figure 6.

Deduced Amino Acid Sequences of BoSP11-18, BrSP11-52, RsSP11-6, and Chimeric SP11s Used for Brassica Transformation.

Boxes indicate the conserved Cys residues, and asterisks show the same amino acid residues between two sequences.

Figure 7.

Analysis of B. rapa Transgenic Plants Expressing Chimeric BoSP11-18(52) and BoSP11-18.

(A) DNA gel blot analysis of the transgenes. DNA was digested with BamHI and HindIII, and probed with a mixture of BoSP11-18 and chimeric BoSP11-18(52). The arrow indicates the 1.5-kb band of the transgenes. DNA size markers in kb are shown to the left. M, molecular marker; WT, nontransgenic BrS-60 plant; T18(52), transgenic plants carrying chimeric BoSP11-18(52); T18, those carrying wild-type BoSP11-18.

(B) RNA gel blot analysis of the transgenes. The same probe as in the DNA gel blot analysis was used. RNA size markers in kb are shown to the left. BoS-18, a BoS-18 homozygote.

Chimeric SP11 genes of RsSP11-6 were produced by replacing Region III and Region V of RsSP11-6 with the sequence of BoSP11-18 [chimeric RsSP11-6(18)] or BrSP11-52 [chimeric RsSP11-6(52)] (Figure 6). These genes were introduced into B. oleracea cv Green Comet, BoS-39/S-15 or BoS-2/S-15 heterozygote. The independent transgenic plants obtained were two with chimeric RsSP11-6(18), i.e., T6(18)-A and T6(18)-B, and one with chimeric RsSP11-6(52), i.e., T6(52)-A. DNA gel blot analysis detected 2.1-kb bands after double-digestion with EcoRI and HindIII in the transgenic B. oleracea plants (Figure 3B). RNA gel blot analysis showed the expression of the SP11 transgenes in all the transgenic plants (Figure 3C). The expression levels of the SP11 transgenes in the transgenic plants were as high as that of the endogenous SP11 in a BrS-52 homozygote. T6(18)-A and T6(18)-B carrying chimeric RsSP11-6(18) were incompatible with the stigmas of the BoS-18 homozygote and the interspecific hybrid of BoS-18/BrS-60, although completely compatible with the hybrid of BrS-52/BoS-15 (Figure 4, Table 6). T6(52)-A carrying chimeric RsSP11-6(52) was incompatible with the stigmas of BrS-52/BoS-15, BoS-18, and BoS-18/BrS-60.

Table 6.

Incompatibility of Pollen Grains of B. oleracea Transgenic Plants Expressing Chimeric RsSP11-6(18) and Chimeric RsSP11-6(52) with the Stigmas of BoS-18 and BrS-52

			Stigmas
Pollen Donors	S Haplotypes	Transgenes	BoS-18	BoS-18/BrS-60	BrS-52/BoS-15	BoS-8
T6(18)-A	S-2/S-15	Chimeric RsSP11-6(18)	−	−	++	++
T6(18)-B	S-2/S-15	Chimeric RsSP11-6(18)	−	−	++	++
T6(52)-A	S-39/S-15	Chimeric RsSP11-6(52)	−−	−−	−	++
WT1	S-39/S-15		++	++	++	++
WT2	S-2/S-15		++	++	++	++

The indices are as follows: −−, completely incompatible (no or few germinating pollen grains on a stigma); −, incompatible (>30 pollen grains on a stigma and <5 pollen tubes penetrating papilla cells); ++, completely compatible (>100 pollen tubes penetrating papilla cells).

DISCUSSION

Regions of SP11 Important for the Recognition by SRK

In this study, BrS-9 and RsS-21, which are an intergeneric pair of S haplotypes having highly similar SP11 and SRK alleles, were demonstrated to have the same recognition specificity of pollen-stigma interaction. It was also found that RsS-6 is incompatible with BoS-18 and partially incompatible with BrS-52. Alignment of the deduced amino acid sequences of BrSP11-9 and RsSP11-21 indicates that there are differences of amino acid residues in Region II, Region IV, and Region VI but not in Region III and Region V. Higher sequence polymorphism in Region IV than those in the other regions has been observed in the interspecific pairs of S haplotypes between B. oleracea and B. rapa (Sato et al., 2003). This observation may suggest that Region IV is not as important as Region III and Region V, which are conserved between the SP11 proteins in the interspecific pairs and the intergeneric pair. However, between BoS-18 and RsS-6, which have the same recognition specificity, there was no biased distribution of different amino acid residues in the SP11 sequences.

The results of the bioassay using chimeric SP11 proteins suggested that Region III and Region V are important for the recognition of BrSP11-52. This inference was supported by the results of the pollination tests using transgenic plants having chimeric SP11 genes, where the pollen grains of the transgenic plants carrying chimeric RsSP11-6(52) and chimeric BoSP11-18(52) were incompatible and partially incompatible, respectively, with the stigmas having BrS-52, whereas those of the transgenic plants carrying chimeric RsSP11-6(18) and wild-type BoSP11-18 were compatible with the BrS-52 stigmas. Using a bioassay with recombinant proteins, Chookajorn et al. (2004) have found that replacing only four continuous amino acid residues is sufficient to alter the recognition specificity from S-6 to S-13 in B. oleracea. These four amino acid residues are located in Region V. On the other hand, the pollen grains of the transgenic plants carrying chimeric RsSP11-6(52) were also incompatible with the stigmas of BoS-18. This result suggests that other regions of RsSP11-6 may also be important for the recognition by the BoS-18 stigmas.

Mishima et al. (2003) have elucidated the solution structure of the SP11 proteins, in which Region III forms an α-helix and Region V corresponds to the region from β2 to β3. Region III and Region V are arranged close to each other by disulfide linkage (Takayama et al., 2001; Mishima et al., 2003). The lower half of Region II is also located close to both Region III and Region V. The tertiary structure formed by these regions may determine the recognition specificity of the SP11 proteins. On the other hand, the loop in Region IV is located opposite the loop formed by Region V in the SP11 proteins. The loop in Region IV is the hypervariable region of the SP11 protein, which has been considered to serve as a specific binding site for SRK by Mishima et al. (2003). This study indicated that Region IV is not as important as Region III and Region V for determining the recognition specificity. It can be inferred that amino acid variations may be accumulated in Region IV because of its low importance of this region in the recognition function.

Evolution of S Haplotypes

The pair of BrS-9 and RsS-21 can be considered to have been derived from the same ancestral S haplotype, and the amino acid differences between these S haplotypes, which would have arisen after the divergence of these two genera, is not considered to have contributed to the alteration of recognition specificity. Although BoS-18, BrS-52, and RsS-6 are also considered to have been derived from the same ancestral S haplotype, remarkably lower sequence similarities between these S haplotypes than those between BrS-9 and RsS-21 may suggest that diversification of BoS-18, BrS-52, and RsS-6 had occurred before the divergence of these genera. If the sequence differences between these S haplotypes arose after the divergence of the genera and the species, the sequence similarity between BrS-52 and BoS-18 would be higher than that between RsS-6 and BrS-52 or between RsS-6 and BoS-18. However, the similarities of nucleotide sequences and deduced amino acid sequences between BrS-52 and BoS-18 were lowest in those between these three S haplotypes. Furthermore, BrS-52 and BoS-18 have different recognition specificity, indicating that they are different from the interspecific pairs of S haplotypes identified so far (Sato et al., 2003).

The consensus sequences of the SP11 and SRK alleles of RsS-6, BoS-18, and BrS-52, which can be regarded as the sequences of the putative ancestral S haplotype (S-X), were determined by selecting the nucleotides that are present in more than two sequences in the alignment of these three S haplotypes. The consensus amino acid sequences of SP11 and the S domain of SRK, except three amino acids for SP11 and five amino acids for SRK, were deduced. The sequence similarities of RsSP11-6, BoSP11-18, and BrSP11-52 to the consensus amino acid sequence of SP11, i.e., SP11-X, were 87.3, 78.2, and 72.7%, respectively, and those of RsSRK-6, BoSRK-18, and BrSRK-52 to the consensus sequence of the S domain of SRK, i.e., SRK-X, were 92.8, 93.8, and 92.3%, respectively (Figures 8A and 8B). The relationships between these sequences suggest that RsSP11-6 is closest to the putative ancestral SP11 sequence and that BrSP11-52 is most distantly related. The lowest amino acid sequence similarity between BoSP11-18 and BrSP11-52 among the four SP11 sequences including SP11-X and the different recognition specificity between BoSP11-18 and BrSP11-52 suggest that BoSP11-18 and BrSP11-52 have evolved so as to acquire different recognition specificity.

Figure 8.

Relationships of the Three S Haplotypes RsS-6, BoS-18, and BrS-52.

(A) Schematic representation of the relationships. S-X shows the putative ancestral S haplotype. Deduced amino acid identities and nucleotide identities of SP11 and SRK of BoS-18, RsS-6, and BrS-52 to those of S-X are shown on and under the bars, respectively. The thick and thin red bars show complete and partial incompatibility, respectively. The green bars show compatibility.

(B) Comparison of deduced amino acid sequences of SP11s. White boxes show the conserved Cys residues. Red boxes show amino acid residues different from those of SP11-X, and yellow boxes show the position in which the amino acid residues of SP11-X were not assumed.

Several models of the generation of a new S haplotype have been presented and discussed (Matton et al., 1999; Uyenoyama and Newbigin, 2000; Uyenoyama et al., 2001; Chookajorn et al., 2004). The relationship of BoS-18, BrS-52, and RsS-6 would provide interesting information to facilitate understanding of the mechanism of the S haplotype diversification. Their relationships of cross-incompatibility, i.e., complete incompatibility between BoS-18 and RsS-6, partial incompatibility between BrS-52 and RsS-6, and compatibility between BoS-18 and BrS-52, suggest that in the diversification of S haplotypes between BoS-18 and BrS-52, BoS-18 has maintained the incompatibility with RsS-6 and that BrS-52 has diverged more, resulting in lowering of its incompatibility with RsS-6. This process of S haplotype diversification supports the model presented by Chookajorn et al. (2004), in which gradual sequence modification maintaining the recognition specificity of SP11 and SRK is considered to contribute to the generation of a new S haplotype. Matton et al. (1999) have proposed that the generation of a new S haplotype passed through an intermediate stage in which it had dual specificity. RsSP11-6 can be regarded as dual-specific SP11 in a broad sense. Although the dual-specific S haplotypes would be eliminated from a population (Uyenoyama and Newbigin, 2000), it can be inferred that reproductive isolation may have contributed to the maintenance of RsSP11-6 in the population. The coexistence of a dual-specific S haplotype and other S haplotypes incompatible with it in a natural population has not been observed in any plant species so far.

The strength or stability of self-incompatibility is an important genetic trait for F1 hybrid breeding of Brassicaceae vegetables. Strength of self-incompatibility is controlled by the S locus as well as by genetic background. Weak incompatibility by the S locus may be caused by low affinity of SP11 and SRK, which was represented as partial incompatibility between the chimeric SP11 proteins and the wild-type SRK protein in this study. Alteration of strength of self-incompatibility might have occurred in the evolution of the S haplotypes.

METHODS

Plant Materials

Homozygotes of S-8, S-9, S-32, S-36, S-45, S-46, S-52, and S-60 in Brassica rapa and those of S-6, S-8, and S-18 in Brassica oleracea were used as plant materials. The S haplotypes in these species have been numbered in England and Japan independently (Nou et al., 1993; Ockendon, 2000). Interspecific hybrids between B. oleracea and B. rapa were raised by ovary culture according to Inomata (1977). B. rapa cv Osome (Takii Seed, Kyoto, Japan), which is a heterozygote of S-52 and S-60, and an S-60 homozygote derived from Osome were used for the transformation of B. rapa. For the transformation of B. oleracea, a broccoli cultivar, Green Comet (S-39/S-15 or S-2/S-15) (Takii Seed), was used.

Construction of the SP11 Genes for Transformation of Brassica

The mature protein region of the wild-type SP11 gene was amplified from SP11 cDNAs of BoSP11-18 (Sato et al., 2002), RsSP11-6, and RsSP11-21 (Okamoto et al., 2004) using primers shown in Supplemental Tables 1 and 2 online. The signal peptide region of SP11 was amplified from BrSP11-9 and BrSP11-52 using a primer pair of 9-ATG-BamHI and 9-S-R and a pair of 52-ATG-BamHI and 52-S-R, respectively. The sequence of BrSP11-9 was used for the construction of RsSP11-21 and that of BrSP11-52 was used for the construction of BoSP11-18 and RsSP11-6. The PCR product of the signal peptide region was mixed with that of the mature protein region. The mixture was used as a template of PCR with a primer pair of 9-ATG-BamHI and 21RS for RsSP11-21, that of 9-ATG-BamHI and 9RS for RsSP11-21(9), and that of 52-ATG-BamHI and 18RS for RsSP11-6 and BoSP11-18. The PCR products were cloned to the pGEM-T vector. Each clone that had the expected sequence was selected by nucleotide sequencing. The coding region of SP11 was ligated to the 1.0-kb sequence of the BrSP11-46 promoter (Sato et al., 2003), and inserted into the binary vector pSLJ491 (Jones et al., 1992) for the transformation of B. rapa, and into the position of the ubiquitin promoter and the spinach (Spinacia oleracea) GPAT gene of the binary vector used in rice (Oryza sativa) transformation (Ariizumi et al., 2002) for the transformation of B. oleracea.

Plant Transformation

The constructs were introduced into Agrobacterium tumefaciens strain EHA105 for B. rapa transformation, and into EHA101 for B. oleracea transformation. Hypocotyls of B. rapa and flower stems of B. oleracea were used as explants. Adventitious shoots from transformed cells were selected by kanamycin for B. rapa and by hygromycin for B. oleracea.

DNA Gel Blot Analysis

Total DNA was isolated from a leaf by DNeasy plant mini kit (Qiagen USA, Valencia, CA). Two micrograms of DNA were digested with appropriate restriction endonucleases, electrophoresed on a 1.0% (w/v) agarose gel, and transferred to a nylon membrane (Nytran N) (Schleicher and Schuell, Dassel, Germany). Hybridization was performed in 5× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate) containing 1.0% blocking reagent (Boehringer Mannheim, Mannheim, Germany), 0.1% sodium-N-lauroyl sarcosinate, and 0.02% SDS at 65°C. The membrane was washed twice in a solution consisting of 0.1× SSC and 0.1% SDS at 65°C for 20 min. A digoxigenin-labeled probe of cDNA was prepared by PCR with PCR DIG labeling mix (Boehringer Mannheim). DNA bands hybridized with the digoxigenin-labeled probe were detected following the supplier's instructions (Boehringer Mannheim).

RNA Gel Blot Analysis

Total RNA was isolated from anthers by ISOGEN (Nippongene, Tokyo, Japan). After denaturation in glyoxal, 12 μg RNA was subjected to electrophoresis on a 1% agarose gel in 10 mM sodium phosphate buffer, pH 7.0, and transferred to Nytran N. Hybridization was performed using a digoxigenin-labeled cDNA probe in 5× SSC containing 1.0% blocking reagent (Boehringer Mannheim), 0.1% sodium-N-lauroyl sarcosinate, and 0.02% SDS at 65°C. The membrane was washed twice in 0.1× SSC containing 0.1% SDS at 65°C for 20 min.

Pollination Test

On the day of anthesis, flowers were cut off from the plants and placed on an agar plate. After emasculation, the stigmas were covered with a layer of pollen grains. The pollinated flowers were kept at 20°C for 6 h. The pistils were immersed in 1 N NaOH for 1 h at 50°C, stained with aniline blue (0.1% aniline blue in 0.1 M K₃PO₄), and mounted in 50% glycerol. Pollen tubes were observed under an ultraviolet light fluorescence microscope. The evaluation of self-incompatibility was conducted using indices based on the number of pollen tubes penetrating stigma papilla cells. The indices are as follows: −−, completely incompatible, i.e., no or few germinating pollen grains observed on a stigma with no pollen tube penetrating a papilla cell; −, incompatible, i.e., >30 germinating pollen grains on a stigma and <5 pollen tubes penetrating papilla cells; +−, partially incompatible, i.e., >5 and <30 pollen tubes penetrating papilla cells; +, compatible, i.e., 30 to 100 pollen tubes penetrating papilla cells; ++, completely compatible, i.e., >100 pollen tubes penetrating papilla cells. Pollination tests were conducted using >12 flowers in each cross-combination, and the most frequent index in each cross-combination was represented.

Production of Chimeric SP11 Genes and Recombinant Proteins

Chimeric SP11 genes were constructed for a bioassay and the transformation of Brassica. Primers and methods used for the construction of the chimeric SP11 genes are shown in Supplemental Tables 1 and 2 online. For example, in the production of chimeric RsSP11-6(18), an RsSP11-6 cDNA clone was used as a template of 1st PCR and the PCR product amplified using the primer pair of 6-S-F and 6III-18R was mixed with the PCR product amplified using three primers of 6III-18F, 52V-2, and 18RS. The mixture was used as a template of 2nd PCR using a primer pair of 6-S-F and 18RS. Each PCR product of chimeric SP11 was cloned into the pGEM-T vector (Promega, Madison, WI), and a clone of each chimeric SP11 was selected by determining the nucleotide sequence.

The plasmids of 52(II-1), 52(IV-2), and 52(V-2) were digested with BamHI and PstI, and digested inserts were cloned into the pQE30 vector (Qiagen). The plasmids of the other clones were digested with BamHI and SalI, and the inserts were cloned into the pQE30 vector. The pQE30 DNA with each chimeric SP11 was introduced into E. coli M15. The synthesized protein of SP11, which was tagged with 6× His at the N-terminal, was purified using nickel-nitrilotriacetic acid agarose (Qiagen).

Bioassay of Recombinant SP11 Proteins

The stigma was treated with a 0.5 μL solution of 300 ng/μL of the recombinant SP11 protein. After being air-dried, the stigma was pollinated by the pollen grains of a compatible S haplotype, i.e., BrS-32, BrS-36, or BrS-46. The levels of incompatibility were rated with indices different from those used for the tests of the transgenic plants and the interspecific hybrids, because the incompatibility exhibited by this assay was weak. The indices of the number of pollen tubes penetrating the stigma are as follows: 1, <10; 2, 10 to 30; 3, 31 to 100; and 4, >100. Means and standard errors of the indices were calculated, and significance of the difference of the means between the treatments and the control experiment was analyzed by t test.