Intrahaplotype Polymorphism at the Brassica S Locus

Corresponding author: J. Mark Cock, Reproduction et Développement des Plantes, UMR 5667 CNRS-INRA-ENSL, Ecole Normale Supérieure de Lyon, 46 allée d'Italie, 69364 Lyon Cedex 07, France., mark.cock{at}ens-lyon.fr (E-mail)

Communicating editor: D. CHARLESWORTH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The S locus receptor kinase and the S locus glycoproteins are encoded by genes located at the S locus, which controls the self-incompatibility response in Brassica. In class II self-incompatibility haplotypes, S locus glycoproteins can be encoded by two different genes, SLGA and SLGB. In this study, we analyzed the sequences of these genes in several independently isolated plants, all of which carry the same S haplotype (S₂). Two groups of S₂ haplotypes could be distinguished depending on whether SRK was associated with SLGA or SLGB. Surprisingly, SRK alleles from the two groups could be distinguished at the sequence level, suggesting that recombination rarely occurs between haplotypes of the two groups. An analysis of the distribution of polymorphisms along the S domain of SRK showed that hypervariable domains I and II tend to be conserved within haplotypes but to be highly variable between haplotypes. This is consistent with these domains playing a role in the determination of haplotype specificity.

MOST flowering plant species are hermaphrodite and possess flowers in which male (stamen) and female (pistil) reproductive organs are in close proximity. However, outcrossing is favored in the majority of these species due to the presence of self-incompatibility (SI) systems that allow the pistil to recognize and reject self-pollen. In Brassica, SI is controlled by a single genetic locus (the S locus) and self-pollen is rejected on the stigma surface (reviewed in COCK 2000 ; NASRALLAH 2000 ).

One of the most striking features of the S locus is its extremely high level of polymorphism. For example, >80 different S locus alleles (or haplotypes) have been identified in Brassica oleracea (RUFFIO-CHABLE and GAUDE 2001 ). This high level of polymorphism is thought to be the result of balancing selection, which maintains large numbers of S haplotypes in Brassica populations over long periods of time (UYENOYAMA 1995 ). Another consequence of this unusual mode of evolution is that S haplotypes exhibit trans-specific evolution; they are often more similar to an S haplotype from another species in the Brassicaceae than to other S haplotypes from the same species (DWYER et al. 1991 ; SAKAMOTO et al. 1998 ).

Physical mapping and sequence analysis of different S haplotypes has revealed large-scale rearrangements that can extend to several hundred kilobases in this region of the genome (BOYES and NASRALLAH 1993 ; BOYES et al. 1997 ; CUI et al. 1999 ; SUZUKI et al. 1999 ). Several expressed genes are located within this highly polymorphic region and two of these genes have recently been shown to play key roles in the SI response. TAKASAKI et al. 2000 used a transgenic approach to show that SRK encodes the female component of the SI response. SRK encodes the S locus receptor kinase, a membrane-anchored stigma glycoprotein that shares structural similarity with animal receptor protein kinases (STEIN et al. 1991 ; DELORME et al. 1995 ; STEIN et al. 1996 ). SCR (or SP11) was shown to encode the male component of the self-incompatibility response (SCHOPFER et al. 1999 ; TAKAYAMA et al. 2000 ). SCR encodes the S locus cysteine-rich protein, a small secreted protein that is thought to be located in the pollen coating (STEPHENSON et al. 1997 ; SCHOPFER et al. 1999 ; TAKAYAMA et al. 2000 ).

While the functions of SRK and SCR seem to be relatively clear, that of a third S locus gene, SLG (S locus glycoprotein), is less well understood. SLG encodes a secreted glycoprotein that closely resembles the extracellular (or S) domain of SRK. The similarity between these two genes is often particularly marked within haplotypes, indicating concerted evolution (STEIN et al. 1991 ; KUSABA et al. 1997 ; NISHIO and KUSABA 2000 ). Initially, several pieces of evidence suggested that SLG functioned in the SI response (TORIYAMA et al. 1991 ; SHIBA et al. 1995 ). However, more recent evidence argues against a role for this protein in the recognition of self-pollen (GAUDE et al. 1995 ; CABRILLAC et al. 1999 ; OKAZAKI et al. 1999 ; NISHIO and KUSABA 2000 ) although SLG may play an auxiliary role by mediating maturation of the SRK protein (DIXIT et al. 2000 ). If this is the case, SLG is probably not the only protein that can carry out this function as self-incompatible plants that lack SLG have been identified (OKAZAKI et al. 1999 ; NISHIO and KUSABA 2000 ). It has been suggested that a truncated form of SRK, eSRK (GIRANTON et al. 1995 ), may substitute for SLG in these plants (DIXIT et al. 2000 ). TAKASAKI et al. 2000 showed that coexpression of SLG and SRK in transgenic plants resulted in a stronger SI response than expression of SRK alone, again supporting the hypothesis that SLG plays an auxiliary role. Another function that has been proposed for SLG is as one of the components on the stigma surface that mediate pollen adhesion (LUU et al. 1999 ). In the latter case, SLG would function independently of the SI system as pollen adhesion is not influenced by the SI response (LUU et al. 1997 ).

Both SLG and SRK are highly polymorphic. The alleles of these genes can be grouped into two distinct classes, class I and class II, and there is a correlation between these classes and observed SI phenotypes. SI in Brassica is sporophytic and a nonlinear hierarchy of dominance relationships has been shown to exist between S haplotypes (THOMPSON and TAYLOR 1966 ). Class I haplotypes tend to be dominant and to confer a strong SI response whereas class II haplotypes tend to be recessive to the class I haplotypes and to confer a weaker SI response. Two different SLG genes, SLGA and SLGB, have recently been shown to be associated with class II S haplotypes (CABRILLAC et al. 1999 ). Both SLGA and SLGB contain a single intron but SLGA differs from SLGB in that its second exon encodes a membrane-spanning domain, allowing the production of alternative transcripts that encode both secreted and membrane-anchored proteins (TANTIKANJANA et al. 1993 ; CABRILLAC et al. 1999 ).

Data from genetic crosses indicate that each S haplotype encodes an independent recognition system. Self-pollen rejection is initiated as a result of the interaction of male and female components encoded by the same S haplotype. In principle, therefore, comparison of different haplotypes should permit the identification of sequence motifs that determine haplotype specificity. This approach has been used profitably, for example, as an aid to functional analysis of polymorphic disease resistance gene loci (ELLIS et al. 1999 ). However, a drawback with applying this approach to self-incompatibility systems is the very high level of polymorphism at the S locus. For example, searches for regions that potentially determine haplotype specificity in the S domain of SRK have identified several extremely polymorphic hypervariable (HV) regions (NASRALLAH et al. 1987 ; KUSABA et al. 1997 ). However, the level of polymorphism throughout the rest of the S domain is also very high and it is possible that the HV regions are simply under a low level of selective constraint and that this is sufficient to explain the higher level of polymorphism in these regions (NASRALLAH 1997 ). The approach that we have taken here, to circumvent this problem, is to compare not only S haplotypes that confer different SI specificities ("between haplotype" comparisons) but also several independently isolated S haplotypes that confer the same SI specificity ("within haplotype" comparisons). Regions under relaxed selective constraint are expected to be more divergent (compared to the rest of the protein) in within-haplotype comparisons whereas regions involved in determining haplotype specificity should be conserved in genes that confer the same SI specificity. Sequence heterogeneity at S locus gene loci has been previously reported for the S₂ haplotype (KUSABA et al. 2000 ). Here we analyze an additional 10 S₂ haplotypes. Two groups of S₂ haplotype were defined, depending on the presence of either only SLGA or only SLGB. We show that, at the sequence level, hypervariable domains HV1 and HV2 are conserved in independent SRK₂ alleles but are divergent between haplotypes, indicating a role for these domains in determining S haplotype specificity.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant material and pollination analysis:
The seven S₂ B. oleracea plants analyzed in this study were from diverse origins and include varieties of kale, broccoli, and Brussels sprouts (Table 1). Incompatibility phenotypes were determined by self- and cross-pollinations using previously described procedures (RUFFIO-CHABLE and GAUDE 2001 ). Homozygous S₆ and S₈ tester lines were from the collection at the Genetic Resources Unit, Horticulture Research International (Wellesbourne, United Kingdom), which has recently been duplicated at INRA, Le Rheu (Rennes, France).

Immunological detection of SLG proteins in stigma extracts:
Protein extraction, separation of proteins by isoelectric focusing, electrotransfer onto nitrocellulose membranes, and immunodetection of antigen with antibodies were performed as described previously (GAUDE et al. 1991 , GAUDE et al. 1993 ). Polyclonal rabbit serum specific for class I SLG proteins and monoclonal antibody MAb 85-36-71, which recognizes class II SLG proteins, have been described previously (GAUDE et al. 1991 , GAUDE et al. 1993 ; GIRANTON et al. 1995 ).

DNA blot analysis:
DNA blot analysis was carried out as described (CABRILLAC et al. 1999 ). Probes corresponding to intron sequences of SLGA and SLGB (CABRILLAC et al. 1999 ) and to SLA (PASTUGLIA et al. 1997B ) have been described. The SRK probe was derived from the third intron of SRK₂ (J. M. COCK, unpublished data) by PCR amplification with SK42 (5'-GTATAAATAATGAAGGAATCACTATGAAAT-3') and SK40 (5'-GATCACTTATACAAAACCAACAGAGCAG-3').

PCR and PCR-restriction fragment length polymorphism segregation analysis:
SLGA and SLGB alleles were amplified from genomic DNA, using the PCR and oligonucleotides PS3 (see below) and SG17 (5'-TGTTCCGTCTGTCAAGTCCCACTGCTGCGG-3') for the progeny of plant i/j and PS4CM (5'-CGGAATATGGTATAAGAAAGTCTCCCA-3') and SG39 (5'- CTTTGCGTTTCAACACGTTGATTCA-3') for the progeny of plant k/l. SRK alleles were similarily amplified using the oligonucleotides PS5CM (5'-GGAATATGGTATAAAAAAGCCCCCTG-3') and SG2 (see below). PCR amplification conditions were 94° for 3 min followed by 30 cycles of denaturation at 94° for 40 sec, annealing at 51° for 40 sec, and extension at 72° for 1 min in a GeneAmp PCR system 9600 cycler (Perkin-Elmer, Shelton, CT). PCR products were digested with either TaqI or NdeI restriction enzymes (GIBCO BRL, Gaithersburg, MD) to detect a restriction fragment length polymorphism (RFLP).

cDNA cloning and DNA sequencing:
Total RNA (1 µg) from stigmas of each of the seven S₂ plants was reverse transcribed with Superscript II reverse transcriptase (GIBCO BRL), using oligonucleotide RA1 (FROHMAN et al. 1988 ) as a primer. SRK and SLG sequences were PCR amplified from the resulting cDNA or from genomic DNA extracted from the same plants, using different combinations of the following oligonucleotides:

• 5' oligonucleotides:

• PS3, 3'-ATGAAAGGGGTACAGAACAT-5' (NISHIO et al. 1996 )

• SDOM1, 3'-GWTGGTAYCTCGGRATRTGGTA-5'

• 3' oligonucleotides:

• SK30, 3'-TTCTCGCCCTCATAAACACAACAG-5'

• SK38, 3'-CTCCAACCTATGATTTTTCCAGT-5'

• SK66DC, 3'-CTCCTCCAAAAGCAGAACACGATAACACTC-5'

• SG2, 3'-GGCCTGCAGCAGCATTCAATCTGAC-5'

PCR products were amplified under the same conditions described above and were cloned into pGEM-T easy (Promega, Madison, WI) and sequenced by the dideoxynucleotide chain termination method of SANGER et al. 1977 on an automatic sequencer (Applied Biosystems, Foster City, CA). Sequence data were analyzed with Lasergene sequence analysis software (DNAstar, London).

Sequence analysis:
Multiple sequence alignments were constructed using the Megalign program (Lasergene; DNAstar). Neighbor-joining trees were constructed from these alignments using ClustalW software on the Pasteur Institute website (http://bioweb.pasteur.fr/#log; THOMPSON et al. 1994 ). Nonsynonymous substitutions were identified using DnaSP, version 3 (ROZAS and ROZAS 1999 ).

The nucleotide sequence data presented in this article have been submitted to the EMBL, GenBank, and DDBJ nucleotide sequence databases under accession nos. AJ306573, AJ306574, AJ306575, AJ306576, AJ306577, AJ306578, AJ306579, AJ306580, AJ306581, AJ306582, AJ306583, AJ306584, AJ306585, AJ306586, AJ306587, AJ306588, AJ306589, AJ306590, AJ306591.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of Brassica plants carrying S₂ haplotypes:
Diallel crosses were carried out to test the self-incompatibility phenotypes of seven putative S₂ B. oleracea plants from diverse origins (including kale, broccoli, and Brussels sprout lines; Table 1). Table 2 shows that a strong incompatible reaction was observed in the majority of cross-pollinations between the seven plants. In several crosses, however, either the SI reaction was weak or the cross was fully compatible. This was often the case, for example, when plant f was used as the female partner or when plant g/h was used as the male partner. This type of variability in the strength of the SI response has been observed previously, particularly with class II haplotypes. For example, RUFFIO-CHÂBLE et al. (1997) described similar levels of variability in the strength of the SI response between different Brassica lines homozygous for the S₁₅ haplotype. This phenomenon is thought to be due to the action of suppresser and modifier genes that are unlinked to the S locus but that modulate the SI response (HINATA et al. 1994 ). In general, however, each of the seven lines exhibited a strong SI response with several of the other members of the group, indicating that they all carried a common S specificity. All of the plants tested, with the exception of plant f, were self-incompatible.

Immunodetection of SLG in stigma extracts of the S₂ plants:
Immunological detection of SLG in stigma extracts separated by isoelectric focusing is a highly discriminative method for typing S haplotypes (RUFFIO-CHABLE et al. 1997 , RUFFIO-CHABLE et al. 1999 ; OKAZAKI et al. 1999 ). Fig 1 shows that two bands, and ß, characteristic of the S₂ haplotype (GAUDE et al. 1995 ), were specifically recognized by the monoclonal antibody (mAb) 85-36-71 in all of the S₂ plants except g/h and i/j. No additional S-locus-specific proteins were detected with either mAb 85-36-71 or with a polyclonal antibody that specifically recognizes class I SLG proteins (DELORME et al. 1995 ; GAUDE et al. 1995 ) in any of the seven S₂ plants.

View larger version (63K): In this window In a new window Download PPT slide   Figure 1. Immunodetection of S-locus-encoded glycoproteins in stigma extracts from the seven S2 plants. Soluble stigma proteins were separated by isoelectric focusing gel electrophoresis. Immunodetection was performed with monoclonal antibody mAb 85-36-7. The - and ß-polypeptides detected by the antibody are indicated.

Variation in the complement of SLG genes in the various S₂ plants:
The immunodetection experiments indicated heterogeneity among the S₂ plants. To investigate this further, we probed genomic DNA blots with probes that specifically detect four S locus genes, SRK, SLGA, SLGB, and SLA (Fig 2). Plants g/h and i/j showed markedly different patterns to the five other S₂ plants. Two bands (2.00 and 1.25 kbp) hybridized to the SRK probe in DNA of both of these two plants and neither of these bands corresponded to that detected in the other five S₂ plants with the same probe (a single fragment of 1.27 kbp). A 0.85-kbp fragment that hybridized to the SLGB probe was detected in DNA of plants g/h and i/j but no hybridization was detected to the SLGA probe. Conversely, a 1.10-kbp SLGA fragment was detected in DNA of the five other S₂ plants but, in most cases, no SLGB sequence was detected. The only exception was DNA from plant k/l, which contained both the 1.10-kbp SLGA and the 0.85-kbp SLGB bands, indicating the presence of both genes. The SLA probe hybridized to the same samples as the SLGA probe, as expected (SLA has been shown to be very closely linked to SLGA; BOYES and NASRALLAH 1995 ; PASTUGLIA et al. 1997B ). These results were confirmed by probing genomic DNA digested with BamHI (data not shown).

View larger version (41K): In this window In a new window Download PPT slide   Figure 2. Detection of SRK, SLGA, SLGB, and SLA alleles in B. oleracea plants carrying S2, S3, and S15 haplotypes. A gel blot of genomic DNA digested with EarI was sequentially hybridized with the probes indicated beneath each panel. The positions of molecular length markers are shown at left in kilobases (kbp).

The presence of two bands that hybridized to the SRK probe in DNA from plants g/h and i/j indicated that the plants analyzed might have been heterozygous at the S locus. To investigate this further we analyzed 18 progeny of the plant i/j. Fig 3A (top) shows that the two sequences detected by the SRK probe segregated with a 4:10:4 ratio (homozygous i:heterozygous i/j:homozygous j) that closely approximates a Mendelian 1:2:1 ratio. When BamHI-digested DNA was probed with the SLGB probe, the pattern obtained was consistent with the presence of two polymorphic allelic sequences. However, an additional, nonpolymorphic fragment was also detected in this experiment (Fig 3A, middle). An alternative approach, based on restriction digestion of PCR products, was therefore used to specifically amplify the SLGB alleles and to detect a polymorphism that was predicted from the sequences of the two SLGB alleles (Fig 3A, bottom). These analyses demonstrated that plant i/j was heterozygous at both the SRK and SLGB loci and allowed the identification of the alleles associated with each haplotype.

View larger version (35K): In this window In a new window Download PPT slide   Figure 3. Linkage analysis of the SLG and SRK alleles in plants i/j and k/l. (a) Segregation of SRK and SLGB alleles in self-progeny of plant i/j. DNA gel blot analysis was performed on 5 µg/lane of EarI- (top) or BamHI-digested (middle) genomic DNA of plant i/j and 18 self-progeny using SRK (top) and SLGB (middle) probes. The bottom shows a PCR-RFLP analysis of the segregation of the two SLGB alleles. SLGB sequences were specifically amplified with oligonucleotides PS3 and SG17 and digested with TaqI to reveal a restriction fragment length polymorphism predicted from sequence analysis of the two SLGB alleles. Bands corresponding to alleles of either SRK or SLGB from the S2i and S2j haplotypes are indicated by either a solid circle or an open circle, respectively. The S2 haplotype(s) (S2i or S2j) carried by each of the progeny is (are) indicated below each lane. The positions of molecular length markers are shown at left in kilobases (kbp) or base pairs (bp). (b) Analysis of the segregation of SLGA, SLGB, and SRK in self-progeny of plant k/l. SLGA, SLGB, and SRK sequences were specifically amplified with oligonucleotides PS3 and SG38, PS4CM and SG39, or PS5CM and SG2, respectively. The S2 haplotype(s) (S2k and/or S2l) carried by each of the progeny is (are) indicated below each lane. The positions of molecular length markers are shown at left in base pairs (bp). (c) Schematic representation showing the complement of genes present in S2a, in S2b, and in the 10 S2 haplotypes carried by the seven plants analyzed in this study (including three plants that were heterozygous at the S locus).

DNA gel blots indicated that plant k/l carried alleles of SRK, SLGA, and SLGB, suggesting that it may have been heterozygous at the S locus (Fig 2). To investigate this further, we screened genomic DNA from 10 progeny of plant k/l for the presence of SLGA and SLGB, using PCR and gene-specific oligonucleotide pairs. SLGA and SLGB segregated independently in this population, confirming that plant k was heterozygous (Fig 3B).

The conclusions that can be drawn from the data presented in Fig 2 and Fig 3 are summarized in Fig 3C. Fig 3C also includes the two S₂ haplotypes characterized by TANTIKANJANA et al. 1993 and KUSABA et al. 2000 , which we have designated as S_2a and S_2b, respectively. The figure shows that two classes of S₂ haplotype could be clearly distinguished. In the first, SRK is associated with SLGA (haplotypes a, c, d, e, f, and k) and in the second SRK is associated with SLGB (haplotypes b, g, h, i, j, and l). None of the plants analyzed carried an S₂ haplotype consisting of SRK, SLGA, and SLGB as has been described in the S₁₅ haplotype.

Cloning and sequence analysis of the S locus genes present in the B. oleracea S₂ plants:
SRK, SLGA, and SLGB sequences were amplified from the seven S₂ plants by reverse transcription (RT)-PCR or by PCR amplification of genomic sequences, using different combinations of six oligonucleotides designed to amplify class II S locus genes. We isolated clones corresponding to all of the SRK, SLGA, and SLGB alleles detected by DNA gel blotting except SRK_2g and SRK_2h. However, note that, based on comparison of SLGB sequences (see below), these two SRK alleles are likely to be highly similar to SRK_2i and SRK_2j. The cloned fragments shared a region corresponding to almost the entire S domain (represented by a shaded bar in Fig 5A). To demonstrate that the sequences that we designated SRK were not derived from SLGA or SLGB genes, we amplified, by RT-PCR, a fragment that included both S domain and kinase domain sequences (data not shown).

View larger version (21K): In this window In a new window Download PPT slide   Figure 4. Sequence comparison and phylogenetic relationships of the S domains of class II SRK, SLGA, and SLGB alleles. (a) Similarity between alleles of SRK, SLGA, and SLGB from 12 S2 haplotypes. Alleles corresponding to each gene are grouped separately and the range of similarity between genes at the nucleotide (nuc.) and amino acid (aa.) levels is indicated. For each gene, alleles were subgrouped on the basis of similarity and the numbers of synonymous/nonsynonymous substitutions observed within each group and subgroup are indicated. All comparisons were carried out using a region of 1299 bp corresponding to the majority of the S domain (see Fig 5A where this region is indicated by a shaded bar). (b) Neighbor-joining tree based on class II SLG and SRK S domain sequences. SRK2a and SLG2a (TANTIKANJANA et al. 1993 ), SRK2b and SLG2b (KUSABA et al. 2000 ), SLG5 (SCUTT and CROY 1992 ), SRK5, SRK15, SLGA15, and SLGB15 (CABRILLAC et al. 1999 ) have been described previously. SFR2 (PASTUGLIA et al. 1997A ) was included as an outgroup. Based on sequence similarity and gene structure, SLG5 is an allele of SLGB. The tree was constructed using the PAM250 amino acid substitution matrix. The numbers next to the branches are bootstrap values expressed as percentage confidence level and based on 2000 repeats. The scale indicates amino acid substitutions/site. Note that this tree is presented as a means of representing the relationships between the sequences shown and probably does not correctly describe the evolutionary events that generated these sequences. The phylogeny of these sequences has almost certainly been obscured in many cases by events that tend to result in concerted evolution within haplotypes.

View larger version (52K): In this window In a new window Download PPT slide   Figure 5. Distribution of nucleotide and amino acid polymorphisms in SRK, SLGA, and SLGB. (a) Schematic representations showing the positions of observed nucleotide and amino acid substitutions along the S domains of SRK, SLGA, and SLGB following intra- and interhaplotype comparisons. Comparisons were carried out for the 10 SRK2 alleles (SRK2), the 6 SLGA2 alleles (SLGA2), the 6 SLGB2 alleles (SLGB2), and between SRK2 and either SRK5 or SRK15 (SRK class II). Vertical bars and vertical bars with open circles indicate the positions of nucleotide and amino acid residues that exhibited polymorphism. In the bottom only amino acid substitutions are shown, either between SRK2 and SRK5 (above the rectangle) or between SRK2 and SRK15 (below the rectangle). Open rectangles represent signal peptides plus S domains; the region that was compared is indicated by a shaded bar. The positions of the DR (deletable region) and the three hypervariable regions (HV1, HV2, and HV3) are indicated beneath the S domains by solid bars. The hatched portions of the S domain represent the two variable regions defined by NASRALLAH et al. 1987 . (b) Frequency of observed amino acid substitutions in different regions of the SRK S domain in either inter- or intrahaplotype comparisons. The mean frequencies of observed amino acid substitutions, expressed as percentage substitutions per amino acid residue, are indicated by lightly shaded and darkly shaded bars for comparisons within the 10 SRK2 sequences and for comparisons between different class II SRK sequences (SRK2/SRK5, SRK2/SRK15, and SRK2/SRK29), respectively. SRK2, SRK5, and SRK15 are from B. oleracea, and SRK29 is from B. rapa. DR, deletable region; HV1, HV2, and HV3, hypervariable regions; rest, rest of the region indicated by the shaded bar in Fig 5A, excluding the DR, HV1, HV2, and HV3 regions.

To estimate the level of sequence polymorphism at each S locus gene locus we compared, at both the nucleotide and deduced amino acid sequence levels, the S domains of the SRK, SLGA, and SLGB alleles from the seven S₂ plants and from the two S₂ haplotypes that have been described previously (TANTIKANJANA et al. 1993 ; KUSABA et al. 2000 ). Fig 4A shows that polymorphisms were detected in SRK, SLGA, and SLGB although SLGA was significantly less polymorphic than SLGB and SRK. Despite this polymorphism, alleles of the three genes could be clearly distinguished at the sequence level. SLGA₂ was less closely related to SRK₂ and SLGB₂ than these latter genes were to each other.

Fig 4B shows a phylogenetic tree that was constructed from an alignment of the deduced amino acid sequences of the S domains of the S₂ sequences with those of other class II S locus genes. The allelic forms of SRK₂ clustered in one part of the tree together with the two previously reported sequences, SRK_2a and SRK_2b. SRK sequences corresponding to other class II alleles fell outside this cluster. Taken together with the data from genetic crosses (Table 2), this confirmed that the seven plants analyzed in this study carried the S₂ haplotype.

Surprisingly, two phylogenetically distinct groups of SRK₂ sequence could be distinguished (Fig 4B). These two groups correlated exactly with the presence of either SLGA or SLGB in each haplotype (compare Fig 4B with Fig 3C).

Sequence analysis of the S₂ plants also provided a possible explanation as to why no SLG proteins were detected in stigma extracts of plants g/h and i/j (Fig 1). The epitope recognized by the mAb 85-36-71 (IYVNTLSSSE) is conserved at the predicted N terminus of the mature SLGA protein, but a variant sequence (TYVNTMSSSE) is predicted to occur in the SLGB protein encoded by the i, j, g, and h haplotypes, and no SLGA gene was detected in these plants. From a practical point of view, heterogeneity concerning the presence or absence of SLG genes in haplotypes that confer the same SI phenotype could present problems for S haplotype typing methods that are based on detection of SLG at either the DNA or protein level (NISHIO et al. 1996 ; RUFFIO-CHABLE et al. 1997 ; SAKAMOTO et al. 1998 ; OKAZAKI et al. 1999 ; RUFFIO-CHABLE et al. 1999 ). In this respect, methods based on determining the SRK genotype are likely to be more reliable (NISHIO et al. 1997 ).

Sequence polymorphism in the HV domains of SRK:
Recognition of self-pollen in self-incompatible Brassica is thought to involve haplotype-specific interaction between SCR on the male side and SRK on the female side. It has, however, been difficult to identify the regions of these two proteins that determine haplotype specificity because of their extremely polymorphic nature. For example, although hypervariable regions have been identified in the extracellular domain of SRK [HV1, HV2, HV3, and a deletable region (DR); NASRALLAH et al. 1987; KUSABA et al. 1997], it is not known whether these domains determine haplotype specificity or whether they are merely regions that are under a low level of selective constraint and therefore are free to diverge to a greater degree. One way of addressing this question is to carry out comparisons both between alleles of the same S haplotype (where domains determining S haplotype specificity should be conserved) and between alleles from different S haplotypes (where domains determining S haplotype specificity should be polymorphic).

When the 10 SRK₂ alleles were compared and polymorphic residues were positioned with respect to the S domain, we noted that none of the substitutions affected residues of the DR, HV1, or HV2 (Fig 5A, top). In contrast, when SRK sequences from different class II haplotypes were compared, polymorphic residues appeared to be clustered in these regions (Fig 5A, bottom). To further investigate this observation, we calculated the percentage of substitutions observed in each region on a per residue basis so that substitution rates could be compared despite the differences in length of each region (Fig 5B). This analysis suggested that the pattern of substitutions observed when the different SRK₂ sequences were compared differed from that observed when SRK sequences were compared between haplotypes. On average, when SRK sequences of the three class II haplotypes were compared, a substitution was more than three times more likely to be found at a residue within the DR, HV1, or HV2 than at a residue elsewhere in the S domain. In contrast, when the SRK₂ sequences were compared, no substitutions were observed in these regions. We used a modified version of the Hudson-Kreitman-Aguadé (HKA) test (HUDSON et al. 1987 ) to determine whether there was a significant difference in the distribution of nonsynonymous substitutions between the DR/HV1/HV2 and the rest of the region analyzed in a comparison of SRK₂ alleles vs. a comparison between the SRK₂ alleles and SRK₁₅ (i.e., comparing within-haplotype and between-haplotype comparisons). We did not include HV3 with the other HV regions in this analysis because it is divergent between some SRK₂ alleles (Fig 5B) and yet conserved in several SRK alleles corresponding to different S haplotypes (see below), suggesting that it has not evolved in the same manner as the other HV domains. Table 3 shows that the HKA test gave a result that was very close to being significant (P = 0.07). This is interesting because the HKA test provides a conservative estimate of significance, particularly in this case, where the number of observed nonsynonymous substitutions is small.

Although the total number of polymorphic sites identified when SLGB₂ sequences were compared was similar to that observed for the SRK₂ sequences, a much higher proportion was located in the DR, HV1, and HV2 (7 out of 14 substitutions, Fig 5A). A Fisher's exact test showed that this difference in the distribution of polymorphisms in SLGB and SRK was highly significant (P = 0.006; Table 4), indicating that these domains are not under the same functional constraints in SLGB as in SRK. These data are consistent with several reports that indicate that SLG does not play a role in haplotype-specific recognition of self-pollen (GAUDE et al. 1995 ; CABRILLAC et al. 1999 ; NISHIO and KUSABA 2000 ).

The difference in the distribution of polymorphisms in SRK and SLGB also supports the hypothesis that the DR, HV1, and HV2 domains have been conserved in the SRK₂ alleles. To investigate this difference further, we carried out a second HKA test, again to compare distributions of nonsynonymous substitutions between the DR/HV1/HV2 and the rest of the region analyzed, but this time comparing the "within SRK₂" data to a comparison between SRK₂ and SLGB₂. Table 5 shows that a significant difference was detected in this comparison (P = 0.03), suggesting that there has been either selection against changes in the DR, HV1, and HV2 of SRK₂ or that there has been positive selection for changes in these regions since the divergence of SRK and SLGB. It will, however, be important to carry out similar analyses on other S haplotypes to ensure that the observed difference is not due to gene conversion having, by chance, preferentially homogenized the non-HV regions in the S₂ haplotype.

To date, intrahaplotype polymorphism has been described in only one other Brassica haplotype, S₁₃ (KUSABA et al. 2000 ). Surprisingly, and in contrast to our findings with the S₂ haplotype, when the S domains of SRK₁₃ and SRK_13-b were compared, one of the three amino acid substitutions was located in HV2. However, note that, if HV1 and HV2 are involved in determining S haplotype specificity, it is possible, and even likely, that only a subset of residues within the HV domains is involved in this function. Six of the seven differences between the predicted mature SLG₁₃ and SLG_13-b proteins were found within HV1, supporting our observation that substitutions are clustered in the HV domains in SLG proteins from the same haplotype.

To further investigate the role of the HV domains in SRK function, we compared the HV domains of all the SRK sequences available in the databases (Fig 6). The sequence of the HV3 region was not conserved in the SRK₂ sequences and, moreover, the HV3 domains of four of the SRK₂ sequences were identical to those of SRK₅ and SRK₄₅ from B. oleracea and B. campestris, respectively. In contrast, HV1 and HV2 were both 100% conserved in the 10 SRK₂ sequences but no two SRK sequences from different haplotypes exhibited the same sequence in these regions. Taken together with the analyses described above, these data argue against DR, HV1, and HV2 merely being regions of the SRK protein that are under a low level of functional constraint. Rather, they indicate that substitutions in these regions are associated with the acquisition of new SI specificities.

View larger version (63K): In this window In a new window Download PPT slide   Figure 6. Sequence comparison of the DR and HV regions from the SRK2 alleles with other class II and class I SRK alleles. Solid bars indicate the positions of each region. Numbers indicate the position of the last residue of each block in relation to the N-terminal end of the predicted unprocessed polypeptide. Residues conserved in a majority of sequences are boxed. *, SRK sequences with identical HV3 regions; • and , two pairs of identical DR sequences; Bo, B. oleracea; Br, B. rapa; Bn, B. napus; DR, deletable region; HV1, HV2, and HV3, hypervariable regions.

The DR was conserved among the SRK₂ sequences but two pairs of class I SRKs shared identical DR domains (BrSRK₈/BoSRK₂₃ and BoSRK₃/BnSRK_A10, where Bo, Br, and Bn refer to the species of origin, B. oleracea, B. rapa, and B. napus, respectively; Fig 6). Therefore, this domain does not determine S haplotype specificity alone although it is possible that it may contribute to specificity by acting in combination with another polymorphic region of the protein such as HV1 or HV2.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

SLG and S locus structure:
In a previous study, we reported the existence of two different SLG genes (SLGA and SLGB) in class II S haplotypes (CABRILLAC et al. 1999). Both SLG genes were present in an S₁₅ haplotype but either only SLGA or only SLGB was detected in an S₅ and in an S₂ haplotype, respectively. More recently, KUSABA et al. 2000 described a novel form of the S₂ haplotype, S_2b, which exhibited both structural and sequence differences compared to the S₂ haplotype originally described by TANTIKANJANA et al. 1993 . In this study, we analyzed the sequences of SLG and SRK genes from 10 additional S₂ haplotypes from diverse genetic backgrounds. On the basis of sequence similarities, the 10 S₂ haplotypes could be divided into two groups that resembled S_2a on one hand and S_2b on the other. In the first group, SRK was associated with SLGA, and SLGB was absent, while in the second group SRK was associated with SLGB, and SLGA was absent.

Taken together, these results indicate that SRK, SLGA, and SLGB were already associated at the S locus in an ancestor of the class II haplotypes. SLGA or SLGB would then have been lost from the S locus in some haplotypes, these events occurring, at least in part, after the divergence of the different class II S haplotype specificities. An alternative scenario has been proposed by KUSABA et al. 2000 . They suggest that SLGA was originally acquired by the S₂ haplotype and then subsequently transmitted to the S₁₅ haplotype. This latter model is based on the fact that SLGB is commonly found in class II haplotypes of both B. oleracea and B. rapa (SCUTT and CROY 1992 ; HATAKEYAMA et al. 1998 ) whereas SLGA has been found only in the S₂ and S₁₅ haplotypes so far. Note, however, that SRK₁₅ is more similar to SLGA₁₅ than to SLGB₁₅ whereas the opposite is true for SRK₂, suggesting that, if SLGA was acquired after divergence of the class II haplotypes, then this probably occurred in the S₁₅ haplotype. A more detailed analysis of independent isolates of additional class II haplotypes such as S₅ should help to distinguish between these different models.

The S locus region has been characterized in detail in several class I S haplotypes and there is currently no evidence for the presence of a second SLG gene in this class of haplotype (BOYES et al. 1997 ; CUI et al. 1999 ; SUZUKI et al. 1999 ). It seems likely, therefore, that the acquisition of a second SLG gene by the class II haplotypes occurred after their divergence from class I haplotypes.

It is difficult to determine precisely how SRK, SLGA, and SLGB originated due to the highly polymorphic nature of these three genes. Comparison of their introns indicates a complex series of recombination and mutation events that tend to obscure their evolutionary history (CABRILLAC et al. 1999 ). It seems probable, however, that SLGA arose as a result of a duplication of SRK (TANTIKANJANA et al. 1993 ). SLGB may have originated in a similar manner, although it is also possible that it was recruited to the S locus from another part of the genome. Analogous translocation events of this type have been observed between distant disease resistance gene clusters in tomato (PARNISKE and JONES 1999 ).

Comparison of the S₂ haplotypes indicated the existence of two distinct phylogenetic groups of SRK₂ alleles and these two groups correlated with the presence of either SLGA₂ or SLGB₂, respectively. The fact that SRK alleles from each group can be clearly distinguished at the sequence level suggests that recombination between these two types of S₂ haplotype has been suppressed or has occurred only rarely. An alternative possibility is that these two groups of S₂ alleles evolved independently in geographically isolated populations and have been brought together only recently. Note, however, that we found no correlation between the presence of SLGA or SLGB and the origin of the plants that were analyzed.

Hypervariable domains and haplotype specificity:
In both sporophytic and gametophytic multiallelic SI systems, genes that encode female components of the SI system have been shown to possess regions of extreme sequence polymorphism, the HV domains. In the Solanaceae, S-RNAse HV domain swapping experiments have provided evidence that these domains are responsible for the difference in specificity between closely related alleles (MATTON et al. 1997 ). However, other regions of these proteins also seem to be important because exchanges of HV domains between more distantly related alleles often result in loss of activity rather than acquisition of new allelic specificities (KAO and MCCUBBIN 1996 ; ZUREK et al. 1997 ).

By analogy to the SI systems of the Solanaceae, and bearing in mind the unusually high number of functional S haplotypes in Brassica, it seems likely that the highly polymorphic HV regions of SRK are also involved in determining S haplotype specificity. However, in the absence of a direct experimental demonstration of the role of these domains, it has been difficult to distinguish between this possibility and a second hypothesis in which the HV regions are under a low level of selective constraint and therefore free to diverge to a greater degree (NASRALLAH 1997 ). In this study we have addressed this problem by comparing the level of polymorphism in these regions either within or between haplotypes. These comparisons show that HV1 and HV2 tend to be conserved in SRK molecules that confer the same SI specificity but are diverged in SRK molecules that confer different SI specificities, indicating a role for HV1 and HV2 in determining haplotype specificity.

In the Solanaceae, comparisons of the rates of nonsynonymous and synonymous substitution in the hypervariable region of S-RNases indicated that these regions have been subjected to positive selection (CLARK and KAO 1991 ). We carried out a similar analysis of HV1 and HV2 in the class II SRK genes but did not find any direct evidence of positive selection acting in these regions (data not shown). Similar results were obtained by AWADALLA and CHARLESWORTH 1999 when they analyzed the HV regions of class I SRK genes. However, these authors noted that, while there was no strong evidence of positive selection acting on the HV regions (K_a/K_s ratios were close to 1), K_a/K_s ratios for these regions were significantly higher than for the rest of the protein. They suggested that evidence of positive selection may be obscured by the high level of substitution in these regions, particularly if adaptive substitutions can occur at only a small subset of the amino acid positions.

The role of SLG in S locus function and evolution:
There is accumulating evidence that SLG is not involved in the haplotype-specific recognition of self-pollen (GAUDE et al. 1995 ; CABRILLAC et al. 1999 ; OKAZAKI et al. 1999 ; NISHIO and KUSABA 2000 ). If this is the case, then it is necessary to explain why SLG and SRK alleles are often highly similar within the same S haplotype (STEIN et al. 1991 ; GORING et al. 1993 ; DELORME et al. 1995 ; KUSABA et al. 1997 ). We have suggested that this concerted evolution of SLG and SRK is due to their close proximity in the genome, which tends to favor gene conversion events (CABRILLAC et al. 1999 ). There is convincing evidence that gene conversion events occur between SLG and SRK, particularly within haplotypes (GORING et al. 1993 ; WATANABE et al. 1994 ; SUZUKI et al. 1997 ; CABRILLAC et al. 1999 ). The DR, HV1, and HV2 regions were highly conserved in the SRK₂ sequences analyzed in this study but this was not the case for SLGB₂, where there was a tendency for substitutions to accumulate in these regions. We suggest that the capacity of SLG to accumulate substitutions in its HV domains may be an important factor in the generation of new S haplotype specificities. Gene conversion events would provide a means of transferring novel HV domain sequences from SLG to SRK and this could accelerate the evolution of novel SI specificities on the female side. Two other factors may be important for the functionality of such a system. The first is the overall similarity between SLG and the S domain of SRK, particularly within haplotypes, which is presumably the result of (reciprocal) sequence homogenization. The second is the maintenance of SLG as a functional protein due to its role in pollen adhesion and/or SRK maturation. In this respect, it is interesting that SLG binds PCPA1, a component of the pollen coat that is related to SCR at the sequence level (DOUGHTY et al. 1998 ).

	ACKNOWLEDGMENTS

We thank Fabienne Deguerry and Anne-Marie Thierry for technical assistance, and Rosemary MacClenaghan, Graham King, Angela Pinnegar, and Dave Astley from HRI Wellesbourne for providing Brassica lines. This work was funded by the Institut National de la Recherche Agronomique, the Centre National de la Recherche Scientifique, and the Ecole Normale Supérieure de Lyon.

Manuscript received May 3, 2001; Accepted for publication July 16, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AWADALLA, P. and D. CHARLESWORTH, 1999  Recombination and selection at Brassica self-incompatibility loci. Genetics 152:413-425[Abstract/Free Full Text].

BOYES, D. C. and J. B. NASRALLAH, 1993  Physical linkage of the SLG and SRK genes at the self-incompatibility locus of Brassica oleracea. Mol. Gen. Genet. 236:369-373[Medline].

BOYES, D. C. and J. B. NASRALLAH, 1995  An anther-specific gene encoded by an S locus haplotype of Brassica produces complementary and differentially regulated transcripts. Plant Cell 7:1283-1294[Abstract].

BOYES, D. C., M. E. NASRALLAH, J. VREBALOV, and J. B. NASRALLAH, 1997  The self-incompatibility (S) haplotypes of Brassica contain highly divergent and rearranged sequences of ancient origin. Plant Cell 9:237-247[Abstract].

CABRILLAC, D., V. DELORME, J. GARIN, V. RUFFIO-CHÂBLE, and J. L. GIRANTON et al., 1999  The S₁₅ self-incompatibility haplotype in Brassica oleracea includes three S gene family members expressed in stigmas. Plant Cell 11:971-986[Abstract/Free Full Text].

CLARK, A. G. and T.-H. KAO, 1991  Excess nonsynonymous substitution of shared polymorphic sites among self-incompatibility alleles of Solanaceae. Proc. Natl. Acad. Sci. USA 88:9823-9827[Abstract/Free Full Text].

COCK, J. M., 2000 A receptor kinase and the self-incompatibility response in Brassica, pp. 270–298 in Advances in Botanical Research, Thematic Vol. 32: Plant Protein Kinases, edited by M. KREIS and J. C. WALKER. Academic Press, London.

CUI, Y., N. BRUGIERE, L. JACKMAN, Y. M. BI, and S. J. ROTHSTEIN, 1999  Structural and transcriptional comparative analysis of the S locus regions in two self-incompatible Brassica napus lines. Plant Cell 11:2217-2231[Abstract/Free Full Text].

DELORME, V., J. L. GIRANTON, Y. HATZFELD, A. FRIRY, and P. HEIZMANN et al., 1995  Characterization of the S locus genes, SLG and SRK, of the Brassica S₃ haplotype: identification of a membrane-localized protein encoded by the S locus receptor kinase gene. Plant J. 7:429-440[Medline].

DIXIT, R., M. E. NASRALLAH, and J. B. NASRALLAH, 2000  Post-transcriptional maturation of the S receptor kinase of Brassica correlates with co-expression of the S-locus glycoprotein in the stigmas of two Brassica strains and in transgenic tobacco plants. Plant Physiol. 124:297-312[Abstract/Free Full Text].

DOUGHTY, J., S. DIXON, S. J. HISCOCK, A. C. WILLIS, and I. A. PARKIN et al., 1998  PCP-A1, a defensin-like Brassica pollen coat protein that binds the S locus glycoprotein, is the product of gametophytic gene expression. Plant Cell 10:1333-1347[Abstract/Free Full Text].

DWYER, K. G., M. A. BALENT, J. B. NASRALLAH, and M. E. NASRALLAH, 1991  DNA sequences of self-incompatibility genes from Brassica campestris and B. oleracea: polymorphism predating speciation. Plant Mol. Biol. 16:481-486[Medline].

ELLIS, J. G., G. J. LAWRENCE, J. E. LUCK, and P. N. DODDS, 1999  Identification of regions in alleles of the flax rust resistance gene L that determine differences in gene-for-gene specificity. Plant Cell 11:495-506[Abstract/Free Full Text].

FROHMAN, M. A., M. K. DUSH, and G. R. MARTIN, 1988  Rapid production of full-length cDNAs from rare transcripts: amplification using a single gene-specific oligonucleotide primer. Proc. Natl. Acad. Sci. USA 85:8998-9002[Abstract/Free Full Text].

GAUDE, T., L. DENOROY, and C. DUMAS, 1991  Use of a fast protein electrophoretic purification procedure for N-terminal sequence analysis to identify S-locus related proteins in stigmas of Brassica oleracea. Electrophoresis 12:646-653[Medline].

GAUDE, T., A. FRIRY, P. HEIZMANN, C. MARIAC, and M. ROUGIER et al., 1993  Expression of a self-incompatibility gene in a self-compatible line of Brassica oleracea. Plant Cell 5:75-86[Abstract/Free Full Text].

GAUDE, T., M. ROUGIER, P. HEIZMANN, D. J. OCKENDON, and C. DUMAS, 1995  Expression level of the SLG gene is not correlated with the self-incompatibility phenotype in the class II S haplotypes of Brassica oleracea. Plant Mol. Biol. 27:1003-1014[Medline].

GIRANTON, J. L., M. J. ARIZA, C. DUMAS, J. M. COCK, and T. GAUDE, 1995  The S locus receptor kinase gene encodes a soluble glycoprotein corresponding to the SRK extracellular domain in Brassica oleracea. Plant J. 8:827-834[Medline].

GORING, D. R., T. L. GLAVIN, U. SCHAFER, and S. J. ROTHSTEIN, 1993  An S receptor kinase gene in self-compatible Brassica napus has a 1-bp deletion. Plant Cell 5:531-539[Abstract].

HATAKEYAMA, K., T. TAKASAKI, M. WATANABE, and K. HINATA, 1998  Molecular characterization of S locus genes, SLG and SRK, in a pollen-recessive self-incompatibility haplotype of Brassica rapa L. Genetics 149:1587-1597[Abstract/Free Full Text].

HINATA, K., A. ISOGAI and K. ISUZUGAWA, 1994 Manipulation of Sporophytic Self-Incompatibility in Plant Breeding. Kluwer Academic Publishers, Dordrecht, The Netherlands.

HUDSON, R. R., M. KREITMAN, and M. AGUADE, 1987  A test of neutral molecular evolution based on nucleotide data. Genetics 116:153-159[Abstract/Free Full Text].

KAO, T. H. and A. G. MCCUBBIN, 1996  How flowering plants discriminate between self and non-self pollen to prevent inbreeding. Proc. Natl. Acad. Sci. USA 93:12059-12065[Abstract/Free Full Text].

KUSABA, M., T. NISHIO, Y. SATTA, K. HINATA, and D. OCKENDON, 1997  Striking sequence similarity in inter- and intra-specific comparisons of class I SLG alleles from Brassica oleracea and Brassica campestris: implications for the evolution and recognition mechanism. Proc. Natl. Acad. Sci. USA 94:7673-7678[Abstract/Free Full Text].

KUSABA, M., M. MATSUSHITA, K. OKAZAKI, Y. SATTA, and T. NISHIO, 2000  Sequence and structural diversity of the S locus genes from different lines with the same self-recognition specificities in Brassica oleracea. Genetics 154:413-420[Abstract/Free Full Text].

LUU, D.-T., P. HEIZMANN, and C. DUMAS, 1997  Pollen-stigma adhesion in kale is not dependant on the self(in)compatibility genotype. Plant Physiol. 115:1221-1230[Abstract].

LUU, D. T., D. MARTY-MAZARS, M. TRICK, C. DUMAS, and P. HEIZMANN, 1999  Pollen-stigma adhesion in Brassica spp involves SLG and SLR1 glycoproteins. Plant Cell 11:251-262[Abstract/Free Full Text].

MATTON, D. P., O. MAES, G. LAUBLIN, Q. XIKE, and C. BERTRAND et al., 1997  Hypervariable domains of self-incompatibility RNAses mediate allele-specific pollen recognition. Plant Cell 9:1757-1766[Abstract].

NASRALLAH, J. B., 1997  Evolution of the Brassica self-incompatibility locus: a look into S-locus gene polymorphisms. Proc. Natl. Acad. Sci. USA 94:7673-7678.

NASRALLAH, J. B., 2000  Cell-cell signaling in the self-incompatibility response. Curr. Opin. Plant Biol. 3:368-373[Medline].

NASRALLAH, J. B., T.-H. KAO, C.-H. CHEN, M. L. GOLDBERG, and M. E. NASRALLAH, 1987  Amino acid sequence of glycoproteins encoded by three alleles of the S-locus of Brassica oleracea. Nature 326:617-619.

NISHIO, T. and M. KUSABA, 2000  Sequence diversity of SLG and SRK in Brassica oleracea L. Ann. Bot. 85:141-146[Abstract/Free Full Text].

NISHIO, T., M. KUSABA, M. WATANABE, and K. HINATA, 1996  Registration of S alleles in Brassica campestris L by the restriction fragment sizes of SLGs. Theor. Appl. Genet. 92:388-394.

NISHIO, T., M. KUSABA, K. SAKAMOTA, and D. J. OCKENDON, 1997  Polymorphism of the kinase domain of the S-locus receptor kinase gene (SRK) in Brassica oleracea L. Theor. Appl. Genet. 95:335-342.

OKAZAKI, K., M. KUSABA, D. J. OCKENDON, and T. NISHIO, 1999  Characterization of S tester lines in Brassica oleracea: polymorphism of restriction length of SLG homologues and isolelectric points of S-locus proteins. Theor. Appl. Genet. 98:1329-1334.

PARNISKE, M. and J. D. JONES, 1999  Recombination between diverged clusters of the tomato Cf-9 plant disease resistance gene family. Proc. Natl. Acad. Sci. USA 96:5850-5855[Abstract/Free Full Text].

PASTUGLIA, M., D. ROBY, C. DUMAS, and J. M. COCK, 1997a  Rapid induction by wounding and bacterial infection of an S gene family receptor-like kinase gene in Brassica oleracea. Plant Cell 9:49-60[Abstract].

PASTUGLIA, M., V. RUFFIO-CHÂBLE, V. DELORME, T. GAUDE, and C. DUMAS et al., 1997b  A functional S locus anther gene is not required for the self-incompatibility response in Brassica oleracea. Plant Cell 9:2065-2076[Abstract