Self-Incompatibility in the Genus Arabidopsis: Characterization of the S Locus in the Outcrossing A. lyrata and Its Autogamous Relative A. thaliana

Journal List > Plant Cell > v.13(3); Mar 2001

Plant Cell. 2001 March; 13(3): 627–644.

PMCID: PMC135518

Self-Incompatibility in the Genus Arabidopsis

Characterization of the S Locus in the Outcrossing A. lyrata and Its Autogamous Relative A. thaliana

Makoto Kusaba,¹^a Kathleen Dwyer,^b Jennifer Hendershot,^b Julia Vrebalov,^c June B. Nasrallah,^a and Mikhail E. Nasrallah²^a

^aDepartment of Plant Biology, Cornell University, Ithaca, New York 14853

^bBiology Department, University of Scranton, Scranton, Pennsylvania 18510

^cFederal Soil and Nutrition Laboratory, United States Department of Agriculture, Ithaca, New York 14853

¹Current address: Institute of Radiation Breeding, National Institute of Agrobiological Resources, Ministry of Agriculture, Forestry, and Fisheries, Ohmiya-machi, Naka-gun, Ibaraki 319-2293, Japan.

²To whom correspondence should be addressed. E-mail men4/at/cornell.edu; fax 607-255-5407

Received October 11, 2000; Accepted January 15, 2001.

This article has been cited by other articles in PMC.

Abstract

As a starting point for a phylogenetic study of self-incompatibility (SI) in crucifers and to elucidate the genetic basis of transitions between outcrossing and self-fertilizing mating systems in this family, we investigated the SI system of Arabidopsis lyrata. A. lyrata is an outcrossing close relative of the self-fertile A. thaliana and is thought to have diverged from A. thaliana ~5 million years ago and from Brassica spp 15 to 20 million years ago. Analysis of two S (sterility) locus haplotypes demonstrates that the A. lyrata S locus contains tightly linked orthologs of the S locus receptor kinase (SRK) gene and the S locus cysteine-rich protein (SCR) gene, which are the determinants of SI specificity in stigma and pollen, respectively, but lacks an S locus glycoprotein gene. As described previously in Brassica, the S haplotypes of A. lyrata differ by the rearranged order of their genes and by their variable physical sizes. Comparative mapping of the A. lyrata and Brassica S loci indicates that the S locus of crucifers is a dynamic locus that has undergone several duplication events since the Arabidopsis–Brassica split and was translocated as a unit between two distant chromosomal locations during diversification of the two taxa. Furthermore, comparative analysis of the S locus region of A. lyrata and its homeolog in self-fertile A. thaliana identified orthologs of the SRK and SCR genes and demonstrated that self-compatibility in this species is associated with inactivation of SI specificity genes.

INTRODUCTION

Self-incompatibility (SI) is the major outcrossing mechanism in the family Brassicaceae (de Nettancourt, 1977). Species in this family have been grouped into 19 tribes on the basis of morphological criteria (Schultz, 1936), and SI has been described in all tribes analyzed to date. When Bateman (1955) surveyed 182 species distributed in 11 tribes, he found that approximately half of these species included self-incompatible accessions. In a survey of 59 taxa in the subtribe Brassicineae of the tribe Brassiceae (which includes Brassica and Raphanus), 50 taxa were self-incompatible (Takahata and Hinata, 1980). In all cases analyzed, SI has been shown to be controlled sporophytically by a single S (sterility) locus, with multiple alleles or variants and complex dominance relationships between alleles (Bateman, 1954, 1955; Thompson and Taylor, 1966): in self-incompatible plants, pollen will not develop on a stigma that expresses the same S alleles as the pollen parent.

Molecular analysis of the Brassica S locus region has shown that this mendelian locus is a gene complex consisting of distinct stigma-expressed and anther-expressed genes that determine SI specificity in stigma and pollen, respectively (reviewed in Nasrallah, 2000). The SRK (for S locus receptor kinase) gene (Stein et al., 1991) encodes a plasma membrane–spanning receptor serine/threonine kinase specific to the stigma epidermis (Stein et al., 1996) and is the determinant of SI specificity in the stigma (Takasaki et al., 2000). The SCR (for S locus cysteine-rich protein) gene, which is the determinant of SI specificity in pollen (Schopfer et al., 1999), is expressed specifically in the anther tapetum and in microspores (Schopfer et al., 1999; Schopfer and Nasrallah, 2000; Takayama et al., 2000) and encodes a small secreted cysteine-rich protein predicted to localize to the pollen coat and to function as a ligand for the SRK receptor. As expected for genes involved in self-recognition, the SRK and SCR genes are highly polymorphic (Stein et al., 1991; Kusaba et al., 1997; Kusaba and Nishio, 1999; Schopfer et al., 1999; Schopfer and Nasrallah, 2000; Watanabe et al., 2000) and appear to have coevolved. A third highly polymorphic gene contained within the S locus in most Brassica S haplotypes is the SLG (for S locus glycoprotein) gene (Nasrallah et al., 1987; Kusaba et al., 1997). This gene exhibits a high degree of sequence similarity to the extracellular domain or the S domain of SRK (Stein et al., 1991; Kusaba et al., 1997) and encodes a glycoprotein specific to the papillar cell and localized in its wall (Kandasamy et al., 1989). The function of SLG is not well understood, and its role in SI has been questioned (Cabrillac et al., 1999; Nasrallah, 2000; Nishio and Kusaba, 2000). However, recent studies have indicated that SLG enhances the intensity of the SI response (Takasaki et al., 2000), possibly by contributing to the proper maturation of the SRK receptor and its accumulation to physiologically relevant levels (Dixit et al., 2000).

Sequence comparisons of Brassica SRK alleles have allowed a grouping of S haplotypes into two classes. Class I consists of S haplotypes that are placed high in the dominance series and that determine a robust SI response; class II consists of S haplotypes that are recessive to class I haplotypes in pollen and that determine a relatively leaky SI response (Chen and Nasrallah, 1990; Kusaba et al., 1997). Class I and class II SRK alleles diverge by >30%, whereas class I and class II SCR alleles are so diverged from one another (Schopfer et al., 1999) that the high number of available class I SCR sequences has not allowed the isolation of class II SCR alleles.

Comparative mapping of different S haplotypes has demonstrated that they vary not only in the sequence of their SI genes but also in the order, relative orientation, and spacing of these genes (Boyes and Nasrallah, 1993; Boyes et al., 1997; Cui et al., 1999; Nasrallah, 2000; Takayama et al., 2000). This structural heteromorphism and the sequence polymorphism of SI genes exhibit trans-specific variation (Dwyer et al., 1991; Kusaba et al., 1997; Schopfer et al., 1999; Uyenoyama, 2000). Thus, the S haplotypes of Brassica are thought to be ancient and to predate speciation in the genus. Estimates of divergence time have indicated that class I and class II S haplotypes diverged ~40 million years ago, whereas S haplotypes within class I and class II diverged ~25 and 7 million years ago, respectively (Uyenoyama, 1995, 2000).

In view of the ancient age of Brassica S haplotypes, an understanding of the origins and evolution of the S locus complex and its SI specificity genes in the crucifer family requires comparisons with S locus genes from tribes of the Brassicaceae other than the tribe Brassiceae. Because a detailed molecular analysis of SI has been performed only in Brassica species, it is not known which of the features elucidated for SI and the S locus in Brassica are specific to lineages within the Brassiceae and which features are common to the S loci of crucifers in general. Application of a phylogenetic approach to the study of SI is likely to provide insight into the structure and composition of the archetypal S locus as well as the origins of S locus polymorphisms.

As a starting point for this analysis, we elected to investigate SI in the genus Arabidopsis. Classically, the genus Arabidopsis has been placed in the tribe Sisymbrieae (Al-Shehbaz, 1984), but more recently it has been included in the tribe Arabideae (Price et al., 1994). In any case, Arabidopsis is thought to have diverged from Brassica ~15 to 20 million years ago (Yang et al., 1999). For our analysis, we used an outcrossing herbaceous perennial plant that was incorporated recently into the genus Arabidopsis on the basis of molecular data (O'Kane and Al-Shehbaz, 1997; van Treuren et al., 1997; Koch et al., 1999). This species, Arabidopsis lyrata (formerly known as Arabis lyrata, Arabis petraea, or Cardaminopsis petraea), has a genome size approximately twice that of A. thaliana. The genes of the two Arabidopsis spp share a high degree of sequence similarity, allowing facile transfer of molecular markers and other data generated by the A. thaliana genome project to A. lyrata (van Treuren et al., 1997). In addition, unlike Brassica spp, A. lyrata is not cultivated and is not closely related to cultivated species. This feature has made it an attractive model for population genetics and ecological studies (Koch et al., 1999) and will allow a study of SI in natural populations that have not been subjected to selective breeding.

Importantly, a molecular analysis of SI in A. lyrata and comparisons with its self-fertilizing close relative A. thaliana are likely to provide insight into the evolution of SI and the genetic basis of transitions in mating systems within the crucifer family. In a previous comparative mapping study of the S locus region of Brassica and its homeolog in A. thaliana, we found that the two regions, although largely colinear, differ by the absence of the SI specificity genes in A. thaliana (Conner et al., 1998). This result suggested two possibilities. One possibility is that the A. thaliana genome does not contain SI specificity genes, which implies that self-compatibility in this species resulted from deletion of these genes. The other possibility is that the SI specificity genes were translocated to different chromosomal locations in the Brassica and Arabidopsis lineages, which leaves open the question of why A. thaliana is self-fertile. Analysis of the S locus and SI specificity genes in A. lyrata is expected to distinguish between these possibilities.

In this article, we report on the analysis of SI in a population of A. lyrata plants segregating for two SI specificities. We describe the isolation of SRK and SCR genes from the two S haplotypes encoding these specificities. The sequence polymorphism exhibited by these genes as well as the complement of genes contained in the A. lyrata S locus and their genomic organization are discussed in relation to our current knowledge of SI in Brassica. Furthermore, on the basis of the chromosomal location of the S locus in Arabidopsis spp and the analysis of A. thaliana sequences orthologous to the SRK and SCR genes, the molecular basis for self-compatibility in this species is inferred.

RESULTS

Cytology, Development, and Genetics of SI in A. lyrata

A. lyrata has been described as a largely outcrossing species on the basis of the inability of isolated plants to self-pollinate and set seed (Schierup, 1998). However, cytological and developmental studies of pollen rejection have not been reported for this species. We examined pollen tube behavior on A. lyrata stigmas after self- and cross-pollination and at different stages of stigma development. This microscopic analysis showed that self-pollen is inhibited at the surface of the stigma epidermis in mature flowers and that the stigmas of immature buds support the development of pollen tubes in self-pollinations (data not shown). Thus, the SI response in A. lyrata is similar to that of other self-incompatible crucifers, both in its site of inhibition and in its developmental regulation. Furthermore, as in Brassica, S locus homozygotes can be generated and maintained by forced self-pollination in immature young buds before the stigma acquires competence for pollen rejection.

To investigate the genetic control of SI in A. lyrata, we started with seed collected from the wild in Michigan and kindly provided by Charles Langley (University of California at Davis), from which we generated a population that segregated for SI phenotype by isolating a single plant and forcing self-pollination on developing floral buds. A population of 54 plants was analyzed by microscopic examination of pollen tube growth in self-pollinations and in reciprocal cross-pollinations among siblings and with the parent plant (see Methods). The data are consistent with the segregation of two SI specificities, Sa and Sb, and with the single locus control of SI. As shown in Table 1, the plants were grouped into three classes based on pollination phenotype: 18 plants were SaSa, 27 plants were SaSb, and nine plants were SbSb. The observed ratios of the three classes approximate the 1:2:1 ratio expected for segregation of a single locus ( equation M1

). In pollen of heterozygotes, Sb is dominant to Sa. In stigmas of heterozygotes, Sa and Sb exhibit codominance, with weakening of Sa activity. The Sa haplotype also conferred a somewhat leaky SI response in SaSa homozygotes, as indicated by increased seed set resulting from spontaneous self-pollination (pseudocompatibility) relative to SbSb homozygotes and SaSb heterozygotes.

Table 1.

Pollination Analysis of Plants Segregating for Sa and Sb Specificities^a

Molecular Cloning of A. lyrata SRKa and SRKb Genes

A cDNA library was constructed in the bacteriophage vector λUniZAP from pistil mRNA isolated from an SaSb plant. The library was screened at low stringency with S domain probes from Brassica SLG and related genes (see Methods). Positive clones that also contained kinase sequences were identified subsequently by hybridization with a probe derived from the kinase domain of Brassica SRK. Several clones that hybridized with both probes were sequenced partially and found to be derived from the same gene. Sequence analysis of the clone with the longest insert revealed one open reading frame (ORF) with the features characteristic of a functional SRK gene (see below). To determine if this gene was in fact derived from the A. lyrata S locus, we performed restriction fragment length polymorphism (RFLP) analysis of plants segregating for the Sa and Sb haplotypes. As shown in Figure 1A, under high-stringency wash conditions (see Methods), the S domain of the putative SRKa gene hybridized with a 1.7-kb restriction fragment in EcoRV digests of DNA from plants carrying the Sa haplotype. This probe did not hybridize with DNA from Sb homozygotes, indicating a high degree of divergence between the two SRK alleles. Furthermore, in the 54 plants analyzed, there was perfect correlation between the presence of this 1.7-kb EcoRV fragment and SI phenotype deduced from pollination assays. These results demonstrate that we had isolated the SRKa allele of A. lyrata. Additional confirmation of this conclusion derives from work done independently by Deborah Charlesworth and co-workers, who amplified an A. lyrata genomic sequence nearly identical to the S domain of the SRKa gene and showed its linkage to the S locus by using an independent segregating population (Schierup et al., 2001).

Figure 1.

DNA Gel Blot Analysis of the A. lyrata SRK Gene.

Genomic clones for the SRKa gene were isolated from libraries of A. lyrata SaSb DNA constructed in the bacteriophage vector λGEM11 and in the bacterial artificial chromosome (BAC) vector pBeloBAC (see Methods). A screen of the λGEM11 library with the S domain of SRKa resulted in the isolation of several λ clones for SRKa. Analysis of sequences flanking the SRKa gene in these clones provided single-copy probes that subsequently were used to probe the BAC library, leading to the isolation of BAC clones derived from the Sa and Sb haplotypes (see Methods).

In our RFLP analysis of plants segregating for the Sa and Sb specificities by using the SRKa S domain probe, we detected an Sb-associated 550-bp XhoI fragment under low-stringency wash conditions (data not shown). A BAC clone derived from the Sb haplotype was digested with XhoI, and a subclone containing the 550-bp XhoI fragment was obtained. This 550-bp XhoI fragment was used to screen the pistil cDNA library, resulting in the isolation of six independent cDNA clones. All of these clones hybridized with a kinase domain probe, suggesting that they contained SRK-like sequences. Further analysis indicated that the six clones were all derived from the same gene. RFLP analysis of our segregating population with the 550-bp XhoI probe confirmed the linkage of this gene to the S locus. Under high-stringency wash conditions, the 550-bp XhoI probe detected a 1.3-kb restriction fragment that cosegregated with Sb specificity, as shown by the EcoRV digests in Figure 1B.

DNA isolated from SaSa and SbSb homozygotes was digested with several restriction enzymes and subjected to gel blot analysis by using S domain probes from the SRKa and SRKb genes, respectively. Figure 1C shows that under high-stringency conditions, the SRKa probe hybridized with a single restriction fragment in SacI and EcoRV digests. Two restriction fragments were detected in XhoI digests due to the presence of an internal XhoI site within the probe. These results indicate that the A. lyrata genome does not contain other genes with a high degree of sequence similarity to SRKa. In particular, a homolog of the SLG gene, which is detectable in Brassica DNA by using SRK S domain probes even under high-stringency conditions, appears to be absent from the Sa haplotype of A. lyrata.

In a similar gel blot analysis of DNA isolated from SbSb homozygotes, the SRKb-derived 550-bp XhoI probe detected an additional restriction fragment under high-stringency wash conditions (Figure 1D), indicating the presence of a sequence with appreciable sequence similarity to the SRKb gene. A restriction fragment of a slightly different size, possibly containing an allelic form of this sequence, also was detected by the 550-bp XhoI probe in SaSa plants (Figure 1B). The RFLP for this sequence cosegregated with the S locus in the 12 plants analyzed, suggesting its possible linkage to the S locus. However, the gene was not located in the S locus BAC contig; therefore, it must lie outside of the S locus proper.

Sequence Polymorphism of the A. lyrata SRK Gene

Database searches showed that the A. lyrata SRK genes share appreciable sequence similarity with Brassica SRK alleles as well as with several other members of the S gene family. The sequence alignment in Figure 2A shows that the SRKa and SRKb genes exhibit the features characteristic of receptor-like kinases of the S gene family: a signal sequence, a predicted extracellular S domain containing 12 conserved cysteine residues, a transmembrane domain, and a kinase domain with conserved residues suggestive of serine/threonine kinase activity. Alignment of the cDNA and genomic sequences revealed that SRKa and SRKb share with Brassica SRK genes (Stein et al., 1991) and A. thaliana S receptor–like kinase genes (Tobias et al., 1992; Dwyer et al., 1994; Tobias and Nasrallah, 1996) a conserved exon–intron structure in which the entire S domain is contained on the first exon, the transmembrane and juxtamembrane domains are encoded by the second exon, and the kinase domain is contained on exons 3 to 7. The six introns of SRKa and SRKb occur at the same positions as in other S receptor kinase genes, but they differ in size between the two genes (arrowheads and numbers in Figure 2A).

Figure 2.

Polymorphism of the A. lyrata SRK Gene and Relationship to the Brassica SRK Genes and A. thaliana ARK Genes.

The SRKa and SRKb ORFs are highly diverged and exhibit only 62.6% amino acid sequence similarity. In Brassica, sequence similarities in the range of 60 to 65% are observed in comparisons between class I and class II SRK alleles, whereas sequence similarities of >80% are typically observed in comparisons of SRK alleles within a class. It is possible that the Sa and Sb haplotypes belong to two distinct classes of A. lyrata SRK alleles. However, the SRKa and SRKb sequences are equally diverged from the class I and class II alleles of Brassica: each allele shares 50 to 60% sequence similarity with the Brassica class I SRK6 and class II SRK2 alleles, and these A. lyrata alleles did not group with either Brassica allele in a preliminary phylogenetic analysis (Figure 2B).

Expression Pattern of the A. lyrata SRK Gene

Reverse transcription–polymerase chain reaction (RT-PCR) analysis using primers specific for the SRKa gene showed that this gene is expressed at its highest level in stigmas (Figure 3A). SRKa expression is maximal at 1 to 2 days before flower opening and persists in stigmas of open flowers. The gene also is expressed at low levels in styles but not in anthers at 1 to 2 days before flower opening. This expression pattern is largely identical to that of Brassica SRK genes. However, one difference between the two taxa may be noted: in A. lyrata, SRK expression was detected only in pistils, whereas in Brassica, SRK also exhibited a low level but physiologically irrelevant expression in anthers (Stein et al., 1991, 1996; Conner et al., 1997; Cui et al., 2000; Takasaki et al., 2000).

Figure 3.

Expression of the A. lyrata SRK Gene.

A. lyrata RNA isolated from SaSa and SbSb pistils was examined by gel blot analysis using S domain probes derived from the SRKa and SRKb genes, respectively. This analysis revealed the presence of two transcript species: a 2.8-kb transcript that corresponds to the size of the full-length SRK transcript, and a shorter 1.6-kb transcript (Figure 3B) that is ~5 and 10% as abundant as the 2.8-kb transcript in SaSa and SbSb pistils, respectively. Brassica stigmas also produce 2.8- and 1.6-kb transcripts that are detected with SRK S domain probes (Stein et al., 1991). However, in most Brassica strains analyzed to date, the steady state amount of the 1.6-kb transcripts exceeds that of the 2.8-kb transcripts by as much as two orders of magnitude. The Brassica 1.6-kb transcripts have been shown to consist predominantly of transcripts derived from the SLG gene and to a lesser extent of SRK transcripts that contain the S domain–encoding exon and terminate within the first intron of the gene (Stein et al., 1991; Giranton et al., 1995). In A. lyrata, the entire 1.6-kb transcript population must be derived from the SRK gene, based on of the absence of an SLG-like gene in the Sa and Sb haplotypes we analyzed.

Molecular Cloning and Characterization of the A. lyrata SCR Genes

We postulated that the SCR gene would be located in close physical proximity to the SRK gene in A. lyrata, as it is in Brassica. Furthermore, the SCR alleles of Brassica exhibit extensive sequence divergence even when S haplotypes are compared, the SRK genes of which share >90% sequence similarity. Because of the >30% sequence divergence exhibited by the SRKa and SRKb genes, we also postulated that a probe containing the SCRa gene would not hybridize with transcripts encoded by the SCRb gene. Therefore, to identify the SCR genes in the Sa and Sb haplotypes, we used subclones of SRKa-containing λ genomic clones and of an SRKb-containing BAC clone to probe gel blots of poly(A)⁺ RNA isolated from SaSa and SbSb anthers. An 8.8-kb XhoI fragment (Figure 4A) from an SRKa-containing λ clone hybridized specifically with an ~500-bp transcript in SaSa anthers (data not shown), suggesting that this fragment contained the SCRa allele. Similarly, a 14.5-kb EagI fragment from an SRKb-containing BAC clone (Figure 4A) hybridized specifically with an ~500-bp transcript in SbSb anthers (data not shown), indicating that it contained the SCRb allele.

Figure 4.

Location, Polymorphism, and Expression of the A. lyrata SCR Gene.

The 8.8-kb XhoI fragment and the 14.5-kb EagI fragment were used to screen anther cDNA libraries prepared from SaSa and SbSb plants, respectively. Several positive clones were isolated from each library. Comparison of cDNA and genomic sequences revealed that the A. lyrata SCR genes have a structure similar to that of the Brassica SCR genes. Each gene is interrupted by one intron that is located toward the 3′ end of the signal sequence (Figure 4B) and measures ~1.5 kb in SCRa and 78 bp in SCRb. The cDNA clones isolated from the SaSa library all contained an identical 258-bp ORF with a structure characteristic of the SCR gene (Figure 4B). This ORF predicts a polypeptide of 86 amino acid residues containing a potential signal sequence that, when cleaved, would result in a mature protein of 63 amino acids having a molecular mass of 7.3 kD and a pI of 8.3. The cDNA clones isolated from the SbSb library also all contained an identical 243-bp ORF that encodes a protein of 81 amino acids containing a potential signal sequence (Figure 4B). The predicted mature SCRb protein is 55 amino acids long and has a molecular mass of 6.5 kD and a pI of 8.3.

The SCRa and SCRb sequences are highly diverged from each other and from Brassica SCR genes. SCRa and SCRb share only 35.2% amino acid sequence identity, a level of intraspecific allelic polymorphism also exhibited by Brassica SCR genes. As shown in Figure 4B, the predicted mature SCRa and SCRb proteins contain the eight cysteines and one glycine residue that are conserved in most Brassica SCR alleles analyzed to date (Schopfer et al., 1999; Schopfer and Nasrallah, 2000; Takayama et al., 2000; Watanabe et al., 2000). Also like their Brassica orthologs, SCRa and SCRb differ in the sequence and length of the segments between and flanking these conserved residues. Interestingly, the SCRa and SCRb alleles also have highly divergent signal sequences and differ in the positions of their predicted signal sequence cleavage sites (Figure 4B). In this respect, they differ from the Brassica SCR alleles isolated to date, all of which contain relatively conserved signal sequences.

The SCRa and SCRb genes also differ in their positions within the S locus relative to the SRK gene. Fine structure mapping of an Sa-derived BAC clone indicated that SRKa and SCRa are located in inverse orientation and within ~1.5 kb of each other (Figure 4A). Unexpectedly, restriction analysis of Sb BAC clones showed that the Sb haplotype contained two copies of SCRb, designated SCRb-1 and SCRb-2 (Figure 4A). These two genes are located within a longer direct repeat that also includes a retroelement-like sequence downstream from SCRb-2 and a truncated duplicate of the element downstream from SCRb-1 (Figure 4A). The SCRb-1 and SCRb-2 genes are arranged in the same orientation as SRKb, with SCRb-1 positioned ~15.5 kb from SCRb-2 and ~21.0 kb from SRKb. The exons of the SCRb-1 and SCRb-2 genes are identical to the SCRb cDNA sequence, and the two genes also are identical to each other over the length of sequence analyzed, namely, from 580 bp upstream of the initiation codon to 280 bp downstream from the stop codon. Therefore, it is likely that SCRb-1 and SCRb-2 are both functional genes and exhibit identical expression patterns.

RNA gel blot analysis demonstrated that the SCRa and SCRb genes are expressed specifically in anthers. As shown in Figure 4C for SCRb, transcripts were detected in both young and mature anthers as well as in isolated microspores. This expression pattern is similar to that described previously for Brassica SCR genes by RNA gel blot analysis (Schopfer et al., 1999), in situ hybridization (Takayama et al., 2000), and reporter gene analysis (Schopfer and Nasrallah, 2000).

Chromosomal Location of the S Locus in Arabidopsis spp

Before our construction and screening of a pistil cDNA from A. lyrata as described above, we had attempted to isolate S locus genes from A. lyrata based on the assumptions that the S locus occupied a similar chromosomal location in Brassica and A. lyrata and that the regions flanking the SI genes would be largely colinear in the two taxa. We previously had mapped the Brassica S locus to an ETR1-linked position and performed a comparative analysis of the Brassica S locus region and its ETR1-linked homeolog in A. thaliana (Conner et al., 1998). This analysis had identified several molecular markers that flank the SI specificity genes in Brassica and are conserved in the homeologous region of the A. thaliana genome (Conner et al., 1998). From these markers, we selected markers that were single copy in Brassica and A. thaliana and that were located on either side of the Brassica SRK–SCR–SLG complex (see Methods). These markers were used to screen a library of A. lyrata genomic DNA constructed in the bacteriophage vector λGEM11, and 50 positive clones were isolated and assembled into contigs of overlapping clones (data not shown). Restriction mapping of the clones confirmed the synteny of the A. thaliana and A. lyrata homeologous regions. Unexpectedly, however, none of the clones hybridized with Brassica SRK-derived probes, suggesting that they did not contain SRK-related sequences. We subsequently found that these clones did not hybridize with S domain probes from A. lyrata SRK and that λGEM11 clones containing the SRKa gene did not hybridize with the molecular markers derived from the S locus region of Brassica. These unexpected results indicated that the S locus occupies a different region of the A. lyrata genome than it does in Brassica.

Database searches using the SRKa sequence showed that it is most similar to an A. thaliana SFR2-like receptor kinase (GenBank accession number CAA20202) that is located on BAC T6K22 in contig fragment 55 of chromosome IV. In addition, database searches with partial DNA sequences of subclones derived from regions flanking the A. lyrata SRKa gene produced hits on chromosome IV of A. thaliana. Specifically, a subset of these sequences exhibited 94 to 96% sequence identity with the S domain of the ARK3 gene and 90% sequence identity with a BRLK-like gene, both of which are contained on BAC T6K22. This ARK3-containing region thus is identified as the A. thaliana homeolog of the A. lyrata S locus region.

Long-Range Comparative Mapping of Two Haplotypes of the A. lyrata S Locus and Its Homeolog in A. thaliana

A map of BAC T6K22 is shown in Figure 5. The region spanned by the BAC clone contains several predicted genes, including a cluster of five genes that exhibit sequence relatedness to kinase genes. Of these predicted kinase genes, three belong to the S receptor kinase class of genes and therefore are candidates for being the A. thaliana counterpart of SRK. These are T6K22.120 (GenBank accession number CAA20204; designated B120 in Figure 5), a sequence related to the B. oleracea BRLK gene; T6K22.110, which is identical to ARK3 (GenBank accession number CAA20203), a gene characterized previously (Dwyer et al., 1994); and T6K22.100 (GenBank accession number CAA20202), a sequence related to a wound-inducible SFR2 receptor-like kinase from B. oleracea (Pastuglia et al., 1997a).

Figure 5.

Comparative Map of the A. lyrata Sa and Sb Haplotypes and the Homeologous Region of A. thaliana.

To compare the organization and gene content of this A. thaliana region to the A. lyrata S locus region and identify possible A. thaliana orthologs of SI specificity genes, we performed a comparative mapping study of A. thaliana BAC T6K22 and A. lyrata BAC clones corresponding to the Sa and Sb haplotypes. The availability of the sequence of BAC T6K22 provided us with a collection of molecular markers that allowed us to derive long-range maps for the A. lyrata Sa and Sb haplotypes and to compare the A. lyrata S locus with its homeolog in A. thaliana. Restriction mapping and partial sequencing of the A. lyrata Sa and Sb BAC clones revealed a high degree of microsynteny between the A. lyrata and A. thaliana homeologous regions within segments that flank the SRK-SCR–containing segment (namely, in the segment defined by the B50 and B80 genes and in the segment defined by the ARK3 and B160 genes). However, the segment between ARK3 and B80, which contains the SRK and SCR genes (i.e., the S locus proper), differs markedly in the three regions analyzed, both in its size and organization. This segment is 32 kb in length in A. thaliana, ~50 kb in the A. lyrata Sa haplotype, and ~100 kb in the A. lyrata Sb haplotype. This segment also differs in the relative orientation and spacing of the SRK and SCR sequences, not only between the A. lyrata Sa and Sb haplotypes, as discussed above, but also between A. thaliana and the two A. lyrata regions. Interestingly, all three regions contain sequences with similarity to pol genes of retrotransposons: the A. thaliana and the A. lyrata Sa haplotype each contains at least one such sequence, whereas the A. lyrata Sb haplotype contains at least five such sequences.

Identification of SRK and SCR Orthologs in A. thaliana

The SFR2-like kinase T6K22.100 most likely represents the A. thaliana ortholog of SRK. This conclusion is based on several lines of evidence. First, the comparative mapping study of the largely syntenous A. lyrata S locus region and its homeolog in A. thaliana described above shows that the T6K22.100 sequence and the A. lyrata SRK genes are located in the same positions relative to flanking genes (Figure 5). Second, orthologs of BRLK-like and ARK3 genes have been identified in A. lyrata Sa and Sb haplotypes and are distinct from the SRK gene. Third, ARK3, which is the S kinase gene that is most tightly linked to SRK in A. lyrata, is unlikely to function in SI because its promoter was found to be expressed at the root–hypocotyl junction and at the bases of lateral roots, axillary buds, and pedicels (Dwyer et al., 1994).

To identify sequences orthologous to SCR, we searched BAC T6K22 for the presence of sequences related to A. lyrata SCR genes. The search identified three SCR-related sequences (Figure 6A) downstream from the SRK ortholog T6K22.100, within a 15.7-kb region lacking predicted exons/ORFs (Figure 5). None of the three candidate SCR orthologs is predicted to encode full-length SCR proteins, and therefore they are designated ΨSCR1, ΨSCR2, and ΨSCR3 (Figure 6A). The ΨSCR1 sequence, located ~700 bp downstream from T6K22.100 between positions 50,496 and 51,472 of the BAC T6K22 sequence, exhibits the exon–intron structure characteristic of SCR genes but contains a truncated ORF (Figure 6A). The ΨSCR2 and ΨSCR3 sequences are located ~8.5 kb downstream from T6K22.100 at positions 43,623 to 43,664 and 43,724 to 43,687 of BAC T6K22, respectively, and share sequence similarity with the SCRa signal sequence (Figure 6A) as well as with the SCRa 5′ untranslated region (data not shown).

Figure 6.

The ΨSRK and ΨSCR Sequences of A. thaliana.

Analysis of the A. thaliana SRK Ortholog

The annotation of the SFR2-like kinase T6K22.100 predicts a functional gene that encodes a full-length receptor kinase containing all of the features of receptor protein kinases in the S gene family, including 12 conserved cysteines in the extracellular S domain and the residues conserved in kinases. This annotation assumes the gene structure shown in Figure 6B, which is typical of Brassica and A. lyrata SRK as well as other receptor kinases in the S gene family.

To determine if this SRK ortholog produces transcripts that have the potential to encode a functional receptor kinase, we screened a cDNA library prepared from A. thaliana floral bud RNA with a fragment containing the S domain of the gene. We isolated three independent cDNA clones derived from the T6K22.100 gene. Sequence analysis of these cDNAs revealed that the gene is unlikely to produce a receptor kinase product. On the basis of sequence analysis of the three cDNA clones, the T6K22.100 gene is predicted to encode a 547–amino acid ORF rather than the 844–amino acid ORF predicted in the database (Figure 6B). This truncated ORF terminates within exon 4 due to the use, at the intron 3–exon 4 boundary, of a splice site different from that predicted in the database. As shown in Figure 6C, the use of this splice site introduces a stop codon six codons into exon 4. The T6K22.100 gene therefore is a nonfunctional ortholog of SRK and is designated ΨSRK.

The isolation of cDNA clones for the ΨSRK gene demonstrated that the gene was transcribed, and it was of interest to investigate its spatial and temporal expression pattern. To characterize the activity of the ΨSRK promoter, we amplified a fragment spanning a region from 23 to ~1600 bp upstream of the initiating methionine codon and fused it to the uidA gene, which encodes the β-glucuronidase (GUS) reporter enzyme (see Methods). Transformation of A. thaliana with the resulting ΨSRK::uidA::nos chimeric gene produced 25 independent transformants. In all transformants, GUS activity was restricted to the pistil. Twenty-three transformed families exhibited intense GUS staining in the stigma epidermis as well as in the transmitting tract of the style and ovary, as shown in Figure 7. In the remaining two families, GUS staining was confined to the stigma epidermis (data not shown). These results indicate that the ΨSRK promoter exhibits a spatial activity pattern in pistils identical to that of the functional SLG and SRK genes of Brassica (Sato et al., 1991; Toriyama et al., 1991; Stein et al., 1996).

Figure 7.

Activity of the ΨSRK Gene Promoter.

DISCUSSION

We have shown that the SI response of A. lyrata is similar to that of Brassica in its sporophytic control by a single specificity-determining locus, in its cytological manifestation by the inhibition of self-pollen at the surface of stigma epidermal cells, and in the molecular basis of self-recognition. These observations are consistent with a monophyletic origin of SI in the crucifer family.

Using a population derived by selfing a plant heterozygous for two S haplotypes that we designated Sa and Sb, we isolated from the A. lyrata S locus orthologs of the Brassica SRK and SCR genes that determine SI specificity in stigma and pollen. Our analysis of the two S haplotypes and the isolation of two alleles each for the SRK and SCR genes have demonstrated that these genes exhibit expression patterns that are largely similar to those of their Brassica orthologs. The two genes also occur in close physical proximity and exhibit extensive sequence polymorphism, as in Brassica. Interestingly, the ~37% sequence divergence exhibited by the SRKa and SRKb genes is comparable to the sequence divergence observed between class I and class II SRK alleles in Brassica, which implies an ancient divergence of the two alleles similar to that suggested for the two classes of Brassica S haplotypes. The Sa and Sb haplotypes might represent extremes in the spectrum of S haplotypic divergence for A. lyrata. The availability of SRK and SCR sequences from these S haplotypes should facilitate the isolation of other SRK and SCR alleles from this species and allow a study of allelic polymorphism and distribution in natural populations. Furthermore, the extensive sequence divergence of the SRKa and SRKb genes suggests that A. lyrata S haplotypes may fall into two distinct classes, as in Brassica. However, analysis of additional A. lyrata S haplotypes is needed to determine the range of S locus variability that exists in the species and to confirm the existence of two haplotypic classes. In any case, the observation that SRKa and SRKb sequences are equally diverged from the class I and class II SRK alleles of Brassica suggests that the Sa and Sb haplotypes of A. lyrata diverged from each other after the Brassica–Arabidopsis split. Alternatively, the phylogenetic relationships might reflect the occurrence of gene conversion or recombination events between SRK alleles within each lineage after speciation.

Even more divergent than the SRK alleles are the SCRa and SCRb alleles. The mature proteins predicted by SCRa and SCRb exhibit ~67% sequence divergence. Significantly, the SCRa and SCRb alleles also differ in their predicted signal peptide sequences. In this respect, they differ from the Brassica SCR alleles isolated to date, all of which are derived from class I S haplotypes and have highly conserved signal sequences (Schopfer et al., 1999; Nasrallah, 2000; Takayama et al., 2000; Watanabe et al., 2000). The signal sequence divergence in the A. lyrata SCR genes is consistent with an ancient divergence for these genes and with the conclusion that Sa and Sb belong to distinct classes of extant S haplotypes. Analysis of the SCRa and SCRb sequences thus adds to the body of evidence that the SCR gene is one of the most highly polymorphic genes in eukaryotes, exceeding the polymorphism of other recognition genes such as the pheromone genes that determine mating specificity in the ciliate Euplotes raikovi (Luporini et al., 1995) and the abalone sperm lysin gene that determines species-specific fertilization (Vacquier, 1998).

It is of interest that the A. lyrata Sa and Sb haplotypes appear to lack an SLG-like gene. SLG, which apparently was formed by a duplication of the SRK gene (Tantikanjana et al., 1993), is found in Brassica and Raphanus spp (Sakamoto et al., 1998). It occurs in tight genetic and physical linkage to the SRK and SCR genes and is expressed at high levels in plants homozygous for most Brassica S haplotypes, but it is absent (Okazaki et al., 1999) or expressed at low levels (Tantikanjana et al., 1993, 1996; Gaude et al., 1995; Cabrillac et al., 1999) in plants homozygous for a few S haplotypes. An SRKb-related sequence that may be genetically linked to the S locus was detected on DNA gel blots of A. lyrata Sa and Sb homozygotes. However, this sequence is unlikely to represent an SLG gene because it lacks a distinguishing feature of SLG, namely, that its alleles exhibit extensive S haplotype–associated polymorphisms that parallel those of the SRK gene. Nevertheless, our data do not rule out categorically the possibility that this SRKb-related sequence may be an SLG ortholog, albeit an atypical one.

The conclusion that the A. lyrata Sa and Sb haplotypes lack a “typical” SLG gene is supported by the observation that A. lyrata stigma RNA did not exhibit the high levels of SRK-homologous 1.6-kb transcripts that characterize most of the Brassica strains containing an SLG gene. The modest amounts of 1.6-kb transcripts that are detected by SRK S domain probes in A. lyrata stigmas likely represent truncated SRK transcripts that correspond to the S domain of the gene. These short transcripts are produced by SRK genes (Stein et al., 1991; Giranton et al., 1995) and SRK-like genes of Arabidopsis (Tobias and Nasrallah, 1996). They are predicted to encode a secreted soluble form of the SRK ectodomain and have been shown to produce such a protein for at least one Brassica SRK allele (Giranton et al., 1995). The function of this class of protein is not known. However, the apparently general use of a polyadenylation signal within the first intron of SRK and SRK-like genes is suggestive of some functional significance for the soluble form of the receptor domain. It is possible that in A. lyrata, as well as in the self-incompatible Brassica plants that lack SLG or express low levels of SLG, the soluble form of SRK fulfills a role similar to that proposed for SLG and functions to enhance the SI response (Dixit et al., 2000; Takasaki et al., 2000).

Thus, analysis of the A. lyrata S locus has revealed both similarities and differences relative to the S locus of Brassica. The absence of an SLG-like gene in A. lyrata suggests that the ancestral S locus complex of crucifers consisted of the two SI specificity-determining genes SRK and SCR. The duplication event that gave rise to SLG must have occurred in a progenitor of the tribe Brassiceae/subtribe Brassicineae (which includes Brassica and Raphanus) and must have been lost subsequently in some members of these taxa. In this context, note that duplications, insertions, and deletions, all of which result in the restructuring and expansion or contraction of chromosomal regions, must have occurred frequently during the evolutionary history of the S locus. Clearly, these processes are limited neither to the SLG gene nor to Brassica S haplotypes. Indeed, in A. lyrata, the Sb haplotype differs from the Sa haplotype in the spacing and orientation of SRK and SCR, not only relative to each other but also relative to flanking markers.

The Sb haplotype also differs from the Sa haplotype by the presence of a tandem duplication of the SCR gene and its flanks, which contributes to but does not fully account for the approximately twofold expansion of the SRK/SCR–spanning region in the Sb haplotype relative to the Sa haplotype. The nature and origin of the remainder of the DNA in the expanded segments of the S haplotype remain to be determined, although some of this DNA consists of retroelement-like sequences, at least five of which were detected in the Sb haplotype. In any case, the structural heteromorphism of S haplotypes and the extreme divergence of SI specificity genes, both of which have been documented extensively in Brassica, appear to be common features of the S locus in the Brassicaceae. We conclude that the S locus of the Brassicaceae is a dynamic locus that has been, and perhaps continues to be, subjected to repeated rearrangements, deletions, and insertions. Consistent with this conclusion, the sequence identity of the SCRb-1 and SCRb-2 genes suggests that the direct repeat found in the Sb haplotype was duplicated quite recently. Note that rearranged gene order and fractured homology often are associated with complex loci and are thought to reduce the frequency of recombination within coadapted gene complexes (Ferris and Goodenough, 1994). Similarly, structural heteromorphism at the S locus might serve to maintain the tight association of the SI specificity genes over time (Boyes et al., 1997; Casselman et al., 2000).

An unexpected result of our study was the finding that the S locus occupies different chromosomal locations in Brassica and Arabidopsis spp. In Brassica, the S locus is located in a region that is syntenous with an ETR1-linked chromosomal segment of A. thaliana chromosome I (Conner et al., 1998), whereas in A. lyrata, the S locus maps to a region that corresponds to contig fragments 54 to 55 of A. thaliana chromosome IV. In both locations, the S locus is recognizable not only by the presence of its polymorphic SI specificity genes but also as an island exhibiting rearranged gene order and variable physical size that lies in the midst of an otherwise highly syntenous chromosomal region. These observations suggest that the S locus gene complex was translocated as a unit between two distant genomic locations.

When might this translocation have occurred? One possibility is that it occurred after the Arabidopsis–Brassica split. Translocation of the S locus presumably would have involved only one S haplotype and thus would represent a severe bottleneck on the number of S haplotypes in the derived lineage. Such an event, followed by diversification of the translocated locus and loss or inactivation of SI genes in the original chromosomal location, would explain the distinct locations of the S loci of Brassica and Arabidopsis. This scenario also is consistent with the phylogenetic analysis of S haplotypes, which, although limited, suggests a more recent divergence between Brassica class I and class II S haplotypes and between the A. lyrata Sa and Sb haplotypes than between Brassica and A. lyrata S haplotypes.

Further comparative mapping studies of the S locus in diverse taxa of the Brassicaceae are required to resolve the direction of S locus translocation events and the frequency with which they occur. Such studies also may elucidate the mechanisms that underlie S locus translocation in different lineages and those that underlie the restructuring of the S locus in a particular lineage. In this context, note that the Sa and Sb haplotypes of A. lyrata, the S locus region of A. thaliana, and several Brassica S haplotypes (Boyes and Nasrallah, 1995; Pastuglia et al., 1997b; Cui et al., 1999; Suzuki et al., 1999) all contain retroelement-like sequences. It is possible that retroelement activity, which is thought to be an effective agent of genome reorganization (Bennetzen, 1996), may have played a role in the translocation and restructuring of the S locus in the Brassicaceae.

The identification of the ARK3-linked region as the location of the S locus in Arabidopsis spp allowed us to identify orthologs of the SRK and SCR genes in A. thaliana and to understand the basis of self-compatibility in this species. The A. thaliana SRK ortholog is a nonfunctional gene that encodes a truncated ORF. We suspect that this ΨSRK gene has become inactivated relatively recently because it has remained transcriptionally active and its promoter has retained a pattern of activity identical to that described previously for the functional Brassica SRK genes in transgenic pistils (Stein et al., 1996). On the other hand, only partial remnants of the SCR gene were recognized in A. thaliana. Interestingly, the A. thaliana ΨSRK and ΨSCR sequences are more similar to the A. lyrata SRKa and SCRa genes than to SRKb and SCRb, and the relative orientation of ΨSRK and the longest ΨSCR sequence, ΨSCR1, is similar to that of SRKa and SCRa. Although the position of ΨSRK and ΨSCR relative to the ARK3 and B80 flanking markers is different from that of the SRKa and SCRa genes, these relationships suggest that the nonfunctional S haplotype in the A. thaliana Columbia strain is derived from an Sa-like functional S haplotype.

The presence of the ΨSRK and ΨSCR sequences in A. thaliana indicates that the initial mutation that caused the shift to self-fertility in this species was not a complete deletion of the SI specificity genes but a more subtle mutation that led to the inactivation of one of these genes. The sequence of events that led to the inactivation of SRK and SCR cannot be surmised. However, because SRK and SCR function exclusively in SI, a mutation in either gene is expected to lead eventually to the decay of the other member of the SI recognition gene complex. Further analysis of different A. thaliana ecotypes will determine if all ecotypes contain the same nonfunctional S haplotype and the same lesions in their SRK and SCR orthologs as does the Columbia ecotype. Such a study might determine if the inactivation of SI genes occurred more than once in this species and might produce estimates of the number of S haplotypes that were included in the founding population of the A. thaliana lineage.

METHODS

Plant Material

Arabidopsis lyrata subsp lyrata is a herbaceous outcrossing perennial whose range in North America extends from Minnesota and Wisconsin south into Missouri, east into Georgia, north into Vermont, and west into Ontario (O'Kane and Al-Shehbaz, 1997). We used A. lyrata plants descended from accessions collected in Michigan (kindly provided by Charles Langley, University of California at Davis) and A. thaliana ecotype Columbia.

Pollination Analyses and Determination of the Self-Incompatibility Phenotype

Examination of pollen tube development at the stigma surface was performed using UV light fluorescence microscopy as described previously (Kho and Baer, 1968). To identify self-incompatibility (SI) genotype, we crossed several plants from a segregating population of 54 plants in all possible combinations. These crosses suggested a model of allelic interactions that was verified subsequently using selected homozygous and heterozygous plants as pollen parents in crosses to all plants in the segregating population to determine cross-compatibility relationships. Three groups of plants were identified as a result of this analysis (Table 1).

DNA Gel Blot Analysis

DNA was isolated from leaves according to Murray and Thompson (1980). The DNA was transferred to GeneScreen Plus nylon membranes (DuPont–New England Nuclear, Boston, MA), and the blots were prehybridized and hybridized at 65°C in 10% (w/v) dextran sulfate, 330 mM sodium phosphate, pH 7.0, 10 mM EDTA, and 5% (w/v) SDS. High-stringency washes were performed in a solution containing 0.2 × SSC (1 × SSC is 0.15 M NaCl and 0.015 M sodium citrate) and 0.1% (w/v) SDS at 65°C (in the analysis of the segregating population) or 68°C (in the restriction enzyme surveys).

RNA Isolation, Reverse Transcription–Polymerase Chain Reaction, and RNA Gel Blot Analysis

Pistils and anthers were dissected from flowers and floral buds at different stages of development. Total RNA was isolated by the Trizol method (Gibco BRL, Bethesda, MD). Poly(A)⁺ RNA was purified using the FastTrack RNA isolation kit (Invitrogen, La Jolla, CA).

Reverse transcription–polymerase chain reaction (RT-PCR) was performed using the Advantage RT-for-PCR kit (Clontech, Palo Alto, CA). For amplification of SRKa, we used 10 ng of pistil poly(A)⁺ RNA, the S domain primer 5′-CAAAGAATGGAACATG-3′, and the kinase domain primer 5′-TGATAGTCTTTTCACAGCAATCT-3′. Amplification of the 820-bp SRKa fragment was performed under the following conditions: 94°C for 5 min, 28 cycles of 94°C for 1 min, 55°C for 2 min, and 72°C for 2 min, and a final extension at 72°C for 5 min. The abundance of actin cDNA also was determined by PCR to verify that the PCRs contained equal amounts of cDNA. For amplification of actin, we used the following degenerate primers that were designed for the amplification of A. thaliana actin genes and kindly provided by M.K. Kandasamy (University of Georgia, Athens): 5′-TCYTTNCTN-ATRTCNACRTCRCAYTTCATDAT-3′ and 5′-GARAARATGACNCARATNATGTTYGARACNTT-3′. The 495-bp product was amplified under the following conditions: 94°C for 5 min, 32 cycles of 94°C for 1 min, 48°C for 2 min, and 72°C for 2 min, and a final extension at 72°C for 5 min.

For gel blot analysis, RNA was subjected to denaturing electrophoresis on 1% (w/v) agarose gels. After capillary transfer of the RNA to GeneScreen Plus membranes (DuPont–New England Nuclear), the blots were prehybridized and hybridized as described for DNA gel blot analysis. High-stringency washes were performed at 65°C in 0.2 × SSC and 0.1% (w/v) SDS (two washes of 10 min each), and low-stringency washes were performed at 65°C in 2 × SSC and 0.1% (w/v) SDS (two washes of 10 min each). The probes used in RNA gel blot analysis were either DNA probes generated by subcloning appropriate restriction fragments from genomic λ and bacterial artificial chromosome (BAC) clones into the pZero plasmid or probes derived from SRK and SCR cDNAs. The relative amounts of RNA loaded in different lanes were estimated by subsequent probing of the blots with a Brassica actin probe or by visualizing the intensity of rRNA staining with ethidium bromide. Quantitation of signal intensity was performed with a Molecular Dynamics PhosphorImager using the ImageQuant software package (Molecular Dynamics, Sunnyvale, CA).

Construction and Screening of cDNA Libraries

Three cDNA libraries were constructed in the bacteriophage vector λUniZAP: (1) a pistil cDNA library prepared with mRNA isolated from the pistils of an SaSb plant; (2) an anther cDNA library prepared with mRNA isolated from mixed stage anthers of an SaSa homozygote; and (3) an anther cDNA library prepared with mRNA isolated from mixed stage anthers of an SbSb homozygote.

The SaSb pistil library was initially screened at low stringency with a mixture of Brassica S domain probes derived from several class I and class II SLG genes and an A. thaliana probe derived from the S locus–related gene AtS1 (Dwyer et al., 1992). Positive clones that also contained kinase sequences were identified by hybridization with a kinase domain probe derived from the Brassica SRK6 gene. This screen resulted in the isolation of the SRKa sequence. Subsequent screening of the pistil library with a 550-bp XhoI fragment derived from an Sb BAC clone resulted in the isolation of the SRKb sequence.

The SCRa and SCRb cDNA clones were isolated by screening the SaSa and SbSb anther cDNA libraries, respectively, with A. lyrata–derived DNA probes that were generated by subcloning appropriate restriction fragments from genomic λ and BAC clones into the pZero plasmid (Invitrogen) or λDashII (Stratagene, La Jolla, CA), respectively, and subsequent purification of the insert.

Construction and Screening of A. lyrata Genomic λ and BAC Libraries

For the construction of a bacteriophage library, A. lyrata SaSb DNA was partially digested with Sau3A1, and a fraction containing fragments of 9 to 20 kb was cloned into the BamHI site of the λGEM11 vector (Promega, Madison, WI). The library was screened with an S domain probe derived from the SRKa cDNA to isolate SRKa-containing clones.

A large insert library of A. lyrata SaSb DNA was constructed in the pBeloBAC vector according to Woo et al. (1994). Because of the existence of a large family of genes related to SRK in crucifer genomes, the isolation of BAC clones for the A. lyrata S locus required the generation of single-copy molecular markers from the region. We searched the A. thaliana BAC T6K22 sequence for sequences that were apparently unique in the A. thaliana genome. One such sequence, T6K22.150, was amplified from BAC T6K22 by using the gene-specific primers 5′-TTGCTTATCGCCGTGGTTCC-3′ and 5′-AGCGACTCGATTTCCTTCTCTGTA-3′. Subsequent gel blot analysis of A. lyrata DNA demonstrated that the T6K22.150 probe produced a relatively strong hybridization signal and appeared to be a single- copy sequence in A. lyrata (data not shown). This probe was used to screen the A. lyrata BAC library. A total of 15 BAC clones were identified that, upon further analysis, were classified into two groups, one group of six clones derived from the Sa haplotype and another group of nine clones derived from the Sb haplotype.

Sequence Analysis and Database Searches

DNA sequencing was performed using an automated sequencer at the sequencing facility of the Cornell University Bioresource Center. Sequences were manipulated and aligned using DNASTAR Lasergene software (DNASTAR Inc., Madison, WI). BLAST searches were performed on the National Center for Biotechnology Information Internet site (http://www.ncbi.nlm.nih.gov). Sequences were aligned using the MegAlign feature of DNASTAR. Parsimony analyses were conducted on the aligned cDNAs or derived protein sequences by using the branch and bound algorithm of PAUP* 4.0b2 (Swofford, 1999).

Construction of Reporter Gene Fusions, Plant Transformation, and Analysis of Transgenic Plants

A DNA fragment spanning a region from 23 to ~1600 bp upstream of the putative initiating methionine codon of the T6K22.100 (ΨSRK) sequence was generated by PCR amplification and two primers, 5′-GGATCCGAGATTAGGTTTGTACATG-3′ and 5′-GGAAAATGAACAACATATGAGGATCC-3′, which incorporated BamHI recognition sites to facilitate cloning. The amplified product was TA cloned into the pCRII vector (Invitrogen) and digested with BamHI. The promoter insert was excised from this vector by BamHI digestion and inserted into the BamHI site of the β-glucuronidase (GUS) expression vector pBI101 (Jefferson et al., 1987) 5′ of the uidA gene, which encodes the GUS protein. A plasmid containing the promoter fragment in the appropriate orientation was selected and introduced into Agrobacterium tumefaciens strain pCIB542/A136 (derived from helper plasmid EHA101; Hood et al., 1986). Transformation of A. thaliana strain C24 (obtained from M. Jacobs, Vrije Universiteit, Brussels, Belgium) with the ΨSRK::uida::nos chimeric gene was performed according to Valvekens et al. (1988). Transgenic plants were selected on the basis of kanamycin resistance, and the stable integration of the transgene was confirmed by DNA gel blot analysis in primary transformants and in progeny plants generated by allowing the plants to self-pollinate.

The ΨSRK::uida::nos primary transformants and their T2 progeny were analyzed for GUS activity by the histochemical method using the chromogenic substrate 5-bromo-4-chloro-3-indolyl glucuronide (Jefferson et al., 1987), as described previously (Toriyama et al., 1991).

Acknowledgments

We thank Charles Langley (University of California, Davis) for providing the A. lyrata seed, Jeffrey Doyle (Cornell University, Ithaca, NY) for assistance with the phylogenetic analysis, M.K. Kandasamy (University of Georgia, Athens) for actin primers, and Bela Kudish and Michelle Battle for work related to the construction of the ΨSRK reporter gene and plant transformation. We thank the Arabidopsis Biological Resources Center for providing BAC T6K22. This work was supported by Grant No. IBN-0077289 from the National Science Foundation.

References

Al-Shehbaz, I.A. (1984). The tribes of the Cruciferae (Brassicaceae) in the southeastern United States. J. Arnold Arbor. Harv. Univ. 65:, 343–373.
Bateman, A.J. (1954). Self-incompatibility systems in angiosperms. II. Iberis amara. Heredity 8:, 305–332.
Bateman, A.J. (1955). Self-incompatibility systems in angiosperms. III. Cruciferae. Heredity 9:, 52–68.
Bennetzen, J.L. (1996). The contributions of retroelements to plant genome organization, function and evolution. Trends Microbiol. 4:, 347–353. [PubMed].
Boyes, D.C., and Nasrallah, J.B. (1993). Physical linkage of the SLG and SRK genes at the self-incompatibility locus of Brassica oleracea. Mol. Gen. Genet. 236:, 369–373. [PubMed].
Boyes, D.C., and Nasrallah, J.B. (1995). An anther-specific gene encoded by an S locus haplotype of Brassica produces complementary and differentially regulated transcripts. Plant Cell 7:, 1283–1294. [PubMed].
Boyes, D.C., Nasrallah, M.E., Vrebalov, J., and Nasrallah, J.B. (1997). The self-incompatibility (S) haplotypes of Brassica contain highly divergent and rearranged sequences of ancient origin. Plant Cell 9:, 1–12.
Cabrillac, D., Delorme, V., Garin, J., Ruffio-Chable, V., Giranton, J.-L., Dumas, C., Gaude, T., and Cock, M.J. (1999). The S15 self-incompatibility haplotype in Brassica oleracea includes three S gene family members expressed in stigmas. Plant Cell 11:, 971–986. [PubMed].
Casselman, A.L., Vrebalov, J., Conner, J.A., Singhal, A., Giovannoni, J., Nasrallah, M.E., and Nasrallah, J.B. (2000). Determining the physical limits of the Brassica S locus by recombinational analysis. Plant Cell 12:, 23–33. [PubMed].
Chen, C.-H., and Nasrallah, J.B. (1990). A new class of S-sequences defined by a pollen recessive self-incompatibility allele of Brassica oleracea. Mol. Gen. Genet. 222:, 241–248. [PubMed].
Conner, J.A., Tantikanjana, T., Stein, J.C., Kandasamy, M.K., Nasrallah, J.B., and Nasrallah, M.E. (1997). Transgene-induced silencing of S locus genes and related genes in Brassica. Plant J. 11:, 809–823.
Conner, J.C., Conner, P., Nasrallah, M.E., and Nasrallah, J.B. (1998). Comparative mapping of the Brassica S-locus region and its homeologue in Arabidopsis: Implications for the evolution of mating systems in the Brassicaceae. Plant Cell 10:, 801–812. [PubMed].
Cui, Y.H., Brugiere, N., Jackman, L., Bi, Y.M., and Rothstein, S.J. (1999). Structural and transcriptional comparative analysis of the S locus regions in two self-incompatible Brassica napus lines. Plant Cell 11:, 2217–2231. [PubMed].
Cui, Y.H., Bi, Y.M., Brugiere, N., Arnoldo, M.A., and Rothstein, S.J. (2000). The S locus glycoprotein and the S receptor kinase are sufficient for self-pollen rejection in Brassica. Proc. Natl. Acad. Sci. USA 97:, 3713–3717. [PubMed].
de Nettancourt, D. (1977). Incompatibility in Angiosperms: Monographs on Theoretical and Applied Genetics 3. (Berlin: Springer-Verlag).
Dixit, R., Nasrallah, M.E., and Nasrallah, J.B. (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. [PubMed].
Dwyer, K.G., Balent, M.A., Nasrallah, J.B., and Nasrallah, M.E. (1991). DNA sequences of self-incompatibility genes from Brassica campestris and B. oleracea: Polymorphism predating speciation. Plant Mol. Biol. 16:, 481–486. [PubMed].
Dwyer, K.G., Lalonde, B.A., Nasrallah, J.B., and Nasrallah, M.E. (1992). Structure and expression of AtS1, an Arabidopsis thaliana gene homologous to the S-locus related genes of Brassica. Mol. Gen. Genet. 231:, 442–448. [PubMed].
Dwyer, K.G., Kandasamy, M.K., Mahosky, D.I., Acciai, J., Kudish, B.I., Miller, J.E., Nasrallah, M.E., and Nasrallah, J.B. (1994). A superfamily of S locus-related sequences in Arabidopsis: Diverse structures and expression patterns. Plant Cell 6:, 1829–1843. [PubMed].
Ferris, P.J., and Goodenough, U.W. (1994). The mating-type locus of Chlamydomonas reinhardtii contains highly rearranged DNA sequences. Cell 76:, 1135–1145. [PubMed].
Gaude, T., Rougier, M., Heizmann, P., Ockendeon, D.J., and Dumas, C. (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. [PubMed].
Giranton, J.L., Ariza, M.J., Dumas, C., Cock, J.M., and Gaude, T. (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. [PubMed].
Hood, E.E., Helmer, G.L., Fraley, R.T., and Chilton, M.D. (1986). The hypervirulence of Agrobacterium tumefaciens A281 is encoded in a region of pTiBo542 outside of T-DNA. J. Bacteriol. 168:, 1291–1301. [PubMed].
Jefferson, R.A., Kavanaugh, T.A., and Bevan, M.W. (1987). GUS fusions: β-Glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6:, 3901–3907. [PubMed].
Kandasamy, M.K., Paolillo, D.J., Faraday, C.D., Nasrallah, J.B., and Nasrallah, M.E. (1989). The S-locus specific glycoproteins of Brassica accumulate in the cell wall of developing stigma papillae. Dev. Biol. 134:, 462–472. [PubMed].
Kho, Y.O., and Baer, J. (1968). Observing pollen tubes by means of fluorescence. Euphytica 17:, 298–302.
Koch, M., Bishop, J., and Mitchell-Olds, T. (1999). Molecular systematics and evolution of Arabidopsis and Arabis. Plant Biol. 1:, 529–537.
Kusaba, M., and Nishio, T. (1999). Comparative analysis of S haplotypes with very similar SLG alleles in Brassica rapa and Brassica oleracea. Plant J. 17:, 83–91. [PubMed].
Kusaba, M., Nishio, T., Satta, Y., Hinata, K., and Ockendon, D.J. (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. [PubMed].
Luporini, P., Vallesi, A., Miceli, C., and Bradshaw, R.A. (1995). Chemical signaling in ciliates. J. Eukaryot. Microbiol. 42:, 208–212. [PubMed].
Murray, M.G., and Thompson, W.F. (1980). Rapid isolation of high molecular weight plant DNA. Nucleic Acids Res. 8:, 4321–4325. [PubMed].
Nasrallah, J.B. (2000). Cell-cell signaling in the self-incompatibility response. Curr. Opin. Plant Biol. 3:, 368–373. [PubMed].
Nasrallah, J.B., Kao, T.-H., Chen, C.-H., Goldberg, M.L., and Nasrallah, M.E. (1987). Amino-acid sequence of glycoproteins encoded by three alleles of the S locus of Brassica oleracea. Nature 326:, 617–619.
Nishio, T., and Kusaba, M. (2000). Sequence diversity of SLG and SRK in Brassica oleracea L. Ann. Bot. 85 (suppl. A):, 141–146.
O'Kane, S.L., and Al-Shehbaz, I.A. (1997). A synopsis of Arabidopsis (Brassicaceae). Novon 7:, 323–327.
Okazaki, K., Kusaba, M., Ockendon, D.J., and Nishio, T. (1999). Characterization of S tester lines in Brassica oleracea: Polymorphism of restriction fragment length of SLG homologues and isoelectric points of S-locus glycoproteins. Theor. Appl. Genet. 98:, 1329–1334.
Pastuglia, M., Roby, D., Dumas, C., and Cock, J.M. (1997a). Rapid induction by wounding and bacterial infection of an S gene family receptor-like kinase in Brassica oleracea. Plant Cell 9:, 1–13.
Pastuglia, M., Ruffio-Chable, V., Delorme, V., Gaude, T., Dumas, C., and Cock, J.M. (1997b). A functional S locus anther gene is not required for the self-incompatibility response in Brassica oleracea. Plant Cell 9:, 2065–2076. [PubMed].
Price, R.A., Palmer, J.D., and Al-Shehbaz, I.A. (1994). Systematic relationships of Arabidopsis: A molecular and morphological perspective. In Arabidopsis, C.R. Somerville and E.M. Meyerowitz, eds (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), pp. 7–19.
Sakamoto, K., Kusaba, M., and Nishio, T. (1998). Polymorphism of the S-locus glycoprotein gene (SLG) and the S-locus related gene (SLR1) in Raphanus sativus L. and self-incompatible ornamental plants in the Brassicaceae. Mol. Gen. Genet. 258:, 397–403. [PubMed].
Sato, T., Thorsness, M.K., Kandasamy, M.K., Nishio, T., Hirai, M., Nasrallah, J.B., and Nasrallah, M.E. (1991). Activity of an S-locus gene promoter in pistils and anthers of transgenic Brassica. Plant Cell 3:, 867–876. [PubMed].
Schierup, M.H. (1998). The effect of enzyme heterozygosity on growth in a strictly outcrossing species, the self-incompatible Arabis petraea (Brassicaceae). Hereditas 128:, 21–31.
Schierup, M.H., Mable, B., Awadalla, P., and Charlesworth, D. (2001). Identification and characterization of a polymorphic receptor kinase gene linked to the self-incompatibility locus of Arabidopsis lyrata. Genetics (in press):.
Schopfer, C.R., and Nasrallah, J.B. (2000). Self-incompatibility: Prospects for a novel putative peptide-signaling molecule. Plant Physiol. 124:, 935–940. [PubMed].
Schopfer, C.R., Nasrallah, M.E., and Nasrallah, J.B. (1999). The male determinant of self-incompatibility in Brassica. Science 286:, 1697–1700. [PubMed].
Schultz, O.E. (1936). Cruciferae. In Die Naturlichen Pflanzenfamilien, Vol. 17b, 2nd ed, A. Engler and H. Harms, eds (Leipzig, Germany: Engelmann), pp. 227–658.
Stein, J.C., Howlett, B., Boyes, D.C., Nasrallah, M.E., and Nasrallah, J.B. (1991). Molecular cloning of a putative receptor protein kinase gene encoded at the self-incompatibility locus of Brassica oleracea. Proc. Natl. Acad. Sci. USA 88:, 8816–8820. [PubMed].
Stein, J.C., Dixit, R., Nasrallah, M.E., and Nasrallah, J.B. (1996). SRK, the stigma-specific S locus receptor kinase of Brassica, is targeted to the plasma membrane in transgenic tobacco. Plant Cell 8:, 429–445. [PubMed].
Suzuki, G., Kai, N., Hirose, T., Fukui, K., Nishio, T., Takayama, S., Isogai, A., Watanabe, M., and Hinata, K. (1999). Genomic organization of the S locus: Identification and characterization of genes in SLG/SRK region of S9 haplotype of Brassica campestris (syn. rapa). Genetics 153:, 391–400. [PubMed].
Swofford, D.L. (1999). PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4.0. (Sunderland, MA: Sinauer Associates).
Takahata, Y., and Hinata, K. (1980). A variation study of subtribe Brassicineae by principal component analysis. In Brassica Crops and Wild Allies: Biology and Breeding, S. Tsunoda, K. Hinata, and C. Gomez-Campo, eds (Tokyo: Japan Science Societies Press), pp. 33–49.
Takasaki, T., Hatakeyama, K., Suzuki, G., Watanabe, M., Isogai, A., and Hinata, K. (2000). The S receptor kinase determines self-incompatibility in Brassica stigma. Nature 403:, 913–916. [PubMed].
Takayama, S., Shiba, H., Iwano, M., Shimosato, H., Che, F.S., Kai, N., Watanabe, M., Suzuki, G., Hinata, K., and Isogai, A. (2000). The pollen determinant of self-incompatibility in Brassica campestris. Proc. Natl. Acad. Sci. USA 97:, 1920–1925. [PubMed].
Tantikanjana, T., Nasrallah, M.E., Stein, J.C., Chen, C.-H., and Nasrallah, J.B. (1993). An alternative transcript of the S locus glycoprotein gene in a class II pollen-recessive self-incompatibility haplotype of Brassica oleracea encodes a membrane-anchored protein. Plant Cell 5:, 657–666. [PubMed].
Tantikanjana, T., Nasrallah, M.E., and Nasrallah, J.B. (1996). The Brassica S gene family: Molecular characterization of the SLR2 gene. Sex. Plant Reprod. 9:, 107–116.
Thompson, K.F., and Taylor, J.P. (1966). Non-linear dominance relationships between S alleles. Heredity 21:, 345–362.
Tobias, C.M., and Nasrallah, J.B. (1996). An S-locus-related gene in Arabidopsis encodes a functional kinase and produces two classes of transcripts. Plant J. 10:, 523–531. [PubMed].
Tobias, C.M., Howlett, B., and Nasrallah, J.B. (1992). An Arabidopsis thaliana gene with sequence similarity to the S-locus receptor kinase gene of Brassica oleracea. Plant Physiol. 99:, 284–290. [PubMed]
Toriyama, K., Thorsness, M.K., Nasrallah, J.B., and Nasrallah, M.E. (1991). An S-locus gene promoter directs sporophytic expression in the anther tapetum of transgenic Arabidopsis. Dev. Biol. 143:, 427–431. [PubMed].
Uyenoyama, M.K. (1995). A generalized least-squares estimate for the origin of sporophytic self-incompatibility. Genetics 139:, 975–992. [PubMed].
Uyenoyama, M.K. (2000). A prospectus for new developments in the evolutionary theory of self-incompatibility. Ann. Bot. 85 (suppl. A):, 247–252.
Vacquier, V.D. (1998). Evolution of gamete recognition proteins. Science 281:, 1995–1998. [PubMed].
Valvekens, D., Van Montagu, M., and Van Lijsebettens, M. (1988). Agrobacterium tumefaciens–mediated transformation of Arabidopsis thaliana root explants by using kanamycin selection. Proc. Natl. Acad. Sci. USA 85:, 5536–5540. [PubMed]
van Treuren, R., Kuitteinen, H., Karkkainen, K., Baena-Gonzalez, E., and Savolainen, O. (1997). Evolution of microsatellites in Arabis petraea and Arabis lyrata, outcrossing relatives of Arabidopsis thaliana. Mol. Biol. Evol. 14:, 220–229. [PubMed].
Watanabe, M.A., et al. (2000). Highly divergent sequences of the pollen self-incompatibility (S) gene in class-I S haplotypes of Brassica campestris (syn. rapa) L. FEBS Lett. 473:, 139–144. [PubMed].
Woo, S.-S., Jiang, J., Gill, B.S., Patterson, A.H., and Wing, R.A. (1994). Construction and characterization of a bacterial artificial chromosome library of Sorghum bicolor. Nucleic Acids Res. 22:, 4922–4931. [PubMed].
Yang, Y.W., Lai, K.N., Tai, P.Y., Ma, D.P., and Li, W.H. (1999). Rates of nucleotide substitution in angiosperm mitochondrial DNA sequences and dates of divergence between Brassica and other angiosperm lineages. J. Mol. Evol. 48:, 597–604. [PubMed].