S-RNase and SLF Determine S-Haplotype–Specific Pollen Recognition and Rejection

Self-incompatibility (SI) is a genetically determined system for recognition and rejection of self-pollen and pollen from closely related plants. Recognition specificity is controlled by the S-locus. This locus is complex in the sense that it contains multiple genes: one controls specificity on the pistil side (pistil S), another on the pollen side (pollen S). It is also highly polymorphic; the S-specificity genes in different haplotypes have highly divergent sequences, giving rise to a set of proteins having a similar function but distinctive recognition characteristics. S-RNase–based SI occurs in the Solanaceae, Rosaceae, and Scrophulariaceae. S-RNase itself controls specificity on the pistil side. S-RNases were discovered in the mid-1980s (Bredemeijer and Blaas, 1981; Anderson et al., 1986), but it took eight more years to provide direct evidence that they are the genes that determine S-specificity in the pistil (Lee et al., 1994; Murfett et al., 1994). It has recently been determined that SLF (S-locus F-box) genes control specificity on the pollen side. SLF genes were first reported in the Scrophulariaceae (Lai et al., 2002). At first, SLF was not thought to be sufficiently polymorphic to be pollen S, the determinant of S-specificity in pollen. However, highly polymorphic SLF genes were described in Rosaceae in 2003 and early 2004 (Entani et al., 2003; Ushijima et al., 2003; Ikeda et al., 2004). Recent months have seen publication of transformation experiments definitively showing that SLF determines specificity on the pollen side in Petunia inflata (Sijacic et al., 2004). These results are supported by transformation experiments showing that the Antirrhinum SLF2 gene is implicated in SI, although it was not possible to directly address S-specificity in this system (Qiao et al., 2004b). The compressed time scale between identification of SLF and definitive proof that it is pollen S is a testament to the hard work and skill of the laboratories involved in research in all three S-RNase–based SI systems. The identification of the S-specificity determining genes on both the pollen and pistil sides brings closure to the first phase of research into the molecular basis of S-RNase–based SI. This is an excellent time to evaluate our understanding of the system, as research moves toward a focus on the biochemistry and physiology of pollen recognition and rejection. This essay describes only the most recent results related to identification of pollen S and highlights select facts that will shape research in the near future. For a more comprehensive review, please see Kao and Tsukamoto (2004). S-RNase–based SI is widely distributed among the angiosperms (Igic and Kohn, 2001; Kao and Tsukamoto, 2004). Fundamentally, it is a mechanism for controlling pollination. It can be described as genetic interaction between the sporophyte and the gametophyte that prevents inbreeding. It is a typical single-locus gametophytic SI system—compatibility is determined by the S-haplotype of pollen, the haploid male gametophyte. Pollen is rejected when its S-haplotype is the same as either of the S-haplotypes in the diploid pistil. Conversely, pollen with any S-haplotype not present in the pistil is compatible (de Nettancourt, 2001). This is the origin of the cumbersome but necessary term S-specific pollen rejection (or S-allele–specific pollen rejection). S-RNase is also implicated in interspecific pollen rejection, but the mechanisms are distinct from SI (Murfett et al., 1996). Thus, S-RNase–based systems are the most widespread and diverse pollen rejection systems known. The genetics of SI are simple. The challenge is to understand how the genetic interaction between the pollen and the pistil is manifested at the biochemical level. Necessarily, the first priority was to identify the biochemical determinants of S-specificity. The approaches used had to be tied to the requirements imposed by the biology and genetics of SI. Putative specificity determinants must meet three criteria: linkage to the S-locus, polymorphism between different S-haplotypes, and expression in the pollen or pistil. S-RNases, the determinants of S-specificity in the pistil, were identified by searching for proteins that met these three requirements. The principle was simple: perform controlled crosses to determine the S-specific pollen rejection phenotypes of a family of plants and analyze for cosegregating proteins. In practice, however, this was difficult. It took years to identify plant materials with protein bands that clearly cosegregated with S-specific pollen rejection (Bredemeijer and Blaas, 1981). Anderson et al. (1986)ultimately obtained N-terminal sequence from a glycoprotein expressed in the Nicotiana alata style that cosegregated with the ability to reject S2-pollen. Numerous similar clones soon were obtained from other S-haplotypes, species, and families. Sequence analysis led to the discovery that these polymorphic proteins have ribonuclease activity; thus, they were labeled S-RNases (McClure et al., 1989). It took eight additional years to show definitively that S-RNase genes met the requirements for the determinants of S-specificity in the pistil. Linkage to the S-locus was determined by genetic analyses showing no recombination between S-RNase genes and the pollen rejection phenotype (Murfett et al., 1994). S-RNase gene sequences, and the surrounding genomic regions, proved to be highly polymorphic between different S-haplotypes (Anderson et al., 1989; Ioerger et al., 1990, 1991). S-RNase expression was restricted to the pistil, and the proteins were secreted into the extracellular matrix forming the path from the stigma to the ovary (Cornish et al., 1987; Anderson et al., 1989). Although these data provided strong correlative evidence that S-RNase was the determinant of S-specificity in the pistil, they were not definitive. Articles published in 1994 provided direct evidence that S-RNase is the product of the pistil S gene. They showed that expression of the N. alata SA2-RNase gene or the P. inflata S3-RNase gene caused transgenic plants to gain the ability to reject SA2- or S3-pollen, respectively (Lee et al., 1994; Murfett et al., 1994). The basic mechanism of S-RNase–based SI was outlined in the early 1990s. Gray et al. (1991)showed that S-RNase can gain access to the pollen cytoplasm and act as a potent translational inhibitor. A tracer experiment using 32P-labeled pollen showed that pollen RNA is degraded after incompatible but not compatible pollinations (McClure et al., 1990). Finally, Huang et al. (1994)showed that ribonuclease activity is required for pollen rejection. The basic model that emerged is that each S-haplotype encodes a unique S-RNase protein that is secreted into the transmitting tract extracellular matrix. An interaction between S-RNase and the pollen-side specificity factor, pollen S protein, determines whether the pollination is compatible or incompatible. In incompatible pollinations, S-RNase gains access to the pollen tube cytoplasm where its ribonuclease activity causes a cytotoxic effect; in a compatible pollination, this cytotoxic effect is evaded. The search for the pollen S gene began shortly after the S-RNase gene was cloned. It quickly became obvious that the pollen S protein is not as abundant as S-RNase and different approaches would be needed. Ultimately, pollen S gene candidates were successfully identified by large-scale genomic DNA sequencing in the vicinity of S-RNase genes. This approach succeeded because it was tied directly to the three strict requirements mentioned earlier: candidates must be linked to the S-RNase gene, highly polymorphic, and expressed in pollen. However, proving that a candidate gene is pollen S was not a trivial matter because the only known phenotypes associated with it are pollen tube growth inhibition in SI and a subtle genetic interaction between S-haplotypes that occurs in heteroallelic pollen. The tight genetic linkage between pollen S and S-RNase provided an obvious route to identifying candidates. However, SI species are not always well suited for genetic experiments, and various SI systems have different strengths and weaknesses. Essential contributions were made by researchers working in Antirrhinum, Nicotiana, Petunia, and Prunus (reviewed in Kao and Tsukamoto, 2004). In Nicotiana and Petunia, historically the best systems for S-RNase studies, a great deal of effort was required to accumulate a collection of pollen-expressed markers linked to S-RNase genes (Li et al., 2000; McCubbin et al., 2000a, 2000b; Wang et al., 2003). In spite of the difficulties, Wang et al. (2003)mapped 13 markers and S-RNase in a large P. inflata population segregating for S1- and S2-haplotypes. Nine markers and the S-RNase gene occurred in a 0.25-centimorgan region and could not be separated by recombination. BAC clones were isolated to allow construction of a physical map. Although contigs amounting to 4.4 Mb were constructed from the S2-haplotype, only one contig contained more than a single marker, revealing that the S-locus is physically very large in Petunia. Ultimately, the pollen expressed genes closest to the P. inflata S2-RNase gene were identified in an 881-kb contig by cDNA selection and direct sequencing of 328 kb surrounding the S2-RNase gene (Wang et al., 2004). In other species, the S-locus proved to be more compact. Markers defining the boundaries of the S-locus in P. dulcis turned out to be only 72 kb apart (Ushijima et al., 2001). In Antirrhinum and Prunus mume, genomic contigs spanning only ∼65 kb proved to contain pollen S candidates (Lai et al., 2002; Entani et al., 2003). Genes that are physically close to the S-RNase gene all fulfill the first requirement for a pollen S candidate; they are linked to the S-locus. Applying the additional requirements of polymorphism and expression in pollen allowed the best candidates to emerge. Results from genomic sequencing studies showed that the S-locus region contains multiple F-box protein genes. Unfortunately, expression in pollen is common; therefore, this is not a discriminating test for pollen S candidates. Sequence polymorphism proved to be a better way to distinguish among candidates. The polymorphic F-box protein genes so identified (i.e., pollen S) are referred to as SLFx (S-locus F-box from haplotype X) genes in Antirrhinum, Petunia, and Prunus mume by some authors (Lai et al., 2002; Qiao et al., 2004b, 2004a; Sijacic et al., 2004), others refer to them as SFB genes (Ushijima et al., 2003, 2004; Ikeda et al., 2004). The Antirrhinum AhSLF2 gene was the first SLF gene reported (Lai et al., 2002). It was shown to be expressed specifically in pollen, but the polymorphism was not as great as was expected for pollen S. AhSLF2 probes revealed restiction fragment length polymorphisms in different S-haplotypes but that the degree of cross-hybridization was high, suggesting less sequence polymorphism than S-RNase genes (Lai et al., 2002). Studies in Prunus strongly pointed to SLF genes as the best pollen S candidates because their sequence polymorphism is comparable to S-RNase. This set of Prunus pollen S candidates (referred to as SLF in P. mume and SFB in P. dulcis) showed S-haplotype–specific hybridization similar to obtained with S-RNase probes (Entani et al., 2003; Ushijima et al., 2003). sequence analyses of 13 of these genes showed a of and with a in recognition as well as evidence of selection in et al., 2004). Ushijima et al. Prunus having that only S-haplotype specificity in pollen pistil The P. expressed a gene present in the gene. In P. mume, the contained an gene with a large that the would be as for selection and the of of SFB genes in strongly a between SLF and pollen S. the data from Antirrhinum and Prunus provided strong correlative evidence (i.e., close linkage to expression in pollen, and polymorphism between that SLF (or and pollen S were the same gene. was a The definitive test is a transformation experiment an SLF gene to be directly for an S-haplotype–specific effect on pollination. studies in P. inflata have provided definitive proof that one of the F-box protein genes is pollen S (Sijacic et al., 2004), and a recent in Antirrhinum also directly implicated an F-box protein gene in SI but not address haplotype specificity (Qiao et al., 2004b). These studies of pollen S candidates are more subtle than to plant genes. A of how pollen in S-RNase–based systems and how S-specificity is expressed in pollen is to understand the transformation in the to a gene in SI and to show that it determines S-specificity. the that S-haplotype–specific pollen rejection is the defining of SI, it is to the that to the cytotoxic effect of S-RNase is also in S-RNase–based SI. For although is to the cytotoxic effect of it is also to all other Thus, although a pollen S-haplotype is rejected by one it is to all The to this some of S-RNase and the transformation to test pollen S candidates were strongly by these and Tsukamoto, 2004). The model the S-specific recognition function (i.e., the genetic of pollen S) and the S-RNase function into a single A model these in pollen S determines and a prevents the cytotoxic activity of S-RNases et al., 2001). The function of pollen S is supported by studies that that of the pollen S gene is to the pollen. et al. in N. the could be by of all or of the S-locus that pollen became The of with of pollen S in the et al. is it is an gene. However, the Ushijima et al. a of Prunus that at one of pollen S is the not to have pollen S The of both and the N-terminal (i.e., F-box of the protein, the were as is these as they the ability to other S-haplotypes but they no show rejection by their S-RNase. It is possible to that the N-terminal of the protein in to S-RNase (i.e., and the in S-haplotype rejection. This is with the but the model is not because it is possible that a The Prunus results (Ushijima et al., the for a of pollen S function in Petunia or where in and transformation can be However, is only the of and inhibition of definitively have been shown to in the same Although the of how pollen the of are not and the different that pollen S transformation experiments had to for the that pollen S is an gene (i.e., it function to the cytotoxic of In or RNA approaches were not Thus, Sijacic et al. Qiao et al. pollen S candidates into pollen a pollen S gene. This is by the in pollen S genes when different S-haplotypes are expressed in the haploid pollen pollen from of the S-locus, as described et al., or in an SI plant is to an SI is (de Nettancourt, The only the pistil to reject S1- and S2-pollen. only occurs in diploid (i.e., and pollen is rejected on both the diploid or pistil in Pollen The S-haplotypes of and are Pollen S-haplotypes by these are shown in with SLF where haploid In gametophytic SI, pollen is rejected in a but it is to the of all other S-RNases (i.e., An is SI occurs when different pollen S-haplotypes (i.e., heteroallelic are An plant with an SLF2 Pollen the SLF2 in SI similar to heteroallelic pollen. S-specific An plant with an SLF2 pollen the the SLF2 with the of SI the with the S2-haplotype, pollen and is similar to pollen from the The experiment S-specificity because occurs when the occurs with one haplotype but not the in Pollen The S-haplotypes of and are Pollen S-haplotypes by these are shown in with SLF where haploid In gametophytic SI, pollen is rejected in a but it is to the of all other S-RNases (i.e., An is SI occurs when different pollen S-haplotypes (i.e., heteroallelic are An plant with an SLF2 Pollen the SLF2 in SI similar to heteroallelic pollen. S-specific An plant with an SLF2 pollen the the SLF2 with the of SI the with the S2-haplotype, pollen and is similar to pollen from the The experiment S-specificity because occurs when the occurs with one haplotype but not the Sijacic et al. Qiao et al. the heteroallelic pollen effect to test pollen S candidates. However, because of the way specificity is are to a candidate in SI and to a candidate pollen S gene in SI pollen with a different S-haplotype (i.e., heteroallelic will in pollen that be This approach can a candidate in SI, but it is S-specificity is only with the additional test of the candidate in with the same It is that pollen S transformation results in the only one after it became from work in Antirrhinum, Petunia, and Prunus that SLF genes were the best candidates for pollen S (Lai et al., 2002; Entani et al., 2003; Ushijima et al., 2003; Kao and Tsukamoto, 2004; Wang et al., 2004). In P. the gene was into an (Sijacic et al., 2004). The SI on the pollen was as function was and only occurred in the Qiao et al. similar experiments with the Antirrhinum AhSLF2 gene. AhSLF2 gene were into SI P. The Antirrhinum gene as and SI on the pollen side was Thus, in SLF expressed in with a different SLF gene (i.e., the SLF gene present at the caused a in SI on the pollen side. This is with on the of the N. alata and the heteroallelic pollen effect However, as the heteroallelic pollen effect is any different pollen S genes SI to of SI would be expected after any gene (i.e., other than pollen S) implicated in pollen SI S-specificity can best be by the of SI in the one where the SLF is from the same S-haplotype as the SLF gene. Sijacic et al. also the gene into an of function was However, all the were either or and the This is as SI when the is expressed in pollen with the The of that the effect of the is S-haplotype it only with other S-haplotypes (i.e., or and has no effect on the Qiao et al. the AhSLF2 gene into P. to test its function in SI. it is not possible to the effect (or of of the on the Antirrhinum S2-haplotype, this approach for S-specificity. the S-haplotype–specific effect of is definitive. SLF is pollen S. The most in S-RNase–based SI are the is at a will toward understanding the biochemistry and physiology of S-specific pollen rejection. only the were known. At the pistil S-RNase is secreted into the transmitting tract extracellular at the other pollen RNA is degraded or in or of pollen S as an F-box protein that the specificity the interaction between pollen S and pistil S is linked to F-box proteins the in the F-box) complex that proteins for et al., 2003). The complex the in of a on proteins that for the F-box proteins can act as specificity to the complex by proteins additional proteins also be with distinct F-box proteins are with a as have an large of F-box protein genes et al., 2002; and 2004), and has been linked to a of SI et al., 2003). have that as an of by it for (Qiao et al., Sijacic et al., 2004; Ushijima et al., 2004). This model is and well be but other are F-box proteins apart from protein For the F-box protein to of growth and is required for proteins et al., 2001). has been to as a it has in and DNA 2001; and 2004). protein is the in S-RNase–based SI, it must be determined the model is or whether a more complex model is needed. to test of this model have been The S-specificity by an interaction between SLF and S-RNase. Qiao et al. and results that show a direct interaction between AhSLF2 and S-RNase and evidence that AhSLF2 a complex with and proteins. Thus, SLF with S-RNase and in an is no biochemical evidence for an S-specific AhSLF2 both and as well as S3-RNase from P. (Qiao et al., 2004b). the interaction between S-RNase and SLF determines S-specificity Thus, either the between SLF and have different of an or the for S-specific interaction have not been research will these Although these experiments a some of the by Qiao et al. to be in other systems they are For a of the model is that of S-RNase after compatible but not incompatible pollination. Qiao et al. this by the of S-RNase in style after and and that S-RNase after compatible pollination. However, this has not been in other studies showed large of S-RNase both compatible and incompatible pollen et al., Gray et al. no evidence for S-RNase using S2-RNase and in N. alata pollen The (Qiao et al., that S-RNase is the of the complex and that it is degraded by the are of but it is also to be and test them in other systems. An complex not be the only that to S-RNase. and a protein from P. that S-RNase. the is expressed in only has been shown to to S-RNase in the It is not known whether the is directly involved in SI or whether it with The it be that Qiao et al. experiments the biochemistry and physiology pollen rejection. studies are and it is very to their experiments to other systems. the of pistil S and pollen S and can to on to studies of the biochemistry and physiology of S-RNase–based SI. Sijacic et al. that SLF is the pollen S gene, and Qiao et al. that SLF S-RNase. The is for the specificity of this interaction using and studies of the S-RNase and SLF genes. The possible of genes at the S-locus be of the of will determine whether the model of to be candidates and required for pollen rejection, as well as that are as Studies of and the origin of specificity are obvious research for the future. However, they will not provide a comprehensive model of S-RNase–based SI. are of S-RNase–based SI of are are also facts that are out of of SI. It is to some of these to the so that they can to a more comprehensive The mechanism of S-RNase is an of tract S-RNase into the extracellular matrix. SLF is very to be a the S-specificity where SLF and S-RNase occurs in the pollen tube This is also the of S-RNase because incompatible pollen tube RNA is degraded Thus, it is to the route from the extracellular matrix to the The only of S-RNase describes of both and into pollen et al., S-RNase was reported to be in the pollen tube However, its is to because were to than and markers for were not Although pollen tube growth is associated with of at the is also a and pollen are in this is the route of S-RNase It is that and are by and 2003). proteins be or as a for into and is for of an F-box protein in et al., 2001). S-RNase is into a it would not have access to either SLF or RNA in the Thus, the of whether S-RNase is into a or whether it gains access directly to the cytoplasm is an that to be It is also how the specificity in SI is manifested at the biochemical level. S-RNase experiments are but they are in to of SI. et al. experiments in of SA2- and from N. alata were Nine S-RNases were although all were were of S-specific pollen rejection. Thus, the S-RNases were of as but could not be Kao and McCubbin similar results in P. one of the N. alata was shown to be in an interspecific pollen rejection system and and the evidence that S-RNase is it is that the to pollen rejection because they are not by pollen At a the experiments show that the S-specificity function of S-RNase is more than its they also that S-specificity and cytotoxic activity are on the pollen side. In the of pollen a that to Conversely, recognition of by SLF could either a to or a Finally, although it is known that other than the specificity determinants are required for SI, it is not known how or how they and protein are of S-RNase–based SI, any is required for pollen rejection is a possible is the only other than S-RNase and SLF that has been identified with (McClure et al., It is a protein secreted into the transmitting tract matrix in Nicotiana, and (McClure et al., et al., et al., 2002). and RNA experiments show that is required for pollen rejection, but its function is (McClure et al., et al., 2002). on both the pollen and pistil side of the SI For for A protein that S-RNase has been but it is not known whether it has a function in SI et al., better of the required for S-RNase–based pollen rejection, it is not possible to a comprehensive of the pollen S gene is the most in S-RNase–based SI in recent have a from to but are more than is the of inhibition of access to the or are the of as determines the specificity of the is the between interaction and pollen tube are the on of S-RNase and SLF of these can be and others for studies to S-RNase and the physiology of growth inhibition are needed. A in pollen rejection system would be a however, is are to understanding S-RNase–based SI than the of for for in the in is supported by