Abstract
Almond has a self-incompatibility system that is controlled by an S locus consisting of the S-RNase gene and an unidentified “pollen S gene.” An almond cultivar “Jeffries,” a somaclonal mutant of “Nonpareil” (ScSd), has a dysfunctional Sc haplotype both in pistil and pollen. Immunoblot and genomic Southern blot analyses detected no Sc haplotype-specific signal in Jeffries. Southern blot showed that Jeffries has an extra copy of the Sd haplotype. These results indicate that at least two mutations had occurred to generate Jeffries: (1) deletion of the Sc haplotype and (2) duplication of the Sd haplotype. To analyze the extent of the deletion in Jeffries and gain insight into the physical limit of the S locus region, ∼200 kbp of a cosmid contig for the Sc haplotype was constructed. Genomic Southern blot analyses showed that the deletion in Jeffries extends beyond the region covered by the contig. Most cosmid end probes, except those near the Sc-RNase gene, cross-hybridized with DNA fragments from different S haplotypes. This suggests that regions away from the Sc-RNase gene can recombine between different S haplotypes, implying that the cosmid contig extends to the borders of the S locus.
SELF-INCOMPATIBILITY (SI) in flowering plants is a genetic system that prevents self-fertilization by enabling the pistil to reject pollen from genetically related individuals. In most cases, SI is controlled by a single multi-allelic locus called the S-locus (de Nettancourt 1977). In gametophytic self-incompatibility (GSI) systems from three plant families, the Rosaceae, Solanaceae, and Scrophulariaceae, the S-locus products in the pistils are basic glycoproteins with ribonuclease (RNase) activity called S-RNases (for reviews, see Golzet al. 1995; McCubbin and Kao 2000). Transformation experiments provided evidence that expression of active S-RNases is essential for SI reaction in the pistil of Petunia inflata and a Nicotiana hybrid (Huanget al. 1994; Leeet al. 1994; Murfettet al. 1994; McCubbinet al. 1997). Although the S-locus products in pistils have been well characterized, the identity of the S-locus products in the pollen is not known.
Studies on pistil- or pollen-specific self-compatible mutants (Thompsonet al. 1991; Golzet al. 1999) and transformation experiments with solanaceous species (Leeet al. 1994; Murfettet al. 1994) imply that the S-RNases do not affect SI phenotype of pollen. We previously analyzed a pistil-specific self-compatible mutant of the rosaceous species, Japanese pear, in which the pistil S4 allele is nonfunctional but the pollen one is functional, and showed that at least a 4-kbp genomic region including the S4-RNase gene had been deleted in the mutant (Sassaet al. 1997). These findings suggested that different genes, the pistil S gene (S-RNase gene) and the unidentified pollen S gene, determine the SI phenotype in pistil and pollen, respectively, and that the two genes behave as if they are a single gene. On the basis of the findings that the S locus of the Solanaceae and Rosaceae is a multigene complex, the term “haplotype” has been adopted to denote variants of the locus and the term “allele” to denote variants of a given polymorphic gene at the S locus (McCubbin and Kao 2000).
In almond (Prunus dulcis), which belongs to the Rosaceae, there is a mutant for SI called “Jeffries.” Jeffries (ScmSd) is a naturally occurring somaclonal mutant and was found as a sport on a tree of the cultivar “Nonpareil” (ScSd). The pistils of Jeffries accept pollen from any heterozygous almond cultivar except self-pollen, whereas the pollen of Jeffries is rejected by the pistils of cultivars carrying the Sd haplotype (Kesteret al. 1994). Tao et al. (1997) showed that the Sd-RNase accumulated in the pistil of Jeffries but the Sc-RNase did not. These findings indicated that Jeffries lacks function of the Sc haplotype both in the pistil and the pollen and suggested that characterization of the mutation in Jeffries could help us identify the pollen S gene. In this study, we carried out immunoblot and genomic Southern blot analyses and showed that deletion of the Sc haplotype and duplication of the Sd haplotype had occurred to generate Jeffries. Aiming to clarify the extent of the deletion in Jeffries, we conducted a chromosome walking experiment starting from the Sc-RNase gene in Nonpareil, walked both upstream and downstream, and constructed a cosmid contig covering ∼200 kbp of the c haplotype region. Genomic Southern blot analyses S with cosmid end probes showed that the entire region covered by the cosmid contig had been deleted in Jeffries. Consequently, the deletion in Jeffries is >200 kbp. Most cosmid end probes, except for those located near the Sc-RNase gene, cross-hybridized with DNA fragments from different S haplotypes on genomic Southern blots, implying that regions away from the S-RNase gene were shared through recombination between different S haplotypes and thus the cosmid contig extends beyond the borders of the S locus.
MATERIALS AND METHODS
Plant materials: Eight cultivars of almond, Nonpareil (ScSd), Mission (SaSb), Wood Colony (SaSc), Sauret no. 2 (SaSc), Merced (SbSc), Sauret no. 1 (SaSd), Monterey (SbSd), and Jeffries (ScmSd) were used. Jeffries is a naturally occurring mutant found as a sport on a Nonpareil tree. Scm stands for the mutation of the Sc haplotype both in pistil and pollen of Jeffries (Kesteret al. 1994; Taoet al. 1997). Leaves were freeze dried and stored at -20° until used for DNA isolation. Pistils were collected from flowers at the balloon stage of development, frozen in liquid N2, and stored at -80° until use.
Expression of recombinant Sc-RNase in Escherichia coli and preparation of antiserum: cDNA encoding the Sc-RNase was cloned into the E. coli expression vector pThioHis (Invitrogen, Carlsbad, CA) to express a fusion protein of Sc-RNase and thioredoxin with a histidine patch. The cDNA encoding the mature peptide sequence of the Sc-RNase (Ushijimaet al. 1998b) was amplified by PCR with primers FScTHis (5′-TAC CTA GTG GAT CTT ATG ACT ATT TTC-3′) and RScTHis (5′-CGA GAT CTT TAT TGA AAC TTG ATG TCA ATT TTA-3′) to incorporate KpnI and BglII restriction sites into the 5′ and 3′ ends of the amplified fragment, respectively, and cloned into the expression vector. The fusion protein was abundant in the insoluble inclusion bodies. Proteins in inclusion bodies were extracted by lysis buffer (O’Farrell 1975) and subjected to two-dimensional polyacrylamide gel electrophoresis (2D-PAGE; Hirano 1982). Fusion protein spots were excised from Coomassie-blue-stained 2D-PAGE gels and recovered by Centriluter with Centricon-10 (Millipore, Bedford, MA). The recovered proteins were mixed with adjuvant (TiterMax Gold, CytRx) and injected into a rabbit to obtain antiserum.
Immunoblot analysis: Stylar proteins extracted from acetone powder with lysis buffer (Sassaet al. 1993) were quantified as described by Ramagli and Rodriguez (1985). Twenty micrograms of total proteins were separated by nonequilibrium pH gradient electrophoresis (NEPHGE) with a slab gel containing 8 m urea, 7.5% acrylamide, 0.3% bis-acrylamide, 5% glycerol, and 2% Ampholine pH 3.5-10 and electroblotted onto a polyvinylidene difluoride (PVDF) membrane (Sassa and Hirano 1998). The blot was stained with Ponceau S to check for equal loading. Free binding sites on the PVDF membrane were blocked with 4% nonfat dried milk in TRIS-buffered saline (TBS) containing 20 mm TRIS-HCl, pH 7.5, and 0.5 m NaCl. The blocked membrane was incubated with the antiserum solution diluted to 1:1000 with 4% nonfat dried milk in TBS. It was then incubated with alkaline phosphatase-conjugated secondary antibody and stained by enzymatic reaction with 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and nitro-blue-tetrazolium (NBT).
Isolation of genomic DNA: Freeze-dried leaves were ground into powder using a mortar and pestle and homogenized in homogenizing buffer (Zhanget al. 1995). The homogenate was centrifuged (1800 × g at 4° for 30 min) to collect the pellet. Genomic DNA was isolated from the pellet by Nucleon Phytopure for plant DNA extraction kit (Amersham Pharmacia, Buckinghamshire, UK) and further purified by chloroform extraction.
Genomic Southern blot analysis: Genomic DNA (5 μg) from the various almond cultivars was digested with restriction enzymes, separated on a 0.8% agarose gel, and blotted onto a nylon membrane (Biodyne Plus, Pall). The blot was probed with cDNA clones encoding the Sc- or the Sd-RNases (Ushijimaet al. 1998b) or cosmid end probes (see below). cDNAs and end regions of cosmid inserts were labeled with dUTP linked to digoxigenin (DIG; Roche Diagnostics, Mannheim, Germany) by PCR using gene-specific primers AS1 II and Amy-C5R for the cDNAs encoding S-RNases (Tamuraet al. 2000) or cosmid end-specific primers (see below). The blots were washed at high stringency conditions and visualized as previously described (Ushijimaet al. 1998a).
Quantitative genomic Southern blot analysis: Genomic DNA of Jeffries and Nonpareil were digested with DraI. Four, six, and eight micrograms of the digested DNA were loaded and separated on a 0.8% agarose gel, and blotted onto a nylon membrane (Biodyne Plus, Pall). The blot was probed with cDNA clones encoding the Sd-RNase (Ushijimaet al. 1998b) or P. dulcis proline-rich protein (PdPRP; H. Sassa, K. Ushijima and H. Hirano, unpublished data), which were labeled by enhanced chemifluorescence (ECF) random prime labeling module (Amersham Pharmacia). Hybridization, high stringency washes, and signal amplifications were conducted according to the manufacturer’s protocols (Amersham Pharmacia). Signal intensities were quantified with a fluorescence scanning system (FluorImager SI, Amersham Pharmacia). The intensity of the Sd haplotype-specific signal was equalized against those of the PdPRP gene, which is a single-copy gene in almond (H. Sassa, K. Ushijima and H. Hirano, unpublished data) and was used as the internal control. The quantitative data were obtained from three independently prepared genomic DNA blots.
Construction and screening of a cosmid library: Genomic DNA of Nonpareil was separated on a 0.3% agarose gel. The gel containing the 35- to 40-kbp region of the genomic DNA was excised and embedded in a 4% low melting point agarose gel. The genomic DNA was concentrated into the 4% agarose gel by electrophoresis, purified from the agarose gel with β-agarase (New England BioLabs, Beverly, MA), blunt-ended with T4 DNA polymerase (Roche Diagnostics), and phosphorylated with T4 polynucleotide kinase (New England BioLabs). The resulting genomic DNA was cloned into the SmaI site of pWEB cosmid vector (Epicentre Technologies, Madison, WI), and packaged in vitro using MaxPlax packaging extract kit (Epicentre Technologies).
The packaged library was plated with E. coli XL1-blue MR and screened by colony hybridization. Filters were hybridized with the cDNA encoding the Sc-RNase or cosmid end probes (see below), which were labeled with DIG. After the high stringency washes (two times 5 min at room temperature with 2× SSC and 0.1% SDS followed by two times 15 min at 65° with 0.1× SSC and 0.1% SDS), colorimetric detection with BCIP and NBT was conducted according to the manufacturer’s protocol (Roche Diagnostics). Putative positive clones were isolated by standard alkaline lysis method and checked by PCR. Because some cosmid end probes cross-hybridized with the DNA fragments from the Sd haplotype (see results), Southern blot or PCR-restriction fragment length polymorphism (RFLP) analyses for the isolated clones were carried out to distinguish the S haplotypes from which the clones were derived, and the clones derived from the Sc haplotype were used to construct the contig.
Preparation of end probes of the cosmid clones and chromosome walking: Both end regions (∼500-700 bp) of cosmid inserts were sequenced with an autosequencer (Model 4000L Sequencer, LI-COR). Cosmid end-specific primers were designed from these sequences. End regions of cosmid inserts were labeled with DIG by PCR using the end-specific primers and used as probes for the library screening and the genomic Southern blot analyses.
Cosmid clones were digested with NotI and electrophoresed on a 0.3% agarose gel to estimate the insert size. Overlap of cosmid clones was determined by amplifying the region from cosmid clones by the Expand Long Template PCR System (Roche Diagnostics) with the vector primers WEBF (5′-GCC AGG GTT TTC CCA GTC AC-3′) and PWR (5′-GCA TAA TAC GAC TCA CTA TAG G-3′), which were designed from sequences of the cloning site of pWEB vector and end-specific primers of the overlapping clones.
RESULTS
The pistil of the Sc haplotype mutant Jeffries accumulates no Sc-RNase but has an elevated level of the Sd-RNase: The Sc haplotype is nonfunctional in both the pistil and the pollen of the mutant Jeffries, although the Sd haplotype is functional in both (Kesteret al. 1994). To compare the accumulation of the S-RNases in Jeffries with those in the original cultivar Nonpareil, we conducted immunoblot analysis for the S-RNases. Stylar proteins of almond cultivars were separated by NEPHGE and subjected to immunoblotting with the Sc-RNase antiserum. The antiserum gave Sc or Sd haplotype-specific signals at the basic region of the blot, indicating that the antiserum reacted not only with the Sc-RNase but also with the Sd-RNase (Figure 1). The Sc and Sd haplotype-specific signals were detected in the stylar proteins of Nonpareil. However, no Sc haplotype-specific signal was detected in proteins from Jeffries. The S d haplotype-specific signal in Jeffries was stronger than that in Nonpareil, although equal amounts (20 μg) of the stylar proteins were subjected to immunoblotting, suggesting that the expression level of the Sd-RNase gene in Jeffries is higher than that in Nonpareil.
Jeffries lacks the Sc-RNase gene and has an extra copy of the Sd-RNase gene: To analyze the mutation of the S-locus region in Jeffries, genomic Southern blot analyses using the cDNAs encoding the Sc- and Sd-RNases as probes were carried out. The cDNA probe for the Sd-RNase gave two Sd haplotype-specific signals in both Jeffries and Nonpareil DNA digested with EcoRI or DraI (Figure 2). The cDNA probe for the Sc-RNase gave no signals with Jeffries DNA but detected the Sc haplotype-specific signals in Nonpareil DNA, indicating that the Sc-RNase gene had been deleted in the mutant Jeffries. The intensities of the Sd haplotype-specific signals in Jeffries also appeared stronger than those in Nonpareil (Figure 2) as was seen earlier in the immunoblot analysis (Figure 1).
—Immunoblot analysis of stylar proteins of almond cultivars with the antiserum against the Sc-RNase. Stylar proteins of six almond cultivars were separated by NEPHGE and electroblotted onto a PVDF membrane. The blots were stained with Ponceau S solution to ensure the equal loading of stylar proteins and then subjected to immunodetection with the antiserum against the Sc-RNase. The arrows of Sc and Sd denote the Sc and Sd haplotype-specific signals, respectively. JF, Jeffries (ScmSd); NP, Nonpareil (ScSd); MS, Mission (SaSb); WC, Wood Colony (SaSc); #1, Sauret no. 1 (SaSd); MT, Monterey (SbSd).
For quantitative analysis, genomic DNA blots were probed with the cDNA encoding the Sd-RNase and the 2.4-kbp signal specific to the Sd haplotype was quantified by a fluorescence scanning system (Figure 3). The intensities for the Sd haplotype-specific signals were equalized using the signal of the PdPRP gene (H. Sassa, K. Ushijima and H. Hirano, unpublished results), which is a single-copy gene in almond genome. This showed that the intensity of the Sd haplotype-specific signal in Jeffries was ∼1.8-fold stronger than that in Nonpareil (Figure 3).
—Genomic Southern blot analysis for the Sc- and the Sd-RNase genes in Jeffries and Nonpareil. Genomic DNAs (5 μg per lane) were digested with EcoRI or DraI and probed with the cDNA encoding the Sc- or Sd-RNases. The arrows of Sc and Sd denote the Sc and the Sd haplotype-specific signals, respectively. JF, Jeffries; NP, Nonpareil.
—Quantitative genomic Southern blot analysis for the Sd-RNase gene in Jeffries and Nonpareil. (A) Genomic Southern blot analysis of the Sd-RNase gene in Jeffries and Nonpareil by ECF system. Four, six, and eight micrograms of the genomic DNAs of Jeffries or Nonpareil were digested with DraI and probed with the cDNA encoding the Sd-RNase and the PdPRP, which were labeled by ECF random prime labeling module. Amount of loaded DNAs and the cultivars were denoted above each lane. (B) Quantitative analysis of intensities of the Sd haplotype-specific signals in Jeffries and Nonpareil. Of two Sd haplotype-specific signals, a 2.4-kbp one was quantified by a fluorescence scanning system, because its intensity was stronger than the 0.8-kbp one. The signals were equalized by the signals of the PdPRP gene used as the internal control and normalized to that of 4 μg in Nonpareil. The data represent the mean of the relative intensities derived from three independent experiments. Error bars indicate ± SE. The Sd haplotype signal ratios (Jeffries/Nonpareil) in 4, 6, and 8 μg of genomic DNAs were 1.86, 1.81, and 1.83, respectively.
Deletion of the Sc haplotype and the flanking regions in Jeffries is >200 kbp: No Sc haplotype-specific signal was detected by the Sc-RNase probe in Jeffries, indicating that the Sc-RNase gene has been deleted from Jeffries. Because Jeffries lacks function of the Sc haplotype not only in the pistil but also in the pollen, the pollen Sc gene should also be deleted in Jeffries. This suggests that the genomic region deleted in Jeffries represents the maximal physical limit of the Sc haplotype. We thus decided to construct a cosmid contig for the Sc haplotype region to determine the extent of the deletion and gain insight into the size of the Sc haplotype. A cosmid library was constructed from genomic DNA of the original cultivar Nonpareil and screened by colony hybridization with the cDNA encoding the Sc-RNase, and then with cosmid end probes to obtain overlapping clones. Finally, we isolated 12 overlapping cosmid clones that covered ∼200 kbp of the Sc haplotype and flanking regions with the Sc-RNase gene roughly in the middle (Figure 4A).
To confirm whether this contig covers the region deleted in Jeffries, genomic Southern blot analyses were conducted with end probes derived from the cosmid clones (Figure 4B). Sc haplotype-associated signals (denoted as Sc in Figure 4B) were detected in Nonpareil by all the cosmid end probes, except for NP179R and NP182R. NP179R and NP182R probes gave smear and multiple banding patterns in genomic DNA from Jeffries and Nonpareil, respectively. In addition to the Sc signal, all the cosmid end probes except NP79F, NP179R, and NP182R gave a single band associated with S d haplotype (denoted as Sd signal). Intensities of the Sd signals in Jeffries were also stronger than those in Nonpareil, as had previously been seen with cDNA probes for the Sd-RNase (Figures 2 and 3). Genomic Southern blot analyses using DNA from different S genotypes were carried out to confirm the identity of Sc or Sd signals (data for NP207R is shown). NP207R cross-hybridized with DNA fragments from all S genotypes tested and showed S haplotype-associated RFLPs on DraI-digested DNA blot, even when hybridization and washing was carried out under high stringency conditions. However, the DraI-digested DNA blot was the only case showing S haplotype-associated RFLP with NP207R probe. NP207R gave a monomorphic pattern with most restriction enzymes tested (Figure 4C and Table 1). Similar to NP207R, all the cosmid end probes except NP79F, NP179R, and NP182R cross-hybridized to one or two bands in all S genotypes tested (Figure 4 B and C). The region corresponding to the cosmid contig was shown to be deleted in Jeffries, because Sc haplotype-associated signals were not detected in Jeffries by any of the cosmid end probes, including the most upstream probe, NP283R, and the most downstream probe, NP207R (Figure 4).
DISCUSSION
The almond cultivar Jeffries is a naturally occurring somaclonal mutant found as a sport on a Nonpareil tree. It was reported to have a nonfunctional Sc haplotype both in the pistil and pollen (Kesteret al. 1994). Tao et al. (1997) showed that the Sc-RNase was not detectable in the stylar proteins of Jeffries by 2D-PAGE analysis. In this article, we further characterized the mutation of Jeffries by immunoblot and genomic Southern blot analyses and showed that no Sc haplotype-specific signals could be detected in Jeffries. The quantitative analysis of the Sd haplotype-specific signal by genomic Southern blot showed that the signal intensity of the Sd-RNase gene in Jeffries was ∼1.8-fold higher than that in Nonpareil. The intensity of the Sd haplotype-associated signals in Jeffries detected by cosmid end probes was also stronger than that in Nonpareil. These results suggested that two mutations had occurred in a somatic cell of Nonpareil to generate Jeffries: (1) deletion of the Sc haplotype and (2) duplication of the Sd haplotype. The result that the Sd haplotype-specific signal in Jeffries was not double but ∼1.8-fold of that of Nonpareil might be explained by postulating that a small amount of wild-type cells remain in Jeffries, reflecting its somaclonal origin. A similar situation is seen in a somaclonally derived, self-compatible mutant of Japanese pear, “Osa-Nijisseiki,” which is probably chimeric for the S4-RNase gene (Sassaet al. 1997). Indeed, the Sc-RNase gene could be amplified from the genomic DNA of Jeffries by PCR (data not shown). A weak Sc haplotype-specific signal could also be detected by genomic Southern blot analyses in Jeffries when an increased amount of DNA was loaded (data not shown). Although Jeffries may retain a small number of wild-type cells, it seems these are not sufficient for the function of the Sc haplotype both in the pistil and the pollen, since Jeffries behaves as if homozygous for the Sd haplotype (Kesteret al. 1994). This is consistent with the finding that high levels of S-RNase expression are required for the pistil to reject self-pollen in an RNase-based GSI system (Clarket al. 1990; Leeet al. 1994).
—Construction of a cosmid contig for almond Sc haplotype region and genomic Southern blot analysis with cosmid end probes. A cosmid library was constructed from the genomic DNA of the original cultivar Nonpareil. The cosmid contig for the Sc haplotype region was constructed by chromosome walking from the Sc-RNase gene to both the upstream and the downstream regions. (A) Schematic representation of the cosmid contig for the Sc haplotype region. The open, solid, and shaded boxes denote the cosmid clones, cosmid end probes used in B, and the Sc-RNase gene, respectively. The “A” nucleotide of putative initiation codon (ATG) of the Sc-RNase gene (Ushijimaet al. 1998b) was positioned to be +1. (B) Genomic Southern blot analyses with cosmid end probes and selected restriction enzymes. The genomic DNAs were digested with selected restriction enzymes, which gave the S haplotype-associated RFLP, and probed with cosmid end probes. The four cosmid end probes, NP201F, 207R, 217R, and 235F, were designed from end sequences of the Sd-derived clones. Locations of the Sd-derived probes on the corresponding Sc-derived clones were estimated by PCR with the end primers and the vector primers WEBF or PWR and the Sc-derived clones as the templates. The restriction enzymes and the cultivars were denoted above each lane. The arrows of Sa, Sb, Sc, and Sd denote the Sa, Sb, Sc, and Sd haplotype-associated signals, respectively. (C) Southern blot with HindIII-digested genomic DNA. Cosmid end probes apart from the Sc-RNase gene tended to give a monomorphic pattern on Southern blots with most restriction enzymes. In addition to a monomorphic band, NP79R gave the Sa haplotype-associated signal. JF, Jeffries (ScmSd); NP, Nonpareil (ScSd); MS, Mission (SaSb); #2, Sauret no. 2 (SaSc); MR, Merced (SbSc); #1, Sauret no. 1 (SaSd); MT, Monterey (SbSd).
If there is a causal relationship between the two mutations in Jeffries, it suggests the possibility that a gene conversion-like event between the Sc and the Sd haplotype occurred in a somatic cell of Nonpareil to generate Jeffries. Gene conversion events in plant somatic cells were observed in DNA double-strand break (DSB) repair via homologous recombination (Gorbunova and Levy 1999). DSBs are critical lesions in eukaryotic genomes and are repaired via nonhomologous end-joining or homologous recombination. Puchta (1999) demonstrated that the homologous recombination with gene conversion occurred between interchromosomal ectopic homologous sequences in somatic cells of Nicotiana tabacum by in vivo induction of DSB with the highly specific restriction enzyme I-SceI. It was also shown that a new allele of the 27-kD zein locus in maize was generated by gene conversion for which the allelic sequences were used as templates (Huet al. 1998). It is also possible that the Sc haplotype region of a somatic cell of Nonpareil had converted into the Sd haplotype region by homologous recombination with gene conversion to generate Jeffries. However, because we did not determine and characterize the junctions of the region deleted in Jeffries or carry out genetic analysis, the possibility that the duplicated copy of the Sd haplotype has translocated at a chromosomal region other than the S locus cannot be excluded.
RFLPs between the Sc and Sd haplotypes detected by the cosmid end probes
We screened a cosmid library of Nonpareil and isolated 12 overlapping cosmid clones that cover ∼200 kbp of the Sc haplotype and flanking regions. Only 2 of 11 cosmid end probes gave smear or multiple bands, suggesting that the S-locus flanking regions of Prunus are not abundant in repetitive sequences. This is in contrast to the solanaceous S-RNase gene that is flanked by abundant repetitive sequences (Coleman and Kao 1992; Mattonet al. 1995), which is consistent with a probable subcentromeric localization (Brewbaker and Natarajan 1960; Golzet al. 1999). Genetic analyses have shown that the S locus of Lycopersicon would be located close to the centromere of chromosome I (Bernacchi and Tanksley 1997). Furthermore, Entani et al. (1999) physically localized the S locus of P. hybrida to the centromeric region of chromosome III by fluorescence in situ hybridization (FISH) using the S-RNase cDNA probe. They also isolated a 20-kbp genomic clone containing the S-RNase gene and showed that the clone contained a repetitive sequence that hybridized to the centromeric regions of all Petunia chromosomes in FISH analysis. The centromeric location of the solanaceous S locus suggested that recombination around the S-locus region may be suppressed by the centromeric heterochromatin (Robbinset al. 2000), consistent with the speculation that the solanaceous S locus is in excess of 1 Mb (McCubbin and Kao 1999). On the other hand, this study shows that the S-locus flanking regions of Prunus are not abundant in repetitive sequences and provides no data implying centromeric localization of the S locus. This suggests the possibility that the extent of the S locus of Prunus may be smaller than that of the Solanaceae.
It was difficult to delimit the S locus by analysis of the extent of the region deleted in Jeffries because it was larger in size than the 200-kbp cosmid contig. However, the organization of the cosmid contig for the Sc haplotype and flanking regions may help us gain insight into the extent of the S locus of Prunus. In SI systems controlled by an S locus consisting of separate pistil S and pollen S genes, such as those of the Solanaceae, Rosaceae, and Brassicaceae, recombination between the pistil S gene and the pollen S gene is thought to lead to loss of SI. Inbreeding depression is then thought to eliminate self-compatible plants from the population, and thus recombination between the two genes seems to be suppressed, whether or not it is actually suppressed. The suppression of recombination within the S locus could allow accumulation of point mutations, rearrangements, and other sequence variations and lead to the S locus being heterogeneous among different S haplotypes. In contrast, nucleotide sequences beyond the boundary of the S locus, i.e., the S-locus flanking region, would be common to many S haplotypes as a result of recombination. Therefore, the S locus may be confined within the region showing sequence diversity among different S haplotypes and flanked by regions whose sequences are relatively conserved among different S haplotypes. This theoretical consideration was supported by recent experimental data for the organization of the S locus and flanking regions of Brassica. Cui et al. (1999) constructed and sequenced two contigs for the S910 and the SA14 haplotypes and flanking regions of Brassica and showed that the downstream regions of SLG (S-locus glycoprotein gene) between two haplotypes showed high sequence similarities, which led them to speculate that the SLL2-SLL1 region located just downstream of the SLG would be one border separating the S locus and the flanking region. In contrast, the nucleotide sequences and gene organization of upstream region of SLG, in which SRK (pistil S gene) was located, was not conserved between S910 and SA14 haplotypes and seemed to correspond to the S locus and contain the pollen S gene. The recently identified pollen S gene of Brassica, SCR/SP11, was exactly located upstream of the SLG and extremely proximal (∼20 kbp) to the SRK (Schopferet al. 1999; Brugièreet al. 2000; Cuiet al. 2000; Takayamaet al. 2000). Analogous organization was observed in the cosmid contig constructed in this study, which consisted of the region containing the Sc haplotype-specific sequences such as the Sc-RNase gene, and the flanking regions containing sequences common to many different S haplotypes. The cosmid end probes that were located at the regions upstream from NP79R and downstream from NP182F crosshybridized with DNA fragments from all S haplotype flanking regions, even when hybridization and washing were done under high stringency conditions. This indicated that nucleotide sequences of the regions upstream from NP79R and downstream from NP182F are relatively conserved among many different S haplotypes as a result of recombination. In contrast, genomic Southern blot analyses showed that the region between NP79R and NP182F contained the Sc haplotype-specific sequences (Sc-RNase gene and NP79F) and repetitive sequences and was poor in sequences shared among many different S haplotypes. These results imply that the ∼70-kbp region between NP79R and NP182F contains the entire Sc haplotype and that the cosmid contig extends beyond both borders of the S locus of Prunus and consequently contains the pollen Sc gene. Further structural, transcriptional, and functional analyses of the cosmid contig would be required to see whether it actually contains the pollen S gene.
Acknowledgments
This work was partially supported by Grants-in-Aid to H.S. from the Ministry of Education, Science, Sports and Culture of Japan. K.U. is a fellow of the Japan Society for Promotion of Science.
Footnotes
-
Communicating editor: M. K. Uyenoyama
- Received August 11, 2000.
- Accepted January 18, 2001.
- Copyright © 2001 by the Genetics Society of America
