We report the integration of the linkage map of tomato chromosome 2 with a high-density bacterial artificial chromosome fluorescence in situ hybridization (BAC–FISH)-based cytogenetic map. The euchromatic block of chromosome 2 resides between 13 and 142 cM and has a physical length of 48.12 μm, with 1 μm equivalent to 540 kb. BAC–FISH resolved a pair of loci that were 3.7–3.9 Mb apart and were not resolved on the linkage map. Most of the regions had crossover densities close to the mean of ∼200 kb/cM. Relatively hot and cold spots of recombination were unevenly distributed along the chromosome. The distribution of centimorgan/micrometer values was similar to the previously reported recombination nodule distribution along the pachytene chromosome. FISH-based physical maps will play an important role in advanced genomics research for tomato, including map-based cloning of agronomically important traits and whole-genome sequencing.
ECONOMICALLY, the Solanaceae compose the third most important plant taxon, and consist of >3000 species. Distinctive aspects of development and the variety of phenotypes and habitats make the Solanaceae good models for investigation of the genetic bases of diversification and adaptation. To this end, the “International Solanaceae Genome Project (SOL)” was launched (Mueller et al. 2005). Tomato is well suited to represent the Solanaceae because it has a relatively small genome and a strong genetics, genomics, and cytogenetics foundation.
Peterson et al. (1999) provided an overview of the DNA content and physical length of all 24 chromosome arms. Tomato has pericentromeric heterochromatin, as do other Solanaceae. The synaptonemal complex karyotype data indicate that 77% of the tomato genome is located in heterochromatin and 23% in euchromatin (Peterson et al. 1996). The genome size (1C) is ∼95 pg of DNA (Michaelson et al. 1991), implying 212 Mb of euchromatin (Bennett and Smith 1976; http://www.sgn.cornell.edu; tomato sequencing scope and completion criteria).
Excellent morphological and molecular genetic maps of the tomato genome are available (Rick and Yoder 1988; Tanksley et al. 1992). For example, >1000 restriction fragment length polymorphisms (RFLPs), mutants, and isozymes have been located on a map that totals >1276 cM (Tanksley et al. 1992). In addition, 67 RFLP and 1175 amplified fragment length polymorphism (AFLP) markers were used to construct a RFLP–AFLP map that totals 1482 cM (Haanstra et al. 1999). To date, 2037 markers have been used to create a map that totals 1460 cM; this map is available from the Solanaceae Genome Network (SGN) database (http://www.sgn.cornell.edu; EXPEN 2000 map) and is used for the “SOL” project. This linkage map, which represents all the chromosomes, does not provide sufficient detail to support genome sequencing. Because linkage map distances are not simply related to physical distances, physical mapping is needed to determine the locations of markers on chromosomes. For this purpose, bacterial artificial chromosome (BAC) fingerprinting and overgo hybridization have been applied. Currently, 3439 contigs have been anchored on the EXPEN 2000 map.
As participants in the international SOL consortium, we are responsible for sequencing the euchromatic region of chromosome 2, the third largest chromosome of tomato (Sherman and Stack 1992). Critical steps in this process are identification of the boundaries of the euchromatin and determination of the physical locations of markers. Pachytene chromosome analysis indicates that the physical size of the euchromatin is 22–26 Mb (Peterson et al. 1996; Chang 2004). The characteristic morphology of chromosome 2, with its nucleolar organizing region (NOR) and acrocentric structure, makes it easily distinguishable from the other chromosomes. Furthermore, the entire euchromatic block is located on the distal region of the long arm of the chromosome and is clearly separated from the pericentromeric heterochromatin. The linkage map of chromosome 2 has been well defined using 308 molecular markers, and its size is estimated as 143 cM (EXPEN 2000 map). A physical map has also been constructed for chromosome 2 using 75 marker-anchored BAC clones (EXPEN 2000 map). However, neither map provides sufficient detail of the physical locations of markers to initiate genome sequencing. “Molecular cytogenetics” can contribute significantly to the genome map by resolving the order of closely linked markers and confirming the physical positions of markers on the linkage groups (Anderson et al. 2004; Van Der Knaap et al. 2004; Chang et al. 2007).
Fluorescence in situ hybridization (FISH) is the most versatile and accurate method for determining the euchromatic–heterochromatic boundaries, the locations of chromosome-specific BAC clones, and the locations of repetitive and single-copy DNA sequences (Fransz et al. 2000; Cheng et al. 2001a,b; Wang et al. 2006). Here, we report the cytological and physical structure of tomato chromosome 2 in relation to the linkage map, using BAC–FISH mapping.
MATERIALS AND METHODS
Tomato (Lycopersicon esculentum cv. Micro-Tom) plants were grown in a controlled-environment room at 26° ± 1° under 16 hr light/8 hr dark.
BAC probe preparations:
All BAC clones used for BAC–FISH were kindly provided by S. Tanksley and J. Giovannoni at Cornell University, Ithaca, New York. Tomato BAC probes were labeled with digoxigenin-11-dUTP or biotin-16-dUTP by nick translation according to the protocols provided by the manufacturer of the labeling kits (Roche, Basel, Switzerland). The Arabidopsis pAtT4 clone (Richards and Ausubel 1988) and the wheat pTa71 clone containing a 9.1-kb fragment of 45S rDNA (Gerlach and Bedrock 1979) were used to detect telomeric and rDNA regions, respectively.
Pollen mother cells (PMCs) were separated using the method of Fransz et al. (2000) with some modification. Immature flower buds were fixed in ethanol:acetic acid (3:1) for 2 hr and stored at 4°. These were rinsed in distilled water and incubated in an enzyme mix (0.3% pectolyase, 0.3% cytohelicase, and 0.3% cellulase) in citrate buffer (10 mm sodium citrate, pH 4.5) for 2 hr. Each bud was softened in 60% acetic acid on an uncoated, ethanol-cleaned microscopic slide kept at 45° on a hot plate. The contents were smeared on the slide, fixed with ice-cold ethanol:acetic acid (3:1), and dried.
The FISH procedure was previously reported by Koo et al. (2004). In brief, chromosomal DNA on the slides was denatured with 70% formamide at 70° for 2.5 min, followed by dehydration in a 70, 85, 95, and 100% ethanol series at −20° for 3 min each. The probe mixture containing 50% formamide (v/v), 10% dextran sulfate (w/v), 5 ng/μl salmon sperm DNA, and 50 ng/μl labeled probe DNA was heated at 90° for 10 min and then kept on ice for 5 min. A 20-μl aliquot of this mixture was applied to the denatured chromosomal DNA and covered with a glass coverslip. The slides were then placed in a humid chamber at 37° for 18 hr. Probes were detected with avidin–FITC and anti-digoxigenin Cy3 (Roche). Chromosomes were counterstained with 1 μg/μl 4′,6-diamidino-2-phenylindole (DAPI) (Sigma, St. Louis). The signals were detected using a cooled CCD camera (CoolSNAP; Photometrics, Tokyo). The images were obtained with a Leica epi-fluorescence microscope equipped with FITC–DAPI two-way or FITC–rhodamine–DAPI three-way filter sets (Leica, Tokyo) and were processed with Meta Imaging Series TM 4.6 software. The final printed images were prepared using Photoshop 7.0 (Adobe, San Jose, CA).
Leaf nuclei were prepared as described by Jackson et al. (1998). A suspension of nuclei was deposited at one end of a poly-l-lysine-coated slide (Sigma) and air dried for 10 min. STE lysis buffer (8 μl) was added, and the slide was incubated at room temperature for 4 min. A clean coverslip was used to slowly drag the contents along the slide. The preparation was air dried, fixed in ethanol:glacial acetic acid (3:1) for 2 min, and baked at 60° for 30 min. The DNA fiber preparation was incubated with a probe mixture, covered with a 22 × 40-mm coverslip, and sealed with rubber cement. The slide was placed in direct contact with a heated surface in an oven at 80° for 3 min, transferred to a wet chamber that had been prewarmed at 80° for 2 min, and then transferred to 37° for overnight incubation. The posthybridization washing stringency was the same as in FISH of chromosome spreads. Signal detection was performed according to Koo et al. (2004).
Chromosome identification and measurement:
The images of 20 DAPI-stained pachytene bivalents at approximately the same stage were captured from different PMCs to study the distributions of heterochromatin, positions of FISH signals, and lengths of pachytene chromosomes. The images were measured directly on the screen using the FISH Image System (Meta Imaging Series TM 4.6).
Cytological architecture of chromosome 2:
The pachytene chromosome 2 of tomato is easily distinguished from the other chromosomes because it is acrocentric and bears a large secondary structure, the NOR, on the short arm (Peterson et al. 1996). The DAPI staining of the pachytene chromosome demonstrated striking differences between the euchromatin and the heterochromatin. Brightly fluorescing heterochromatic regions were detected next to the centromere of the long arm and over the entire short arm (Figure 1A). Weakly fluorescing euchromatin was observed on the long arm (Figure 1A). Chromosome 2 at meiotic prophase I was a fully paired bivalent with a mean length of 70.22 μm, based on 20 independent measurements. The lengths of the euchromatic and heterochromatic regions (including the NOR) were 48.12 ± 3.17 and 22.1 ± 1.23 μm, respectively. Previous studies estimated the size of the euchromatic region of chromosome 2 as 22–26 Mb (Sherman and Stack 1992; Peterson et al. 1996; Chang 2004). We used the largest size estimation (26 Mb), following guidelines of the International Tomato Genome Sequencing Project. Thus, we considered the euchromatin of the pachytene chromosome to have an average of 540 kb/μm.
Determination of the euchromatin borders:
Several BAC probes anchored at each end of linkage group 2 were hybridized to pachytene bivalents, and the physical locations of the BAC–FISH signals were examined. The marker cLER-1-H17 was mapped onto 0.0 cM, which is the north end of linkage group 2. The FISH signal for the BAC clone LE_HBa0007F24 anchored by cLER-1-H17 was detected on the distal end of the short arm of chromosome 2 where it covered the NOR (Figure 2A), and minor signals were also detected on the pericentromeric regions of some pachytene chromosomes (data not shown). The other BAC clones anchored to molecular markers located between 0 and 12 cM gave multiple FISH signals in the pericentromeric heterochromatic regions of all chromosomes (data not shown). The BAC–FISH signal of the T1238 (13 cM)-anchored LE_HBa0303I24 BAC was seen only near the boundary of the euchromatin and pericentromeric heterochromatin. This was located at 3.5 ± 1.3 μm from the pericentromeric heterochromatic region of the long arm (Figure 1B). The south end of the euchromatin was verified by the T1554 (142 cM)-anchored BAC clone LE_HBa0177F12. The FISH signal for LE_HBa0177F12 was detected at the distal end of the long arm of the pachytene chromosome 2 (Figure 1B). Sequencing revealed that SL_MboI0006E22, containing telomere-specific repeated sequences, is located 100 kb from LE-HBa0177F12 (Figure 2, B–D). The biotin-labeled SL_MboI0006E02 (green) was detected at the distal ends of several pachytene chromosomes, including chromosome 2 (Figure 2B). The digoxigenin-labeled Arabidopsis telomere-specific probe (pAtT4, red) was colocalized with green signals generated from SL_MboI0006E02 (Figure 2C). These data taken together identified the euchromatin between 13 and 142 cM as suitable for our study.
Integration of the cytogenetic and linkage maps:
To construct an integrated high-density cytogenetic map of chromosome 2, we selected 28 BAC clones anchored to molecular markers that were dispersed along the entire linkage group 2 (Table 1). These BAC-derived probes yielded strong FISH signals in pachytene chromosome 2 and clearly demonstrated the corresponding position and order of the selected BAC clones (Figure 3, A1–A4). The order of the selected marker-anchored BAC clones was the same as in the linkage map, except for an inversion of the markers located at 66 and 70 cM (Red dotted circles in Figure 3, A2). In addition to the identification of this inversion, the BAC–FISH map sometimes resolved loci that were not resolved on the linkage map. For example, T1555 and T1535 resided at the same position in the linkage map, but, in the BAC–FISH map, the signals anchored to the two markers were visibly separated (pink dotted circles in Figure 3, A3).
Physical gaps in linkage map 2:
The initial selection of marker-anchored BAC clones suggested four gaps (i.e., the absence of molecular marker-anchored BAC regions) that were ≥10 cM long: 18–28 cM, 46–66 cM, 96–106 cM, and 112–130 cM. However, the BAC–FISH analysis demonstrated that the gap that occurred between 46 and 70 cM (note the inversion of loci at 66 and 70 cM) was the only real physical gap (blue dotted circle in Figure 3, A2). The distance between the two FISH signals observed from LE_HBa0072A04 (46 cM) and LE_HBa0329G05 (70 cM) was 7.28 μm (data not shown), implying a physical distance of ∼3.93 Mb (i.e., 7.28 μm × 540 kb/μm). On the basis that 1 cM = 185 kb, the 20-cM interval is estimated to be 3.70 Mb. Thus, the two calculations gave similar physical distances for the large gap on chromosome 2.
Estimation of the base pair/centimorgan relationship:
Molecular marker-anchored BAC–FISH mapping was used to determine the relationship between base pairs and centimorgans (Cheng et al. 2002; Chang et al. 2007). We used five BAC-derived probes for the global determination of the base pair/centimorgan relationship. The five BAC–FISH signals were detected and easily separated into four physical blocks, G1, G2, G3, and G4 (Figure 3, A), according to genetic distances of 33, 27, 33, and 36 cM, respectively. The physical portions of these four blocks composed 24% (G1, 5.8 Mb), 20.9% (G2, 5.1 Mb), 35.3% (G3, 8.5 Mb), and 19.6% (G4, 4.7 Mb) of the entire euchromatic portion of chromosome 2. The base pair/centimorgan relationships calculated from these observations were 176 kb/cM (G1), 189 kb/cM (G2), 258 kb/cM (G3), and 131 kb/cM (G4).
In addition to these rather global measurements of the base pair/centimorgan relationship, we measured more local base pair/centimorgan relationships using 28 BAC clones covering the entire chromosome 2 (Table 2). For most of chromosome 2, the base pair/centimorgan relationship was <200 kb/cM (Figure 4). Some hot spots were detected between 120 and 140 cM, and some cold spots between 72 and 73 cM (Figure 4). Both ends of the euchromatin block have less recombination than the rest of chromosome 2 (Figure 4).
Recombination nodules (RNs) represent real crossovers in the genome and are available for tomato (Sherman and Stack 1995). Therefore, we compared the centimorgan/micrometer relationship with the average number of RNs along chromosome 2. The calculated centimorgan:micrometer ratio trends are similar to the RN distribution redrawn from Sherman and Stack (1995) (Figure 5B).
BAC–FISH-identified chromosome 2-specific BAC clones:
Because plant genomes contain many repetitive and redundant sequences, the first task in sequencing the entire euchromatic region of chromosome 2 is the selection of chromosome 2-specific anchored BAC clones. Thus, we first selected all BAC clones anchored to molecular markers located between 13 and 142 cM in linkage group 2. In total, 69 BAC clones were selected from the SGN database (http://www.sgn.cornell.edu). BAC–FISH analyses of these BAC clones demonstrated several different types of hybridization patterns (Table 3): a single FISH signal was located on pachytene chromosome 2 for 37 BACs; multiple FISH signals were located on the pericentromeric heterochromatin regions of the pachytene chromosomes, including chromosome 2, for three BACs; no FISH signal was located on pachytene chromosome 2, but was located on other chromosomes for one BAC; and no FISH signal was observed on any chromosome for 28 BACs. BAC–FISH identified 37 BAC clones that could be used as “seed” BACs for sequencing. This analysis also indicated that one BAC clone (LE_HBa0258N07) that was previously assigned to chromosome 2 by overgo hybridization actually occurs on other chromosomes, but not on chromosome 2 (data not shown).
BAC– and fiber-FISH used for the confirmation of “next” BAC candidates:
After identifying the sequences of seed BAC clones, we selected “next” BAC clones using a BLASTN search of the BAC end-sequence (BES) database of the SGN (http://www.sgn.cornell.edu/tools/blast/). Of the next BAC candidate clones that matched seed BAC sequences, those having minimal overlap with seed BAC sequences were selected as next BAC clones. We also filtered repetitive BES to reduce the occurrence of false clone picks.
However, it is possible that seed BAC sequences and BES could contain additional repetitive sequences that we had not previously curated. Therefore, Dual-color BAC–FISH analyses using seed BAC clones and next BAC clones were used to confirm the selection of next BAC clones (Figure 6). For example, the digoxigenin-labeled seed BAC clone LE_H168N10 overlapped with the biotin-labeled next BAC clone LE_M045L06 (Figure 6A). However, extended DNA fiber-FISH clearly resolved the two BAC clones (Figure 6B).
We defined the cytological architecture of tomato chromosome 2, including the location of the euchromatic block, and we integrated the recombination-based linkage map and the FISH-based physical map. This information is important in establishing a guide for genome sequencing and map-based gene cloning.
Microscopic observation of DAPI-stained pachytene chromosome 2 indicated a mean length of 70.22 μm. Of this, 48.12 μm was classified as a single euchromatic region. Taking 26 Mb as the total euchromatin of chromosome 2 gave a base pair/micrometer value for euchromatin of the pachytene chromosome of 540 kb/μm. Previous studies reported that the DNA compactness in euchromatin corresponds to 0.6 Mb/μm (Arumuganathan and Earle 1991; Peterson et al. 1996; Budiman et al. 2004). Thus, although different chromosomes and independent chromosome-spreading techniques were used, similar DNA compactness was determined, indicating that the mean base pair/micrometer relationship of euchromatin of the tomato pachytene is 0.5–0.6 Mb/μm.
To assess whether molecular marker loci existed for the distal ends of the chromosome arms, BAC probes corresponding to the ends of linkage group arms were hybridized to pachytene bivalents, and the physical locations of the FISH signals were examined. Most of the BACs derived from markers located between 0.0 and 12 cM showed multiple FISH signals in the pericentromeric heterochromatic region of all chromosomes, including the NOR or the pericentromeric heterochromatin of chromosome 2, except for one BAC (LE-HBa0155E05, 2 cM), which showed a single FISH signal in the heterochromatic region of the short arm of chromosome 2 (data not shown). Both FISH and DAPI data indicate that this segment may be composed of clusters of repetitive sequences. This result forced us to ignore this segment of the euchromatic region when selecting BAC clones for further sequencing projects. Thus, we used the euchromatic block between 13 and 142 cM for this project.
Comparing the linkage map EXPEN 2000 with our BAC–FISH map revealed an inversion in chromosome 2 (Figure 3, A2). This inconsistency in the position of the loci may have been caused by variation among the strains examined. Such apparent inversions have been reported in maize and tomato (Peterson et al. 1999). The order of the loci on the genetic and cytological maps was generally the same, as expected.
BAC–FISH maps sometimes allowed us to resolve the locations of markers that were not resolved on the linkage map. For example, both LE-HBa0213A01 and LE-HBa0164H08 were located at 88 cM in the linkage map. However, the BAC–FISH signals were separable (Figure 3, A3). This inconsistency in the position of the loci implies a low rate of recombination in the interval between them. The higher resolution of FISH mapping revealed physical gaps that could be troublesome for sequencing projects. FISH has been used to estimate the positions and sizes of gaps in the physical maps of rice and Arabidopsis. The linkage map of tomato reported by Tanksley et al. (1992) contains more markers per centimorgan than any other plant linkage map. However, chromosome 2 still contains four gaps (i.e., the absence of molecular marker-anchored BAC regions) that span >10 cM. BAC–FISH identified only one physical gap of 3.7–3.9 Mb, occurring between 46 and 66 cM, that is considered troublesome for further genome sequencing (Figure 3, A2).
Our data based on five BAC clones that spanned the entire euchromatin of chromosome 2 indicated that, on the average, 1 cM corresponds to 189 kb (range: 131–258 kb) to be compared with previous estimations of 330–1150 kb/cM (Tanksley et al. 1992; Giovannoni et al. 1995; Alpert and Tanksley 1996; Tor et al. 2002). Thus, our mean value is 16–56% as large as previous estimates. However, previous studies estimated the base pair/centimorgan relationship of specific regions, rather than of the entire chromosome (Tanksley et al. 1992; Giovannoni et al. 1995; Alpert and Tanksley 1996; Tor et al. 2002). Furthermore the genetic map they used was less saturated then the EXPEN 2000 map. The twofold differences in the base pair:centimorgan ratio between the G3 and G4 regions can be explained by higher recombination ratios in G4 than in G3. This is also consistent with the RN map of tomato (Richards and Ausubel 1988; Sherman and Stack 1995). To measure more localized base pair:centimorgan ratios of chromosome 2, we used 28 BAC clones that covered the entire chromosome 2 (Table 2 and Figure 3). Although most of the regions gave values close to the mean of 200 kb/cM, we observed a relatively hot spot (120–141 cM) and some cold spots (72–73, 106–108, and 141–142 cM) (Figure 4). Similar variation was observed in rice and Arabidopsis (Umehara et al. 1994; Schmidt et al. 1995).
Because RNs on synaptonemal complexes represent the site of crossovers (Anderson and Stack 2005) and are visible under transmission electron microscopy, we can estimate the accuracy of the integrated map of physical distance by comparing the centimorgan/micrometer relationship with the average number of RNs along the chromosome. Sherman and Stack (1995) measured the absolute positions of RNs on chromosome 2 and reported the number that occurred in each 0.1-μm segment. In Figure 5B we compare the RN distribution with the local centimorgan:micrometer ratios and find the trends in the two data sets to be similar. The DNA at both ends of the euchromatin block showed less recombination frequency than in the central region.
This may have occurred because recombination is low near the large 45S rDNA loci or telomeric repeats. This integration map will be valuable for estimating the physical structure of chromosome 2 before the chromosome is sequenced completely.
In the initial stage of genome sequencing, the marker-anchored BAC clones should be selected and termed as seed BAC clones. Because plants generally have many repetitive sequences the selection of accurate seed BAC clones is important. We used BAC–FISH analyses to select chromosome-2-specific seed BACs. We screened 69 seed BAC candidates that were previously selected by the Giovannoni and Tanksley groups at Cornell University, using overgo hybridization. FISH analysis identified 37 BACs that exhibited single strong FISH signals on pachytene chromosome 2; these were selected as the seed BACs. However, 46% of BAC clones identified by overgo hybridization resided on chromosomes other than chromosome 2. Some BAC–FISH signals detected on multiple chromosomes may be explained by the presence of repetitive sequences, so we do not know which chromosome segment was represented by these BAC clones. These FISH analyses imply that identifying the physical location of chromosome-specific BAC clones is important for this type of sequencing project.
BAC-based fingerprint contig (FPC), iterative hybridization, and sequence tag contig have been used successfully in sequencing the human, rice, and Arabidopsis genomes (Marra et al. 1997, 1998; Mayer et al. 1999; Mozo et al. 1999; McPherson et al. 2001; Chen et al. 2002; Pampanwar et al. 2005; Sasaki et al. 2005). A tomato FPC map was constructed using 88,640 BAC clones that covered the tomato genome 10 times. The map comprises 4385 contigs and 22,945 singletons; 82% of contigs are composed of <25 contig members. The small number of contig members prohibits the application of the FPC map to the selection of next BAC clones. Alternatively, we used a BLASTN search of the SGN database for BES to select next BAC candidates that overlapped the sequences of seed BAC clones. Considering the complexity of plant genomes, a BLASTN search using <1 kb BESs has the potential to identify the wrong chromosomal segment. To overcome this problem, we used BAC– or fiber-FISH to confirm the accuracy of the BLASTN search results. Dual-color BAC–FISH using seed BAC clones and next BAC clones was successful (Figure 6A). Because fiber-FISH has higher resolution than BAC–FISH, it showed a clearer relationship between seed and next BAC clones (Figure 6B).
Compared to the three other techniques presently used for physical mapping, namely, DNA contigs (Mozo et al. 1999), cytogenetic stocks (Kunzel et al. 2000), and in situ hybridization (Cheng et al. 2001a,b), the use of BAC–FISH for the integration of cytogenetic and genetic linkage maps, as in this study, is advantageous in terms of speed, accuracy, and applicability to a broad spectrum of organisms. Polyploidy and heterochromatin reduce the usefulness of contigs. For some species, the production of cytogenetic stocks is time consuming or impossible. Moreover, the resolution of the physical map obtained is comparatively low.
In situ hybridization techniques give varied resolutions depending upon such factors as genome size and ploidy. As shown in our study, a combination of cytogenetic and genetic methods can yield a high-resolution physical map.
We thank the Solanaceae Genome Network (Jim Giovanoni and Steve Tanksley) at Cornell University for providing all the tomato BAC resources. This work was supported by grants from the Crop Functional Genomics Center (CG1221) of the 21st Century Frontier Research Program funded by the Ministry of Education and Science of the Korean government.
↵1 These authors contributed equally to this study.
Communicating editor: F. W. Stahl
- Received March 25, 2008.
- Accepted May 10, 2008.
- Copyright © 2008 by the Genetics Society of America