RESULTS

Experimental approach: To compare M26 hotspot activity in meiotic ectopic and allelic recombination, we performed multiple pair-wise crosses with strains carrying a single copy of the ade6 gene, either at the endogenous ade6 locus, or integrated at a novel chromosomal site (Figure 1). At each site three different alleles, M26, M375, or 469, were inserted separately. Each allele is a single-bp mutation that creates a translation termination codon and an adenine auxotroph. Except for the different single-bp mutations, the structures at each chromosomal site are identical for all three alleles. Strains carrying either the M26 hotspot or M375 control mutation were crossed with strains carrying a 469 test allele. These markers are separated by a distance of ~1300 bp (Szankasiet al. 1988). Intragenic recombinants were detected as adenine prototrophs among spore colonies. M26 hotspot activity is the ratio of the recombinant frequency of an M26 × 469 cross divided by the recombinant frequency of an M375 × 469 cross involving the same loci. Heteroallelic crosses were performed between strains with ade6 alleles at different chromosomal sites (ectopic recombination) or at the same site (allelic recombination).

Allelic recombination and hotspot activity: To study the effects of chromosomal context on allelic recombination, we performed heteroallelic crosses with artificially integrated ade6 alleles at multiple chromosomal sites. For the random integrations of pJV8 (z7 and z15), the ade6 alleles are surrounded by plasmid DNA on each end (Figure 1). Allelic recombinant frequencies have been reported previously (Virginet al. 1995). The same crosses were performed independently as controls in the current study (shown on the diagonals in Tables 2 and 3), with less than twofold variation between the results of the two studies. In the ade6-M375 × ade6-469 crosses there is up to an 80-fold variation in the allelic recombinant frequencies for crosses with the same set of marker alleles at the different integration sites (Table 2). However, most of the variation is attributable to the markedly elevated recombination in the ura4-147(Sac) integration. Excluding this integration, there is a sixfold range in recombinant frequencies. At the ura4 locus, several different ade6 integration structures were studied, and the recombinant frequencies varied up to ~80-fold for different structurally related integrations, consistent with previous findings (Virginet al. 1995).

The M26 hotspot showed chromosomal context dependence in allelic recombination. M26 was active at the natural ade6 locus, where it was originally created (Gutz 1971), and at the ura4-108(5.9) and ura4-142(Spe) integrations (Table 4; Virginet al. 1995), with 6–22-fold induction of intragenic recombination relative to M375. It was inactive at all other integration sites tested. As discussed previously, these and other studies indicate that chromosomal context is an important determinant of recombinant frequency and recombination hotspot activity (Virginet al. 1995; Foxet al. 1997).

Ectopic recombination and hotspot activity: To determine if ectopic recombination is subject to the same chromosomal context effects as allelic recombination, we crossed strains with ade6 integrations at different chromosomal positions. Ectopic recombination is much less frequent than allelic recombination. The ectopic recombinant frequencies in the ade6-M375 × ade6-469 crosses reflect recombination without the influence of the M26 hotspot. The frequencies vary over a nearly 300-fold range (from 2.7 × 10⁻⁴ to 10⁻⁶ recombinants per viable spore), and most of the frequencies were reduced 100–1000-fold relative to allelic recombination (Table 2). The greatest reduction relative to allelic recombination was 3000-fold [ura4-147(Sac) × ade6-469], and the least reduction was ~3-fold [z15-2 × ura4-152(Spe)]. In all but three crosses [z15-2 × ura4-152(Spe), z15-2 × ura4-145(Sac), and z15-3 × ura4-147(Sac)] the ectopic recombinant frequency was less than 10% of either corresponding allelic recombinant frequency.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Integration sites and structures. (A) Genomic map of the different loci studied. Heavy vertical lines represent the three S. pombe chromosomes. The relative chromosome lengths and centromere positions (○) are taken from Munz et al. (1989). The integration sites are placed on the map according to sequence data (see materials and methods) and Hoheisel et al. (1993) and Munz et al. (1989). The solid arrow indicates the direction of transcription of the ade6 gene relative to cenIII (Grimmet al. 1994). Dashed arrows indicate the inferred orientations of the ade6 genes at the different integration sites based on the pattern of viable translocations (see results). The ura4 integrations include both orientations, with ade6 transcription directed toward the centromere [ura4(3.0) and ura4(5.9)] and away from the centromere [ura4(Sac) and ura4(Spe)]. (B) The plasmid pJV8 was linearized with Nar I prior to transformation of GP981 (h⁻ ade6-D1 ura4-D18 leu1-32) to Ura⁺. Independent random integrations, designated z7 and z15, were isolated. (C) Plasmids were constructed with different lengths of S. pombe genomic DNA containing the ade6 gene, together with several hundred base pairs of bacterial plasmid DNA, inserted into the EcoRV site of the ura4 gene. These constructs were integrated into the chromosomal ura4 locus by homologous transplacement. See text for discussion of allele designations. For details of construction, see Virgin et al. (1995). Solid arrows, ade6 gene; open boxes, ura4 gene; heavy solid lines, bacterial plasmid DNA; thin solid line, S. pombe genomic DNA. Arrowheads indicate the direction of transcription. Numbers below the lines represent distances in kilobase pairs. For clarity, portions of the plasmid are not drawn to scale. E, EcoRV; N, NarI; P, PvuII; Sc, SacI; Se, SpeI.

The pattern of M26 hotspot activity was the same in ectopic and allelic recombination. As discussed above, M26 is active in allelic recombination at the natural ade6 locus and in the ura4-108(5.9) and ura4-142(Spe) integrations (Tables 2, 3 and 4; Virginet al. 1995). It was active in these same integrations in ectopic recombination with all other loci tested, with hotspot values ranging from 7–22-fold (Table 4). Likewise, in integrations where M26 is inactive in allelic recombination, it was also inactive in ectopic recombination. In some cases of ectopic recombination, hotspot values of 2–3-fold were observed. These small increases were inconsistent and of doubtful significance. These results imply a similar mechanism of M26 action in ectopic and allelic recombination.

To further compare M26 hotspot activity in ectopic and allelic recombination, we determined the locus that was converted to ade6⁺ for some of the crosses. Gutz (1971) has shown by tetrad dissection that in allelic recombination at the endogenous ade6 locus, the M26 allele is predominantly a recipient of wild-type information in conversion events, whereas the M375 allele shows parity in the direction of information transfer. We wanted to compare the direction of information transfer in ectopic recombination with the previous results from allelic recombination. However, with the low frequencies of ectopic recombination, tetrad dissection is impractical. As an alternative method of determining the direction of information transfer, we used test crosses to determine the location of the ade6⁺ allele in the recombinants. Recombinants were crossed with two different tester strains, each with an ade6⁺ allele at one of the parental integration sites (Figure 2). With recombinants resulting from nonreciprocal recombination, only one of these two test crosses is expected to generate red Ade⁻ spore colonies, because one test cross will have ade6⁺ alleles at the same locus in both parents (the Ade⁺ recombinant and the Ade⁺ tester strain). Thus, in test crosses that segregate Ade⁻ spores, the Ade⁺ tester strain identifies the locus that was the donor in the original cross (Figure 2A). Using this assay, we determined which allele was a recipient in a subset of the crosses.

In crosses in which the M26 hotspot was active it was a recipient of wild-type information in 97–99% of the recombination events (Table 5, lines 1 and 3). This result is consistent with tetrad analysis of allelic recombination (Gutz 1971). However, in crosses in which the M26 hotspot was inactive there was incomplete correlation with hotspot activity. There was parity in the direction of information transfer with M26 in the z15 integration, in which the hotspot was inactive (Table 5, line 12). In other crosses in which the hotspot integrated at ura4 was inactive, M26 was predominantly a recipient in 86–95% of the recombination events (Table 5, lines 5, 8, and 10). Furthermore, the nonhotspot ura4-154::ade6-M375(3.0) allele was also a recipient in 79–88% of recombination events (Table 5, lines 6, 9, and 11). These results suggest that locus-specific rather than allele-specific factors determine the direction of information transfer in ectopic recombination involving the ura4 locus. To further test this idea, we reversed the orientation of the alleles in the ura4 × z7 cross (Table 5, line 7). The ura4 locus remained predominantly a recipient. The ade6 locus showed a similar pattern in which it was predominantly a recipient with either the M26 or M375 allele (Table 5, lines 1–4). The direction of information transfer in ectopic recombination appears to depend on locus-specific factors, which may obscure allele-specific effects.

We confirmed the genetic analysis by comparison of the results with a Southern blot (Figure 3). Each lane represents ClaI-digested DNA from either parental (lanes 1, 20, 21, and 40) or recombinant (lanes 2–10, 12–19, 22–29, and 31–39) isolates from an ade6-M26 × z7-3::ade6-469 cross, probed with pJV8. Since there is no ClaI site in the ade6 gene or in the integration plasmid, a single band represented the ade6 and z7 loci in the parental strains. Above each lane is the designation for reciprocal (R) or nonreciprocal (N) recombinant, based on the segregation patterns in the test crosses (Figure 2). In the nonreciprocal recombinants one or both parental bands were present. One band represents the recombinant ade6⁺ allele, and the other band represents a parental allele, which should segregate at random. In recombinants with one band at the position of a parental allele (lanes 8, 22–25, 32, 33, 36, and 37), this locus could be assigned as the recombinant ade6⁺ allele (the recipient of wild-type information), since it must be present in the recombinant prototroph. In all these lanes the single band comigrated with the parental ade6 band, which was the recipient by the genetic assay in all the isolates represented (data not shown). This confirmation of the results from the test crosses indicates that the complexities of information transfer in ectopic recombination are not due to technical artifacts.

Crossing over and reciprocal translocations: As another means of comparing the mechanisms of ectopic and allelic recombination, we determined the frequency of crossing over associated with ectopic intragenic recombination. Crossing over between similarly-oriented loci on nonhomologous chromosomes results in a balanced translocation, the only type of rearrangement likely to produce a viable haploid spore. An assessment of the frequency of crossing over is gained from test crosses to strains with ade6⁺ integrations (Figure 2 and Table 5). Recombinants with balanced translocations should produce only Ade⁺ progeny in both test crosses because the translocation chromosomes show “pseudolinkage;” i.e., the translocation chromosomes must segregate together to produce a viable haploid spore (Figure 2; Szankasiet al. 1986). As confirmation, Southern blot analysis of a subset of these recombinants showed a pattern consistent with reciprocal translocations (Figure 3). In each lane with a reciprocal recombinant, as determined by the test crosses, both parental bands were lost and two new recombinant bands appeared. In crosses with such recombinants the calculated frequency of crossing over associated with ectopic intragenic recombination was 34–58% (Table 5), consistent with previous observations in allelic ade6 recombination (Grimmet al. 1994). The proportion of crossovers was similar for M26 and M375. In two crosses, only one or two of ~90 recombinants examined did not segregate Ade⁻ spores in either test cross (Table 5, lines 1 and 10). The nature of these rare recombinants has not been further studied. Based on the pattern of translocations obtained with the other crosses, simple balanced translocations that create pseudolinkage were not expected from the crosses in lines 1 and 10 (see below).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Segregation patterns in test crosses. The possible segregation patterns for the first (reductional) meiotic division are shown for test crosses between Ade⁺ recombinants from an original cross (e.g., Tables 2 and 3) and Ade⁺ tester strains (Table 1). The two chromatids are shown as a single bar and the second (equational) meiotic division is omitted. In the examples shown, recombinants originated from a cross between a strain carrying an ade6-M26 allele at the endogenous ade6 locus on chromosome III (gray bar) and a strain carrying an ade6-469 allele at the z7 integration site on chromosome II (black bar). The M26 allele was converted to ade6⁺, without (A) or with (B) associated crossing over. Open circle, centromere; open box, ade6 allele; open box with +, ade6⁺ allele. (A) Nonreciprocal recombination (conversion without crossing over) in the original cross. The Ade phenotype (+ or −) of each spore from the test crosses is indicated. Only two of the four spores are shown for each segregation pattern. The other two spores generated from the second meiotic division would be identical to the pair illustrated. The Ade⁺ tester strain that produces Ade⁻ spore colonies in a cross with the recombinant identifies the donor locus in the original cross. (B) Reciprocal recombination (conversion with associated crossing over) in the original cross, generating a balanced translocation. Both the Ade phenotype and predicted spore viability are indicated. Because the translocation chromosomes must segregate together to produce a balanced genome and a viable spore, no viable Ade⁻ spores are produced.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Southern blot with meiotic recombinants. DNA from parental (lanes 1, 20, 21, and 40) and recombinant (lanes 2–10, 12–19, 22–29, and 31–39) strains was digested with ClaI prior to electrophoresis and hybridization with pJV8 (Figure 1). Lanes 11 and 30 represent phage lamba DNA digested with HindIII. Above each lane is the designation for reciprocal (R) or nonreciprocal (N) recombinant, according to results from test crosses (Figure 2). In the lanes with DNA from parental strains, the bands representing the ade6 and z7 loci are indicated. In the lanes with DNA from recombinants, the bands corresponding to the translocation chromosomes are indicated. The band assignments in the reciprocal recombinants were made by probing the same blot with probes specific for the S. pombe genomic DNA at each end of the z7 integration (data not shown).

The production of viable translocations, or lack thereof, was combined with chromosomal position information to infer the orientation of the different integrations relative to the centromere (Figure 1). All of the Ade⁺ recombinant progeny tested from some crosses produced Ade⁻ spores in one test cross (Table 5, lines 2, 5–7, 11, and 12), inferring that the original crosses did not produce viable balanced translocations. When the two recombining loci in the original cross were on different chromosomes, then the lack of translocations was most likely due to the ade6 alleles being in opposite orientations relative to their respective centromeres. In this case, crossing over would result in a dicentric chromosome and an acentric fragment and spore inviability. Therefore, the ade6 and z15 loci are inferred to be in opposite orientations relative to cenIII and cenI, respectively, and ade6 and z7 are inferred to be in the same orientations relative to cenIII and cenII, respectively. The ura4-117(3.0) and ura4-154(3.0) integrations produced translocations with z15 but not z7, a pattern opposite to the endogenous ade6 gene. Combining this information with the known orientation of the ade6 gene relative to cenIII (Grimmet al. 1994), the inferred orientations of the ade6 gene in the different integrations are shown in Figure 1.

In crosses between the endogenous ade6 locus and the ura4-117(3.0) and ura4-154(3.0) integrations (Table 5, lines 10 and 11), which are on opposite arms of chromosome III and in opposite orientations relative to cenIII, crossing over would result in a dicentric chromosome III and an acentric fragment. A dicentric chromosome III resulting from crossing over between the ade6 locus and the ura4 integrations might result in a viable spore if one centromere were inactivated. However, a dicentric chromosome III would not be expected to cause pseudolinkage in the same way as a heterochromosomal translocation.