This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yeadon, P. J. Articles by Catcheside, D. E. A. Search for Related Content PubMed PubMed Citation Articles by Yeadon, P. J. Articles by Catcheside, D. E. A.

Alleles of the Hotspot cog Are Codominant in Effect on Recombination in the his-3 Region of Neurospora

School of Biological Sciences, Flinders University, Adelaide, South Australia, 5001 Australia

¹ Corresponding author: School of Biological Sciences, Flinders University, Box 2100, Adelaide, South Australia, 5001 Australia.
E-mail: david.catcheside{at}flinders.edu.au

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 
There are two naturally occurring functional alleles of the recombination hotspot cog, which is located 3.5 kb from the his-3 locus of Neurospora crassa. The presence of the cog⁺ allele in a cross significantly increases recombination in the his-3 region compared to a cross homozygous for the cog allele. Data obtained shortly after discovery of cog⁺ suggested that it was fully dominant to cog. However, a dominant cog⁺ conflicts with observations of hotspots in Saccharomyces cerevisiae and Schizosaccharomyces pombe, in which recombination is initiated independently of homolog interactions, and suggests recombination mechanisms may differ in Neurospora and yeast. We present evidence that cog alleles are codominant in effect on both allelic recombination in his-3 and crossing over between loci flanking his-3. In addition, we show that genetic background variation has at least a twofold effect on allelic recombination. We speculate that variation in genetic background, together with the complexities of recombination in crosses bearing close mutant alleles, accounts for the previous conclusion that cog⁺ is dominant to cog.

Crossovers are not randomly distributed along chromosomes but tend to be clustered (LICHTEN and GOLDMAN 1995; BAUDAT and NICOLAS 1997; JEFFREYS et al. 1998) in regions termed hotspots (HOLLIDAY 1968). Tetrad analyses indicate that crossovers are more likely close to a locus that has experienced conversion (OLIVE 1959; STADLER 1959; MURRAY 1960; LISSOUBA et al. 1962). In addition, evidence of gene conversion has been found close to hotspots for crossing over (BORTS and HABER 1989; JEFFREYS et al. 2001; GUILLON and DE MASSY 2002; JEFFREYS and NEUMANN 2002). The association between conversion and crossing over suggests that the same mechanism is responsible for both outcomes and this has been a central feature of recombination models (SZOSTAK et al. 1983; SUN et al. 1991; NASSIF et al. 1994; PâQUES and HABER 1999). Recent studies suggest that crossover and noncrossover products arise from different pathways and that the decision between the two outcomes is made after initiation but before production of a recombination intermediate (ALLERS and LICHTEN 2001; HUNTER and KLECKNER 2001). However, both conversion and crossing over are thought to be initiated by the same mechanism.

In Saccharomyces cerevisiae, initiation of recombination is by a double-strand break (DSB) in one homolog, probably generated by the Spo11 protein (KEENEY et al. 1997). SPO11 homologs have been found in all eukaryotes in which they have been sought, including flies (MCKIM and HAYASHI-HAGIHARA 1998), worms (DERNBURG et al. 1998), mammals (ROMANIENKO and CAMERINI-OTERO 1999), plants (GRELON et al. 2001), and the filamentous ascomycete Neurospora crassa (F. J. BOWRING, P. J YEADON, R. J. STAINER and D. E. A. CATCHESIDE, unpublished results), suggesting conservation of the initiation mechanism.

The Neurospora recombination hotspot cog, located centromere-distal of the his-3 locus (BOWRING and CATCHESIDE 1991; YEADON and CATCHESIDE 1995a, 1998), influences allelic recombination within his-3 and crossing over in the chromosomal segments surrounding the gene (ANGEL et al. 1970). Two cog phenotypes, high (cog⁺) and low (cog) frequency recombination, have been described (ANGEL et al. 1970), with the chromosome that bears cog⁺ almost exclusively experiencing conversion (CATCHESIDE and ANGEL 1974; YEADON and CATCHESIDE 1998). Although there are multiple differences between the cog region (YEADON and CATCHESIDE 1995a) sequences of cog^Ea, cog^LA, and cog^EA, all of which are cog, and cog^La, the only naturally occurring cog⁺ allele known (YEADON and CATCHESIDE 1995b, 1999), a 10-bp sequence including two single-nucleotide polymorphisms (SNPs) is required for the high-frequency recombination phenotype (YEADON and CATCHESIDE 1998). Recombination is known to be initiated >2.2 kb from the 3' end of his-3 (YEADON et al. 2001) and a peak in conversion close to these SNPs (YEADON and CATCHESIDE 1998) suggests that initiation may occur at this location, which is 3.4 kb from his-3.

The presence of cog⁺ increases allelic recombination frequency 6-fold and crossovers between his-3 and the centromere-distal gene, ad-3, 4-fold when compared to similar crosses in which cog is homozygous (CATCHESIDE and ANGEL 1974). The trans-acting rec-2 gene imposes an additional level of regulation of recombination in this region of LG I as the dominant allele, rec-2⁺, has an epistatic effect (SMITH 1968; CATCHESIDE 1979). In the presence of rec-2⁺, recombination between his-3 alleles is reduced 30-fold in crosses containing cog⁺ and 4-fold in crosses of homozygous cog to the same low level (ANGEL et al. 1970). In addition, recombination events that occur in the presence of rec-2⁺ appear to be initiated at the 5' end of his-3 and not at cog (CATCHESIDE and ANGEL 1974; YEADON and CATCHESIDE 1998). It seems likely that the allelic recombination frequency attributable to each cog allele reflects the frequency with which recombination is initiated there and that the rec-2⁺ product prevents initiation at either cog allele (CATCHESIDE and ANGEL 1974).

Study of haploid meiosis indicates that the timing, frequency, and distribution of DSBs are independent of interhomolog interaction in S. cerevisiae (DE MASSY et al. 1994; GILBERTSON and STAHL 1994) and in Schizosaccharomyces pombe (YOUNG et al. 2002), suggesting that recombination is initiated independently at each allele of a particular hotspot. If recombination is initiated by a DSB at cog in Neurospora and each initiation is independent of initiation at the other cog allele, the six-fold increase in allelic recombination in crosses heterozygous for cog⁺/cog (CATCHESIDE and ANGEL 1974) implies that DSBs occur 11 times more frequently at cog⁺ than at cog and predicts that a heteroallelic cross homozygous for cog⁺ would yield close to twice as many recombinants as one in which cog⁺ is heterozygous. However, ANGEL et al. (1970) found that, in crosses heteroallelic for his-3 K26/K874, there was little difference in the allelic recombination frequency between cog⁺ homo- and heterozygotes, leading to the conclusion that cog⁺ is fully dominant to cog. Although rec-2⁺ reduces allelic recombination in cog/cog diploids fourfold, there is no apparent decrease in crossovers between his-3 and ad-3 in the same diploids, even though CATCHESIDE and ANGEL (1974) estimated that it should have been detectable.

In yeasts, no naturally polymorphic recombination hotspots have yet been found. The ade6-M26 mutation in S. pombe increases conversion in ade6 10- to 15-fold when compared to the closely linked ade6-M375 mutation (GUTZ 1971; FOX and SMITH 1998). During recombination M26 is preferentially converted to wild type (GUTZ 1971), so the M26 chromosome, like that carrying cog⁺, is usually the recipient of information. The use of an opal suppressor mutation, sup9, allowed measurement of meiotic intragenic recombination with M26 heterozygous or homozygous (PONTICELLI et al. 1988). PONTICELLI et al. (1988) concluded that crosses homozygous for M26 yielded 10 times more recombinants than those lacking M26. Moreover, recombination frequency in M26 homozygotes was approximately the sum of the two heterozygous frequencies. However, since ade6-M26 sup9 spores form colonies only 50% as efficiently as ade6⁺ sup9 spores, PONTICELLI et al. (1988) doubled the numbers of recombinants in the homozygous assay to reach this conclusion, so codominance of M26 and wild-type ade6 hotspot alleles is far from certain.

Like cog⁺, M26 increases crossing over nearby. The substitution of M26 for M375 results in a 2.5-fold increase in intrachromosomal crossing over, from 0.3 to 0.8% (SCHUCHERT and KOHLI 1988). In contrast, the ura4-aim–tps16 genetic interval flanking ade6 is not strongly affected by the presence of M26. The genetic distance increases to 12.5 cM, compared to the 11.8 cM measured in the absence of M26 (ZAHN-ZABAL et al. 1995). ZAHN-ZABAL et al. (1995) found that M26 convertants experience exchange between ura4-aim and tps16 at the same frequency as M375 convertants and concluded that the slight increase in crossing over is due to the higher frequency of conversion at M26 and the resultant increase in conversion-associated crossovers.

With the exception of ade6-M26, all other artificial hotspot polymorphisms have been generated by deletion of part of the promoter region of the gene in which recombination was studied. Deletion of the ade6 promoter removes the hotspot activity of M26 only when the deletion is in cis to M26, with no effect of the deletion in trans (ZAHN-ZABAL et al. 1995). Strangely, the conversion frequency in ade6 in the absence of M26 is unaffected by the same deletion, whether in cis or in trans to M375.

In S. cerevisiae, homozygous deletion of a poly(dA·dT) tract in the promoter region of ARG4 (9) reduced conversion of the arg4-RV mutation to 0.8% from the wild-type level of 7.4% (NICOLAS et al. 1989). Unlike the ade6 promoter deletions in S. pombe (ZAHN-ZABAL et al. 1995), diploids heterozygous for 9 yielded a similarly low frequency of arg4-RV conversion of 1% (6 tetrads of 562; NICOLAS et al. 1989). To our knowledge, there appear to be no data on the effect of ARG4 9 on crossing over.

Regions in which the crossover rate is elevated also exist in the human genome (JANSON et al. 1991; OUDET et al. 1992; HUBERT et al. 1994). Recombination hotspots have been identified at several human loci (CHAKRAVARTI et al. 1984; OUDET et al. 1992; YIP et al. 1999), including within the human major histocompatibility complex class II region (CULLEN et al. 1995). At one of the six hotspots in this region (JEFFREYS et al. 2001), DNA2, the FG11G/A polymorphism, appears to alter the crossover frequency in sperm (JEFFREYS and NEUMANN 2002). Haplotype-specific PCR primers were used to amplify recombinant DNA molecules from sperm taken from men heterozygous for various SNPs within the 5.5-kb amplification region. Two FG11G/G homozygotes yielded recombinant molecule frequencies of 0.1 x 10^–5 and 0.7 x 10^–5 while three of five A/G heterozygotes gave frequencies of between 2 x 10^–5 and 3 x 10^–5 (with the most extreme values in the other two men at 0.9 x 10^–5 and 10 x 10^–5), suggesting that FG11A significantly increases the activity of the DNA2 hotspot (JEFFREYS and NEUMANN 2002). In addition, in recombinant molecules from FG11A/G heterozygotes, the FG11G SNP is overrepresented, suggesting that, as in other hotspot allele heterozygotes (GUTZ 1971; CATCHESIDE and ANGEL 1974; NICOLAS et al. 1989), the FG11A strand is usually the recipient of information (JEFFREYS and NEUMANN 2002). Since the crossover frequency (2.6 x 10^–5) in the single A/A homozygote falls in the middle of the range for A/G heterozygotes, one might conclude that the FG11A hotspot allele is fully dominant to the FG11G allele. However, the wide range of frequencies in the heterozygotes shows that factors other than the FG11 SNP have a large effect on crossing over at DNA2, making conclusions drawn from a single homozygote highly unreliable.

In recent years, we have constructed Neurospora strains that carry cog⁺ in cis to his-3 mutant alleles K480, K504, and K1201, allowing analysis of additional allele pairs, which are farther apart than the 215 bp separating K26 and K874 (YEADON and CATCHESIDE 1999). K26/K874 diploids yield a low frequency of His⁺ progeny compared to those with more distant pairs of mutant sites (ANGEL et al. 1970), increasing the chance that measured recombination frequencies will be confounded by random variation. Higher recombination frequencies yielded by distant allele pairs reduce the impact of random factors and are likely to aid differentiation between His⁺ frequencies from cog⁺ hetero- and homozygotes. We have also constructed both cog⁺ and cog stocks that are mutant in the centromere-proximal gene lys-4, and cog⁺, cog, and rec-2⁺ strains that are mutant in the centromere-distal gene ad-3, allowing measurement of the effect of all possible cog and rec-2 genotypes on exchange in the his-3 region.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 
Construction of strains:
The genotypes of all strains used in this study are listed in Table 1. The his-3 mutation K26 was generated in a Lindegren Y8743 strain (ANGEL et al. 1970) and thus all nonrecombinant K26 strains have the high-frequency (cog⁺) recombinator allele, cog^La. K480, K874, and K1201 mutations were generated in Emerson a strains (ANGEL et al. 1970) and so all nonrecombinant strains bearing these alleles have the low-frequency (cog) recombinator allele, cog^Ea. The cog allele found in St. Lawrence 79a, cog^S79a (YEADON and CATCHESIDE, 1995a), has the same sequence as cog^La and is also cog⁺ (YEADON and CATCHESIDE 1995b). Lindegren A carries an allele of cog with a sequence different from that of cog^La and cog^Ea (YEADON and CATCHESIDE, 1995b, 1999), but appears to have a low-frequency recombination phenotype identical to that of cog^Ea.

T11317, T12010, T12011, T12012, and T12013 are descendants of T11281 and are therefore all K1201 cog⁺. Likewise, T10997, T10998, and T9149 are descendants of T6275 and are K874 cog⁺. T10997 carries the dominant rec-2⁺, so crosses to this strain have substantially reduced recombination in the his-3 region (SMITH 1968; ANGEL et al. 1970). T12078–T12081 are His⁺ progeny of T11805 and T10998. T11997, a K1201 descendant of T11317 and phenotypically cog, was generated by an unselected conversion event within cog, which was detected by sequencing. T11311 and T11313 (his-3 K874 cog) each have his-3 from Emerson a and cog^LA. F3300, from the collection of D. G. Catcheside, is rec-2⁺ and supposedly cog⁺, although SNP analysis suggests it is cog.

T11039, T11041, and T11043 [Fungal Genetics Stock Center (FGSC) nos. 6526, 6077, and 6098, respectively] were generated in St. Lawrence 74A (OVERTON et al. 1989) as was cog (YEADON and CATCHESIDE 1995a). T11058 and T11059 are progeny of T11039, T11061, and T11062, and T11063 of T11041, T11065, and T11066 from FGSC no. 6085 (A, his-3 1-226-0503, cog^Ea rec-2⁺), also generated in St. Lawrence 74A (OVERTON et al. 1989), and T11067 of T11043. Thus, these strains carry a variety of mutant his-3 alleles, but are all cog.

T11782 and T11789 were made by replacement transfection of the his-3 K458 recipient strains T11644 and T11630 (YEADON et al. 2001), respectively, with a PCR product including his-3 K26.

Culture methods and media:
These were as described by BOWRING and CATCHESIDE (1996), except that crosses were supplemented with 200 µg/ml L-histidine, 500 µg/ml L-alanine, 500 µg/ml L-arginine, 200 µg/ml adenine, and 400 µg/ml L-lysine as required. Vegetative cultures were supplemented with 200 µg/ml L-histidine, 500 µg/ml L-arginine, 500 µg/ml L-alanine, 400 µg/ml adenosine, and 400 µg/ml L-lysine as required.

Recombination assays:
Ascospores were harvested from a single crossing tube in distilled water. After estimation of the number of spores by hemocytometer, an appropriate volume was added to 20-ml layer agar (0.8% Difco agar, 2% sucrose, 2% Vogel's N medium) kept at 60°. Following serial dilution in layer agar and incubation at 60° for 45–70 min, 3-ml aliquots of the highest and lowest dilutions were plated onto selective and nonselective medium, respectively. Plates were incubated overnight at 20° and then moved to 34° for 24–48 hr to express cot-1 and to restrict colony size. The dilution factor between selective and nonselective plates varied from 1/1600 to 1/100 for allelic recombination assays and from 1/10 to 1/100 for intergenic assays, depending on the recombination frequency.

Statistical analysis of recombination data:
For his-3 K1201/K874 heterozygotes (Figure 1), his-3 K26/K874 heterozygotes (Figure 6), lys-4 ad-3 (Figure 4), and his-3 ad-3 trans-heterozygotes (Figure 5), data from crosses homozygous for cog⁺ were compared to those from crosses heterozygous for cog⁺/cog. The significance of any difference between frequency distributions was determined by a two-tailed t-test. Since the data are expressed as frequencies, each frequency was transformed (P sin^–1 ) before comparison. Each comparison of frequency distributions was also subjected to an f-test to assess the level of variance in each distribution and thus to determine whether to perform a t-test for equal or unequal variances. To determine heterogeneity of recombination frequencies from repeats of a single cross or within a single genotype, a ² test was used to assess the probability that the colony counts could differ by chance. Where repeat crosses gave homogeneous counts and the data set was large (Figures 1 and 2), repeat counts were combined to give a single recombination frequency.

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   FIGURE 1.— Allelic recombination in diploids heteroallelic for K1201/K874. The centromere is at the left of the figure. The mutant sites are separated by 1605 bp (YEADON and CATCHESIDE 1999). PF is the frequency of His+ progeny yielded by each cross, multiplied by 105. His+ frequencies were obtained from cog+ homozygotes (A: T12010 x T4395, T12011 x T4395, T12012 x T4395, T12013 x T4395, T12010 x T11126, T12010 x T11153, T12011 x T11113, T12011 x T11126, T12011 x T11132, T12011 x T11153, T11805 x T4395, T11805 x T10998, T11317 x T6275, and T11317 x T9149), cog/cog+ (B: T11997 x T4395, T11997 x T11110, T11997 x T11113, T11997 x T11125, T11997 x T11132, T11997 x T11153, and T11801 x T10998), cog+/cog heterozygotes (C: T11805 x T11089, T11317 x T11045, T11281 x T11311, and T11281 x T11313), cog homozygotes (D: T11801 x T11089, T9144 x T11311, T9144 x T11313, and T9144 x T11045), and a rec-2+ heterozygote (E: T11805 x T10997).

View larger version (31K): [in this window] [in a new window] [Download PPT slide]   FIGURE 6.— Allelic recombination in crosses heteroallelic for K874/K26. The centromere is at the left of the figure. The mutant sites are separated by 215 bp (YEADON and CATCHESIDE 1999). PF is the frequency of His+ progeny yielded by each cross, multiplied by 105. His+ frequencies were obtained from cog+ homozygotes (A: T10989 x T11113, T10989 x T11110, T10989 x T11125, T10989 x T11126, and T10989 x T11132), cog/cog+ (B: T11318 x T4395, T11318 x T11113, T11318 x T11125, and T11318 x T10998), cog+/cog heterozygotes (C: T10989 x T11092, T10989 x T11093, T10989 x T11099, T10989 x T11104, T10989 x T11105, T10989 x T11117, and T10989 x T11127), and cog homozygotes (D: T11318 x T4396, T11318 x T11093, T11318 x T11099, T11318 x T11104, T11318 x T11105, and T11318 x T11127).

View larger version (22K): [in this window] [in a new window] [Download PPT slide]   FIGURE 4.— Crossing over between lys-4 and ad-3. The centromere, to the left of the figure, is represented by a solid disc. The figure is not to scale. The frequency of Lys+ Ad+ progeny is given as a fraction of the total number of viable spores. Lys+ Ad+ frequencies were obtained from four replicate crosses of a single cog+ homozygote (A: T11805 x T10998), three cog+/cog heterozygotes (B: T11089 x T11805, T10998 x T11801, and T10998 x T11802), two cog homozygotes (C: T11089 x T11801 and T11089 x T11802), and three rec-2+ heterozygotes (D: T10997 x T11801, T10997 x T11802, and T10997 x T11805).

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   FIGURE 5.— Crossing over between his-3 and ad-3. The centromere is on the left. The figure is not to scale. The frequency of His+ Ad+ progeny is given as a fraction of the total number of viable spores. His+ Ad+ frequencies were obtained from cog+ homozygotes (A: T12080 x T11789, T12078 x T11782, T12081 x T11789, and T12079 x T11782), cog+/cog heterozygotes (B: T11081 x T11058, T4400 x T4395, T11081 x T11089, T11081 x T11057, T11668 x T11782, T11667 x T11789, and T4400 x T11113), cog homozygotes (C: T4400 x T4396, T4400 x T11067, T4400 x T4396, T4400 x T11066, T4400 x T11059, T4400 x T11063, T4400 x T11062, T4400 x T11065, T4400 x T11061, and T4400 x T11058), and rec-2+ heterozygotes (D: T4400 x T11039, T4400 x T11043, and T4400 x T11041).

View larger version (29K): [in this window] [in a new window] [Download PPT slide]   FIGURE 2.— Allelic recombination in diploids heteroallelic for K874/K480. The centromere is to the left of the figure. The mutant sites are separated by 888 bp (YEADON and CATCHESIDE 1999; YEADON et al. 2002). PF is the frequency of His+ progeny yielded by each cross, multiplied by 105. His+ frequencies were obtained from a cog+ homozygote (A: T11125 x T11760), two cog/cog+ heterozygotes (B: T4396 x T11760 and T4396 x T11764), a cog+/cog heterozygote (C: T11125 x T11681), and a cog homozygote (D: T4396 x T11681).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 
Allelic recombination frequency is elevated in cog⁺ homozygotes:
While it was thought that cog⁺ is dominant to cog (CATCHESIDE and ANGEL 1974), we find that the His⁺ frequency is higher when cog⁺ is homozygous than when it is heterozygous (Figures 1–3), with the exception of crosses heteroallelic for K26/K874 (Figure 6), discussed below. For crosses heteroallelic for K1201/K874 (Figure 1), the His⁺ frequency from crosses homozygous for cog⁺ (Figure 1A; 786/10⁵ viable spores) is greater (P = 4 x 10^–9) than that (343/10⁵) when cog⁺ and K874 are in cis (Figure 1B). When cog⁺ and K1201 are in cis (Figure 1C), the recombination frequency is somewhat lower (218/10⁵), but still higher than that from crosses homozygous for cog (32/10⁵; Figure 1D). Similarly, for diploids heteroallelic for K874/K480, the recombination frequencies are 257/10⁵, 171/10⁵, 160/10⁵, and 16/10⁵ (Figure 2, A–D, respectively). For diploids heteroallelic for K1201/K26, the recombination frequencies are 253/10⁵, 140/10⁵, 112/10⁵, and 41/10⁵ (Figure 3, A–D, respectively). Therefore, it appears that cog⁺ is not dominant to cog but that the hotspot alleles operate codominantly to influence the frequency of allelic recombination at his-3.

View larger version (29K): [in this window] [in a new window] [Download PPT slide]   FIGURE 3.— Allelic recombination in diploids heteroallelic for K1201/K26. The centromere is to the left of the figure. The mutant sites are separated by 1390 bp (YEADON and CATCHESIDE 1999). PF is the frequency of His+ progeny yielded by each cross, multiplied by 105. His+ frequencies were obtained from cog+ homozygotes (A: T11317 x T10989 and T11281 x T10989), cog/cog+ (B: T9194 x T10988 and T9194 x T11286), cog+/cog heterozygotes (C: T11317 x T11318 and T11281 x T11318), and a cog homozygote (D: T9194 x T11284).

Heterogeneity of assay data and the effect of genetic background on recombination:
Repeat assays of the same cross almost invariably yield consistent recombination frequencies. For example, three assays of T11997 x T11110 gave His⁺ frequencies of 327/10⁵, 330/10⁵, and 346/10⁵ (² heterogeneity test gives P = 0.88). Four repeat Lys⁺ Ad⁺ assays of T11805 x T10998 also gave homogeneous crossover frequencies (P = 0.07), and similar results were obtained for 13 other repeat assays (P values range from 0.02 to 0.77). These data show that technical variations, sampling error, and other random factors have little effect on the recombination frequencies measured in this study. (Recombination assay data are available in Tables A1 and A2 in an electronic appendix at http://www.genetics.org/supplemental/).

In contrast, recombination frequencies from crosses of strains with the same cog, his-3, and rec-2 alleles can be variable. Since recombination frequencies do not vary between repeats of the same cross, other factors must be affecting recombination in these crosses. For crosses of the his-3 K1201 cog strain T11997 to T11110, T11125, and T11132 (Figure 1C), three his-3 K874 cog⁺ strains extracted from a single cross, recombination frequencies are homogeneous (² heterogeneity test gives P = 0.32). A similar result is obtained for crosses of T10989 (his-3 K26 cog⁺) to the same three strains (Figure 6A; P = 0.10). However, if we include crosses of T11997 and T10989 to T11113 (Figures 1B and 6A, respectively), which has the same parents as T11110, T11125, and T11132, in each case the data become less homogeneous (P = 0.0002 and 0.02, respectively). When T10989 is crossed to T11092, T11093, T11099, T11104, T11105, T11117, and T11127 (Figure 6C), which are his-3 K874 cog strains with the same parents as T11110, T11113, T11125, and T11132, the recombination frequencies are substantially heterogeneous (P = 1.6 x 10^–8). In addition, crosses of the his-3 K1201 cog⁺ strain T12011 to his-3 K874 cog⁺ strains T11113, T11126, T11132, and T11153 (Figure 1A) yield homogeneous recombination frequencies (P = 0.22), but data from T12011 x T4395 (a his-3 K874 cog⁺ strain made >40 years ago; ANGEL et al. 1970; Figure 1A) are substantially heterogeneous (P = 1 x 10^–17). These data suggest the existence of more than one gene, each with a small effect on recombination, and that the parents of the K1201 and K874 strains described above carried different alleles of these genes.

Crossover frequency also varies within crosses of a single known genotype (Figures 4 and 5). T12078, T12079, T12080, and T12081 are his-3⁺ cog⁺ ad-3 progeny of a cross between T11805 and T10998. The frequency of His⁺ Ad⁺ progeny from crosses of these strains to T11782 and T11789 (his-3 K26 cog⁺) falls into two distinct groups (² yields P = 4 x 10^–9), with crosses to T12078 and T12080 yielding frequencies of 8.8% (P = 0.99) and those to T12079 and T12081 yielding lower frequencies of 6.6 and 6.0%, respectively (P = 0.29; Figure 5). A likely explanation is that alleles of a gene that affects the frequency of crossing over in the his-3 region are segregating in the progeny of T11805 and T10998. T12078 and T12080 received the higher-frequency allele, and T12079 and T12081 the lower-frequency allele.

Allelic recombination in K874/K26 heterozygotes:
Crosses homozygous for cog and heteroallelic for K26/K874 yield a very low frequency of His⁺ progeny (3/10⁵; Figure 6D), so we expect the contribution of initiation at cog to have little effect on recombination in cog/cog⁺ heterozygotes. When cog⁺ and K874 are in cis (Figure 6B), the His⁺ frequency is 29/10⁵-44/10⁵ and when cog⁺ and K26 are in cis (Figure 6C), it is 14/10⁵–30/10⁵. From crosses homozygous for cog⁺ (Figure 6A), the His⁺ frequency is 21/10⁵–31/10⁵ viable spores. Thus, there is no apparent increase in His⁺ frequency when cog⁺ is homozygous (for A and B, a t-test yields P = 0.03; A and C, P = 0.39; B and C, P = 0.01). These data confirm the previous results (ANGEL et al. 1970; CATCHESIDE and ANGEL 1974) and show how analysis of a few K874/K26 heterozygotes led to the conclusion that cog⁺ is fully dominant to cog.

However, His⁺ frequencies in K26 by K874 crosses are heterogeneous for all genotypes in which cog⁺ is present (Figure 6, A–C; ² heterogeneity tests yield P = 0.02, P = 0.001, and P = 1.6 x 10^–8, respectively), suggesting the influence of factors other than cog.

Flanking marker exchange in progeny experiencing allelic recombination:
Forty-five percent (57/128) of His⁺ progeny of a K1201/K874 diploid, homozygous for cog⁺, were recombinant for flanking markers lys-4 and ad-3. Similar frequencies (P = 0.92) were obtained from K1201/K874 diploids heterozygous for cog⁺/cog (44% or 55/126), homozygous for cog (43% or 51/119), or heterozygous for rec-2⁺ (40% or 50/124).

The effect of rec-2⁺ on allelic recombination and on crossing over:
As seen in crosses heteroallelic for K1201/K874 (Figure 1), the presence of rec-2⁺ significantly reduces allelic recombination compared to that measured in crosses homozygous for cog (P = 8 x 10^–4). In contrast, rec-2⁺ does not significantly reduce the frequencies of either Lys⁺ Ad⁺ or His⁺ Ad⁺ spores from those seen in crosses homozygous for cog (Figures 4 and 5) with two-tailed t-tests for equal variances yielding values of P = 0.4 and 0.3, respectively.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 
cog alleles are codominant:
Two copies of cog⁺ yield twice as many local crossovers as a single copy. The lys-4–ad-3 interval is 20 cM in crosses homozygous for cog⁺, 9.4 cM in those heterozygous for cog⁺/cog, and 2 cM in those homozygous for cog (Figure 4). Similarly, the his-3–ad-3 interval is 15.5 cM in crosses homozygous for cog⁺, 7.5 cM in crosses heterozygous cog⁺/cog, and 1.4 cM in those homozygous for cog (Figure 5). Alleles of cog are therefore codominant in effect on local crossing over.

In addition, in heteroallelic crosses, providing that the mutant alleles are distant (Figures 1–3), two copies of cog⁺ result in approximately twice the average frequency of allelic recombination as that of a single copy. Thus we conclude that cog⁺ is not dominant to cog, but rather that the two alleles operate independently of one another to attract recombination events.

The relative frequency of recombination initiation at cog and cog⁺:
In any comparison of crosses carrying the same pair of mutant his-3 alleles, the His⁺ frequency should be directly related to the rate of initiation at the cog hotspot. This rate, in the absence of rec-2⁺, is dependent on the cog alleles in the cross.

For crosses heteroallelic for K1201/K874, the average His⁺ frequency (786/10⁵ viable spores) from crosses homozygous for cog⁺ is 25 times higher than that (32/10⁵) from crosses homozygous for cog. For crosses heteroallelic for K874/K480, the ratio is 16 (257/16) and for those heteroallelic for K1201/K26, only 6 (253/41). The previous estimate of an 11-fold increase in DSBs at cog⁺ relative to cog, based on data from crosses heteroallelic for K874/K26 (ANGEL et al. 1970), is consistent with our estimate using the same allele pair (9-fold, or 27/3) and falls in the middle of the range of our estimates. It is possible therefore that recombination is initiated as much as 25 times more frequently at cog⁺ than at cog.

Genetic background variation alters recombination frequency:
Although recombination frequency in the his-3 region depends upon which alleles of cog and rec-2 are present, it seems that these are not the only factors involved. His⁺ frequency, for a single pair of mutant his-3 alleles, varies over a twofold range in crosses with identical cog and rec-2 genotypes (Figures 1– 3 and 6). Crossover frequency varies in a similar way (Figures 4 and 5). Analysis of recombination frequencies from crosses between strains with the same or similar genetic backgrounds suggests that genes with small effects on recombination segregate in our laboratory strains. Such an effect has been detected previously, where allelic recombination at the nit-2 locus was found to vary with parental provenance (CATCHESIDE 1970).

The effect of rec-2⁺ on crossing over in the his-3 region:
In a cog homozygote, recombination in his-3 is reduced fourfold when rec-2⁺ is present (Figure 1 and CATCHESIDE and ANGEL 1974) and that which persists appears to be initiated from the rec-2⁺-independent hotspot at the 5' end of his-3 (CATCHESIDE and ANGEL 1974). Therefore, if crossovers resulting from initiation at this or other rec-2⁺-independent hotspots in the lys-4–ad-3 interval occur autonomously, unaffected by those generated by initiation at cog, the absence of rec-2⁺ in a cross should increase crossing over in this region. However, between lys-4 and ad-3, crossovers occur at an average frequency of 2.0% in cog rec-2 homozygotes and 1.7% in crosses including rec-2⁺ (Figure 4, C and D). In the his-3–ad-3 interval, the equivalent average frequencies are 1.4 and 1.0%, respectively (Figure 5, C and D). The crossover frequency for each interval in cog homozygotes is thus unaffected by the presence of rec-2⁺ (P = 0.4 and 0.3, respectively). How can this be?

Perhaps we simply failed to detect the increase in crossing over due to the absence of rec-2⁺. We consider this unlikely, as we detected a difference between 0.8 and 0.3% in mean His⁺ frequency for crosses heteroallelic for K1201/K874 (Figure 1, A compared to B; P = 2 x 10^–6), despite highly heterogeneous data (for A, P = 4 x 10^–19; for B, P = 3 x 10^–250). To obtain the crossover data, a similar number of colonies were counted but spore suspensions experienced fewer dilutions than in estimations of the yield of His⁺ progeny from K1201/K874 crosses, thus decreasing sampling error and increasing our chance of differentiating between the frequencies. In addition, the His⁺ Ad⁺ frequencies (Figure 5, C and D) are no more heterogeneous (for C, P = 1 x 10^–88, and for D, P = 1 x 10^–9) than the K1201/K874 His⁺ frequencies, so it seems improbable that a real difference in frequency due to the presence of rec-2⁺ has been confounded by variation in genetic background.

If events initiated at cog were more likely than those initiated at cog⁺ to proceed by synthesis-dependent strand annealing (NASSIF et al. 1994; PâQUES and HABER 1999), the absence of rec-2⁺ would stimulate conversion but not crossing over in a cog homozygote. If this were the case, His⁺ progeny of cog homozygotes would experience fewer crossovers than His⁺ progeny extracted from crosses where cog⁺ is present. However, the frequency of flanking marker exchange is the same in His⁺ progeny from all crosses heteroallelic for K1201/K874 (P = 0.92).

We must therefore conclude that our original assumption, that rec-2⁺-independent and cog-associated crossovers occur autonomously, is incorrect and that reduction of the latter yields an increase in the former type of crossover. It may be that events are initiated autonomously but that, as cog-associated crossovers decrease, events initiated elsewhere have an increased chance of yielding crossovers, the phenomenon of crossover interference (MULLER 1916). Alternatively, since competitive interaction between two nearby hotspots in S. cerevisiae has been observed to reduce the activity of both (XU and KLECKNER 1995; FAN et al. 1997), the lack of DSBs at cog when rec-2⁺ is present may increase the frequency of initiation at other locations in the his-3 region.

cog⁺ appears dominant in K26/K874 heterozygotes:
In crosses heteroallelic for distant his-3 alleles, a cog⁺ homozygote gives a His⁺ frequency close to the sum of the heterozygotes, suggesting that the His⁺ frequency is determined by the frequency of recombination initiation. In contrast, in crosses heteroallelic for K26 and K874 (Figure 6A), which are 215 bp apart (YEADON and CATCHESIDE 1999), cog⁺ homozygotes yield a His⁺ frequency lower than that of one of the heterozygotes (P = 0.03), despite a presumed doubling in initiation frequency. Alleles in close proximity experience co-conversion more often than widely separated sites (HILLIKER et al. 1994; YEADON et al. 2002). Therefore, a recombination event initiated at cog (YEADON et al. 2001) on the chromosome bearing K874 is less likely to terminate between K874 and K26 to yield a His⁺ spore than is a more distant allele pair. In addition, when nearby alleles are co-converted, both mismatches may be included in a single repair tract (MODRICH and LAHUE 1996). Thus, the probability that a conversion event initiated on the K26 chromosome and covering K874 and K26 will result in a His⁺ spore is reduced compared to a similar event involving remote alleles. However, despite this effect, a cog⁺ homozygote should yield His⁺ progeny at the sum of the two heterozygous frequencies, whether the mutant alleles are close or distant.

Our His⁺ frequencies are in most cases heterogeneous within a single known genotype, suggesting segregation of genes, unlinked to cog, that affect recombination. K26/K874 heterozygotes are no exception, with the frequencies yielded by K874 cog/K26 cog⁺ crosses especially heterogeneous (P = 1.6 x 10^–8). We therefore suggest that variation in genetic background is responsible for the apparent lack of additivity of our K26/K874 recombination frequencies. Since our strains are descendants of those of D. G. Catcheside, genetic background variation is also a likely explanation for the previous nonadditive data (ANGEL et al. 1970; CATCHESIDE and ANGEL 1974). It is clear that recombination involving K874 and K26, the only allele pair available to CATCHESIDE and ANGEL (1974) to test the dominance relationship of cog and cog⁺, is a special case, and that this relationship is more easily investigated by analysis of crosses heteroallelic for more distantly separated alleles.

We have shown that naturally occurring alleles of the recombination hotspot cog are codominant. Since initiation of recombination is thought to be independent of interaction between homologs (DE MASSY et al. 1994; GILBERTSON and STAHL 1994; YOUNG et al. 2002), co-dominance of hotspot alleles was predicted, but not demonstrated prior to this study. This work supports the conclusion that the frequency of conversion in crosses homozygous for ade6 M26 in S. pombe is equal to the sum of the two heterozygous frequencies (PONTICELLI et al. 1988), making it likely that all hotspot alleles operate codominantly to influence recombination nearby.

We also present evidence that genes that influence both conversion and crossing over are polymorphic in laboratory strains of N. crassa. Recombination at the naturally polymorphic human DNA2 hotspot has yielded data (JEFFREYS and NEUMANN 2002) that suggest that genetic background may have a similar effect on recombination in humans. Identification and investigation of the polymorphic genes involved in the genetic background effect in Neurospora may assist with identification of similar polymorphisms present in humans.

	ACKNOWLEDGEMENTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 
This work was supported by grants from the Australian Research Council and the Flinders University Research Budget.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS LITERATURE CITED

 

ALLERS, T., and M. LICHTEN, 2001 Differential timing and control of noncrossover and crossover recombination during meiosis. Cell 106: 47–57.[CrossRef][Medline]

ANGEL, T., B. AUSTIN and D. G. CATCHESIDE, 1970 Regulation of recombination at the his-3 locus in Neurospora crassa. Aust. J. Biol. Sci. 23: 1229–1240.[Medline]

BAUDAT, F., and A. NICOLAS, 1997 Clustering of meiotic double-strand breaks on yeast chromosome III. Proc. Natl. Acad. Sci. USA 94: 5213–5218.[Abstract/Free Full Text]

BORTS, R. H., and J. E. HABER, 1989 Length and distribution of meiotic gene conversion tracts and crossovers in Saccharomyces cerevisiae. Genetics 123: 69–80.[Abstract/Free Full Text]

BOWRING, F. J., and D. E. A. CATCHESIDE, 1991 The initiation site for recombination cog is at the 3' end of the his-3 gene in Neurospora crassa. Mol. Gen. Genet. 229: 273–277.[CrossRef][Medline]

BOWRING, F. J., and D. E. A. CATCHESIDE, 1996 Gene conversion alone accounts for more than 90% of recombination events at the am locus of Neurospora crassa. Genetics 143: 129–136.[Abstract]

CATCHESIDE, D. E. A., 1970 Control of recombination within the nitrate-2 locus of Neurospora crassa: an unlinked dominant gene which reduced prototroph yields. Aust. J. Biol. Sci. 23: 855–865.[Medline]

CATCHESIDE, D. E. A., 1979 Effect of rec-2⁺ on the formation of double mutant recombinants in Neurospora crassa. Aust. J. Biol. Sci. 32: 261–265.

CATCHESIDE, D. G., and T. ANGEL, 1974 A histidine-3 mutant, in Neurospora crassa, due to an interchange. Aust. J. Biol. Sci. 27: 219–229.[Medline]

CHAKRAVARTI, A., K. H. B