Evidence for Negative Interference: Clustering of Crossovers Close to the am Locus in Neurospora crassa Among am Recombinants

Corresponding author: David E. A. Catcheside, School of Biological Sciences, Flinders University, Bedford Park, South Australia, 5042., bidec{at}cc.flinders.edu.au (E-mail)

Communicating editor: P. J. PUKKILA

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In response to a conflict between two mapping studies in the predicted orientation of the allele map with respect to the centromere, Fincham proposed that recombination events at the Neurospora am locus rarely have an associated crossover. Fincham considered that the elevated levels of crossing over between flanking markers in am recombinants resulted from negative interference, an increased probability of a nearby second event, and on this basis predicted a clustering of crossing over near am in these recombinants. In this article we reevaluate the data from three mapping studies of the am locus and report molecular evidence that shows crossovers to be clustered immediately proximal to am in am recombinants.

DESPITE the elegant simplicity of a model for meiotic recombination that accounted for both gene conversion and crossing over (HOLLIDAY 1964 , HOLLIDAY 1974 ), a challenge to the ubiquity of a mechanistic association between gene conversion and crossing over was soon offered by FINCHAM 1974 in considering an apparent conflict between his (FINCHAM 1967 ) and SMYTH's (1973) mapping data for the Neurospora am locus.

am⁺ progeny selected at random from crosses of the type P m¹ m²⁺ D by p m¹⁺ m² d, where P/p and D/d are proximal and distal flanking markers, respectively, while m¹ and m² are different auxotrophic mutations within am, were examined in both mapping efforts. The relative proportions in the two classes with nonparental association of flanking markers, R1 (p m¹⁺ m²⁺ D) and R2 (P m¹⁺ m²⁺ d), differed between studies. Although both authors found that R1 R2 in the majority of their crosses, where there was disparity, Smyth found that R1 was larger than R2, while in certain of Fincham's pairings, R2 exceeded R1. As the class in excess was assumed to indicate which of the two possible gene orientations minimized the number of exchange events required, this disparity led each author to propose a different orientation for the am gene with respect to the centromere. While the am data suggest that the R1:R2 disparity cannot be used to orient genetic maps (BOWRING and CATCHESIDE 1995 ), the "polarity" criterion that relies on the relative proportions of recombinant progeny with a parental configuration of flanking markers describes a reasonably accurate allele order (see Figure 1; BOWRING and CATCHESIDE 1995 ).

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. Flanking marker exchange among am+ progeny from three mapping studies. The correction for incidental exchange is that of PERKINS 1979 : (q - d)/(1 - d) where q is the proportion of am+ recombinants with flanker exchange and d is the standard map distance in morgans. In A, q = 0.22 (90 crosses), d = 0.025 (RAMBOSEK and KINSEY 1983 ); in B, q = 0.295 (59 crosses), d = 0.08 (SMYTH 1973 ); in C, q = 0.399 (65 crosses), d = 0.148 (FINCHAM 1967 ). P of the Pearson r value is shown below each line.

FINCHAM 1974 attempted to reconcile the conflicting conclusions regarding the orientation of am and, without introducing complicating modifications to the Holliday model, proposed that genesis of the R2 class resulted from gene conversion at either m¹ or m² and an independently initiated crossover on the far side of the nonconverted site. This, coupled with the size of the R class (R1 + R2) that accounted for 40 and 30% of recombinants in FINCHAM's (1967) and SMYTH's (1973) mapping efforts, respectively, led Fincham to conclude that recombination at am is due predominantly to conversion without a mechanistically associated crossover. However, given that exchange in the flanking intervals was substantially higher among am⁺ progeny than standard map distances dictated, Fincham proposed a variant of PRITCHARD 1960 "negative interference," suggesting that conversion events in am increased the likelihood of a separately initiated crossover nearby. Fincham therefore predicted that crossovers among am convertants would be clustered near am.

Despite the formulation of sophisticated models in which the R2 class can arise from a single recombination intermediate (e.g., MESELSON and RADDING 1975 ; ORR-WEAVER and SZOSTAK 1985 ), further studies of recombination at am have failed to counter Fincham's early disquiet. Fincham and Smyth used different combinations of flanking markers and, in a later mapping effort, RAMBOSEK and KINSEY 1983 used a third combination. In Figure 1, flanking marker exchange among am⁺ progeny (q) from numerous different allelic pairings in each study (see Figure 1 legend) is plotted against the distance to flanking markers (d) for each of the three studies. Data in the upper line (Observed) have not been corrected for incidental exchange, those crossovers normally expected between flanking markers, so a rise in flanker exchange among am⁺ recombinants corresponding to an increase in flanker separation is expected here. Data in the lower line, however, (Corrected) have those incidental crossovers expected to have occurred by chance removed. The correction for incidental exchange is that of PERKINS 1979 : (q - d)/(1 -d), where q is the proportion of am⁺ recombinants with flanker exchange and d is the standard map distance in morgans. The correction is imperfect in that d was not reported for each of the 214 pairings comprising the three studies; however, the d values used are those reported by the author of each article and to our knowledge represent reasonable estimates of interval size. Corrected frequencies should therefore represent only crossovers associated with the recombination event in am. Because it is hard to envisage a mechanism whereby increasing flanker separation could influence the nature of recombination events in am, the rise in this line is unexpected. Indeed, this rise suggests that the distance to flanking markers among am⁺ recombinants is larger than that among unselected progeny or, put another way, that recombination events in am promote (negatively interfere with) crossovers in the flanking intervals without being mechanistically associated.

We repeated one of Smyth's crosses and followed the segregation of both tightly linked molecular flanking markers and conventional flanking markers among am⁺ progeny (cross B163; BOWRING and CATCHESIDE 1996 ). Where conventional flanking markers were recombined, the majority of events were outside of the region bounded by molecular markers, and we concluded that a maximum of 7% of am conversions had a crossover that could be considered mechanistically associated. While the distance to flanking markers was elevated among am⁺ progeny, the elevation was just short of statistical significance, and although an absence of molecular markers proximal to am precluded examination of the position of crossovers here, there appeared to be no clustering of crossovers in the distal interval. We now have molecular markers in the proximal interval and, in accordance with Fincham's prediction, report a clustering of crossovers immediately proximal to am among B163 recombinants.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

B163 is a cross between F6325 and F11089 (BOWRING and CATCHESIDE 1996 ). B539 is a cross between strains B530 (a, his-3 K874, cog⁺; cyh-2, leu-5, sp B132, rec-2) and B537 (A, his-3 K874, cog⁺, ad-3 K118; cot-1 C102t; rec-2⁺, ure-2 D74, am⁶), which was carried out at 25° in SC medium (DAVIS and DE SERRES 1970 ) supplemented with 2% sucrose, histidine (0.2 g/liter), adenine (0.4 g/liter), leucine (0.2 g/liter), and alanine (0.5 g/liter). leu⁺ am⁺ recombinants from cross B539 were selected on Vogel's N medium (DAVIS and DE SERRES 1970 ) lacking both leucine and alanine but with added histidine and adenosine. Glycine (1.5 g/liter) was also added to suppress leaky growth of am mutants.

Restriction-site polymorphism was detected by probing EcoRV-digested Neurospora DNA with cosmids G18:A7 or G4:B5, which are from the pMOcosX library (ORBACH 1994 ). Genomic DNA was prepared either as described (BOWRING and CATCHESIDE 1996 ) or using the mini-prep method of IRELAN et al. 1993 . EcoRV digestion of genomic DNA was carried out according to the manufacturer's instructions (Bresatec, Adelaide, South Australia). Probe labeling, Southern transfer, hybridization, and membrane stripping were carried out according to protocols included with the Phototope kit (New England Biolabs, Beverly, MA). Probe was synthesized using linearized cosmid DNA as a template.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Of 205 progeny from cross B163 that experienced a recombination event in am, 21 had a crossover between HP and sp (BOWRING and CATCHESIDE 1998 ), markers 8.3 kb (BOWRING and CATCHESIDE 1996 ) and 450 kb (F. J. BOWRING and D. E. A. CATCHESIDE, unpublished results; T. SONE, J. BOK, F. J. BOWRING, D. E. A. CATCHESIDE and A. J. F. GRIFFITHS, unpublished results) proximal to am, respectively. To probe for clustering of crossovers in these recombinants, two restriction-site polymorphisms (RSPs) were used to define three contiguous subintervals immediately proximal to am. The closest (subinterval I) is bordered by HP and G18E, which is 80 ± 10 kb proximal; the next (subinterval II) extends from G18E to G4E (at 145 ± 15 kb proximal); and the most remote, G4E to sp (subinterval III), represents the remainder of the HP to sp interval (Figure 2).

View larger version (24K): In this window In a new window Download PPT slide   Figure 2. Position of relevant linkage group V markers segregating in B163 and B539.

The same RSPs distinguish strain B530 from strain B537 (cross B539, Figure 2). Of 153 leu-5⁺ am⁺ recombinants from this cross, 67 were also sp⁺ and thus had a crossover between sp and am. The segregation of G18E and G4E was examined in 64 of these recombinants. Because HP was not scored in progeny from cross B539, subinterval I is 8.3 kb larger in this cross than in B163. However, as the HP to am⁶ genetic distance is estimated to be <0.09 cM (BOWRING and CATCHESIDE 1996 ), any effect is expected to be slight.

Figure 3 shows the position of crossovers among the subset of recombinant progeny from crosses B163 and B539 that experienced an exchange proximal of am in regions I, II, and III. The progeny of cross B163, selected for conversion in am, have a significantly different distribution of crossovers in these intervals from the progeny of cross B539 in which am conversion was not selected for (² = 11.0, d.f. = 2, P = 0.004). Because subinterval III accounts for 57% of crossovers among progeny from both pairings, the difference appears to be wholly contained within the two subdivisions most closely proximal to am. In cross B163, 38% of crossovers occurred in subinterval I, immediately proximal to am, compared with only 11% in subinterval I for cross B539. The situation is reversed in subinterval II where, in cross B163, only 5% of crossovers occurred compared with 32% of crossovers in cross B539.

View larger version (12K): In this window In a new window Download PPT slide   Figure 3. Position of crossovers proximal to am in am+ recombinants (B163) and recombinants selected for an exchange between sp and am (B539). *, subinterval I is 8.3 kb larger in B539 (see text).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We compared the distribution of crossovers across three chromosome segments proximal to am in two crosses. In cross B163, where we first selected for a conversion event in am, there was clustering of crossovers in the interval immediately proximal to am: 8 of 21 (38%) crossovers within the 450-kb sp to HP interval were between G18E at +80 ± 10 kb and HP at +8.3 kb. In contrast, in cross B539, where there was no selection for conversion in am, only 7 of 63 (11%) crossovers occurred in the equivalent interval (G18E to am⁶ at +15 bp).

Although these data show a significant clustering of crossovers close to am in progeny that experienced conversion in am, they may underestimate the degree of clustering if the number of crossovers designated as associated with conversion is an overestimate. BOWRING and CATCHESIDE 1996 assessed the likelihood of association on the basis of proximity. HP and another molecular flanking marker 5.7 kb distal to am (HD) were chosen to define an interval outside of which crossovers were thought unlikely to be associated with conversion in am. However, those inside this interval are not necessarily associated. For example, the principal criterion that led FINCHAM 1974 to consider crossovers among the R2 class as being separately initiated events was a converted segment separated from a crossover by a nonconverted segment. A total of 8 of the 14 crossovers between HP and HD in B163 recombinants satisfy this criterion (BOWRING and CATCHESIDE 1998 ) and on these grounds could be considered independent. Moreover, a higher density of markers in am and the associated increase in resolution could conceivably increase this number and further reduce the estimated number of crossovers associated with the conversion tract.

We previously considered the possibility that at am, where the frequency of crossovers per kilobase is low, recombination events might extend over a large distance. If this were the case, our estimate that a maximum of 7% of conversion events in am have a mechanistically associated crossover would be an underestimate. One of several reasons for our earlier rejection of this possibility was that there appeared to be no clustering of crossovers distal to am among B163 recombinants (BOWRING and CATCHESIDE 1998 ). However, as we have now found clustering of crossovers in the proximal interval, the possibility seems worthy of further consideration. Could this clustering be due to the spreading of recombination events initiated at am to a point beyond HP? Indeed, GILBERTSON and STAHL 1996 have proposed a model that could account for the clustering of crossovers proximal to am. In this model, which was formulated to account for the observed tendency of crossovers to occur on one side of the ARG4 recombination initiation site, cleavage of the two Holliday junctions comprising the hypothesized recombination intermediate is both asynchronous and strand specific. If HP was contained in symmetrical heteroduplex following migration of the Holliday junction beyond this marker, the apparent separation of conversions from crossovers could be accounted for. However, as HP is 8.3 kb from the peak of conversion frequency in cross B163 (BOWRING and CATCHESIDE 1998 ), migration would need to cover at least this distance. Moreover, if our interpretation of the data presented in Figure 1 is sound, explanation of the disparity between the corrected frequencies of flanking marker exchange in Smyth's (B) and Fincham's (C) studies requires that Holliday junctions migrate much farther and in the opposite direction. Because both Fincham and Smyth used sp as their proximal marker, the additional crossovers observed by Fincham are likely to have occurred distal to Smyth's distal marker (his-1) but proximal to Fincham's distal marker (inl). his-1 is probably farther than 300 kb, but certainly no less than 150 kb, distal to am (BOWRING and CATCHESIDE 1995 ), so if these additional crossovers were mechanistically related to the recombination event in am, the recombination event would need to span at least this distance.

We cannot rule out the existence of a polymorphic recombination hotspot in the appropriate subinterval of the appropriate diploid, or a hitherto unknown trans-acting regulator of recombination, as an explanation for the different distribution of crossovers in the two pairings. However, no such genes are known to influence the frequency of crossing over near am. The data presented here and elsewhere do, however, fit Fincham's hypothesis.

FINCHAM 1974 hypothesized that recombination events in am rarely had an associated crossover and we have shown this to be the case (BOWRING and CATCHESIDE 1996 ). Fincham further hypothesized, on the basis of his and Smyth's mapping efforts, that the increase in flanker exchange observed among am⁺ progeny was driven by negative interference. The addition of data from a subsequent mapping study is consistent with this (Figure 1). Finally, Fincham predicted that because of negative interference, crossovers would be clustered about am in am recombinants. We report here that crossovers are clustered proximal to am in am recombinants.

FINCHAM 1974 proposed that local synapsis, established during recombination in am, might drive negative interference by increasing the probability of additional synapsis and thus additional recombination events nearby. While not proposed by Fincham nor addressed by our data, we interpret Figure 1 to suggest that negative interference may extend beyond the immediate vicinity of am. Conceivably, this could be driven by local synapsis at am, increasing the probability of additional synapsis over an extended distance with the effect decreasing with distance from am and ultimately subsiding somewhere outside of the interval bounded by sp and inl.

The counting model for chiasma interference (FOSS et al. 1993 ) also predicts clustering of crossovers about a converted locus. In this model, an organism-specific number (m) of conversion events (Co) must occur between each conversion event that is resolved as a crossover (Cx) and, based upon an estimate of ~0.3 for the proportion of conversion events that have an associated crossover in both Saccharomyces cerevisiae and Neurospora crassa, m was thought to be 2 in these fungi. Within the framework of the counting model, recombination in am signals the completion of one Co. Clustering results from an increased likelihood of a crossover nearby, because the number of subsequent Co's that must occur prior to a Cx is now reduced by one. While a test of the model in S. cerevisiae failed, for Neurospora, where the model accurately accounts for certain recombination data, its applicability was not ruled out (FOSS and STAHL 1995 ).

Clustering of crossovers in the subinterval immediately proximal to am in am⁺ recombinants might be driven by local synapsis or by the action of a counting machine. Indeed, this clustering is explicable within the framework of existing molecular models of meiotic recombination if extensive Holliday junction migration is invoked. However, such models have crossovers that are mechanistically associated with conversion events as a foundation, and we wonder if this foundation has been rigorously tested. FOGEL et al. 1979 point out that an assumption inherent in the method used to estimate the proportion of conversion events that have an associated crossover is an absence of conversion-driven interference. While conversion-driven positive interference has been ruled out (STADLER 1973 ), conversion-driven negative interference has not. Indeed, whereas a mechanistic association between gene conversion and crossing over and the operation of negative interference are both expected to increase the level of flanking marker exchange about a converted locus, it would not be difficult to miss this latter phenomenon. Perhaps the possibility of conversion-generated negative interference deserves more experimental attention than it has received.

	ACKNOWLEDGMENTS

This work was supported by a grant from the Australian Research Council.

Manuscript received November 30, 1998; Accepted for publication March 19, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BOWRING, F. J. and D. E. A. CATCHESIDE, 1995  The orientation of gene maps by recombination of flanking markers for the am locus of Neurospora crassa. Curr. Genet. 29:27-33[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].

BOWRING, F. J. and D. E. A. CATCHESIDE, 1998  Analysis of conversion tracts associated with recombination events at the am locus of Neurospora crassa.. Curr. Genet. 34:43-49[Medline].

DAVIS, R. H. and F. J. DE SERRES, 1970  Genetic and microbiological research techniques for Neurospora crassa.. Methods Enzymol. 17A:79-143.

FINCHAM, J. R. S., 1967  Recombination within the am gene of Neurospora crassa.. Genet. Res. 9:49-62.

FINCHAM, J. R. S., 1974  Negative interference and the use of flanking markers in fine-structure mapping in fungi. Heredity 33:116-121.

FOGEL, S., R. MORTIMER, K. LUSNAK, and F. TAVARES, 1979  Meiotic gene conversion: a signal of the basic recombination event in yeast. Cold Spring Harbor Symp. Quant. Biol. 43:1325-1341.

FOSS, E. J. and F. W. STAHL, 1995  A test of a counting model for chiasma interference. Genetics 139:1201-1209[Abstract].

FOSS, E., R. LANDE, F. W. STAHL, and C. M. STEINBERG, 1993  Chiasma interference as a function of genetic distance. Genetics 133:681-691[Abstract].

GILBERTSON, L. A. and F. W. STAHL, 1996  A test of the double-strand break repair model for meiotic recombination in Saccharomyces cerevisiae. Genetics 144:27-41[Abstract].

HOLLIDAY, R., 1964  A mechanism for gene conversion in fungi. Genet. Res. 5:282-304.

HOLLIDAY, R., 1974  Molecular aspects of genetic exchange and gene conversion. Genetics 78:273-287[Free Full Text].

IRELAN, J., V. MIAO, and E. U. SELKER, 1993  Small scale DNA preps for Neurospora crassa. Fungal Genet. Newsl. 40:24.

MESELSON, M. and C. RADDING, 1975  A general model for genetic recombination. Proc. Natl. Acad. Sci. USA 72:358-361[Abstract/Free Full Text].

ORBACH, M. J., 1994  A cosmid with a Hy^R marker for fungal library construction and screening. Gene 150:159-162[Medline].

ORR-WEAVER, T. L. and J. W. SZOSTAK, 1985  Fungal recombination. Microbiol. Rev. 49:33-58[Free Full Text].

PERKINS, D. D., 1979  Crossing over of flanking markers is less than 50% among meiotic convertants. Genetics 91(Suppl.):94.

PRITCHARD, R. H., 1960  Localized negative interference and its bearing on models of gene recombination. Genet. Res. 1:1-24.

RAMBOSEK, J. A. and J. A. KINSEY, 1983  Fine structure mapping of the am (GDH) locus of Neurospora. Genetics 105:293-307[Abstract/Free Full Text].

SMYTH, D. R., 1973  A new map of the amination-1 locus of Neurospora crassa and the effect of the recombination-3 gene. Aust. J. Biol. Sci. 26:1355-1370[Medline].

STADLER, D. R., 1973  The mechanism of intragenic recombination. Annu. Rev. Genet. 7:113-127[Medline].