Spore-Killing Meiotic Drive Factors in a Natural Population of the Fungus Podospora anserina

Corresponding author: Alfons J. M. Debets, Laboratory of Genetics, Wageningen University, Dreyenlaan 2, 6703 HA Wageningen, The Netherlands., fons.debets{at}fungen.el.wau.nl (E-mail)

Communicating editor: R. H. DAVIS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In fungi, meiotic drive is observed as spore killing. In the secondarily homothallic ascomycete Podospora anserina it is characterized by the abortion of two of the four spores in the ascus. We have identified seven different types of meiotic drive elements (Spore killers). Among 99 isolates from nature, six of these meiotic drive elements occurred in a local population. Spore killers comprise 23% of the natural population of P. anserina in Wageningen, The Netherlands, sampled from 1991 to 1997. One Spore-killer type was also found in a French strain dating from 1937. All other isolates found so far are sensitive to spore killing. All seven Spore killer types differ in the percentage of asci that show killing and in their mutual interactions. Interactions among Spore killer types showed either mutual resistance or dominant epistasis. Most killer elements could be assigned to linkage group III but are not tightly linked to the centromere.

SEGREGATION distorters are genetic elements that show meiotic drive, a phenomenon in which one member of a pair of heterozygous alleles is transmitted in excess of the expected Mendelian ratio of 50% (SANDLER and NOVITSKI 1957 ; LYTTLE 1991 ). Well-known examples of segregation distorters are the sex-ratio chromosomes (SR) in Drosophila, a male sex chromosomal drive system, and the t-haplotype in mice and segregation distorter (SD) in Drosophila, both male autosomal drive systems (LYTTLE 1991 ). In Drosophila and mouse, the meiotic drive systems minimally involve two closely linked loci, a distorter and its cis-acting target. All distorters are associated with polymorphic chromosomal structures, such as inversions. Their ratio of distortion in these examples can exceed 90% and they are closely linked to the centromere. It is not known for most drive systems whether they involve two closely linked loci. Likewise, distortion ratios for Drosophila and mice in nature may vary greatly. Meiotic drive systems in these organisms showing <90% distortion are harder to detect. Furthermore, classes of insensitive target or suppressor alleles have accumulated to counter these selfish elements (LYTTLE 1991 ).

Meiotic drive allows deleterious alleles to spread through populations if the frequency gain from their segregation advantage more than compensates the frequency loss due to elimination by natural selection. Thus it threatens adaptive evolution and it is therefore of great interest to obtain information on the extent of meiotic drive in natural populations. This is not easy to study because in animals and plants a driving genetic element requires a specific phenotype to be observable. For this reason it is understandable that an appreciable number of known cases of meiotic drive involve genes affecting the sex ratio. However, fungi in which the haploid nuclei resulting from meiosis are linearly arranged within an ascus provide unique opportunities to analyze abnormal segregation, for precisely the same reason that they have played such a big role in the classical experiments by Lindegren and others on fundamental aspects of linkage, meiotic recombination, and gene conversion (see WHITEHOUSE 1973 ; PERKINS 1992 ). Any meiotic drive system in such fungi—provided the elimination of the nuclei containing the nondriving allele occurs in an early stage after the completion of meiosis, as it does in all known meiotic drive systems—will be observed in a cross between a driving and a sensitive strain as spore killing: the degeneration and early abortion of half the ascospores in a certain proportion of the asci. This is not the only distinguishing feature of drive systems in fungi. The ascospores are the products of meiosis as well as the progeny. Thus distortion in fungi also affects the number of offspring produced and reduces the fecundity, which has important consequences for the population genetics of meiotic drive in fungi (NAUTA and HOEKSTRA 1993 ).

The earliest analysis of two segregation distorters in fungi, then called ascospore abortion factors, is by PADIEU and BERNET 1967 in the ascomycete Podospora. TURNER and PERKINS 1979 , TURNER and PERKINS 1991 identified such abortion factors in Neurospora as Spore killers. Other fungi in which distorters have been found are Gibberella fujikuroi (= Fusarium moniliforme) and Cochliobolus heterostrophus (see RAJU 1994 , RAJU 1996 for a review). However, the best-studied example of meiotic drive in ascomycetes is Spore killer (Sk) in Neurospora. Haploid Spore killer strains of Neurospora were originally identified because asci always contained four viable black and four small inviable unpigmented spores in crosses with standard wild-type strains. All the viable spores carry the Sk^K allele. In crosses homozygous for a killer allele (Sk^K x Sk^K) each ascus contains eight viable black ascospores, as in normal sensitive crosses (Sk^S x Sk^S), indicating that killing occurs only in crosses heterozygous for the killing factor (TURNER and PERKINS 1979 , TURNER and PERKINS 1991 ).

Several Spore killer types have been characterized in Neurospora: Sk-1^K from Neurospora sitophila and a Sk-2^K and Sk-3^K from N. intermedia. Only Sk-1^K occurs widespread in nature (TURNER and PERKINS 1979 ). Both Sk-2^K and Sk-3^K were introgressed into the genetically better-characterized N. crassa and both mapped to a region of 30 map units across the centromere of linkage group III. This region, 3% of the total genomic map, was found to contain a recombination block (CAMPBELL and TURNER 1987 ). No evidence was found for large inversions or chromosome rearrangements, though small inversions might exist between markers (BOJKO 1988 ; TURNER and PERKINS 1991 ). The killer complex must therefore be considered as a haplotype. Whether Sk-1^K is associated with a recombination block is unknown (TURNER and PERKINS 1979 , TURNER and PERKINS 1991 ).

Meiosis is normal in crosses between Spore killers and sensitives. Both nuclear types coexist within the same ascus cytoplasm and ascus development is typical until after postmeiotic mitosis when the nuclei are enclosed by ascospore walls. Both nuclear types can coexist as well in vegetative heterokaryons, as is apparent from rare occasions when they are included together in the same ascospore (RAJU 1979 ; RAJU and PERKINS 1991 ; TURNER and PERKINS 1991 ).

Sk-2 and Sk-3 have been introgressed into the secondarily homothallic N. tetrasperma, which normally makes asci with four large spores that are heterokaryotic for mating type and any other centromere-linked markers that are heterozygous in the cross. Crosses of N. tetrasperma heterozygous for the centromere-linked killers Sk-2 and Sk-3 all produced four-spored asci as predicted from the behavior of these killers in the eight-spored species. The sensitive nuclei were protected in heterokaryotic Sk^K + Sk^S ascospores, but killing occurred in this species when exceptional small homokaryotic ascospores were formed (RAJU and PERKINS 1991 ).

Podospora anserina grows on dung of herbivores and is also a secondarily homothallic ascomycete. It also produces four binucleate spores per ascus (Fig 1). For the behavior of Spore killers in P. anserina the following aspects of ascospore formation are relevant. (1) Programming of nuclear positioning in the Podospora ascus is such that following meiosis and postmeiotic mitosis, the two ascospores in one-half of the ascus each receive two nonsister nuclei descending from the two meiotic products from the same half tetrad. The other two ascospores each contain two nonsister nuclei from the other half tetrad. As a result, ascospores are homokaryotic for all markers showing first divison segregation (FDS) and heterokaryotic for those markers that show second division segregation (SDS; see ESSER 1974 ; RAJU and PERKINS 1994 ). Due to an obligate single crossover between the centromere and the mating-type locus, (nearly) all spores are heterokaryotic for mating type (mat+ and mat-). (2) In crosses heterozygous for a Spore killer element, ascospores that receive only sensitive nuclei abort, whereas ascospores that are homo- or heterokaryotic for a killer nucleus survive. Therefore, the frequency of asci showing two viable and two inviable ascospores (two-spored asci) reflects the frequency of FDS for the Spore killer element (Fig 1). This fraction of two-spored asci is hereafter referred to as the spore-killing percentage. This study describes the results of our search for meiotic drive elements in a natural P. anserina population. We have identified and characterized seven different types of Spore killers, indicated by the abortion of half of the ascus progeny, among 99 recently isolated Dutch strains of P. anserina and 3 older isolates originating from France. Six of these Spore killer types can be attributed with certainty to meiotic drive elements. We have also analyzed the interactions between the different killer types. Finally, mapping data of the Spore killer elements are presented.

View larger version (31K): In this window In a new window Download PPT slide   Figure 1. Model to explain spore killing in P. anserina as the segregation of a meiotic drive element. The figure shows a cross between two strains carrying a Spore killer () and a sensitive (•) allele. Two nonsister nuclei descending from a half tetrad are enclosed in each ascospore. FDS for the Spore killer element results in an ascus with two surviving spores, each homokaryotic for the killer element, and two aborted spores, each homokaryotic for the sensitive alleles. SDS results in a four-spored ascus, in which each ascospore survives because it carries both a nucleus with the killer allele and one with the sensitive allele.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and culture methods:
P. anserina strain S, isolated in Normandy, France, in 1937 was used as a standard tester strain. Spore killer strains Y and Z originated from Picardy, France, in 1937 (BELCOUR et al. 1997 ). P. comata Spore killer strain T (Picardy, France, 1937) was previously described as P. anserina but was renamed later on the basis of morphology and mitochondrial type. It is interfertile with P. anserina (PADIEU and BERNET 1967 ; BELCOUR et al. 1997 ). All other P. anserina Spore killer strains (Wa numbers) were collected in Wageningen, The Netherlands, during 1991–1997 from dung (VAN DER GAAG et al. 1998 ). Recombinants of the Spore killer strains (XS numbers) with the genetic background of sensitive strain S were obtained by five recurrent backcrosses.

Marker strains used in this study were Cs3 (16% SDS, LG I, cold sensitive), 136 (0.5% SDS, LG II, green spores), Cs2 (3% SDS, LG III, cold sensitive), 187 (76% SDS, LG III, green spores), rd1 (84% SDS, LG III, round spores), Lys2 (0% SDS, LG IV, lysine requiring), As7 (0% SDS, LG V, paromomycin resistant), Cs18 (0–7% SDS, LG VI, cold sensitive), Cs12 (0–3% SDS, LG VII, cold sensitive, paromomycin hypersensitive; PICARD 1971 ; MARCOU et al. 1990 ). All marker strains were derived from wild-type strain S.

Culture conditions and media have been described by ESSER 1974 . Cornmeal agar was used as a standard growth medium with 100 mg/liter lysine added for the Lys2 marker. Tests for the Lys2 marker were performed on minimal medium 2 (MM2) without lysine. Paromomycin hypersensitivity and resistance were tested on MM2 supplemented with 270 µg/ml paromomycin (Sigma, St. Louis). All cultures were grown at 27°. Cold sensitivity was tested at 11° (PICARD-BENNOUN and LE COZE 1980 ).

Crosses were performed on moistened copromes (horse dung tablets that were sterilized by -irradiation; WOOD and COOKE 1984 ) on a sterile filter paper on top of the agar in a plate to improve crossing ability. Crossing occurred either by spermatization of monokaryotic strains with microconidia or by confrontation of mycelia of opposing mating type. Backcrosses of progeny to the parental strains were performed using the spermatization technique.

Genetic mapping:
Methods of genetic analysis have been described by ESSER and KUENEN 1967 . In short, Spore killer strains were crossed with centromere-linked marker strains to identify the linkage group. Two-spored asci resulting from FDS for the Spore killer element (see Fig 1) were tested for the occurrence of the marker. Such markers show FDS, and thus the two surviving ascospores will be homokaryotic either for the marker or the wild-type allele. Nonparental ditypes (NPD, i.e., the two surviving spores show the marker) and parental ditypes (PD, i.e., the two surviving spores are wild type) will be equally frequent in the progeny when killer and marker are unlinked. When linked, NPDs appear very rarely, because they require a four-chromatid double crossover between the killer locus and the marker. When using a centromere-linked tester strain, SDS for the marker is rare and can be neglected.

Markers more distal from the centromere were used for the establishment of linkage on chromosomal arms. Four-spored asci, resulting from SDS for the Spore killer element, were tested for the absence or occurrence of the marker. A low percentage of SDS for only the Spore killer element compared to SDS for both Spore killer and marker indicates linkage to that chromosomal arm (depending on the distance of the markers).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Spore killing in Podospora reflects meiotic drive:
To assess the extent of meiotic drive in a natural population we sampled the local P. anserina population of Wageningen, The Netherlands. During the period 1991–1997 we obtained a total of 99 new isolates of P. anserina. Species determination was based on morphological criteria, as well as on fertility with the standard strain S, or with wild-type strains sexually compatible with S (MIRZA and CAIN 1969 ; VAN DER GAAG et al. 1998 ). We observed that 23 isolates produced up to 95% two-spored asci in these crosses, instead of the expected four-spored asci (Table 1, Fig 2). In addition to the new wild-type strains, we analyzed three French Podospora strains, T, Y, and Z, isolated in 1937 and showing similar spore abortion in specific crosses (BELCOUR et al. 1997 ). Strain T was previously characterized as a Spore killer strain (PADIEU and BERNET 1967 ; TURNER and PERKINS 1991 ) and was recently renamed P. comata on the basis of morphological and molecular data (BELCOUR et al. 1997 ). Progeny grown from two-spored and four-spored asci were backcrossed to both mating types of the parental strains to confirm that the observed spore killing is caused by meiotic drive. The results from such an analysis of strain Wa58 (Psk-7) are shown in Table 2 and may be summarized as follows:

1. Selfing of the progeny from two-spored ascus progeny always yielded normal four-spored asci (no abortion).

   
   

   View larger version (168K): In this window In a new window Download PPT slide   Figure 2. Rosettes of asci from crosses between P. anserina strains. Asci with darker ascospores are more mature than those with lighter spores. (A) A rosette from a normal cross showing only four-spored asci. (B) A cross of Wa6 (Psk-1) x S showing a high percentage of two-spored asci. (C) A cross of Wa28 (Psk-2) x S showing 70% two-spored asci. An ascus containing two dikaryotic and two small monokaryotic killer spores can be seen (black arrow). Both monokaryotic sensitive spores have been aborted. (D) A cross of Wa58 (Psk-7) x S showing 50% two-spored asci. (E) A cross of Wa20 (Psk-3) x Wa16 showing rosettes with different killing percentages. Aborted spores are also visible within the asci (black arrows). (F) A cross between Wa52 (Psk-1) x Wa58 (Psk-7) showing 30% two-spored asci. The five-spored ascus (black arrow) indicates that both smaller mononucleate spores contain a killer locus (as expected for a parental ditype). A four-spored ascus (white arrow) containing a mononucleate spore indicates SDS for one of the killer loci, resulting in the segregation of an (aborted) sensitive nucleus.

   
   

   View this table: In this window In a new window   Table 1. Spore killer types found in isolates of Podospora

   
   

   View this table: In this window In a new window   Table 2. Progeny tests of cultures from two- and four-spored asci of the cross Wa58 (Psk-7K) x S (Psk-7S)

2. Backcrossing of the progeny of the two-spored asci to the S⁺ and S^- strains showed spore killing, whereas backcrossing to the Wa58⁺ and Wa58^- strains gave normal four-spored asci.

3. Selfing of the four-spored progeny showed ascospore abortion.

4. Backcrosses of the four-spored progeny produced spore killing in the backcross to S^- and Wa58⁺ or to S⁺ and Wa58^-, but not to both.

5. In addition to these observations, it must be added that there is no effect of the mating type or of the sexual role (maternal or paternal) of the strains involved.

These data support the meiotic drive model of spore killing as presented for Neurospora (TURNER and PERKINS 1979 ) and applied to the genetic system of P. anserina (Fig 1). Spores from a cross between a strain carrying a Spore killer element and a sensitive strain would be either homokaryotic (i.e., show FDS for the Spore killer element, reflecting no crossover in the centromere proximal region) or heterokaryotic for the Spore killer element (SDS, reflecting a crossover). Only spores carrying a Spore killer nucleus survive; thus FDS for the Spore killer results in two-spored asci, the aborted spores carrying only sensitive alleles. SDS would result in four-spored asci. The sensitive nuclei in the heterokaryotic four-spored asci are viable as can be seen in the selfings and backcrosses. Results similar to those described in Table 2 were found for all other Spore killer isolates that were tested. In all cases, standard strain S behaved as the sensitive isolate.

There are at least six different Spore killer types in the Wageningen population of P. anserina:
The Spore killer strains were initially classified on the basis of (1) spore killing frequency in a cross to a standard sensitive strain (FDS percentage) and (2) the interaction between the Spore killers (Table 1 and Table 3). In this way at least six types of Spore killers could be identified among the 99 natural isolates. An additional seventh type was discovered in the French P. anserina strain Y. All Spore killer strains of the same type showed a constant and repeatable spore-killing frequency when crossed to strain S and absence of spore killing when intercrossed. When intercrossed (Table 3), however, Spore killer strains of different types show killing, similar to the behavior of Sk-2^K x Sk-3^K in Neurospora.

Fig 2 shows some rosettes of a normal cross and different spore-killing reactions. The killer types Psk-1 and Psk-5 show the highest frequency of two-spored asci, >90% (Fig 2B). Psk-4, Psk-6, and Psk-7 produce the lowest killing percentage; only half of the asci contain two spores; the remaining asci carry four spores as is the normal condition (Fig 2D). The Psk-2 strain and P. comata strain T show intermediate levels of spore killing; 75% of two-spored asci are found (Fig 2C). Spores homokaryotic for the sensitive allele can be observed only for a short time in these killer crosses. They completely degrade at the start of spore wall formation.

The group of Psk-3 strains is different from the others in that the frequency of spore killing is highly variable among perithecia of the same cross (Fig 2E). Fruiting bodies with any combination of two- and four-spored asci can be observed. Furthermore, ascospore abortion is found only in crosses between specific strains and even between some Psk-3 strains (Table 4). Another distinguishing feature of the Psk-3 group is that the aborted spores do not disintegrate as in the other killer types, but remain in the asci as tiny, shriveled spores. Because of the erratic expression of spore abortion, further genetic analysis of these Psk-3 isolates has not been performed. It is therefore not certain that this group contains true meiotic drive elements.

Spore killers are stable upon recurrent backcrossing:
Several strains belonging to different Spore killer types were backcrossed five times with the sensitive strain S to assess the stability of the Spore killers and, at the same time, to obtain a more identical genetic background and to increase fertility for further analysis. The fraction of two-spored asci of the fifth recurrent backcross did not differ from the percentage found in the first cross for Psk-1, Psk-2, Psk-5, and Psk-7 (Table 1). These Spore killer types all show a stable percentage of two-spored asci. We were not able to proceed in backcrossing Psk-4 and Psk-6 killers, owing to fertility problems.

High frequency and diversity of Spore killers in a natural population of Podospora:
The incidence of Spore killer strains in the P. anserina population of Wageningen appears remarkably high. Of the 99 Wa strains isolated between 1991 and 1997, 23 contain a driving element. Spore killers were found during all years of isolation, except for 1995. In 1996 no strains were isolated. Psk-1, Psk-2, and Psk-6 strains were found over several years in the population; Psk-4 and Psk-7 were isolated only in 1994 (Table 5).

Among six strains isolated in 1937 in Picardy, France, two contained a meiotic drive element. Also the P. comata Spore killer strain T was isolated on that occasion (BELCOUR et al. 1997 ). No Spore killer strains were reported among isolates from other French regions (L. BELCOUR, personal communication). Both the Dutch Spore killer strain Wa58, isolated in 1994, and the French strain Z, originating in 1937, belong to the same killer type Psk-7 (Table 1).

In contrast to Neurospora, no neutral strains, i.e., strains that are not killed but do not themselves kill, were found. However, only a selection of the Wageningen isolates before 1994 was tested against every new Spore killer isolate. Exceptions are the strains from the Psk-3 group that show killing behavior only in specific crosses (Table 4). Other nonkiller strains are sexually incompatible, or produce four-spored asci in crosses with Psk-3 members, and seem to act like neutral strains. Among the different Spore killer types some killer types are resistant to killing by other killer types. This is discussed below in more detail.

Spore killer types show dominant epistatic or mutual resistant interactions:
We have crossed the Spore killer strains to each other and measured the fraction of spore killing. The results of the initial crosses are shown in Table 3 for one representative of each killer type. Results for other strains were similar. No empty perithecia or completely aborted progeny were found in any of these crosses. Two types of interaction between killer elements were observed. (1) Most commonly, the fraction of two-spored asci was similar to that found for one of the parents when crossed to a sensitive tester strain. We refer to this type of interaction as dominant epistasis. For example, the cross between Psk-1 and Psk-2 shows the killing percentage characteristic of Psk-1 (Table 3). Backcrosses of two-spored progeny from these interactions produced results similar to those for backcrosses of two-spored progeny from Psk-1 and a sensitive tester. (2) In some interactions between killers, a much lower percentage of two-spored asci was observed than was found for any of the parents when crossed to the standard testers. This type of interaction can be described as mutual resistance of the unlinked killer elements (for explanation see below). For example, in the cross Psk-1 x Psk-7 a relatively low percentage of two-spored asci was observed (Fig 2F and Table 3). We have analyzed the two-spored progeny by backcrosses to the parental Spore killer strains, by selfing, and by crosses with a sensitive strain. The results of the analysis of the Psk-1 x Psk-7 progeny are shown in Table 6 as an example of mutually resistant Spore killers. These results can be summarized as follows:

1. Backcrosses with both parental strains did not show any killing. Every nucleus in the two-spored progeny therefore contains both killers.

   
   

   View this table: In this window In a new window   Table 6. Progeny tests of cultures from two-spored asci of crosses between Psk-1 and Psk-7

2. Crosses with the sensitive strain produced a similar killing percentage as the original cross between Psk-1 and Psk-7 did. This verifies that the two surviving ascospores contain both killers.

3. Selfing of the progeny from two-spored asci yielded only four-spored asci. This is also consistent with the surviving ascospores being homokaryotic for both killer elements.

Spore killing between mutually resistant Spore killers is reminiscent of the interaction of unlinked duplicate genes. Ascospores survive when at least one killer element is present, and spore killing is limited to the NPD class of asci with FDS for both killer elements. In these asci, the surviving two spores carry both Psk-1 and Psk-7. The two aborted spores carry neither. The low percentage (23%) of two-spored asci is consistent with the two killer loci being unlinked. Tightly linked Spore killers would have given a high percentage of two-spored asci. Psk-4 and Psk-6 were also found to be mutually resistant.

Not every backcross with each wild-type strain was fertile, but usually enough information could be extracted when backcrosses of other strains of the same killer type were taken into consideration. This was, however, not the case with analysis of the four-spored asci, although the incomplete results obtained in these backcrosses did not contradict the analysis of the two-spored asci.

The relation of the Spore killer types found in Wageningen to each other is summarized in the scheme below:

Psk-1Psk-4 | | |Psk-2 | | |Psk-7Psk-6The straight line represents a mutually resistant interaction, whereas the arrow indicates a dominant epistatic interaction. All Spore killer types to the right of the arrowhead are sensitive to killing by the killer types on the left. Thus Psk-7 and Psk-1 are the most effective killer types in the Wageningen population. They are mutually resistant but kill Psk-2, Psk-4, and Psk-6. Psk-4 and Psk-6 are killed by all the other killer types but are mutually resistant. The French Psk-5 Spore killer type has a more complex interaction to the other killer types. Psk-5 is sensitive to Psk-7, mutually resistant to Psk-1, and kills all other killer types.

Most Spore killers are assigned to LG III, but not tightly linked to the centromere:
All Spore killer strains were crossed to centromere-linked marker strains to identify the linkage group of the killer element. The linkage analysis for one representative per Spore killer type is presented in Table 7. Other strains of the same killer type showed similar results. All Psk-1, Psk-2, Psk-5, and Psk-7 killer strains showed linkage to the linkage group III centromere marker Cs2, as indicated by the low percentage of NPD asci. Linkage could not be established with certainty for Psk-6, but LG III (10.8% NPD) seems more likely than LG IV (18% NPD). Spore killer type Psk-4 seems to be located on linkage group IV, although a possible location on LG V cannot be excluded. Not all unlinked markers, however, show an equal 50% segregation pattern (Table 7). This can be due to negative fitness aspects, which lead to an underrepresentation of the marker involved. Furthermore, interference as often observed in P. anserina may hamper linkage group assignment for certain genes (M. PICARD, personal communication).

Psk-1 and Psk-5 were crossed to a strain with two other markers on both sides of the LG III centromere to determine their location on the chromosome arm. A green spore marker, 187, is located on the left arm of the chromosome, whereas rd1, a round spore marker, is situated on the right arm. Results of these crosses indicated a strong interference (Table 8). The SDS percentage for rd1 was reduced from >80% to 25% when a crossover for Psk occurred. This reduction was found in all combinations of killer and markers (data not shown). A control cross of rd1 with 187 did not show this interference. Whether the interference is due to, or merely detected by, the presence of the Spore killer cannot be concluded from these data. These results can be explained by strong chromosomal interference across the centromere or by positive chromatid interference, if a crossover for rd1 is followed by a specific second crossover involving the same chromatids. In the first case, the killer must be present on the left arm, whereas in the second case the killer is present on the right arm of the linkage group. Alternatively, centromere misdivision or occurrence of spindle overlap at the second division could simulate double crossovers across the centromere with positive interference (D. D. PERKINS, personal communication).

Also, attempts were made to further localize Psk-7 in crosses with LG III markers rd1 and 187 (Table 8). Chromosomal arm linkage would be most strongly indicated by a low tetratype (T) fraction of two-spored asci (FDS for Psk-7) for that marker. The T class for 187 is combined with the PD class, since neither can be distinguished phenotypically. If the PDs comprise a considerable part of the combined classes, linkage to the left arm is possible. The large T class for the rd1 marker also indicates linkage to the left chromosomal arm.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Spore killer types in Podospora:
We have found segregation distorters showing meiotic drive in natural isolates of the secondarily homothallic ascomycete P. anserina. These meiotic drive elements cause the abortion of two of the four spores in the ascus. The two surviving ascospores contain the distorter, whereas the aborted spores contain alleles sensitive to it. Other causes of ascospore abortion, such as translocations (PERKINS 1974 ; PERKINS and BARRY 1977 ; BRONSON 1988 ), lethal mutations (DELANGE 1981 ), or spore color mutants (MARCOU et al. 1990 ; RAJU 1994 ), can be excluded. Reciprocal translocations are not expected to produce two-spored asci in Podospora, and nonreciprocal translocations could produce at most 50% of such asci. Lethal genotypes and spore color mutants can be excluded because both killer and sensitive strains behave normally during selfing and in homokaryotic condition. Our data on backcrosses, killing percentages, and localization of the killer complex demonstrate that the observed spore killing in Podospora is caused by meiotic drive elements.

Rescue of sensitive alleles was first shown in crosses of Neurospora Sk-2 and Sk-3 with the mutant Banana, which is sensitive to both killers. These crosses produce giant ascospores containing four killer and four sensitive nuclei of both types (RAJU 1979 ). Also, when Sk-2 and Sk-3 were introgressed into pseudohomothallic N. tetrasperma, no killing was found in heterokaryotic ascospores. This fungus produces four binucleate spores, but unlike P. anserina, the spores are heterokaryotic for centromere-linked markers. The sensitive nuclei are saved here because tight linkage of the killer genes to the centromere ensures that every ascospore receives a killer nucleus (RAJU and PERKINS 1991 ). As in Neurospora, sensitive alleles can be rescued in Podospora by inclusion in an ascospore with a nucleus containing the killer allele. This is the case in four-spored asci. The sensitive nucleus proved fully functional in backcrosses of the four-spored progeny with the sensitive and killer parents and shows that no irreversible damage occurs to the nucleus during or after meiosis.

Segregation distortion in Podospora differs in an important common aspect from other meiotic drive systems found in nature (LYTTLE 1991 ). Absence of tight linkage to the centromere was observed, which is reflected in the lower fraction of asci in which ascospores are killed in four of the seven Spore killer types. Also the failure to detect resistant or suppressor alleles differs from the other drive systems. The Podospora ascospore abortion factors resemble in their behavior the Spore killer complexes found in the sibling species P. comata (PADIEU and BERNET 1967 ) and the heterothallic ascomycetes C. heterostrophus (BRONSON et al. 1990 ), G. fujikuroi (= F. moniliforme; KATHARIOU and SPIETH 1982 ), N. intermedia, and N. sitophila (TURNER and PERKINS 1979 ). However, a significant difference in the reproductive system exists between the secondarily homothallic Podospora and the other ascomycetes, which are heterothallic. This sexual strategy has consequences for the presence of Spore killers. As in N. tetrasperma, each ascospore of Podospora contains two nuclei and is heterokaryotic for mating type. SDS for the Spore killer locus results in shielding of the sensitive nucleus within the ascospores and the formation of a normal four-spored ascus whose spores are heterokaryotic for the Spore killer. Killing frequencies in this secondarily homothallic fungus can range from 0% (complete SDS) to 100% (complete FDS), as exemplified by this study. Thus, the percentage of SDS for the Spore killer locus influences the percentage of two-spored asci found. In contrast, the ascospores of heterothallic ascomycetes are homokaryotic and a crossover only leads to a shift in the linear order of the nuclei in the ascus. All four sensitive nuclei are still killed, and the four spores containing the killer nuclei remain. Thus the percentage of SDS does not have any influence on the killing percentage in heterothallic fungi.

An SDS percentage of nearly 100% in Podospora, as found for the mating-type locus, would automatically lead to a nonkilling phenotype. This led Perkins and co-workers to propose that secondary homothallism (in N. tetrasperma) evolved as a mechanism to escape Spore killer elements in the heterothallic precursor species (TURNER and PERKINS 1991 ; RAJU and PERKINS 1994 ). However, this proposition only holds for driving elements in N. tetrasperma that are closely linked to the centromere. A Spore killer located more distantly from the centromere would still give a spore-killing phenotype in N. tetrasperma, though such high frequencies as found in P. anserina would not be expected. No Spore killer elements have been identified in N. tetrasperma. However, ascospore abortion is high in outcrosses between wild-collected N. tetrasperma strains (JACOBSON 1995 ), and the basis of the ascospore death remains undetermined. Drive elements are not excluded.

We have identified seven different Spore killer types, six of which occurred in a sample of 99 wild-collected strains from Wageningen, The Netherlands. Psk-2 shows a percentage of killing comparable to the percentage found in P. comata Sk-1 (or -a2; PADIEU and BERNET 1967 ; TURNER and PERKINS 1991 ). It is probable that Spore killer complexes found in P. anserina and P. comata are related, since the two species are relatively interfertile.

One set of Spore killers, the Psk-3 group, possesses some unique properties different from other killer types. First, the two aborted spores remain visible within the ascus as small unpigmented spores together with the two normal-sized black ascospores. Second, the percentage of killing varies between fruiting bodies within the same cross. Last, Psk-3 killers show the spore-killing phenotype only in crosses with specific strains. Most other strains are apparently resistant to Psk-3 killing. This variable killing percentage superficially resembles the ascospore abortion found in Podospora by BERNET 1965 in crosses between strains S and s. Ascospore abortion occurred in the s perithecia at the restrictive temperature of 18°. The amount of killing found varied over time; perithecia that were initiated later had a decreased amount of two-spored asci. No killing was found in the S perithecia, nor at normal growth temperatures. All the surviving spores in the two-spored asci belonged to the s genotype. The maternal effect is associated with the s prion, which is also involved in the heterokaryon incompatibility reaction in this fungus (COUSTOU et al. 1997 ). However, the Psk-3 killing occurs at normal growth temperatures (27°) and does not have a maternal effect.

In the heterothallic G. fujikuroi, a mixed Spore killer type, Sk^mx, was also found. This killer type causes the abortion of half of the spores in 23–70% of the asci. The remaining asci are normally eight-spored (KATHARIOU and SPIETH 1982 ; SIDHU 1984 ). Crosses between Sk^mx strains result in a variety of asci containing two, four, six, or eight viable spores (SIDHU 1984 , SIDHU 1988 ). Sk^mx strains are also partially resistant to normal Sk strains. A variable killing percentage occurs also in N. intermedia with certain partially sensitive or resistant strains. Strains are called resistant in Neurospora when at least 25% of the asci contain eight spores, but the partially resistant strains found in nature produced at least 50% eight-spored asci (TURNER 1977 and personal communication). It is possible that the killing reaction of the Podospora Psk-3 group is caused by a few remaining partially resistant strains, whereas the other strains are fully resistant. The killing of the Psk-3 group with other specific strains, but not the variability of killing, can also be explained by synthetic lethals (THOMPSON 1986 ), i.e., epistatic genes that affect viability only in specific combinations.

Interaction between Spore killer types:
The Spore killer types found in P. anserina show either dominant epistasis or mutual resistance. In a dominant epistatic interaction one killer strain behaves like a killer and the other like a normal sensitive strain. In the Wageningen population, Psk-1 and Psk-7 show dominant epistasis to all other killer types, whereas Psk-4 and Psk-6 are sensitive to killing by all the other Spore killers. The dominant epistatic interaction resembles the interaction between Sk^mx and Sk in G. fujikuroi, where Sk^mx appears to be dominant epistatic to Sk, even though Sk^mx kills less efficiently than Sk (KATHARIOU and SPIETH 1982 ).

Interactions between mutually resistant Spore killers exhibit a much lower killing percentage than that observed in either parent when crossed to a normal sensitive. Ascospores from two-spored asci from these crosses are recombinant (NPD) types that now possess both killer elements. These recombinant double killers are less efficient distorters, since sensitive alleles can be rescued by each single killer type. We did not find double killer strains in our sample, but the strain studied by PADIEU and BERNET 1967 contained two unlinked killer elements. We have localized six of the Spore killer types by crossing killer strains with sensitive centromere-linked marker strains. Remarkably, almost all Spore killer types are found in linkage group III. Also the het-s locus involved in the above-mentioned spore killing between strains s and S is located in linkage group III. The only exception, Psk-4, is probably in LG IV.

Recombination can easily occur between unlinked or distally linked killer types, as found for the interactions Psk-4 x Psk-6 and Psk-1 x Psk-7. However, the observed recombination percentage of some interactions, e.g., Psk-1 x Psk-5, is hard to understand. Both killer types show a high percentage of FDS and are possibly located on the same arm of LG III. Normal recombination cannot produce such high recombination values for tightly linked markers. A very specific interference type has to be assumed, or perhaps some other factor interferes with the spore killing pattern. In C. heterostrophus, the analysis of a Spore killer was complicated by the presence of a translocation (TAGA et al. 1984 ; BRONSON 1988 ; BRONSON et al. 1990 ).

No mutually sensitive killer strains were found, in contrast to N. intermedia, in which Spore killer Sk-2 and Sk-3 kill each other when crossed (TURNER and PERKINS 1979 , TURNER and PERKINS 1991 ). However, TURNER and PERKINS 1991 in their analysis of the data from PADIEU and BERNET 1967 with the P. comata killer strains show that the results are consistent with mutual killing of the a and b genotypes. Apparently a mutually sensitive reaction may exist in Podospora, although such Spore killer types have not been encountered in our P. anserina sample.

Natural populations:
Of the 99 newly isolated P. anserina strains from Wageningen, 23% contain a meiotic drive element. As argued in the Introduction, fungi with ordered tetrads linearly arranged in asci provide a unique possibility to observe the genome-wide extent of meiotic drive because any meiotic drive element present in a cross heterozygous for the driving allele will cause spore killing. Viewed in this way, meiotic drive can be concluded to be common in this population. On the other hand, assuming that the number of coding genes per genome is in the order of 10⁴, the probability per locus of a segregation-distorting allele is in the order of 10^-5, implying that non-Mendelian segregation at nuclear loci is rare indeed. Data on spore killing in other fungal populations show roughly a similar picture. In N. sitophila, the overall incidence of Sk-1 is 19%, but geographic regions exist where sensitive or killer strains were not obtained. The frequencies of Sk-2 and Sk-3 in N. intermedia in nature are extremely low and killer strains are restricted to the southeast Asian archipelago. The highest number of Spore killers was found in worldwide isolates of G. fujikuroi var. moniliforme. Here, a total frequency of 88% Sk and Sk^mx killers was observed (KATHARIOU and SPIETH 1982 ). However, a later study by SIDHU 1988 of midwestern United States isolates showed a reduced frequency of <50% Sk and Sk^mx. For G. fujikuroi var. subglutinans SIDHU 1984 found results comparable to those of KATHARIOU and SPIETH 1982 . The worldwide incidence of Spore killers in C. heterostrophus is 50% in Race O field isolates, but no killers have been found in the Race T isolates. Spore killers in Race O were restricted to the United States mainland and could be subdivided into regions that were polymorphic or consisted only of killer isolates (BRONSON et al. 1990 ).

Most of the Spore killer types originating from Wageningen could be recovered over several years, indicating a relative stability of the killer genes within the natural population. The finding of the Psk-7 killer type in the recently isolated Dutch population and the French strains isolated almost 60 years earlier also supports the idea of a stable population of killers and sensitives. A prerequisite for the maintenance of a stable polymorphism of driving and sensitive alleles at a distorter locus in fungi is the existence of neutral or resistant strains as predicted by the model of NAUTA and HOEKSTRA 1993 . No such neutral strains have been found for P. anserina yet, though several killer types are resistant to other killer types. The same situation exists in G. fujikuroi where Sk^mx is partially resistant against Sk (KATHARIOU and SPIETH 1982 ). No resistant types were observed in C. heterostrophus (BRONSON et al. 1990 ) and strains resistant for Sk-1^K are rare in N. sitophila (TURNER and PERKINS 1991 ; B. C. TURNER, personal communication). In N. intermedia, fully and partially resistant strains to Sk-2^K and Sk-3^K or both were found in nature. Genes conferring resistance to killing have been mapped closely linked to the killer complex. r(Sk-2)-1 is at the left end of the recombination block, while two interacting resistance genes were mapped at loci flanking the right end of the recombination block. Widespread resistance for Sk-2^K has been found in N. crassa and in N. intermedia even though the Sk-2 killer haplotype has not been found in N. crassa (B. C. TURNER, personal communication).

P. anserina and N. tetrasperma are able to reproduce by selfing and do not depend on outcrossing. Selfing protects the offspring from being harmed by Spore killers both because it avoids the introduction of killer elements from other strains and because the program of ascus development results in ascospores that are heterokaryotic for genes far from the centromere (P. anserina) or near the centromere (N. tetrasperma). Chances for meiotic drive by spore killing to occur depend on the occasional outcross of a sensitive strain with a Spore killer. This aspect of the reproductive system will affect the population genetics of meiotic drive and has not been taken into account in the model of spore killing analyzed by NAUTA and HOEKSTRA 1993 . Also the chromosomal location of the distorter locus relative to the centromere was not considered in this population genetics model. One would expect that any linked suppressor of recombination between a distorter and the centromere would be selected, since centromere-linked distorters are maximally effective. Remarkably, we have found some distorters with appreciable amounts of SDS. It is of interest to study further the population genetics of meiotic drive in P. anserina, not only taking the reproductive system and the location into account, but also the implications of interactions between abundant Spore killer types for retention of sensitive alleles. Segregation distorters, once established in a population, may probably linger on for a longer time than in heterothallic species.

To understand the evolutionary consequences of spore killing, it is important to know more about the ecology of spore killing. In this study we detected Spore killers in roughly one-quarter of the natural isolates. However, all crosses were done under standardized laboratory conditions at a constant temperature of 27°, while, e.g., the het-s locus of P. anserina only shows meiotic drive when strain s is used as maternal parent in a cross to S at low temperature (18°). It is important to analyze the effect of fluctuations in environmental conditions such as temperature on spore killing. Also fitness consequences of Spore killers should be studied; e.g., are there differences in size or number of spores from perithecia of crosses homozygous or heterozygous for Spore killers as compared to sensitive strains?

Finally, a mechanistic understanding of spore killing awaits molecular characterization of the genes and gene products involved. At the same time such a molecular approach may yield insight into the evolutionary origins of meiotic drive in fungi.

	ACKNOWLEDGMENTS

We thank A. A. Haspels for assistance with the analysis of the Spore killer strains; L. Belcour for providing us with the French Spore killer strains T, Y, and Z; and M. Picard for the marker strains. Furthermore, many thanks to D. Perkins, N. Raju, and B. Turner for numerous helpful suggestions to improve the manuscript. This work was supported by a grant from the Dutch organization for scientific research (NWO-ALW).

Manuscript received March 20, 2000; Accepted for publication May 5, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BELCOUR, L., M. ROSSIGNOL, F. KOLL, C. H. SELLEM, and C. OLDANI, 1997  Plasticity of the mitochondrial genome in Podospora. Polymorphism for 15 optional sequences: group-I, group-II, intronic ORFs and an intergenic region. Curr. Genet. 31:308-317[Medline].

BERNET, J., 1965  Mode d'action des gènes de `barrage' et relation entre l'incompatibilité cellulaire et l'incompatibilité sexuelle chez Podospora anserina.. Ann. Sci. Nat. Bot. Veg. 6:611-768.

BOJKO, M., 1988  Presence of abnormal synaptonemal complexes in heterothallic species of Neurospora.. Genome 30:697-709[Medline].

BRONSON, C. R., 1988  Ascospore abortion in crosses of Cochliobolus heterostrophus heterozygous for the virulence locus Tox1. Genome 30:12-18.

BRONSON, C. R., M. TAGA, and O. C. YODER, 1990  Genetic control and distorted segregation of T-toxin production in field isolates of Cochliobolus heterostrophus.. Phytopathology 80:819-823.

CAMPBELL, J. and B. C. TURNER, 1987  Recombination block in the Spore killer region of Neurospora.. Genome 29:129-135[Medline].

COUSTOU, V., C. DELEU, S. SAUPE, and J. BEGUERET, 1997  The protein product of the het-s heterokaryon incompatibility gene of the fungus Podospora anserina behaves as a prion analog. Proc. Natl. Acad. Sci. USA 94:9773-9778[Abstract/Free Full Text].

DELANGE, A. M., 1981  The mutation Sk(ad-3A) cancels the dominance of ad-3A⁺ over ad-3A in the ascus of Neurospora.. Genetics 97:237-246[Abstract/Free Full Text].

ESSER, K., 1974 Podospora anserina, pp. 531–551 in Handbook of Genetics I, edited by R. C. KING. Plenum Press, New York.

ESSER, K., and R. KUENEN, 1967 The Genetics of Fungi. Springer-Verlag, Berlin/Heidelberg/New York.

JACOBSON, D. J., 1995  Sexual dysfunction associated with outcrossing in Neurospora tetrasperma, a pseudohomothallic ascomycete. Mycologia 87:604-617.

KATHARIOU, S. and P. T. SPIETH, 1982  Spore killer polymorphism in Fusarium Moniliforme.. Genetics 102:19-24[Abstract/Free Full Text].

LYTTLE, T. W., 1991  Segregation distorters. Annu. Rev. Genet. 25:511-557[Medline].

MARCOU, D., M. PICARD-BENNOUN and J.-M. SIMONET, 1990 Genetic map of Podospora anserina, pp. 3.58–3.67, in Genetic Maps: Locus Maps of Complex Genomes, Ed. 5, edited by S. O'BRIEN. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MIRZA, J. H. and R. F. CAIN, 1969  Revision of the genus Podospora.. Can. J. Bot. 47:1999-2048.

NAUTA, M. J. and R. F. HOEKSTRA, 1993  Evolutionary dynamics of Spore killers. Genetics 135:923-930[Abstract].

PADIEU, E. and J. BERNET, 1967  Mode d'action des gènes responsibles de l'avortement de certaine produits de la méiose chez l'Ascomycete Podospora anserina.. Compt. Rend. Hebd. Séances Acad. Sci., Sér. D 264:2300-2303. (English translation in TURNER and PERKINS 1991)..

PERKINS, D. D., 1974  The manifestation of chromosome rearrangements in unordered asci of Neurospora.. Genetics 77:459-489[Abstract/Free Full Text].

PERKINS, D. D., 1992  Neurospora: the organism behind the molecular revolution. Genetics 130:687-701[Medline].

PERKINS, D. D. and E. G. BARRY, 1977  The cytogenetics of Neurospora.. Adv. Genet. 19:133-285[Medline].

PICARD, M., 1971  Genetic evidences for a policistronic unit of transcription in the complex locus "14" in Podospora anserina I. Genetic and complementation maps. Mol. Gen. Genet. 111:35-50[Medline].

PICARD-BENNOUN, M. and D. LE COZE, 1980  Search for ribosomal mutants in Podospora anserina: genetic analysis of cold-sensitive mutants. Genet. Res. 36:289-297[Medline].

RAJU, N. B., 1979  Cytogenetic behavior of Spore Killer genes in Neurospora.. Genetics 93:607-623[Abstract/Free Full Text].

RAJU, N. B., 1994  Ascomycete Spore killers: chromosomal elements that distort genetic ratios among the products of meiosis. Mycologia 86:461-473.

RAJU, N. B., 1996  Meiotic drive in fungi: chromosomal elements that cause fratericide and distort genetic ratios. J. Genet. 75:287-296.

RAJU, N. B. and D. D. PERKINS, 1991  Expression of meiotic drive elements Spore killer-2 and Spore killer-3 in asci of Neurospora tetrasperma.. Genetics 129:25-37[Abstract].

RAJU, N. B. and D. D. PERKINS, 1994  Diverse programs of ascus development in pseudohomothallic species of Neurospora, Gelasinospora and Podospora.. Dev. Genet. 15:104-118[Medline].

SANDLER, L. and E. NOVITSKI, 1957  Meiotic drive as an evolutionary force. Am. Nat. 91:105-110.

SIDHU, G. S., 1984  Genetics of Gibberella fujikuroi. V. Spore killer alleles in G. fujikuroi. J. Hered. 75:237-238[Abstract/Free Full Text].

SIDHU, G. S., 1988 Gibberella spp., pathogens of many crop species, pp. 159–167 in Genetics of Plant Pathogenic Fungi, Advances in Plant Pathology, Vol. 6, edited by G. S. SIDHU. Academic Press, New York.

TAGA, M., C. R. BRONSON, and O. C. YODER, 1984  Non-random abortion of ascospores containing alternate alleles at the Tox-1 locus of the fungal plant pathogen Cochliobolus heterostrophus.. Can. J. Genet. Cytol. 27:450-456.

THOMPSON, V., 1986  Synthetic lethals: a critical review. Evol. Theory 8:1-13.

TURNER, B. C., 1977  Resistance to Spore killer genes in Neurospora strains from nature. Genetics 86(Suppl.):S65-S66. [Abstract].

TURNER, B. C. and D. D. PERKINS, 1979  Spore killer, a chromosomal factor in Neurospora that kills meiotic products not containing it. Genetics 93:587-606[Abstract/Free Full Text].

TURNER, B. C. and D. D. PERKINS, 1991  Meiotic drive in Neurospora and other fungi. Am. Nat. 137:416-429.

VAN DER GAAG, M., A. J. M. DEBETS, H. D. OSIEWACZ, and R. F. HOEKSTRA, 1998  The dynamics of pAl2-1 homologous linear plasmids in Podospora anserina.. Mol. Gen. Genet. 258:521-529[Medline].

WHITEHOUSE, H. L. K., 1973 Towards an Understanding of the Mechanism of Heredity, Ed. 3. Edward Arnold, London.

WOOD, S. N. and R. C. COOKE, 1984  Use of semi natural resource units in experimental studies of coprophilous fungi. Trans. Br. Mycol. Soc. 83:337-374.

This article has been cited by other articles:

				N. B. Raju, R. L. Metzenberg, and P. K. T. Shiu Neurospora Spore Killers Sk-2 and Sk-3 Suppress Meiotic Silencing by Unpaired DNA Genetics, May 1, 2007; 176(1): 43 - 52. [Abstract] [Full Text] [PDF]

				D. D. Perkins A fratricidal fungal prion PNAS, May 27, 2003; 100(11): 6292 - 6294. [Full Text] [PDF]

				H. J. P. Dalstra, K. Swart, A. J. M. Debets, S. J. Saupe, and R. F. Hoekstra Sexual transmission of the [Het-s] prion leads to meiotic drive in Podospora anserina PNAS, May 27, 2003; 100(11): 6616 - 6621. [Abstract] [Full Text] [PDF]