Copyright 0 1991 by the Genetics Society of America

Cloning the Mating Types of the Heterothallic Fungus Podospora anserina: Developmental Features of Haploid Transformants

Carrying Both Mating Types

Marguerite Picard, Robert Debuchy and Evelyne Coppin Laboratojre de Ginitiqw, URA 1354 du CNRS, Universiti Paris XZ, 91405 Orsay cedex, France

Manuscript received July 30, 1990 Accepted for publication April 1, 199 1

ABSTRACT DNAs that encode the mating-type functions (mat+ and mat-) of the filamentous fungus Podospora

anserina were cloned with the use of the mating-type A probe from Neurospora crassa. Cloning the full mat information was ascertained through gene replacement experiments. Molecular and functional analyses of haploid transformants carrying both mating types lead to several striking conclusions. Mat+ mat- strains are dual maters. However, the resident mat information is dominant to the mat information added by transformation with respect to fruiting body development and ascus production. Moreover, when dual mating mat+ mat- strains are crossed to mat+ or mat- testers, there is strong selection, after fertilization, that leads to the loss from the mat+ mat- nucleus of the mat information that matches that of the tester. Finally, the mat locus contains at least two domains, one sufficient for fertilization, the other necessary for s.porulation.

T HE genetic control of sexual reproduction in lower eucaryotes is an intricate process. In many

heterothallic species, sexual reproduction requires both mating-type differences and differentiation of sexual organs or gametes. Second, in the simplest (bipolar) situations, the mating types are specified by one locus which, in the traditional nomenclature, is represented by a pair of alleles. However, multiple sex-related genes, both structural and regulatory, may reside at these mat loci. Third, there are several ex- amples of functions, apparently unrelated to sexual reproduction, controlled by mat loci: vegetative in- compatibility in Neurospora crussa (BEADLE and COON- RADT 1944), life span in Podospora anserina (RIZET 1953), organelle gene inheritance in the alga Chlam-

ydomonas reinhardtzi (SAGER 1954), pathogenicity in the basidiomycete Ustilago maydis (PUHALLA 1970).

P. anserina is a haploid heterothallic filamentous fungus. The mating types are called mat+ and mat- (previously designated mt+ and mt-). There is no visible difference between the wild-type strains of the two mating types, but mat+ mycelia senesce later than mat- ones (RIZET 1953; MARCOU 1961). Both kinds of strains give rise to male gametes (microconidia) and differentiate female organs (ascogonia). Matings occur only between male and female organs of opposite mating types. Fertilization induces a series of events leading to the development of the female organ into a fruiting body or perithecium. The mature perithe- cium contains many asci with ascospores bearing the haploid products of meioses. All the asci of a perithe- cium originate from only one male and one female nucleus. Genetics 148: 539-547 uuly. 1991)

Several laboratories began studying the mating-type loci of filamentous fungi at the molecular level by cloning the mat genes from ascomycetes such as N . crassa (VOLLMER and YANOFSKY 1986; GLASS et al. 1988; METZENBERG and GLASS 1990; STABEN and YANOFSKY 1990; GLASS, GROTELUESCHEN and METZ- ENBERG 1990) and from basidiomycetes, e.g., Schizo- phyllum commune (GIASSON et a l . 1989), U. maydis (KRONSTAD and LEONG 1989; SCHULZ et al. 1990) and Coprinus cinereus (MUTASA et a l . 1990). The mecha- nism by which mating-type genes control sexual re- production has been intensively studied in the yeast Saccharomyces cerevisiae (reviewed by HERSKOWITZ 1989).

Here we describe the cloning of the mat+ and mat- loci of the filamentous ascomycete Podospora anserina. We cloned the Podospora mat- DNA by using the Neurospora mating-type A DNA as a probe.

Transformation experiments yielded haploid strains carrying both mat+ and mat- information. Molecular and physiological analyses of these unnat- ural strains give new insights into the genetic control of heterothallism in filamentous fungi. The data are discussed with respect to the general model for the actions of the S . cerevisiae MAT locus and to the biology of filamentous fungi.

MATERIALS AND METHODS

Fungal strains and transformations: Genetics and bio- logical properties of P . anserina were first described by RIZET and ENGELMANN (1949) and have been reviewed by EWER (1974). The asci contain four spores, each formed around two non-sister nuclei after a post-meiotic mitosis. A

540 M. Picard, R. Debuchy and E. Coppin

few asci contain five spores among which two are smaller and develop around one nucleus. These uninucleate spores give rise to homocaryotic mycelia, which have been used for molecular and genetic analysis. When it was necessary, whole asci were investigated in order to know more precisely what happened during (or before) meiosis. The mating type (mat) is controlled by two haplotypes mat+ and mat- that display a 98% second division segregation so that the five- spore asci generally contain a mat+ and a mat- small spore. The 193 mutation, unlinked to mat, which prevents spore and mycelium pigmentation, exhibits a high first division segregation (PICARD 197 1).

The recipient strains for transformation experiments car- ried either the leul-1 mutation or the u r d - 6 mutation (RAZANAMPARANY and BECUERET 1986). The first one was used for transformations with cosmids from the mat- li- brary: in this case, the pHSU8 vector (DEBUCHY et al. 1988) carries the opal suppressor tRNA su8-I (DEBUCHY and BRY- COO 1985) which suppresses the leul-1 mutation. The sec- ond one was used for transformations with cosmids from the mat+ library: in that case, the cosmidic vector (pHC79- ura5) carries the u r d gene (BECUERET et al. 1984). The (ura') transformants were crossed to ura5-6 tester strains while the (leu+) transformants were crossed to tester strains carrying the 193 mutation, which allows one to analyze easily the inheritance of the su8-1 selective marker. Fur- thermore, the transforming su8-I marker suppresses the spore color mutation 193 in such a way that the three kinds of genotypes are easily scored : 193 su8+ (white spores), 193 su8-1 (green spores) and 193+ (black spores).

Transformation experiments were performed as de- scribed (BRYCOO and DEBUCHY 1985) with the exception that the spheroplasts were heat shocked (5 min at 48") before adding the DNA (BERCES and BARREAU 1989).

Bacterial strains: Cloning and plasmid preparations were done with both Escherichia coli BJ 5 183 (HANAHAN 1983) or HB 101 (BOYER and ROULLANDDUSSOIX 1969).

Plasmids and vectors: The plasmid ~ M T A G - ~ contains a 1.2-kb EcoRI-Bcll fragment from the A mating-type specific region of N . crassa (GLASS et al. 1988). The plasmid pCSN4 has a 1.9-kb BamHI-EcoRV fragment from the a mating type-specific region of N . crassa (GLASS et al. 1988). The pHSU8 vector (DEBUCHY et al. 1988) contains the su8-I gene that suppresses both the leul-1 and the 193 spore color UGA mutations. The pHC79-ura5 vector (TURCQ, DENAY- ROLLES and BECUERET 1990) contains the u r d + gene.

Genomic libraries: The P. anserina mat- lambda library (DEBUCHY et al. 1988) and the P. anserina mat- cosmid library (PICARD et al. 1987) were described previously. The P. anserina mat+ cosmid library was constructed by TURCQ, DENAYROLES and BECUERET (1 990).

Genomic DNA preparation and analysis: Genomic DNA was prepared from lyophilized mycelium according to a miniprep method (COPPIN-RAYNAL, PICARD and ARNAISE 1989). Genomic DNA was treated overnight with restriction enzymes. DNA fragments were separated on 0.5% agarose gels and transferred to Hybond Nylon membrane (Amer- sham Corporation) by the Southern method.

Hybridization procedures: Hybridizations with N . crassa A and a probes were performed at low stringency. The 1.2- kb EcoRI-BamHI fragment from pMTAG-2 ( A idiomorph) and the 1.9-kb EcoRV-EcoRV fragment from pCSN4 (a idiomorph) were gel purified, isolated by freeze squeeze and randomly labeled (FEINBERC and V~CELSTEIN 1983, 1984); activities were 2 - 10' cpm/pg and 1 O9 cpm/pg, respectively. Nylon filters were prehybridized overnight at 37" in 43% formamide, 5 X SSC, 0.1% bovine serum albumin, 0.1 % Ficoll40, 0.1 % polyvinylpyrrolidone 350, 20 rg/ml salmon

1 2 3 4

4.3,

FIGURE 1 .-Southern blots of Podospora genomic mat- (lanes 1 and 3) and mat+ (lanes 2 and 4) DNAs digested with EcoRl and probed at low stringeny with the 1.2-kb EccoRI-BamHI fragment from PMTAG-2 (A idiomorph) (lanes 1 and 2) and the 1.9-kb EcoRV-EcoRV fragment from pCSN4 (a idiomorph) (lanes 3 and 4). The 7-kb band present in lanes 1 and 3 is contamining pBR322- like DNA hybridizing with trace amounts of vector present in the A and a probes.

sperm. The probe was added and filters were incubated for 48 hr, then washed in 2 X SSC containing 0.1 % SDS for 2 hr at 37". High stringency hybridizations performed for screening the cosmid libraries and for analyzing genomic DNAs from the transformants were as described by MAN- IATIS, FRITSCH and SAMBROOK (1 982) except that BLOTTO (JOHNSON et al. 1984) was used instead of Denhardt's solu- tion.

RESULTS

Cloning the mating" specific regions: Ge- nomic DNA from mat+ and mat- strains was probed at low stringency with the A and a mating type specific regions of N. crassa. The A specific probe hybridizes to a mat- specific sequence on a 18-kb EcoRI fragment (Figure 1). In the case of the a specific probe, short exposure of the blot revealed a 13-kb EcoRI fragment present in both mat+ and mat- strains. Overexposure of the blot revealed very faint additional fragments that made a reliable search for a mating type-specific fragment impossible (Figure 1).

The mat- specific sequence was detected in a bac- teriophage library of a mat- strain probed at low stringency with the A-specific DNA. The insert of the corresponding phage (X9-1, Figure 2) was used as a probe to screen a mat- cosmid library and allowed the isolation of four cosmids: N8, N9, N10 and N1 1. The deduced physical map of the genome in this region and of the cosmids is shown in Figure 2A. The 18-kb EcoRI fragment present in cosmids N9, N10 and N11 hybridized with the A-specific probe (data not shown), confirming its identity with the genomic DNA fragment that hybridized with the same probe (Figure 1). Hybridization at high stringency of the 18- kb EcoRI fragment with the total DNA of a mat- strain indicates that no similar sequence is present on

Podospora Dual-Mater Transformants 54 1

A 21 m 3 1.5 1.2 6 4.3 18 5 4.5 1.8

I I 1 II I I 1 4 , I -9.1

N8 N9

N10 NI1

1.2 6 22 B UI -

FIGURE 2.-EcoRI maps of the mat- (A) and mat+ (B) loci. Number above each fragment indicates its size in kb. Superposed numbers indicate the sizes of unordered fragments. Lambda and cosmid inserts are shown. Mat- cosmid inserts (N8, N9, N I O , N 1 1) end with EcoRI sites. Lambda (mat-) and U1 (mat+) cosmid inserts end with Sau3A sites.

a different sized EcoRI fragment (data not shown). Hybridization of the 18-kb EcoRI DNA fragment

from the mat- strain with the DNA from a mat+ strain reveals a 22-kb EcoRI fragment (data not shown). Therefore, besides its mat- specific region, the 18-kb fragment has some sequences present in both mat+ and mat- genome. We supposed that these common sequences were linked to the mat-specific region and could be used to identify cloned DNA from the mat locus of a mat+ strain. Cosmid U1 of a mat+ library hybridizes to the 18-kb EcoRI probe. This cosmid contains an EcoRI fragment of 22 kb (Figure 2B). Total DNA of mat+ strains does not contain any other fragments that hybridize with the 22-kb EcoRI fragment of U1 (data not shown).

Mat- fertilization functions are encoded in the 18-kb EcoRI fragment: Cosmids N8, N9, NIO and N11 (see Figure 2) transform the l e d - I mat+ strain to leucine prototrophy with similar efficiency. Ten transformants bearing each cosmid were tested for fertilization and ascospore production (Table 1). All N8 primary transformants mated only as mat+: they did not cross with the mat+ tester strain. However, most N9 (6/10), N10 (9/10) and N11 (8/10) trans- formants induce formation of perithecia in both mat+ and mat- tester strains (Table 1). These data confirm that the DNA from a Podospora mat- strain isolated with the Neurospora A probe confer the ability to mate with a mat+ strain on a mat+ recipient. The gene(s) necessary for fertilization is (are) totally in- cluded in the 18-kb EcoRI fragment; it is the only fragment common to N9, N 10 and N I 1. Transfor- mation with the 18-kb fragment subcloned in pHSU8 confirms that this fragment contains mat- DNA; 54 of 84 transformants tested mate with mat+ and mat- testers.

Transformants bearing cosmids N9, N 10 or N11 that display a mat+ mat- phenotype were expected to contain a tandem duplication of mat DNA, one copy mat+ and one mat-. Their EcoRI restricted genomic DNA was fractionated by agarose gel electrophoresis and the gel blots were hybridized with radiolabelled N9, N 10 or N 1 1 cosmids. The EcoRI restriction maps

TABLE 1

Developmental features of mat+ mat-* transformants

Fertililation to tion with

Cosmid formants mat+ mat- mal+ mat-

N 8 I O - +++ NF +++ N10 1" - +++ NF +++

1 +++ +++ - ? 8 +++ +++ - +++

N9 4 - +++ NF +++ 4 +++ +++ + +++ 1 +++ ? + ? 1 +++ ? ?

No. of trans-

- NI 1 2 - +++ NF +++

4 +++ +++ + +++ 2 +++ +++ - +++ 2 +++ ? + ?

The transformants were used as male partners in all the crosses. - = no mating or no ascus production. + = poor ascus production, high frequency of infertile perithecia. +++ = similar to reference strain. NF = lack of fertili~ation prevents assay. The question marks indicate that we were unable to recover transformed nuclei from the cross between the primary transformant and the mat- tester strain. This can be explained by a low ratio of transformed/untrans- formed nuclei in these transformants or because the transformed nuclei are unable to go through sexual reproduction in crosses to mat- strains.

Cosmid integrated at an ectopic position.

wt wt - + 1 2 3 a b + - N11

22- - 18/

- 7.8 -7.5

L * -1.2

A FIGURE 3.-Southern blots of genomic DNA from N9B2 (A),

NI IBI (B) transformants and from mat+ and mat- wild-type strains. The DNAs were digested with EcoRl and hybridized with "P-labeled N9 (A) or N 1 1 (B) cosmid. A lane 1, N9B2 mat+ mat-* primary transformant; lane 2, N9B2 mat+ mat-* transformant sexual progeny; lane 3, N9B2 mat-* progeny. B lane a, N I 1 BI mat+ mat-* primary transformant; lane b N I 1B1 mat-* progeny. The size of fragments are given in kb.

of the mat+ and mat- idiomorphs and adjacent re- gions differ as shown in Figure 2. The mat+ infor- mation is in a 22-kb fragment, which is replaced by two fragments of 4.3 and 18 kb in mat- DNA. This polymorphism is seen in blots of DNA from the mat+ and mat- untransformed strains in Figure 3. The Southern analysis of N9B2 and N 1 1 B1 transformants is shown in Figure 3, lanes AI and Ba. The two

542 M. Picard, R. Debuchy and E. Coppin

transformants contain the 22-kb fragment specific to the resident mat+ DNA and the 4.3- and 18-kb frag- ments specific to the transforming mat- DNA as well as the 7.8-kb EcoRI fragment corresponding to inte- grated cosmid vector. The vector fragment was close to (N 1 1) or superimposed (N9) on a genomic fragment (see the map in Figure 2). The presence of the vector was checked by using the labeled vector without insert as a probe. Besides the polymorphic mating-type bands and the vector band, the pattern of the trans- formants corresponded to the pattern of the untrans- formed strain. No additional fragment was found, indicating that cosmid DNA had integrated by ho- mologous recombination at the resident locus. Sub- sequent genetic analysis of the transformants con- firmed that the transforming DNA had integrated at mat. The integration event gave rise to a tandem duplication, one copy containing the mat- informa- tion, the other one containing the mat+ information. All the transformants that were submitted to Southern analysis (four N9, two N10 and five N11 transform- ants) bore tandem duplications. Furthermore, molec- ular analysis of one transformant that exhibited the mat+ phenotype alone (as the recipient strain) indi- cated that only the mat+ 22-kb fragment was present; the 4.3- and 18-kb fragments were not detected. This suggests that a part of the cosmid sufficient to com- plement the Zeul-1 mutation had integrated, but with- out the mat- information.

Dual mat+, mat--* information can be inherited through a cross to a mat- strain: Primary transform- ants contain both transformed and untransformed nuclei. In order to obtain homocaryotic transformed strains, it is necessary to go through sexual reproduc- tion (in Podospora, uninucleate microconidia are un- able to germinate and act only as male gametes in the fertilization process). Most primary transformants that display the mat+ mat-* phenotype' are fully fertile when crossed to a mat- strain (see Table 1). The transforming marker (su8-1) suppresses the spore color mutation 193, so crosses were performed with a 193 mat- (white spores) strain. Ten to 30 uninucleate spores displaying the 193 su8-1 (green spores) phe- notype were isolated from each cross. Analysis showed that, in all cases, cosmids integrated at (or very near) the mat+ locus. In fact, most spores carried the selec- tive (su8-1) marker and displayed the dual mat+ mat-* phenotype.

Southern blots were performed on DNA from prog- eny exhibiting this mat+ su mat-* phenotype. In the case of N9 transformants, the restriction pattern did not differ from that of the primary transformants shown in Figure 3, lanes 1 and 2 for N9B2. Although

I The star indicates that mat DNA was added by transformation. For instance, mat+ mat-* indicates that a recipient mat+ was transformed with a cosmid carrying the mat- information.

they remain dual maters, progeny of the two N11 transformants that have been analyzed lost the 4.3-kb EcoRI fragment (data not shown). Therefore, the mat+ and mat-* information, artificially associated in the same nucleus, can be inherited through a cross to a mat- strain and, as expected from transformation with NlO, the 4.3-kb EcoRI fragment is not essential for the dual mating phenotype.

The mat--* information can be lost through a cross to a mat- strain: Although mat-* can be inher- ited through a cross, extensive analysis of offspring from mat+ su8-1 mat-* X mat- crosses (described in the previous paragraph), revealed that some spores lost the mat- phenotype. These strains are in no way different from a mat+ reference strain. These mat+ strains belong to two classes: those which retained and those which lost the su8-1 marker. The second class is probably due to recombination/excision of tandemly duplicated sequences that occurs mostly before pre- meiotic replication in Podospora (COPPIN-RAYNAL, PICARD and ARNAISE 1989). Molecular analysis of the first class of strains, which no longer expressed the mat- information although still containing the trans- formation marker, showed that they retained the vec- tor part of the cosmid (data not shown). The mecha- nisms giving rise to these progeny are under study and will be discussed elsewhere. The results of mat+ mat-* X mat- crosses are summarized in Figure 5 (left part, progeny class 1 to 3).

The 4.3-kb EcoRI fragment contains information necessary for perithecium development and ascus production: As noted above (see also Table l), most transformants obtained with N9, N10 and N11 cos- mids mate with mat+ strains as well as with mat- strains, in the sense that fertilization occurred. Fur- thermore, all transformants produced many asci when crossed to a mat- strain. The transformants could be divided into two classes on the basis of their behavior when they were crossed to a mat+ strain. None of the nine N10 transformants gave rise to progeny in this kind of cross; in contrast, the N9 (five among six) and N 1 1 (six among eight) transformants gave a few fertile perithecia when crossed to mat+ strains (see Table 1). Fertile perithecia represented less than 1% of those formed but the fertile perithecia contained a normal number of asci with viable spores. We do not know why so few perithecia were fertile when the (N9 and N1 1) mat+ mat-* transformants were crossed to a mat+ strain while all perithecia were fertile when the same transformants were crossed to a mat- strain. However, low fertility was inherited by the mat+ mat- progeny in the N9 secondary (purified) transformants. Although secondary N11 transformants that lost the 4.3-kb EcoRI fragment and retained the 18-kb EcoRI fragment were still able to fertilize a mat+ strain, they no longer generated fertile perithecia. These obser-

Podospora Dual-Mater Transformants 543

vations, along with the fact that N 10 transformants never gave rise to fertile perithecia in a cross to mat+ strains, suggest that the 4.3-kb EcoRI fragment con- tains information necessary for ascus production. In fact, the only difference between N9 and N 10 cosmids is that the first contains and the second lacks this fragment (see Figure 2).

The mut-* information can replace the mat+ information through a cross to a mut+ strain: Prog- eny from the few fertile perithecia of mat+ mat-* X mat+ crosses were analyzed genetically. All three N9 and two of four N11 transformants gave rise to a minority of spores that still expressed the selective (su8-I) marker. However, most asci had lost this marker. Genetic analysis showed that these spores no longer expressed the mat+ information but, instead, behaved as mat- strains. They mated perfectly with mat+ strains (as female and as male), and gave rise to fertile perithecia whose progeny were identical to those from classical mat+ X mat- crosses. The results of mat+ mat-* X mat+ crosses are summarized in Figure 5 (right part).

Several mat- progeny from the cross of a mat+ su8- I mat-* transformant to a mat+ strain were submitted to molecular analysis (Figure 3, lanes 3 and b). The loss of the 22-kb mat+ and 7.8-kb vector DNA frag- ments indicated that the resident mat+ information was excised together with the vector DNA containing the su8-I selective marker. The 7.8-kb fragment in lane 3 corresponded only to genomic DNA since it did not hybridize to a vector probe. An event during the cross led to the replacement of the mat+ resident information by the mat- transforming sequence. We confirmed that the mat- sequence had integrated in the mat+ chromosomic context by using a restriction fragment length polymorphism that distinguishes DNA from mat+ and mat- origins (COPPIN-RAYNAL, PICARD and ARNAISE 1989). DNA probes from that region reveal a 2.2-kb EcoRV fragment in mat+ DNA and a 2.4-kb fragment in mat- DNA (Figure 4). Only the 2.2-kb mat+ specific band was detected in N9B2 mat+ mat-* and in its mat- progeny (Figure 4). The same results were obtained with N 1 1 B1. This control eliminates the possibility that the mat- spores are contaminants.

Therefore, through this two-step process (transfor- mation followed by recombination/excision) it is pos- sible to obtain a mat- strain from a mat+ strain. This observation proves that cosmids N9 and N 1 1 contain all the specific sequences necessary to confer the mat- identity. These sequences lie in the 4.3- and 18-kb EcoRI fragments which are the only ones shared by these two cosmids.

The mut+ su8-I mat-* strains are not true hom- othallic strains: Homocaryotic strains from uninu- cleate spores that carry information for both mating

2.4

wt - + 1 2 3

”” -

2.2

FIGURE 4.--Southern blots of genomic DNA from N9B2 trans- formants and mat+ and mat- wild-type strains. The DNAs were digested with EcoRV and hybridized with ‘2P-labeled T 7 cosmid (COPPIN-RAYNAL, PICARD and ARNAISE 1989). Lanes 1 and 2, N9B2 mat+ mat-* primary and purified transformant; lane 3, N9B2 mat-* progeny. The polymorphic 2.2- and 2.4-kb fragments are indicated.

types look like self-fertile (homothallic) strains with respect to the fertilization process. In fact, at the end of growth, the plates are covered with fertilized peri- thecia. This self-mating makes it very difficult to use these strains as female partners in crosses. Apparently, self-fertilization is so efficient that very few female organs remain available to male gametes of another strain. Therefore, at the level of fertilization, these mat+ mat-* strains, like heterocaryotic strains where the mat+ and mat- information are expressed in two nuclei, are self-maters. However, only the heterocar- yotic strains give progeny; the unnatural homocar- yotic mat+ mat-* transformants are self-sterile. Fur- ther studies, especially cytological data, will clarify this block in the development of fertilized perithecia.

A converse story for mut+: The U 1 cosmid isolated from the mat+ library was used to transform a recip ient mat- strain. In this case, the selective marker carried by the cosmid was the Podospora ura5+ gene (BEGUERET et al. 1984) and the recipient was ura5-. Among 28 (ura+) transformants, 11 exhibited only the recipient mat- phenotype but 17 induced perithe- cium formation with both mat+ and mat- tester strains. Seven of these mat+* mat- transformants were analyzed further. In three cases the cosmid in- tegrated at an ectopic position. In four other trans- formants it integrated at the mat locus, probably through a homologous recombination event. In all cases, these mat+* mat- strains gave progeny when crossed to a mat+ strain. On the contrary, the cross was far less fertile when mat+* information from the cosmid was required (mat+* mat- X mat-). One of the ectopic transformants produced perithecia devoid of asci in such a cross. The ectopic integration prob- ably disrupted the mat+* sporulation information while keeping intact the sequences required for fertil- ization.

Progeny from the mat+* mat- X mat+ crosses show that mat+* can be transmitted through meiosis for

544 M. Picard, R. Debuchy and E. Coppin

homologous and ectopic transformants. In the latter case, as the ectopic mat+* segregated independently of the mat locus, spores containing two copies of mat+ information (mat+, mat+*) were recovered. They did not display any phenotype different from standard mat+ strains. As for the homologous transformants, recombination/excision events occurred during crosses giving rise to mat- spores (when they were crossed to a mat+ strain) and to mat+ (when they were crossed to a mat- strain). The mat+ spores correspond to a replacement of the resident mat- information by the mat+ information borne by the cosmid.

The data obtained with mat+ are very similar to those described for mat-. Although DNA from mat+* mat- transformants has not been analyzed, the mat+ data can be summarized in the following five points. First, the ectopic expression of mat+ and the gene replacement experiments indicate that the U 1 cosmid contains all information necessary for mat+ activities. Second, two copies of mat+ DNA do not disturb the life cycle of the fungus. Third, the mat+* mat- strains are dual maters. Fourth, ascus production needs more information than fertilization does as evidenced by one of the ectopic transformants. Fifth, with respect to sporulation (perithecium development and ascus production) there is a better expression of the resident DNA than of the transforming DNA.

There is only one difference in our results with mat+ and mat- transforming DNAs. While we did not obtain progeny from the self-fertilized perithecia in the case of mat+ mat-* strains, one of the U1 homol- ogous transformants was partly self-fertile. The spores from this selfing mat+* mat- strain kept the same mat+* mat- phenotype as the parental strain, which closely resembles homothallism. Further experiments are needed to understand why mat+*mat- and mat+ mat-* strains do not behave in the same way with regard to selfing.

DISCUSSION

The A specific probe from N . crassa permitted the cloning of P. anserina DNA that encodes mat- func- tions. The cloned fragment contains all sequences necessary to mat- activities (mating, perithecium de- velopment and ascus production) as evidenced by mat+ to mat- substitution experiments.

DNA encoding the Podospora mat+ determinants was identified in a genomic library using a mat- probe containing DNA sequences common to mat+ and mat- genomes. Mat- to mat+ replacement experi- ments showed that the cloned sequences contain all information required for mat+ activites.

Probing total DNA from mat+ and mat- strains with DNA fragments encoding either mat+ or mat-- activities revealed that these sequences are present only as a single copy in each haploid genome. There-

fore, unlike yeasts (NASMYTH and TATCHELL 1980; STRATHERN et al. 1980; BEACH 1983), Podospora contains no silent mating information. Podospora shares this feature with the four characterized heter- othallic filamentous fungi: the ascomycete N . crassa (GLASS et al. 1988) and the basidiomycetes S. commune (GIASSON el al. 1989), U. maydis (KRONSTAD and LEONG 1989; SCHULZ et al. 1990) and C. cinereus (MUTASA et al. 1990).

The a specific probe of Neurospora hybridizes to a 13-kb EcoRI fragment present in both mat+ and mat- Podospora genomic DNAs. The EcoRI restriction map of 84-kb flanking the mat+/mat- region does not reveal any 13-kb fragment (Figure 2). Therefore, Podospora contains a sequence, related to the Neu- rospora a idiomorph, which is not mat specific. This a related sequence may correspond to one of the nu- merous incompatibility genes identifed in Podospora (RIZET and ESSER 1953; BERNET 1967). In fact, the mt a-1 polypeptide is not only responsible for mating functions but also for vegetative incompatibility. These two activites might be controlled by two differ- ent domains of the polypeptide (STABEN and YANOF- SKY 1990). Furthermore, the lack of mat+/a hybridi- zation is well explained by preliminary sequence com- parisons (R. DEBUCHY and E. COPPIN, unpublished data). Although weak homology was found at the DNA level, extensive similarity was detected at the level of the translation products. This indicates that mat+ may be the functional counterpart of mt a al- though DNA sequences have strongly diverged.

The transformation experiments described in this paper give new insights into the genetic control of mating types in heterothallic ascomycetes. The data can be discussed according to four ideas.

Mat+ mut- strains are dual maters: Neurospora and Podospora mating differ strikingly from mating in two other heterothallic organisms where informa- tion for both mating types in the same nucleus is documented. In S . cerevisiae, the presence of active MATa and MATa information in the same haploid or diploid nucleus leads to sterility (the cells are unable to mate) (HABER and GEORGE 1979; KLAR, VOGEL and MacLeod 1979). However, in the alga C. rein- hardtii, diploids heterozygous at the mating-type locus (mt+/mt-) behave as mt- cells (EBERSOLD 1967). The sterile phenotype of yeasts heterozygous at MAT is well understood and explained through the regulatory activities of the MATa2 and MATaZ genes (see HER- SKOWITZ 1989, for a review). In Chlamydomonas. it is not known why mt- is dominant to mt+. In Neuro- spora, transformants that contain both A and a mat- ing-type sequences mate as both A and a, although exhibiting a growth inhibited phenotype unless vege- tative inconlpatibility associated with mating-type is suppressed (GLASS et al. 1988; STABEN and YANOFSKY

Podospora Dual-Mater Transformants 545

1990). In Podospora, the mat+ mat- strains grow as well as reference strains and mate as both mat+ and mat-.

It is now clear that the mat loci of filamentous fungi contain regulatory genes (GLASS, GROTELUESCHEN and METZENBERC 1990; STABEN and YANOFSKY 1990; R. DEBUCHY and E. COPPIN, unpublished data). How- ever these regulatory genes might not act in the same way as in yeasts. In fact, filamentous fungi have other means at their disposal in order to regulate matings since they develop sexual organs whatever the mating t Y F

Ascus production is dependent upon information beyond that necessary for fertilization: Transfor- mation of a mat+ recipient with the three cosmids N9, N 10 and N 1 1 from a mat- library generates two kinds of transformants: those able to mate as mat- and to give fertile perithecia when crossed to mat+ and those able to mate as mat- but unable to give progeny when crossed to mat+ (Table 1). Comparison of the cosmid maps (Figure 2) leads to the conclusion that complete information for fertilization lies in the 18-kb EcoRI fragment while the 4.3-kb EcoRI fragment contains information required for sporulation, i.e. all the proc- esses happening between fertilization and ascus pro- duction. Here again, the traditional nomenclature that represents mating-types as alleles is inadequate. The data presented in this paper indicate the existence of several (at least two) closely linked sexual-related functions in the mating-type locus region. This situa- tion is like that in S. cerevisiae (STRATHERN, HICKS and HERSKOWITZ 198 l), Schizosaccharomyces pombe (KELLY et al. 1988), Chlamydomonas (GALLOWAY and GOOD- ENOUGH 1985) and Neurospora (STABEN and YANOF- SKY 1990; GLASS, GROTELUE~CHEN and METZENBERC 1990). It is also tempting to speculate that bipolar fungi (whose mating types are specified by one locus) have information at one locus that tetrapolar basidi- omycetes (whose mating-type functions are encoded by two complex loci) have in two loci.

Resident information for ascus production is dominant to information added by transformation: The ten transformants (four N9 and six6 N1 1 , Table 1) that mate as both mat+ and mat-, give only a few fertile perithecia when crossed to mat+ but all the perithecia are fertile when the cross is performed to a mat- strain. Conversely, the resident mat- infor- mation is expressed better than mat+ when mat- recipients are transformed with mat+. Neurospora transformants behave in the same way, at least when A is the resident information and a the additional information (STABEN and YANOFSKY 1990).

Therefore, the mating-type information required for ascus production does not follow the rule of the apparent coexpression observed for the fertilization information. Further genetic and molecular data are

' - 1 I

i n h a i w a I

I

m t g s u w * mprem nore I M I ~ * mm+-E)* ntnr(+)m&*

FIGURE 5.-Formation of progeny of various mating-types in crosses of mat+ su8-1 mat-* transformants with mat- and mat+ reference strains. The indicated phenotypes are those of the spores carrying the ex-transformant locus. The mat locus of the tester partner segregates normally. Asterisks mark the mat information from the transforming cosmid. @ = full mat+ expression. 0 = full mat- expression. (+) = mat+, high fertilizer, poor ascus productor. + = mat+ high fertilizer, no ascus production. - = mat- high fertilizer, very poor ascus productor. su = su8-1, transformation marker carried by the cosmids. The thickness of the arrows roughly indicates the relative frequency of each kind of progeny from the crosses. DNA of progeny class 1 is shown in Figure 3A (lane 2) and Figure 4 (lane 2); class 4 in Figure 3A (lanes 3 and b) and Figure 4 (lane 3).

needed to understand this situation. The infertility of self-mating Podospora and Neurospora transformants also distinguishes these ascomycetes from the basidi- omycete U. maydis , where heterozygosity at the b locus in a haploid nucleus allows the fungus to com- plete the sexual stage of its life cycle (KRONSTAD and LEONC 1989).

Perithecial development, a glow in the black box? Figure 5 summarizes the different kinds of progeny observed when a mat+ mat--" strain is crossed either to a mat+ or a mat- tester. The striking fact is that there is a strong selection which goes in the opposite directions for the two crosses. When the cross is performed with a mat- strain, progeny containing the transformant chromosome always express full mat+ function. The mat- information is either lost or kept in its original state (good mater, poor ascus produc- tion). When the cross is performed with a mat+ strain, the resident mat+ information of the mat+ mat-* transformant is either lost or transmitted with a new status (good mater, poor ascus producer). Therefore, the tandemly duplicated DNA formed by homologous recombination of the cosmid in the mat locus is the target of different kinds of events that change the structure and (or) the expression of the dual infor- mation. We presume that within the developing peri- thecium, those nuclei in which the DNA re- arrangements preserve the information necessary for fertility with the tester partner are the ones that undergo caryogamy, meiosis and ascus development.

After more than 40 years and despite increasingly precise cytological analysis, the developing perithe- cium remains a black box (ZICKLER 1973; SIMONET and ZICKLER 1978). After fertilization, the male and the female nuclei divide, giving rise to a heterocar- yotic syncytium inside the fruiting body. Then dicar-

546 M. Picard, R. Debuchy and E. Coppin

yotic cells emerge from the syncytium. These dicar- yons divide in such a way as to preserve the dicaryotic structure except in the apex where caryogamy occurs. Each diploid cell that will undergo meiosis and spor- ulation processes to give rise to an ascus contains mat+ and mat- information. However it is not known whether the mat+ and mat- nuclei recognize each other at the time of dicaryon formation or at the time of caryogamy. Furthermore, nobody knows how this recognition is achieved. Cytological analysis of the sterile perithecia induced in mat+ mat-* X mat+ crosses or after self-fertilization of the mat+ mat- strains may enlighten us with respect to this old prob- lem.

Structural and functional analyses of the mat+ and mat- information required for mating and perithe- cium development should help determine whether the mat genes of filamentous fungi are structural or reg- ulatory genes (or both). Comparison of the Podospora mating-type genes to those of other fungi should help explain how these genes and the control of sexual reproduction evolved among ascomycetes and basidi- omycetes.

We are much indebted to YVES BRYGOO for initiating the work o n Podospora mating types and to LOUISE GLASS, CHUCK STABEN, CHARLES YANOFSKY and BOB METZENBERC for their kind gift of Neurospora A and a clones. We thank JOEL BEGUERET for providing us the mat+ cosmid library. We acknowledge the technical assistance of ARLETTE .4DoUTTE and SIMONE LARGEAULT. We are also grateful to V~RONIQUE BERTEAUX for her participation to genetic and molecular analysis of the transformants. This work was supported by lNSERM and Fondation pour la Recherche Medicale.

LITERATURE CITED

BEACH, D. H., 1983 Cell type switching by DNA transposition in fission yeast. Nature 305 682-683.

BEADLE, G. W., and V. L. COONRADT, 1944 Heterocaryosis in Neurospora crassa. Genetics 2 9 291-308.

BEGUERET, J., V. RAZANAMPARANY, M. PERROT and C. BARREAU, 1984 Cloning gene ura5 for the orotidylic acid pyrophospho- rylase of the filamentous fungus Podospora anserina: transfor- mation of protoplasts. Gene 3 2 487-492.

BERGES, T., and C. BARREAU, 1989 Heat shock at an elevated temperature improves transformation effkiency of protoplasts from Podospora anserina. J. Gen. Microbiol. 135: 601-604.

BERNET, J., 1967 Les syst&mes d’incompatibilit0 chez le Podospora anserina. C. R. Acad. Sci. 265: 1330-1333.

BOYER, H. W., and D. ROULLAND-DUSSOIX, 1969 A complemen- tation analysis of the restriction and modification of DNA in Escherichia coli. J. Mol. Biol. 41: 459-472.

BRYGOO, Y., and R. DEBUCHY, 1985 Transformation by integra- tion in Podospora anserina. l . Methodology and phenomenol- ogy. Mol. Gen. Genet. 200 128-131.

COPPIN-RAYNAL, E., M. PICARD and S. ARNAISE, 1989 Transformation by integration in Podospora anserina. 111. Replacement of a chromosome segment by a two-step process. Mol. Gen. Genet. 219 270-276.

DEBUCHY, R., and Y. BRYGOO, 1985 Cloning of opal suppressor tRNA genes of a filamentous fungus reveals two tRNAser genes with unexpected structural differences. EMBO J. 4: 3553- 3556.

DEBUCHY, R., E. COPPIN-RAYNAL, D. LE COZE and Y. BRYGOO, 1988 Walking on a chromosome towards a centromere in the filamentous fungus Podospora anserina: cloning of a sequence lethal at a twwopy state. Cum. Genet. 13: 105-1 11.

EBERSOLD, W. T., 1967 Chlamydomonas reinhardtii: heterozygous diploid strains. Science 151: 447-449.

ESSER, K., 1974 Podospora anserina, pp. 531-551 in Handbook of Genetics, Vol. 1, edited by R. C. KING. Plenum, New York.

FEINBERC, A. P., and B. VWELSTEIN, 1983 A technique for radiolabeling DNA restriction endonuclease fragments to high specificity activity. Anal. Biochem. 132 6-13.

FEINBERC, A. P., and B. VWELSTEIN, 1984 A technique for radiolabeling DNA restriction endonuclease fragments to high specificity activity. Addendum. Anal. Biochem. 137: 266-267.

GALLOWAY, R. E., and U. W. GOODENOUCH, 1985 Geneticanalysis of mating-type locus linked mutations in Chlamydomonas rein- hardtii. Genetics 11 1: 447-461.

GIASSON, L., C. A. SPECHT, C. MILCRIM, C. P. NOVOTNY and R. C. ULLRICH, 1989 Cloning and comparison of Aa mating-type alleles of the Basidiomycete Schirophyllum commune. Mol. Gen. Genet. 218 72-77.

GLASS, N. L., J. GROTELUESCHEN and R. L. METZENBERC, 1990 The Neurospora crassa A mating-type region. Proc. Natl. Acad. Sci. USA 87: 4912-4916.

GLASS, N. L., S. J. VOLLMER, C. STABEN, J. GROTELUESCHEN, R. L. METZENBERG and C. YANOFSKY, 1988 DNAs of the two mat- ing-type alleles of Neurospora crassa are highly dissimilar. Sci- ence 241: 570-573.

HABER, J. E., and J. P. GEORGE, 1979 A mutation that permits the expression of normally silent copies of mating-type infor- mation in Saccharomyces cerevisiae. Genetics 93: 13-35.

HANAHAN, D., 1983 Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166 557-580.

HERSKOWITZ, I., 1989 A regulatory hierarchy for cell specializa- tion in yeast. Nature 342 749-757.

JOHNSON, D. A., J. W. GAUTSCH, J. R. SPORTSMAN and J. H. ELDER, 1984 Improved technique utilizing nonfat dry milk for analy- sis of proteins and nucleic acids transferred to nitrocelulose. Gene Anal. Tech. 1: 3-8.

KELLY, M., J. BURKE, M. SMITH, A. KLAR and D. BEACH, 1988 Four mating-type genes control sexual differentiation in the fission yeast. EMBO J. 7: 1537-1547.

KLAR, A. J. S., S. FOCEL and K. MACLEOD, 1979 MARI-a regu- lator of the HMa and H M a loci in Saccharomyces cerevisiae. Genetics 93: 37-50.

KRONSTAD, J. W., and S. A. LEONG, 1989 Isolation of two alleles of the b locus of Ustilago maydis. Proc. Natl. Acad. Sci. USA

MANIATIS, T., E. F. FRITSCH and J. SAMBROOK, 1982 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

MARCOU, D., 1961 Notion de longivite et nature cytoplasmique du dkterminant de la senescence chez quelques champignons. Ann. Sci. Nat. Bot. 11: 653-764.

METZENBERC, R. L., and N. L. GLASS, 1990 Mating-type and mating strategies in Neurospora. BioEssays 1 2 53-59.

MUTASA, E. S., TYMON, A. M., GOTTCENS, B., MELLON, F. M., LITTLE, P. F. R. and L. A. CASSELTON, 1990 Molecular organisation of an A mating-type factor of the basidiomycete fungus Coprinus cinereus. Curr. Genet. 18: 223-229.

NASMYTH, K. A,, and K. TATCHELL, 1980 The structure of trans- posable yeast mating-type loci. Cell 1 9 753-764.

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

PICARD, M., R. DEBUCHY, J. JULIEN and Y. BRYCOO, 1987 Transformation by integration in Podospora anserina.

8 6 978-982.

Podospora Dual-Mater Transformants 547

11. Targeting to the resident locus with cosmids and instability of the transformants. Mol. Gen. Genet. 210 129-134.

PUHALLA, J. E.,1970 Genetics studies of the b incompatibility locus of Ustilago maydis. Genet. Res. 1 6 229-232.

RAZANAMPARANY, V., and J. BEGUERET, 1986 Positive screening and transformation of ura5 mutants in the fungus Podospora anserina : characterization of the transformants. Curr. Genet.

RIZET, G., 1953 Sur la longi.viti des souches de Podospora anser- ina. C. R. Acad. Sci. 237: 1106-1 109.

RIZET, G., and C. ENGELMANN, 1949 Contribution i l’i.tude gi.ni.- tique d’un ascomycete ti.trasport: Podospora anserina (Ces) Rehm. Rev. Cytol. Biol. Veg. 11: 201-304.

RIZET, G., and K. ~ E R , 1953 Sur des phtnomenes d’incompati- biliti. entre souches d’origines differentes chez Podospora anser- ina. C. R. Acad. Sci. 237: 760-761.

SAGER, R., 1954 Mendelian and non-Mendelian inheritance of streptomycin resistance in Chlamydomonas. Proc. Natl. Acad. Sci. USA 40: 356-363.

SCHULZ, B., F. BANNETT, M. DAHL, R. SCHLESINGER, W. SCHAFER, T. MARTIN, 1. HERSKOWITZ and R. KAHMANN, 1990 The b alleles of U. maydis, whose combinations program pathogenic

1 0 81 1-817.

development, code for polypeptides containing a homeodo- main-related motif. Cell 6 0 295-306.

SIMONET, J.-M., and D. ZICKLER, 1978 Genes involved in cary- ogamy and meiosis in Podospora anserina. Mol. Gen. Genet.

STABEN, C., and C. YANOFSKY, 1990 Neurospora crassa a mating- type region. Proc. Natl. Acad. Sci USA 87: 491 7-492 1.

STRATHERN, J. N., SPATOLA, E., MCGILL, C. and J. B. HICKS, 1980 Structure and organization of transposable mat- ing-type cassettes in Saccharomyces yeasts. Proc. Natl. Acad. Sci. USA 77: 2839-2843.

STRATHERN, J., J. HICKS and I. HERSKOWITZ, 1981 Control of cell type in yeast by the mating-type locus. The al-aZ hypothesis. J. Mol. Biol. 147: 357-372.

TURCQ, B., M. DENAYROLES and J. BEGUERET, 1990 Isolation of the two allelic incompatibility genes s and S of the fungus Podospora anserina. Curr. Genet. 17: 297-303.

VOLLMER, S. J., and C. YANOFSKY 1986 Efficient cloning of genes of Neurospora crassa. Proc. Natl. Acad. Sci. USA 83: 4869- 4873.

ZICKLER, D., 1973 La msiose et les mitoses au cours du cycle de quelques Ascomycetes. These d’ttat, Orsay.

Communicating editor: R. L. METZENBERG

162: 237-242.