Heredity 68 (1992) 405—410 Received9May 1991 Genetical Society of Great Britain Evolution of reproductive systems in filamentous ascomycetes. I. Evolution of mating types M. J. NAUTA & R. F. HOEKSTRA Department of Genetics, Agricultural University, Ore yenlaan Z 6703 HA Wageningen, The Netherlands Inthe ascomycete family of Sordariaceae both heterothallism (with two mating types) and homo- thallism (without mating types) are common. A population genetic model is made in an attempt to find out under which conditions evolution from one system to the other is conceivable. Analysis shows that evolution from hetero- to homothallism is possible but evolution from homo- to hetero- thallism is improbable. As in these haploid fungi self-fertilization has other consequences than in diploid organisms, homothallism seems to have little disadvantage. It is found that polymorphism in homo- and heterothallism can be stable, although this has not yet been found in Sordariaceae in nature. Keywords:ascomycetes,evolution, heterothallism, homothallism, mating types, reproduction. thallic species are self-sterile and possess mating types. Introduction Here 'mating types' is defined as 'two different sexes Mostpopulation genetic models about the evolution of without morphological sex-differentiation'. These sex and mating systems concern animals and plants mating types receive different names in different (e.g. Maynard Smith, 1978; Bell, 1982; Stearns, 1987; species but are called +and—inthis study. Michod & Levin, 1988). The fungi are largely over- Note that the terms monoecy and dioecy are confus- looked. Some of the reasons for this may be the relative ing in this context. In plants these terms refer to species lack of knowledge about their population structure and in which individuals produce gametes of only one sex genetics, the complex life cycle of many fungi and the or of both sexes. All Sordariaceae make both, indepen- puzzling variety in reproductive systems. This varia- dent of mating type. [The implication of this will be tion, however, also offers an opportunity for compara- discussed in a subsequent paper (M. J. Nauta & R. F. tive studies of the evolutionary forces that shape the Hoekstra, 1992).] different mating systems. A remarkable phenomenon is the occurrence of This study presents a model of the evolution of both homo- and heterothallic species within many mating types in filamentous ascomycetes, exemplified related ascomycete genera and families. This means by the family Sordanaceae. This family includes some that homo- and/or heterothallism must have evolved genetically well-known species, e.g. Neurospora crassa, independently quite often. One may suspect, then, that Podospora anserina and Sordaria fimicola, which live the threshold for switching from one system to another on rotten plant material or herbivore dung. They show cannot be too high. relatively simple life cycles (see below). Some popula- The purpose of this study is to discover the condi- tion genetic (Perkins & Turner, 1988) and molecular tions, defined in general fitness parameters under (Glass et a!., 1990) data are also available and provide which homothallism can evolve to heterothallism and useful information. The model will probably also be vice versa. valid for many other ascomycete species, but these are not treated explicitly here. Themodel In the Sordariaceae (as in many other ascomycete families) roughly two mating systems exist: homo- Themodel is based on a typical Sordariaceae life cycle thallism and heterothallism. Homothallic species are as presented in Fig. 1. Note the following character- self fertile and have no mating types, whereas hetero- istics. 405 406 M. J. NAUTA & R. F. HOEKSTRA LIL..mycelium ASEX ascospore conidia . / \ ascus SE/ MEIOSIS KARYOGAMY N ascogonium young ascus Fig.1 Lifecycleof the heterothallic model organism. A haploid mycelium contains either nuclei of mating type +(•)orof mating type —(0). Theconidia can develop asexually into a new mycelium or fertilize ascogonia of the opposite mating type. After karyogamy and meiosis an ascus with eight ascospores (four of each mating type) is formed. The homothallic model organism has the same life cycle, but has no mating types (self-fertilization is possible). 1 The life cycle is haploid. There is only a very short There are two heterothallic mating types (+ and —) stage of diploidy (in the young ascus) which is immedi- withfrequencies x1 and x2, and one homothallic ately followed by meiosis. 'mating type' with frequency x3 (x1 + x2 + x3 =1). 2 Each individual mycelium forms both conidia and These three types are assumed to be determined by ascogonia, that is both male and female gametes. As three alleles at one locus. The fitness of a heterothallic stated above this is completely independent of mating cross + X— equals1, the crosses x —and x + type. have a fitness w1 (w1 1). The homothallic crossing 3 The conidia serve as both male gametes and x has a fitness w2 when it concerns outcrossing asexual spores. [This is a simplification of the situation (frequency 1 —s)and w3 when selfing (frequency s). found in N. crassa, where micro- as well as macro- All mycelia produce the same amount of ascogonia conidia exist. The first seem to serve mainly as a ferti- and conidia. There is an excess of conidia formed, so lizing agent and the second as an asexual spore (Perkins all ascogonia are fertilized. [This can be compared with & Turner, 1988). In the laboratory, however, both can ovules and pollen in higher plants (Charlesworth & perform the two functions.] Charlesworth, 1978).] 4 Because of haploidy self-fertilization does not The conidia disperse randomly over the area. Some imply recombination. [A similar phenomenon in ferns land on unoccupied substrate and have a chance to is called intragametophytic selfing (Klekowski, 1979; germinate. Others land on a mycelium and are able to Hedrick, 1987).] From a genetic point of view then the fertilize ascogonia. There is an active attraction formation of selfed spores is equivalent to forming between unlike mating types as in, for example, Podo- asexual spores. spora anserina (Esser, 1959) and Bombardia lunata Furthermore, the following assumptions are made in (Zickler, 1952). Identical mating types do not attract the model. each other. This means that both a + conidium landing 408 M. J. NAUTA & R. F. HOEKSTRA p0: (C>0 and A+B+C<0)or (C>0and now become A+B+C>0 and A> —+B>O and q<2) or (C<0 andA+B+C<OandA< —4B<0andq>c2), A=2(1—s)(w2—w1), fd: C<OandA+B+C>0, B'2sw3—(1 +s)w,, he: C<OandA+B+C<Oandnot(A< —+B<0 and C' w1. q>c2), ho: C>OandA+B+C>Oandnot(A> —+B>Oand After the right mutation the second mating type can q<2). invade under the same conditions as the first [condition Before discussing some special cases, note that the (5)J.Toachieve a better impression of these formulae, invasion of heterothallism in a homothallic population some special cases will be considered. (See Fig. 2 for is unlikely to happen with two mating types (two illustrations.) simultaneous mutations) at once. That is, one has to 1 No selfing: S consider the introduction of one mating type first. This Homothallism can invade if w1>2/3and hetero- means that x =0 or x2 =0. Elaborating this case gives thallism if w2<w1. condition (5) again for heterothallic invasion. The 2 All homothallics are selling: s=1 frequency q must be 0.5atleast, because the mating Homothallism can invade if Wi> 1 —w3and type must also serve as a heterothallic partner. heterothallism if w1>2w3. The expressions A, B and C in formulae (3) and (4) (a) 1 pa (b) 1 Wi W1 0 0 0 0 1 w2 (c) 1 (d) pa ha Wi W1 he fd 0 0 0 1 0 1 W3 Fig. 2 Equilibrium states in four different cases. po stable polymorphism, he =heterothallism,ho =homothallism, fd =frequencydependent. (a) s=0(noselling, discussed as case 1 in the main text), (b) s=1(all homothallics are selfing, case 2), (c) s =0.5andw2 =w3(homothallic crosses are equally fit), (d) s =0.5and w1 =w2(outbreeding crosses with homothallics are equally fit). EVOLUTION OF MATING TYPES IN ASCOMYCETES 409 3 All outcrossing sex has the same fitness: with the present model is the second condition, which w1= w2 = 1 is less severe here (w1 >2/3). The reason for this is the (a) selling mildly deleterious: w30.5. Hetero- gamete differentiation in the present model and the fact thallism can never invade, homothallism is that no ascogonia are lost by incompatible fusions. stable. It does not mean, however, that the evolution of (b) selling strongly deleterious: w3<0.5.Both heterothallism has become easier. It is hard to find homo- and heterothallism can invade, poly- convincing reasons why w1, w2 and/or w3 should be morphism is stable. considerably smaller than 1. It is clear that conditions for homothallism to invade The idea that heterothallism must have preceded a heterothallic population will be much more easily homothallism in evolution is supported by DNA realized than conditions for heterothallism to invade. A sequencing of the mating type genes of Neurospora heterothallic population can only be stable with strong crassaandthe comparison of these sequences with selection pressure against homothallism and/or selling. other Sordariaceae (Glass et a!., 1990; Metzenberg & The model seems to suggest, therefore, that evolu- Glass, 1990). It has been found that + and —(called tion from hetero- to homothallism may be possible but A and a in Neurospora) are dissimilar and that most that evolution from homo- to heterothallism is (but not all) homothallic species carry homologous expected to be rare. At the same time it shows that, sequences of both mating types in one haploid genome. when evolution from homo- to heterothallism or vice Mating type switching, as in yeast (Herskowitz, 1988), versa occurs, one should also expect to find popula- is improbable in the Sordariaceae.