Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types

Article (PDF Available)inBiological Reviews 86(2):421-42 · May 2011with135 Reads
DOI: 10.1111/j.1469-185X.2010.00153.x · Source: PubMed
Abstract
The advantage of sex has been among the most debated issues in biology. Surprisingly, the question of why sexual reproduction generally requires the combination of distinct gamete classes, such as small and large gametes, or gametes with different mating types, has been much less investigated. Why do systems with alternative gamete classes (i.e. systems with either anisogamy or mating types or both) appear even though they restrict the probability of finding a compatible mating partner? Why does the number of gamete classes vary from zero to thousands, with most often only two classes? We review here the hypotheses proposed to explain the origin, maintenance, number, and loss of gamete classes. We argue that fungi represent highly suitable models to help resolve issues related to the evolution of distinct gamete classes, because the number of mating types vary from zero to thousands across taxa, anisogamy is present or not, and because there are frequent transitions between these conditions. We review the nature and number of gamete classes in fungi, and we attempt to draw inferences from these data on the evolutionary forces responsible for their appearance, loss or maintenance, and number.
BRV brv˙153
B
Dispatch: July 29, 2010 Journal: BRV
Journal Name Manuscript No. Author Received: No of pages: 22 TS: Suresh
Biol. Rev. (2010), pp. 000000. 1
doi: 10.1111/j.1469-185X.2010.00153.x
Having sex, yes, but with whom? Inferences
from fungi on the evolution of anisogamy
and mating types
Sylvain Billiard
1,
,ManuelaL
´
opez-Villavicencio
2
, Benjamin Devier
3
,Michael
E. Hood
4
,C
´
ecile Fairhead
5
and Tatiana Giraud
3
1
Universit´e Lille Nord de France, USTL, GEPV, CNRS, FRE 3268, F-59650 Villeneuve d’Ascq, France
2
Origine, Structure, Evolution de la Biodiversit´e, UMR 7205 CNRS-MNHN, Mus´eum National d’Histoire Naturelle, CP39, 57 rue Cuvier,
75231 Paris Cedex 05, France
3
Ecologie, Syst´ematique et Evolution, Universit´e Paris-Sud, F-91405 Orsay cedex, France; CNRS F-91405 Orsay cedex, France
4
Department of Biology, McGuire Life Sciences Building, Amherst College, Amherst, MA 01002-5000, USA
5
Institut de G´en´etique et Microbiologie, Universit´e Paris-Sud, UMR CNRS 8621, F-91405 Orsay cedex, France
(Received 3 November 2009; revised 15 July 2010; accepted 16 July 2010)
ABSTRACT
The advantage of sex has been among the most debated issues in biology. Surprisingly, the question of why sexual
reproduction generally requires the combination of distinct gamete classes, such as small and large gametes, or gametes
with different mating types, has been much less investigated. Why do systems with alternative gamete classes (i.e. systems
with either anisogamy or mating typ es or bo th) ap pear even though they restrict the probability of finding a compatible
mating partner? Why does the number of gamete classes vary from zero to thousands, with most often only two classes?
We review here the hypotheses proposed to explain the origin, maintenance, number, and loss of gamete classes. We argue
that fungi represent highly suitable models to help resolve issues related to the evolution of distinct gamete classes, because
the number of mating types and anisogamy vary from zero to thousands across taxa, and because there are frequent
transitions between these conditions. We review the nature and number of gamete classes in fungi, and we attempt to draw
inferences from these data on the evolutionary forces responsible for their appearance, loss or maintenance, and number.
Key words: sexual reproduction, gametes, syngamy, mitochondrial inheritance, cytoplasmic genetic elements, gamete
classes, mating types.
CONTENTS
I. Introduction ................................................................................................ 2
II. Models explaining the origin, number and loss of gamete classes ........................................... 3
(1) Origin of gamete classes ................................................................................ 3
(a) The ‘‘by-product’’ model ........................................................................... 3
(b) The ‘‘selfish element’’ model ........................................................................ 4
(c) The ‘‘inbreeding depression avoidance’’ model ..................................................... 4
(d) The ‘‘sex advantage enhancer’’ model .............................................................. 4
(e) The ‘‘ploidy’’ model ................................................................................ 4
(f ) The ‘‘organelle inheritance’’ model ................................................................. 4
(g) The ‘‘gamete size’’ model ........................................................................... 5
(h) The ‘‘anisogamy consequence’’ model .............................................................. 6
(2) The optimal number of gamete classes ................................................................. 6
* Address for cor respondence: (E-mail: sylvain.billiard@univ-lille1.fr).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
2 Sylvain Billiard and others
(a) The number of gamete classes depends on negative frequency-dependent selection and the effective
size of the population ............................................................................... 6
(b) The number of gamete classes depends on their mechanism of origin .............................. 6
(c) The mating type number depends on the mating system ........................................... 6
(d) The mating type number depends on mating dynamics ............................................ 7
(e) The mating type number depends on genomic constraints ......................................... 7
(3) The loss or maintenance of gamete classes ............................................................. 7
(a) Gamete classes can be lost when population or ecological factors render universal compatibility
advantageous ....................................................................................... 8
(b) Gamete classes can be maintained for the same reasons they appeared ............................. 8
III. Mating types, anisogamy and sex in fungi .................................................................. 8
(1) Anisogamy and mitochondrial inheritance in fungi .................................................... 8
(2) Mating types in fungi, homothallism and heterothallism ............................................... 9
IV. Origin, number and loss of gamete classes in fungi ......................................................... 10
(1) The origin and maintenance of gamete classes in fungi ................................................. 10
(a) The ‘‘organelle inheritance’’ model ................................................................. 14
(b) The ‘‘inbreeding depression avoidance’’ model ..................................................... 14
(c) The ‘‘sex advantage enhancer’’ model .............................................................. 14
(d)Othermodels ....................................................................................... 15
(e) An evolutionary scenario for the origin of gamete classes in fungi and in other taxa ................ 15
(2) The optimal number of gamete classes in fungi ........................................................ 16
(3) The loss and maintenance of gamete classes in fungi ................................................... 17
V. Conclusions ................................................................................................ 17
VI. Acknowledgements ......................................................................................... 18
VII. References .................................................................................................. 18
VIII. Supporting Information .................................................................................... 22
I. INTRODUCTION
Sexual reproduction involves a variety of costs relative to
asexual or clonal reproduction (e.g. energy, time, infection
risk, and the ‘‘cost of males’’ in anisogamous species), and
many investigations have attempted to find evolutionary
benefits to sex that balance these costs (Otto, 2009). These
investigations have mainly focused on the consequences of
sex in terms of the genetic variation generated by recombi-
natorial meiosis during the transition from the diploid to the
haploid phase of the sexual life cycle. Syngamy, the counter-
part of meiosis that restores the diploid phase, has received
much less attention, although the regulation of syngamy is
central to the evolution of sexual eukaryotes. Indeed, syn-
gamy is generally not random, occurring almost exclusively
through the combination of distinct and alternative gamete
classes, such as small a nd large gametes (anisogamy, i.e. a
system regulating syngamy between different gamete classes
that is dependent upon size) or gametes with different mating
types (hereafter referring to a molecular mechanism of the
gamete allowing discrimination for syngamy, independent of
size dimorphism). We use the generic term ‘‘gamete classes’’
to refer to forms of gametes that are intercompatible but that
cannot undergo syngamy within the same class, including
both anisogamy and mating types as properties that may
distinguish gamete classes (Fig. 1).
Many questions are raised by the existence of distinct
gamete classes: how can a system that restricts the possibil-
ity of syngamy to only a subset of existing gametes in the
population evolve by natural selection? In the extreme, when
there are two gamete classes, each gamete can only fuse with
one half of the potential gamete partners. How, then, can a
system with two gamete classes arise from a system with no
restrictions? And why are there thousands of gamete classes
in some species while others have reverted to no size or
mating type restrictions on syngamy? The most problematic
case, with the existence of two gamete classes, appears para-
doxically the most frequent among eukaryotes (male and
female, or two mating types). Any new rare gamete class,
being compatible with all other existing classes, should be
selected for, leading to an increase in the number of gamete
classes. Both why gamete classes exist at all and what the
forces are that act upon the number of classes need to b e
understood from an evolutionary point of view.
Broadly satisfying answers to the questions raised above
are still lacking, and groups such as fungi represent highly
suitable organisms to test various hypotheses that have been
put forward based upon theoretical grounds. Indeed, fungi
constitute a u nique kingdom where all combinations of
gamete classes are represented, and with frequent transitions
between different numbers of mating types and/or gamete
size dimorphism. Importantly, anisogamy is decoupled from
mating types in fungi, allowing separate consideration of
the evolutionary forces acting on the two determinants of
gamete classes. Finally, data on the number of mating types,
anisogamy, and on the genetics underlying these traits are
abundant for fungi.
Our aims in this review are to investigate three specific
questions: (1) why did mating types and anisogamy evolve?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 3
Universal compatibility
Mating types
Anisogamy
Anisogamy and mating types
Option 2:
biparental
inheritance
Gametogenesis
and meiosis
Syngamy
Option1:
uniparental
inheritance
Karyogamy
Adults
Gametes
Zygotes
Syngamy limitations:
gamete classes
Typical sexual reproduction
Nuclear DNA
Cytoplasmic DNA
Small gamete
Large gamete
Different mating types
(molecular mechanism of
recognition, no size difference)
Possible syngamy
Anisogamy
Fig. 1. Illustration of typical sexual reproduction, and its consequences on cytoplasmic inheritance and mate limitation, and of
anisogamy and mating types. We have attempted to use unambiguous terminology. ‘‘Anisogamy’’ is defined as a system regulating
syngamy between different gamete classes that is dependent upon size, i.e. with small and large gametes. Gamete size can be
determined by either the haploid genotype (e.g. in mosses) or the diploid genotype (e.g. in other plants and animals) that produces the
gamete. The term ‘‘mating types’’ describes a molecular mechanism allowing discrimination for syngamy at the level of the gamete,
independent of size dimorphism. Mating type determination can also depend on the haploid genotype (e.g. in fungi, and pollen in
gametophytic self-incompatibility in plants) or the diploid genotype (e.g. sporophytic self-incompatibility in plants) that produced the
gamete. We finally use the generic term ‘‘gamete classes’’ as classes of gametes that are intercompatible but that cannot undergo
syngamy within the same class, to include both anisogamy and mating types.
(2) What are the evolutionary or proximate forces controlling
the number of gamete classes? In particular, why are systems
with two mating types or two gamete sizes so prevalent,
despite being apriorithe least favourable case? (3) Why are
gamete classes sometimes lost and sometimes maintained
over long periods? We review the theoretical investigations
performed so far, emphasizing their underlying hypothe-
ses and predictions, as well as their limitations. We finally
attempt to validate or refute these theories using data on
fungi available in the literature.
II. MODELS EXPLAINING THE ORIGIN,
NUMBER AND LOSS OF GAMETE CLASSES
(1) Origin of gamete classes
Numerous non-exclusive evolutionary models have been pro-
posed to explain the emergence of distinct gamete classes.
These models generally assume that all individuals were ini-
tially able to mate with any other in the population and that
zygotes produced by the fusion of two gametes of different
classes had the highest fitness. Below we outline different
models attempting to explain the emergence of gamete
classes, highlighting their central assumptions.
(a) The ‘‘by-product’’ model
This model assumes that mating types appeared as
a by-product of a molecular system ensuring syngamy
(Hoekstra, 1982, 1987; Hoekstra, Iwasa & Weissing, 1990).
A bipolar system of recognition between gametes evolves
in a sexual species by the reciprocal and differential loss
of one of the two recognition molecules (a pheromone and
pheromone receptor) initially carried by all gametes for
signalling the initiation of syngamy. These losses lead to
the coexistence of two gamete classes: one having lost the
receptor and the other having lost the pheromone.
The central assumption of this model is that the molecular
mechanism for recognition is originally symmetrical, with a
pheromone receptor and a pheromone produced by every
gamete, all gametes thus being undifferentiated. It is further
assumed that the probability of mating between undifferen-
tiated gametes may be lower than between differentiated
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
4 Sylvain Billiard and others
gametes, because of self-saturation of receptors by
self-produced pheromones in undifferentiated gametes.
(b) The ‘‘selfish element’’ model
Mating types appeared together with sexual reproduction
and were selected for by providing a means for selfish genetic
elements to transmit horizontally (Hoekstra, 1990; Bell,
1993). In this model a cytoplasmic genetic element that can
promote fusion of its host cell with another cell appears in an
asexual species. This genetic element invades the population
because it is transmitted both verti cally and horizontally,
even if it causes a fitness cost for infected cells. This genetic
element can initiate fusion of its host cell with either infected
or uninfected cells. A subsequent mutation allowing a variant
of this genetic element to promote fusion only with uninfected
cells can then invade. The population then consists of two
categories of cells: uninfected cells and i nfected cells that can
discriminate between themselves and uninfected cells.
The model assumes that the selfish genetic element can
be transmitted vertically during cell division and horizon-
tally after cell fusion. Another assumption is that the selfish
genetic element can be lost with some probability, which
prevents its fixation. The fusion between cells is assumed to
be costly, which allows the discriminating selfish genetic ele-
ments, inducing fusions with only uninfected cells, to increase
in frequency when rare. The model predicts an asymmetry
between mating types regarding the molecular recognition
mechanism:onlyonetypeofgameteisexpectedtocarrythe
genetic element inducing syngamy and allowing discrimina-
tion, and it should be transmitted to both daughter cells.
(c) The ‘‘inbreeding depression avoidance’’ model
Mating types were selected for by providing a means to
avoid selfing and inbreeding depression (Charlesworth &
Charlesworth, 1979; Uyenoyama, 1988a, b). If selfing rates
and inbreeding depression are high enough, mating types
that function to discriminate at the level of the diploid indi-
vidual can invade a panmictic population. This is because
they prevent fertilization between individuals bearing the
same allele (i.e. promote disassortative mating), preventing
selfing and thus inbreeding depression by reducing the pro-
portion of the genetic load in a homozygous condition for
genes required during the diploid phase of the life cycle.
These models assume inbreeding depression, i.e. that selfed
or inbred offspring have a lower fitness than outbred off-
spring. Individuals are assumed to be diploid and inbreeding
depression arises from exposure to selection of rare deleteri-
ous mutations that become homozygous in inbred offspring.
These models can only be applied to species that are diploid
and whose mating type determinism is controlled at the
diploid level, such as plants and animals (Fig. 2).
(d) The ‘‘sex advantage enhancer’’ model
Mating types were selected for by maximizing the recom-
binatorial advantages of sex (Cz
´
ar
´
an & Hoekstra, 2004).
Several hypotheses have been invoked for the advantage of
sex (DNA repair, breakdown of detrimental genetic associa-
tions and increase of genetic variance i n offspring, red queen
hypothesis; Agrawal, 2006). If sex is indeed advantageous for
any of these reasons, syngamy must occur between gametes
that are not identical haploid clones (Nauta & Hoekstra,
1992). Even DNA repair needs intact sequence fragments
as templates to repair damaged genes. Cz
´
ar
´
an & Hoekstra
(2004) proposed that mating types were selected for because
they prevented syngamy between identical haploid clones.
They showed that, if the spatial distribution of identical hap-
loids is clumped, mating types determined at the haploid
level can evolve to ensure mating between different hap-
loid genotypes. This model is different from the ‘‘inbreeding
depression avoidance’’ described above, which is based upon
the exposure of deleterious mutations to selection when
inbreeding r enders them homozygous.
Within this model no deleterious mutations are assumed,
the disadvantage of mating among haploid clones of the same
genotype is that costs of sex are p aid (loss of time and energy),
while no b enefit is gained by recombining between identical
haploids. Sex is assumed to b e advantageous. Gametes are
assumed to have limited dispersal ability.
This model can only be applied to organisms like fungi or
algae in which the life cycle involves replication of haploid
genotypes and allows syngamy between identical haploid
clones and for which the genetic determinism of mating
type occurs at the haploid level (Fig. 2). The model predicts
that mating types should be present in species exhibiting
the following characteristics: mating types determined at
the haploid level, possibility of syngamy between identical
haploids, and limited gamete dispersal.
(e) The ‘‘ploidy’’ model
Mating types were selected for by allowing cells to respond
appropriately to their ploidy status (Haag, 2007). Mating
types evolved to provide a cellular signal that indicates
the genomic condition of diploidy. This prevents further
syngamy and indicates the potential to proceed with meiosis.
Mating types are selected for that allow gametes to undergo
syngamy only with another haploid cell. This model has only
been articulated as a verbal argument so far.
(f ) The ‘‘organelle inheritance’’ model
Mating types and/or anisogamy were selected for by reg-
ulating the inheritance of cytoplasmic genetic elements
and avoiding the cost of intra-genomic conflict (Hurst &
Hamilton, 1992; Hutson & Law, 1993; Yamauchi, 2003).
One of the most important consequences of syngamy is
the mixing of cytoplasmic genetic elements from different
individuals. It is generally accepted that within-cell compe-
tition between different cytoplasmic elements can result in
conflicting levels of selection that impair the fitness of the
cell. For instance, mitochondrial mutants with an impaired
contribution to the cell’s performance but with more rapid
within-cell replication will be able to invade. Such cases have
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 5
Selfing
Intra-haploid mating
Outcrossing
(Mitoses)
2n
2n
2n
2n
Diploid 1 Diploid 2
Zygotes
Meioses
Tetrads
(Mitoses)
nnn
nn
nn
nn
nn
n
Prevented by heterothallism in fungi (existence of mating types determined at the haploid level)
Prevented by separate germ lines for male and female functions in plants and animals
Prevented by self-incompatibility systems in plants (exi
stence of mating types determined at the diploid level)
Always possible in fungi (mating types determined at the haploid level)
nnn
nn
nn
nn
nn
n
Possible step of clonal
multiplication of gametes
or extended haploid phase
for haploid organisms
Fig. 2. Diagram showing the products of meiosis from two diploid genotypes, which are often able to proliferate mitotically in
fungi and contribute eventually to mating. Various m o des of mating are indicated by arrows originating from the far-left meiotic
product. Homothallism does not require genetic differences for mating and allows union between any two cells, often including
descendants of the same meiotic product as s hown here (called intra-haploid mating) and leading to genome-wide homozygosity.
Selfing is most commonly used to indicate mating between cells derived from two different meiotic tetrads of the same diploid
genotype (although homothalllism and intra-haploid mating can be considered subcategories of selfing). Outcrossing is the m ating
between cells derived from meioses in two different diploid genotypes. Intra-haploid mating is possible in homothallic fungi, but
is prevented in heterothallic species by the existence of mating types determined at the haploid level. Plants and animals cannot
perform intra-haploid mating because of the existence of separate germ lines for male and female functions. Mating types and
heterothallism cannot prevent selfing in fungi, while self-incompatibility systems prevent selfing in plants.
in fact been reported (Hintz, Anderson & Horgen, 1988;
Taylor, Zeyl & Cooke, 2002). This model considers that
mating types have evolved in isogamous species because it
was advantageous for the nuclear genome to limit genetic
conflicts that arise through the mixing of diverse cytoplasmic
elements by enforcing uniparental inheritance of cytoplasm.
Anisogamy has also been proposed to have evolved to avoid
such intra-genomic conflicts by enforcing uniparental inher-
itance of cytoplasm via only the larger gametes (Hurst &
Hamilton 1992; Hurst, 1996).
This model predicts that whenever cytoplasmic mixing
occurs at syngamy, gamete classes should e xist and should
be associated with uniparental control over inheritance of
organelles.
These models generally assume at least three steps: appear-
ance of a selfish cytoplasmic genetic element, appearance of a
mutation allowing uniparental inheritance of the cytoplasm,
and finally invasion by a mutation allowing syngamy to take
place only between two different types of gametes, with one
transmitting cytoplasm and the other not. A second assump-
tion of these models is that zygotes formed by the fusion of
gametes conferring uniparental inheritance of organelles are
the fittest.
(g) The ‘‘gamete size’’ model
Anisogamy is assumed to have evolved because of fitness
properties directly linked to gamete sizes (motility, zygote
viability, and gamete numbers) (Parker, Baker & Smith,
1972; Maynard Smith, 1978, 1982; Bell, 1978; Charlesworth,
1978; Hoekstra, Janz & Schilstra, 1984; Hoekstra, 1984;
Bulmer, 1994; Matsuda & Abrams, 1999; Randerson &
Hurst, 1999; Dusenbery, 2000, 2006; Maire, Ackermann
& Doebelli, 2001; Bulmer & Parker, 2002; Bonsall, 2006;
Iyer & Roughgarden, 2008; Togashi et al., 2009). This set
of models considers three main features: (1) the relationship
between gamete size and gamete number: small gametes
are cheaper to produce and can thus be generated in larger
numbers; (2) the relationship between zygote size and zygote
fitness: a larger zygote has a higher fitness, imposing selection
on gamete sizes for the ability to form a viable zygote; (3) the
motility of gametes: the existence of both small motile and
large immobile gametes increases the rate of encounters when
fertilization i s external. Most of these models assume the pre-
existence of two mating types upon which the evolution of
gamete size is superimposed.
All these models agree that under more or less stringent
conditions two classes of gamete sizes can evolve. However,
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
6 Sylvain Billiard and others
when the assumption of the pre-existence of two mating
typesisrelaxed(i.e. when the mating is not limited apriorito
fertilization between small and large gametes), the conditions
for the evolution of anisogamy are more stringent (e.g. Iyer &
Roughgarden, 2008). Furthermore, mating can occur only
between small and large gametes in all anisogamous species
and not between two large gametes (Hoekstra, 1987), but it is
difficult to imagine why this could be disfavoured. Differences
in motility between small and large gametes should favour
fusion between different-sized gametes, but why mechanisms
preventing syngamy between large gametes evolved remains
to be explained (see next section). Another limitation is that
these models assume that mating types predate the evolu-
tion of anisogamy, and thus how gamete classes originated
remains enigmatic.
(h) The ‘‘anisogamy consequence’’ model
It is assumed that mating types evolved as a consequence
of anisogamy (Charlesworth, 1978; Parker, 1978; Hoekstra,
1984; Maire et al., 2001; Iyer & Roughgarden, 2008; Togashi
et al., 2009). In these models, anisogamy evolved following
the scenario above but syngamy could still occur between
same-size gametes, and mating types thus evolved to restrict
fusion between same-size gametes. They assume that syn-
gamy between small gametes is disadvantageous because the
resulting small zygote size impairs its fitness.
The problem still remains to find conditions under which
the fusion between large gametes is disadvantageous. To
our knowledge, only verbal models have been proposed for
the evolution of a population with two mating types from a
panmictic anisogamous population, each mating type asso-
ciated with a given gamete size and difference in motility.
Hoekstra (1984) showed that selection should favour linkage
among mating type loci and loci affecting motility and size
dimorphism. But this model still assumes that two mating
types predated this linkage to anisogamy, and it still does not
explain why syngamy between two large immobile gametes
should be disfavoured.
(2) The optimal number of gamete classes
The number of gamete classes is highly variable among
species, even between closely related taxa, ranging from only
two to several thousands. We review here the factors that
have been suggested to affect the number of gamete classes
(see Fig. 3 for a summary).
(a) The number of gamete classes depends on negative
frequency-dependent selection and the effective size of the population
The main consequence of the existence of gamete classes is
that a given gamete can mate with only a fraction of the
available gametes in the population. Hence, in systems where
gamete classes are inter-compatible, any new gamete class
appearing in a population should be favoured by selection
in outcrossing species because it can mate with a larger
proportion of the population than can the resident gamete
classes. This advantage remains as long as it is rarer than the
other gamete classes (Iwasa & Sasaki, 1987). The dynamics of
gamete classes in a population is thus governed by negative
frequency-dependent selection, predicting the evolution of
unlimited numbers of gamete classes. However, populations
are finite and some gamete classes will be lost by drift. Finally,
the number of gamete classes in a population is a balance
between mutations (appearance of novel gamete classes),
drift, and negative frequency-dependent selection.
In this scenario a cost to waiting for a mate is assumed to
exist (Iwasa & Sasaki, 1987).
The balance between mutations (appearance of novel
gamete classes), drift and negative frequency-dependent
selection results in the following predictions for outcrossing
species: (1) the higher the effective population size, the higher
the number of gamete classes (Yokoyama & Hetherington,
1982); (2) gamete classes can be maintained during long evo-
lutionary periods, leading to trans-specific polymorphism (i.e.
with alleles older than species ages, being maintained across
speciation events; Richman, 2000).
(b) The number of gamete classes depends on their mechanism of origin
Limitation of the number of gamete classes is inherent in
the p roposed mechanism for their origin for several of the
models described in Section II.1, in particular the ‘‘selfish
element’’ model, the ‘‘gamete size’’ model, the ‘‘by-product
model’’ and the ‘‘anisogamy consequence’’ model. Indeed,
these models explain the evolution of two gamete classes,
differentiated either by the recognition molecule produced
in a unifactorial system, by the ability to maximize horizontal
transmission of selfish genetic elements, or the advantage of
producing small or large gametes. In those cases, a third
mating type or an intermediate gamete size cannot invade
the population (Parker et al., 1972; Hoekstra, 1982). Under
the by-product model, the appearance of a new receptor
and/or pheromone will not allow the appearance of a new
mating type, because the mutant would not be recognized by
any of the two extant gamete types (Hoekstra, 1987). Where
gamete classes evolve to regulate cytoplasmic conflicts (the
‘‘organelle inheritance’’ model), it has b een shown that the
evolution of more than two classes is also highly unlikely
(Hurst, 1996). Indeed, while it is easy to imagine mecha-
nisms allowing uniparental inheritance of cytoplasm when
there are two gamete classes (one transmitting its cytoplasm,
the other not), it is much harder to imagine a mechanism by
which more than two gamete classes can efficiently manage
to establish rules of uni parental inheritance, although rare
cases have been r eported (Kawano, Kuroiwa & Anderson,
1987; Meland et al., 1991; Moriyama & Kawano, 2003). The
hypotheses of ‘‘inbreeding depression avoidance’’ and ‘‘sex
advantage enhancer’’ by contrast do not predict aprioriany
limitation of the number of gamete classes.
(
c) The mating type number depends on the mating system
Mating systems should also be able to limit the number
of mating types when they are determined at the haploid
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 7
0
2
++
-Advantage of mating with non-identical haploid
-Inbreeding depression (but not for haploids)
-Cytoplasmic conflict (if organelle inheritance linked to mating types)
Homothallism in fungi: for universal compatibility
- Frequency-dependent selection if outcrossing
-Sheltered load
-Selfing
-Small effective size
-Constraint on the evolution of novel specificities
-Anisogamy (if linked to mating types)
-Cytoplasmic conflict (if organelle inheritance linked to mating types)
-Sheltered load
-Selfing
-Small effective size
Number of
mating types
Fig. 3. Diagram summarizing the evolutionary pressures invoked to explain the appearance, maintenance and loss of mating type
systems as well as the evolution of the number of mating types. ‘‘0’’ means no restriction at syngamy, i.e. no gamete classes, ‘‘2’’
means two gamete classes, ‘‘++’’ means more than two gamete classes. The text associated with each arrow lists the evolutionary
pressures suggested to explain the transition in the indicated number of gamete classes. The texts associated with the blocked arrows
list evolutionary pressures that can prevent the transition in the number of gamete classes indicated.
level (Giraud et al., 2008). Mating types determined at the
haploid level, such as i n fungi, do not prevent diploid self-
ing, i.e. syngamy between gametes produced by the same
diploid individual (Giraud et al., 2008). Diploid individuals
in such species are necessarily heterozygous at the mating
type locus and thus produce inter-compatible gamete classes.
Some fungi with haploid-determined mating types in fact
exhibit high rates of selfing (Giraud et al., 2008). Under such
highly selfing mating systems and when gamete dispersal is
low, the advantage for rare alleles would also be low. Low
gamete dispersal reduces the chances of encountering unre-
lated mating partners, and because individuals self anyway,
acquisition of an allele that is new to the population grants
little increase in mating opportunities. This hypothesis has
only been proposed verbally so far (Giraud et al., 2008).
(d) The mating type number depends on mating dynamics
The temporal dynamics of searching for a mating partner is
a factor capable of maintaining only two mating types: when
gametes can wait for a suitable mate, a rare mating type has
little advantage (Iwasa & Sasaki, 1987). This scenario assumes
that the period for mate searching should be long and the cost
of mate searching should be low for the advantage of being
rare to decrease, limiting invasion by new mating types.
(e) The mating type number depends on genomic constraints
The molecular mechanisms underlying mating specifici-
ties may impede the emergence of new mating types. It
may just be too difficult to generate a new variant that
is simultaneously compatible with both pre-existing mating
types by one-step mutations: for instance, mutations may be
needed simultaneously in both the pheromone and the recep-
tor genes, or the new variant will not be able to mate. If recog-
nition between a specific pheromone and a specific receptor is
needed for syngamy, any mutants in the pheromone or recep-
tor genes may even be incompatible with all extant mating
types (Hoekstra, 1987). The number of mating types can also
be greatly influenced by recombination suppression around
mating type loci where recessive deleterious alleles are fixed
in the permanently heterozygous state. Such ‘sheltered’ load
can impede invasion by new mating types because of shared
deleterious mutations in linkage to the ancestral and the
derived mating types (Uyenoyama, 1997; Richman, 2000).
In the extreme case, a whole chromosome p air evolves into
highly dimorphic sex chromosomes with very little recombi-
nation between them. The larger the degree of divergence
between the sex c hromosomes the more likely that a new
gamete class will be incompatible with half of the potential
gametes because of exposing deleterious mutations that were
previously sheltered in the permanently heterozygous condi-
tion, hence limiting the appearance rate of new mating types.
In species with only two mating types, it is predicted that
sheltered load will be found near mating type loci, or that
new mating types are difficult to evolve by simple mutations.
(3) The loss or maintenance of gamete classes
While the number of gamete classes is thus under a variety of
potential evolutionary pressures, the question of why gamete
classes are maintained at all remains highly enigmatic under
multiple scenarios mentioned above. We define here the
‘‘loss of gamete classes’’ as evolution leading to a lack of
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
8 Sylvain Billiard and others
discrimination for syngamy: every haploid cell can undergo
syngamy with any other haploid cell in the population,
including with genetically id entical haploids. We review
here the factors that may be responsible for the loss or the
maintenance of gamete classes (see Fig. 3 for a summary).
(a) Gamete classes can be lost when population or ecological factors
render universal compatibility advantageous
When there are gamete classes, a given gamete cannot mate
with a fraction of the population. If there is a cost to waiting
for a mate, a universally compatible gamete could invade
the population because of the advantage of being able to
cross with all individuals, including itself. The conditions
allowing for loss of gamete classes have been extensively
investigated in the case of self-incompatibility systems in
plants (Charlesworth & Charlesworth, 1979; Porcher &
Lande, 2005). If, for instance there is a limited probabil-
ity of encountering mates and thus a risk of not finding
a compatible mate at all, being able to self may represent
reproductive insurance. This is similar to self-compatibility
evolving in plants due to pollen limitation, where female
reproductive output is decreased because all ovules cannot
be fertilized (Porcher & Lande, 2005). The advantage of find-
ing compatible mates increases when the number of gamete
classes is small (Charlesworth & Charlesworth, 1979). If there
are two gamete classes in the population, a universally com-
patible gamete, when rare (after its appearance by mutation
for instance), may mate with twice as many individuals as
either of the two existing gamete classes. By contrast, if there
is an infinitely large number of gamete classes, a universally
compatible gamete has no outcrossing advantage relative to
other gamete types. So, if the advantage of being universally
compatible is so high, why are gamete classes maintained,
sometimes over long e volutionary periods?
(b) Gamete classes can be maintained for the same reasons they
appeared
Anisogamy is certainly maintained because gametes of
an intermediate size would be disfavoured (Parker et al.,
1972). By contrast, the only theoretical models investigating
the maintenance of mating type systems focused on self-
incompatibility systems in plants and invoked genetic load as
the main force (Porcher & Lande, 2005). We have found no
models that investigate the importance of inbreeding depres-
sion or the advantage of sex when the life cycle is mainly
haploid, but it seems reasonable to suggest that if these forces
can promote the emer g ence of gamete classes, then they can
also contribute to their maintenance.
III. MATING TYPES, ANISOGAMY AND SEX IN
FUNGI
The different models outlined above regarding the origin, loss
or maintenance, and number of gamete classes make differ-
ent assumptions and predictions, which we aim here to assess
using available data for fungi. Fungi are of particular interest
regarding the rules of compatibility between gametes, and
it has been a long-standing challenge to integrate concepts
of sex in fungi with those developed for plants and animals
(Link, 1929). We summarize below the most prominent fea-
tures of the regulation of syngamy i n fungi, which necessarily
leads to some over-simplification. Most terms are defined
andillustratedinFigs1and4.Wecompiledliteraturedata
regarding systems of mating type determination, mode of
reproduction, and number of mating types in the two main
groups of fungi, the ascomycetes (many moulds and yeasts)
and the basidiomycetes (smut, rusts and mushrooms), and
we present these traits in a phylogenetic context (Fig. 5;
supporting information available online, see Section VII).
(1) Anisogamy and mitochondrial inheritance
in fungi
Filamentous ascomycetes are most often anisogamous
(Debuchy, Berteaux-Lecellier & Silar, 2010), while yeasts and
basidiomycetes are frequently isogamous, although excep-
tions exist in both groups. Many filamentous ascomycetes
produce small ‘‘spermatia’’ as dispersing gametes (also called
microconidia) and more complex and sessile ‘‘female’’ struc-
tures (e.g. ascogonia), functioning as large gametes (Debuchy
et al., 2010). Mitochondria are then inherited via the
female-like mating partner (i.e. analogous to maternal inher-
itance), although some paternal leakage can occur (Barroso
& Labarere, 1995; Barr, Neiman & Taylor, 2005; Xu, 2005).
Importantly, anisogamous a scomycetes are almost never
dioecious; haploid individuals of each mating type produce
both large and small gametes, except for a very few cases
with ‘‘female-sterile strains’’ (Leslie & Raju, 1985; Tredway,
Stevenson & Burpee, 2003).
In isogamous yeasts, mitochondria are inherited from both
cells involved in syngamy (‘‘biparentally’’ hereafter) and then
actively segregate during subsequent divisions (Berger &
Yaffe, 2000). In basidiomycetes, sexual reproduction is most
often achieved by the fusion of two haploid mycelia or similar-
sized cells. In this group isogamy is the rule and most often
there are no differentiated cells playing the role of gametes
(Casselton, 2002). Mitochondria can be inherited either uni-
or biparentally in the fused cells (Barr et al., 2005; Xu, 2005;
Barroso & Labarere, 1995). When inheritance is uniparental
at plasmogamy, it can be controlled by mating type. When
inheritance is biparental at plasmogamy, they nevertheless
usually segregate rapidly during the subsequent mitoses, most
probably by an active mechanism (Jin & Horgen, 1994). The
‘‘Buller phenomenon’’ in basidiomycetes describes the case
where organelle uniparental inheritance is ensured without
mating type having a role: a heterokaryotic mycelium (i.e.
with two different haploid nuclei per cell) can fertilize a
homokaryotic mycelium (Buller, 1931; Callac et al., 2006).
Mitochondria in the resulting heterokaryotic mycelium are
then those of the recipient homokaryotic mycelium. There
are some cases of anisogamy in basidiomycetes, such as in
some rusts (Bruckart et al., 2010), but the rules of mitochon-
drial inheritance are unknown.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 9
most heterothallic ascomycetes
(almost always only two mating types)
MAT1 MAT2
MAT1 MAT2
Anisogamous-
Isogamous-
Bipolar or unifactorial (one locus, two mating types or more)
Tetrapolar or bifactorial (two loci, numerous mating types )
most heterothallic basidiomycetes
MAT1 MAT2
MAT3 MAT4
Possible syngamy
MAT1
MAT2
Alleles or idiomorphs that deter
mine syngamy compatibility,
encoding transcriptional regulators of pheromones,
pheromone receptors and other mating determination genes
(in ascomycetes) or encoding directly pheromones and their
receptors (in basidiomycetes)
Gametes (spores)
Hyphae
Legend
(A)
MAT3
MAT4
(B)
Homothallism due to the presence of the two mating types in the same
nucleus is the most frequent case
MAT1
MAT2
MAT1
MAT2
Ascomycetes
Homothallism due to MAT switching: A single MAT is expressed at any
one time but the cell can switch MAT; this is frequent in yeasts
Ascomycetes
(& basidiomycetes)
MAT1
(MAT2)
(MAT1)
MAT2
Basidiomycetes
& ascomycetes
MAT1
MAT1
Homothallism due to mating between cells of the same mating type, sometimes
called same-sex mating; found in rare cases only
Pseudo-homothallism due to heterokaryon with two nuclei of opposite mating
types in a spore; found in rare cases only
MAT1 MAT2 MAT1 MAT2
Ascomycetes &
basidiomycetes
Fig. 4. Summary of modes of syngamy in fungi. (A) Heterothallism; (B) different forms of homothallism.
(2) Mating types in fungi, homothallism
and heterothallism
In addition to anisogamy when it is present, strict molecular
mechanisms exist in most fungi governing which haploid cells
can fuse, controlled by mating type (MAT ) loci that p resent
two or many alternative forms. Mating types are not asso-
ciated with anisogamy (Fig. 4) and all genotypes in anisoga-
mous fungi are able to produce large and small gametes.
Mating type determinism is regulated strictly in the haploid
stage in fungi, in contrast to most well-studied eukaryotes, in
which syngamy compatibility is determined by developmen-
tal differences between the sexes or male and female organs
in the diploid stage. In fungal species where syngamy is
restricted by the MAT locus, successful fusion of gametes can
occur between haploids carrying different mating type alleles;
this is called heterothallism. Compatibility can be determined
by alleles at a single locus, a condition called unifactorial (or
bipolar) heterothallism, or by alleles at two unlinked loci,
called bifactorial (or tetrapolar) heterothallism. Most het-
erothallic ascomycete fungi present a unifactorial system,
regulated by highly dissimilar mating type genes present at
the same locus, r eferred to as ‘‘idiomorphs’’ rather than alle-
les due to the uncertainty of their origin by common descent
(Debuchy et al., 2010; Butler, 2007). These genes code for
transcription factors that induce, in particular, the production
of pheromones and pheromone receptors. Although most
of the ascomycete species present only two alternative
idiomorphs, rare cases of a single MAT locus with more than
two mating types have been reported in this group (Fig. 5A).
By contrast, mating types in basidiomycetes correspond to
different alleles of the same genes (i.e. they are not idiomorphs;
Devier et al., 2009), coding for pheromones, p heromone
receptors, and homeodomain proteins. Most basidiomycete
fungi present bifactorial systems, with the pheromone-related
and homeodomain genes at unlinked loci. The MAT loci
in bifactorial basidiomycetes are often highly polymorphic,
yielding species with many allelic forms that combine to pro-
duce thousands of mating types (Casselton & Kues, 2007).
Roughly a quarter of basidiomycete species present a unifac-
torial heterothallism, which is either due to the tight linkage
of the two loci containing pheromone-related genes and
homeodomain genes, or the loss of function of one of the
two loci contributing to mating type specificity (Fraser et al.,
2007). Unifactorial basidiomycetes show either two or more
mating types, while bifactorial basidiomycetes always show
more than two (Fig. 5C).
In contrast to heterothallism described above, in which
syngamy can only occur between haploid cells carrying
different mating type alleles, many fungi are homothallic,
especially among ascomycetes, meaning in the strict historical
(experimental) sense that a haploid individual is able to mate
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
10 Sylvain Billiard and others
with its own mitotic descendants. Homothallism is often
called ‘‘self-compatibility’’ in the fungal literature, but we
avoid this term here. ‘‘Self-compatibility’’ indeed can be
confused with selfing ability in diploid plants and animals,
which involves syngamy between the products of separate
meioses and has different genetic consequences than the
genome-wide homozygosity resulting from ‘‘intra-haploid
mating’’ by homothallic fungi (Giraud et al., 2008) (Fig. 2).
Furthermore, the term ‘‘self-compatible’’ may suggest that
the evolutionary cause responsible for homothallism was
to allow for intra-haploid mating, which is not necessarily
true. Homothallism can indeed be viewed also as a lack of
discrimination at syngamy: each haploid is compatible with
all other haploids in the population in homothallic species,
including genetically identical haploids.
Three different proximal mechanisms can cause homoth-
allism, which occurs mostly in ascomycetes (Fig. 5A). The
most frequent proximal cause is that each haploid possesses
the two idiomorphs in its genome (Yun et al., 1999; Lin
& Heitman, 2007). It has never been investigated how-
ever whether both mating type idiomorphs are expressed
in all cells or if they are expressed alternately. Recent
data on mutants however suggest that both pheromones
and pheromone receptors would be constitutively expressed
(Alby, Schaefer & Bennet, 2009). In some fungi, mainly
yeasts, homothallism may arise from mating-type switching:
in some yeasts the expressed MAT locus can be replaced
during mitotic divisions by translocation involving either one
of the two flanking silent MAT idiomorphs (Strathern et al.,
1982; Butler, 2007). Another, rarer, type of homothallism
is to allow for mating between cells carrying a single and
identical MAT allele, which is called ‘‘same-sex mating’’ (Lin
& Heitman, 2007). Phylogenies of ascomycetes and basid-
iomycetes show that homothallism has evolved at least a
dozen times independently. Some other fungal species are
referred to as ‘‘pseudohomothallic’’, which involves cyto-
logical processes that package nuclei from a single meiosis,
but with different mating types, into a single spore (Lin &
Heitman, 2007; Raju & Perkins, 1994; Rizet, 1943). Thus
pseudohomothallism is more analogous to diploid selfing,
and even intra-tetrad mating, and does not lead to the
same genetic consequences as intra-haploid mating (Lin &
Heitman, 2007; Raju & Perkins, 1994; Rizet, 1943).
IV. ORIGIN, NUMBER AND LOSS OF GAMETE
CLASSES IN FUNGI
(1) The origin and maintenance of gamete classes in
fungi
Fungi are relevant organisms to the understanding of why
systems might evolve that impose limitations at syngamy, i.e.
to study the emergence and maintenance of gamete classes,
particularly since anisogamy and mating types in fungi are
decoupled (Fig. 4) and variable (Fig. 5). Among all the models
Fig. 5. Phylogenies indicating mating type systems in fungi and the number of mating types. For further details and references, see
supplementary material available online (see Section VIII for details). (A) Phylogeny of ascomycetes based on previously published
phylogenies showing the diversity of reproductive systems. In all ascomycetes, compatibility is determined by a single mating-type
locus with two different idiomorphs/alleles, except in Glomerella cingulata where multiple alternative alleles exist. In heterothallic
species only one of the two MAT idiomorphs is present in each haploid. Pseudohomothallic species produce spores w ith two nuclei
of opposite mating types. Homothallic species bear two MAT genes in a single haploid either linked in the same chromosome or
in different chromosomes. Same-sex-mating species are homothallic species where only one MAT idiomorph has been found in
the genome. We use the term ‘‘same-sex m ating’’ both for species undergoing mating between identical haploids carrying a single
mating type, such as in some homothallic Neurospora species, and for species where genetically different cells carrying the same
mating type idiomorph/allele are able to fuse, such as in Cryptococcus (see C). Marginal systems occur in some species where low
proportions of strains present a different system. We also indicate species where regions of suppressed recombination around the
mating locus have been descr ibed. Anisogamy is the rule in ascomycetes, is independent of mating type, and mitochondria are
maternally inherited in most cases. Where data were not available for a particular character, no symbol is shown. (B) Phylogeny
of yeast ascomycetes (Saccharomycotina) based on previously published phylogenies showing the diversity of reproductive systems.
The two MAT alleles are idiomorphs coding for regulators of transcription. Homothallic species bear three MAT genes in a single
haploid, one copy being expressed, the other two being silent and able to replace the expressed copy. In heterothallic species only
one of two MAT idiomorphs is present in each haploid, except in the species with the symbol ‘‘3’’ that bear three copies but
without being able to switch mating types. Same-sex mating species are homothallic species where only one MAT idiomorph has
been found in the genome. Yeasts are isogamous, mitochondria are inherited biparentally and are then actively segregated. Where
data were not available for a particular character, no symbol is represented. (C) P hylogeny of basidiomycetes based on previously
published phylogenies showing the diversity of reproductive systems. In heterothallic species the MAT locus includes genes coding
for pheromone receptors and pheromones, as well as homeodomain proteins, either at a single locus (unifactorial species) or at two
distinct loci (bifactorial species). Pseudohomothallic species produce spores with two nuclei of opposite mating type. Same-sex mating
species are homothallic species with a single MAT allele, but able to mate with cells carrying identical MAT alleles. The numbers of
MAT alleles are indicated (two or multiple alleles). We also indicate species where regions of suppressed recombination around the
mating locus have been described. Most basidiomycetes are isogamous, plasmogamy occurring by fusion between hyphae, without
differentiation of particular cells into gametes. Mitochondria can be inherited either uni- or biparentally at plasmogamy (indicated
by different arrow types) in the fused cells. When inheritance is biparental, mitochondria from the two parents are however usually
segregated rapidly during the subsequent mitoses, probably by an active mechanism. Where data were not available for a particular
character, no symbol is represented.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 11
Podospora anserina
Chaetomium aureum
N. galapagosensis
Neurospora africana
N. dodgei
N. lineolata
Gelasinospora calospora
G. tetrasperma
N. sublineolata
G. cerealis
N. pannonica
N. terricola
N. crassa
N. sitophila
N. tetrasperma
N. intermedia
N. discreta
Sordaria fimicola
S. macrospora
S. brevicollis
S. sclerogenia
C. elatum
C. funicula
C. globosum
Truncated MAT A-2
Truncated MAT A-2
Truncated MAT A-3
Truncated MAT a-1
Truncated MAT a-1, no MAT A-3
Truncated MAT a-1
Arnium arizonense
F. graminearum
F. asiaticum
F. acaciae-mearnsii
F. boothii
F. mesoamericanum
F. cortaderiae
F. brasilicum
F. austroamericanum
F. meridonale
F. lunulosporum
F. cerealis
F. culmorum
F. pseudograminearum
Magnaporthe grisea
Gibberella fujikuroi
Fusarium oxysporum
Glomerella cingulata
G. lindemuthiana
G. graminicola
Ophiostoma novo-ulmi
Cryphonectria parasitica
Sordariomycetes
C. coerulescens
Chromocrea spinulosa
Ceratocystis eucalypti
C. fimbriata
C. fagacearum
C. pinicola
Leotiomycetes
Sclerotinia trifoliorum
S. sclerotiorum
S. minor
Botrytis cinerea
B. globosa
B. porri
B. convoluta
Botryotinia squamosa
Pyrenopeziza brassicae
Eurotiomycetes
Dothideomycetes
To Saccharomycotina
To Basidiomycota
No MAT a-1
Heterothallic
Pseudohomothallic
Homothallic (two MAT idiomorphs)
Apomictic, asexual, putatively asexual
Homothallic (a single MAT idiomorph,
same-sex mating)
More than two mating types
Suppressed recombination around
MAT locus
Marginal system only present in some
strains
Suggested acquisition of MAT genes
by lateral gene transfer
C. virescens
O. ulmi
A
Fig. 5. Legend on previous page.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
12 Sylvain Billiard and others
Sordariomycetes
Cochliobolus heterostrophus
Bipolaris sacchari
C. luttrellii
C. carbonum
C. victoriae
C. homomorphus
C. ellisii
C. cymbopogonis
C kusanoi
C. intermedius
Stemphylium herbarum
S. gracilariae
S. majusculum
S. astragali
S. eturmiunum
S. botryosum
S. lancipes
S. solani
S. callistephi
S. triglochinicola
S. paludiscirpi
S. sarciniforme
S. lotii
S. trifolii
S. xanthosomatis
M. graminicola
Aspergillus fumigatus
Neosartorya fischeri
N. spathulata
N. otanii
N. fennelliae
N. nishimurae
N. udagawae
Petromyces alliaceus
Penicillium chrysogenum
Talaromyces flavus
T. stipitatus
T. derxii
Penicillium marneffei
Histoplasma capsulatum
Coccidioides immitis
C. posadasii
Unlinked MAT
Dothideomycetes
To Saccharomycotina
To Basidiomycota
Eurotiomycetes
Mycosphaerella fijiensis
M. eumusae
M. musicola
Eurotiomycetes
N. indohii
N. tsurutae
A. nidulans
A. flavus
A. oryzae
Unlinked MAT
Linked MAT
A. parasiticus
Heterothallic
Pseudohomothallic
Homothallic (two MAT idiomorphs)
Apomictic, asexual, putatively asexual
Homothallic (a single MAT idiomorph,
same-sex mating)
More than two Mating Types
Suppressed recombination around
MAT locus
Marginal system only present in some
strains
Suggested acquisition of MAT genes
by lateral gene transfer
Fig. 5. (Cont.)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 13
Heterothallic
Pseudohomothallic (spores including nuc
lei with different mating types)
Homothallic by mating type switching
Homothallic by same sex mating
Coprinellus disseminatus
Pholiotoa nameko
Amylostereum areolatum
Heterobasidion insulare
Heterobasidion annosum
Ganoderma boninense
Thyphula ishikariensis
Thanatephorus cucumeris
Cryptococcus neoformans
Sporisorium reilianum
Ustilago maydis
Ustilago hordei
Microbotryum violaceum
Rhodosporium turoloides
Tetrapolar
Bipolar
Multiple mating types
Two mating types
Region of supressed recombination
Schizophyllum commune
Coprinus cinereus
Pleurotus djamor
Laccaria bicolor
2
Agrocybe aegerita
Stropharia rugosoannulata
2
2
2
2
2
2
Uniparental mitochondrial inheritance at plasmogamy
Biparental mitochondrial inheritance
at plasmogamy
Agaricus bisporus
Armillaria bulbosa
Agaricus bitorquis
Lentinula edodes
Uniparental mitochondrial inheritance at
plasmogamy controlled by mating type
Marginal system only present in
some strains
Saccharomyces cerevisiae
Saccharomyces paradoxus
Candida glabrata
Kluyveromyces delphensis
Candida castellii
Zygosaccharomyces rouxii
Kluyveromyces thermotolerans
Saccharomyces kluyveri
Kluyveromyces lactis var. lactis
Eremothecium (Ashbya) gossypii
Candida albicans
Candida tropicalis
Candida parapsilosis
Lodderomyces elongisporus
Candida (Pichia) guilliermondii
Debaryomyces hansenii
Candida (Clavispora) lusitaniae
Yarrowia lipolytica
Schizosaccharomyces pombe
3
3
3
Heterothallic
Homothallic by MAT switching (silent cassettes are present)
Homothallic without switching by same-sex mating
Putatively asexual
Three MAT loci are present despite lack of MAT switching
3
B
C
Fig. 5. (Cont.)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
14 Sylvain Billiard and others
proposed to explain the evolution of gamete classes, several
do not seem consistent with observations in fungi. We first
argue that one of the most widely a ccepted hypotheses, the
‘‘organelle inheritance’’ model, certainly does not hold true
for fungi. We then evaluate the other models, particularly
the ‘‘inbreeding depression avoidance’’ and ‘‘sex advantage
enhancer’’ models. Finally, we propose an evolutionary
scenario for the emergence of gamete classes in fungi.
(a) The ‘‘organelle inheritance’’ model
The ‘‘organelle inheritance’’ model is one of the most gener-
ally accepted hypotheses to explain the emergence of gamete
classes. This model predicts an association between gamete
classes and the control of inheritance of organelles in species
where plasmogamy occurs at syngamy. Most ascomycetes
are anisogamous, with inheritance of mitochondria through
the larger gamete, even if some paternal leakage of mito-
chondrial inheritance through the smaller gamete does occur
(Barr et al., 2005; Xu, 2005; Milgroom & Lipari, 1993; Lee
& Taylor, 1993). This case could therefore appear consistent
with the ‘‘organelle inheritance’’ model (Hurst & Hamilton,
1992). It is however not simply the size difference between
the gametes that allows maternal inheritance in ascomycetes
because there is an active molecular mechanism ensuring
that mitochondria are inherited from only one gamete even
when isogamy is forced (Lee & Taylor, 1993). Anisogamy
therefore does not seem necessary to ensure uniparental
inheritance and appears only to be associated with molecular
mechanisms controlling organelle transmission.
In fact, fungi show numerous cases of mitochondrial uni-
parental inheritance that are completely independent of any
gamete class (either anisogamy or mating type), such as in
yeasts and in basidiomycetes. In yeasts, syngamy is isoga-
mous and mitochondria actively segregate during the first
few rounds of cell division (Berger & Yaffe, 2000). Regard-
ing basidiomycetes, Hurst & Hamilton (1992) argued that
they did not undergo fusional sex, i.e. that there was no
cytoplasmic mixing and therefore no need to control selfish
genetic elements. This may be true for some mushrooms
in which plasmogamy is immediately followed b y recip-
rocal backwards nuclear migration (but not mitochondrial
migration) along the mycelia, but not for many other basid-
iomycete species (Barroso & Labarere, 1995; Alexopoulos,
Mims & Blackwell, 1996). In many basidiomycetes, mating
occurs between yeast-like cells and cytoplasmic fusion occurs
(Wilch, Ward & Castle, 1992). Furthermore, even when mat-
ing occurs between mycelia, there are many cases where a
filamentous heterokaryon grows after the syngamy of hap-
loid cells, containing the two parental haploid nuclei and
possibly the two parental cytoplasms, without any nuclear
migration (Hintz et al., 1988). The mode of inheritance of
cytoplasmic elements is therefore highly relevant in most
species of basidiomycetes. In several species the mating type
loci in fact control mitochondrial inheritance (Barr et al.,
2005; Wilch et al., 1992; Yan & Xu, 2003) (Fig. 5C). There
are nevertheless a number of basidiomycete species with
biparental mitochondrial inheritance, at least i mmediately
following syngamy (Barr et al., 2005; Wilch et al., 1992; Aanen
et al., 2004) (Fig. 5C). Mitochondrial heteroplasmy usually
resolves after a number of mitotic divisions in the hyphae,
more rapidly than expected by simple stochastic p rocesses,
suggesting that another active mechanism exists to regulate
cytoplasmic conflicts, here again unrelated to gamete classes
(Hintz et al., 1988; Barroso & Labarere, 1995, 1997). In cases
where a germinating spore or a heterokaryotic mycelium fer-
tilizes a haploid mycelium (Buller, 1931; Callac et al., 2006),
there also appear to be rules controlling mitochondrial inher-
itance independent of the size of the cells or their mating type.
Basidiomycetes therefore also seem to stand in contradiction
to the ‘‘organelle inheritance’’ model, because gamete classes
do not appear to be needed to regulate cytoplasmic conflicts.
(b) The ‘‘inbreeding depression avoidance’’ model
The ‘‘inbreeding d epression avoidance’’ models were
originally developed to explain the appearance of
self-incompatibility systems in angiosperms. Some investi-
gations have shown that inbreeding depression exists in fungi
(Leslie & Raju, 1985; Xu, 1995). Ascomycete fungi, however,
are haploid for most of their life cycle, which should purge
deleterious mutations and thus prevent an inbreeding depres-
sion effect. Furthermore, these models cannot fully be applied
to fungi because unifactorial, and to a lesser extent bifactorial,
haploid mating types cannot prevent diploid selfing (Giraud
et al., 2008; Fig. 2). Indeed, because each diploid genotype is
necessarily heterozygous at the MAT locus in heterothallic
species, it will produce haploid gametes with different mating
types that are thus able to mate (i.e. equivalent to diploid
selfing). In fact several fungi are known to undergo high rates
of diploid selfing while having mating types (Giraud et al.,
2008) (Fig. 2). Inbreeding depression therefore cannot be a
force acting on the origin or maintenance of mating types
determined at the haploid level, such as in fungi.
(c) The ‘‘sex advantage enhancer’’ model
The ‘‘sex advantage enhancer’’ model, where mating
between unlike gametes facilitates a variety of fitness ben-
efits afforded by sexual reproduction, seems more relevant
to fungi because many identical haploid clones of a given
gamete can in fact be produced and may not necessar-
ily disperse (Day & Garber, 1988). Mating types may well
then have evolved in fungi to prevent intra-haploid mat-
ing, i.e.
syngamy between strictly identical haploid clones,
which would have no advantage over asexual reproduc-
tion. Recombination between genetically identical haploid
yeasts indeed seems to be costly (Birdsell & Willis, 1996).
Homothallic fungi are capable of intra-haploid mating, but
they may not constitute an invalidation of this hypothe-
sis because they may still outcross most often in nature if
gametes disperse before mating. In fact, it has been shown
for some homothallic species that mating is preferentially
disassortative (i.e. between genetically different individuals);
this is called ‘‘relative heterothallism’’ (Pontecorvo et al.,
1953). Also, in some species considered as homothallic in the
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 15
laboratory the great majority of syngamy in nature occurs
between cells that are disti nct meitoic produc ts rather than
mitotic descendants of a common clone. The yeast Saccha-
romyces cerevisae for instance, capable of MAT switching in the
laboratory, undergoes mostly intra-tetrad mating in some
natural populations, i.e. syngamy between cells that are dis-
tinct meitoic products rather than mitotic descendants of
a common clone (Katz Ezov et al., 2010). However, some
homothallic fungi, such as the homothallic Neurospora species,
rely mainly on intra-haploid mating for reproduction. These
species do not produce asexual conidia and they do not seem
to outcross in nature (Glass, Jacobson & Shiu, 2000). Sexual
reproduction by intra-haploid mating could provide some
advantages. This is the case for instance when other traits,
like resistance, become linked to sexual reproduction. Sexual
spores are indeed often more resistant and can survive more
adverse conditions than asexual conidia (Aanen & Hoekstra,
2007). Sex is also associated with cell rejuvenation, avoid-
ing senescence (Sinclair, Mills & Guarente, 1998; Haedens,
Malagnac & Silar, 2005). For some species, it has been found
that slightly deleterious mutations accumulate at a lower rate
when individuals reproduce by intra-haploid mating com-
pared with asexual reproduction (Bruggeman et al., 2003;
Aanen & Hoekstra, 2007). Finally, in some ascomycetes
infected by RNA viruses, the transmission of the virus will
be vertical only via the asexual conidiospores while no trans-
mission is observed to the sexual progeny (Rogers, Buck &
Brasier, 1986; Coenen, Kevei, & Hoekstra, 1997).
(d) Other models
The assumptions or predictions of the other hypotheses
invoked to explain the origin of mating types do not seem to
hold in fungi.
The ‘‘by-product’’ model predicts that a given mating type
should produce either pheromones or receptors, but not both.
All fungal species in fact can produce both pheromones and
receptors. Furthermore, the existence of fungal species capa-
ble of ‘‘same-sex mating’’ (Fig. 5), i.e. having a single MAT
allele in their genome and able to mate with cells carrying the
same MAT allele, indicates that there is no absolute require-
ment to harbour different molecules to undergo syngamy.
The ‘‘selfish element’’ model also appears inconsistent with
symmetry in the signalling for mating compatibility in most
fungal systems, where members of all gamete classes carry
both pheromones and pheromone receptors. For the ‘‘self-
ish element’’ model, only infected cells should be expected
actively to provide signalling mechanisms that promote syn-
gamy. Also, for this model to operate, the selfish genetic
element must itself induce syngamy and be transmitted to all
daughter cells, or at least to more than half of the daugh-
ter cells. However, in all fungal mating type systems, the
transmission of mating types is Mendelian. The MAT genes
thus do not benefit from syngamy more than the rest of the
nuclear genome.
The ‘‘ploidy’’ model is consistent with the fact that one
of the current roles of mating types in fungi is to induce
changes in the physiology and gene expression patterns of
the organism (Stanton & Hull, 2007). However, this model
does not seem compatible with observations, as the numerous
homothallic fungi manage to respond developmentally in an
appropriate manner to when they are haploid or diploid, in
the absence of any differences at their mating type locus.
The ‘‘gamete size’’ model and the ‘‘anisogamy conse-
quence’’ model are difficult to test in fungi because there are
no available data on the relationship between zygote fitness
and its size and there have not been multiple transitions
between anisogamy and isogamy.
(e) An evolutionary scenario for the origin of gamete classes in fungi
and in other taxa
Support for the ‘‘organelle inheritance’’ model was claimed
from various taxa: in fact, a correlation between uniparental
inheritance of organelles and anisogamy, or with mating
types in isogamous species, seems to exist (Hurst & Hamilton,
1992). However, such a correlation is not indicative of
cause and effect, and an alternative hypothesis is that the
mechanisms controlling organelle inheritance evolved by
being superimposed upon pre-existing gamete classes that
evolved for other reasons. This evolutionary sequence was
suggested by Maynard Smith & Szathmary (1995, p. 162).
In fact, there are several examples in fungi where mito-
chondrial inheritance is uniparental although apparently not
controlled by gamete classes, either anisogamy or mating
type. The hypothesis that gamete classes pre-date contr ol
over organelle inheritance is also consistent with paternal
inheritance or biparental inheritance of mitochondria and
chloroplasts in gymnosperms (Reboud & Zeyl, 1994) and
some chytrid fungi (Borkhardt & Olson, 1983), because then
it is not the gamete class (in this case, gamete size) that forces
uniparental inheritance. In the myxomycete Didymium iridis,
uniparental inheritance is frequent while not being associ-
ated with any known gamete class (Silliker, Liles & Monroe,
2002; Scheer & Silliker, 2006).
Under such an evolutionary sequence, it is still impor-
tant to explain the emergence of gamete classes before the
evolution of uniparental inheritance of organelles. We argue
that two evolutionary models appe ar the best candidates for
explaining the emergence of mating types: the ‘‘inbreeding
depression avoidance’’ model and ‘‘sex advantage enhancer’’
model. Which of these two models is the most relevant may
depend on the organism being considered and whether mat-
ing compatibility is determined at the haploid or diploid
level in particular. The ‘‘inbreeding depression avoidance’’
model has indeed received much attention in the literature
as an explanation for the e volution of self-incompatibility in
angiosperms, but it may not be generally applicable as the
haploid determinism of mating compatibility in fungi does
not prevent diploid selfing. Indeed, deleterious mutations can
still be exposed to selection by selfing in heterothallic fungal
species: the requirement for mating between haploid cells
carrying different mating types d oes not prevent inbreed-
ing depression. The ‘‘sex advantage enhancer’’ model has
received much less attention, but we expect that it could
be of fundamental importance for organisms like fungi or
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
16 Sylvain Billiard and others
some algae having a mating type determined at the haploid
stage. For such organisms, mating types would be needed to
benefit from recombination, by preventing syngamy between
identical haploids. In organisms such as plants or animals,
intra-haploid mating is i mpossible, rendering the ‘‘sex advan-
tage enhancer’’ model irrelevant to explain the maintenance
of mating types. For such organisms, the ‘‘inbreeding depres-
sion avoidance’’ model appears the most valid. Regarding
the evolution of anisogamy, we suggest that the best can-
didate would be the ‘‘gamete size’’ model: large and small
gametes evolved by disruptive selection on gamete size.
(2) The optimal number of gamete classes in fungi
Beyond the evolutionary forces responsible for the emer-
gence of gamete classes, it is interesting to investigate those
acting on their numbers, and in particular why some species
retain only two gamete classes even though i t appears apriori
to be the least favourable case. The existence of only two
gamete sizes is understandable as it has been shown that
selection will not then favour the evolution of a third class of
intermediate size. The most puzzling problem is therefore the
abundance of species with only two mating types. It is striking
on Fig. 5 that almost all hemiascomycetes and ascomycetes
have only two mating types or are homothallic (except Glom-
erella cingulata,Fig.5A,B),whilethenumberofmatingtypes
varies from two to thousands in basidiomycetes, with fre-
quent transitions b etween two and many (Fig. 5C). Fungi
thus are highly suitable to investigate the factors responsible
for variability in mating type number.
Frequency-dependent selection should act to increase
the number of mating types because any new rare allele
could potentially mate with any other gamete in the
population. Several of the evolutionary forces that may
explain the emergence of gamete classes would also limit their
number to two, in particular the ‘‘by-product’’ model, the
‘‘selfish element’’ model, the ‘‘gamete size’’ model, and the
‘‘anisogamy consequence’’ model. However, as we have seen
above, these models are probably not the best explanations
for the emergence of mating types in fungi. Under the
‘‘organelle inheritance’’ model, it also seems unlikely that
more than two gamete classes can evolve. However,
this model cannot explain the maintenance of only two
mating types in ascomycetes, hemiascomycetes, and some
basidiomycetes because inheritance of mitochondria in these
groups is not controlled by mating type. Hence, it appears
that the factors limiting the number of mating types in many
fungi to only two are not generally also the evolutionar y forces
responsible for their origin. It may nevertheless be the case
that once uniparental inheritance is associated with mating
type, as in some basidiomycetes (Fig. 5), the emergence of a
third mating type would be selected against (Hurst, 1996).
How can we explain the prevalence of only two mat-
ing types in hemiascomycetes and ascomycetes whereas
their number varies between two and thousands in basid-
iomycetes? The main difference between ascomycetes a nd
basidiomycetes lies in the genetic determinism of mating
types. In basidiomycetes, it can be imagined that different
specificities can easily evolve because haploid cells need only
to carry different alleles to be compatible with extant mating
types (Kohte, Gola & Wendland, 2003). In ascomycetes,
mating types are idiomorphs that encode transcription fac-
tors inducing pheromone and receptor gene expression.
The emergence of a third mating type in ascomycetes
depends on the mechanism of recognition at syngamy. If
syngamy does not occur unless a specific pheromone has
activated its specific corresponding receptor, a third mating
type cannot evolve, because mutants will not b e able to
mate with any extant mating type (Hoekstra, 1987). The
maintenance of only two mating types in ascomycetes may
thus only be due to proximal constraints, although more
experimental work is required. There is a single example
in ascomycetes (G. cingulata) where they have overcome this
constraint (Fig. 5A), although the underlying molecular basis
is not known.
Within basidiomycetes, the number of mating types is
highly variable, ranging from two to thousands (Fig. 5C).
Frequency-dependent selection is likely to be responsible for
increasing the number of mating types, but several forces
could act to prevent an increase. The genetic basis of mating
type determinism is relevant: for a given number of alleles at
each of the two loci involved in mating type determination,
unifactorial systems will generate less mating type pheno-
types i n the population than bifactorial systems, which have
recombination between the loci. Some unifactorial species
however have more than a dozen mating types and link-
age between the two MAT loci has evolved several times
independently (Fig. 5C), suggesting that genetic determinism
(i.e. unifactorial versus bifactorial) does not represent a strong
constraint, and may even be selected for promoting a high
or a low number of mating types.
Restriction of the number of mating types to only two in
some basidiomycetes may be due to mating systems that are
highly selfing, where the existence of mating types prevents
fusion between identical haploids but a third mating type
would have little advantage because each diploid individual
is heterozygous for mating type and can self (Giraud et al.,
2008). Unfortunately, data to link the number of mating types
with the selfing rate in natural populations of basidiomycetes
are still too scarce to test this hypothesis. In basidiomycetes,
hyphae can grow short distances to contact mates and thus
some mate-finding effort i s possible which could limit the
advantage of rare mating types. However, data are not
available on the costs associated with waiting for a mate.
Finally, r egions of suppressed recombination are known
to exist around mating type loci in several fungal species
(Fraser & Heitman, 2005; Hood, 2002) (Fig. 5). Deleterious
alleles linked to mating type have been reported, particularly
in species with only two mating types (Oudemans et al.,
1998; Thomas et al., 2003; Callac et al., 2006). Phylogenies
of mating type alleles in the basidiomycete Coprinus cinereus
have longer terminal branches than expected (May et al.,
1999), which could b e a signature of a decreasing rate of
mating type appearance because of deleterious mutations
associated with the mating type loci. Such sheltered load
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 17
may be a major force limiting the number of mating types
(Uyenoyama, 1997), perhaps even to two. Several forces may
therefore act on the number of mating types in fungi, but
more data are needed.
(3) The loss and maintenance of gamete classes
in fungi
The loss of gamete classes is evolution leading to a lack
of discrimination for syngamy and this should be favoured
when there is a low probability or a cost to waiting to find
a compatible mate. Fungi appear to be unusual in that mul-
tiple i ndependent losses of gamete classes have occurred.
Same-sex mating in fungi ind eed conforms to our definition
of loss of gamete class (Figs 4 and 5). The most frequent form
of homothallism, resulting from the presence of two inter-
compatible MAT idiomorphs in the haploid genome is one
mechanism by which this is achieved (Figs 4 and 5). How-
ever it is not known whether differential temporal expression
occurs in these cases. By c ontrast, hemiascomycete species
that exhibit mating type switching are not compatible with all
other cells in the population but only with those expressing
the alternative mating type. The phylogenies of ascomycetes
and basidiomycetes show a large number of independent
origins of homothallism (Fig. 5), and rare reverse transi-
tions from homothallism to heterothallism have also been
proposed in ascomycetes (Galagan et al., 2005).
Because of the potential advantages of universal compati-
bility during mating, homothallism can easily be selected for
in fungi (Nauta & Hoekstra, 1992). In homothallic species,
there is no restriction at syngamy except anisogamy, which is
not genetically determined to distinguish haploid genotypes:
every haploid has the potential to mate with any other hap-
loid in the population, including itself. Homothallism is often
presented as an adaptation in fungi to allow a haploid-selfing
mating, but the reverse interpretation is also interesting
to consider: homothallism may have evolved in outcross-
ing species to allow universal compatibility among gametes
(Giraud et al., 2008). An experimental study has shown
that, under experimental conditions mimicking natural con-
ditions, pseudo-homothallic spores of the basidiomycete
Agaricus bisporus do not self but instead fertilize homokaryons
(Callac et al., 2006). In fact, little evidence of pseudoho-
mothallic reproduction was found in a spatial-temporal anal-
ysis of natural populations (Xu, Desmerger, & Callac, 2002).
These results are consistent with homothallism or pseudo-
homothallism having evolved to achieve universal compati-
bility and not to allow selfing or intra-haploid mating. O ther
observations however are more consistent with homothal-
lism having evolved, or at least being maintained, to allow
intra-haploid mating and to benefit from the advantages of
sex not related to recombination. For instance, homothal-
lic Neurospora species do not produce any trichogynes or
conidia (Glass & Kuldau, 1992), i.e. the male and female
organs allowing cross-fertilization. Mating between different
mycelia is therefore prevented; intra-haploid mating is the
only possible sexual reproduction in these species, and occurs
regularly.
Again, when comparing the phylogenies of basidiomycetes
and ascomycetes, it is striking that the evolution of homoth-
allism differs between the two groups (Fig. 5). In basid-
iomycetes, all species are heterothallic with a few exceptions,
while in ascomycetes homothallism evolved several times
independently, different proximal causes likely being respon-
sible (linkage of idiomorphs, same-sex mating, and mating
type switching). Reverse transition from homothallism to het-
erothallism has also been proposed in ascomycetes (Galagan
et al., 2005). In fact, it seems that ascomycetes and b asid-
iomycetes use different strategies to increase the probability
of compatibility at syngamy: many basidiomycetes have high
numbers of mating types, facilitating the finding of compat-
ible mates but preventing intra-haploid mating. By contrast,
in ascomycetes molecular constraints linked to the molec-
ular mechanisms of gamete recognition may mean that
homothallism is the only way to achieve universal compat-
ibility. Homothallism however allows intra-haploid mating,
precluding many of the advantages of sex. If gametes are
dispersed and do not remain clumped together, this may not
be a significant problem. Evolution towards homothallism
or the maintenance of heterothallism may therefore depend
on gamete dispersal in ascomycetes. Unfortunately, no data
are available to test this suggestion. Inbreeding depression
could also play a role in the persistence of mating type
systems in fungi. Ascomycetes are haploid for most of their
life cycle while basidiomycetes are dikaryotic, which could
allow the accumulation of more deleterious alleles in the lat-
ter, preventing the emergence of homothallism in the latter
group. Again, however, not enough data are available yet to
disentangle these hypotheses.
V. CONCLUSIONS
(1) Hurst & Hamilton (1992) defined gamete classes
based on their evolutionary origin: ‘‘sexes’’ when
gamete classes determine the uniparental inheritance
of organelles, and ‘‘mating types’’ when gamete classes
allow the avoidance of i nbreeding depression. We
suggest that these definitions are not adequate for
several reasons. First, there remain doubts about the
sequence of events leading to the evolution of gamete
classes and uniparental inheritance. Second, at least in
some organisms, uniparental inheritance of organelles
is not associated with a particular gamete class. Hence,
even when uniparental inheritance is observed, it is
not possible to define different sexes on the basis of
this observation alone. Third, in some taxa it can be
difficult to determine whether organelle inheritance
is uniparental, and whether it is associated with a
gamete class, such that this definition of ‘‘sex’’ is not
appropriate. Finally, the definition of ‘‘mating types’’
as a means for avoiding inbreeding depression can
neglect other evolutionary forces acting at the origin
of gamete classes, for instance in the ‘‘sex advantage
enhancer’’ model. In conclusion, it is not appropriate
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
18 Sylvain Billiard and others
or useful to define ‘‘sexes’’ and ‘‘mating types’’ on the
basis of the evolutionary causes at their origin. Instead,
we propose the definition of ‘‘sexes’’ as gamete classes
based on morphological differences, such as size, and
‘‘mating types’’ as cases where molecular mechanisms
restrain gamete compatibilities independent of size
dimorphism. These definitions are more relevant to
empiricists and are in more general agreement with
how these terms are used across other branches of
biology. Moreover, such a definition for ‘‘sexes’’ makes
simple the relationships between gamete classes and
gender evolution.
(2) Regarding the forces at the origin of gamete classes,
we have attempted to show that Fungi represents a
unique kingdom to test various hypotheses. We con-
clude that the ‘‘organelle inheritance’’ model may
not be as general as previously considered to explain
either the emergence of mating types or anisogamy.
Uniparental inheritance of organelles may have acted
on pre-existing gamete classes. The most plausible
explanations for the evolution of mating types are the
‘‘inbreeding depression avoidance’’ model and ‘‘sex
advantage enhancer’’ model, and the most acceptable
model for the evolution of anisogamy is the ‘‘gamete
size’’ model.
(3) Regarding the forces responsible for the number of
mating types, we argue that Fungi also allow us to dis-
entangle various hypotheses. High numbers of mating
types may have evolved in basidiomycetes in response
to selection for increased compatibility at syngamy.
However, some basidiomycete species retain two mat-
ing types, possibly due to constraints applied by their
mating system, genetic load around MAT loci, or
organelle inheritance being controlled by mating type.
Ascomycetesmaynotbeabletoevolvemultiplemating
types due to proximal constraints, but have increased
their compatibility at syngamy by homothallism (i.e.
reversion to zero functional mating types). Heterothal-
lism is maintained in other species, probably by the
same forces that promoted the emergence of mating
types, in particular the prevention of intra-haploid
mating.
(4) Further work in fungi clearly are needed, for example
on selfing rates in natural populations, the natural fre-
quency of intra-haploid matings in homothallic species,
the cost of waiting for a mate, the dispersal range of
gametes, the existence and strength of inbreeding
depression, and the shape of functions linking zygote
size and fitness.
VI. ACKNOWLEDGEMENTS
We thank Robert Debuchy, Philippe Silar, Jacqui Shykoff,
Xavier Vekemans, Phillipe Callac, Pascal Frey, Laurence
Hurst, Thomas Lenormand and all the scientists who par-
ticipated in the workshop funded by the GDR ComEvol
for useful and stimulating discussions. We acknowledge
grants ANR-09-0064-01 and NSF-DEB 0747222. S.B.
acknowledges post-doctoral g rants from the CNRS. We
apologize to all those colleagues whose work we may have
failed to cite in this article.
VII. REFERENCES
References preceded by *, ** and *** were used to construct Fig. 5A, B,
and C, respectively, and may not have been cited in the main text. See
Section VIII for further details.
Aanen,D.K. & Hoekstra, R. F. (2007). Why sex is good: on fungi and beyond.
In Sex in fungi: molecular determination and evolutionary implications (ed. J. Heitman,
J. W. Kronstad,J.W.Taylor and L. A. Casselton) Washington, DC: ASM
Press. pp. 527534.
Aanen,D.K., Kuyper,T.W., Debets,A.J.M. & Hoekstra, R. F. (2004). The
evolution of non-reciprocal nuclear exchange in mushrooms as a consequence of
genomic conflict. Proceedings of the Royal Society of London B 271, 12351241.
*Aguileta,G., Marthey,S., Chiapello,H., Lebrun,M.-H., Rodolphe,F.,
Fournier,E.,Gendrault-Jacquemard,A.&Giraud, T. (2008). Assessing the
performance of single-copy genes for recovering robust phylogenies. Systematic Biology
57, 613627.
Agrawal, A. F. (2006). Similarity selection & the evolution of sex: Revisiting the red
queen. PLoS Biology 4, e265.
***Aimi,T., Yoshida,R., Ishikawa,M., Dapeng,B. & Kitamoto, Y. (2005).
Identification and linkage mapping of the genes for the putative homeodomain
protein (hox1) and the putative pheromone receptor protein homologue (rcb1) in a
bipolar basidiomycete, Pholiota nameko. Current Genetics 48, 184194.
Alby,K., Schaefer,D. & Bennet, R. J. (2009). Homothallic and heterothallic
mating in the opportunistic p athogen Candida albicans. Nature 460, 890–3.
Alexopoulos,C.J., Mims,C.W. & Blackwell
, M. (1996). Introductory Mycology.
John Wiley & Sons, Inc.
*Arzanlou,M., Crous,P. & Zwiers, L. H. (2010). Evolutionary dynamics of
mating-type loci of Mycosphaerella spp. occurring on banana. Eukaryotic Cell 9, 164172.
***Bakkeren,G., Jiang,G., Warren,R.L., Butterfield,Y., Shin,H.,
Chiu,R., Linning,R., Schein,J., Lee,N., Hu,G., Kupfer,D.M., Tang,Y.,
Roe,B.A.,Jones,S.,Marra,M.&Kronstad, J. W. (2006). Mating fa ctor link-
age and genome evolution in basidiomycetous pathogens of cereals. Fungal Genetics
and Biology 43, 655 666.
***Bakkeren,G.&Kronstad, J. W. (1994). Linkage of mating type loci distinguishes
bipolar from tetrapolar mating in basidiomycetous smut fungus. Proceedings of the
National Academy of Sciences U.S.A. 91, 7085 7089.
***Banuett, F. (1995). Genetics of Ustilago maydis, a fungal pathogen that induces
tumors in maize. Annual review of genetics 29, 179208.
Barr,C., Neiman,M. & Taylor, D. (2005). Inherita nce and recombination of
mitochondrial genomes in plants, fungi and animals. New Phytologist 168, 3950.
Barroso,G. & Labarere, J. (1995). Molecular organization and heredity of the
mitochondrial genome in Basidiomycetes. Mycoscience 36, 135142.
***Barroso,G. & Labarere, J. (1997). Genetic evidence for nonrandom sorting
of mitochondria in the basidiomycete Agrocybe aegerita. Applied and Environmental
Microbiology 63, 4686 4691.
***Begerow,D., Stoll,M. & Bauer, R. (2006). A phylogenetic hypothesis of
Ustilaginomycotina based on multiple gene analyses and morphological data.
Mycologia 98, 906916.
Bell, G. (1978). The evolution of anisogamy. Journal of Theoretical Biology 73, 247270.
Bell, G. (1993). The sexual nature of the eukaryote genome. Journal of Heredity 84,
351359.
Berger,K.H.&Yaf fe, M. P. (2000). Mitochondria l DNA inheritance in Saccharomyces
cerevisiae. Trends in Microbiology 8, 508513.
Birdsell,J. & Willis, C . (1996). Significant competitive advantage conferred by
meiosis and syngamy in the yeast Saccharomyces cerevisiae. Proceedings of the National
Academy of Sciences U.S.A. 93, 908912.
Bonsall, M. B. (2006). The evolution of anisogamy: The adaptive significance of
damage, repair and mortality. Journal of Theoretical Biology 238, 198210.
Borkhardt,B.&Olson, L. W. (1983). Paternal inheritance of the mitochondrial-
DNA in interspecific crosses of the aquatic fungus Allomyces. Current Genetics 7,
403404.
Bruckart, W. L. III, Eskandari,F.M.,Berner,D.K.&Aime, M. C. (2010). Life
cycle of Puccinia acroptili on Rhaponticum (
= Acroptilon) repens. Mycologia 102, 6268.
Bruggeman,J., Debets,A.J.M., Wijngaarden,P.J., deVisser,J.A.G.M. &
Hoekstra, R. F. (2003). Sex slows down the accumulation of deleterious mutations
in the homothallic fungu s Aspergillus nidulans. Genetics 164, 479485.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Biological Reviews (2010) 000 000 © 2010 The Authors. Biological Reviews © 2010 Cambridge Philosophical Society
Having sex, yes, but with whom? Inferences from fungi on the evolution of anisogamy and mating types 19
Buller, A. H. R. (1931). Researches on fungi, vol. IV. Longmans, Green & Co., London,
United Kingdom.
Bulmer, M. (1994). Theoretical Evolutionary Ecology. Sunderland, MA: Sinauer.
Bulmer,M.G. & Parker, G. A. ( 2002). The evolution of anisogamy: a game
theoretic approach. Proceedings of the Royal Society of London B 269, 23812388.
*Burt,A., Carter,D.A., Koenig,G.L., White,T.J. & Taylor, J. W. (1996).
Molecular markers reveal cryptic sex in the human pathogen Coccidioides immitis.
Proceedings of the National Academy of Sciences of the United States of America 93, 770 773.
Butler, G. (2007). Evolution of the mating-type locus: the ascomycetes. In Sex in fungi:
molecular determination and evolutionary implications (ed. J. Heitman,J.W.Kronstad,
J. W. Taylor and L. A. Casselton) Washington, DC: ASM Press. pp. 318.
**Butler,G., Kenny,C., Fagan,A., Kurischko,C., Gaillardin,C. &
Wolfe, K. H. (2004). Evolution of the MAT locusanditsHoendonucleasein
yeast species. Proceedings of the National Academy of Sciences of the United States of America
101, 16327.
**Butler,G., Rasmussen,M.D., Lin,M.F., Santos,M.A.,Sakthikumar,S.,
Munro,C.A., Rheinbay,E., Grabherr,M., Forche,A., Reedy,J.L.,
Agrafioti,I.,Arnaud,M.B.,Bates,S.,Brown,A.J.,Brunke
,S.,Costanzo,
M. C., Fitzpatrick,D.A., de Groot,P.W., Harris,D., Hoyer,L.L.,
Hube,B., Klis,F.M., Kodira,C., Lennard,N., Logue,M.E., Martin,R.,
Neiman,A.M., Nikolaou,E., Quail,M.A., Quinn,J., Santos,M.C.,
Schmitzberger,F.F.,Sherlock,G.,Shah,P.,Silverstein,K.A.,Skrzypek,
M. S., Soll,D., Staggs,R., Stansfield,I., Stumpf,M.P., Sudbery ,P.E.,
Srikantha,T.,Zeng,Q.,Berman,J.,Berriman,M.,Heitman,J.,Gow,N.A.,
Lorenz,M.C.,Birren,B.W.,Kellis,M.&Cuomo, C. A. (2009). Evolution of
pathogenicity a nd sexual reproduction in eight Candida genomes. Nature 459,