SEARCH

SEARCH BY CITATION

Keywords:

  • post-zygotic isolation;
  • prezygotic isolation;
  • secondary contact;
  • specialization;
  • sympatric speciation;
  • tempo of speciation;
  • time course of speciation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Generally, stronger reproductive isolation is expected between sympatric than between allopatric sibling species. Such reproductive character displacement should predominantly affect premating reproductive isolation and can be due to several mechanisms, including population extinction, fusion of insufficiently isolated incipient species and reinforcement of reproductive isolation in response to low hybrid fitness. Experimental data on several taxa have confirmed these theoretical expectations on reproductive character displacement, but they are restricted to animals and a few plants. Using results reported in the literature on crossing experiments in fungi, we compared the degree and the nature of reproductive isolation between allopatric and sympatric species pairs. In accordance with theoretical expectations, we found a pattern of enhanced premating isolation among sympatric sibling species in Homobasidiomycota. By contrast, we did not find evidence for reproductive character displacement in Ascomycota at similar genetic distances. Both allopatric and sympatric species of Ascomycota had similarly low levels of reproductive isolation, being mostly post-zygotic. This suggests that some phylogeny-dependent life-history trait may strongly influence the evolution of reproductive isolation between closely related species. A significant correlation was found between degree of reproductive isolation and genetic divergence among allopatric species of Homobasidiomycota, but not among sympatric ones or among Ascomycota species.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Different types of reproductive isolation among closely related species may evolve depending on their geographical distribution. Incipient species evolving in different geographic areas (allopatry) have no opportunities to mate with each other; so, reproductive isolation is expected to arise gradually and slowly as a result of independent mutation, genetic drift and indirect effects of natural selection driving local adaptation (Coyne & Orr, 2004, pp. 83–110). By contrast, models of speciation in sympatry usually involve the rapid selection of reproductive isolation (Coyne & Orr, 2004, pp. 125–178). When incipient species that have diverged in allopatry secondarily come into sympatry, several outcomes are possible: one of the two may go extinct, they may fuse if reproductive isolation is weak enough, or there may be a rapid evolution of premating isolation by selection for avoiding interspecific crosses, if the progeny is unfit because of post-zygotic isolation (Kirkpatrick & Ravigné, 2002). Altogether, stronger reproductive isolation is therefore expected between sympatric than allopatric sibling species. Such reproductive character displacement is expected to mainly affect premating (or prezygotic) reproductive isolation (but see Coyne & Orr, 2004, pp. 365–366) and has extensively been studied in the theoretical literature dealing with reinforcement (Noor, 1999; Servedio & Noor, 2003).

Enhanced premating reproductive isolation in sympatry compared with allopatry has been detected in many natural cases, for instance among Drosophila, damselflies, frogs, fish, crickets, toads, birds, marine organisms and rodents (Coyne & Orr, 2004, pp. 357–360; Lukhtanov et al., 2005; Smadja & Ganem, 2005; and references therein). Experimental data on several taxa have therefore confirmed the theoretical expectations on reproductive character displacement, but are almost exclusively restricted to animals and a few plants (Armbruster et al., 1994; Coyne & Orr, 2004). Further, several studies have failed to detect enhanced isolation in Drosophila, toads and grasshoppers (Coyne & Orr, 2004, pp. 361–362). These inconsistencies may reflect the rarity of reinforcement in some taxa or may only be the consequence of looking at sister species with different histories, for instance with different evolutionary ages, or that have been in contact for different periods of time. Coyne & Orr (2004) therefore advocated that we need comparative studies to test for the frequency of reproductive character displacement, controlling for genetic divergence.

Fungi are interesting models for the study of evolution of reproductive isolation (Burnett, 2003). Many can be cultured and crossed under laboratory conditions and mycologists have long reported numerous mating experiments among fungal species. Second, fungi display a huge variety of life cycles, potentially allowing exploration of the influence of parameters such as dispersal ability on the speciation process. Third, numerous species complexes are known in fungi, encompassing multiple recently diverged sibling species. In the present study, we used previously published data on crossing experiments within fungal species complexes to investigate reproductive isolation patterns among close species in the two main groups of true fungi, Ascomycota and Basidiomycota. Our aims were to assess whether a pattern of reproductive character displacement existed in Fungi and whether reproductive isolation increased with genetic distance, among allopatric and/or sympatric species.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Data set

We searched the bibliographic data bases Web of Knowledge (http://isiknowledge.com/) and Pubmed (http://www.ncbi.nlm.nih.gov/) from the last 20 years for all papers with ‘hybrid* and fung*’, ‘mating and fung*’, ‘cross* and fung*’, ‘biological species and fung*’ and ‘reproductive isolation and fung*’ in the title, keyword or abstract. Older papers were found in the citation lists of these articles.

We only considered reproductive isolation within species complexes, i.e. between sibling species nearly indistinguishable morphologically, often grouped under a single species name, but for which evidence of genetic isolation had been reported, such as molecular evidence of restricted gene flow and/or some level of reproductive isolation under laboratory conditions. Molecular evidence of independent evolution among sibling species were mainly from sequence data showing the existence of divergent monophyletic clades with bootstrap supports higher than 70, but also from isozymes, RFLP, AFLP and DNA/DNA hybridization data showing lack of gene flow (see references in Table 1).

Table 1.   Species complexes.
Species complexLife style and hosts for parasitesSympatryAllopatryITS accession numbersReference
Number of species pairs Mean IR ± SE Genetic distance ± SENumber of species pairs Mean IR ± SE Genetic distance ± SE
  1. For each species complex used in the study, lifestyle and hosts for parasites, number of species pairs used in this study, mean degree of reproductive isolation detected in vitro and mean genetic distance among sympatric and allopatric species pairs, respectively, and references (A, Ascomycota, B, Basidiomycota).

(A) Ascomycota
Ascochyta sp.Parasite (legumes)40.56 ± 0.260.014 ± 0.006   DQ383950, 52-54Kaiser et al. (1997), Hernandez-Bello et al. (2006), Peever et al. (2007)
Ascobolus sp.Saprophyte11.00 20.88 ± 0.13 NoMeinhardt et al. (1984)
Botrytis cinereaParasite (fruits, legumes, flowers)10.750.00   E. Fournier pers. comm.Fournier et al. (2005), E. Fournier and P. Leroux, pers. comm.
Ceratocystis sp.Parasite (trees)150.32 ± 0.080.023 ± 0.00340.25 ± 0.180.025 ± 0.005U75618, 20, 26, AY214001, AY233907, 21, 25Harrington & McNew (1998), Harrington et al. (1996)
Cochliobolus sp.Parasite (grasses and cereals)90.81 ± 0.040.015 ± 0.005   AF071332, AF158105, 9, 10, AF163074, AB179836Berbee et al. (1999), Nelson (1960)
Cryphonectria sp.Parasite (chestnut trees)11.000.087   AY141873, AY143076Hoegger et al. (2002)
Epichloë sp.Parasite (grasses)130.50 ± 0.000.027 ± 0.004110.50 ± 0.000.023 ± 0.003L07131, 36, 38, U57665, X62987, AB105953Moon et al. (2004), Schardl et al. (1997), Schardl & Leuchtmann (1999)
Erysiphe graminisParasite (grasses)60.50 ± 0.110.022 ± 0.006   AJ313137-40Hiura (1978), Wyand & Brown (2003)
Giberella fujikuroiParasite (cereals)131.00 ± 0.000.032 ± 0.00721.00 ± 0.000.029 ± 0.021U34555-60, U34568Leslie (1991, 1995), O’Donnell et al. (1998)
Histoplasma capsatulatumParasite (humans, animals)20.63 ± 0.130.028 ± 0.00610.500.02AF322377, 78, 86, 87Kasuga et al. (1999), Taylor et al. (1999)
Neurospora sp.Saprophyte90.50 ± 0.040.00210.50 AY681192-93Dettman et al. (2003)
Ophiostoma piceaParasite (trees)31.00 ± 0.000.035 ± 0.00821.00 ± 0.00 AY194505, AY573250, AY618239Brasier & Kirk (1993), Uzunovic et al. (2000)
Ophiostoma ulmi, O. novo-ulmi and O. himal-ulmiParasite (trees)10.500.00720.50 ± 0.000.016 ± 0.00AF198232, 34, 36Brasier (2001), Brasier & Mehrotra (1995)
Phomopsis sp.Saprophyte11.00    NoBrayford (1990)
Saccharomyces sp.Saprophyte130.31 ± 0.060.017 ± 0.00220.25 ± 0.000.018 ± 0.002AY130306-09, AY046148, EF457568Naumov (1996), Liti et al. (2006), Murphy et al. (2006)
Venturia inaequalisParasite (Apple trees, Pyracantha)10.500.018   B. Le Cam, pers. comm.Le Cam et al. (2002)
(B) Basidiomycota
Amylostereum sp.Saprophyte/parasite (trees)31.00 ± 0.000.033 ± 0.00320.53 ± 0.080.016 ± 0.00DQ383950, 52-54Boidin & Lanquetin (1984)
Armillaria sp.Saprophyte/parasite (trees)381.00 ± 0.000.028 ± 0.006400.96 ± 0.020.032 ± 0.006NoAnderson et al. (1989)
Collybia dryophilaSaprophyte121.00 ± 0.00 330.76 ± 0.08 E. Fournier pers. comm.Vilgalys & Johnson (1987)
Flammulina sp.Saprophyte/parasite (trees)90.91 ± 0.080.031 ± 0.00360.91 ± 0.090.031 ± 0.004U75618, 20, 26, AY214001, AY233907, 21, 25Hughes et al. (1999), Petersen et al. (1999)
Fomes pinicolaSaprophyte/parasite (trees)11.00 20.75 ± 0.00 AF071332, AF158105, 9, 10, AF163074, AB179836Mounce & Macrae (1938)
Hebeloma sp.Ectomycorrhizal531.00 ± 0.000.010 ± 0.002   AY141873, AY143076Aanen & Kuyper (1999), Aanen et al. (2000)
Heterobasidion annosumParasite (trees)30.92 ± 0.040.017 ± 0.00160.66 ± 0.130.014 ± 0.002L07131, 36, 38, U57665, X62987, AB105953Harrington et al. (1989), Johannesson & Stenlid (2003), Stenlid & Karlsson (1991)
Hyphoderma setigerumSaprophyte81.00 ± 0.000.136 ± 0.00961.00 ± 0.000.110 ± 0.020AJ313137-40Nilsson et al. (2003)
Laccaria laccataEctomycorrhizal41.00 ± 0.00 61.00 ± 0.00 U34555-60, U34568Mueller (1991)
Lentinula sp.Saprophyte31.00 ± 0.000.11660.50 ± 0.220.098AF322377, 78, 86, 87Mata et al. (2001)
Microbotryum violaceumParasite (Caryophyllacea)260 ± 0.00NA   AY681192-93Le Gac et al. (2007a, b)
Marasmius androsaceusSaprophyte11.00 ± 0.00 20.50 ± 0.25 AY194505, AY573250, AY618239Gordon & Petersen (1997)
Peniophora sp.Saprophyte11.00 ± 0.00 20.45 ± 0.23 AF198232, 34, 36Chamuris (1991)
Pleurotus sp.Saprophyte231.00 ± 0.000.124120.42 ± 0.150.059NoVilgalys & Sun (1994)
Polyporus sp.Saprophyte/parasite (trees)101.00 ± 0.000.124 ± 0.01390.94 ± 0.040.110 ± 0.020AY130306-09, AY046148, EF457568Hoffmann (1978), Macrae (1967), McKay (1962),
Serpula himantioidesSaprophyte11.000.01511.000.009 ± 0.00B. Le Cam, pers. comm.Kauserud et al. (2006)
Sistotrema sp.Ectomycorrhizal71.00 ± 0.00 41.00 ± 0.00 NoUllrich (1973)

We aimed at comparing patterns of in vitro reproductive isolation between allopatric and sympatric species pairs. It can, however, be difficult in some cases to assess whether distributions are allopatric or sympatric. To be conservative regarding the detection of reproductive character displacement, we chose to consider species pairs as sympatric only when: (1) it was clearly indicated in at least one paper; or (2) the geographic distribution of the individuals sampled strongly suggested overlapping distribution areas (e.g. individuals of the two species sampled in the same city, county, country or mountain).

Fungal life cycles and experimental assessment of pre- versus post-mating isolation

Fungi having very diverse life cycles (Alexopoulos et al., 1996), the description below is obviously an oversimplification of reality, but is given to introduce the terms used in the fungal literature and includes only the main features of fungal life cycles that are relevant for the nature of reproductive isolation.

In Ascomycota, sexual reproduction occurs between two strains of opposite haploid mating types (spores or mycelium). Plasmogamy occurs and the dikaryotic stage can be maintained during a few cellular divisions. Karyogamy then takes place, followed by meiosis, leading to asci formation containing haploid ascospores that can disperse. In most pathogenic Ascomycota, haploid ascospores germinate on the host and the resulting mycelium grows within or on its host, where sexual reproduction takes place (Alexopoulos et al., 1996).

Regarding Homobasidiomycota, a haploid mycelium grows into the ground or a substrate and when two compatible mycelia come into contact, they fuse and produce clamp connections. A dikaryotic mycelium is then formed. The dikaryotic stage can persist for a long time. Fruiting bodies (basidia) are typically produced in basidiocarps (i.e. mushrooms) and yield basidiospores that can passively disperse, germinate and give rise to haploid mycelia.

The life cycle of the parasitic Microbotryum violaceum, the single Basidiomycota species complex not belonging to Homobasidiomycota in our study, can be considered as more similar to Ascomycota, with sexual reproduction occurring on the host just after meiosis and with plasmogamy occurring between single cells (Le Gac et al., 2007a).

For Basidiomycota, the degree of reproductive isolation was calculated as 1 − p, where p was the proportion of crosses showing evidence of plasmogamy initiation, i.e. clamp connections in Homobasidiomycota and conjugations in M. violaceum. This represents a quantitative measure of the degree of premating reproductive isolation, very similar to the measure of prezygotic isolation used by Coyne & Orr (1989). As basidiocarps (i.e. fruits) are very difficult to obtain in vitro, data were not available for calculating levels of post-mating reproductive isolation in Homobasidiomycota.

Quantitative results on the success of crosses are usually not reported in Ascomycota. For in vitro crosses, male and female gametes are usually mixed in Petri dishes, where the formation of asci and ascospores is observed. Data in publications report the formation, viability and fertility of ascospores, without quantitative information. We thus considered reproductive isolation in Ascomycota as a discrete measure where 1, 0.75, 0.50, 0.25 and 0 indicate, respectively, lack of ascus formation, asci without ascospores, abnormal ascospores, viable but sterile ascospores and fertile ascospores. Premating reproductive isolation is thus indicated by an RI = 1. The other categories (0, 0.25, 0.50 and 0.75) refer to different levels of post-mating isolation. As the measure of post-zygotic isolation in Coyne & Orr (1989), the classes of reproductive isolation we scored in Ascomycota are discrete. However, our index corresponds to the sequential developmental stages of the post-mating isolation rather than to the possibility of obtaining viable and fertile hybrids of both sexes with the two possible reciprocal crosses, as was carried out by Coyne & Orr (1989). The distinction between sexes in the success of interspecific crosses is indeed not commonly made in fungi.

Most of the data obtained from the literature are best suited for comparing patterns of reproductive isolation at the species level (for example, species x and y are sympatric and display a degree of reproductive isolation of RI = u), rather than at the population level (for example, species x population a and species y population a are sympatric and RI = u, but species x population a and species y population b are allopatric and RI = v). For only two species complexes, Neurospora and Heterobasidion, were data clearly available at the population level, and these indicated stronger reproductive isolation in sympatry than in allopatry. In the Heterobasidion species complex, one species pair exhibited a mean RIs = 0.87 when the strains had the same geographic origin and a mean RIa = 0.17 when they had different geographic origins (Capretti et al., 1990; Stenlid & Karlsson, 1991). We performed the statistical analyses (see below) considering this species pair in three different ways: (1) excluding this species pair; (2) considering this species pair as sympatric with RI = 0.52 (i.e. the mean RI = (RIs + RIa)/2); and (3) including this species pair both as sympatric with RI = 0.87 and as allopatric with RI = 0.17. The three ways of considering this species pair led to very similar results. We only present the results with exclusion of this pair. In the Neurospora species complex, Dettman et al. (2003) identified a higher level of reproductive isolation among sympatric strains than among allopatric. The difference in the success of crosses between sympatric and allopatric strains was, however, quantitative, which did not change the class of reproductive isolation of the Ascomycota species pairs as we scored it.

Genetic distances

We used ITS sequences available in GenBank to compute genetic distances between sibling species pairs. Sequences were available for 133 species, distributed in 25 of the 33 fungal complexes used in our study. Alignments were performed using Bioedit v6.0.7 (Tom Hall, Isis pharmaceuticals Inc., Carlsbad, CA, USA) and corrected by hand when necessary. Genetic distances between species were calculated using Mega 3.1 (Kumar et al., 2004) using two models, Kimura-2-parameter (K2P) and p-distance. Distances obtained with these models were very similar; we therefore present our results using only the K2P distance.

Data analyses

Because different measures of reproductive isolation had to be used in Ascomycota and Basidiomycota, and because previous reviews had suggested that intersterility may evolve more easily in Basidiomycota than in Ascomycota (Burnett, 2003; Kohn, 2005), analyses were performed separately on the two groups. The single species complex of Basidiomycota not belonging to Homobasidiomycota in our data, M. violaceum, had a life cycle more close to that of Ascomycota for some aspects and its measures of reproductive isolation also corresponded to different phenomena (see above). We therefore performed analyses on Ascomycota and Homobasidiomycota separately.

Statistical analyses were performed using JMP (SAS Institute Inc., SAS Campus Drive, Cary, NC, USA) on each of two data sets to compare patterns of in vitro reproductive isolation between species living in sympatry and allopatry. The first data set considered the results of the crossing experiments performed between all the species pairs as the statistical units, each pair being classified as either allopatric or sympatric, and being assigned a measure of reproductive isolation and a genetic distance when available. This data set including all species pairs, however, presents the potential problem of having many points that are not phylogenetically independent. We therefore used a second data set that considered the species complexes as the statistical units. For each species complex, the mean degree of reproductive isolation (±SEM) and the mean genetic distance (±SEM) of the species pairs were computed, in sympatry and/or allopatry. However, although some of the species pairs within species complex are not phylogenetically independent, we believe that the evolution of reproductive barriers could be considered as independent events, even within a species complex. Indeed, reproductive isolation has to evolve de novo for each speciation event and is not expected to be directly influenced by the type of reproductive barrier that evolved in the preceding speciation event. Evidence for the independent evolution of reproductive barriers is provided by the different types of reproductive barriers that are found within single species complexes.

Levels of reproductive isolation corresponded to discrete values in Ascomycota and to continuous values in Homobasidiomycota. To obtain similar statistical powers when performing statistical analyses within Ascomycota and within Homobasidiomycota, we transformed the reproductive isolation values obtained for Homobasidiomycota into discrete values prior to analyses. Reproductive isolation values of 0–0.24, 0.25–0.49, 0.50–0.74, 0.75–0.94 and 0.95–1.00 were, respectively, put in the classes 0, 0.25, 0.5, 0.75 and 1.

Nonparametric Wilcoxon rank tests were performed to assess whether there was a difference in degree of reproductive isolation between sympatric and allopatric species of Ascomycota on the one hand and Homobasidiomycota on the other hand. These tests were performed using both the data available for all species pairs and for the species pairs displaying a genetic distance <0.05 (see Results). For both Ascomycota and Homobasidiomycota, t-tests were performed to assess whether there was a difference in genetic distance between allopatric and sympatric species pairs. Finally, Spearman rank correlation tests were performed to investigate the relationship between reproductive isolation and genetic distance, in sympatry and allopatry, within Ascomycota and within Homobasidiomycota, using the data available both for all species pairs and for the species pairs with a genetic distance <0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

We obtained data on crossing experiments among sibling species for 16 species complexes of Homobasidiomycota and 16 species complexes of Ascomycota (Table 1). We could estimate genetic distances between sibling species within nine of the Homobasidiomycota species complexes and 14 of the Ascomycota species complexes (Table 1). The mean genetic distance among Homobasidiomycota species pairs (distH = 0.054 ± 0.005) was greater than that among Ascomycota species pairs (distA = 0.023 ± 0.002) (Fig. 1). To perform the analyses on species complexes of similar ages for Homobasidiomycota and Ascomycota, we used a restricted data set including only species pairs with a genetic distance below 0.05 (Fig. 2). In this restricted data set, genetic distances were not significantly different between allopatric and sympatric species pairs or species complexes, both within Ascomycota and Homobasidiomycota (Table 2). There was therefore no age bias in allopatry vs. sympatry comparisons.

image

Figure 1.  Reproductive isolation between species pairs as a function of genetic distance. Plot of the degree of reproductive isolation as a function of genetic distance (K2P), between all species pairs; in (a) Homobasidiomycota, (b) Ascomycota, with indication of the geographical situation: black triangles for allopatry and white circles for sympatry.

Download figure to PowerPoint

image

Figure 2.  Reproductive isolation between species pairs as a function of genetic distance, for genetic distances lower than 0.05. Plot of the degree of reproductive isolation as a function of genetic distance (K2P), between all species pairs having a genetic distance below 0.05; (a) Homobasidiomycota, (b) Ascomycota, with indication of the geographical situation: black triangles for allopatry and white circles for sympatry. Mean values and standard errors are represented separately for allopatry (black square) and sympatry (white square).

Download figure to PowerPoint

Table 2.   Results of the t-test for assessing whether genetic distances are different between allopatric and sympatric species pairs or species complexes, within Ascomycota and within Homobasidiomycota respectively.
  Genetic distancetd.f.P-value
In sympatryIn allopatry
AscomycotaComplexes0.017 ± 0.0120.018 ± 0.0060.288170.78
Pairs0.019 ± 0.0110.021 ± 0.0080.923770.36
HomobasidiomycotaComplexes0.018 ± 0.0100.016 ± 0.0080.414130.69
Pairs0.013 ± 0.0020.014 ± 0.0020.1781030.86

Reproductive character displacement in Homobasidiomycota

Reproductive isolation data were available for 310 species pairs within 16 complexes of Homobasidiomycota (Table 3, Figs 1–3). Complete or nearly complete reproductive isolation was found between virtually all sympatric species pairs: 170 pairs showed complete reproductive isolation (RI = 1) and seven pairs nearly complete reproductive isolation (RI between 0.84 and 0.97). A single sympatric species pair, from the genus Flammulina, exhibited high compatibility (RI = 0.24). Among allopatric species pairs, 93 pairs presented complete reproductive isolation (RI = 1), three pairs showed strong isolation (RI between 0.84 and 0.97), 17 pairs had intermediate levels of compatibility (RI between 0.21 and 0.75) and 19 pairs showed full compatibility (RI = 0).

Table 3.   Results of the nonparametric Wilcoxon rank test to assess the difference in reproductive isolation between sympatric and allopatric species complexes or species pairs, within Ascomycota and within Homobasidiomycota.
  NsNaRIsRIaZP-value
  1. Numbers of sympatric (Ns) and allopatric (Na) units, average reproductive isolation in sympatry (RIs) and allopatry (RIa) ± standard error, absolute Z scores, and P-values are indicated (A) on the complete data set; (B) on species pairs with genetic distances below 0.05.

(A)
AscomycotaPairs94270.58 ± 0.030.55 ± 0.050.450.65
Complexes1690.68 ± 0.06 0.60 ± 0.100.910.36
HomobasidiomycotaPairs1781320.99 ± 0.010.79 ± 0.036.66<0.0001
Complexes16140.99 ± 0.010.70 ± 0.083.520.0004
(B)
AscomycotaPairs60190.52 ± 0.040.45 ± 0.050.620.54
Complexes1360.58 ± 0.07 0.50 ± 0.110.850.40
HomobasidiomycotaPairs58470.98 ± 0.010.72 ± 0.064.20<0.0001
Complexes780.97 ± 0.020.64 ± 0.142.040.04
image

Figure 3.  Reproductive isolation as a function of genetic distance within species complexes, for genetic distances lower than 0.05. Plot of the degree of reproductive isolation as a function of genetic distance (K2P), within complexes whose species pairs had a genetic distance below 0.05; in (a) Homobasidiomycota, (b) Ascomycota, with indication of the geographical situation: black triangles for allopatry and white circles for sympatry. Mean values and standard errors are represented separately for allopatry (black square) and sympatry (white square).

Download figure to PowerPoint

Reproductive isolation was found to be significantly stronger in sympatry than in allopatry for Homobasidiomycota, considering either the species pairs or the species complexes as the statistical units (Table 2). Genetic distances between species pairs could be estimated for 165 species pairs. All pairs with a genetic distance higher than 0.05 presented total incompatibility (RI = 1, n = 60, Fig. 1). Considering only the species pairs with a genetic distance below 0.05, reproductive isolation was still significantly stronger in sympatry than in allopatry (Table 3).

Lack of reproductive character displacement in Ascomycota

Results of crossing experiments were available for 120 species pairs within 16 complexes of Ascomycota (Table 3, Figs 1–3). Among sympatric species pairs, only 24 pairs showed complete reproductive isolation (RI = 1), whereas 64 showed only post-zygotic isolation (23 pairs with RI = 0.25, 23 pairs with RI = 0.5 and 18 pairs with RI = 0.75) and six pairs presented full compatibility (RI = 0).

Among allopatric species pairs, five pairs showed complete reproductive isolation (RI = 1), 20 pairs showed post-zygotic isolation (three pairs with RI = 0.25, 15 pairs with RI = 0.5 and two pairs with RI = 0.75) and two pairs had full compatibility (RI = 0).

There was no significant difference between sympatric and allopatric pairs of Ascomycota in the degree of reproductive isolation, considering either the species pairs or the species complexes as the statistical units (Table 3). Genetic distances were obtained for 89 species pairs. The 10 pairs with a genetic distance higher than 0.05 showed complete reproductive isolation (RI = 1). Considering only the pairs with a genetic distance below 0.05, there was still no significant difference in the degree of reproductive isolation between sympatric and allopatric species pairs (Table 3).

Reproductive isolation as a function of genetic distance

We investigated whether the degree of reproductive isolation and genetic distance were correlated, in sympatry and in allopatry, for both Ascomycota and Homobasidiomycota. We only found a significant correlation among allopatric species pairs of Homobasidiomycota in the complete data set (Table 4).

Table 4.   Results of the Spearman rank correlation test between genetic distance and degree of reproductive isolation, among sympatric or allopatric species pairs or species complexes, within Ascomycota and within Homobasidiomycota respectively.
  SympatryAllopatry
ρP-valueρP-value
  1. (A) On the complete data set; (B) on species pairs with genetic distances below 0.05.

(A)
AscomycotaComplexes0.430.130.310.55
Pairs0.180.14−0.190.42
HomobasidiomycotaComplexes0.210.590.10.82
Pairs0.120.230.46<0.0001
(B)
AscomycotaComplexes0.070.82−0.620.19
Pairs−0.120.35−0.410.09
HomobasidiomycotaComplexes−0.480.280.120.77
Pairs−0.010.520.230.12

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Reproductive character displacement

Our results suggest that there are contrasting patterns of reproductive character displacement in fungi. Homobasidiomycota showed evidence for reproductive character displacement, with stronger reproductive isolation in sympatry than in allopatry, but Ascomycota did not. This was observed for similar genetic distances of Ascomycota and Homobasidiomycota species pairs, and the species complexes had similar genetic distances in allopatry and in sympatry in the two groups. This indicates that the difference between Ascomycota and Homobasidiomycota in the pattern of reproductive isolation was not because of biases in ages of the species.

Data gathered in the literature may be subject to publication bias. For instance, one can imagine that existence of reproductive isolation can be reported more often than the lack of a reproductive barrier, especially between sympatric species pairs. However, the numerous examples of Ascomycota species pairs reported in the literature that are genetically isolated, but do not display any form of premating reproductive isolation, or even any reproductive isolation at all, indicate that such a bias is minimum and there is no reason why it should be present in Homobasidiomycota but not in Ascomycota.

In Homobasidiomycota, reproductive isolation was stronger between sympatric species pairs than between allopatric species pairs, and this was due to enhanced premating reproductive isolation due to sexual partner recognition. All but one of the sympatric species pairs indeed displayed very strong premating reproductive isolation. This suggests the existence of a strong selective pressure for the evolution of premating barriers between closely related species in sympatry, such as reinforcement in response to low hybrid fitness (Noor, 1999). Alternatively, reproductive character displacement can result from other mechanisms, such as the extinction of one population or the fusion of insufficiently isolated incipient species after a secondary contact.

A pattern of reproductive character displacement has been shown to occur in several taxa, but studies generally focused on particular animal species pairs, by comparing reproductive isolation between sympatric and allopatric populations (Coyne & Orr, 2004, pp. 357–365). By contrast, we provide evidence for reproductive character displacement at the scale of a large taxonomic group. Homobasidiomycota are thus good models for studying speciation, in particular for studying the evolution of premating reproductive isolation. However, they are virtually never considered in theoretical work or in reviews (see, for instance, Coyne & Orr, 2004). Many examples of fungal speciation have, however, been studied in detail (Burnett, 2003; Kohn, 2005; Dettman et al., 2003; Le Gac et al., 2007a).

In contrast to Homobasidiomycota, we found no clear pattern of reproductive character displacement in Ascomycota. In most of the cases, both allopatric and sympatric species pairs had similarly low levels of reproductive isolation and the reproductive barriers were mostly post-mating. It is in fact intriguing to find in Ascomycota many cases of pairs of closely related species in sympatry that are not isolated by a strong premating barrier. Theory usually predicts that such species pairs should fuse, undergo reinforcement or one of the species should go extinct (Noor, 1999). The lack of premating isolation among sympatric pairs of Ascomycota may be due to a peculiarity of their life cycle. The mycelium of most of the Homobasidiomycota grows into the ground where it may encounter mycelia from other species. On the contrary, the mycelium of most of the Ascomycota grows within or on a specific substrate (a host) and because their gametes have very low dispersal abilities, Ascomycota can mate only where their parent mycelium is able to grow. As a result, species specialized on different hosts rarely have the opportunity to mate even in sympatry. Thus, such species may be functionally allopatric. A recent model based on such a life cycle predicts that in these parasites sibling species can be maintained in sympatry through strong specialization alone, even in the absence of premating isolation due to recognition (Giraud, 2006; Giraud et al., 2006). Most of the Ascomycota species complexes included in this study were in fact plant pathogens (12 of 16). Further, the above mechanism of ecological speciation can be generalized to nonparasite Ascomycota species specialized in different habitats. We believe that the contrasted pattern found here, i.e. reproductive character displacement in Homobasidiomycota but not in Ascomycota, is consistent with the mechanism of speciation by mere specialization described by Giraud et al. (2006) in which specialization pleiotropically generates both adaptation and reproductive isolation.

Alternative explanations for the maintenance of sibling species in the same geographic area without evolution of strong premating barriers include: (1) the possibility that sexual reproduction is not frequent in nature, impeding strong selection for enhanced premating isolation and fusion of insufficiently isolated populations; (2) the possibility that the sibling species only share a restricted geographic contact zone (parapatry). In this case, the local selective pressure in the contact zone may not be able to overcome the gene flow from nonsympatric areas.

Tempo of speciation

In addition to comparing the degree of reproductive isolation among sympatric and allopatric fungal species, we also looked at the evolution of reproductive isolation with genetic divergence. Reproductive isolation among allopatric incipient species is expected to rise gradually and slowly as a result of independent mutation, genetic drift and indirect effects of natural selection driving local adaptation (Coyne & Orr, 2004, pp. 83–110). Positive correlations between reproductive isolation and genetic distance, commonly taken as a surrogate for time, have in fact been reported among several allopatric taxa, such as insects (Coyne & Orr, 1989; Presgraves, 2002; Christianson et al., 2005), frogs (Sasa et al., 1998), birds (Price & Bouvier, 2002; Tubaro & Lijtmaer, 2002; Lijtmaer et al., 2003), fish (Mendelson, 2003; Russell, 2003; Bolnick & Near, 2005) and angiosperms (Moyle et al., 2004). Here, we found a significant correlation between reproductive isolation and genetic distance among allopatric species in Homobasidiomycota considering all species pairs and the whole range of genetic distances. The lack of significant correlation between genetic distance and reproductive isolation in Ascomycota seems to be because of the lower level of divergence within these species complexes (Fig. 1). Indeed, only 10 species pairs of Ascomycota had a genetic distance higher than 0.05, and they all had reached complete premating isolation. We considered only Homobasidiomycota species pairs with a genetic distance below 0.05, the test was not significant either.

Few studies have examined the time course of speciation in sympatry. Here, we did not find significant correlations between genetic distance and reproductive isolation among sympatric species in either Ascomycota or Homobasidiomycota. The reason may be once again different in the two groups. In Homobasidiomycota, reproductive isolation seems to have evolved rapidly in sympatry, with almost all species pairs completely isolated, so that the degree of reproductive isolation cannot increase further with time. In Ascomycota, reproductive isolation was weak in sympatry as well as in allopatry and the lack of correlation between reproductive isolation and time is certainly due here again to a young age of the species complexes considered. This again shows the difference in Ascomycota and Homobasidiomycota for the evolution of reproductive isolation between sympatric species.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

This study is, to our knowledge, the first one to report that large taxonomic groups of organisms can behave completely differently as regards the evolution of reproductive isolation. It seems that only species with a strong premating isolation due to sexual partner recognition can coexist in sympatry within Homobasidiomycota, whereas sympatric sibling species often coexist without premating isolation within Ascomycota. This suggests that some phylogeny-dependent life-history trait may strongly influence the evolution of reproductive isolation between closely related species. Identifying those life-history traits that influence the evolutionary processes responsible for these different patterns requires comparative studies between phylogenetically independent taxa, having a broad assortment of life-history traits.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

We thank J. A. Shykoff, Y. Brygoo, J. Enjalbert, M. E. Hood, A. P. Møller, E. Fournier, M. Kirkpatrick, R. Yockteng, J. Tyerman and two anonymous referees for helpful discussions and comments and for improving the English of the manuscript. We thank P. Leroux, E. Fournier, B. Le Cam and P. Gladieux for sharing unpublished data. This research was funded by an ACI Jeunes Chercheurs from the French Ministère de la Recherche.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References