Divergent adaptation promotes reproductive isolation among experimental populations of the filamentous fungus Neurospora
© Dettman et al; licensee BioMed Central Ltd. 2008
Received: 11 July 2007
Accepted: 31 January 2008
Published: 31 January 2008
An open, focal issue in evolutionary biology is how reproductive isolation and speciation are initiated; elucidation of mechanisms with empirical evidence has lagged behind theory. Under ecological speciation, reproductive isolation between populations is predicted to evolve incidentally as a by-product of adaptation to divergent environments. The increased genetic diversity associated with interspecific hybridization has also been theorized to promote the development of reproductive isolation among independent populations. Using the fungal model Neurospora, we founded experimental lineages from both intra- and interspecific crosses, and evolved them in one of two sub-optimal, selective environments. We then measured the influence that initial genetic diversity and the direction of selection (parallel versus divergent) had on the evolution of reproductive isolation.
When assayed in the selective environment in which they were evolved, lineages typically had greater asexual fitness than the progenitors and the lineages that were evolved in the alternate, selective environment. Assays for reproductive isolation showed that matings between lineages that were adapted to the same environment had greater sexual reproductive success than matings between lineages that were adapted to different environments. Evidence of this differential reproductive success was observed at two stages of the sexual cycle. For one of the two observed incompatibility phenotypes, results from genetic analyses were consistent with a two-locus, two-allele model with asymmetric (gender-specific), antagonistic epistasis. The effects of divergent adaptation on reproductive isolation were more pronounced for populations with greater initial genetic variation.
Divergent selection resulted in divergent adaptation and environmental specialization, consistent with fixation of different alleles in different environments. When brought together by mating, these alleles interacted negatively and had detrimental effects on sexual reproductive success, in agreement with the Dobzhansky-Muller model of genetic incompatibilities. As predicted by ecological speciation, greater reproductive isolation was observed among divergent-adapted lineages than among parallel-adapted lineages. These results support that, given adequate standing genetic variation, divergent adaptation can indirectly cause the evolution of reproductive isolation, and eventually lead to speciation.
Adaptation to different environments is postulated to play a large role in speciation and adaptive radiation . With ecological speciation, reproductive isolation between populations evolves incidentally as a by-product of adaptation to divergent environments [2, 3]. There are three potential sources of this reproductive isolation. First, the traits that are under divergent selection may indirectly reduce reproductive success by altering mate choice. Second, ecological mechanisms may contribute directly to reproductive isolation among species. If two species are specialized to different environments, interspecific hybrids may have reduced fitness in both environments because they lack the full complement of adaptive factors possessed by each of the parental species and exhibit a maladaptive intermediate phenotype. Third, divergent adaptation may promote reproductive isolation through genetic incompatibilities, as predicted by the Dobzhansky-Muller model [4, 5]. When populations adapt to divergent environments, selection promotes sorting and fixation of different alleles in the alternate populations. Subsequent hybridization creates novel combinations of alleles that have not been tested and approved by selection, creating the possibility of negative, genetic interactions . Under the Dobzhansky-Muller model, this antagonistic epistasis between incompatible alleles at interacting loci causes reductions in reproductive success. Given enough time, two completely separated populations will inevitably develop genetic incompatibilities due to mutation and genetic drift. Does divergent adaptation accelerate or increase the likelihood of the evolution of genetic incompatibilities, and thus speciation?
The amount of standing genetic variation in the ancestral population before speciation has a major effect on the evolution of reproductive isolation. Low-diversity populations may be restricted to adaptive peaks that are proximal to that of their conspecific ancestors, thereby limiting their ability to explore alternative adaptive peaks. High-diversity populations created by interspecific hybridization may possess greater segregational variance, with the potential for significant negative or positive epistasis among genetic factors. Positive epistasis may allow hybrid populations to access higher-fitness peaks elsewhere on the fitness landscape, thereby facilitating greater and more rapid evolutionary responses to novel environments. Furthermore, the level of potential divergence as a result of differential sorting between lineages increases with genetic variance. Hybrid speciation theory therefore suggests that the greater genetic variation associated with hybrid populations may increase the likelihood and rate of ecological speciation .
To test these key predictions of ecological and hybrid speciation, we developed an experimental evolution system using the filamentous fungus Neurospora. Several attributes make Neurospora ideal for this study. Like other microbes, it is amenable to experimental evolution because of short generation time, and ease of indeterminate propagation and archival storage . The amount of genetic variation in artificial populations can be controlled by crossing parents of known relatedness. Also, the role of ecological specialization in promoting species divergence in this genus finds support in phylogenetic reconstructions .
Most importantly, Neurospora has a well-characterized sexual system with identifiable events associated with mating and sexual development [10, 11]. Defects that cause reproductive isolation can be easily recognized at numerous stages in the sexual cycle [12–14]. In outbreeding Neurospora, mating occurs between haploid individuals of opposite mating type (mat A or mat a). Each self-sterile individual produces receptive "female" mating structures (protoperithecia) and fertilizing "male" propagules (conidia). Successfully fertilized protoperithecium enlarge rapidly and develop into mature fruiting bodies called perithecia. At this point, prezygotic isolation can be assessed by perithecial production. Karyogamy, meiosis, and mitosis occur to produce the haploid progeny (ascospores) which are then forcibly ejected from the perithecium. At this point, postzygotic isolation can be assessed by progeny viability.
Most experimental-evolution studies of reproductive isolation have focused on premating isolation and mating preference in Drosophila [3, 15, 16]. This form of isolation does not play an important role in speciation in fungi and plants because these organisms do not have "behaviour" in the classical, zoological sense. Here we focus on the incompatibilities that occur after mating (gamete transfer), which are central to reproductive isolation in all eukaryotic organisms.
In a recent study of the unicellular yeast, Saccharomyces cerevisiae, divergent adaptation resulted in both extrinsic ecological isolation and intrinsic genetic incompatibilities . It is not known, however, if microbes with different levels of biological complexity respond to divergent, ecological selection in similar ways . Neurospora has a filamentous rather than unicellular growth form, with more complex mating and sexual development than yeast. Here we study speciation with a similar experimental approach to that of ref. 17 but use a multicellular microbial system. In brief, we evolved replicate lineages of Neurospora in one of two sub-optimal, selective environments. The asexual fitness of lineages was then measured to test for adaptation to the selective environment and trade-offs in the alternate environment. To test for the development of reproductive isolation (both prezygotic and postzygotic), the success of matings between lineages was assessed using two different measures. By varying the amount of standing genetic (and phenotypic) diversity in the founding populations, we tested if increased variation caused by hybridization facilitates adaptation and the evolution of reproductive incompatibilities. To test if divergent adaptation increases the likelihood of the development of reproductive isolation, the effects of chance and divergent selection were differentiated by comparing the reproductive success between parallel-adapted lineages and divergent-adapted lineages.
The average sequence divergence between the parental individuals in Exp. H was 4.87%, nearly three times greater than that in Exp. L (1.63%; ref. 9). Replicate lineages of Neurospora were experimentally evolved for seven cycles of propagation, with each cycle consisting of a sexual phase followed by an asexual phase (Fig. 1). Divergent selection was applied during asexual reproduction by exposing the lineages to one of two sub-optimal, selective environments: high salinity or low temperature. For Exp. L, lineages evolved in high salinity or low temperature were named LN or LT, respectively. For Exp. H, lineages evolved in high salinity or low temperature were named HN or HT, respectively. In total, each experimental lineage underwent 7 meiotic generations and ~1500 mitotic generations.
To test for trade-offs, growth rate was assayed in the alternate, selective environment. The evolved lineages displayed an average fitness reduction of 1.2% (± 1.3 SE), relative to fastest ancestor. Only the LT lineage class had a significantly reduced growth rate in high salinity (Fig. 2; one tailed, df = 5, p = 0.04). In the permissive environment (Fig. 2), each lineage class had a significantly greater growth rate than the fastest parent (two tailed, df = 5, p ≤ 0.04 in each case), with an average fitness gain of 6.8% (± 0.8 SE).
In nature, the growth rate of evolved lineages in relation to each other, rather than in relation to the ancestor, is what determines competitive fitness. When assayed in the high-salinity environment, high-salinity-evolved lineages had a greater mean growth rate than low-temperature-evolved lineages (one tailed, df = 5; LN > LT, p < 0.007; HN > HT, p = 0.37). Similarly, low-temperature-evolved lineages had a greater growth rate than high-salinity-evolved lineages when assayed in the low-temperature environment (LT > LN, p = 0.10; HT > HN, p = 0.02). When modeled with ANOVA, the interaction between assay environment and selective environment was significant for Exp. L (df = 1, p = 0.0038), but not for Exp. H (p = 0.19). Despite these mixed significance levels, the overall pattern suggested a negative correlation of fitness gains across environments.
Sexual reproductive success – progeny viability
To monitor the effects of environmental selection on the sexual reproductive success among lineages, we assayed progeny viability during the course of evolution. Assay points 1–7 were designated at the end of each of the seven sexual-asexual propagation cycles. Progeny viability assays were performed by crossing each lineage with the five other evolved lineages and measuring ascospore germination under permissive conditions. Given the reciprocity of matings regarding gender (male/female) and mating type (mat A/mat a), 4 different matings were performed for each of the 15 pairwise combinations of lineages within each experiment. Thus, each assay comprised 60 matings and the scoring of approximately 10,000 ascospores.
Evolution in sub-optimal conditions initially reduced progeny viability, however, regular sexual cycles subjected the experimental lineages to selection for greater reproductive success. As a response, progeny viabilities increased an average of 0.06 (± 0.06 SE) and 0.03 (± 0.02 SE) per successive assay point for Exp. L and H, respectively. Over the course of the evolution regimen, progeny viabilities had nearly recovered to the values observed for the original, parental crosses (Exp. L parental cross = 0.78; Exp. H parental cross = 0.17).
Summary of analysis of variance for progeny viability.
Assay Point 6
Assay Point 7
L (low diversity)
H (high diversity)
Progeny viability in selective environments
Sexual reproductive success – perithecial production
Progeny viability measures postzygotic isolation and reflects the quality of progeny produced at the completion of sexual reproduction. Perithecial production measures prezygotic isolation and reflects the success of fertilization events. Perithecial number also influences the quantity of progeny produced, assuming that more perithecia produce more progeny.
Genetic basis of incompatibility phenotype
The genetic basis of the interactions causing the reductions in perithecial production was investigated further. Sixty f1 progeny were isolated from crosses between HT2 and HN2 lineages. These progeny were mated back to the HT2 and HN2 parents, in both gender roles, and perithecial production was assayed.
When the f1 progeny were grown as females and fertilized by the HT2 and HN2 males, perithecial production values formed a bimodal distribution (not shown). The discriminatory phenotype, defined as a female producing >3 times more perithecia when fertilized by HT2 males than when fertilized by HN2 males, segregated 30:30 (χ2, p = 1.0). These two lines of evidence suggested that the phenotype was controlled by a single, Mendelian locus. This locus was named dfe for "discriminatory female", and the alleles that confer discrimination or non-discrimination were named dfe d and dfe n , respectively.
When these progeny were grown as males and used to fertilize discriminatory HT2 females, perithecial production values again formed a bimodal distribution (not shown). The discriminated phenotype, defined as the inability of a male to activate an HT2 female to produce >1 perithecia/mm2, was observed in 26 of the 60 f1 progeny (χ2, p = 0.30). This segregating locus, which controls the "discriminated male" phenotype, was named dma and the alleles that confer discrimination or non-discrimination were named dma d and dma n , respectively.
Mean perithecial production for matings between all combinations of the 4 two-locus genotypes.
dfe d dma d
dfe d dma n
dfe n dma d
dfe n dma n
dfe d dma d
dfe d dma n
dfe n dma d
dfe n dma n
We were able to reconstitute the original, parental HT2 and HN2 genotypes, and associated phenotypes, by crossing recombinant f1 genotypes (dfe n dma n × dfe d dma d ). Among 24 f2 progeny, we obtained 13 parental (9 dfe d dma d ; 4 dfe n dma n ) and 11 recombinant (6 dfe d dma n ; 5 dfe n dma d ) genotypes (χ2, p = 0.51). Again, the two alleles at both loci segregated evenly in these f2 progeny (χ2, p > 0.22 in both cases).
The founding populations of Neurospora were composed of multiple individuals with standing genetic variation. As evolution progressed, the genetic composition of these initially identical lineages could change by the sorting of pre-existing and new alleles, or by the creation of novel allele combinations through recombination. How these processes interact with the different types and strengths of selective pressures during the evolution regimen ultimately determines the phenotype of the end-point lineages. In this study, we investigated the effects of these processes on adaptive increases in asexual fitness and the evolution of reproductive isolation.
The best evidence for adaptation is when the evolved lineages have greater fitness than their most-fit ancestor (Fig. 2). Although specialization to one environment may be associated with fitness reductions in other environments [19, 20], we found only limited evidence for such fitness trade-offs (Fig. 2). These results are fully consistent with the recent experimental study of speciation in yeast . Taken together, the mycelial growth rate data demonstrate that the negative correlation of fitness gains in the two selective environments was a result of divergent environmental selection. Surprisingly, divergent selection also caused a correlated response of an increase in general growth rate in the permissive environment (Fig. 2). Furthermore, environmental adaptation in the asexual phase had a correlated effect on ascospore germination in those selective environments (Fig. 5). This pattern is expected if ascospore germination is considered to be the first step in mycelial growth.
Compared to previous experimental studies with fungi [17, 21–24], the levels of environmental adaptation we observed here were unexpectedly low. We used mycelial growth rate as a proxy for fitness because it was the most obvious trait under selection; the mycelial sector with the fastest growth rate was transferred as the founder for the next culture tube. Other asexual traits, such as biomass accumulation, may have also been under selection, but were not measured. In addition, selection during sexual cycles may have counteracted selection during the asexual phase, causing a reduction in the net asexual fitness gains [25–27].
Another factor potentially impeding adaptation is the coenocytic growth form of this filamentous fungus. Hyphae of Neurospora are densely packed with nuclei that are capable of migrating between hyphal compartments. The interconnectedness of the mycelial network may have a homogenizing effect, interfering with competition among genetically different nuclei. Individual nuclei have, at best, only partial control of the phenotype and are not fully exposed to selection. Propagation of lineages by conidia, in which individual nuclei are more exposed to selection than in hyphal tips, may have allowed for more efficient competition and a greater adaptive response . Further comparative, experimental studies are needed to determine the relationship between growth form and adaptive potential of filamentous fungi.
When ecological speciation via divergent adaptation occurs, there are two possible underlying causes of sexual reproductive isolation. First, direct selection on the sexual phase may cause specialization to a particular environment, such that sexual reproduction is disrupted when not within that environment. In this case, reductions in reproductive success are caused by interactions between adapted genotypes and the environment ("ecological isolation"). Second, divergent selection during the asexual phase may cause the accumulation of negatively interacting loci, which consequently have indirect effects on the sexual phase. In this case, reductions in reproductive success are caused by antagonistic epistasis between genetic factors in the adapted genomes ("genetic isolation").
To discriminate the direct effects of divergent adaptation during the sexual phase from the indirect effects of divergent adaptation during the asexual phase, these two selective pressures were experimentally decoupled. The Neurospora populations were exposed to selective, divergent environments only during the asexual phase. Most organisms spend the majority of their life in the asexual phase, so adaptive evolutionary responses to environmental selection typically occur in this phase. At no point in our evolution regimen was reproductive isolation the target of selection. The sexual phase was performed under identical, permissive, conditions for all lineages. Thus, any differential patterns of sexual reproductive success could be attributed to the indirect effects of divergent selection on non-sexual traits that had overpowered the unifying effects of parallel selection on the sexual phase.
Our experimental design also allowed us to identify genetic isolation. When sexual reproductive success was assayed in the permissive environment, any patterns of differential reproductive success could be attributed solely to genetic isolation, and not to ecological isolation. One may argue that, in nature, genetic isolation is far more important than ecological isolation because the latter is, by definition, dependent upon the environment. Incompatibilities caused by intrinsic genetic interactions may be permanent despite changing environments.
We could determine if the development of reproductive isolation was caused by random effects of chance or by deterministic effects of divergent selection. If the evolution of reproductive incompatibilities was random, these phenotypes should not be correlated with the selective environments: parallel and divergent selection should have similar effects on reproductive isolation. Thus, divergent selection can be implicated as the causal factor only when greater reproductive isolation is observed between divergent-adapted lineages than between parallel-adapted lineages (Figs. 3, 4, 6).
In general, the conclusion that divergent adaptation promotes the development of reproductive isolation was supported by results from progeny viability assays, particularly from the later stages of the experiments (Fig. 3). Differential progeny viability that was consistent with ecological speciation was observed in almost all comparisons made at the level of lineage class as female (Fig. 4). The two lineage classes (LT and HN) that drove the trend were evolved in different environments, and were derived from founding populations with different levels and sources of genetic variation. Thus, differential progeny viability was not dependent upon a particular selective environment, or particular genetic components possessed by the progenitors. Rather, divergent selection in divergent environments caused the patterns of differential reproductive success.
Perithecial production assays for Exp. H, but not Exp. L, were consistent with divergent adaptation promoting the development of reproductive isolation (Fig. 6). For Exp. H, females produced more perithecia when fertilized by parallel-adapted males than when fertilized by divergently-adapted males. When HT lineages functioned as the female, the discrepancy was quite dramatic: PE matings produced, on average, 9.39 times more perithecia than DE matings.
The incompatibility phenotype and its inheritance pattern support a two-locus, two-allele model with asymmetric, antagonistic epistasis. The two loci, dfe and dma, interact in a gender-role-specific fashion to modulate the progression through the sexual cycle. Antagonistic interactions between the dfe d and the dma d alleles cause major reductions in perithecial production, but only when the female possesses the dfe d allele and the male possesses the dma d allele. The reciprocal pairing, and all other allele pairings, do not result in the incompatibility phenotype. This model is fully compatible with the Dobzhansky-Muller model.
Reductions in reproductive success are compounded over the successive stages of a single life cycle. When perithecial production and progeny viability are combined, PE matings have a 7.64-fold advantage over DE matings for Exp. H. For Exp. L, in which the degree of differential reproductive success was much lower, PE matings have only a 1.13-fold (13%) advantage over DE matings. Although this is not a large difference, one must consider that these effects are compounded over successive generations, allowing minor disparities in reproductive success to have substantial consequences on gene flow between populations.
We are aware of only six other experimental evolution studies in which the effects of divergent selection and genetic drift could be separated. Recent work with experimental populations of yeast clearly demonstrated that divergent adaptation resulted in significantly greater reproductive isolation than parallel adaptation . The other five studies all used Drosophila as the experimental system, but obtained mixed results. In four studies [29–32], divergent selection had greater effects than parallel selection. In another study , however, the predictions of ecological speciation were not supported.
This study revealed the strong influence that the maternal partner has on the success of a mating. These results are explicable for Neurospora because the female bears nearly the whole burden of sexual reproduction. The maternal partner forms complex structures for mating, actively senses the presence of a male, and produces most of the tissue comprising a mature perithecium. In contrast, the only role of the male is to passively donate nuclei to the female.
Increased genetic variance caused by interspecific hybridization allows for both negative and positive epistasis, thus expanding the fitness distribution in both directions. The Dobzhansky-Muller model focuses on negative interactions between heterospecific genomes that may cause less fit genotypes. Hybrid speciation theory focuses on positive interactions (and transgressive segregation) that may create more fit genotypes. In this sense, the two theories are opposite sides of the same coin. Here we tested whether the potential benefits of positive epistasis could overpower the potential detriments of negative epistasis, regarding the overall fitness of lineages and their ability to adapt; the results of our experiment suggest they cannot. Even though the multiple rounds of sex and recombination should have allowed the separation and purging of negatively interacting factors, lineages from the high-diversity Exp. H did not adapt to the selective environments more than the lineages from the low-diversity Exp. L. When mycelial growth rates in all three environments are considered (Fig. 2), it appears that Exp. L lineages actually had greater responses (both direct and correlated) to selection than Exp. H lineages.
Results from progeny viability and perithecial production assays supported predictions from hybrid speciation theory: differential reproductive success was more common and more striking in Exp. H than in Exp. L (Figs. 3, 4, 6). Exp. H was initiated from a cross between two species, N. crassa and N. intermedia, with nearly complete reproduction isolation [12–14]. The founding population likely contained considerable genetic variance for reproductive success. We hypothesize that, as variation sorted out during the course of evolution, lineages became fixed for particular combinations of compatible genomic regions. Parallel adaptation during the asexual phase promoted the fixation of similar genomic regions. Divergent adaptation promoted the fixation of dissimilar genomic regions, thereby increasing the likelihood of genetic incompatibilities among lineages.
Identifying the genomic regions and, ultimately, the genes that interact to produce the incompatibility phenotypes is critical to understanding the causes of reproductive isolation. Further research should investigate the genomic composition of the evolved lineages, and its correlation with adaptation and incompatibility. Genomic regions consistently shared among parallel-selected lineages could point to the genes or mutations under selection and involved in adaptation. The most discrete differences in reproductive success were in perithecial production, so it might be possible to determine the location and characteristics of the two genetic factors (dfe and dma). At this point, it is unclear if these loci encode single genes or sets of tightly linked genes. By genetic manipulation and controlled crosses, one could determine precisely the nature and mechanism of the antagonistic interaction that causes the Dobzhansky-Muller incompatibility.
This study demonstrated three important aspects of speciation in a developmentally complex eukaryotic microbe. First, divergent selection on a common, founding population resulted in divergent adaptation and environmental specialization. Evolved populations typically had higher fitness in their selective environment than in the alternate environment in which they had not evolved. This negative correlation of fitness across environments provides the phenotypic variation upon which ecological mechanisms of isolation can act. Second, divergent adaptation resulted in greater reproductive isolation than parallel adaptation. Evidence for differential reproductive success was observed at two stages of the sexual cycle: perithecial production and progeny viability. The genetic basis for one of the observed incompatibility phenotypes can be explained by a two-locus, two-allele model with asymmetric, antagonistic epistasis. Third, the effects of divergent adaptation on reproductive isolation were more pronounced for populations with greater initial genetic variation. Interspecific hybridization may produce novel variation that allows the populations to follow previously inaccessible evolutionary trajectories. Taken together, our results support the hypothesis that, given adequate standing genetic variation, divergent adaptation can indirectly cause the evolution of differential reproductive success, and eventually lead to speciation.
Two strains of N. crassa subgroup NcA (D143 [mat A, FGSC 8903] and D69 [mat a, FGSC 8829]) and one strain of N. intermedia subgroup NiA (D73 [mat a, FGSC 8833]) were used in this study. Species tester strains of N. crassa (FGSC 6682 [mat A] and FGSC 6683 [mat a]) were used for mating-type determination. Information on these strains can be obtained from refs. 9 and 14, or the Fungal Genetics Stock Center (FGSC ). Unless otherwise stated, strains were cultured on Vogel's Minimal (VM) medium (1× Vogel's salt solution, 1.5% sucrose, 1.5% agar).
Crossing and progeny isolation
The Neurospora species used in this experiment are heterothallic (self-sterile). Each individual can act as a "female" or "male", depending upon how the individual was prepared for mating and how the cross was performed. Gender is determined by the direction of fertilization and nuclear transfer. The individual or population to act as the female was grown in 13 × 100 mm test tubes containing slanted Westergaard's Synthetic Crossing (SC) medium (1× SC salts, 1% sucrose, 2% agar) to promote the development of specialized, female mating structures (protoperithecia). The male was grown on VM medium to produce mitotic propagules (conidia). After 4 d, the female was fertilized with a dense suspension of conidia harvested from the male. Hair-like protrusions (trichogynes) from the protoperithecia sense and actively grow towards nearby males through chemotaxis. Successfully fertilized hyphae proliferate within the enlarging protoperithecium that develops into a mature fruiting body (perithecium). Within the perithecium, apical segments of the fertilized hyphae (asci) undergo karyogamy, meiosis, and one mitotic division, with each ascus yielding eight haploid progeny (ascospores). When mature, the asci forcibly eject the ascospores out of an apical opening (ostiole) in the perithecium: the ascospores adhere to the inside test tube wall. Ascospores were harvested 18 d after fertilization, then heat-shocked for 40 min at 60°C to activate the germination process. Ascospores were spread on colony-restricting sorbose medium (1× Vogel's salt solution, 2.0% sorbose, 0.05% dextrose, 2.0% agar) and allowed to germinated for 24 h. Viable colonies formed by individual ascospores were transferred to separate test tubes. The mating type of progeny was determined by spotting conidial suspensions on lawns of female-receptive mat A and mat a tester strains.
To produce the low-diversity, founding populations for Experiment L, an intraspecific cross between N. crassa D143 and N. crassa D69 was performed. For Experiment H, an interspecific cross between N. crassa D143 and N. intermedia D73 was used produce the hybrid, high-diversity, founding populations. Note that N. crassa D143 was a parent common to both experiments. The sequence divergence between parental individuals was the mean uncorrected genetic distance from four independent loci . For each experiment, 25 progeny of each mating type were isolated. The individuals in each mating-type class were grown to the point of conidiation, harvested, then mixed in approximately equal proportions in sterile distilled water. Other than the hybrid origin of the founding populations in Exp. H, all other protocols were the same for Exp. L and H.
For Exp. L, the 25-genotype, mat A founding population was divided into two groups of three sub-lineages, designated LN1-A, LN2-A, LN3-A, and LT1-A, LT2-A, LT3-A. Similarly, the mat a founding population was divided into sub-lineages designated LN1-a, LN2-a, LN3-a, and LT1-a, LT2-a, LT3-a. Lineages consisted of paired sub-lineages of opposite mating type; for example, lineage LN1 consisted of sub-lineages LN1-A and LN1-a. The same was done for Exp. H, except the lineages were given HN and HT names (see Fig. 1). To initiate each of the 12 sub-lineages, a 50 μl aliquot of the founding population mixture, containing approximately 5 million conidia, was inoculated into the end of glass tubes (2.5 × 30 cm) filled horizontally with 20 ml of solid media (1.5% agar). Conidia typically contain 2–3 nuclei, so the inoculum represents between 10–15 million nuclei. The LN and HN lineages were propagated asexually in a high salinity environment (VM medium with 0.7 M NaCl), whereas the LT and HT lineages were propagated in low temperature environment (VM medium incubated at 12°C). The sub-lineages of opposite mating type were propagated separately to avoid the loss of one mating type by drift. When cultures neared the end of tubes, a 15 × 20 mm strip of the entire leading colony front was transferred to a new tube with fresh medium. Not only were hyphae growing on the surface of the medium, but significant amounts of submerged and aerial hyphae were present. Thus, we assumed that 100% of the 300 mm2 of two-dimensional area was colonized. Given that the average hyphal strand is 12 μm wide, and using the calculated nuclear densities (see below), we estimated that our asexual propagation regimen transferred approximately 25 million nuclei between successive culture tubes.
Experimental lineages underwent a total of 450 cm of asexual growth. Determination of the number of mitotic generations the lineages had evolved is not straight-forward for Neurospora. First, the coenocytic mycelium has numerous nuclei per hyphal compartment, and nuclei can migrate freely between compartments. Second, mitosis is not synchronized across nuclei. Third, nuclear division is not restricted to the growing front of the colony. These characteristics make tracking the history of individual nuclei impossible. Despite these difficulties, we estimated the number of cell cycles by two methods. The first estimate relied on the correlation between cell cycle duration and mycelial extension rate. Given a cell cycle time of 1.5 h  and a growth rate of 2.0 mm/h at 22°C on VM medium, 16.67 cell cycles are completed for each 5 cm of growth. The second estimate relied on the number of nuclei per unit length of hyphal filament. Neurospora hyphae are densely nucleated, with an estimated 1000 nuclei per 1 mm of hyphae [36, 37]. A single hyphal filament of 5 cm in length, which approximates the length of the actively growing colony front, would contain ~50,000 nuclei. Assuming a single initial nucleus, and that all nuclei remain mitotically active, 15–16 cell cycles would be needed to generate 50,000 nuclei and 5 cm of growth. Thus, the two methods gave similar results, and we estimate that 450 cm of asexual growth represents approximately 1500 mitotic generations. Our generation calculations (16.67 cell cycles in 5 cm) are very conservative compared to estimates from other filamentous fungi (e.g. ref 38; 115 cell cycles in 3 cm), so we may in fact be greatly underestimating the generation number. Moreover, our estimates do not include any of the mitotic generations that occurred during the sexual cycles and progeny isolation.
Asexual evolution was interspersed with six cycles of sexual reproduction (Fig. 1), with an average of 68 cm of asexual growth between cycles. To maintain the genetic cohesiveness of lineages, genetic exchange was achieved by crossing each sub-lineage with its paired sub-lineage of opposite mating type. To reconstitute each sub-lineage, 20 progeny of the appropriate mating-type were isolated and mixed in approximately equal proportions in sterile distilled water. Further asexual evolution was re-initiated from approximately 5 million conidia in solution. Including the initial founding cross, lineages underwent a total of 7 sexual cycles.
Mycelial growth rate
Cultures were inoculated near the perimeter of petri plates (100 × 15 mm) containing the appropriate medium. To allow cultures to acclimate to the assay environment and reach a steady-state growth rate, measurements were started only after the culture had grown approximately 2 cm or more. Rates were averaged over the entire assay period, the duration of which depended upon the environment. Growth rates were determined for each sub-lineage by averaging three replicated measurements. The permissive environment was VM medium at 22°C. Mycelial growth rates were measured at the end of the evolution regimen.
To test for adaptation, the growth rate of an evolved sub-lineage was divided by the growth rate of the fastest parental progenitor. The N. crassa D143 parent was used in all cases because it had a greater growth rate than the other parent, for both Exp. L and H, and in all environments. The statistical significance of deviations from 1.0 was determined by t-tests. Tests for adaptation (>1) or trade-offs (<1) were one-tailed, whereas tests for correlated responses in the permissive environment were two-tailed.
Sexual reproductive success – progeny viability
Progeny viability assays were performed by mating sub-lineages from the same experiment. A culture from each sub-lineage was grown for 3 d on VM slants, then used to inoculate 5 SC slants for mating. The physiological states of cultures prior to mating were similar because all cultures were grown outside of the selective environments, under permissive conditions, for two transfers: first on VM slants, then on SC slants. These steps ensured that the stress of growing in the selective environments did not affect reproductive success or confound the assessment of genetic isolation. Progeny viability assays consisted of 60 matings: each of the 12 sub-lineages was fertilized by the 5 different sub-lineages of opposite mating type. Each pairing of two sub-lineages consisted of two reciprocal matings: one with the mat A partner as female, one with the mat a partner as female. This design allowed the assessment of female function to be separated from male function.
For each mating, heat-shocked ascospores were spotted on sorbose medium and allowed to germinate at 22°C for 24 h. Progeny viability was quantified as the proportion of ejected ascospores that germinated to produce hyphae. For the full assays in the permissive environment, we scored an average of 170 ascospores from each mating, resulting in an average of 10,168 ascospores per assay. The significance of differences between mating classes was determined from germination/non-germination frequencies using Fisher Exact tests. Matings were weighted in inverse proportion to the number of ascospores scored, such that each mating contributed equally to the statistic. To represent the variation among matings within classes, germination counts were converted to proportions and standard errors were calculated.
When progeny viability was scored in the selective environments, we used only the 30 matings in which the mat A partner was female. Analyses of full assays showed that the mating type of the female did not affect reproductive success. These assays were performed at assay point 7 for Exp. L.
Sexual reproductive success – perithecial production
The female was inoculated and grown on SC plates for 4 d to develop receptive protoperithecia. A 2.5 μl aliquot of a conidial suspension made from the male was spotted onto the female. After 8 d, we counted the maximum number of perithecia in the field of view of a dissecting scope at 50× magnification (19.63 mm2), from which the number of perithecia per mm2 was calculated. At this developmental stage, successfully fertilized protoperithecia have commenced sexual development, as indicated by a darkening of the perithecial wall and the formation of beaks or beak-like protrusions. Such fruiting bodies were classified as perithecia. Protoperithecia that had not commenced sexual development by day 8 never matured into perithecia, as verified by inspection of the same areas 7 d later.
The concentration of conidia in the fertilizing solution was great enough that all protoperithecia presumably had access to conidia. When a subsample of fertilizations was repeated with 1/10 and 1/100 dilutions of conidial solutions, the same number of perithecia were produced. Thus, females were over 100× saturated by males.
This work was supported by Discovery grants (JBA, LMK) and a Postdoctoral Fellowship (JRD) from the Natural Sciences and Engineering Research Council of Canada.
- Schluter D: The Ecology of Adaptive Radiation. 2000, Oxford: Oxford University PressGoogle Scholar
- Schluter D: Ecology and the origin of species. TREE. 2001, 16: 372-380.PubMedGoogle Scholar
- Rundle HD, Nosil P: Ecological speciation. Ecol Lett. 2005, 8: 336-352. 10.1111/j.1461-0248.2004.00715.x.View ArticleGoogle Scholar
- Dobzhansky T: Genetics and the Origin of Species. 1937, New York: Columbia University PressGoogle Scholar
- Muller HJ: Isolating mechanisms, evolution, and temperature. Biol Symp. 1942, 6: 71-125.Google Scholar
- Orr HA, Turelli M: The evolution of postzygotic isolation: Accumulating Dobzhansky-Muller incompatibilities. Evolution. 2001, 55: 1085-1094.View ArticlePubMedGoogle Scholar
- Seehausen O: Hybridization and adaptive radiation. TREE. 2004, 19: 198-207.PubMedGoogle Scholar
- Elena SF, Lenski RE: Evolution experiments with microorganisms: The dynamics and genetic bases of adaptation. Nat Rev Genet. 2003, 4: 457-469. 10.1038/nrg1088.View ArticlePubMedGoogle Scholar
- Dettman JR, Jacobson DJ, Taylor JW: A multilocus genealogical approach to phylogenetic species recognition in the model eukaryote Neurospora. Evolution. 2003, 57: 2703-2720.View ArticlePubMedGoogle Scholar
- Raju NB: Genetic control of the sexual cycle in Neurospora. Mycol Res. 1992, 96: 241-262.View ArticleGoogle Scholar
- Davis RH: Neurospora: Contributions of a Model Organism. 2000, New York: Oxford University PressGoogle Scholar
- Perkins DD, Turner BC: Neurospora from natural populations: toward the population biology of a haploid eukaryote. Exp Mycol. 1988, 12: 91-131. 10.1016/0147-5975(88)90001-1.View ArticleGoogle Scholar
- Turner BC, Perkins DD, Fairfield A: Neurospora from natural populations: a global study. Fungal Genet and Biol. 2001, 32: 67-92. 10.1006/fgbi.2001.1247.View ArticleGoogle Scholar
- Dettman JR, Jacobson DJ, Turner E, Pringle A, Taylor JW: Reproductive isolation and phylogenetic divergence in Neurospora: comparing methods of species recognition in a model eukaryote. Evolution. 2003, 57: 2721-2741.View ArticlePubMedGoogle Scholar
- Rice WR, Hostert EE: Laboratory experiments on speciation: What have we learned in 40 years?. Evolution. 1993, 47: 1637-1653. 10.2307/2410209.View ArticleGoogle Scholar
- Coyne JA, Orr HA: Speciation. 2004, Sunderland, Massachusetts: Sinauer Associates, IncGoogle Scholar
- Dettman JR, Sirjusingh C, Kohn LM, Anderson JB: Incipient speciation by divergent adaptation and antagonistic epistasis in yeast. Nature. 2007, 447: 585-588. 10.1038/nature05856.View ArticlePubMedGoogle Scholar
- Kohn LM: Mechanisms of fungal speciation. Ann Rev Phytopath. 2005, 43: 279-308. 10.1146/annurev.phyto.43.040204.135958.View ArticleGoogle Scholar
- Cooper VS, Lenski RE: The population genetics of ecological specialization in evolving Escherichia coli populations. Nature. 2000, 407: 736-739. 10.1038/35037572.View ArticlePubMedGoogle Scholar
- MacLean RC, Bell G, Rainey PB: The evolution of a pleiotropic fitness tradeoff in Pseudomonas fluorescens. Proc Natl Acad Sci USA. 2004, 101: 8072-8077. 10.1073/pnas.0307195101.PubMed CentralView ArticlePubMedGoogle Scholar
- Papa KE: Continuous selection for increased linear growth rate in Neurospora crassa. J Hered. 1971, 62: 87-89.PubMedGoogle Scholar
- Zeyl C, Vanderford T, Carter M: An evolutionary advantage of haploidy in large yeast populations. Science. 2003, 299: 555-558. 10.1126/science.1078417.View ArticlePubMedGoogle Scholar
- Clark TA, Anderson JB: Dikaryons of the basidiomycete fungus Schizophyllum commune: Evolution in long-term culture. Genetics. 2004, 16: 1663-1675. 10.1534/genetics.104.027235.View ArticleGoogle Scholar
- Goddard MR, Charles H, Godfray J, Burt A: Sex increases the efficacy of natural selection in experimental yeast populations. Nature. 2005, 434: 636-640. 10.1038/nature03405.View ArticlePubMedGoogle Scholar
- da Silva J, Bell G: The ecology and genetics of fitness in Chlamydomonas VI. Antagonism between natural selection and sexual selection. Proc Roy Soc, London B. 1992, 249: 227-233. 10.1098/rspb.1992.0109.View ArticleGoogle Scholar
- Xu JP: Estimating the spontaneous mutation rate of loss of sex in the human pathogenic fungus Cryptococcus neoformans. Genetics. 2002, 162: 1157-1167.PubMed CentralPubMedGoogle Scholar
- Zeyl C, Curtin C, Karnap K, Beauchamp E: Antagonism between sexual and natural selection in experimental populations of Saccharomyces cerevisiae. Evolution. 2005, 59: 2109-2115.View ArticlePubMedGoogle Scholar
- Schoustra SE, Slakhorst M, Debets AJM, Hoekstra RF: Comparing artificial and natural selection in rate of adaptation to genetic stress in Aspergillus nidulans. J Evol Biol. 2005, 18: 771-778. 10.1111/j.1420-9101.2005.00934.x.View ArticlePubMedGoogle Scholar
- de Oliveira AK, Cordeiro AR: Adaptation of Drosophila willistoni experimental populations to extreme pH medium II. Development of incipient reproductive isolation. Heredity. 1980, 44: 123-130.View ArticleGoogle Scholar
- Kilias G, Alahiotis SN, Pelecanos M: A multifactorial genetic investigation of speciation theory using Drosophila melanogaster. Evolution. 1980, 34: 730-737. 10.2307/2408027.View ArticleGoogle Scholar
- Dodd DMB: Reproductive isolation as a consequence of adaptive divergence in Drosophila pseudoobscura. Evolution. 1989, 43: 1308-1311. 10.2307/2409365.View ArticleGoogle Scholar
- Rundle HD, Chenoweth SF, Doughty P, Blows MW: Divergent selection and the evolution of signal traits and mating preferences. PloS Biol. 2005, 3: 1988-1995. 10.1371/journal.pbio.0030368.View ArticleGoogle Scholar
- Rundle HD: Divergent environments and population bottlenecks fail to generate premating isolation in Drosophila pseudoobscura. Evolution. 2003, 57: 2557-2565.View ArticlePubMedGoogle Scholar
- Fungal Genetics Stock Center. [http://www.fgsc.net]
- Martegani E, Tome F, Trezzi F: Timing of nuclear division cycle in Neurospora crassa. J Cell Sci. 1981, 48: 127-136.PubMedGoogle Scholar
- Namboodiri AN, Lowry RJ: Vegetative nuclear division in Neurospora. Am J Bot. 1967, 54: 735-748. 10.2307/2440951.View ArticleGoogle Scholar
- Freitag M, Hickey PC, Raju NB, Selker EU, Read ND: GFP as a tool to analyze the organization, dynamics and function of nuclei and microtubules in Neurospora crassa. Fungal Genet Biol. 2004, 41: 897-910. 10.1016/j.fgb.2004.06.008.View ArticlePubMedGoogle Scholar
- Schoustra SE, Debets AJM, Slakhorst M, Hoekstra RF: Reducing the cost of resistance; experimental evolution in the filamentous fungus Aspergillus nidulans. J Evol Biol. 2006, 19: 1115-1127. 10.1111/j.1420-9101.2006.01102.x.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.