|Title||On the evolution of allorecognition and somatic fusion in ascomycete filamentous fungi|
|Source||Wageningen University. Promotor(en): Bas Zwaan, co-promotor(en): Duur Aanen; Fons Debets. - Wageningen : Wageningen University - ISBN 9789462572973 - 128|
Laboratory of Genetics
|Publication type||Dissertation, internally prepared|
|Keyword(s)||ascomycota - schimmels - genetica - moleculaire herkenning - celgroei - evolutie - ascomycota - fungi - genetics - molecular recognition - cell growth - evolution|
Cooperation -behaviour that benefits other individuals- can be beneficial at the level of the group. For example, a large group is better protected against predators than a small group, and a group of individuals dividing certain tasks may be more efficient than a group of individuals that perform all tasks separately. A cooperating group can thus reach a higher reproduction than a group of non-cooperating individuals. However, even though cooperation increases fitness at the level of the group, it is difficult to explain the evolution of cooperation: within a cooperative group, non-cooperative individuals, which still do profit from other, cooperating, individuals will have more resources to spend on reproduction, and thus are predicted to have a higher fitness. Natural selection will thus select these cheaters until there are no cooperating individuals left.
Hamilton’s kin-selection theory predicts that stable cooperation can evolve when cooperation is directed with a higher probability towards genetically related individuals (kin) than towards unrelated individuals. In his formulized condition (rB - C > 0), cooperation is stable when the cost of helping (expressed as the number of offspring not produced because of the cooperative allele) is lower than the benefit (the number of additional offspring as a consequence of the cooperative allele), multiplied by the average relatedness of the individuals that receive the help. Kin-selection theory is not the only theory that can explain cooperation, but is generally accepted as an important factor in many forms of cooperation. One mechanism to direct help preferentially towards kin is population viscosity. Little dispersal results in progeny and ancestors staying close together in a group. If individuals are more motile, active kin discrimination becomes important. In order to direct cooperation preferentially towards kin, many organisms have developed specialized genetic kin recognition mechanisms, based on one or more polymorphic recognition loci. These organisms only cooperate with individuals that partially or fully match their own recognition genes. Crozier made a model based on marine invertebrates that form colonies. When colonies are close to each other they cooperatively fuse with neighbouring colonies when they are clonally related, or actively compete when they are less related. The decision to cooperate is based on genetic allorecognition. Crozier’s model predicted that if fusion increases fitness, common alleles will be favoured since individuals with common recognition alleles will fuse more often. This will lead to selection of the most frequent recognition alleles until recognition polymorphism disappears completely. Thus, there is a cost of allorecognition that may reduce the genetic variation upon which allorecognition crucially relies, a prediction now known as ‘Crozier’s paradox’. An important hypothesis that can solve this paradox is to incorporate the effect of cheating. Cheating will lead to a cost to the group of cooperating individuals and therefore can impose selection pressure to maintain allorecognition. Another hypothesis is that allorecognition diversity may be selected for another function.
Multicellularity is an extreme example of cooperation: the cells of an individual usually show division of labour, and altruism is strong because only a fraction of the cells reach the germline. A multicellular individual thus essentially is a cooperating group of cells, and evolution acts at different levels. The multicellular individual gets selected based on its fitness compared to the multicellular individuals it competes with, while at the same time the cells within the multicellular individual are under natural selection. This can lead to a potential conflict, where cheating cells evolve that have a higher fitness relative to other cells within the individual, but at the same time reduce the fitness of the multicellular individual. Theory predicts that a high relatedness among the cells of an individual reduces the opportunities for such cheating cells. Consistent with this hypothesis, there are some important mechanisms, which maintain high relatedness observed in multicellular organisms. One mechanism is regular single-celled bottlenecks in the lifecycle such as spores or seeds or zygotes from which multicellular individuals develop by mitotic division. Another important mechanism to maintain the high relatedness after the single-celled bottleneck is allorecognition to prevent fusion with non-self cells. Allorecognition is found in most multicellular organisms, but seems most relevant for organisms in which fusion between individuals or aggregation of cells is a notable part of their lifecycle.
In this thesis, I have used filamentous ascomycete fungi as a model for the evolution of stable multicellularity and allorecognition. The fungi have regular single-celled bottlenecks in the form of spore formation, from which they develop by clonal division of the nuclei to form a tubular network known as the fungal mycelium or colony. An interesting aspect of fungal growth is that the mycelium is not clearly divided into cell compartments, with the result that cytoplasm and nuclei can freely move through parts of the colony. This implies that organelles (nuclei, mitochondria), and not cells, are the main potential selective unit below the individual. Another important feature of fungi is that fungal colonies can fuse. Whether neighbouring colonies fuse or reject each other is determined by a highly polymorphic genetic allorecognition mechanism.
In this thesis, I have used these fungi to address the theoretical problem identified by Crozier, the evolutionary stability of genetic kin recognition. We first tested whether fusion between colonies indeed is beneficial compared to allorecognition, and whether this can lead to erosion of allorecognition diversity (chapter 2). We used the ascomycete fungus Neurospora crassa, a well-established model species for genetic research, of which numerous strains are available of different allotype. We found that cultures grown from a single allotype have a higher spore production than cultures grown from a mixture of different allotypes. This shows that fusion is beneficial relative to allorecognition. We determined the precise causes of this relative cost of allorecognition, by using a fusion mutant that partially mimics the effect of allorecognition. Colonies remain separated from each other, similar to colonies separated due to allorecognition. However, in contrast to confrontations between colonies with a different allotype, in which part of the mycelium is sacrificed in a cell death reaction, in confrontations between different colonies of the fusion mutant, there is no active rejection. This comparison showed that the benefit of fusion is due not only to absence of mutual antagonism, which occurs upon allorecognition, but also to an increase in colony size per se. We then experimentally demonstrated that the benefit of fusion selects against allorecognition diversity, as predicted by Crozier. We show that there is a positive correlation between the frequency of an allotype and its competitive fitness, thus showing that positive frequency dependent selection acts on allotype diversity, thus leading to erosion of allotype diversity.
In the remaining part of the thesis, I have used different ascomycete fungus models to test various hypotheses to explain the evolutionary stability of allorecognition. One hypothesis considers allorecognition as a means to protect against cheating genotypes, genotypes that have a competitive advantage in combination with a wildtype genotype, but that reduce total reproductive output (chapter 3). According to recent theoretical models that simulate the evolution of allorecognition in combination with the possibility of somatic cheating, high allorecognition diversity can evolve in combination with low frequencies of cheating. The main condition is that cheating can evolve from cooperative genotypes. In order to test the hypothesis that cheating is a realistic threat to multicellular growth in fungi, we used an experimental evolution approach with N. crassa that maximised the potential for cheating genotypes by selecting under low relatedness, conditions: a high inoculation density, complete mixing at each transfer and in the absence of allorecognition. Within less than 300 generations, all eight replicate lines we evolved under these conditions significantly decreased their average asexual spore production. This yield reduction was caused by genotypes that matched the criteria for cheating: they had increased competitive fitness relative to a cooperative ancestral type, but spore production was significantly decreased when grown in mono culture or together with a cooperative type. A parallel control experiment, in which relatedness was kept high within the colony by using a fusion mutant, did not result in a reduction in asexual spore yield, showing that maintaining high relatedness provides efficient protection against cheating. From these results we can conclude that cheating can evolve quickly from cooperative genotypes, but that cheating only is selected when relatedness is low. This explains that cheating genotypes are generally not picked up from nature, since relatedness will usually be higher under natural conditions. First, the extremely high density used in our experiments is unlikely to occur in nature, so that there is more clonal outgrowth relative to fusion. Second, the high diversity of allorecognition alleles observed in nature will increase the average relatedness among the nuclei of a single individual. At the same time, the threat of cheating creates selection pressure to maintain allorecognition.
A different hypothesis, specific to fungi, is the possibility that allorecognition provides protection against cytoplasmic cheaters (chapter 4). Usually, mitochondria are restricted in their movement by cell compartments, so that there is selection at the level below that of the cells. In fungi, mitochondria can move through the mycelium similar to the nuclei. For this reason, mitochondria can be selected within a fungal colony similar to the way nuclei can be selected within a fungal colony. We studied the evolutionary dynamics of mutant mitochondria that cause senescence in Neurospora intermedia, a species closely related to N. crassa. The mitochondria mutate under the influence of a natural occurring mitochondrial plasmid that acts as a mutagen. The mutated mitochondria have a selective advantage within the fungal colony, which allows them to increase in frequency at the cost of colony fitness. Once the mutated mitochondria reach a high frequency, the colony dies. Therefore, these mitochondrial mutants are typical cheaters, which increase their own relative fitness at the cost of the colony. We performed evolution experiments where we varied relatedness by varying fusion and bottleneck size. We show that reduction of the bottleneck size reduces the predictability of selection of mutant mitochondria. Then, we show that evolution with a fusion mutant effectively selects against mutant mitochondria and prevents senescence of the cultures. In a following experiment we then show that allorecognition can prevent or delay senescence in a similar way as what happens in cultures with a fusion mutant. These experiments confirmed that cheating mitochondrial genotypes provide a realistic threat to fungal multicellularity and that allorecognition can help keeping these mutants at a low frequency.
Although the selective pressure by cheating appears to be sufficient to maintain the allorecognition diversity observed in fungi, it does not exclude the hypothesis that allorecognition diversity can also be the result of selection for another function. In chapter 5, I describe the highly polymorphic het-c locus in Podospora anserina. The het-c locus determines allorecognition together with two unlinked loci termed het-d and het-e. Each het-c allele is incompatible with a specific subset of the het-d and het-e alleles. We found that the het-c allorecognition gene is under diversifying selection and more polymorphic than most other fungal allorecognition genes. Several aspects hint to a possible function in pathogen recognition for the het-c, het-d and het-e allorecognition system, such as its high variability and structural and sequence homologies to plant defence genes. Therefore, we argue that diversity in these genes may be selected for both maintaining allorecognition and pathogen recognition. The characteristics of these genes seem an exception and have not been found for other fungal allorecognition genes. The functioning of these genes in pathogen recognition and defence remains to be demonstrated. So although these results are interesting, cheating remains the most probable solution to explain the evolution of allorecognition diversity.
The results described in this thesis emphasize the influence of somatic cheating on the evolution of allorecognition in fungi. Fungi are economically and medically very important for society. Therefore, the results described in this thesis are very useful since they give new insight in how high relatedness can keep fungal growth stable if this is desired and how cheating might be useful to use against undesired fungal growth. Finally, I discuss that cheating is a risk in most multicellular organisms and that allorecognition is very important to prevent such cheating genotypes from spreading between individuals.