Advances in understanding the evolution of fungal genome architecture

Hypermutation

Mutation rates estimate the number of genetic changes occurring within a cell per generation. Isolates with significantly higher mutation rates than those of closely related isolates or laboratory reference strains of the same species are said to display a hypermutator phenotype. Increased mutation rates have been extensively characterized in the yeast Saccharomyces cerevisiae and have provided the foundation for identifying genetic mechanisms underlying human diseases, including a form of hereditary cancer known as Lynch syndrome^32–36. Hypermutation in fungi has been shown to promote rapid adaptation to novel environmental conditions, to influence overall fitness and genomic structural changes, and to be subsequently tempered by the emergence of anti-mutator suppressor alleles, among other consequences^8,9,37,38.

Recently, naturally occurring fungal hypermutator strains have been identified and characterized. Clinically and environmentally isolated strains of the human fungal pathogens Cryptococcus deuterogattii, Cryptococcus neoformans, and Candida glabrata have been found to harbor loss-of-function mutations in DNA mismatch repair components and thus display hypermutator phenotypes^9,39–43. Defects in the mismatch repair protein Msh2 are especially prevalent and typically produce single base pair insertions or deletions in homopolymeric nucleotide runs^9,44. These isolates display higher rates of genomic evolution at the nucleotide level and have the ability to more rapidly adapt to novel stressors in vitro, including the ability to develop de novo resistance to antifungal drugs in vitro^9,39–42,45. Incompatibilities in the mismatch repair components Mlh1 and Pms1 have also been reported to contribute to hypermutation in S. cerevisiae laboratory strains^46,47. Studies have identified several S. cerevisiae strains isolated from humans that encode these same MLH1-PMS1 incompatible alleles⁴⁸. Although the S. cerevisiae diploids with incompatible alleles isolated from humans were not hypermutators themselves, their meiotic progeny displayed a wide range of mutation rates and included hypermutators, thus providing an opportunity to produce progeny with variable fitness under stressful conditions^47,48. Another mechanism of hypermutation caused by a mutation in a DNA polymerase delta subunit was also described in C. neoformans; this mechanism of hypermutation resulted in genome-wide increases in transition and transversion mutations but reduced viability and virulence^49,50.

Mobilization of transposable elements throughout the genome can mediate hypermutation at a larger scale than the single-nucleotide polymorphism changes in mismatch repair and DNA polymerase mutants. Within the past year, two studies have reported instances of hypermutation via mobilization of transposons under stress conditions, illustrating that host infection can trigger transposition in the human pathogen Cryptococcus deneoformans and in the plant pathogen Zymoseptoria tritici^51,52. Another study has also characterized the evolutionary dynamics and genomic impacts of similar bursts of transposon expansion within the genomes of Microbotryum species⁵³. These instances of hypermutation provide important examples of how elevated genome-wide mutation rates influence rapid genomic microevolution, pathogenesis, and the ability of fungi to rapidly adapt to novel environments.

Despite the fitness benefits of hypermutation in stressful or changing conditions, spontaneous mutations can be largely deleterious in normal conditions, and thus hypermutation might be beneficial only from a short-term evolutionary perspective but disadvantageous in the long term⁵⁴. All organisms, including fungi, must therefore strike a balance between high mutation rates that provide variation for natural selection to act upon and high genomic integrity that allows for long-term survival. Research has shown that, in contrast to hypermutation, some fungal species, particularly those with relatively long lifespans, can have ultra-low mutation rates, such as those observed in the fairy-ring mushroom species Marasmius oreades and the pathogenic forest fungus Armillaria gallica^55,56. On the other hand, some fungi have been found to persist in the long term with high mutation rates and compromised genome integrity. For example, several novel Hanseniaspora species have recently been identified and contributed to the characterization of two Hanseniaspora lineages with different rates of evolution: a fast-evolving lineage and a slow-evolving lineage⁵⁷. While species in both lineages have lost a number of genes involved in regulating the cell cycle and maintaining genomic integrity, the fast-evolving lineage has undergone rapid genomic changes and represents a group of hypermutator species. These species provide a perspective on the long-term evolutionary consequences of sustained elevated mutation rates and could provide answers to critical questions in the field⁵⁷.

Fast-evolving regions of the genome

Genomic evolution can occur at different rates that are dependent on genomic context. Rapid evolution within a genome is driven by the adaptive potential of alleles and relaxed selective constraints as well as the propensity of regions to undergo change. Fungal genomic regions that typically experience faster evolution than other loci include introns, intergenic sequence, centromeres, and subtelomeric regions^58–65.

The evolution of introns throughout fungal lineages is dictated by retention of ancestral introns, frequent intron loss events, and relatively few instances of intron gain^58,59,66,67. Due to the large variation in intron content across fungal genomes, introns are considered to be one of the more rapidly evolving genomic elements in fungi^58,59,66. Spliceosomal twin introns or “stwintrons” were recently identified in several orders of ascomycete species^68,69. These nested introns display intriguing trends of gain and loss across fungi and either are removed by complex, multi-step splicing events or trigger alternative splicing and exon skipping. The evolutionary dynamics of these newly identified stwintrons may provide insight into the evolution of intron gain and loss, thus further informing the known evolutionary trajectory of fungal genomes.

The breadth and depth of RNA sequencing in fungi has revealed that long non-coding RNAs (lncRNAs) are abundant and diverse within fungal genomes, are transcribed from intergenic and genic regions alike, and carry out a number of distinct roles^60,61,70. Only recently have fungal lncRNAs been identified and characterized in fungi other than the model yeasts S. cerevisiae and Schizosaccharomyces pombe. These newly characterized fungal lncRNA transcriptomes show little conservation between species as well as little consensus in post-transcriptional modifications^71–73. Across fungi, lncRNAs have been shown to regulate gene expression in cis and in trans, influencing mating behaviors, sexual development, morphogenesis, cellular metabolism, virulence, drug resistance, and genomic stability^61,70,74. One class of lncRNAs are antisense lncRNAs (ASlncRNAs), which overlap with the open-reading frames of genes in an antisense orientation and regulate expression of their cognate transcripts. It has been shown that loss of RNAi in fungi is correlated with expansion and increased abundance of ASlncRNAs, suggesting that these elements play an important role in modulating gene expression and therefore in the stability and evolution of fungal genomes^75–77. Additional deep RNA sequencing among numerous and diverse fungi complemented with molecular genetic analyses will provide further insight into the diversity of lncRNAs and their structure, function, evolution, and impact on overall genomic evolution.

Despite having a conserved role in proper chromosome segregation, centromeres are one of the most rapidly evolving sequences among eukaryotes⁶². Previous studies focused on centromere identification in relatively few ascomycete fungal species, including S. cerevisiae, S. pombe, and Candida albicans^78,79. These studies revealed that centromere length in fungi could range from a few hundred nucleotides to 300 kb in length and varied significantly among related species. These studies also established that centromeres are defined by the binding of a conserved histone H3 variant, CENP-A, along with many other conserved kinetochore proteins. Over the past several years, whole-genome sequencing with long-read Nanopore and PacBio sequencing platforms has led to the identification of centromeres in many other fungi and a better understanding of centromere structure and evolution. One such study revealed that centromeres across the Schizosaccharomyces group harbor a similar architecture but vary in their sequence, suggesting sequence is not critical for centromere function²⁵. Centromeres in the plant-pathogenic ascomycete Magnaporthe oryzae were found to consist of long AT-rich sequences, a feature conserved in related species²⁶.

Characterization of centromeres in the Cryptococcus pathogenic species complex revealed that centromeres in Cryptococcus are long, rich with retrotransposons, and syntenic among different species^80,81. However, centromere lengths vary significantly among three pathogenic species in this complex due to the truncation of retroelements as a result of RNAi loss in one of the species⁸¹. A more recent study in another basidiomycete group of Malassezia species identified short centromeres (less than 1 kb), which were highly AT-rich and devoid of any repeat content⁸². Following the pattern in Schizosaccharomyces and the Cryptococcus species complex, centromeres in Malassezia are also syntenic between closely related species.

The nature of centromeres in the Mucoromycota phyla of fungi have also been characterized using Mucor circinelloides, an emerging model organism⁸³. It is important to note that M. circinelloides lacks CENP-A but harbors other conserved kinetochore proteins. Identification of centromeres in this species revealed a mosaic nature of centromeres where kinetochore binding was restricted to a small region, flanked by multiple copies of a retrotransposon. Each of these studies emphasizes the fact that fungal centromeres are highly diverse with few conserved features and are rapidly evolving, even among closely related species. Interestingly, all centromeres identified thus far are in ORF-free regions that are mostly devoid of transcription, and many centromeres are AT-rich in nature. Combining these features together might help in predicting centromeres in fungal species where experimental tools have not been applied or are not yet available.

Other genomic regions in which rapid evolution has been observed are the subtelomeres and the repeat-rich regions of plant pathogens that harbor effector genes mediating virulence. Subtelomeric loci are prone to rapid expansion and contraction and often encode virulence-associated genes as well as biosynthetic gene clusters for many secondary metabolites. The impact of rapid subtelomeric evolution on virulence has been highlighted in the human fungal pathogens Candida auris and Aspergillus fumigatus as well as in the plant fungal pathogen Pyrenophora teres f. teres^63–65. Interestingly, in the emerging fungal pathogen C. auris, a particular genetic lineage that has not been involved in outbreaks exhibits high mutation rates and genomic instability, especially in subtelomeric regions. This subtelomeric variation is hypothesized to be attributable to a loss-of-function mutation in DCC1, a gene involved in regulating telomere length, genome instability, and the cohesion of sister chromatids in yeast^64,84. Similar to subtelomeric regions, the regions harboring effector genes, also known as effector compartments, in fungal plant pathogens are also prone to frequent expansion and contraction events^10,11. These regions are enriched in transposable elements and are regularly located near chromosome rearrangement breakpoints, which contributes to their largely lineage-specific identity¹¹. Effector compartments make up the rapidly evolving portion of the “two-speed” genomes of many plant fungal pathogens. Rapid evolution in these loci is attributed to the expansion, contraction, and recombination mediated by repetitive transposons within the region as well as repeat-induced mutation (RIP) and epigenetic silencing¹⁰.

Allelic transmission

The transmission of alleles between species can have a profound impact on genomic evolution. Instances of inter- and intra-kingdom HGT, genetic introgression following hybridization, and meiotic drive elements have all been shown to be strong forces of genomic evolution and have thus received increased attention over the last decade. Robust whole-genome sequencing of isolates within a species as well as the increasing diversity of fungal species sequenced have allowed for the identification of these allelic transmission events and mechanisms and also shed light on their origins^{12–14,85–91}.

HGT refers to the lateral exchange and integration of genetic material between two species, unlike the vertical inheritance pattern between parent and offspring⁹². Previously, prokaryotic species were thought to experience frequent HGT events, while relatively few HGT events were thought to have occurred in fungi. However, advances in whole-genome sequencing have, by increasing efficiency and reducing costs, expanded the number of fungal genome sequences available and allowed the identification of numerous HGT events across the fungal kingdom; many instances of HGT events in fungi have been extensively characterized and reviewed^16–19. HGT events appear to be more frequent in organisms that have undergone significant niche adaptation such as commensal organisms like the saprophytic skin commensal Malassezia species, which have acquired a bacterial flavohemoglobin as well as many other bacterially derived genes⁹³. Similarly, obligate intracellular pathogens belonging to the Microsporidia and Cryptomycota phyla have experienced significant HGT, in which up to 2% of genes are estimated to be derived from HGT events⁸⁵. This work has helped to markedly clarify the number of HGT events in Microsporidia because only a small number of HGT events involving bacteria- and animal-derived genes had been previously documented, and these events were thought to be rare^16,94–96. In addition to the relatively common HGT of bacterially derived genes in fungi, an exciting new study has identified instances of trans-kingdom genetic transfer between an arbuscular mycorrhizal fungal species and its symbiotic plant partner⁹⁷. Another case of trans-kingdom HGT was shown in basal fungi (Chytrids) where a key cell-cycle regulator protein evolved from a viral protein⁸⁶. This study provides key insights into the evolution of cell-cycle regulating machinery that evolved differently in fungi and animals. Further characterization of genomic diversity within and among species across all domains of life will help to identify additional instances of HGT.

Introgression, also known as introgressive hybridization, refers to the integration of genetic alleles from one genetic lineage into another following hybridization. Population-based studies in fungi have revealed the high frequency with which introgression can occur. Two recent examples include 1) the loss and regain of fermentation capacity by Kluyveromyces lactis, which was demonstrated to be a result of introgression of a subtelomeric locus from Kluyveromyces marxianus, and 2) finding that a lichen-forming Rhizoplaca species likely originated from an introgressive hybridization event^12,13. In another intriguing study, introgression following hybridization was shown to be the primary force of evolution in the fungal pathogen responsible for Dutch elm disease, driving changes in host–pathogen interactions via introgression of virulence-associated genes and changes in sexual reproduction via introgression of mating-type loci¹⁴.

Population genomics approaches have also revealed the prevalence of meiotic drive elements throughout the fungal kingdom. Meiotic drive elements are genetic elements that cause non-Mendelian patterns of inheritance by either biasing their inheritance in sexual progeny or by killing progeny that do not inherit the element, also known as killer meiotic drive elements. Fungal meiotic drive elements were first identified in Neurospora many decades ago⁹⁸. Subsequent studies have further characterized the meiotic drive elements in Neurospora, known as Spore Killers, and have revealed other Spore Killer meiotic drive elements in Podospora^{87,88,99–102}. Over the past several years, elegant and thorough research in Schizosaccharomyces species has identified several novel meiotic drive elements^{15,89,90,103,104}. Studies have also shown that in addition to the genomic loci that can act as meiotic drive elements, supernumerary chromosomes in the plant pathogen Z. tritici show evidence of meiotic drive⁹¹. The ability of genetic elements to bias their inheritance without conferring any adaptive benefits or potentially reducing fitness can greatly influence the genomic landscape and evolutionary trajectory of a species.

Chromosome structural variations

Structural variations or chromosome rearrangements occur when chromosome segments are shuffled across the genome. These structural variations include deletions, inversions, duplications, and translocations. Chromosome rearrangements are known to occur in all eukaryotic genomes^105–107. These rearrangements can occur within a single chromosome (intra-chromosomal) as well as across different chromosomes (inter-chromosomal). Many different mechanisms are known to mediate rearrangements, including meiotic recombination, double-stranded break (DSB) repair during mitosis, and transposable elements^108,109.

The prevalence and mechanisms of chromosome rearrangements in fungi have been known for a long time, but recent advances in long-read sequencing have highlighted the importance of these events in multiple cellular processes. Several studies comparing the genome evolution of closely related species revealed that chromosome translocations can drive the evolution of secondary metabolite-producing gene clusters, and many of these translocations coincide with transposable elements^110–112. Some of these rearrangements are also correlated with gene loss, the generation of genetic diversity, and changes in pathogenicity¹¹³. A study in the basal Taphrina genus showed that chromosome rearrangements can drive the evolution of effector superfamily genes and host adaptation as well as speciation¹¹⁴. Another study comparing the genomes of 71 ascomycete species suggests that the rate of chromosome rearrangements accelerated the process of speciation in these species¹¹⁵. A comparison of two different subphyla showed that the number of rearrangements is directly correlated with species richness in each respective subphylum. The same study also showed that whole-genome duplication and pathogenicity might elevate gene order divergence and genome rearrangements¹¹⁵.

Using a direct experimental approach, we recently showed that centromere-mediated chromosome rearrangements, induced by targeting DSBs at centromeric retrotransposons with CRISPR, can alone drive reproductive isolation and thus play a role in speciation¹¹⁶. While centromere-mediated translocation has been proposed to occur based on genome comparisons, we showed its direct impact on the sexual cycle. Such centromere-facilitated chromosome rearrangements have been observed in Malassezia and Candida and are proposed to play a role in both chromosome and karyotypic evolution in these species^82,117,118. Centromere-mediated chromosomal translocation has also been proposed as a driving force during the evolution of the mating-type locus in Cryptococcus from an ancestral outcrossing tetrapolar system to a derived bipolar inbreeding state⁸⁰.

Combined, these studies show that structural variations play a critical role in the evolution of a diverse array of functions in fungal species, including metabolism, pathogenicity, host adaptation, and speciation. Their role in the evolution of secondary metabolite-producing clusters has been observed in multiple fungal species complexes, suggesting that this is a common mechanism via which metabolic potential can be altered or expanded. Additionally, the role of chromosome translocation in speciation has received significant support as a result of these studies. Future studies will reveal the unexplored significance of these chromosome structural changes in fungi. One key aspect to focus on will be the impact of these changes on pathogenicity and how they contribute towards the evolution of virulence in fungal pathogens.

Polyploidy

Whole-genome duplication or polyploidy refers to scenarios in which a cell or organism possesses three or more complete sets of chromosomes compared to the standard haploid karyotype. Polyploidy is found across the fungal kingdom, and one route to its generation involves whole-genome duplication^23,24. Indeed, analysis of the whole-genome sequence of Saccharomyces revealed that one such whole-genome duplication event was responsible for the evolution of this genus and likely involved an interspecies hybridization event^30,119. Despite the presence of polyploidy across the fungal kingdom, studies analyzing the polyploidy of fungal genomes have started to re-emerge²³. Many clinically relevant fungi show variation in ploidy, mainly in response to stressful or new environmental conditions. Some fungi can also exist in multiple different ploidy states and exhibit ploidy transitions from one state to another. Studies in S. cerevisiae, as well as C. albicans, found that ploidy variation can drive rapid adaptation in response to environmental conditions and vice versa^120,121.

A well-studied process of ploidy transition is the parasexual cycle in C. albicans¹²². During this process, tetraploid C. albicans cells undergo ploidy reduction to generate more stable diploid cells, which can be selectively advantageous¹²³. This process is accompanied by concerted chromosome loss and generates aneuploidy¹²⁴. Recent studies have provided intriguing insights into this process. Tetraploid C. albicans cells were found to be metabolically hyperactive, generating reactive oxygen species and DNA DSBs, resulting in increased genomic instability¹²⁵. Another study found that the genome reduction in this parasexual cycle involves recombination, similar to meiosis observed in other organisms, and thus has been termed para-meiosis¹²⁶. A closely related species, Candida tropicalis, was found to exhibit a combination of same- and opposite-sex mating to generate polyploidy, which in turn increases genetic diversity¹²⁷.

Another human fungal pathogen, C. neoformans, has been shown to exist in haploid or hybrid diploid states. C. neoformans can also exist in a polyploid state, known as titan cells, produced in response to host conditions^128,129. These titan cells produce haploid and aneuploid progeny and thereby enhance stress adaptation¹³⁰. A concerted effort from three independent studies established that this phenotype can be induced by multiple environmental as well as genetic signals^131–133. Similar to C. albicans, the ploidy reduction in these titan cells is mediated by the activation of meiotic genes¹³⁴. While these studies explore the reduction of a polyploid genome to a more stable state, the exact role of polyploidy and the factors responsible for inducing polyploidy are not well understood. The presence of different levels of ploidy in S. cerevisiae natural isolates hints toward its role in adaptation to environmental conditions, but such natural resources have not been explored in other fungal species yet.

Aneuploidy

Aneuploidy refers to the stage when a cell deviates from its balanced genome by gaining or losing chromosomes. Aneuploidy has been described in a large number of fungi but mainly with respect to stress conditions. While aneuploidy is commonly thought of as a deleterious state, and in humans is invariably associated with disease states or is lethal, many studies have found naturally occurring aneuploid isolates in different fungal species^20,24. It has been proposed that aneuploidy may represent an intermediate, transient state that can promote evolutionary adaptation^135,136. Experimental evolution of multiple S. cerevisiae strains showed that aneuploidy frequently occurs during the course of evolution¹³⁷. Similar results were observed when 47 non-laboratory isolates of S. cerevisiae were analyzed¹³⁸. The gene SSD1 and its contributions to mitochondrial physiology were identified as the genetic basis of aneuploidy tolerance in these natural isolates¹³⁹. A genome-wide genetic deletion screen in S. cerevisiae found that genes involved in a ubiquitin-mediated proteasomal degradation pathway also confer tolerance to aneuploidy^140,141.

Aneuploidy has been previously known to play a crucial role in drug resistance in pathogenic fungi like C. albicans and C. neoformans^142–144. In both cases, drug resistance involved amplification of fluconazole target genes like ERG11 and genes responsible for the expression of drug efflux pumps, either by isochromosome formation in C. albicans or aneuploidy in C. neoformans. Newer studies have shown that aneuploidy can play roles in osmotic stress, flocculation, and ethanol survival as well as responses to starvation conditions in S. cerevisiae^145–147. We have also shown that aneuploidy can be essential to overcome reproductive barriers between two strains harboring chromosomal translocations¹¹⁶. Overall, aneuploidy seems to be a conserved fungal response to stress or harsh environmental conditions. However, the advances discussed here reveal a role for aneuploidy in adaptation to growth conditions as well as pathogenicity in fungal species. With their well-characterized role in drug resistance, mechanisms leading to aneuploidy also present potent targets for drug development.

B chromosomes

Accessory, supernumerary, or B chromosomes are all different names for extra chromosomes that are present in a cell beyond the reference karyotype. These extra chromosomes segregate in a non-Mendelian fashion, are dispensable, and do not undergo recombination or pairing with the main, essential (also known as A) chromosomes^21,22. Accessory chromosomes are well studied in animals and plants but have also been observed in several plant fungal pathogens. Earlier reports suggested that these chromosomes are required for virulence and can be horizontally transmitted^148,149. More recent studies have focused on the characterization of these chromosomes at the molecular level in Z. tritici and M. oryzae. In Z. tritici, B chromosomes were found to be epigenetically indistinguishable from the core set of chromosomes, whereas the B chromosome (also known as a mini-chromosome) in M. oryzae was found to be more enriched with transposons relative to the core chromosomes^150,151. B chromosomes in both of these species were also found to harbor active centromeres^26,150. However, the B chromosomes in Z. tritici are unstable and frequently lost during vegetative growth¹⁵².

Interestingly, the stability of B chromosomes also depends on the host genotype and thus plays a vital role in host–pathogen interactions¹⁵³. The mini-chromosome in M. oryzae was proposed to be involved in the shuffling of AVR effector genes between the mini-chromosome and core chromosomes, thus aiding in host adaptation¹⁵¹. While the exact function of genes present on these accessory chromosomes is still unknown, they significantly influence the pathogenicity of the fungi in which they are present. For this purpose, an in-depth characterization of accessory chromosomes is essential and should include genetic as well as epigenetic approaches. The identification of B chromosomes in more species will advance the field and may reveal new insights into their roles in pathogenesis, adaptation, evolution, and possibly speciation.