Molecular analyses identify hybridization‐mediated nuclear evolution in newly discovered fungal hybrids

Abstract Hybridization may be a major driver in the evolution of plant pathogens. In a high elevation Alpine larch stand in Montana, a novel hybrid fungal pathogen of trees originating from the mating of Heterobasidion irregulare with H. occidentale has been recently discovered. In this study, sequence analyses of one mitochondrial and four nuclear loci from 11 Heterobasidion genotypes collected in the same Alpine larch stand indicated that hybridization has increased allelic diversity by generating novel polymorphisms unreported in either parental species. Sequence data and ploidy analysis through flow cytometry confirmed that heterokaryotic (n + n) genotypes were not first‐generation hybrids, but were the result of multiple backcrosses, indicating hybrids are fertile. Additionally, all admixed genotypes possessed the H. occidentale mitochondrion, indicating that the hybrid progeny may have been backcrossing mostly with H. occidentale. Based on reticulate phylogenetic network analysis by PhyloNet, Bayesian assignment, and ordination tests, alleles can be defined as H. irregulare‐like or H. occidentale‐like. H. irregulare‐like alleles are clearly distinct from all known H. irregulare alleles and are derived from the admixing of both Heterobasidion species. Instead, all but one H. occidentale alleles found in hybrids, although novel, were not clearly distinct from alleles found in the parental H. occidentale population. This discovery demonstrates that Alpine larch can be a universal host favouring the interspecific hybridization between H. irregulare and H. occidentale and the hybridization‐mediated evolution of a nucleus, derived from H. irregulare parental species but clearly distinct from it.

are all mechanisms known to have generated new pathosystems. The same drivers above are also known to be responsible for the recent comingling of allopatric pathogens or of pathogens with differential host preference, until recently distinctively characterized by nonoverlapping ranges. The movement of pathogenic species with nonoverlapping geographic ranges is particularly relevant because allopatric species are often reported to have maintained the ability to successfully mate with congeneric ones, facilitating the generation of interspecific hybrids (Kohn, 2005).
Indeed, a crucial role in the generation of novel plant pathogens has been ascribed to hybridization between species, a process in which the combination of two or more genomes in a single organism can lead to rapid adaptive evolution (Brasier, 2000;Depotter et al., 2016;McDonald & Stukenbrock, 2016;Olson & Stenlid, 2002;Schardl & Craven, 2003;Stukenbrock, 2016aStukenbrock, , 2016b. Some of the more interesting examples of microbial hybridization include tree pathogens, for example, Phytophthora × alni (Brasier, Cooke, & Duncan, 1999), Melampsora × columbiana (Newcombe, Stirling, McDonald, & Bradshaw, 2000), and Ophiostoma novo-ulmi (Brasier & Buck, 2001). The evolutionary outcomes of interspecific hybridization among plant pathogens are still far from being clearly deciphered; however, a growing body of evidence supports the notion that plant pathogens may develop greater virulence through rapid evolution mediated by hybridization events (Brasier, 2000;Stukenbrock, 2016a). According to this scenario, hybrids can be either transient, but still acting as ephemeral "genetic bridges" between parental species (Brasier, 2000), or they can be viable, as fit as parental individuals, and repeatedly generated as part of hybrid swarms (Gonthier & Garbelotto, 2011). Hybrid swarms that are not reproductively isolated from parental populations and that include ecologically fit and fertile first-generation hybrids are bound to represent the first step in a process leading to populations comprising individuals with genomes admixed between the two parental species (Gonthier & Garbelotto, 2011). On the other hand, hybridization followed by reproductive isolation has been reported to contribute to rapid speciation in yeast (Leducq et al., 2016), and the same has been hypothesized to occur in some filamentous fungi (Gladieux et al., 2014;Kohn, 2005) and in fungus-like oomycetes (Schardl & Craven, 2003).
In North America, H. irregulare generally attacks pines and junipers (Juniperus spp), while the host range of H. occidentale comprises Abies, Picea, Tsuga, Pseudotsuga, Sequoia, and Sequoiadendron (Garbelotto & Gonthier, 2013). Heterobasidion irregulare is present throughout North America, whereas H. occidentale is only present in Western North America (Garbelotto & Gonthier, 2013). Even when both species coexist in the same region, they are often found in different stands due to their different host preference. When they are found in the same site, they are normally partitioned on different hosts; their true comingling appears to be closely associated with the creation of stumps through logging, a practice that has allowed for the establishment of both species on the same substrate (Garbelotto & Gonthier, 2013). Primary infection and colonization of new forest stands is in fact affected by basidiospores on fresh woody surfaces, such as stumps (Rishbeth, 1959). Basidiospores germinate and the fungus saprobically colonizes the stump including its root system, as well as the roots systems of adjacent individuals, thus infecting neighboring standing trees and more stumps (Garbelotto & Gonthier, 2013). As a result of root-to-root secondary infection, Heterobasidion-induced tree mortality appears in groups known as root disease centers that progressively expand in time (Garbelotto & Gonthier, 2013). Human activities have favored the spread of the fungus not only by creating the primary infection substrate, that is, fresh stumps, but also by excluding fires favoring a change in tree species composition and by decreasing/arresting timber harvest operations, thus increasing stand density, which favor tree-to-tree contagion.
Heterobasidion species are known to retain a certain degree of interfertility. In vitro experiments have indicated the rate of observed interfertility ranges about 5%-98% depending on the species combinations (Garbelotto & Gonthier, 2013;Harrington, Worrall, & Rizzo, 1989;Korhonen & Stenlid, 1998). Hybridization processes in nature have also been documented to occur between pairs of taxa within the species complex (Garbelotto & Gonthier, 2013). The generation of hybrids in Heterobasidion spp. in nature, as in the majority of basidiomycetous fungi, can be achieved by plasmogamy of two haploid (n) mycelia of interfertile species, which can generate a heterokaryotic mycelium (n + n) characterized by the co-occurrence of haploid nuclei from both parental species in the same cell. Since karyogamy is delayed, the generated heterokaryotic mycelium often represents the main growth phase of the hybrid isolate. Only once it is well established on a substrate and when environmental conditions are favorable, the fruiting bodies, in which karyogamy and meiosis occur, may be produced. Meiosis produces haploid (n) basidiospores responsible for the infection of new stumps.
In 1996, a natural hybrid genotype between H. irregulare and H. occidentale was found on a ponderosa pine (Pinus ponderosa Laws.) stump in California and in adjoining western juniper (Juniperus occidentalis Hook.) and ponderosa pine trees (Garbelotto, Ratcliff, Bruns, Cobb, & Otrosina, 1996). Both parental species were also isolated from the same ponderosa stump. The hybrid was regarded as a first-generation hybrid due to the presence of both H. irregulare and H. occidentale isozyme alleles at each of 10 loci (Garbelotto et al., 1996). Surprisingly, later experimental evidence suggested that F1 Heterobasidion hybrids may be diploids (2n) rather than heterokaryons (n + n ploidy), indicating that the first step of hybridization affected ploidy (Garbelotto, Gonthier, Linzer, Nicolotti, & Otrosina, 2004), as reported for many plant and animal hybrids (Mallet, 2007).
The discovery of both parental species in many stumps across California (Otrosina, Chase, & Cobb, 1992) and of a large hybrid genotype in a stump (Garbelotto et al., 1996) has led to the hypothesis that H. irregulare × H. occidentale hybridization may require either a universal host or a common substrate where both Heterobasidion species can thrive (Garbelotto et al., 1996). The hypothesis that a novel "common" substrate or host may be necessary for hybridization to occur was further reinforced by the finding that fitness of natural and artificial Heterobasidion hybrids is reduced on hosts preferred by each parental species . In California, simultaneous colonization of ponderosa pine stumps by both species (Garbelotto et al., 1996) (Linzer et al., 2008) suggests that hybridization must have predated the era of commercial logging (e.g., the mid 1800s) and must have occurred in the absence of stumps.
Almost 15 years after the first discovery of a natural hybrid  (Lockman, Mascheretti, Schechter, & Garbelotto, 2014). This second report of a hybrid between H. irregulare and H. occidentale suggested that Alpine larch may be a host for both Heterobasidion species, as described for naturally infected ponderosa pine stumps in California and for artificially inoculated Sitka spruce (Picea sitchensis [Bong.] Carr.) seedlings . Because no logging has occurred at this high elevation site, the finding potentially provided an opportunity to study natural interspecific hybridization in Heterobasidion in North America independent of direct anthropogenic effects. Additionally, if indeed hybridization has been ongoing in these high elevation forests at a significant rate, the study of naturally formed genotypes with admixed genomes may allow to understand the evolutionary implications of hybridization, a topic that has received relatively little attention for the entire fungal kingdom.
This study describes the result of a second more exhaustive sampling of two mortality centers in the same infested Alpine larch stand where the original Lockman et al. (2014) finding had occurred.
Sampling included direct isolation of Heterobasidion genotypes from infected wood, fruiting bodies, and airspora collected on woody spore traps. Sequence analysis of one mitochondrial and four nuclear loci was then performed on all Heterobasidion genotypes obtained through such sampling scheme.
The aims of this study were as follows: a To collect multiple genotypes and determine their genetic makeup (e.g., pure vs. admixed nuclei and mitochondrial type); b To determine which genomic makeup may be dominant in the area; c To genetically characterize any admixed genotype and determine the level and parental origin of the admixture; d To determine whether hybridization may have resulted in an increase in genetic variability.

| Study site and sampling
In September 2014, two distinct root disease centers were sampled in a stand of the Bitterroot Mountains, on the shores of Gem Lake, south of Darby, Montana (elev. 2,530 m;Lat. 45.893528°, Long. −114.278322°). One site was the same where the first hybrid Heterobasidion fruiting body had been original collected (Lockman et al., 2014), the second was adjacent to the first, but clearly separated from it by a treeless buffer, approximately 100 m wide. Each site corresponded to a classical Heterobasidion root disease center (Garbelotto & Gonthier, 2013) characterized by dead, dying, and symptomatic larch trees, roughly circular in shape, and with a diameter of about 50 m each.
In each root disease center, several adjacent Alpine larches (L. lyallii) were either dead or displayed thin crowns ( Figure 1).
The stand is mature, and in the two study sites, 79% of trees were Alpine larches, 17% were whitebark pines (Pinus albicaulis Engelm.), and 4% were subalpine firs (Abies lasiocarpa [Hooker] Nuttall), but only larches were symptomatic. Approximately 43% of the larches were dead, 22% were in a stage of advanced decline, and 35% were healthy. The stand immediately below the one surveyed was characterized by an overwhelming majority of subalpine firs.
In each disease center, isolations were performed directly from fruiting bodies or wood symptomatic for decay, and by subculturing Heterobasidion colonies growing on woody traps. A total of fifteen traps per site were placed in groups of three at 10-m intervals along a 50-m-long linear transect and exposed to the air for a total sampling period of 24 hr.
In order to isolate fungal mycelia from wood or fruiting bodies, fragments of wood, and fruiting bodies 25 mm 2 in size were surface sterilized in laboratory by using 70% Ethanol and placed on 9-cm Petri dishes containing MEA (dilute Malt Extract Agar; 2 g malt extract, 10 g glucose, 2 g peptone, 20 g agar, 1 L distilled water, amended with 0.05 g/L ampicillin). Purification of isolates from contaminants was performed under a dissecting microscope (20× magnification) by transferring putative mycelium of Heterobasidion spp. from the original Petri dishes onto sterile Petri dishes filled with MEA.
In order to collect isolates from aerial spores, spores of Heterobasidion spp. were trapped using the wood-disk exposure method as previously described (Gonthier, Nicolotti, Linzer, Guglielmo, & Garbelotto, 2007). Briefly, wood disks of 11-13 cm diameter were individually placed in 15-cm Petri dishes containing sterile filter papers dampened with 3.5 ml of sterile water, to prevent drying during exposure. A total of 45 open Petri dishes were placed along three transects with five trapping points on the ground at 10 m from one another, each including three woody traps placed at 1 m from one another at the vertices of an imaginary triangle, for a total of five trapping points per transect. After a 24-hr exposure, filter papers were replaced and dampened again with 3 ml of sterile water. Wood disks were sprayed with a benomyl solution (0.010 g benomyl, 500 μL methanol, and 1 L sterile water) and incubated at 24°C for 15 days. Isolations were made under a dissecting microscope (20× magnification) by transferring colonies of Heterobasidion spp. in its conidial stage (Spiniger) onto Petri dishes filled with MEA and amended with 0.3 g/L streptomycin (Kuhlman & Hendrix, 1962).
Spore load was calculated as spores/m 2 per hour (spores/m 2 h) according to Gonthier, Garbelotto, and Nicolotti (2005). All obtained isolates were subsequently grown at 24°C on 6-cm Petri dishes filled with MEA.

| DNA content analysis by flow cytometry
DNA contents were measured by flow cytometry for five Heterobasidion isolates from wood samples randomly chosen. The isolate (ID: Awr400) identified in 1996 by Garbelotto et al. (1996) as a first-generation hybrid between H. irregulare and H. occidentale was included in the analysis as a putative diploid control. Heterobasidion isolates were taken from MEA slants and grown in test tubes containing 5 ml filter-sterilized 5% clarified V8 broth (Englander & Roth, 1980) for 3 days at 21°C. The mycelium was checked for the absence of conidia under the dissecting microscope, then harvested, and washed three times with sterile water. Arabidopsis thaliana Col-0 and Aspergillus fumigatus CEA10 (FGSC A1163) were chosen as the internal DNA reference standards with a genome size of 1C = 157 Mb (Dolezel & Bartos, 2005) and 1C = 29.7 Mb (Veselska, Svoboda, Ruzickova, & Kolarik, 2014), respectively. A modified protocol derived from Bertier, Leus, and D'hondt, de Cock, and Höfte (2013) was then followed. In brief, extraction of nuclei was done using the  com/solut ions/flowjo), and DNA content was subsequently inferred using a quadratic regression with the ratios between the peak positions of the Heterobasidion sample and the two standards.

| DNA extraction and PCR conditions
For DNA extraction, isolates were grown in 250-ml flasks filled with malt extract broth (2% w/v), at room temperature and in the dark Five genomic loci were selected as markers to conduct the phylogenetic analysis. These five loci have proven to be reliable for phylogenetic studies of the genus Heterobasidion, and have all been widely used in the past (Chen, Cui, Zhou, Korhonen, & Dai, 2015;Dalman, Olson, & Stenlid, 2010;Johannesson & Stenlid, 2003;Linzer et al., 2008). The markers were the mitochondrial ATPase subunit 6 (atp6), the glyceraldehyde-3-phosphate dehydrogenase (gpd), the RNA polymerase II 2nd largest subunit (RPB2), the translation elongation factor 1-alpha (EF-1α), and the nuclear ribosomal internal Note: Multiple alleles were named by using letters. The tags "Hi" and "Ho" represents novel "Heterobasidion irregulare-like" and "Heterobasidion occidentale-like" alleles, respectively, while the tag "Hocc" means "H. occidentale" allele. Polymorphisms related and unrelated (private) to putative parental species over total number of detected SNPs (referred to putative parental species) are also indicated.

| Sequencing and phylogenetic analyses
The purified PCR products of atp6, gpd, and RPB2 amplicons were Sanger-sequenced in house at the Forest Pathology and Mycology Laboratory (Berkeley, USA) and at BMR Genomics S.R.L. (Padua, Italy). The 96 purified cloned products of EF-1α and ITS were also Sanger-sequenced in the same laboratories. All amplicons were forward and reverse sequenced with the related primers, and consensus sequences were generated by using the Geneious software, version 9.0.5 (Biomatters, Ltd). Chromatograms of each sequence were analyzed by using both Geneious and SnapGene® Viewer.
The minimum acceptable Phred score considered per base was 20.
For the two not-cloned loci, that is, gpd and RPB2, ambiguous bases showing two overlapping peaks with equal signal intensity were assigned manually based on the allele frequencies of the putative parental species, that is, H. irregulare and H. occidentale. When a SNP in the heterozygotic sequence could not be assigned as being derived from one of the two parents, it was marked as the same ambiguous nucleotide on both alleles. This process allowed to extract with confidence the two homozygotic alleles, each derived from one of the two parents, from uncloned heterozygotic sequences.
For each locus, a multiple sequence alignment was built using the repetitions were performed, with the parameter set as "Admixture Model" and "Allele frequencies independent," and without any prior information of the origin of individual samples. The determination of ΔK based on the highest likelihood of the data (LnP(D)) was used to infer the K value best representing the observed data under the implemented model (Evanno, Regnaut, & Goudet, 2005) using Structure Harvester Web v0.6.94 (Earl & vonHoldt, 2012).
In order to test sequences from Montana Larch isolates for signs of hybridization and introgression events, the software PhyloNet (Than, Ruths, & Nakhleh, 2008)

| Sampling and obtaining of fungal isolates
After sampling and direct isolation, a total of seven heterokaryotic (ploidy = n + n) isolates were obtained from wood samples and one heterokaryotic isolate was obtained from a fruiting body. Six of them were used for molecular analyses (Table 1). Five homokaryotic (ploidy = n) colonies were isolated from spore traps and also used in the molecular analyses; four were pure H. occidentale, and one was admixed (Table 1).
The spore load resulted as high as 18 spores/m 2 h. All isolates from wood samples, fruiting bodies and from spores used in the molecular analyses were deposited at the Mycotheca Universitatis Taurinensis (MUT). Their related accession numbers are listed in Table 1.

| Chromosome analysis by flow cytometry
Flow cytometry with two size standards, that is, A. fumigatus and A. thaliana gave a haploid genome estimate of 34.7 Mbp ( Figure S1), a result in good agreement with the genome assembly size of H. irregulare (33.1 Mbp) (Olson et al., 2012). All five isolates showed two main peaks, with sizes corresponded to 1C (unreplicated haploid genome at G1 cell cycle phase) and 2C (replicated haploid DNA at G2 and M phases; Figure 2). On the other hand, a first-generation hybrid isolate Awr400 21B, which a previous study suggested may be diploid (Garbelotto et al., 2004), showed two peaks corresponding to 2C and 4C of diploid cells.

| Sequence analysis
Sequencing of amplicons resulted in a total of 11 atp6 sequences for the 6 heterokaryotic and the 5 homokaryotic isolates. After sequence trimming, 445 nucleotides per atp6 sequence were aligned.
No ambiguous bases were detected in this set of sequences. All sequences were identical, showing no SNPs within them (Table 2).
Eleven gpd sequences, one for each isolate, were obtained and in total four different alleles were identified from heterokaryons (ploidy n + n) while one additional allele was identified from the homokaryons (ploidy n).  Figure   S3). As for gpd, no ambiguous bases were observed in the five identical sequences from the five homokaryotic isolates ( As shown in Table 2, by comparing sequences of hybrids with those of parents, it was possible to identify two significant mechanisms resulting in de novo nucleotide polymorphisms: (a) SNPs were derived from either parent through hybridization creating a new admixed sequence and (b) SNPs were novel and not derived from either parent, suggesting rapid evolution occurring during hybridization but not directly derived by admixing of the parental alleles.

| Phylogenetic analyses
Alignments of mitochondrial (atp6) and nuclear (gpd, RPB2, EF-1α, and ITS) gene sequences were used to infer the phylogenetic placement of the 11 Heterobasidion genotypes collected from wood samples, fruiting bodies, and spore traps.
Phylogenetic analysis of atp6 sequences indicated that all 11 isolates exclusively possessed the H. occidentale mitochondrial genome ( Figure S4). Conversely, trees obtained from heterozygotic gpd and RPB2 sequences, which harbored ambiguous bases, showed heterokaryotic isolates as being placed in an intermediate position in between reference sequences from the putative parental species H. irregulare and H. occidentale (Figures S5 and S6). However, when ambiguity in alleles was resolved by manual assignment of bases based on parental allele frequencies, the alignment of sequences showed that each heterokaryotic isolate possessed two distinct types of gpd and RPB2 alleles, one clearly related to H. irregulare and one clearly related to H. occidentale (Figures S2 and S3).
The phylogenetic analysis of EF-1α cloned sequences showed that all heterokaryotic isolates possessed at least one H. occidentale allele ( Figure S7). The other alleles found in the isolates grouped together as a separate, statistically well supported, H. irregulare-like cluster sister to the H. irregulare cluster ( Figure S7).
In addition, sequences of isolates MH3006 and MH3007 formed a distinct group but nested within the H. irregulare-like cluster ( Figure S7). Sequences from homokaryotic genotypes from spores grouped with the H. occidentale cluster, with the exception of the sequence of the genotype III2B, which grouped within the H. irregulare-like subcluster and together with the heterokaryotic larch genotypes ( Figure S7).
Topology of the tree based on ITS cloned sequences was similar to that obtained with EF-1α sequences, with a statistically well-supported H. irregulare-like cluster distinct from H. irregulare including one allele from each heterokaryotic isolate ( Figure S8). However, each heterokaryotic isolate also had at least one H. occidentale allele ( Figure S8). Three sequences of isolate MH3006 and one sequence of MH3007 did not cluster with the others but remained unresolved between the two large H. occidentale-and H. irregulare-like clusters.

| Multidimensional scaling, Bayesian, and PhyloNet analysis
The

| Heterobasidion genotypes from Alpine larch are heterokaryotic hybrids harboring two different nuclei, but a single mitochondrial genome
Our survey of two distinct root disease centers in a high elevation mixed forest stand dominated by Alpine larch indicated that, in both instances, disease and mortality of Alpine larch were associated with infection by Heterobasidion. Unexpectedly, all fungal genotypes infecting larch were heterokaryotic (n + n) hybrids that included pairs of nuclei of different origin. These nuclei harbored novel and unreported nuclear alleles, but a comparison with known DNA sequences showed that in these hybrids, one nucleus pos-  (Garbelotto & Gonthier, 2013), thus providing the necessary habitat for interspecific mating to occur, and (c) Hybrid genotypes may be at an advantage when colonizing larch.
Although further research needs to confirm the three hypotheses above, the hybrid genotypes we collected provided us with a unique opportunity to study the molecular evolution of interspecific hybrids. Ecological partitioning between the two Heterobasidion species, and the abundance of subalpine fir (a H. occidentale host) in and immediately below the study sites, suggest only H. occidentale should be present at high altitudes (Otrosina & Garbelotto, 2010).
Our air sampling confirmed that expectation. On the other hand, H. irregulare may have arrived at such high altitude because of an increase in its population size due to the relative recent increased logging of ponderosa pine at lower altitudes in the Bitterroot Mountain Range (Lockman, 2006). At the study site, H. irregulare seems to be particularly rare, as expected of a population that is not native to the site, but may have reached it thanks to the sporadic effects of ascending air currents. However, our trapping produced only four H. occidentale spores (corresponding to 14 spores/m 2 h) suggesting populations of H. occidentale may be small, probably because the site is at the altitudinal limit for the survival of this fungal species.
Numbers of spores trapped in other studies using comparable sampling approaches are significantly higher than the number reported here (Gonthier et al., 2005Gonthier, Lione, Giordano, & Garbelotto, 2012). Hence, hybridization may have been favored by demographic conditions, for example, by low numbers of individuals of both parental species increasing the chances of interspecific encounters and mating (Seehausen, 2004).

| Nuclei of hybrid isolates had either H. occidentale alleles or alleles derived from both H. irregulare and H. occidentale
Sequences of four nuclear loci indicated that allelic variation was large in hybrids and most sequences were not a perfect match for pure H. occidentale or pure H. irregulare GenBank sequences. This suggested that the hybrid genotypes were not F1 hybrids as originally thought (Lockman et al., 2014), but rather they may represent the progeny of hybrids backcrossed with other genotypes. Flow cytometry results confirmed this observation and showed that none of the genotypes from wood or from fruiting bodies were diploid, as it would have been expected for first-generation Heterobasidion hybrids (Garbelotto et al., 2004).

| Loss of concerted evolution and presence of intermediate alleles between H. irregulare and H. occidentale suggest that hybridization processes are ongoing
The loss of concerted evolution of ITS copies (nrDNA) of both H. irregulare-like and H. occidentale-like alleles might be a footprint of rampant ongoing hybridization (Muir, Fleming, & Schlötterer, 2001;Xu, Zeng, Gao, Jin, & Zhang, 2017). After hybridization processes, recombination of ITS regions followed by homogenization of repeats generally occurs, as observed in several basidiomycetous fungi (Hughes & Petersen, 2001;Kauserud, Svegården, Decock, & Hallenberg, 2007). In the current study, the retained heterogeneity of ITS copies regarded as an incomplete concerted evolution might suggest that the hybridization process is still ongoing (Buckler, Ippolito, & Holtsford, 1997 spp. may play a role in shaping the nuclear genome (Giordano, Sillo, Garbelotto, & Gonthier, 2018). Rapid evolution and speciation by hybridization have been described in detail for plants, animals, fungi, and oomycetes (Gross & Rieseberg, 2005;Mallet, 2007;Schumer, Rosenthal, & Andolfatto, 2014 Gladieux et al., 2014).
Although it cannot be excluded that different three dimensional folding of novel alleles during transcription may favor their selection (Chothia & Finkelstein, 1990), all of the novel alleles identified in this study were synonymous to ones previously identified in parental populations. This further reinforces the argument that the primary driver in the hybridization-mediated evolution acting on H. irregulare nuclei was mostly a consequence of its repeated recombination with H. occidentale. The creation of new alleles through intralocus recombination reported in this work has also been documented in Heterobasidion annosum × H. irregulare and Flammulina hybrids (Gonthier & Garbelotto, 2011;Hughes & Petersen, 2001). We believe this to be one of the first studies reporting hybridization-mediated evolution of one of the two parental nuclei due to unbalanced presence of parental individuals. We suggest that evolution of alleles in the H. occidentale nucleus was less evident because of the unidirectional backcrossing between hybrids and pure H. occidentale ( Figure 6).

| Hybridization-mediated evolution resulted in a novel taxonomic entity associated with Alpine larch in Montana, but with potentially far-reaching consequences
The fact that one of the nuclei in heterokaryotic hybrids has evolved to be clearly distinct from its progenitors makes this an irreversible evolutionary process apparently associated with infection of Alpine larch by what may be a novel Heterobasidion taxon.
Given the breadth of previous surveys on this genus, it is likely the distribution of this novel taxonomic entity may be somehow limited, maybe in association with high elevation Alpine larches.
Although the distribution of hybrids may be limited, unidirectional introgression of alleles from the H. irregulare-like nuclei into H. occidentale may not be limited, especially for those alleles that may confer an adaptive advantage to the recipient species (Currat, Ruedi, Petit, & Excoffier, 2008 (Garbelotto et al., 1996;Linzer et al., 2008).

| CON CLUS IONS
It is now widely recognized that hybridization is an important evolutionary process and may play a crucial role in speciation (Gross & Rieseberg, 2005;Mallet, 2007;Schumer et al., 2014;Taylor & Larson, 2019). The molecular characterization of H. irregular × H. occidentale hybrids presented in this study suggests that an anthropogenic disturbance may have lead to the hybridization-mediated evolution of a novel hybrid pathogen affecting Alpine larch. Furthermore, hybridization appears to have disproportionately affected the evolution of alleles of one of the two parents, and a higher rate of evolution was detected for the more rare parental species.
A broader sampling in the area of Bitterroot Mountains is necessary to monitor the range and the impacts on Alpine larch of the hybrid taxon here described. In addition, a large-scale genomic analysis will be pivotal to detect which chromosomal blocks and genes may be subjected to higher recombination and evolution levels, in order to link genomic and ecological traits of hybrids.

ACK N OWLED G M ENTS
We thank Gregg and Laura De Nitto for essential assistance during fieldwork and Maria Friedman for her contribution to generate molecular data while at U.C Berkeley. MG's salary was provided by the University of Torino during the laboratory analysis of data.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
MG and BL performed the fieldwork. MG and FS performed DNA amplification, allelic cloning, and sequencing. FS performed molecular analyses. FS, MG, and PG critically reviewed the manuscript. TK performed flow cytometry. FS and MG wrote the manuscript. All authors critically reviewed and edited the draft.

DATA ACCE SS I B I LIT Y
All data used in this study are included in the paper and in the