Understanding how organisms adapt to extreme living conditions is central to evolutionary biology. Dark septate endophytes (DSEs) constitute an important component of the root mycobiome and they are often able to alleviate host abiotic stresses. Here, we investigated the molecular mechanisms underlying the beneficial association between the DSE Laburnicola rhizohalophila and its host, the native halophyte Suaeda salsa, using population genomics. Based on genome-wide Fst (pairwise fixation index) and Vst analyses, which compared the variance in allele frequencies of single-nucleotide polymorphisms (SNPs) and copy number variants (CNVs), respectively, we found a high level of genetic differentiation between two populations. CNV patterns revealed population-specific expansions and contractions. Interestingly, we identified a ~20 kbp genomic island of high divergence with a strong sign of positive selection. This region contains a melanin-biosynthetic polyketide synthase gene cluster linked to six additional genes likely involved in biosynthesis, membrane trafficking, regulation, and localization of melanin. Differences in growth yield and melanin biosynthesis between the two populations grown under 2% NaCl stress suggested that this genomic island contributes to the observed differences in melanin accumulation. Our findings provide a better understanding of the genetic and evolutionary mechanisms underlying the adaptation to saline conditions of the L. rhizohalophila–S. salsa symbiosis.
The process by which organisms become better adapted to extreme environments and how natural selection shaped organismal phenotypes and genotypes in response to these intense selective pressures are central to evolutionary biology. Melanin-based pigmentation has long been used as a trait to understand adaptive evolution in extreme habitats across all forms of life [1,2,3,4]. Like other organisms, fungi are known to adapt to many stressful environments . Survival mechanisms often involve a higher melanin biosynthesis [6,7,8], leading to increased cell wall rigidity, resistance to many types of environmental stresses, including inactivation of the plant defense systems . Genome-based approaches have been used to investigate the evolution and adaptation of melanized fungi, mostly model species of plant and human pathogens, and soil fungi [9,10,11,12,13]. As of today, our understanding of the molecular mechanisms underlying the role of melanin in the adaptation of plant-associated mutualistic fungi to extreme environments is scarce [14,15,16].
Endophytic fungi often represent a very prominent component of the plant root mycobiome  and their diversity and ecological significance have only recently been widely appreciated . In particular, the root DSEs have received increased attention because they are ubiquitous in terrestrial ecosystems , and can decay organic matter and promote host stress tolerance [20,21,22]. Of note, DSEs are highly abundant in extreme environments experiencing low/high temperatures, drought, or/and high salinity. This prominence suggests that they might possess unique, undiscovered functional traits . The main unifying and conspicuous feature of DSEs is their ability to produce melanized hyphae and microsclerotia in the colonized roots of their host(s) [19, 24]. It has thus been suggested that melanization is a key trait for salt tolerance . Therefore, the study of the evolution of melanin biosynthesis pathways in DSEs should provide a better understanding of the molecular mechanisms underlying ecological adaptations.
Prior to this study, we performed a survey of the endophytic fungi colonizing roots of the native halophyte Suaeda salsa , a widely distributed plant in the saltmarsh areas along the western coast of the Yellow Sea in China. We found that a novel haploid DSE species, Laburnicola rhizohalophila sp. nov. (Pleosporales, Ascomycota) was a core member of the root mycobiome [26, 27]. Intriguingly, L. rhizohalophila isolates displayed a high degree of within-population phenotypic and physiological variations . This finding was unexpected owing to the fact that genotyped individuals were isolated from S. salsa seedlings collected less than one km apart. To address the potential influence of salt stress on L. rhizohalophila population structure, we confirmed the population structure by genome sequencing. Then, we assessed whether some genomic regions are experiencing positive selection and are responsible for the control of adaptive phenotypic traits. Furthermore, we characterized the functional relevance of such sequence variations to provide experimental evidence for selection-associated traits. Our work serves as a case study for investigating the mechanisms underlying the adaptation to abiotic stresses in a poorly investigated, but ecologically important, guild of plant-associated endophytes. This work also emphasizes the importance of taking a “reverse-ecology” approach linking “a priori” genome scan and “a posteriori” phenotypic characterization , albeit this approach is not in wide use.
Materials and methods
L. rhizohalophila isolates
S. salsa is an annual herbaceous euhalophyte that grows in saline soils. The present isolates were sampled along the coastal area in Dongying (N37°23′43′′, E118 °55′25′′), Shandong province, China. We collected S. salsa roots in July 2013 and 2014, and in June 2015. The distance between the three randomly selected sampling sites was less than 1.0 km. All endophytic L. rhizohalophila strains were isolated from surface-sterilized root tissues. The detailed isolation protocol was previously described . In total, 29 isolates were used for the present population genomics analysis. Ex-type living cultures of L. rhizohalophila are preserved in the China General Microbiological Culture Collection Center (CGMCC 3.19607–CGMCC 3.19616 and CGMCC 8756) .
Fungal DNA preparation and genome sequencing
We used L. rhizohalophila JP-R-44 as the reference isolate for high-quality de novo genome sequencing because JP-R-44 is the holotype for this novel species . Protocols used for genomic DNA extraction, DNA library construction, genome sequencing, genome assembly, and annotation are provided in the Supporting Information text S1.
Whole-genome and whole-population Illumina sequencing
The nuclear genomes of 29 L. rhizohalophila isolates were subjected to Illumina sequencing. A library with insert sizes of 300 bp was constructed using the NEBNext Ultra II DNA Library Prep Kit (New England BioLabs, Massachusetts, USA). The libraries were then sequenced on a HiSeq X Ten System (Illumina) using a PE-150 module.
SNP calling and population genetic analyses
For SNP calling, filtered high-quality DNA sequencing reads of the 29 individuals were aligned to the JP-R-44 reference genome using BWA software v0.7.17 and the MEM module . SAM alignment files were sorted and converted to BAM files with SAMtools v1.3 . Next, HaplotypeCaller, CombineGVCFs, GenotypeGVCFs, SelectVariants, and VariantFiltration in the GATK package were used to call, filter, and select SNPs . The filter parameters were as follows: “QD < 2.0 | | FS > 80.0 | | MQ < 20.0 | | SOR > 3.0 | | MQRankSum < −12.5 | | ReadPosRankSum < −8.0 | | QUAL < 40.0”.
To infer the phylogenomic relationships among individuals, a neighbor-joining phylogenetic tree was constructed by TreeBest v1.9.2. (http://treesoft.sourceforge.net/treebest.shtml) under the p-distances model using population-scale SNPs. The population structure was first studied using principal component analysis (PCA). We performed PCA with population-scale SNPs using the software plink  and gcta . A Bayesian population structure assessment was inferred using ADMIXTURE v1.3.0  with maximum likelihood estimation and the block relaxation algorithm. Simulations were run with 1000 bootstraps and tenfold cross-validation. We increased the pre-defined genetic clusters from K = 2 to K = 10 (number of ancestral populations). Prior to running ADMIXTURE, we removed SNPs that were in high linkage disequilibrium (LD). Specifically, we used plink to prune SNPs using a windowed approach, along with an R2 threshold of 0.8.
Analysis of LD decay and recombination detection
To determine whether the population structure of L. rhizohalophila was clonal, panmictic, or selfing, we performed an LD decay analysis. Pairwise LD was calculated between all SNPs from 1 and 10,000 bp apart using vcftools with the settings “vcf All.recode.vcf–keep cleda1.lst–ld-window-bp 10,000–min-r2 0–geno-r2–out cleda1.10k”. The mean value of LD across 10-kbp sliding windows from any given SNP was then calculated. Loess curves were fitted to the mean LD decay values for the sliding windows with the R function “loess”. To determine the approximate distance required to reach 50% LD decay, we used the window start of the closest point above 50% of the maximum linkage. In the scenario that one of the populations had a significantly higher number of SNPs than the other, we would have a finer resolution to detect recombination events. To address this scenario, we used the identical number of SNPs (170,000) randomly sampled along the genome from group 1 and group 2 for the LD analysis.
To determine the level of recombination between groups, we visualized potential recombination events in SplitsTree v4.14.4 (http://splitstree.org/)  for the whole dataset using pairwise distances with the Kimura K3ST model, which allows for the identification of reticulated evolution rather than a strictly bifurcating path.
Identification of the mating-type (MAT) locus and flanking regions in L. rhizohalophila populations
The sequence of the MAT locus and its flanking regions in the JP-R-44 genome was determined by performing a local tBlastn search . The identities of the annotated genes were further determined by BLASTp against the National Center for Biotechnology Information (NCBI) non-redundant protein database .
Estimation of population differentiation
We used three indices to assess population differentiation: SNP analysis using Fst (Wright’s fixation index, mean genetic differentiation)  and Dxy (absolute genetic divergence) (equation 10.20) , and copy number variation (CNV) analysis using Vst (a population differentiation estimator similar to Fst). We selected overlapping 10-kbp windows with a 1-kbp step size to identify regions with increased genetic divergence (Fst) across the two groups as indicated by the ADMIXTURE analysis. This window size was chosen to provide a reasonably large number of SNPs per window (average 108 SNPs per window in L. rhizohalophila population) as recommended in . The gl.fst.pop script was used to calculate pairwise Fst values in the R package StAMPP with 1000 bootstrap replications . Dxy was calculated with an in-house script. We Z-transformed the distribution of Fst and the ZFst values were standardized using R . Genomic windows with high ZFst values (the top 1% and 5%) were used as the significance levels and analyzed for gene content. Candidate loci were visualized using Manhattan plots. Gene ontology (GO) term enrichment analysis in the list of differentiated genes was performed with clusterProfiler v3.4.4 . We set the significant threshold of GO terms enrichment to p value ≤ 0.05.
To identify whether CNVs were also divergent between the two populations, we first identified windows showing significant variation in normalized read depth. Here, we refer to CNV of 0 as an “absence” rather than a deletion because the absence of a locus in an analyzed genome could be the result of a gain in the reference genome and not solely a deletion in the analyzed genomes . To screen genomes for CNVs, we used the cn.MOPS R package with the haplocn.mops algorithm (the default bin size was 1 kbp), which is specifically designed for haploid organisms . We plotted CNV heatmaps for the entire genome as well as heatmaps for genes of interest. In addition, we calculated two parameters, including the polymorphic index content (PIC) and Vst, to measure CNV diversity and divergent CNV profiles between the groups, respectively .
Detection of genomic signatures of selection
To identify the potential evolutionary processes affecting population differentiation, we used population-based methods to detect selection on coding sequences. We tested for departures from neutrality using Tajima’s D statistics  to infer neutrality and/or selection at genomic regions with ZFst outliers, implying divergent natural selection. Tajima’s D detects departures from the expected site-frequency spectrum under neutral assumptions by comparing two measures of nucleotide genetic diversity (θ) —segregating sites (θW) and nucleotide diversity (θπ)—which were calculated using VariScan v2.0.2 software . As Fst statistics are sometimes affected by demography, migration, and small sample size , we applied a nonparametric empirical permutation approach to obtain a null distribution of Tajima’s D statistics . Then, we compared this distribution to the observed Tajima’s D statistics to identify candidate regions under selection for each group. Briefly, when the statistical values of a given locus were below the fifth percentile of the genomic and null distribution, we concluded that the signal of positive selection was active and strong [51, 52]. To further confirm the positive selection signals identified, we used the composite likelihood ratio (CLR) test  as implemented in the program SweeD v3.3.2 . An overlapping 10-kbp window across the whole genome was used for scanning with a cross-population composite likelihood (XP-CLR) test. Finally, selective sweeps often leave large areas surrounding the site of selection in LD. We used Haploview v4.1  to visualize LD in the candidate genomic regions where extremely negative Tajima’s D values were observed.
Mycelial growth and melanin biosynthesis under salinity conditions
We used the mycelial growth and melanin biosynthesis under saline conditions as proxies for fitness. Seven isolates from group 1 or group 2 of L. rhizohalophila were randomly selected for phenotypic measurements. We used potato dextrose broth (PDB) medium with or without 2% (w/v) NaCl for culturing the vegetative mycelium. All fungal cultures were kept in an incubator without shaking for up to 2 weeks at 25 °C. Each treatment was conducted in four biological replicates. Detailed methods for melanin extraction and measurement have been described in .
Biomass accumulation and colony diameter were used to measure fungal growth. Biomass measurements were taken as described above using the same culture conditions, but isolates were grown on PDA with 2% or 6% NaCl. Colony diameters were measured in triplicate after two weeks at 25 °C. The software OriginLab 10.0 (OriginLab Corporation, Northampton, USA) was used for statistical analysis and plotting. Significant differences between the two L. rhizohalophila groups were evaluated with a Student’s t-test at p ≤ 0.05. All data were expressed as mean values with standard deviations (SD).
The genome assembly and annotation of L. rhizohalophila JP-R-44 are available at the NCBI (BioProject number: PRJNA517533, BioSample number: SAMN10836428). The whole-genome sequencing (WGS) project has been deposited at DDBJ/ENA/GenBank under the accession SELF00000000. The raw genome resequencing data have been deposited at the NCBI in the sequence read archive (SRA) database under the accession numbers: SRR8569110 to SRR8569138. Raw VCF file (SNP variants) is available via Data Dryad (https://doi.org/10.5061/dryad.18931zcvw).
Genome and population structure of L. rhizohalophila
The L. rhizohalophila reference genome was assembled into 42 scaffolds with a cumulative length of 61.4 Mbp and N50 of 2.53 Mbp. It contains 14,646 protein-coding genes. The quality of the genome assembly and gene annotation were evaluated with BUSCO (Benchmarking Universal Single-Copy Orthologs v4.0.1, Fungi odb10 dataset). The 90.2% completeness suggests that the genome assembly and annotation are of suitable high quality (Table S1). Whole-genome resequencing from 29 individuals produced > 0.9 billion reads with an average sequencing depth of 72× per individual, ranging from 41× to 159× (Table S2). Sequencing reads were aligned to the reference genome for SNP calling. We recovered a total of 1,241,238 SNPs across the 29 L. rhizohalophila genomes after singleton filtering, with 2.0% of the genome as variable sites. The distance-based NJ phylogenetic tree revealed that the L. rhizohalophila population was well separated into three major subgroups with group 1 genetically closer to group 3, and distant to group 2 (Fig. 1A).
A genetic structure with three subpopulations was further confirmed by using PCA (Fig. 1B) and admixture-based model (Figs. 1C and S1) analyses; the first two principal components (axis 1 and axis 2) accounted for ~59% of the cumulative genetic variation (Fig. 1B). Group 3 was excluded from subsequent analysis due to the small sample size (N = 2).
Recombination pattern analysis and MAT gene structure
The decrease in LD (R2) with physical distance for the two major L. rhizohalophila subpopulations is shown in Fig. 2A. We observed higher LD values in group 2 compared with group 1. LD decay was relatively slower for group 2 with an LD50 of 520 bp, while LD decayed rapidly and reached half its maximum value (LD50) at ~120 bp in group 1. This contrasting LD decay pattern suggests that the frequency of genetic recombination also differed between the two groups.
Organization of the mating-type (MAT) loci and its flanking regions revealed that all individuals in group 1 and group 2 carry both mat1-2-1 and mat1-1-1 idiomorphs in a homothallic gene arrangement (Fig. 2B). We further assessed their evolutionary relationships while considering the possible occurrence of a recombination signal within or between groups. Group 1 was characterized by a complex network with many reticulations; by contrast, a star-shaped pattern was recovered in group 2 (Fig. 2C). Thus, this recombination pattern suggests that group 1 is more likely to be panmictic, while group 2 is more likely to be clonal. The disparity in nucleotide diversity between the two groups supported the conclusion. The genetic diversity of group 1, estimated by the average pairwise divergence (θπ = 0.0028) and the proportion of polymorphic sites per base (θw = 0.0025), was much higher than that in group 2 (θπ = 0.0009 and θw = 0.0010) (Wilcoxon’s test with p value < 2.2 × 10−16) (Table 1).
Genome-wide divergence of SNP and CNV profiles
We then assessed the genetic differences between the two subpopulations, searching for loci under adaptive selection. A total of 696,045 SNPs were called from all individuals from L. rhizohalophila groups 1 and 2. We found that differentiation between the subpopulations was high with an average Fst = 0.563 (95% CI 0.562–0.565) and the mean divergence Dxy = 0.27% (95% CI 0.26–0.27%). Unexpectedly, only approximately 5.3% of the SNPs were shared between the two groups and 11.9% (83,136) of the sites were reciprocally fixed for different alleles in the two groups. The remaining 82.8% of the SNPs were variable in one group and fixed in the other (Table 2). This distribution of shared polymorphism and fixed differences, together with high levels of divergence, may suggest an ancient split between the two groups.
We also assessed the pattern, selection, and diversity of CNVs using a sequence read-depth approach (Fig. S2). We found a total of 317 CNVs (139 duplications and 218 deletions) (Table S3 and Fig. S3). In total, 13.97 Mbp (22.7%) of the L. rhizohalophila genome contains CNVs, whereas only 1.24 Mbp (2%) is affected by SNPs. We detected a total of 1279 genes with significant CNV differences (Welch’s two-sample t-test, false discovery rate corrected p-value < 0.01) between the two groups, including 718 genes with putative functions and 561 genes with unknown functions. We observed a striking decrease in gene copies in group 1 and an expansion of gene duplications in group 2 (Fig. 3A). We further characterized patterns of CNV within populations by independently calculating the Polymorphic Information Content (PIC). The average PIC values were 0.095 and 0.242 for groups 1 and 2, respectively, reflecting the lower level of CNV in group 1 (Fig. 3B, C). Overall, our results suggest a faster turnover of copy numbers for group 1 compared with group 2, which can be explained by more frequent recombination in group 1 (Fig. 2C).
Estimates of Vst for CNVs revealed a number of highly differentiated loci that may be indicative of group-specific selective pressures; the average Vst was 0.255 (95% CI 0.219–0.290) (Fig. 3D). We found that a total of 36 CNVs overlapped 117 genes with extreme Vst values of 1.0. Although no GO term was enriched, small subsets of genes were functionally related to salinity adaptation (ion channels, melanin biosynthesis, etc.). For example, the polyketide synthase (pks1) and the transmembrane osmosensor are thought to be putatively involved in melanin biosynthesis and the high osmolarity-glycerol signaling pathway, respectively.
Footprints of positive selection in the genome
We calculated Tajima’s D to identify deviation from the neutral model of molecular evolution. Tajima’s D values were negative across the entire L. rhizohalophila population (−0.74) and the entire genome of group 2 (−0.32), perhaps as a result of demographic processes (population expansion) affecting the entire genome equally. In contrast, Tajima’s D was slightly positive for group 1 (0.54) (Table 1). In total, we found 1176 genes with outlier ZFst values exceeding the 95th percentile (Fig. 4A). Of these, 173 and 123 genes in group 1 and group 2, had Tajima’s D values below −1.0, respectively. GO enrichment analysis revealed that cellular components including the categories “intrinsic component of membrane,” “integral component of membrane,” and “membrane part” were significantly enriched in group 1 specifically (uncorrected p value < 0.05, Fig. S4), suggesting that the salinity-responsive systems have experienced group-specific activation and potentially positive selection.
To further identify patterns of genomic variation indicative of positive selection, we calculated genome-wide distribution of Tajima’s D using both sliding window and nonparametric empirical permutation approaches , to confirm that these peaks of population differentiation represent foci of selection. Briefly, we defined regions of the genome experiencing putative positive selection when statistic values of a given locus fell below the fifth percentile of the genomic and null distribution (Fig. S5). These criteria resulted in a cut-off of Tajima’s D statistic values of ≤ −1.31 and −1.34 for group 1 and group 2, respectively. If this stringent criterion is used, most of the salinity-responsive genes maintained the signature of positive selection.
Of note, we identified a genomic region (~20 kbp in length) located on scaffold 6 that had a strong signal of positive selection in group 1 individuals. This genomic region appears to encode a ~47 kbp melanin biosynthetic gene cluster, containing six genes related to melanin metabolism (polyketide synthase pks2, protein ID 2362; melanocyte-stimulating hormone receptor mshr, protein ID 2359), transport (major facilitator superfamily transporter mfs, protein ID 2361), transcriptional regulation (transcription factors ID 2358 and 2363) and cell wall remodeling (expansin, protein ID 2360). This region had low levels of genetic variation compared to the genome-wide average and had extremely negative Tajima’s D values in both groups (Fig. 4B). This pattern is consistent with recent and strong positive selection, although low absolute nucleotide diversity might also stem from the reduced mutation rates in these regions. An XP-CLR test also revealed a high but narrow CLR peak at the sweeping site of this region under the 0.01 significance level in group 1 (Fig. 4B). More specifically, the genes pks2 (XP-CLR = 85.4, top 1% cutoff = 15.09, Tajima’s D = −1.85) and mshr (XP-CLR = 39.7, Tajima’s D = −1.83) showed strong signatures of selective sweeps. Subsequently, pairwise LD was examined in this specific region using Haploview to determine the extent of the putative selective sweep. A strong pattern of LD surrounding the pks2 and mfs genes was observed in group 1 but not in group 2 (Fig. 4C, D). Moreover, the pks2 gene contained 10 non-synonymous SNPs fixed or nearly fixed in group 1 with a major allele frequency of 90%, while only three non-synonymous SNPs in group 2 (Fig. S6).
Impact of salinity on melanin synthesis and mycelial growth
Owing to the high sequence divergence of the genomic island encoding a putative melanin biosynthetic gene cluster, we hypothesized that melanin production and growth under salinity stress may also differ dramatically between the two subpopulations. We thus measured the mycelial growth and melanin accumulation in seven representative isolates of each group across two salinity levels. We found that isolates of group 1 accumulated melanin to a lesser extent than group 2 isolates under the absence of NaCl, but the difference was not statistically significant. In contrast, average colony diameter and biomass accumulation were significantly higher in group 1 isolates compared with group 2 isolates (two-tailed t-test, p value = 0.007) in the 2% NaCl treatment (Fig. 5A, B). The melanin content of group 1 isolates was also significantly higher (two-tailed t-test, p value = 0.013) than that of group 2 isolates in the 2% NaCl medium (Fig. 5C), indicating that group 1 is more adapted to saline stress than group 2. The 6% NaCl treatment hampered fungal growth and measurements of melanin content cannot be completed.
Stress and evolution are always heavily intertwined . In this work, we aimed to understand how the L. rhizohalophila–S. salsa symbiosis adapts to high saline soils by conducting population genomics analyses. We identified cryptic subpopulations of L. rhizohalophila, referred to as groups 1 and 2. We characterized a high level of genetic differentiation between these two lineages, suggesting a limited gene flow between subpopulations. Each group displayed a distinctive pattern of genetic recombination. Although the homothallic nature of group 2 individuals likely explains the observed low level of genetic recombination due to frequent inbreeding , it is still unclear why and how the extensive recombination is taking place in group 1. One possibility is that genetic exchange through parasexual recombination would occur and contribute to the observed genetic diversity .
The lack of physical barriers to migration and gene flow in our fine-scale study area suggests that the observed phenotypic differences are likely driven by microhabitat heterogeneity of saline soils imposing strong selection pressures. For instance, sea tides generate dramatic seasonal changes in salinity and other physicochemical factors in salt marsh soils . Such environmental changes could potentially promote the evolution of stable gene polymorphisms. Allelic variants at specific loci would then reflect an ongoing process of adaptation to heterogeneous conditions [59, 60]. This has been suggested to explain the fine-scale genetic divergence of the teleost fish Fundulus heteroclitus populations, inhabiting three concurrent microhabitats within a single salt marsh . In our study, we were not able to measure the physicochemical characteristics of the soils where S. salsa with its L. rhizohalophila endophytes was collected. This should be carried out in future studies to assess the ecological gradient taking place in this habitat.
Melanin biosynthesis is considered a major adaptive trait in fungi [7, 9]. Our results suggested that L. rhizohalophila is undergoing adaptation to salinity stress and this process appears to involve a ‘genomic island of divergence’ related to melanin biosynthesis. This gene arrangement may reduce the ability of recombination to break up favorable combinations of alleles as the frequency of recombination is greater in group 1 (Fig. 2) . Of these genes, pks2 is of particular interest, as most filamentous fungi synthesize melanin via a PKS pathway (Fig. 6) [62, 63]. We also identified a gene encoding a melanocyte-stimulating hormone receptor (MSHR) which is under positive selection. As an integral component of the plasma membrane, MSHR has been found only in mammals where it contributes to skin pigmentation . However, there is a unified cellular principle for melanization in mammals and fungi . We thus speculate that both pks2 and mshr genes are involved in the increased melanin biosynthesis in L. rhizohalophila. In fungi, gene clusters for the synthesis of secondary metabolites generally encode transporter proteins, especially major facilitator superfamily (MFS) transporters . Thus, the mfs gene found in the PKS cluster might be involved in the transport of melanin from the cytosol to the cell wall. The role of the expansin-like gene located in the PKS cluster remains unclear. Expansins act as loosening agents to increase plant cell growth and fungal expansins function in plant root colonization . Given the fact that physically linked genes usually belong to the same metabolic network, it is tempting to speculate that the expansin-like protein loosens the fungal cell wall for increased melanin deposition. We propose that the genomic island contains a melanin biosynthetic gene cluster involved in melanin biosynthesis (PKS2 and MSHR), membrane trafficking (MFS), regulation (two transcription factors), and remodeling (loosening cell wall regulated by expansin) (Fig. 6).
Surprisingly, it seems that deletion of pks1 does not reduce the melanin production in group 1 when challenged by salinity stress. The predicted coding sequences of pks1 (protein ID 12252) and pks2 were 741 bp and 6339 bp in length, respectively. Interestingly, in silico analysis using antiSMASH 5.0 revealed that the pks1 only had a ketosynthase (KS) domain but lacked the essential acyltransferase (AT) and acyl carrier protein (ACP) domains (Fig. 6), indicating that the function of pks1 may be impaired.
Our data showed that group 1 isolates exhibited a growth advantage and higher melanin accumulation relative to group 2 isolates when incubated under 2% NaCl conditions. This suggests that increased melanization facilitates the adaptation to high-salinity environments, although it remains to assess how L. rhizohalophila hyphae inhabiting the host roots are affected by the external, high soil salinity.
In this study, we uncovered the fine-scale population structure and ecological divergence of the endophyte L. rhizohalophila associated with the roots of the plant halophyte S. salsa. Both functional and genetic evidence support the role of a genomic island encoding a gene cluster, involved in biosynthesis, membrane trafficking, regulation, and localization of melanin, in shaping a melanization-associated phenotype (Fig. 6). Differences in mycelial growth yield and melanin biosynthesis between L. rhizohalophila subpopulations grown under saline stress conditions suggest that this genomic island contributes to the observed differences in melanin accumulation. Our findings provide a better understanding of the genetic and evolutionary mechanisms (e.g., melanin synthesis) underlying the adaptation to saline conditions of the L. rhizohalophila–S. salsa symbiosis. Population genomics studies of additional DSE symbioses will confirm whether melanization is an adaptation to extreme environments.
Hoekstra H. Genetics, development and evolution of adaptive pigmentation in vertebrates. Heredity. 2006;97:222–234.
McNamara ME, Rossi V, Slater TS, Rogers CS, Ducrest AL, Dubey S, et al. Decoding the evolution of melanin in vertebrates. Trends Ecol Evol. 2021; https://doi.org/10.1016/j.tree.2020.12.012.
Roulin A. Melanin-based colour polymorphism responding to climate change. Glob Chang Biol. 2014;20:3344–3350.
Laurent S, Pfeifer SP, Settles ML, Hunter SS, Hardwick KM, Ormond L, et al. The population genomics of rapid adaptation: disentangling signatures of selection and demography in white sands lizards. Mol Ecol. 2016;25:306–323.
Naranjo-Ortiz MA, Gabaldón T. Fungal evolution: major ecological adaptations and evolutionary transitions. Biol Rev Camb Philos Soc. 2019;94:1443–1476.
Cordero RJ, Casadevall A. Functions of fungal melanin beyond virulence. Fungal Biol Rev. 2017;31:99–112.
Treseder KK, Lennon JT. Fungal traits that drive ecosystem dynamics on land. Microbiol Mol Biol Rev. 2015;79:243–262.
Kejžar A, Gobec S, Plemenitaš A, Lenassi M. Melanin is crucial for growth of the black yeast Hortaea werneckii in its natural hypersaline environment. Fungal Biol. 2013;117:368–379.
Singaravelan N, Grishkan I, Beharav A, Wakamatsu K, Ito S, Nevo E. Adaptive melanin response of the soil fungus Aspergillus niger to UV radiation stress at “Evolution Canyon”, Mount Carmel, Israel. PLoS ONE. 2008;3:e2993.
Krishnan P, Meile L, Plissonneau C, Ma X, Hartmann FE, Croll D, et al. Transposable element insertions shape gene regulation and melanin production in a fungal pathogen of wheat. BMC Biol. 2018;16:78.
Pereira D, Croll D, Brunner PC, McDonald BA. Natural selection drives population divergence for local adaptation in a wheat pathogen. Fungal Genet Biol. 2020;141:103398.
Desjardins CA, Giamberardino C, Sykes SM, Yu CH, Tenor JL, Chen Y, et al. Population genomics and the evolution of virulence in the fungal pathogen Cryptococcus neoformans. Genome Res. 2017;27:1207–1219.
Robertson KL, Mostaghim A, Cuomo CA, Soto CM, Lebedev N, Bailey RF, et al. Adaptation of the black yeast Wangiella dermatitidis to ionizing radiation: molecular and cellular mechanisms. PLoS ONE. 2012;7:e48674.
Knapp DG, Németh JB, Barry K, Hainaut M, Henrissat B, Johnson J, et al. Comparative genomics provides insights into the lifestyle and reveals functional heterogeneity of dark septate endophytic fungi. Sci Rep. 2018;8:6321.
Fernandez CW, Koide RT. The function of melanin in the ectomycorrhizal fungus Cenococcum geophilum under water stress. Fungal Ecol. 2013;6:479–486.
Redman RS, Sheehan KB, Stout RG, Rodriguez RJ, Henson JM. Thermotolerance generated by plant/fungal symbiosis. Science. 2002;298:1581.
Peay KG, Kennedy PG, Talbot JM. Dimensions of biodiversity in the earth mycobiome. Nat Rev Microbiol. 2016;14:434–447.
Rodriguez RJ, White JF, Arnold AE, Redman RS. Fungal endophytes: diversity and functional roles. N Phytol. 2009;182:314–330.
Yuan ZL, Su ZZ, Zhang CL. Understanding the biodiversity and functions of root fungal endophytes: the ascomycete Harpophora oryzae as a model case. In: Irina S Druzhinina IS, Kubicek CP editors). The mycota Vol. IV: environmental and microbial relationships. 3rd ed. Springer; 2016, pp 205–214.
Berthelot C, Leyval C, Foulon J, Chalot M, Blaudez D. Plant growth promotion, metabolite production and metal tolerance of dark septate endophytes isolated from metal-polluted poplar phytomanagement sites. FEMS Microbiol Ecol. 2016;92:fiw144.
Hill PW, Broughton R, Bougoure J, Havelange W, Newsham KK, Grant H, et al. Angiosperm symbioses with non-mycorrhizal fungal partners enhance N acquisition from ancient organic matter in a warming maritime Antarctic. Ecol Lett. 2019;22:2111–2119.
Mateu M, Baldwin A, Maul J, Yarwood S. Dark septate endophyte improves salt tolerance of native and invasive lineages of Phragmites australis. ISME J. 2020;14:1943–1954.
Porras-Alfaro A, Herrera J, Sinsabaugh RL, Odenbach KJ, Lowrey T, Natvig DO. Novel root fungal consortium associated with a dominant desert grass. Appl Environ Microbiol. 2008;74:2805–2813.
Qin Y, Pan XY, Kubicek CP, Druzhinina IS, Chenthamara K, Labbé J, et al. Diverse plant-associated pleosporalean fungi from saline areas: ecological tolerance and nitrogen-status dependent effects on plant growth. Front Microbiol. 2017;8:158.
Gostinčar C, Grube M, de Hoog S, Zalar P, Gunde-Cimerman N. Extremotolerance in fungi: evolution on the edge. FEMS Microbiol Ecol. 2010;71:2–11.
Yuan ZL, Druzhinina IS, Labbé J, Redman R, Qin Y, Rodriguez R, et al. Specialized microbiome of a halophyte and its role in helping non-host plants to withstand salinity. Sci Rep. 2016;6:32467.
Yuan ZL, Druzhinina IS, Wang X, Zhang X, Peng L, Labbé J. Insight into a highly polymorphic endophyte isolated from the roots of the halophytic seepweed suaeda salsa: Laburnicola rhizohalophila sp. nov. (Didymosphaeriaceae, Pleosporales). Fungal Biol. 2020;124:327–337.
Ellison CE, Hall C, Kowbel D, Welch J, Brem RB, Glass NL, et al. Population genomics and local adaptation in wild isolates of a model microbial eukaryote. Proc Natl Acad Sci USA. 2011;108:2831–2836.
Li H, Durbin R. Fast and accurate long-read alignment with Burrows–Wheeler transform. Bioinformatics 2010;26:589–595.
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The sequence alignment/map format and SAMtools. Bioinformatics. 2009;25:2078–2079.
Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, Del Angel G, Levy-Moonshine A, et al. From FastQ data to high confidence variant calls: the genome analysis toolkit best practices pipeline. Curr Protoc Bioinform. 2013;43:11.10.1–11.10.33.
Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7.
Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88:76–82.
Alexander DH, Novembre J, Lange K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009;19:1655–1664.
Huson DH, Bryant D. Application of phylogenetic networks in evolutionary studies. Mol Biol Evol. 2006;23:254–267.
Wilken M, Steenkamp E, Wingfield M, De Beer ZW, Wingfield B. Which MAT gene? Pezizomycotina (Ascomycota) mating-type gene nomenclature reconsidered. Fungal Biol Rev. 2017;31:199–211.
Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402.
Wright S. The genetical structure of populations. Ann Eugen. 1951;15:323–354.
Nei M (ed). Molecular evolutionary genetics. Columbia University Press; 1987.
Carlson CS, Thomas DJ, Eberle MA, Swanson JE, Livingston RJ, Rieder MJ, et al. Genomic regions exhibiting positive selection identified from dense genotype data. Genome Res. 2005;15:1553–1565.
Pembleton LW, Cogan NOI, Forster JW. StAMPP: an R package for calculation of genetic differentiation and structure of mixed-ploidy level populations. Mol Ecol Resour. 2013;13:946–952.
Axelsson E, Ratnakumar A, Arendt ML, Maqbool K, Webster MT, Perloski M, et al. The genomic signature of dog domestication reveals adaptation to a starch-rich diet. Nature. 2013;495:360–364.
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics. 2012;16:284–287.
Zhao S, Gibbons JG. A population genomic characterization of copy number variation in the opportunistic fungal pathogen Aspergillus fumigatus. PLoS ONE. 2018;13:e0201611.
Klambauer G, Schwarzbauer K, Mayr A, Clevert DA, Mitterecker A, Bodenhofer U, et al. cn.MOPS: mixture of Poissons for discovering copy number variations in next-generation sequencing data with a low false discovery rate. Nucleic Acids Res. 2012;40:e69.
Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 1989;123:585–595.
Nei M, Li WH. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc Natl Acad Sci USA. 1979;76:5269–5273.
Hutter S, Vilella AJ, Rozas J. Genome-wide DNA polymorphism analyses using VariScan. BMC Bioinform. 2006;7:409.
Wagner DN, Baris TZ, Dayan DI, Du X, Oleksiak MF, Crawford DL. Fine-scale genetic structure due to adaptive divergence among microhabitats. Heredity. 2017;118:594–604.
Rech GE, Sanz-Martín JM, Anisimova M, Sukno SA, Thon MR. Natural selection on coding and noncoding DNA sequences is associated with virulence genes in a plant pathogenic fungus. Genome Biol Evol. 2014;6:2368–2379.
Sterken R, Kiekens R, Coppens E, Vercauteren I, Zabeau M, Inzé D, et al. A population genomics study of the Arabidopsis core cell cycle genes shows the signature of natural selection. Plant Cell. 2009;21:2987–2998.
Yu F, Keinan A, Chen H, Ferland RJ, Hill RS, Mignault AA, et al. Detecting natural selection by empirical comparison to random regions of the genome. Hum Mol Genet. 2009;18:4853–4867.
Nielsen R, Williamson S, Kim Y, Hubisz MJ, Clark AG, Bustamante C. Genomic scans for selective sweeps using SNP data. Genome Res. 2005;15:1566–1575.
Pavlidis P, Živkovic D, Stamatakis A, Alachiotis N. SweeD: likelihood-based detection of selective sweeps in thousands of genomes. Mol Biol Evol. 2013;30:2224–2234.
Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–265.
Zhan F, He Y, Zu Y, Li T, Zhao Z. Characterization of melanin isolated from a dark septate endophyte (DSE), Exophiala pisciphila. World J Microbiol Biotechnol. 2011;27:2483–2489.
Taylor JW, Hann-Soden C, Branco S, Sylvain I, Ellison CE. Clonal reproduction in fungi. Proc Natl Acad Sci USA. 2015;112:8901–8908.
McGuire IC, Davis JE, Double ML, MacDonald WL, Rauscher JT, McCawley S, et al. Heterokaryon formation and parasexual recombination between vegetatively incompatible lineages in a population of the chestnut blight fungus, Cryphonectria parasitica. Mol Ecol. 2005;14:3657–3669.
Szulkin M, Gagnaire PA, Bierne N, Charmantier A. Population genomic footprints of fine-scale differentiation between habitats in Mediterranean blue tits. Mol Ecol. 2016;25:542–558.
Hamilton JA, De la Torre AR, Aitken SN. Fine-scale environmental variation contributes to introgression in a three-species spruce hybrid complex. Tree Genet Genomes. 2015;11:817.
Yeaman S. Genomic rearrangements and the evolution of clusters of locally adaptive loci. Proc Natl Acad Sci USA. 2013;110:E1743–E1751.
Kroken S, Glass NL, Taylor JW, Yoder OC, Turgeon BG. Phylogenomic analysis of type I polyketide synthase genes in pathogenic and saprobic ascomycetes. Proc Natl Acad Sci USA. 2003;100:15670–15675.
Woo PCY, Tam EW, Chong KT, Cai JJ, Tung ET, Ngan AH, et al. High diversity of polyketide synthase genes and the melanin biosynthesis gene cluster in Penicillium marneffei. FEBS J. 2010;277:3750–3758.
Kameyama K, Montague PM, Hearing VJ. Expression of melanocyte stimulating hormone receptors correlates with mammalian pigmentation, and can be modulated by interferons. J Cell Physiol. 1988;137:35–44.
Upadhyay S, Xu X, Lowry D, Jackson JC, Roberson RW, Lin X. Subcellular compartmentalization and trafficking of the biosynthetic machinery for fungal melanin. Cell Rep. 2016;14:2511–2518.
Coleman JJ, Mylonakis E. Efflux in fungi: la pièce de résistance. PLoS Pathog. 2009;5:e1000486.
Cosgrove DJ. Microbial expansins. Annu Rev Microbiol. 2017;71:479–497.
This work was financially supported by the Fundamental Research Funds for the Central Non-profit Research of the Chinese Academy of Forestry (CAFYBB2020ZY002-1) and the National Natural Science Foundation of China (No. 31722014). Research in FMM’s laboratory is supported by the Laboratory of Excellence ARBRE (ANR-11-LABX-0002-01) and the Beijing Advanced Innovation Center for Tree Breeding by Molecular Design. We extend our sincere gratitude to Dr. Fabien Halkett (INRAE, Université de Lorraine), Dr. Fengyan Bai (State Key Laboratory of Mycology, Chinese Academy of Sciences), and Dr. Shu Zhao (University of Massachusetts) for many useful comments.
Conflict of interest
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The original online version of this article was revised due to missing affiliation of Prof. Yves Van de Peer.
About this article
Cite this article
Yuan, Z., Druzhinina, I.S., Gibbons, J.G. et al. Divergence of a genomic island leads to the evolution of melanization in a halophyte root fungus. ISME J 15, 3468–3479 (2021). https://doi.org/10.1038/s41396-021-01023-8
This article is cited by
Culturable endophytic fungi community structure isolated from Codonopsis pilosula roots and effect of season and geographic location on their structures
BMC Microbiology (2023)
Two new root endophyte and nematode cyst parasite species of the widely distributed genus Laburnicola
Mycological Progress (2022)
Growth-promoting effects of dark septate endophytes on the non-mycorrhizal plant Isatis indigotica under different water conditions