Genotyping of Russian isolates of fungal pathogen Trichophyton rubrum, based on simple sequence repeat and single nucleotide polymorphism

The Trichophyton rubrum species group consists of prevalent causative agents of human skin, nail and hair infections, including T rubrum sensu stricto and T violaceum, as well as other less well‐established or debatable taxa like T soudanense, T kuryangei and T megninii. Our previous study provided limited evidence in favour of the existence of two genetic lineages in the Russian T rubrum sensu stricto population.

of 70 strains in the sample. This sample shared 19 isolates with our earlier study. 16 Major part of strains originated from toenail onychomycosis and tinea pedis cases (87%), 6% were from tinea corporis, 4% from tinea cruris and 3% from fingernail onychomycosis.  Genetic Analyzer (Thermo Fisher Scientific, USA).

| Antifungal susceptibility testing
Antifungal susceptibility testing was performed according to EUCAST E.Def 9.3.1 document, 22 in RPMI-1640 medium with the addition of glucose to a final concentration of 2%. We tested the activity of terbinafine, itraconazole, fluconazole, posaconazole and voriconazole against six T rubrum isolates with newly obtained WGS sequences.

| Molecular strain typing
We employed two strain typing approaches. Firstly, we probed the G793A nucleotide substitution in protein-coding locus TERG_02941 (A7C99_6411) by Sanger DNA sequencing. The amplification was carried out under the following conditions: 5 minutes at 95°C; 39 cycles of 30 seconds at 95°C, 1 minute 5 seconds at 64°C and final elongation for 5 minutes at 72°C. All other steps were performed as described in the section 2.2 for ITS region sequencing. To optimise the genotyping process, we developed TaqMan SNP genotyping assay with two allele-specific probes (Table 1). Samples were run using a C1000 Touch thermal cycler equipped with a CFX96 real-time PCR detection system (Bio- Rad) according to the following programme: initial denaturation step for 3 minutes at 95°C, and 39 cycles of 10 seconds at 95°C and 30 seconds at 59°C. The assay generated fluorescent signal from both probes in every sample, specific signal reached the threshold value with 4-cycle outrunning.
Secondly, we performed microsatellite typing. We used eight microsatellite loci from the papers by Gräser et al 7

| Whole-genome sequencing
Genomic DNA of six T rubrum isolates was extracted using GeneJET Quality control of the reads was performed using FastQC 0.11.8. 25 It was followed by trimming low-quality reads and TruSeq2 adapters with Trimmomatic 0.39. 26 De novo assembly was performed with SPAdes 3.9.0 with the '--careful option'. 27 Raw reads were submitted to the Sequence Read Archive of National Center for Biotechnology Information (BioProject ID PRJNA552357).

| Phylogenomic reconstruction
We analysed six original whole-genome sequences, as well as

| Analysis of genetic polymorphism
A matrix of mutation probability-based genetic distances 30 TERG_02406 gene coding for α-domain protein was probed by AB759084 sequence. 13 The absence of HMG protein-coding gene was verified by the searches with T simii HMG gene, GenBank accession number AB605766. 39

| Quantitative RT-PCR
We hypothesised the influence of number of STRTr9 repeats on FBA gene expression under normal conditions. The sample of six

| RE SULTS
The majority of 69 T rubrum isolates from Russia and the type strain CBS 392.58 had the same ribosomal ITS region sequence, identical to KT285224. There were two exceptions, one of the isolates had 247insC mutation, and another isolate had additional TA element in the microsatellite near the 3'-end of the sequence. Our preliminary study showed discrepancies between the results of strain typing by TERG_02941 locus polymorphism and the microsatellite typing assay of Gong et al, 8 based on eight loci. 40 We hypothesised that this discrepancy was due to insufficient informativeness of the assay and used four additional polymorphic loci, up to a total of 12. The Our first approach for calculating microsatellite polymorphism-based genetic distances employed the probability of mutation events, described with the use of stepwise mutation model. 30 The split network, inferred using mutation probability-based distances, was divided into three major partitions. One partition contained TERG_02941 793A isolates and another two partitions contained TERG_02941 793G isolates ( Figure 1A). The second approach was based on the comparison of the lengths of microsatellite loci. 32 The resulting phylogenetic network also had pronounced division into TERG_02941 793A and 793G parts ( Figure 1B). In Rozenfeld distances-based network, further subdivision of TERG_02941 793G partition was less evident, but still discernable. From 22 TERG_02941 793G isolates, 18-21 were located in the same subgroups in both networks.
For the two genetic lines, the average lengths of PCR products of Tr001, Tr002, Tr003, Tr004, Tr005, Tr007, Tr008, Tr009, Tr(CT) 20 and Tr(GA) 25 loci differed by the length of less than one repeat unit. Average Tr006 lengths differed by a single unit length.
This difference should be obviously explained by the pressure of natural selection. Since the analysis of intraspecific divergence in most cases is carried out using neutral loci, 41 we performed calculation of genetic distances without STRTr9 length data. This broke clear grouping of isolates in both phylogenetic networks.
However, a certain tendency to clustering of TERG_02941 793G isolates persisted ( Figure 1C,D).
The analysis of molecular variance with 12 microsatellite loci revealed neglectable contribution of TERG_02941 793G subdivision to intergroup and overall variability (Table 3). With the removal of STRTr9 from the analysis, the percent of variability, explained by the differences between genetic lineages, dropped drastically. But, this was accompanied by significant increase in explanatory power of TERG_02941 793G group subdivision. A half of the between-group variability in the data set with 11 loci depended on it.
In the sample of 70 isolates, the association index value r d = 0.0248 laid beyond the right boundary of the distribution of expected frequencies. The p-value for r d was estimated at 0.002, which made it possible to reject the null hypothesis of the absence of connections between the markers. Hence, the studied T rubrum population was considered to be clonal. The CAAG repeats of STRTr9 ended at the position −278 in the 5'-region of fructose 1,6-bisphosphate aldolase gene TERG_06744, wherein the adenine of the ATG translation initiation codon is considered + 1. Therefore, we hypothesised that STRTr9 locus promotes microevolution in T rubrum through regulation of aldolase gene expression, which in its turn is one of the key enzymes of glycolysis. 42  These isolates were tested for susceptibility to antifungals ( Table 4).
The phylogenomic tree included T rubrum sensu stricto genomes from Europe, Asia and North America ( Figure 3) and contained  Since urease activity has been regarded phylogenetically important trait in the dermatophytes, 43  Note: The numbers reflect percent of variability among TERG_02941 793A and 793G groups of isolates from the total amount of variability between and within groups. To obtain three groups, TERG_02941 793G isolates were further divided according to the topology of Bruvo distances-based network (left-hand numbers in columns with data on 3 lineages) or by Rozenfeld distances-based network (right-hand numbers).

| D ISCUSS I ON
Until recently, low genetic diversity of T rubrum precluded persuasive description of its population structure. For this species, our earlier work seemingly represented the first successful attempt to obtain congruent results by two independent molecular strain typing techniques. 16 46,47 or filamentous ascomycetes with mixed reproduction mode, with 2%-13% of total variability explained by differences between intraspecific groups. 48,49 Intraspecific genetic lines in fungi sometimes demonstrate restricted geographic ranges. [50][51][52][53] Here, we compiled the map of global distribution of T rubrum genetic lineages (Figure 4). We largely built upon existing literature on NTS typing in T rubrum and departed from the observation that single 450-bp PCR product of TRS-1 locus is associated with ST1 genetic lineage. 16 We did not observe a specific distribution pattern of the two genetic lineages, but in most parts of the World, the derived ST1 lineage of T rubrum presumably had higher prevalence and therefore demonstrated evolutionary success. They concluded that virulence is a strain-specific rather than species-specific feature. Still, lineage-specific differences in STRTr9 locus length may hallmark different adaptation strategies, given that glycolysis pathway regulation is connected with mechanisms of keratin degradation. 66 Urease activity is an ancestral trait, commonly observable in geophilic dermatophytes, 43 and we expected to find it only in basal ST2 lineage. However, in our sample one of three urease-positive isolates belonged to ST1 lineage. Also, genetic diversity in the two lineages proved to be the same, which is contrary to what might be expected if ST2 is an ancestral lineage. It can be probably due to a sort of bottleneck, occurred in the population on the way from South-East Asia to Russia. In future studies, it is necessary to sample isolates from South and South-East Asia, in order to find the territory with highest genetic diversity and thereby conclusively determine the geographic region of the origin of the species.
Thus, we provided the description of the two intraspecific global lineages of T rubrum sensu stricto. They can be revealed by SNP, SSR and WGS-based typing techniques (and NTS, with a certain amount of ambiguity). The species is polymorphic in terms of ecology and physiology, but this polymorphism is not associated with genetic population structure.

ACK N OWLED G EM ENTS
The authors would like to thank the CBS-KNAW culture collection of Westerdijk Fungal Biodiversity Institute, Utrecht, the Netherlands, for providing the strain CBS 392.58. We also appreciate valuable comments from Adéla Čmoková and Ali Rezaei-Matehkolaei.

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