Genetic Diversity and Population Structure of Fusarium commune Causing Strawberry Root Rot in Southcentral China

Strawberry plants and fruits are vulnerable to infections by a broad range of pathogens and pests. However, knowledge about the epidemiology of pathogens causing strawberry diseases is limited. In this study, we analyzed Fusarium commune, a major fungal pathogen causing strawberry root rot, from diseased strawberry root tissues in southcentral China. A total of 354 isolates were obtained from 11 locations that spanned about 700 km from both south to north and east to west. Multilocus genotypes of all isolates were obtained using seven polymorphic simple sequence repeat markers developed in this study. Our analyses revealed significant genetic diversity within each of the 11 local populations of F. commune. STRUCTURE analysis revealed that the optimal number of genetic populations for the 354 strains was two, with most local geographic populations containing isolates in both genetic clusters. Interestingly, many isolates showed allelic ancestry to both genetic clusters, consistent with recent hybridization between the two genetic clusters. In addition, though alleles and genotypes were frequently shared among local populations, statistically significant genetic differentiations were found among the local populations. However, the observed F. commune population genetic distances were not correlated with geographic distances. Together, our analyses suggest that populations of F. commune causing strawberry root rot are likely endemic to southcentral China, with each local population containing shared and unique genetic elements. Though the observed gene flow among geographic regions was relatively low, human activities will likely accelerate pathogen dispersals, resulting in the generation of new genotypes through mating and recombination.


Introduction
The commercially cultivated strawberry (Fragaria × ananassa) is an economically important berry fruit crop worldwide. It is rich in vitamins, minerals, anthocyanins, antioxidants, and other nutrients, and is a highly recommended healthy food for humans [1]. Though wild strawberries are found on all continents, most are of limited agricultural and economic importance. The first strawberry plants representing the current commercial strawberry cultivars were bred in France in the mid-1700s using germplasm originated in the Americas [2]. These garden varieties were then spread across Europe and introduced to other places, including to China at the beginning of the 20th century. The area of planting strawberry has increased greatly in recent years. According to data released by the Food and Agriculture Organization (FAO), from 2008 to 2018, global strawberry production increased by 39.4%, with total production reaching about 33,276 million tons in 2019 [3]. southcentral China. Furthermore, this information was used to infer the epidemiology of Fusarium root rot and to suggest potential future areas of research.

Sample Collection
From April to May 2019, we collected strawberry plants with typical symptoms of root rot disease (wilting and discolored roots) from 11 sites in the following nine counties/municipalities in southcentral China: Chenzhou, Hengyang, Xiangtan, Yueyang, Yiyang, Changde, Zhangjiajie, and Xiangxi in Hunan Province, and Laifeng in Hubei Province. The sampling sites spanned from Chenzhou, Hunan, in the south to Yueyang in the north, and from easternmost Hengyang and Chenzhou to Xiangxi in Hunan and Hubei's Laifeng in the west. The sampled sites included a variety of terrains and climates ( Figure 1). Details about these sites are presented in Table 1, including their location, latitude and longitude, time of sampling, and abbreviated code used elsewhere in this study. Among these 11 sites, three (GY, HN, and XX) are situated in the lowland rolling hill regions; four (YJ, HR, WL1, and WL2) are in the flood plains of Dongting Lake; and four (YD1, YD2, LS, and LF) are in the mountains. At each site, ten wilted plants were collected, with individual diseased plants located at least 5 m from each other. The diseased plants were taken back to the laboratory for cleaning and screened for pathological features associated with root rot, including blackened fibrous roots, brown rot spots on root stems, and red or blocked main root pith. The confirmed diseased plants were stored in the refrigerator at 4 • C until the tissue samples were processed for isolating the putative infecting pathogens.

Isolation of F. commune Strains
To isolate the infecting pathogen, the strawberry plant root system was divided into three parts, corresponding to fibrous roots, main root epidermis, and main root stem pith. From each of the 11 sites, 30 pieces of diseased tissue were taken for each of the three root tissue types for pathogen isolation. Here, for samples from fibrous roots, each piece of diseased tissue represented a distinct infected fibrous root. For samples from main root epidermis, each piece of diseased tissue represented a distinct site along the main root. Similarly, for samples from the main root pith, each piece of diseased tissue represented a distinct section within the main root stem. Up to five pieces of infected tissues from each plant for each of the three root tissue types were used for isolating the infecting pathogen. The different root tissue samples were used to investigate whether there were tissue-specific differences among the alleles and genotypes of the infecting pathogen populations.
For pathogen isolation, strawberry roots with typical symptoms of root rot were washed with clean tap water to remove soil remnants and soft rotten tissues. Each diseased tissue was then surface sterilized with 75% ethanol for 1 min, followed by soaking in 2% sodium hypochlorite solution for 10 min, and washed with sterile pure water 3 times. The surface-sterilized diseased tissue was then dried using sterile filter paper, cut using a sterile scalpel into small pieces of about 0.01 cm 3 , transferred onto a potato dextrose agar (PDA) medium plate containing 50 µg/mL streptomycin, and placed in an incubator at 25 • C. Three tissue pieces were placed on each plate. When the fungal colony grew to about 2 cm in diameter, a sterile toothpick was used to pick a small clump of hyphae from the edge to cultivate pure culture. Fungal colonies with vigorously growing aerial hyphae and macroconidia resembling a spindle were selected for downstream analyses.

DNA Extraction, PCR Amplification, and Species Identification
For species identification, we followed the method of O'Donnell et al. [29] and obtained the DNA sequence of the translation elongation factor 1 α subunit (tef1α) gene from each isolate. Briefly, mycelia were collected from 5-day-old cultures grown on PDA for each Fusarium isolate. The total genomic DNA was extracted following the protocol described in Fang et al. [10]. PCR amplification was performed in a final volume of 25 µL containing 1 µL genomic DNA, 1 µL upstream and downstream primers (10 nM), and 12.5 µL Taq PCR MasterMix (Shengong Bioengineering Co., Ltd., Shanghai, China). The tef1α barcode region was amplified with primers EF1 (5 ATGGTAAGGARGAGAAC 3 ) and EF2 (5 GGARGTCAGTSATCATG 3 ). PCR amplification was performed in a T100TM PCR instrument (Bio-Rad Laboratories, Hercules, Wilmington, DE, USA). The initial denaturation was performed at 94 • C for 5 min, followed by 35 amplification cycles. Each amplification cycle included denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s, extension at 72 • C for 1 min, and finally extension at 72 • C for 10 min. The amplification was verified by agarose gel electrophoresis (1× agarose buffer). The amplified PCR products were purified and sequenced by Tsingke Biotechnology (Changsha, China). The tef1α DNA sequence obtained from each isolate was compared with those in the GenBank database. After BLAST comparison, we retrieved the representative GenBank tef1α sequences of Fusarium species that were closely related to our sequences, especially those within the Fusarium oxysporum species complex. These retrieved sequences were combined with our own sequences for phylogenetic analysis to confirm the species identity of our strains, using the MEGA 7.0 software [30]. The maximum likelihood (ML) algorithm was used to generate the phylogenetic tree. Branch support was inferred from 1000 bootstrap repetitions, and gaps in sequence alignment due to nucleotide insertions and deletion were excluded from analysis [30]. The output phylogenetic tree was edited using the online drawing tool iTOL (Interactive Tree of Life, https://itol.embl.de, accessed on 15 March 2022) [31].

Development of SSR Markers for F. commune
Because there is currently no reported multilocus polymorphic marker for F. commune, we first developed new SSR markers for genotyping F. commune strains. To develop polymorphic SSR markers, we first retrieved the genome sequence of F. commune from NCBI. The F. commune genome data were mined using the MISA v2.1 (http://pgrc.ipkgatersleben.de/misa/misa.html, accessed on 1 April 2021) program. The default setting of the MISA program defined SSR markers as containing one of the following: 1 base repetitions ≥10 times; 2 base repetitions ≥6 times; 3 base, 4 base, 5 base, and 6 base repetitions ≥5 times. When the distance between two SSR markers is less than 100 bp, they form a composite SSR marker and were included as a single SSR marker for testing. After SSR marker identification, we used Primer 3.0 (http://primer3.sourceforge.net/) software (version 2.5.0, accessed on 15 April 2021) to design primers (default parameters).
To identify potential polymorphic markers for our strains of F. commune, we selected eight strains representing different tef1α sequence types and screened 30 pairs of SSR primers, following protocols described previously [32]. The information about the 30 primer pairs, including their amplification success and allele numbers among the eight strains, is summarized in Supplementary Table S1. All primers were synthesized by Tsingke Biotechnology (Changsha, China). The primer pairs showing 100% amplification success and with >4 alleles each were chosen for further genotyping the total population. The following seven primer pairs met the criteria: CM15, CM18, CM22, CM23, CM25, CM27, and CM28. The primer sequences and other information about these seven SSR markers are presented in Table 2.

SSR Genotyping for F. commune Isolates
To obtain multilocus genotype information for all strains in the population, one primer in each of the seven SSR primer pairs was fluorescently labeled with FAM or HEX (Tsingke Biotechnology). For each isolate at each SSR marker, the PCR reaction (15 µL) consisted of 7.5 µL 2× Tsingke MasterMix, 1 µL of each primer (10 nM), and 1 µL template DNA. PCR cycles consisted of a 5 min denaturation (at 94°C), followed by 30 cycles of denaturation (94°C), annealing (50-55°C), and extension (at 72°C), each for 30 s. PCRs were terminated after a final 5 min extension at 72°C. Successful amplifications of the SSR markers were confirmed by agarose gel electrophoresis. After confirmation, 5 µL of each amplification product was mixed with 10 µL of a mixture containing highly deionized formamide and rox-500 fluorescent molecular-weight internal standard as recommended by the manufacturer. The mixtures were then denatured at 95°C for 5 min, placed on ice for 10 min, and loaded onto an ABI 3730xl automatic DNA analyzer for capillary electrophoresis. The output files were analyzed using GeneMapper 4.1 software to identify the lengths of amplified fragments of F. commune isolates and using fluorescent marker standards for each run as fragment size references and following the default setting as recommended by the program. An example graph for allele calling is shown in Supplementary Figure S1.

Data Analysis
The strain relationships based on SSR genotype data were inferred based on the UPGMA algorithm using MEGA7 software. For each of the 11 local populations, the mean numbers per locus for the following diversity features were calculated using PowerMarker v3.25: (i) observed alleles, (ii) alleles with a frequency ≥5%, (iii) effective alleles, (iv) the Shannon's diversity index, (v) private alleles, (vi) locally common alleles (Freq. ≥5%) found in 25% or fewer populations, (vii) locally common alleles (Freq. ≥5%) found in 50% or fewer populations, (viii) expected heterozygosity, and (ix) unbiased expected heterozygosity. The optimal number of genetic populations (K) in the total sample was inferred by STRUCTURE 2.3.4. The STRUCTURE analysis used the admixture model, with 10 replicated runs of K = 1 to 10 after an initial burn-in of 100,000 generations followed by a run length of 1,000,000 generations. The optimal K value was identified following the method described in Evanno et al. [33], using Structure Harvester (http://taylor0.biology.ucla.edu/structureHarvester, accessed on 10 January 2022).
For genetic analyses between local geographic populations such as the analysis of molecular variance (AMOVA), pairwise population F ST , and principal components, we used GenAlEx v 6.5 software [34]. Two types of samples were investigated in our population genetic analyses. The first sample type included all 354 strains. The second sample type used clone-corrected sample where only one representative strain of each multilocus SSR genotype was included in each local geographic population. The three types of root-tissuebased F. commune samples (fibrous root, main root epidermis, main rhizome pith) were similarly analyzed to identify their potential contributions to the total observed genetic variation. The overall relationship between geographic and genetic distances between all pairs of the 354 strains was analyzed using the nonparametric Mantel test through GenAlEx v6.5 software (Melbourne, Australia) [34].

Isolation and Molecular Identification of F. commune by tef1α Sequencing
Based on colony and microscopic morphological characteristics and DNA sequences at the tef1α locus, we obtained a total of 354 F. commune isolates. The number of F. commune isolates from each location is shown in Table 3. Among the 11 geographic locations, the number of isolates ranged from 11 in Longshan (LS) to 66 in Langzhou, Wuling District (WL1). F. commune strains were isolated from all three types of root tissues, with 174 strains from full fibrous roots, 120 from the main root epidermis, and 60 from the main rhizome pith. The tef1α sequence of all 354 F. commune isolates have been deposited in GenBank under accession numbers MZ041792-MZ042021. Figure 2 shows the relationships between our representative strains for the 14 tef1α sequence types and those of closely related Fusarium species based on tef1α sequences. The relationships among all 354 strains based on tef1α sequences are shown in Supplementary Figure S2.

SSR Marker Development
Using MISA to screen the whole-genome sequence of F. commune, we identified a total of 2432 SSR loci belonging to the nine types of SSR. The most abundant type was single nucleotide repeats (46.6%), followed by di-(19.2%), tri-(20.4%), quad-(4.2%), penta-(3.0%), and hexa-(1.8%) nucleotide repeats. From these, we randomly chose 30 to design primer pairs to screen for potential polymorphisms using eight geographically diverse strains from our collection. Among these 30 SSR markers, seven showed 100% amplification success, an unambiguous single amplified product for each strain at each marker, and more than four alleles each among the eight tested strains (Supplementary Table S1). Information about the seven SSR loci, including their primer sequences, amplified fragment lengths, repeat types, and allelic lengths are shown in Table 2. BLAST analyses of the seven SSR loci sequences against the assembled F. commune genome sequence showed that these seven SSR loci were all located on different contigs. Among these seven loci, their amplified fragment sizes ranged from~100 to~400 bp. These seven pairs of primers were then used to analyze our samples of 354 F. commune strains.

Allele and Genotype Characteristics of the Seven SSR Markers
Our analyses revealed that a single PCR-amplified product was detected at each SSR locus in each of the 354 isolates, consistent with all strains being haploid and that each amplified sequence was likely present in a single copy in each isolate. In the total sample of 354 isolates, the average number of alleles was 16 per locus, ranging from six (for marker locus CM23) to 36 (for marker locus CM22) ( Table 2). Together, the seven SSR marker loci revealed a total of 115 alleles across the seven loci. All seven SSR loci were polymorphic in all 11 geographic populations.
The seven SSR markers identified 158 multilocus genotypes (MLGs) among the 354 isolates (Supplementary Table S2 and Figure S3). Among the 158 MLGs, 94 were represented by one isolate each while the remaining 64 MLGs were each represented by two or more isolates in our sample. Among these 64 MLGs, 52 were each represented by two or more isolates from the same geographic location only and together these 52 MLGs represented 188 isolates total in our sample. The most frequent MLG was MLG#20, shared by 19 isolates with all of them from Xiangxiang County. For the remaining 12 MLGs, 10 were shared between two of the 11 local populations and included a total of 54 isolates. The remaining two MLGs were each shared among three geographic populations and together they represented a total of 18 isolates. The longest-distance (~700 km) shared genotypes were between Laifeng County in Hubei and Guiyang County in Hunan, where three MLGs were shared (MLG#33, #96, and #103) with MLG#103 shared among 17 isolates between these two locations. The relationships among all 354 isolates based on the SSR genotypes, including the distributions of the shared MLGs among geographic and root tissue samples, are depicted in Supplementary Figure S3 and Table S2.

SSR Marker Development
Using MISA to screen the whole-genome sequence of F. commune, we identified a total of 2432 SSR loci belonging to the nine types of SSR. The most abundant type was single nucleotide repeats (46.6%), followed by di-(19.2%), tri-(20.4%), quad-(4.2%), penta-(3.0%), and hexa-(1.8%) nucleotide repeats. From these, we randomly chose 30 to design primer pairs to screen for potential polymorphisms using eight geographically diverse strains from our collection. Among these 30 SSR markers, seven showed 100% amplification success, an unambiguous single amplified product for each strain at each marker, and more than four alleles each among the eight tested strains (Supplementary  Table S1). Information about the seven SSR loci, including their primer sequences, amplified fragment lengths, repeat types, and allelic lengths are shown in Table 2. BLAST analyses of the seven SSR loci sequences against the assembled F. commune genome sequence showed that these seven SSR loci were all located on different contigs. Among these seven loci, their amplified fragment sizes ranged from ~100 to ~400 bp. These seven pairs of primers were then used to analyze our samples of 354 F. commune strains.

Allelic and Genotypic Variations within Geographical Populations of F. commune
Our analyses revealed that overall, there was considerable genetic diversity of F. commune causing strawberry root rot within individual local populations as well as in the total sample in southcentral China. For each of the seven SSR loci, two or more alleles were found within each of the 11 local geographic populations. In addition, private allele(s) were found within 10 of the 11 local geographic populations. The only exception was YD2, where no private allele was found while in the WL1 population a total of nine private alleles were found at the seven SSR loci. Overall, among the 11 local populations, the YD1 population had the lowest numbers of observed alleles and effective alleles, and the lowest gene diversity per locus. In contrast, the GY population had the highest number of observed alleles (7.9 per marker locus) while the YJ population had the highest number of effective alleles (4.3 per marker locus) and the highest expected heterozygosity within the population. The details about allelic richness and allelic frequency distribution patterns within each of the 11 local geographic populations are shown in Table 4. Together, our anal-yses revealed that there was high-level allelic diversity, with indigenous genetic elements within most of the 11 local populations of F. commune analyzed here.  Similar to the diversity of alleles observed within each geographic population, a range of MLGs was found among the 11 local populations of F. commune. The lowest number of MLGs was found in the LS and YD1 populations (with 7 MLGs each) while the highest was found in the WL1 population (28 MLGs). However, the observed variation in the number of MLGs was likely due to sample size differences among the local populations because the number of MLGs was positively correlated with sample size (Pearson correlation coefficient = 0.88, p < 0.05). As described in the previous section, among the total 158 MLGs, most (i.e., 146 of the 158 MLGs) were only found in one of the 11 local populations while 12 were shared between two to three local populations. Two local populations (XX and YD1) shared no MLG with other geographic populations. These two local populations were about 300 km from each other and located close to the middle of our sampled regions (Figure 1). In contrast, the LF population located at the northwest corner of our sampled region contained eight shared MLGs with other local populations, including the GY population located in the southeast corner of our sampled sites. The other eight local populations contained 1-4 shared MLGs each. The total number of MLGs, private number of MLGs, and shared number of MLGs within each of the local population are shown in Supplementary Table S3.

Relationships among Geographic Populations of F. commune
In this study, the relationships among populations based on allele and genotype frequencies were analyzed for two sample types, one included the total samples and the second included clone-corrected samples. In the total sample without clone-correction, the analysis of molecular variance (AMOVA) revealed that geographic separation among the 11 local populations contributed 13% to the total genetic variance while individual populations contributed 87%. Both the within-and among-population contributions to the total genetic variation were statistically significant (p < 0.001). Though still statistically significant (p < 0.001), after clone-correction the percentage of contribution to the total genetic variance by geographic separation was reduced to 6% (Table 5). In the pairwise local population comparisons without clone-correction, we found that all 55 ((11 × 10)/2 = 55) pairwise comparisons showed statistically significant differences at p < 0.05, with 53 of the 55 total pairwise population comparison p values smaller than 0.001 ( Table 6). The pairwise F ST values ranged from 0.036 (between LS and WL2) to 0.333 (between YD1 and XX). Here, the F ST values of 0.036 and 0.333 represent the percentages (3.6% and 33.3% respectively) of the total genetic variance contributed by geographic separation between pairs of local population. Overall, based on the total sample of 354 isolates, the YD1 population showed the highest genetic differentiation from other geographic populations (Table 6). The clone-corrected sample showed a similar pattern with all pairwise comparisons showing statistically significant genetic differentiations from each other and with the YD1 population being the overall most differentiated from other geographic populations (Table 7). However, compared to the total samples without clone-correction, all the pairwise F ST values in the clone-corrected sample were lower.  While statistically significant genetic differentiation was found among geographic populations, there was no evidence for significant correlation between genetic difference between strains of F. commune in southcentral China and their geographic distances from each other. Specifically, in the strain-based Mantel test, the R 2 value was 0.0006 and the p value was 0.051 ( Figure 3). A principal coordinate analysis also showed an overall mixing of strains from most geographic areas ( Figure 4A). However, small clusters of strains from individual geographic populations were found in the PCoA graph ( Figure 4A), a result consistent with the genotype-sharing among strains being predominantly from within individual geographic locations as described earlier (Supplementary Figure S3 and Table S2). When PCoA analyses were conducted at the local geographic population level, the separation of geographic populations from each other was clearer than that based on individual strains ( Figure 4B). In Figure 4B, local population YD1 was distinctly separated from other 10 location populations in the first coordinate. While statistically significant genetic differentiation was found among geographic populations, there was no evidence for significant correlation between genetic difference between strains of F. commune in southcentral China and their geographic distances from each other. Specifically, in the strain-based Mantel test, the R 2 value was 0.0006 and the p value was 0.051 ( Figure 3). A principal coordinate analysis also showed an overall mixing of strains from most geographic areas ( Figure 4A). However, small clusters of strains from individual geographic populations were found in the PCoA graph ( Figure 4A), a result consistent with the genotype-sharing among strains being predominantly from within individual geographic locations as described earlier (Supplementary Figure S3 and Table  S2). When PCoA analyses were conducted at the local geographic population level, the separation of geographic populations from each other was clearer than that based on individual strains ( Figure 4B). In Figure 4B, local population YD1 was distinctly separated from other 10 location populations in the first coordinate.
Different from the significant geographic contributions to the genetic variation, we found no statistically significant evidence for root tissue-based genetic differences in our samples of F. commune in southcentral China. Specifically, in both the total and the clonecorrected samples, less than 1% of the total genetic variance could be attributed to the separation of the samples into three root tissue types (Table 5, p > 0.10). Together, this result suggests no evidence of strawberry root tissue-specific specialization for strains of F. commune in southcentral China.

STRUCTURE Results
Bayesian clustering using STRUCTURE revealed that the estimated optimal K value was two for the 354 F. commune isolates ( Figure 5). The analysis showed that within each geographic population, there were genetic elements of both genotypic clusters. However, the distributions of isolates belonging to these two genetic clusters differed among the geographic populations. For example, isolates of the YD1 and XX geographic populations predominantly belonged to one genetic cluster (shown as red in Figure 5) while those in the GY and YD2 geographic populations mainly belonged to a different genetic cluster (shown as green). Interestingly, many isolates contained alleles of both genetic clusters. We note that the second-best estimated optimal K value was six ( Figure 5A). Similar to  Different from the significant geographic contributions to the genetic variation, we found no statistically significant evidence for root tissue-based genetic differences in our samples of F. commune in southcentral China. Specifically, in both the total and the clonecorrected samples, less than 1% of the total genetic variance could be attributed to the separation of the samples into three root tissue types (Table 5, p > 0.10). Together, this result suggests no evidence of strawberry root tissue-specific specialization for strains of F. commune in southcentral China.

STRUCTURE Results
Bayesian clustering using STRUCTURE revealed that the estimated optimal K value was two for the 354 F. commune isolates ( Figure 5). The analysis showed that within each geographic population, there were genetic elements of both genotypic clusters. However, the distributions of isolates belonging to these two genetic clusters differed among the geographic populations. For example, isolates of the YD1 and XX geographic populations predominantly belonged to one genetic cluster (shown as red in Figure 5) while those in the GY and YD2 geographic populations mainly belonged to a different genetic cluster (shown as green). Interestingly, many isolates contained alleles of both genetic clusters. We note that the second-best estimated optimal K value was six ( Figure 5A). Similar to that observed for when K = 2, most local geographic populations contained strains belonging to different genetic clusters, including many strains with mixed ancestries (i.e., those with a mixture of colors; Figure 5B and Figure S4).
Genes 2022, 13, x FOR PEER REVIEW 15 of 20 that observed for when K = 2, most local geographic populations contained strains belonging to different genetic clusters, including many strains with mixed ancestries (i.e., those with a mixture of colors; Figures 5B and S4).

Discussion
This study developed an SSR genotyping system for F. commune and investigated the diversity and distributions of SSR alleles and multilocus genotypes of F. commune samples causing strawberry root rots in southcentral China. Our analyses revealed abundant genetic variation within each of the 11 local geographic populations. Though multilocus genotype sharing was found between strains from several different local populations, all 11 local populations were statistically significantly differentiated from each other. However, the levels of genetic differentiation were not significantly correlated with geographic distance. STRUCTURE analysis revealed that the total sample consisted of two genetically distinct clusters, and with all 11 local populations containing elements of both genetic clusters, including many isolates showing evidence of having genetic ancestries from both clusters. Together, our results suggest that the populations of F. commune causing strawberry root rot are highly genetically diverse within most locations in southcentral China. However, gene flow has occurred among the local populations.
Before the name F. commune was introduced as a separate new species in 2003, strains of F. commune were identified and reported as F. oxysporum or the F. oxysporum species

Discussion
This study developed an SSR genotyping system for F. commune and investigated the diversity and distributions of SSR alleles and multilocus genotypes of F. commune samples causing strawberry root rots in southcentral China. Our analyses revealed abundant genetic variation within each of the 11 local geographic populations. Though multilocus genotype sharing was found between strains from several different local populations, all 11 local populations were statistically significantly differentiated from each other. However, the levels of genetic differentiation were not significantly correlated with geographic distance. STRUCTURE analysis revealed that the total sample consisted of two genetically distinct clusters, and with all 11 local populations containing elements of both genetic clusters, including many isolates showing evidence of having genetic ancestries from both clusters.
Together, our results suggest that the populations of F. commune causing strawberry root rot are highly genetically diverse within most locations in southcentral China. However, gene flow has occurred among the local populations.
Before the name F. commune was introduced as a separate new species in 2003, strains of F. commune were identified and reported as F. oxysporum or the F. oxysporum species complex [35]. However, since its description in 2003, F. commune has been reported from many parts of the world, including the US, Japan, and China. This fungal pathogen has been found to cause diseases in a variety of agricultural crops and vegetables such as root rot in lotus, tobacco, potato, kiwi, rice, and cowpeas, leading to significant crop yield losses [36][37][38][39][40][41]. However, so far, F. commune has not been reported to cause strawberry root rot. In contrast, in California, F. oxysporum f. sp. fragariae has been reported to cause of Fusarium wilt of strawberry. It should be noted that the absence of report linking F. commune to strawberry root rot does not mean that F. commune has not caused strawberry root rot. We believe that it is highly likely that some of the previously (before 2003) identified strains of "F. oxysporum" causing strawberry diseases (including root rot) likely belonged to F. commune. Regardless, our results here clearly indicated that a diversity of F. commune genotypes can cause strawberry root rot.
Molecular markers such as SSR have been increasingly used for studying the origins and spread of plant fungal pathogens, including tracking the pathogens responsible for crop disease outbreaks, epidemics, and pandemics [42]. For Fusarium species, SSR markers have been used to study the epidemiology of several plant pathogens such as F. proliferatum [43], F. culmorum [44], F. oxysporum, F. fujimorium [10], and F. mexicanum [45]. SSR markers are based on copy number variations in one or a few nucleotide repeats within specific genomic regions. They typically have much higher mutation rate than base substitutions. Consequently, they have been extensively used for studying the patterns of recent migrations and dispersals [42]. In this study, SSR molecular markers were used for the first time to study the genetic diversity of F. commune populations. The results showed that these seven SSR markers were highly polymorphic and together, they helped reveal the diversity and distributions of F. commune genotypes across 11 local populations in southcentral China. We note that within the seven SSR markers, aside from the nucleotide repeat number differences among alleles, single nucleotide insertions/deletions outside of the repeated regions but within our amplified DNA fragments were also present ( Table 2). Such differences were detected using an automatic DNA analyzer through capillary electrophoresis, contributing to allelic polymorphisms at all seven loci. The seven SSR markers developed here should be able to help future studies of F. commune impacting strawberries in other parts of the world and affecting other crops within and outside of southcentral China.
Overall, our analyses identified many local population-specific genotypes and a few shared ones between/among the 11 local populations. The shared genotypes likely represent recent gene flows between/among local populations. As described previously and above, F. commune can grow saprophytically in the soil and on a diversity of plants [42,46]. Consequently, gene flow in F. commune could be mediated by both natural factors and anthropogenic activities. For example, under certain conditions, F. commune can produce abundant asexual spores (conidia) that can easily become airborne and disperse by wind. However, if wind-aided natural dispersal were prevalent, we should expect a significant correlation between genetic distance and geographic distance among our strains. The absence of such a statistically significant correlation between genetic and geographic distances suggest that wind-aided dispersals were likely uncommon. Alternatively, human travel between regions, strawberry seedling trade, and other agricultural activities could bring strains of F. commune from one location to another. The longest-distance shared genotypes identified here between the GY and LF populations (about 700 km from each other; Figure 1) were likely mediated through anthropogenic activities. However, these two possibilities are not mutually exclusive, and both could have contributed to the observed genotype-sharing. Overall, our current data reported here indicated that anthropogenic ac-tivities likely contributed more than wind-aided dispersal to the observed genotype-sharing among F. commune strains causing strawberry root rot in southcentral China.
Among the 11 local geographic populations of F. commune from southcentral China, the YD1 populations were the most distinct from others. No strain in the YD1 population shared a multilocus genotype with those in any other geographic population (Supplementary Figure S3 and Table S2). YD1 also had the lowest mean number alleles per locus, the lowest number of effective alleles, the lowest expected heterozygosity, and the lowest number of multilocus genotypes (Table 4 and Table S2). At present, the reason(s) for the distinctiveness of the YD1 population is not known. This population was similar to several other local populations in both geographic and ecological characteristics ( Figure 1 and Table 1). Broader sampling is needed in order to understand the potential contributors to the genetic distinctiveness of the YD1 population.
The observed high allelic and genotypic diversities are consistent with the local populations of F. commune existed long before strawberries were introduced to southcentral China less than 100 years ago. Thus, the strawberry root rot caused by these pathogens most likely represents a (relatively) recent shift from a saprophyte or a pathogen on other plants/crops to strawberries. Indeed, the two or more genetic clusters identified in our sample and the diverse genotypes found in most geographic populations suggest that multiple switches by F. commune have likely occurred in southcentral China to colonize and infect strawberries. Furthermore, the differential distributions of different MLGs, with some being more common and more broadly distributed than others, suggest that some of the F. commune genotypes identified here are likely more adapted to strawberries than other F. commune genotypes. Whether the frequently observed genotypes of F. commune are also more frequently found in the soil and/or in other plants at these sites remain to be investigated. In a study of Fusarium graminearum from different sources in Lithuania and Latvia, low but statistically significant genetic difference (raw data, F ST = 0.0245, p < 0.05; clonecorrected data, F ST = 0.0245, p < 0.05) were found between the soil population and those from living plants [47]. However, no statistically significant difference was found among F. graminearum populations from different groups of host plants or from different years. Their results suggested evidence for potential differential infectiousness/pathogenicity among strains in the soil population of F. graminearum [47]. At present, whether a similar phenomenon exists in F. commune in their pathogenicity toward strawberries is not known. To address this question, comparable number of soil strains of F. commune from each of the 11 corresponding strawberry fields analyzed here will need to be isolated and analyzed using the same molecular markers for comparison. In addition, if differences were found, the differentially distributed alleles and genotypes of F. commune strains between these two ecological niches should be tested to confirm their potential pathogenicity differences on strawberries. Regardless, the high-frequency and broadly distributed genotypes identified in our study represent ideal materials from which to investigate the genetic bases of F. commune pathogenicity causing strawberry root rot and as priority test genotypes for developing root rot-resistant strawberries.
Because of structural differences among the three types of root tissues in strawberries, we hypothesized that strains of F. commune infecting each tissue type might be genetically different from the other two types. However, based on the seven SSR markers, no evidence for significant genetic difference was found among the three types of samples of F. commune based on their infecting root tissue types in southcentral China. It should be noted that our data (i.e., the lack of genetic differentiation among F. commune samples from the three root tissue types) do not exclude the possibility that there might be genes in F. commune responsible for a certain degree of root tissue-specific trophism and pathogenesis. Interestingly, despite relatively equal efforts spent in trying to isolate F. commune from the three root tissue types, the isolation success ratio from fibrous roots, the main root epidermis, and the main rhizome pith was about 3:2:1. This ratio likely reflects the infection process of strawberry root rot where fibrous roots are likely the easiest for the pathogen to infect and grow inside, followed by the main root epidermis, and then inside the main rhizome. However, the frequent sharing of F. commune genotypes among the three root tissues and the lack of overall genetic differentiation among them are consistent with most F. commune strains and genotypes identified here being capable of infecting all three strawberry root tissues in southcentral China.
In summary, our study provides the first insights into the genetic diversity and population structure of F. commune causing strawberry root rot in southcentral China. Our analyses revealed high genetic diversity and significant differences among the 11 local populations of F. commune. The SSR markers developed here should help facilitate future studies on the epidemiology of strawberry root rot in other geographic regions and the relationships between F. commune populations from strawberries and other crops. Information from such studies could help identify the origins and route(s) of spread of genotypes, especially the high-frequency virulent genotypes, which in turn can help design better strategies to minimize their spread and disease impact.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/genes13050899/s1, Figure S1: Example of SSR allele scoring using GeneMapper. Figure S2: Relationships among 354 Fusarium commune strains based on tef1α sequences. Figure S3: Relationships among 354 Fusarium commune strains based on data at seven SSR loci. Figure S4: Distributions of geographic strains among six genetic clusters (k = 6) with each color representing a different genetic cluster. The population labels of 1-11 on the X-axis correspond to GY, HN, HR, LF, LS, WL1, WL2, XX, YD1, YD2, and YJ respectively. Table S1: Information about the 30 tested SSR markers using eight strains of Fusarium commune. Table S2: Isolate ID, geographic origin, source site, and SSR genotype of all 354 Fusarium commune strains from strawberry root rot samples in southcentral China. Table S3