Genetic structure of populations of Rhizoctonia solani AG-4 from five provinces in Iran
Rhizoctonia solani anastomosis group 4 (AG-4) is a serious pathogen causing damping off and root rot in many important crop plants. A total of 190 isolates of R. solani AG-4 HG-I were collected from host fields in five provinces of Iran. The genetic structure of this pathogen was evaluated using seven microsatellite loci, focusing particularly on geographic differentiation. Most of the multilocus genotypes (MLGTs) were unique, with few MLGTs shared among populations. High to moderate levels of gene flow among populations was indicated by low to moderate differentiation between pairs of populations based on the fixation index (FST). Gametic equilibrium of most pairs of microsatellite loci and moderate genotypic diversity were found for two out of five populations, indicating that these populations were sexually recombining in structure. High genotypic diversity, moderate clonal fractions and site-specific genotypes were consistent with mixed reproductive systems for the remaining populations. The findings of departures from Hardy–Weinberg (HW) equilibrium, gametic disequilibrium and a significant excess of homozygotes in half or more than half of the loci were probably caused by the presence of null alleles and the Wahlund effect. This is the first study to consider the population genetics of the root and crown rot pathogen R. solani AG-4.
Thanatephorus cucumeris is a basidiomycete species complex composed of different genetic or anastomosis groups (AGs) with a distinct degree of host specificity (Gonzalez et al., 2001). To date, 14 AGs have been identified (AG-1 to AG-13 and AG-BI). AG-4 is one of the most widely recognized pathogens of diseases incited by R. solani. In Iran, isolates of R. solani identified as AG-4 cause seed rot, post-emergence damping off and root rot of important crop plants including cucumber, bean, tomato, alfalfa, sainfoin, cotton, soyabean, sugar beet and pistachio (Safaie et al., 1999). The pathogen usually overwinters as mycelium or sclerotia in the soil and in or on infected perennial plants or propagative material. In some hosts the fungus may even be carried in the seed. The fungus spreads via rain, irrigation or floodwater; via tools or anything else that carries contaminated soil; and via infected or contaminated propagative materials (Agrios, 2005). Combined analyses of the internal transcribed spacer (ITS), 28S rRNA large subunit and β-tubulin genes revealed that the clade corresponding to AG-4 is as a phylogenetic species within the R. solani species complex (Gonzalez et al., 2006). Three homogeneous groups (HGs), HG-I, HG-II and HG-III, within AG-4 can be differentiated based on colour of colonies and sclerotia formation on potato dextrose agar (PDA). These HGs were also shown to be three distinct phylogenetic species in a parsimonious tree based on phylogenetic analyses of rDNA-ITS sequences by Gonzalez et al. (2001).
Knowledge of the genetic diversity within and among populations could be useful for investigating patterns of migration and cryptic recombination. Also, information on genetic structure is important for elucidating the evolutionary potential of populations and for making predictions regarding the relative risks posed by the pathogen in terms of fungicide applications and breakdown of resistance genes (McDonald & Linde, 2002).
In most R. solani AGs and subgroups, the asexual stage is often predominant in the life history, although the sexual stage of many R. solani AGs (e.g. AG-1 to AG-5) is frequently observed in agricultural fields (Cubeta & Vilgalys, 1997). Through the production of recombinant genotypes, sexual populations maintain higher genotype diversity than asexual populations that may have the same gene diversity (Ciampi et al., 2008). Several AGs in R. solani possess either homothallic (self-fertile) or bipolar heterothallic mating systems. At least four AGs (AG1-IC, AG-2-2 IV, AG-4 and AG-8) have heterothallic bipolar mating systems controlled by a single genetic factor with multiple alleles (Toda & Hyakumachi, 2006). Homothallic mating behaviour has been reported for several AGs, including AG-1 to AG-4. Field isolates are generally assumed to be heterokaryotic (Cubeta et al., 1993). If recombination plays an important role in the population structure of this pathogen, heterokaryotization through fusion of somatic mycelium from two (monokaryotic) basidiospores ought to occur prior to infection. The in vitro detection of mating reactions and their effects on population structure are not easy to determine for most R. solani populations. Molecular genetic markers can be used to infer mating systems to assess the relative contributions of sexual vs asexual reproduction using powerful statistical tests (Cubeta & Vilgalys, 1997). A clearer understanding of the biology and ecology of R. solani is emerging, but basic questions remain about the nature of populations and individuals for most R. solani anastomosis groups. To date, information has only been available on the genetic structure of R. solani AG-1 IA populations of rice in Texas (Rosewich et al., 1999), Louisiana (Bernardes de Assis et al., 2008), India (Linde et al., 2005) and China (Bernardes de Assis et al., 2008, 2009). There is also documentation of AG-1 IA populations of maize in America (González-Vera et al., 2010), AG-3 in solanaceous plants (Ceresini et al., 2007) and R. oryzae-sativae in populations of rice in California (Chaijuckam et al., 2010).
There is currently no information on the role of genetic population structure and sexual stages in the epidemiology of R. solani AG-4. The aims of this study were to determine the genetic diversity within and between populations of R. solani AG-4 and to look for evidence of recombination within each population. The relative contributions of asexual and sexual reproduction in the pathogen populations were also evaluated. Two null hypotheses were tested: (i) Iranian geographical populations (IGPs) of R. solani AG-4 have no genetic subdivision or differentiation; and (ii) IGPs of R. solani AG-4 are recombining in structure.
Materials and methods
Fungal isolates were recovered from diseased roots, crowns and hypocotyls of cucumber, aubergine, pepper, snap bean, squash and sugar beet collected in 2009 from fields in five provinces (Alborz, Isfahan, Khorasan, Lorestan and Western Azarbaijan) approximately 150–1100 km apart, using stratified random sampling (transect sampling by walking through the field) (Table 1). Samples were taken at the four- to six-leaf growth stage from seven to 10 transects, at least 10 m apart, in each field. A sample consisted of a single plant with brown lesions on its root or crown and one isolate. Samples from different host species <3 km apart were combined together to avoid small sample sizes. The anastomosis group of each isolate was determined using the clean-slide technique as previously described (Parmeter et al., 1969). Colour of colonies, and formation of sclerotia on potato dextrose agar (PDA) were used to identify the subgroup of each isolate. The identities of 30 isolates were confirmed by direct sequencing of polymerase chain reaction (PCR) products of the ITS regions. ITS1, 5·8S rDNA and ITS2 regions were amplified by primers ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) and ITS5 (5′-GGAAGTAAAAGTCGTAACAAGG- 3′) (White et al., 1990). The PCR amplification reactions were performed in a thermal cycler (Techne TC-512) according to the following programme: 15 min at 95°C; 35 cycles of 30 s at 94°C, 45 s at 53°C and 1 min at 72°C; then 10 min at 72°C as a final extension step. PCRs were carried out in 15-μL volumes containing 10–15 ng genomic DNA, 1·5 μL 10 × PCR buffer (100 mm Tris-HCl, 15 mm MgCl2, 500 mm KCl pH 8), 1·5 mm MgCl2, 0·4 mm dNTP, primers at 0·3 μm each and 1 U Taq DNA polymerase (Roche Co.). PCR products were electrophoresed in 1·2% agarose gels in Tris–borate EDTA buffer at 85 V for 1·5 h and visualized by ethidium bromide staining on a UV transilluminator. PCR products were purified with a QIAquick PCR Purification kit (QIAGEN) according to the manufacturer’s instructions. The purified PCR products were sequenced with the same primers used for PCR amplifications (Macrogene Co.) and the sequences were compared with sequences of Rhizoctonia isolates available in GenBank.
Table 1. Rhizoctonia solani AG-4 HG-I populations collected in this study
|ISF||Isfahan||101||Cucumber, aubergine, pepper, snap bean||2009|
|LOR||Lorestan||38||Cucumber, snap bean, squash, sugar beet||2009|
|ALB||Alborz||25||Snap bean, squash, sugar beet||2009|
|KHO||Khorasan||12||Aubergine, sugar beet||2009|
|AZAR||Western Azarbaijan||14||Sugar beet||2009|
DNA extraction and PCR
DNA from R. solani AG-4 HG-I was isolated using a modification of the procedure developed by Murray & Thompson (1980). Twelve co-dominant microsatellite loci designated by Zala et al. (2008) and Ferrucho et al. (2009) were used for genotyping isolates. Amplification of non-specific bands at loci TC-AG3-6, TC-AG3-7, TC-AG3-19 and TC-AG3-29 made it difficult to distinguish main alleles from non-specific bands and caused interpretation difficulties in all isolates. No allele was observed at locus TC-02, so these loci were omitted. Seven polymorphic co-dominant microsatellite loci were used in genotyping each isolate. The primers used are listed in Table 2. PCRs were performed separately for each SSR locus in 15-μL volumes containing 3 μL genomic DNA (final concentration 5–15 ng), 1·5 μL 10 × PCR buffer, 0·4 mm dNTP, primers at 0·3 μm each and 1 U Taq DNA polymerase (Roche Co.). Reactions were carried out in a thermal cycler (Techne TC-512) according to the following programme: 15 min at 95°C; 35 cycles of 30 s at 94°C, 30 s at 60°C and 1 min at 72°C; then 10 min at 72°C as a final extension step. PCR products were separated by electrophoresis in denaturing polyacrylamide gels (6% polyacrylamide, 5 m urea) using a Sequi-Gen GT system (Bio-Rad) and stained with silver nitrate (Sambrook & Russell, 2001).
In all analyses, it was assumed that R. solani AG-4 is a functional diploid (i.e. a dikaryon). All data were consistent with this assumption. Multilocus genotyping data were obtained by considering different PCR product sizes as different alleles at each locus. One or two alleles per locus were present and were scored as homozygote or heterozygote. The data were arranged in two data sets. The first set included data from all isolates. In the second, clone-corrected, data set, isolates with the same multilocus microsatellite genotype within each population were considered only once to eliminate the effect of repeated samplings of the same clone on the association among loci.
A multilocus microsatellite genotype (MLMG) for each isolate was determined using the program genotype (Meirmans & Van Tienderen, 2004). Isolates with the same MLMG were assumed to be clones. The clonal fraction was defined as a proportion of asexually derived fungal isolates and was calculated as 1−[(number of different genotypes)/(total number of isolates)]. Several indices of clonal diversity were computed in genodive (v. 2.0b20) (Meirmans & Van Tienderen, 2004), including: the number of genotypes per population (num) and Stoddart & Taylor’s genotypic diversity, Go = 1/Σpi2, where pi is the frequency of the ith genotype (Stoddart & Taylor, 1988), and its evenness (Go scaled by the maximum number of expected genotypes) with Nei’s genetic diversity (div) corrected for sample size.
Heterozygosity and distribution of gene diversity
The effective number of alleles and expected homozygosity and heterozygosity were calculated according to Hartl & Clark (1997) in popgene (v. 1.32) (Yeh & Boyle, 1997). Nei’s unbiased gene diversity was estimated as n/(n−1) × (1−Σipi2), where p is the observed frequency of the ith allele (Nei, 1978). Using clone-corrected data, allelic richness (as the mean number of alleles per locus) was calculated with fstat (v. 188.8.131.52) (Goudet, 1995) using a rarefaction index to eliminate a dependence on sample size (Petit et al., 1998).
Data were analysed for clone-corrected and uncorrected data sets. The degree of population subdivision and the distribution of gene diversity among geographical populations of R. solani AG-4 HG-I were assessed by pairwise comparisons of populations using analysis of molecular variance (amova). A hierarchical analysis of variance was used to partition the covariance components into that among geographical populations and that within geographical populations. In addition, the pairwise fixation index (FST) between groups of geographical populations was calculated to examine population differentiation. Both analyses were performed by a non-parametric approach with 1023 permutations in arlequin (v. 3.11) (Excoffier et al., 2005). Population assignment was implemented with genodive v. 2.0b20. Likelihood ratio tests (LRTs) were used for five populations based on the Monte Carlo test, with an alpha value of 0·002, 100 replicated data sets, and a total of 19 000 resampled individuals (Paetkau et al., 1995).
Reproductive mode: Hardy–Weinberg and gametic equilibrium tests
To assess the contribution of recombination to the genetic structure of R. solani AG-4 HG-I populations, the association within loci was determined by the Hardy–Weinberg equilibrium (HWE) exact test (Guo & Thompson, 1992) using a Markov chain (chain length: 100 000; dememorization: 1000) in arlequin v. 3.11. The association between pairs of loci was investigated using the likelihood ratio test (Excoffier & Slatkin, 1998) with 1023 permutations in arlequin v. 3.11 (Excoffier et al., 2005). The inbreeding coefficient (FIS) (Weir & Cockerman, 1984) and the presence of null alleles, two possible reasons for deviation from HWE and GE, were also investigated. The inbreeding coefficient (FIS) was calculated using arlequin v. 3.11 (Excoffier et al., 2005) with 1023 permutations of the data set. A significant deficit or excess of heterozygotes was tested for by using the program genepop (http://genepop.curtin.edu.au/) with the MCMC algorithm, dememorization of 1000, 100 batches and 1000 iterations/batch. A clone-corrected data set was used for these analyses. Estimates of the frequency of null alleles were performed with micro-checker v. 2.2.3 (van Oosterhout et al., 2004).
All isolates were assigned to AG-4 HG-I based on morphological characteristics. To confirm these results, the ITS regions of 30 isolates of AG-4 HGI were sequenced. The amplification of rDNA-ITS from all selected isolates using the ITS4/ITS5 primer set produced a single fragment of approximately 700 bp. The ITS sequences of all isolates were 98·0–99·95% similar to R. solani AG-4 HGI sequences in the NCBI database (e.g. accession no. JQ343830.1). The similarity between isolates from different host species was 97·9–99·98%. A total of 190 isolates from five geographical populations were analysed in this study.
Gene and genotypic diversity
Seven primers pairs amplified microsatellite loci from R. solani AG-4 HG-I. Three to seven alleles were observed at each locus across populations. AG-3-9 was the least polymorphic locus and TC06 was the most polymorphic. Measures of genotypic and gene diversity for the five populations are summarized in Tables 2 and 3. A total of 88 MLMGs were found in R. solani AG-4 HGI populations. The clonal fractions ranged from 0·14 in population AZAR to 0·53 in population ISF (Table 3). The expected heterozygosity (HE = Nei’s unbiased gene diversity) varied from 0·44 to 0·51 across populations. The genotypic diversity (Go) ranged from 15·25 in population ISF to 5·01 in population LOR (Table 4).
Table 3. Measures of genotypic diversitya in five populations of Rhizoctonia solani AG-4 HG-I from different geographical regions in Iran
Table 4. Summary of heterozygosity statistics and genetic variation statistics for all loci
Hierarchical distribution of genetic variation and population differentiation
According to amova, distinct geographical populations were significantly differentiated (Table 5). Pairwise FST values among populations ranged between −0·01 and 0·12 for the clone-corrected data set (Table 6). A case of no significant differentiation was detected between population ALB and population AZAR (FST = −0·02, P = 0·92). The highest FST values were found between populations ISF and KHO (FST = 0·12, P = 0·00) and ALB and KHO (FST = 0·11, P = 0·00). Using genodive software, seven individuals were tagged as migrants, based on the pre-population thresholds (data not shown).
Table 5. Hierarchical distribution of gene diversity among Iranian populations of Rhizoctonia solani AG-4 HG-I
|Among populationsa||57·32||0·21||10·73|| FST = 0·11||0·00*|
|Among individuals within populations||307·69||−0·03||−1·78|| FIS = −0·02||0·82|
|Within individuals||336·50||1·79||91·05|| FIT = 0·10||0·00*|
|Total||701·52||1·96|| || || |
Table 6. Pairwise fixation index (FST) of Iranian Rhizoctonia solani AG-4 HG-I populations for clone-corrected data
|ISF|| || || || || |
|LOR||0·05*|| || || || |
|ALB||0·07*||0·01*|| || || |
|KHO||0·12*||0·04*|| 0·11*|| || |
Reproductive mode; Hardy–Weinberg and gametic equilibrium tests
Clone-corrected data were analysed for HWE and disequilibrium between pairs of loci (Table 7). The null hypothesis that population structure conformed to random mating was tested. The number of loci in HWE was one to five. Only population KHO was in HWE for five of seven polymorphic loci. Significant departures from HWE were detected at four of seven loci for populations ALB and LOR and three of seven loci for population AZAR. Population ISF showed significant deviations from HWE for six loci. Non-significant negative FIS values were observed for all five populations. The proportion of loci with significant pairwise deviations from equilibrium was 14·3–38·1% in five populations. Estimates of the frequency of null alleles are given in Table 8. The locus TC_AG3_1, showed both high FIS values and high frequencies of null alleles in samples ALB, KHOR and ISF. No significant evidence was found for the presence of the null alleles at TC01, TC_AG3_9 and TC_AG3_29 loci (Table 8).
Table 7. Hardy–Weinberg equilibrium (HWE) and gametic disequilibrium (GD) tests for populations of Rhizoctonia solani AG-4 HG-I in Iran
Table 8. Summary of population-specific FIS indices per polymorphic locus and estimated frequency of null alleles (r) at seven loci for five populations of Rhizoctonia solani AG-4 HG-I in Iran
| r ||0·06||−0·29||0·00||−0·23||0·01|
| FIS ||0·21**||−0·26||NA||−0·22||0·66**|
| r ||0·15*||0·19*||0·01||−0·27||0·06|
| FIS ||0·32**||0·47**||0·11||−0·32||0·21**|
| r ||0·23*||−0·01||−0·06||−0·09||−0·37|
| FIS ||0·50**||0·06||−0·09||−0·13||−0·30|
| r ||0·26*||0·19*||0·30*||0·27*||0·09|
| FIS ||0·57**||0·44**||0·72**||0·69**||0·28**|
| r ||−0·63||−0·48||−0·62||−1·00||−0·76|
| FIS ||−0·75||−0·72||−0·67||−1·00||−0·85|
| r ||0·10*||0·29*||0·23*||0·09||−0·24|
| FIS ||0·23**||0·68**||0·35**||0·12**||−0·19|
| r ||−0·21||−0·78||−0·05||−1·00||−0·15|
| FIS ||0·12 **||−0·91||0·02**||−1·00||−0·10|
High to moderate gene flow was detected among Iranian R. solani AG-4 HGI populations, consistent with low to moderate population subdivision (FST). The FST values indicated various degrees of significant gene flow among all populations (Table 6). Gene flow causes the establishment of new alleles or genotypes in new locations through introduction, reproduction and survival of the introduced organism. Gene flow tends to homogenize alleles among populations and keeps geographic populations genetically interconnected (Rogers & Rogers, 1999). In the population assignment test, seven cases of migrant genotypes were found among Iranian populations. Two of seven genotypes in population AZAR were assigned to population LOR and one genotype in population AZAR was inferred to population ALB. These isolates were recovered from sugar beet roots. Four genotypes of population ISF were related to population LOR; these were isolated from snap bean roots and crowns. Genotype flow refers to the movement of entire genotypes (usually clones or clonal lineages) between distinct populations and occurs only for organisms that have a significant asexual component to their life cycle (http://www.apsnet.org/edcenter/advanced/topics/Popgenetics/pages/GeneGenotypeFlow.aspx). Long-distance dispersal of asexual propagules of R. solani might be achieved by movement of contaminated machinery, seed or irrigation water (Agrios, 2005).
Previous studies showed that AG-4 has a bipolar heterothallic mating system controlled by a single genetic factor with multiple alleles. Basidiospores may not have epidemiological significance for most Rhizoctonia root and seedling diseases during a growing season but may have important ecological and population consequences (Cubeta & Vilgalys, 1997). The evaluation of whether recombination is taking place in R. solani is important because even a small amount of recombination may have significant effects on population structure (Milgroom, 1996). Several AGs with heterothallic mating systems can also exhibit homothallic behaviour, through a process known as homokaryotic fruiting (Cubeta & Vilgalys, 1997). Both heterothallic and homothallic AGs may also undergo recombination through the process of heterokaryon–homokaryon (di-mon) mating originally described by Buller (Cubeta & Vilgalys, 1997). Sexual spores produced as a result of heterothallic mating might constitute an important aspect of the fungal life history that contributes to the genetic diversity of field populations of R. solani AG-4. Analysis of the variation found for neutral markers in field populations supports this assertion. In this study, intermediate to high degrees of genetic diversity and low to moderate clonal fractions and gametic disequilibrium (GD) were observed. In all populations many MLMGs were unique. It is postulated here that unique MLMGs are generated by sexual reproduction. The lowest clonal fraction was 0·14 in population AZAR. A possible cause for this observation is small population size that minimizes the chances of finding shared clones. Significant Hardy–Weinberg disequilibrium was observed in most of the studied populations. After clone correction, populations ISF, LOR and ALB showed the lowest proportions of loci in HWE (14·28–50% of the loci) and moderate GD (38·1, 23·8 and 14·3% of the locus pairs, respectively). The two remaining populations had higher proportions of loci in HWE (57·14–71·42% of the loci) and a high degree of gametic equilibrium (81·0–95·3% of the locus pairs). The basis for heterozygote deficiency in a population could be attributed to inbreeding, the Wahlund effect (pooling populations with different allele frequencies), selection or null alleles (Rosewich et al., 1999). To test whether non-random mating was responsible for departure from HWE and GD, the FIS (inbreeding coefficient) was calculated, but all populations exhibited non-significant negative values. If inbreeding occurred in populations ISF, LOR and ALB, one would expect similar heterozygosity deficiencies in all markers (genome-wide effect) in each population studied. Two, four and six of the seven analysed loci showed a statistically significant heterozygote deficit in populations ALB, LOR and ISF, respectively. Some of the heterozygote deficits detected in populations ISF (at four of six loci), LOR (at three of four loci), ALB (at two of three loci) and KHO (at one of two loci) were probably caused by the presence of null alleles (Table 8) not amplified because of the mutations at the primer annealing sites (Chapuis & Estoup, 2007). In these populations, samples from different host species were collected (Table 1). Therefore, the heterozygote deficits detected may also have resulted from the Wahlund effect (referring to subpopulations in a sample) (Rosewich et al., 1999).
Gametic disequilibrium can be generated by several processes including linkage, population admixture, genetic drift, population expansion, non-random mating and selection (Milgroom, 1996). Allelic richness values were similar across all populations of the fungus, leading to the conclusion that bottlenecks did not cause the observed disequilibrium. The role of selection as a possible cause for the observed departure from HWE and GD in populations ALB, ISF and LOR was investigated. A neutrality test was performed on microsatellite data and no evidence was detected for selection at any of the seven loci (data not shown). It is suggested that admixture (the Wahlund effect) may be one of the possible causes of observed GD. Overall, the results found in populations ALB, ISF and LOR (the presence of GD and the significant deviations from HWE with heterozygote deficiency) are possibly the result of the host specialization effect (the Wahlund effect) and null alleles.
The observed high clonal fractions and moderate to high genotypic diversity shown here are consistent with a considerable asexual reproductive component and a sexually recombining component of reproduction, respectively. Thus, the overall interpretation is that populations of R. solani AG-4 in Iran have a mixed reproductive mode.
Novel genotypes are generated by sexual reproduction. Once favourable genotypes are formed, selection can act upon such individuals and multiply them through asexual reproduction within fields. Selected genotypes then spread as clones across short distances, via sclerotia or mycelium. Long distance dispersal of sclerotia may be by irrigation water, in soil or on farm equipment.
Mixed reproductive modes were also reported for R. solani AG-1–IA in Louisiana and Brazil (Bernardes de Assis et al., 2008) and for R. oryzae-sativae in California (Chaijuckam et al., 2010). In this study, analysed populations showed moderate to high levels of gene flow and evidence of mixed reproductive systems. According to the risk model framework proposed by McDonald & Linde (2002), they have high evolutionary potential. It is recommended that fungicides should be applied with caution and the use of shared irrigation systems, contaminated machinery or infested seeds should be avoided to minimize the spread of sclerotia and gene flow among populations. This study provided initial data on the population genetics of R. solani AG-4 from various geographical areas in Iran. Investigations are in progress on host differentiation of these populations using more microsatellite loci.
We would like to thank Tarbiat Modares University for financial support and Isfahan University of Technology for providing the laboratory tools and equipment.