Arbuscular mycorrhizal fungi (AMF) are an ecologically important group of plant symbionts and their species richness has been shown to influence plant diversity and productivity (Van der Heijden et al., 1998). Genetic diversity within AMF species is important as genetically different isolates have been shown to differentially affect plant growth and nutrition (Munkvold et al., 2004; Koch et al., 2006). The study of AMF diversity in ecosystems, particularly identifying which AMF species associate with different host plants, requires reliable identification of different AMF. It has long been recognized that identifying AMF across broad geographical ranges requires molecular tools for fast and reliable genotyping directly from soil material.
Previously, genotyping methods for distinguishing AMF species have mostly been restricted to ribosomal DNA (rDNA) sequences. The advantage of these loci is the potential for cross-species amplification using universal primers, and the relative ease of amplification from different material (e.g. colonized root pieces, single spores, etc.). A large body of studies have identified the species composition of AMF communities in many different ecosystems (Öpik et al., 2006; Rosendahl, 2008). However, studies of genetic variability within AMF species are important for understanding the basic biology, genetics and ecology of AMF fungi, which cannot be addressed at the community level. For example, a hierarchical study of genetic variability from the local scale within populations right up to an inter-continental scale is lacking. Such hierarchically designed studies could lay the foundation that will allow us to answer fundamental questions about the biology of AMF, their genetics, whether they form recombinant populations, the amount of genetic exchange among populations, the importance of drift and selection in AMF species, and the distribution of genetic and functional diversity in AMF over different geographic scales, and allow us to examine the co-evolutionary relationships between AMF genotypes and their host plant genotypes.
For most of these applications ribosomal markers are unsuitable because of a lack of sufficient within-species variability and are potentially problematic because of confounding intra-sporal variability (Sanders et al., 1995) and copy number polymorphism (Corradi et al., 2007). A population genetics approach to the study of AMF requires multilocus genotyping of nonribosomal loci. Stukenbrock & Rosendahl (2005a,b) first developed and applied this approach by amplifying three different loci in a large set of spores of three Glomus species harvested from the field. However, ideally, multilocus genotyping should comprise a much larger number of loci. Two simultaneously published studies (Croll et al., 2008; Mathimaran et al., 2008), describing genetic markers for AMF, should now make this possible. Both studies identified multiple loci that were variable among isolates of a commonly studied AMF, Glomus intraradices. Length differences among the alleles were used to identify genetic differences. Part, but not all, of the variation was found in repeat regions, and both studies referred to the markers as either microsatellites or simple sequence repeat markers. The simultaneous publication of the two studies might lead to some confusion for researchers who may now want to use these markers. Here, our aim is to clarify how many new and different loci have actually been identified and which loci are likely to be suitable for population genetics studies, to highlight potential problems with the genotyping techniques used, and to discuss future approaches to their use in AMF population biology.
The study by Mathimaran et al. (2008) identified 18 loci and Croll et al. (2008) showed polymorphism in 13 loci, of which two had previously been identified by Raab et al. (2005). The two studies used similar, but not identical, strategies to identify repetitive DNA stretches by searching publicly available databases (Table 1). Candidate sequences were then amplified in a set of isolates and potential length polymorphism was scored. In both studies, loci were amplified in a number of isolates from different geographic locations. It should be noted that one locus described by Mathimaran et al. (2008) is the same as one polymorphic locus identified by Croll et al. (2008) but has been given two different designations. The variation in two more loci reported by Mathimaran et al. (2008) is documented in previously published work. We hope that Table 1 will help researchers who intend to use these markers to identify the different loci for which primers have been developed and prevent unintentional studies of the same locus under two different names.
|Locus||Accession no.||Database||Function||Type||Length polymorphism||Reference|
|Bg32||CG431930||GSS||Unknown||Probably noncoding||Indels||Croll et al. (2008)|
|Bg42||CG431913||GSS||Unknown||Probably noncoding||(TA) repeat + other indels||Croll et al. (2008)|
|Bg62||CG431880||GSS||RNA polymerase II large subunit||Proximate coding region||(TAAAA) repeat + other indels||Croll et al. (2008)|
|Bg196||CG431972||GSS||Unknown||Probably noncoding||Several repeat motifs + other indels||Croll et al. (2008)|
|Bg235||CG432041||GSS||Unknown||Probably noncoding||Several indels||Croll et al. (2008)|
|Bg273||CG432137||GSS||Unknown||Probably noncoding||(T) + (A) repeats + other indels||Croll et al. (2008)|
|Bg276||CG432062||GSS||Unknown||Probably noncoding||Several indels||Croll et al. (2008)|
|Bg303||CG432175||GSS||Unknown||Probably noncoding||Several indels||Croll et al. (2008)|
|Bg348||CG432294||GSS||Predicted protein of unknown function||Proximate coding region||(TAA) + (TAAA) repeats + other indels||Croll et al. (2008)|
|Bg355||CG432269||GSS||Unknown||Probably noncoding||Several indels||Croll et al. (2008)|
|Nuclear intron||BE603853||EST||Intron in gene of unknown function||Proximate coding region||(T), (A) + (TAA) repeats||Croll et al. (2008)|
|mtLSU int1||AJ973189-193||Standard||Intron in mitochondrial LSU gene||Proximate coding region||Several indels||Raab et al. (2005); Croll et al. (2008)|
|mtLSU int2||AJ973189-193||Standard||Intron in mitochondrial LSU gene||Proximate coding region||Indel||Raab et al. (2005); Croll et al. (2008)|
|Glint01||CG432086+113*||GSS||Unknown||Coding||(AAAT) repeat + other indels||Mathimaran et al. (2008)|
|Glint02||DT883628||EST||Unknown||Coding||(GAA) repeat only?||Mathimaran et al. (2008)|
|Glint03||BI452162||EST||Unknown||Coding||(TTAT) repeat? + other indels||Mathimaran et al. (2008)|
|Glint04||BM959176*||EST||Unknown||Coding||(TTA) repeat? + other indels||Mathimaran et al. (2008)|
|Glint05||BE603957*||EST||Putative cell wall protein||Coding||(TAT) repeat? + other indels||Mathimaran et al. (2008)|
|Glint06||BM959329||EST||Unknown||Coding||(CAT) repeat? + other indels||Mathimaran et al. (2008)|
|Glint07||BE603778*||EST||Unknown||Coding||(TTA) repeat? + other indels||Mathimaran et al. (2008)|
|Glint08 (same asBg348)||CG432294||GSS||Predicted protein of unknown function||Proximate coding region||(AATA) repeat? but see Bg348 above||Mathimaran et al. (2008)|
|Glint09||AM118108||Standard||P-Type IID ATPase||Coding||(AATG) repeat? + other indels||Corradi & Sanders (2006); Mathimaran et al. (2008)|
|Glint10||BM027318||EST||Unknown||Coding||(AATGGT) repeat? + other indels||Mathimaran et al. (2008)|
|Glint11||BI452145||EST||Unknown||Coding||(CAA) repeat only?||Mathimaran et al. (2008)|
|Glint12||BM959214||EST||Unknown||Coding||(CAA) repeat + other indels||Mathimaran et al. (2008)|
|Glint13||BM959443*||EST||Unknown||Coding||(AAT) repeat? + other indels||Mathimaran et al. (2008)|
|Glint14||BM027461*||EST||Unknown||Coding||(T) repeat only?||Mathimaran et al. (2008)|
|Glint15||BM959581*||EST||Unknown||Coding||(T) repeat only?||Mathimaran et al. (2008)|
|Glint16||CG431704+705*||GSS||Unknown||Probably noncoding||(A) repeat only?||Mathimaran et al. (2008)|
|Glint17||CG431789+901*||GSS||Unknown||Probably noncoding||(T) repeat only?||Mathimaran et al. (2008)|
|Glint18 (same as Glint09)||AM118108||Standard||P-Type IID ATPase||Coding||(A) repeat only?||Corradi & Sanders (2006); Mathimaran et al. (2008)|
Locus Glint08 identified by Mathimaran et al. (2008) is identical to locus Bg348 from Croll et al. (2008), even though the primers are located at different distances from the repetitive sequence region. Loci Glint09 and Glint18 identified by Mathimaran et al. (2008) were previously published by Corradi & Sanders (2006) and described as genes encoding P-type IID ATPases. Corradi & Sanders (2006) reported polymorphism in a population of G. intraradices based on a comparison of different alleles at the same locus. Furthermore, the gene was found to exist in two variants in each of several isolates and in three variants within one isolate (Corradi & Sanders, 2006). Locus Glint09 is based on the sequence of the third variant; however, the primers designed by Mathimaran et al. (2008) are not specific for this particular variant. As a consequence, the primers based on locus Glint09 potentially amplify up to three different locations in the genome within a single isolate. Locus Glint18 was identified in an assembled sequence (contig) that matches the P-type IID ATPase variants. However, the resulting consensus sequence does not exactly match any of the original P-type IID ATPase variants, probably as a consequence of the contig being assembled from several different variants (i.e. a chimaeric contig). Consequently, primers for locus Glint18 do not specifically amplify one of the several variants. Loci Glint09 and Glint18 are separated by approx. 500 bp. In our opinion, these two loci are unsuitable for most population genetic studies because of the multi-copy nature of the gene they are located in, unless primer sequences are chosen that restrict the amplification to one variant.
The studies of Mathimaran et al. (2008) and Croll et al. (2008) both describe polymorphic loci exhibiting size differences of 1 or 2 bp among some alleles. Scoring such a polymorphism is potentially problematic even if PCR products are separated on a capillary sequencer, Spreadex polymer or polyacrylamide gels. These methods offer a high resolution of allele length differences, but the amplification of repeat motifs often leads to the presence of stutter peaks (or shadow bands) as a result of DNA polymerase error. Where small length differences are observed among alleles, it is advisable to obtain sequences that verify that the differences are real and not an artifact of the electrophoresis. This was not done for all loci showing 1- or 2-bp differences in the study by Mathimaran et al. (2008) and we suggest more rigorous testing of these differences before using these markers in genotyping studies. If large sets of isolates need to be analysed, the risk of artifacts in the allele identification may be dramatically reduced by using only loci with 3-bp or longer repeat motifs.
Assuming that the length differences are accurate, most of the markers identified by Croll et al. (2008) and Mathimaran et al. (2008) are useful for demonstrating genetic differences among G. intraradices isolates. This does not, however, mean that they are suitable for studying all aspects of AMF population biology. Mutation rates vary across the genome and it is generally assumed that noncoding regions evolve at a higher rate than coding regions, as a result of selective constraints on proteins encoded by the genes. Therefore, it is important to identify the location of the loci in the genome to predict their suitability for particular studies. Mathimaran et al. (2008) mostly identified length polymorphism in expressed sequence tags (ESTs). Repeat motifs identified in ESTs are likely to be under selective pressure to maintain functional integrity of the protein. However, most of the markers reported by Croll et al. (2008) and some of those reported by Mathimaran et al. (2008) originate from sequences obtained in a genome survey, where regions throughout the genome were randomly sequenced. Because of their random location in the genome, these sequences are likely to be outside of coding regions. However, G. intraradices was shown to have a relatively small genome of approx. 15 Mb (Hijri & Sanders, 2004) and, therefore, gene density could be relatively high. Neutral loci are preferable for population genetic studies, as the polymorphism more likely reflects random genetic processes such as mutation, migration or drift. As expected, a majority of the loci from both studies show length polymorphism in the repeat motif. However, a large number of indels and substitutions were also found outside the repeat motif (Table 1). Therefore, the markers do not represent pure simple sequence repeats (or microsatellites) and length differences among alleles should be considered carefully. However, the presence of a large number of substitutions enables researchers to use these markers for a variety of applications such as single nucleotide polymorphism (SNP) genotyping.
Genotyping on a large scale requires amplification of DNA from single spores directly collected from the field, instead of passing through the laborious process of in vitro cultivation. However, the small size of G. intraradices spores poses a challenge for the amplification of genetic markers because of the very low amount of DNA. Stukenbrock & Rosendahl (2005b) and Mathimaran et al. (2008) propose two different approaches to solve this problem. In the first study, a nested PCR was performed and up to five different loci could be amplified. However, it is not known whether this method would perform well with the comparatively small spores of G. intraradices. One additional concern is the number of loci that can be amplified simultaneously. Mathimaran et al. (2008) chose a promising method called whole-genome amplification (WGA), providing a higher number of template copies of each locus. This method is increasingly used for amplification of DNA from single cells (Spits et al., 2006), unculturable bacteria (Stepanauskas & Sieracki, 2007) or filamentous fungi (Foster & Monahan, 2005), including AMF (Gadkar & Rillig, 2005a,b). While the potential exists to create many template loci from minute samples of cells or spores, several factors may bias the WGA. Notably, WGA is very sensitive to template contamination by other microorganisms as a result of the indiscriminate DNA amplification; a very real concern for spores from pot cultures or the soil (Hijri et al., 2002; Corradi et al., 2004). Furthermore, some parts of the genome tend to be better amplified than others, creating a representation bias in the final product and potentially null alleles (Pinard et al., 2006). In order to apply whole-genome amplification to field-collected spores, the method should be rigorously tested by using well-defined in vitro cultivated material as a comparison to whole-genome amplification from single spores of the same culture.
If successfully applied, highly discriminatory markers combined with large-scale hierarchical sampling could elucidate the extent of clonal networks within field sites and resolve patterns of genetic diversity at larger geographic scales. Furthermore, the co-evolution between AMF and their host plants could be studied in detail by identifying spatial distributions of particular genotypes. These areas of investigation have become even more relevant in the context of globally applied inoculum in the absence of data on ecological competitiveness and the potential to persist in the field among native AMF (Schwartz et al., 2006). While the global population genetics of plant pathogenic fungi has received much attention in recent years, studies on plant symbionts will hopefully catch up soon.