Improved PCR primers for the detection and identification of arbuscular mycorrhizal fungi


  • Editor: Jim Prosser

Correspondence: Jaikoo Lee, Department of Biology (Area 3), The University of York, PO Box 373,York YO10 5YW, UK. Tel.: +44 1904 328 631; fax: +44 1904 328 505; e-mail:


A set of PCR primers that should amplify all subgroups of arbuscular mycorrhizal fungi (AMF, Glomeromycota), but exclude sequences from other organisms, was designed to facilitate rapid detection and identification directly from field-grown plant roots. The small subunit rRNA gene was targeted for the new primers (AML1 and AML2) because phylogenetic relationships among the Glomeromycota are well understood for this gene. Sequence comparisons indicate that the new primers should amplify all published AMF sequences except those from Archaeospora trappei. The specificity of the new primers was tested using 23 different AMF spore morphotypes from trap cultures and Miscanthus sinensis, Glycine max and Panax ginseng roots sampled from the field. Non-AMF DNA of 14 plants, 14 Basidiomycota and 18 Ascomycota was also tested as negative controls. Sequences amplified from roots using the new primers were compared with those obtained using the established NS31 and AM1 primer combination. The new primers have much better specificity and coverage of all known AMF groups.


Arbuscular mycorrhizal fungi (AMF) colonize the roots of the majority of land plants and can improve plant growth. They are obligate biotrophs that can only be cultured in the presence of their host plant. Different AMF species usually occur in the same roots, but the limited variation in hyphal morphology in either plant or soil makes identification extremely difficult. The fundamental problems of identification and classification in AMF limit the study of natural communities. Approximately 160 species of AMF have been described by spore characteristics or morphology according to the International Vesicular Arbuscular Mycorrhiza (INVAM) culture collection ( However, in nature there can be great variation in spore morphology even within an AMF species (Walker & Vestberg, 1998), and many AMF may reproduce only vegetatively without producing spores (Helgason et al., 2002). Molecular analysis provides a way around this obstacle as it has the potential to identify actively growing fungi in field root samples independently of morphological criteria. The rRNA genes have been used in the majority of AMF molecular ecology studies (Simon et al., 1992; Clapp et al., 1995; Helgason et al., 1998; Rosendahl & Stukenbrock, 2004), and these have generally agreed with classification based on spore morphology (Morton & Redecker, 2001; Schwarzott et al., 2001; Walker et al., 2004). It has therefore been suggested that rRNA provides suitable tools for identification and phylogenetic study in AMF (Simon et al., 1992; Gehrig et al., 1996; Simon, 1996; Clapp et al., 1999; Schwarzott et al., 2001).

Because selective PCR amplification depends on the specificity of the primers, there have been several attempts to design primers specific to AMF (Simon et al., 1992; Helgason et al., 1998). However, further studies revealed that these cannot amplify sequences from some AMF groups that have been described more recently, and they sometimes amplify the DNA of other organisms (Clapp et al., 1995, 1999; Helgason et al., 1999). PCR primers for subgroups or certain target taxa of AMF are relatively easy to design and have been used successfully in several studies (Sanders et al., 1995; Bago et al., 1998; Lanfranco et al., 1999; Kjøller & Rosendahl, 2000; Redecker, 2000; Gamper & Leuchtmann, 2007). However, comprehensive studies of AMF community structure and biodiversity require PCR primers that can amplify all AMF, and such primers have not hitherto been described.

Primer design is an iterative process as new sequence information becomes available. As the phylogeny of a group of organisms becomes better understood, more reliable primers can be designed. On the basis that Glomeromycota is monophyletic (Schüßler et al., 2001), we supposed that there would be shared derived features of AMF sequences that would distinguish them from non-AMF sequences that will be present with AMF DNA in field samples.

Internal transcribed spacer (ITS) sequences have been used extensively for molecular taxonomy (Redecker, 2000; Renker et al., 2006), but they exhibit a high level of variation within AMF species and even within single spores (Sanders et al., 1995; Lloyd-Macgilp et al., 1996). Thus, it may be hard to find distinctive features shared by all AMF, but not found in other organisms. Therefore, small subunit rRNA (SSU rRNA) gene was used for AMF-specific primers in this study because it is less variable than ITS, but allows enough resolution down to the species level in AMF.

Here, we designed a set of specific PCR primers for all AMF. Because the 3′ region of a primer is critical for specific amplification, the distinctiveness of this region was used to discriminate against non-AMF sequences. In addition, the most variable region of the AMF SSU rRNA gene was selected in order to achieve high sequence resolution within the AMF. The primer sites were compared with all available reference sequences of SSU rRNA gene from AMF. The specificity of the new primers was tested using DNA from AMF spores and AMF-colonized roots from the field and counter-tested using DNA from non-AMF organisms. Also, sequence data from the new primers in this study were compared with sequence data from PCR with primer AM1 (Helgason et al., 1998) using the same DNA samples.

Materials and methods

Plants and fungal materials

Twenty-three distinct morphotypes of AMF spores were obtained from trap cultures in Korea (Table 1) and used to test the new set of primers. The 14 species of Basidiomycetes, 18 species of Ascomycetes and 14 plant species listed in Table 2 were used to test for specificity. Non-AMF were obtained from pure cultures, and plant leaves were collected in the campus of Korea National University of Education (KNUE). AMF-colonized roots of three plant species (Miscanthus sinensis, Glycine max and Panax ginseng) were obtained from the field in Chungbuk, Korea.

Table 1.   Sampling sites and host plants of arbuscular mycorrhizal fungal spores used in this study
Specimen no.*Sampling siteHost plant in fieldHost plant in trap culture
  • *

    Each number corresponds to a number presented in Fig. 5.

0626-2Yeonpung, Goesan, Chungbuk, KoreaGlycine maxSorghum bicolor
0702-3Yeonpung, Goesan, Chungbuk, KoreaMiscanthus sinensisSorghum bicolor
0706-3Buyoung, Cheongwon, Chungbuk, KoreaMiscanthus sinensisSorghum bicolor
0711-6Dolsan, Yeosu, Jeonnam, KoreaMiscanthus sinensisSorghum bicolor
0715-3Sangmo, Goesan, Chungbuk, KoreaAllium fistulosumSorghum bicolor
0715-4Sangmo, Goesan, Chungbuk, KoreaAllium fistulosumSorghum bicolor
0715-5Sanmo, Goesan, Chungbuk, KoreaAllium fistulosumSorghum bicolor
0716-2Jangyeon, Goesan, Chungbuk, KoreaMiscanthus sinensisSorghum bicolor
0716-8Mt Deokyu, Jeonbuk, KoreaMiscanthus sinensisSorghum bicolor
0827-1Muju, Jeonbuk, KoreaMiscanthus sinensisSorghum bicolor
0827-4Cheongwon, Chungbuk, KoreaSorghum bicolorSorghum bicolor
0828-1Buyoung, Cheongwon, Chungbuk, KoreaMiscanthus sinensisSorghum bicolor
0828-4Mt Gyeryong, Chungnam, KoreaMiscanthus sinensisSorghum bicolor
0828-5Mt Gyeryong, Chungnam, KoreaMiscanthus sinensisSorghum bicolor
0904-1Udo, Jeju, KoreaMiscanthus sinensisSorghum bicolor
0904-2Bija forest, Jeju, KoreaMiscanthus sinensisSorghum bicolor
0905-2Mt Hanla, 1100 m elevation, Jeju, KoreaMiscanthus sinensisSorghum bicolor
0905-3Udo, Jeju, KoreaArachis hypogaeaSorghum bicolor
0911-4Mokryeon park, Cheongju, Chungbuk, KoreaZoysia japonicaSorghum bicolor
0912-1Mokryeon park, Cheongju, Chungbuk, KoreaRhododendron mucronulatumSorghum bicolor
0912-2Mokryeon park, Cheongju, Chungbuk, KoreaRhododendron mucronulatumSorghum bicolor
0912-5Jangsu, Jeonbuk, KoreaMiscanthus sinensisSorghum bicolor
0912-8Jangsu, Jeonbuk, KoreaMiscanthus sinensisSorghum bicolor
Table 2.   Specificity test of primers AML1 and AML2
  • *

    +, −/+ or − indicates strong, faint or no amplification observed, respectively.

  • Pure cultures have been maintained in the fungal culture collection in Korea National University of Education (KNUE).

 Sorghum bicolor−/+KNUE campus
Glycine max 
Capsicum annuum 
Chrysanthemum morifolium 
Trifolium repens 
Lagenaria leucantha 
Lilium tigrinum−/+ 
Viola mandshurica 
Sedum erythrostichum 
Elsholtzia ciliata 
Kummerowia striata 
Lonicera japonica 
Glycine soja 
Clematis mandshurica 
 Boletus violaceofuscus−/+Pure culture
Suillus granulatus 
Heimiella japonica 
Amanita vaginata 
Clitocybe clavipes 
Lactarius scrobiculatus 
Agaricus silvaticus 
Cryptoporus volvatus 
Cantharellus lutescens 
Ramaria flaccida 
Sarcodon scabrosus 
Naematoloma fasciculare−/+ 
Coprinus atramentarius−/+ 
Tricholoma matsutake 
 Orchid mycorrhizal fungiFrom orchid roots,/pure culture
Penicillium sp.From Korean nuruk, pure culture
Monascus purpureusFrom Chinese nuruk, pure culture
Monascus kaoliang 
Monascus fuliginosus 
Monascus anka 
Hymenoscyphus sp. (C-1)From Rhododendron sp., pure culture
Hymenoscyphus sp. (E-1) 
Hymenoscyphus sp. (F-1) 
Fusarium equisetiNon-AMF from roots of Glycine max and/or Miscanthus sinensis from Chungbuk, Korea
Fusarium oxysporum 
Myrothecium sp. 
Cordyceps takaomontana 
Chaetomium globosum 
Lewia infectoria 
Dichostereum pallescens 
Keissleriella cladophila 
Melanomma sanguinarium 

Design of AMF-specific primers

A set of 92 near-full-length AMF SSU rRNA gene sequences was obtained from the GenBank database and aligned to determine the pattern of sequence conservation (Fig. 1). A dataset of AMF and non-AMF SSU rRNA gene sequences was prepared (Figs 2 and 3). The representative AMF dataset comprised 28 AMF sequences (Schüßler, 2007) that encompass all AMF phylogenetic groups; one to four individual sequences were selected from each of the eight families. Non-AMF sequences included six from plants frequently used as hosts in AMF studies, and five from root pathogenic fungi (Pyrenomycetes) that frequently occur in samples collected from some habitats (Clapp et al., 2002). Additional non-AMF fungal sequences were obtained from roots of M. sinensis and G. max using NS31 and AM1 or NS31 and the degenerate primer 5′-GGT TTC CCR TRA GGY GCC G-3′, which was modified from AM1 (5′-GT TTC CCG TAA GGC GCC GAA-3′: differences underlined) in order to match published SSU rRNA gene sequences of more AMF. Because sequences directly obtained from PCR products were not long enough to compare primer-annealing sites, the most related sequence in GenBank was used as alternative. These two datasets were used to compare annealing sites with primers VANS1 (Simon et al., 1992) and AM1 (Helgason et al., 1998), and also used to search by eye for the best specific primer sites for all AMF using genedoc (Nicholas et al., 1997). Potential sites were selected not only for agreement with AMF sequences but also for mismatches with non-AMF sequences, particularly at the 3′ end. Preference was given to longer amplified products that included the regions that are most variable among AMF sequences. The two best specific priming sites for AMF were found near 300 and 1100 bp on the SSU rRNA gene (Fig. 1) and named AML1 and AML2, respectively. These sites were compared with all 113 AMF SSU rRNA gene sequences in GenBank ( that included these regions. Oligonucleotides for the new primers were synthesized at Bioneer, Korea.

Figure 1.

 Variation of AMF SSU rRNA gene and position of PCR primers. The nucleotide diversity, Pi, is the average number of differences per base between sequences compared pairwise, plotted for 10-base windows (Nei, 1987, equations 10.5 or 10.6). An alignment of 92 published AMF sequences was used. Amplified sequence length is about 1100, 795 and 550 bp with primer pairs NS1–NS4, AML1–AML2 and NS31–AM1, respectively.

Figure 2.

 Comparison of primer AM1 annealing site with AMF (top) and non-AMF (bottom) 18S rRNA gene sequences. Sequences of non-AMF, matching with those obtained by amplification from Miscanthus sinensis and Glycine max roots in this study, are marked with asterisks.

Figure 3.

 Comparison of primer AML1 and AML2 annealing sites with sequences of representative AMF (top) and non-AMF (bottom) SSU rRNA gene. Sequences of non-AMF, matching those obtained by amplification from Miscanthus sinensis and Glycine max roots in this study, are marked with asterisks.

Molecular analysis

DNA was extracted from non-AMF, AMF-colonized roots and plant leaves using cetyltrimethylammonium bromide (Ausubel et al., 1999). DNA from single AMF spores was extracted by crushing spores in a PCR tube using a needle, and used directly as template for PCR. Partial SSU rRNA gene fragments were amplified using nested PCR (Van Tuinen et al., 1998) with the universal eukaryotic primers NS1 and NS4 on a MyGene32 (Bioneer Inc., Seoul, Korea). PCR was carried out using 0.1 mM dNTPs, 10 pmol of each primer, 5 U of Taq DNA polymerase and the supplied reaction buffer (Promega Inc., Seoul, Korea) in total volume of 20 μL as follows: initial denaturation at 94 °C for 3 min, followed by 30 cycles at 94 °C for 30 s, 40 °C for 1 min, 72 °C for 1 min, followed by a final extension period at 72 °C for 10 min.

The first PCR product was diluted 1/100 with 1 × Tris EDTA (TE) buffer. The dilutions were used as template DNA in a second PCR reaction performed using the newly designed primers AML1 (5′-ATC AAC TTT CGA TGG TAG GAT AGA-3′) and AML2 (5′-GAA CCC AAA CAC TTT GGT TTC C-3′) as follows: 3 min initial denaturation at 94 °C, followed by 30 cycles of 1 min denaturation at 94 °C, 1 min primer annealing at 50 °C and 1 min extension at 72 °C, followed by a final extension period of 10 min at 72 °C. For comparison, the same dilutions of the first PCR product were amplified using primers NS31 and AM1 (Helgason et al., 1998) with the same PCR regime except annealing at 58 °C. Primer positions used in this study are shown in Fig. 1. PCR products from AMF spores were sequenced directly after purification using PCR purification kit (Bioneer). In the case of amplifications from plant roots, bands of around 550 bp for NS31-AM1 and 800 bp for AML1–AML2 were cut out and DNA was extracted with a gel purification kit according to the manufacturer's instructions (Bioneer). Purified DNA was cloned into pGEM-T Easy Vector (Promega) and transformed into Escherichia coli DH5α. Thirty positive transformants were selected randomly from each of the second PCR amplification reactions and digested by restriction enzymes AluI, HinfI and AsuC21 according to the manufacturer's instructions (Bioneer). One clone of each restriction fragment length polymorphism (RFLP) type was screened and sequenced using sequencing primers SP6 and T7 on an ABI PRISM 3730 × l DNA Analyzer System at Macrogen (Seoul, Korea).

The effect of primer mismatch on PCR amplification efficiency

To determine the effect on PCR amplification of base mismatches in different positions within a primer, we set up a test based on the DNA of a cloning vector and sequencing primers. T7 promoter primer as forward primer was coupled with T3 promoter primer or one of 13 primers modified from the T3 primer in the 5′, middle or 3′ region (Table 3). The predicted melting temperatures of these sequences were all in the range 46.6–48.1 °C. Oligonucleotides were synthesized at Bioneer. pBluescript II KS(−) vector DNA (Koram Biotech Inc., Seoul, Korea) was amplified as a template with 18 cycles of PCR at an annealing temperature of either 46 or 55 °C.

Table 3.   The primers used to test PCR amplification efficiency
PrimerSequence (5′→3′)TmAmplification*
  • *

    See Fig. 4.

  • T7 primer (5′-GTA ATA CGA CTC ACT ATA GGG C-3′) was used with each of these reverse primers in turn. Capital bold letters indicate the bases changed from the original primer T3.


Effectiveness of new primers for field studies

Two plant species, Anthriscus sylvestris and Phleum pratense were sampled from grassland at Ravenstonedale, Kirkby Stephen, Cumbria, UK. Roots dried at 50 °C for 16 h were ground in a Mixermill MM301 (Retsch) at 24 Hz, three times for 3 min. DNA was extracted from 20 mg of milled roots using plant DNA extraction kit (Qiagen) followed by ethanol precipitation, and redissolved in 20 μL TE buffer. After optimization, the following standard protocol was adopted for amplification of AMF SSU rRNA gene directly from root DNA using primers AML1 and AML2. Each 20-μL reaction mixture contained 1 μL of 1/10 diluted root DNA sample, 10 pmol of each primer, 0.725 U of Taq polymerase (Qiagen) in the manufacturer's reaction buffer and 0.2 mM of each dNTP (Invitrogen). PCR was carried out in a PTC100 (MJ Research) using following cycle regime: 94 °C for 15 min, followed by 30 cycles of 94 °C for 30 s, 58 °C for 40 s, 72 °C for 55 s and additional extension at 72 °C for 5 min. PCR product (5 μL) was loaded and checked on a 1% agarose gel. Cloning and sequencing was carried out as described above.

Data analysis and nucleotide sequence accession numbers

All sequences obtained from this study were aligned using clustalx (Thompson et al., 1994) along with the representative 28 AMF sequences from GenBank. The aligned SSU rRNA gene dataset was trimmed to the primer terminal ends (c. 800 bp). Neighbour-joining (NJ) and maximum likelihood phylogenies (ML) were constructed using njplot and phyml, respectively. Distances for the NJ tree were computed using the Kimura 2-parameter model with 1000 bootstraps. For the ML analysis, general time reversible (GTR) nucleotides substitution model was used with 100 bootstraps using nonparametric analysis. A consensus phylogenetic tree was computed using Mortierella polycephala (accession no. X89436) and Endogone pisiformis (accession no. X58724) sequences as outgroups. DNA sequences obtained in this study were registered in GenBank under the accession numbers EU332706EU332743.


Comparison of priming sites

The annealing site of primer VANS1 (Simon et al., 1992) has multiple mismatches with the AMF dataset (data not shown), while AM1 showed several mismatches with some AMF sequences (Fig. 2). Notably, disagreement at the one or two 3′ terminal bases of AM1 was found with the basal lineages of the AM fungal group, Archaeosporaceae, Ambisporaceae and Paraglomaceae (AA vs. AG or GG, Fig. 2). Primer AM1 also showed disagreements with the non-AMF dataset, but this primer still amplifies several non-AMF sequences. Interestingly, non-AMF sequences that could be amplified using primer AM1 (marked × in Fig. 2) generally match at the 3′ terminal region of primer AM1, although there are several mismatches in the middle or 5′ region, consistent with the importance of matching the primer 3′ terminal for PCR amplification. By contrast, the new primers AML1 and AML2 showed much better matches with all AMF sequences in the dataset but mismatched the non-AMF sequences in several positions, especially in the 3′ terminal one or two bases (Fig. 3). If agreement in the 3′ terminal region of the primer is critical for specific amplification, then we would expect that these non-AMF sequences would not be amplified. A wider comparison of priming sites with 113 SSU rRNA gene of AMF revealed that all sequences were well matched with primers AML1 and AML2 (data not shown). Although a few mismatches were found in the middle or 5′ regions, these would not be expected to prevent amplification (Fig. 3). However, one base mismatch in the 3′ terminal was observed in three accessions (Y17634, AJ006801, AJ006800) of one species, Archaeospora trappei (Fig. 3).

Decisive effect of match at primer 3′ end for specific amplification

Although the degenerate primer 5′-GGT TTC CCR TRA GGY GCC G-3′ had been intended to amplify more AMF sequences, all sequences obtained from plant roots by use of this primer were non-AMF sequences: Lewia infectoria, Keissleriella cladophila, Melanomma sanguinarium, Fusarium oxysporum from M. sinensis roots and F. oxysporum, Myrothecium sp., Dichostereum pallescens from G. max roots. Presumably, this was because these non-AM fungi were abundant in the roots samples and the 3′ terminal of this primer matches these non-AMF sequences too (see the sequences of these non-AMF in Fig. 2). It is clear that effective specific primers must not only match AMF sequences, but they must also be designed to avoid non-AMF.

Our investigation of the effect of primer mismatches showed that PCR amplification intensity was significantly affected by the position of mismatched bases. Mismatches in the middle (H11, H12 and H13 in Fig. 4) or the 5′ terminal (H1, H2, H3, H7 and H8 in Fig. 4) did not greatly affect amplification, but amplification decreased dramatically when there were mismatches in the 3′-terminal region (H4, H5, H6 and H9 in Fig. 4). When the two 3′-terminal bases were mismatched, no amplification was observed even with an annealing temperature as low as 46 °C (data not shown). Although H10 does not have a 3′-terminal mismatch, weak amplification might reflect its potential to form an interfering hairpin structure (ccctc … gaggg). These results confirm that matching at the 3′ region of the primer acts as a critical factor for PCR amplification, and must be considered when designing specific primers.

Figure 4.

 Amplification test using primers with different mismatching base positions. Amplification of pBluescript II KS(−) vector DNA was conducted with T7 forward primer and 18 cycles at annealing temperature 55°C. Each lane corresponds to a reverse primer listed in Table 3.

Coverage of Glomales by new PCR primers

The new primers AML1 and AML2 generated a single clear PCR product of around 800 bp from each of 23 different AMF spore morphotypes, and from colonized roots. Phylogenetically, the sequences of all these products fell within the phylum Glomeromycota (Fig. 5). Six different families were represented, including the Ambisporaceae and Paraglomaceae, which are not amplified using primer AM1. Only the Gerdemanniaceae and Archaeosporaceae were not represented, as appropriate samples were not available.

Figure 5.

 Phylogenetic tree of AMF partial SSU rRNA gene sequences representing the AMF coverage by primers AML1 and AML2. * and † represent amplified sequences from AMF spores and AMF-colonized roots, respectively. Capital letters M, G and P represent the sequences from Miscanthus sinensis, Glycine max and Panax japonicus roots, respectively. Origin of the AMF spores is given in Table 1. Sequences amplified by the new primers AML1 and AML2 and representative reference sequences from each group are included in this analysis. The support values shown at the branches are from NJ (1000 bootstraps)/ML (100 bootstraps) analyses, and thick lines indicate over 95% bootstrap support by both NJ and ML.

Specificity of new PCR primers

Primers AML1 and AML2 showed a good ability to avoid amplification of non-AMF DNA. Even with less stringent PCR conditions (annealing at 50 °C), only two plant samples (Sorghum bicolor and Lilium tigrinum) out of 14 species and three Basidiomycota (Boletus violaceofuscus, Naematoloma fasciculare, Coprinus atramentarius) out of 14 species were very faintly amplified, and nothing was amplified with 18 species of Ascomycota (Table 2). Nonspecific amplification would be further reduced with more stringent PCR conditions because several mismatches were found with their sequences (e.g. S. bicolor in Fig. 3). Sequenced PCR products obtained using primers NS31 and AM1 revealed nonspecific amplification. Eight AMF sequences (17 clones, 60.7%), three non-AM fungal sequences (10 clones, 35.7%) and one plant sequence (one clone, 3.6 %) were amplified from roots of M. sinensis (total 28 clones), and one AMF (one clone, 3.3%) and three non-AM fungal (29 clones, 96.7%) from roots of G. max. By contrast, with the same DNA samples, non-AMF DNA was not detected among 58 clones using primers AML1 and AML2, except a single clone of plant DNA from M. sinensis, suggesting the effective ability of the new primers to avoid nontarget DNA. The reduced incidence of nontarget amplification with primers AML1 and AML2 rather than NS31 and AM1 was highly significant (χ2=19.68 in M. sinensis, and χ2=50.34 in G. max, both P<10−5, 1 d.f.). From the roots of P. ginseng, four different RFLP sequence types of AMF were found using the new primer set, and four out of 12 clones represented Paraglomus sequence (P01 sequence in Fig. 5).

Effectiveness of new primers for studying field communities of AMF

DNA was extracted from the root systems of 15 plants of A. sylvestris and 15 of P. pratense taken directly from the field. Fungal rRNA gene sequences were directly amplified from these samples by PCR using the new primers AML1–AML2 (Fig. 6) and also, for comparison, with the widely used primers NS31–AM1. All of the sequences obtained with AML1 and AML2 were in the Glomeromycota clade (98 clones), whereas 6.5% of sequences from NS31 and AM1 were non-AMF or unknown sequences. It is noteworthy that about 20% of sequences obtained with AML1 and AML2 represented Paraglomus, but no sequences of this AMF group were obtained using NS31 and AM1.

Figure 6.

 Direct PCR amplification from two plant roots using the primer pair AML1 and AML2. Lane 1, Anthriscus sylvestris; lane 2, Phleum pratense; M, DNA smart ladder (Eurogentec).


Since the study conducted by Helgason et al. (1998), the primer pair NS31 and AM1 has contributed substantially to the field-based study of AMF communities. AMF sequences from various geographic locations (Öpik et al., 2006) or different ecosystems (Daniell et al., 2001; Husband et al., 2002a, b; Öpik et al., 2003; Scheublin et al., 2004; DeBellis & Widden, 2006) have been examined using NS31–AM1, and have provided important insights into the relationship between AMF and plants in different environments. However, primer AM1 can amplify only the ‘classical’ AMF (Daniell et al., 2001), now classified in the orders Glomerales and Diversisporales, and not the more divergent Paraglomerales and Archaeosporales (Schüßler et al., 2001). Under some circumstances, it will also amplify non-AMF sequences (Helgason et al., 2002; Douhan et al., 2005; Rodriguez-Echeverria & Freitas, 2006; Santos et al., 2006; Santos-González et al., 2007). Therefore, studies based on primer AM1 are studies of the Glomerales and Diversisporales; the other AMF orders are not detected. This could result in a significant bias in determination of AMF diversity (DeBellis & Widden, 2006). Moreover, the potential to amplify non-AMF sequences when AMF colonization is weak means that care must be taken when using nonsequence-based methods such as terminal RFLP (Vandenkoornhuyse et al., 2003; Douhan et al., 2005). The new primers that we have designed address these issues and allow a more accurate representation of the AMF community in field studies.

We ascertained that mismatch of one or two bases in the middle of the primer or in the 5′ region matters little to PCR amplification, but mismatch of one or two bases in the 3′ region prevents amplification (Table 3 and Fig. 4); hence, we applied this as a strategy for specific PCR primer design. The duplex at the 3′ end of the primer is the substrate for the extension reaction, and therefore a good match to the template in this region is an essential prerequisite for PCR amplification. This is a well-recognized principle of PCR primer design; for example, the effect of different 3′-terminal mismatches on primer extension has been investigated in detail (Huang et al., 1992), and such mismatches have been implicated in biased amplification of bacterial sequences from community samples (Hongoh et al., 2003). We chose representative SSU rRNA gene sequences of AMF from each phylogenetic group (Schüßler, 2007); to encompass all AMF, the primer sites should at least agree with these representative sequences. Therefore, agreement with all representative sequences but disagreement with non-AMF sequence in the 3′ region of the primer were our basic criteria for designing Glomeromycota-specific PCR primers. Despite efforts to satisfy all the conditions, we could not find a specific primer site that completely matched all AMF SSU rRNA gene sequences currently in GenBank. However, the best priming regions that we found were a good match to all 110 AMF sequences except for three sequences of the species A. trappei, while differing significantly from all non-AMF sequences that were examined. Archaeospora trappei sequences matched the primers except for a mismatched base at the 3′ terminal of the forward primer, AML1. Because this mismatch is in the critical 3′ position, amplification of this species would not be expected under stringent PCR conditions.

We showed that the primer pair that we have designed has significant advantages in studies of AMF communities. Firstly, it has improved specificity and largely avoids amplification of non-AMF sequences that may be mixed with AMF sequences in the same samples. This includes not only non-AMF sequences available in GenBank, but also non-AMF sequences that were amplified from field-grown plant roots using less specific primers. Secondly, it detects the great majority of AMF, including fungi in the families Ambisporaceae and Paraglomaceae, whose diversity and distribution are currently little known because previous primers (Clapp et al., 1995; Helgason et al., 1999) did not amplify their sequences. Sequence data from spores (0905-2 and 0912-8 in Fig. 5 and Table 1) and from the roots of P. ginseng (P01 in Fig. 5) revealed that the new set of primer successfully amplifies these divergent groups of AMF. Thirdly, the region amplified using this set of primers is relatively long (c. 795 bp) and contains the most variable region in SSU rRNA gene, providing relevant resolution between different AMF sequences. The DNA sequence amplified using primers AML1 and AML2 includes the sequence amplified using primer NS31 and AM1, but is c. 240 bp longer. This allows direct comparison with the very extensive database of published sequences obtained using NS31 and AM1, and also offers the possibility of a more reliable phylogenetic placement of some environmental AMF sequences.

The new primers AML1 and AML2 allow robust amplification of AMF sequences directly from field roots (Fig. 6), comparable to that obtained with the widely used NS31–AM1 combination. However, the new primers have two significant advantages: they are much more specific to the Glomeromycota and they provide better coverage across the Glomeromycota. They offer the first opportunity to survey the relative abundance of all the known orders of the Glomeromycota using a single tool, and this will open new windows into the diversity of arbuscular mycorrhizal communities.


We thank Donghun Kim for providing the non-AM fungal materials in pure culture. We also thank Thorunn Helgason for proofreading this manuscript. Testing on field samples was supported by the Natural Environment Research Council.