ABSTRACT
An inorganic phosphate transporter gene sequence (852-bp section) allowed discrimination between 10 Glomus fungal species represented by 25 strains. It was particularly valuable in differentiating between morphologically similar species with nucleotide and amino acid sequence differences higher than 3%. This gene is proposed as a reliable barcode for the Glomeromycetes.
The morphological identification of species of arbuscular mycorrhizal (AM) fungi in the genus Glomus remains a difficult task. Currently, AM fungi are molecularly characterized using sections of the nuclear rRNA internal transcribed spacers ITS1 and ITS2 and the 5.8S gene. The small-subunit (SSU) rRNA (28), the large-subunit (LSU) rRNA (3, 8, 23), and a combination of all rRNA gene sequences (32), even though useful for glomalean phylogeny, are inappropriate for species identification due to a high degree of sequence variation. The intrasporal variability in ribosomal DNA (rDNA) sequences (16, 17, 20, 27) is sometimes much higher than the interspecific variation (7, 25), and the phylogenetic overlap of sequences across species is a real problem with respect to identification. The use of protein-encoding genes for actin (15), tubulin (9, 10, 19), and P-type H+-ATPase (9, 11) has already been investigated, but with limited success, while genes for ef1-alpha, V-H+-ATPase, and F-ATPase β-subunit (30) seem efficacious only for discriminating between unrelated Glomus species. The expression profiles of a phosphate transporter (PT) gene of three Glomus species have been studied (4, 18), and the gene was considered a promising avenue for the identification of AM fungi (13). We sequenced an 852-bp fragment from the coding region of a transmembrane phosphate transporter gene (14) of 25 Glomus strains (Table 1) from the Canadian in vitro and in vivo Glomeromycota collection. In vitro strains were propagated on Ri T-DNA-transformed carrot (Daucus carota) roots on M medium (12), and in vivo strains of Glomus mosseae and G. coronatum were pot cultured on Allium porrum (30). Total DNA was extracted from spores using the FastDNA kit (Qbiogene Inc.). The 25-μl PCR mix contained 20 mM Tris (pH 8.4), 50 mM KCl, 3 mM MgCl2, 0.25 mM each deoxynucleoside triphosphate (dNTP), 1 μM each primer, 1 unit Taq Platinum polymerase (Invitrogen Canada Inc.), and 3 μl of a 1:10 dilution of 10 ng/μl gDNA. Primer pairs (Table 1) were P3F (5′-ATGTCTACATCCGATAGAGTAAC-3′) and P3R (5′-GGATTTTTATATTCTCCCAATTTATCG-3′) (15), P4F (5′-GAATTRATGATYATYATTGTYGC-3′) and P4R (5′-AYYCTTTCTTTCCTTTCTTCAACG-3′), P6F (5′-AGTATTTGCTATGCAAGGATTT-3′) and P6R (5′-GTCCACCAATGTCTTTTAGTTT-3′), and P7mF (5′-GTATTCGCGATGCAGGGATTC-3′) and P7mR (5′-GGTCCACCAATGTCTTTTAGTTT-3′). The primers were based on the inorganic phosphate transporter gene sequences AF359112, AY037894, and U38650. Primers P4F/R and P7mF/R were designed manually, and primers P6F/R were designed using Primer3 software (26). The parameters for PCR, carried out on an Eppendorf thermocycler, were as follows: 3 min at 94°C; 40 cycles of 30 s at 94°C, 45 s at 50, 54, or 55°C (Table 1), and 90 s at 72°C; and 7 min at 72°C. The PCR amplicons were purified with ExoSAP-IT (Affymetrix Inc.) prior to sequencing on a 3130XL Genetic Analyzer using BigDye v3.1 (Applied Biosystems). Amplicons were sequenced at least twice, resulting in no ambiguous base positions.
Best primer pairs used to amplify an 852-bp section of a phosphate transporter gene from Glomus species and GenBank accession numbers of the obtained ampliconsa
Sequences were aligned using ClustalX v.1.83 (33) and lengths adjusted with Se-Al (version 2.0a11; A. Rambaut, http://tree.bio.ed.ac.uk). Pairwise nucleotide and amino acid sequence identities were determined two by two using BLASTn (1) and BLASTp (2) (http://www.blast.ncbi.nlm.nih.gov). The phosphate transporter DNA substitution model was determined for each codon using the hierarchical likelihood ratio test implemented in MODELTEST 3.06 (22). Bayesian analysis was conducted using MrBayes v. 3.1.2 (24). Four Markov chains were run over 1 million generations, 10,000 trees were saved, and the burn-in period was set to 1,000.
The nucleotide sequences of Glomus irregulare (strain 1; Table 2) (6) and all strains of Glomus sp. A1 (strains 2 to 5; Table 2), including DAOM197198, were identical. Furthermore, these strains share at least 99.2% of their nucleotides and 100% of their amino acids with a Glomus sp. (strains 6 to 10; Table 2). Their similar spore morphologies are consistent with these findings. Strain 10 is also included in G. irregulare, as it shows only a 0.7% nucleotide variation and no amino acid variation with G. irregulare and the slight morphological differences are insufficient to define a separate species. The sequence similarities between G. irregulare and strain 11 are 99.0% for nucleotides and 99.3% for amino acids. The two latter strains are also morphologically identical. Consequently, strains 2 to 11, including the arbuscular mycorrhizal model fungus DAOM197198 (30), should all be considered to be G. irregulare. The original G. intraradices (KS906, strain 14) is clearly distinct from all the other species studied, particularly the isolate DAOM197198, confirming the results of Stockinger et al. (31). The sequences AF359112 and AY037894 in GenBank (18), identified as belonging to G. intraradices, are completely identical to those of G. irregulare, and their identity should be corrected.
Relative similarities between a fragment of 852 bp of the gene encoding an inorganic phosphate transporter of arbuscular mycorrhizal fungi of the genus Glomusa
Morphologically, strains 17 to 20 belong to the same yet-to-be-described species, and their sequence relatedness supports this classification. The small, pale-colored spores of these four strains also differ morphologically from spores of other described species (5). Sequence U38650 (14) is identical to that of strains 17 to 19, making its original identification as G. versiforme doubtful. The next level of sequence similarity occurs between Glomus sp. strain 11 and G. aggregatum (strain 12), which share 97.0% of nucleotides and 96.8% of amino acids. Glomus aggregatum (strain 12) and G. diaphanum (strain 13) share 96.6% of their nucleotides and 96.1% of their amino acids, providing a clear difference between these two closely related morphological species. Glomus clarum (strain 15), G. custos (strain 16), and G. proliferum (strain 21) are clearly segregated from the others (Table 2). All three G. mosseae strains (strains 22 to 24) have 100% identical sequences. A 30% difference in sequence distance with the other species is consistent with the morphological differences between G. mosseae and G. coronatum and the other Glomus species. The phylogenetic tree (Fig. 1) confirms the percent relatedness (Table 2).
Bayesian phylogenetic tree of the nucleotide sequences of an 852-bp phosphate transporter-encoding gene of some Glomus species. Bayesian values of >80% are indicated on the nodes. Strain numbers are followed by GenBank sequence accession numbers. Scale indicates nucleotide substitution per site. Group 1 (groups indicated in bold and with vertical lines on right), Glomus irregulare: yellow spores, ovoid to irregular shape, with a three-layered spore wall made of a semipermanent outer wall, reactive to Melzer's reagent, and with a semiflexible median wall and a laminated inner wall. Group 2, G. aggregatum: similar to G. irregulare except that the inner laminated wall is strongly reactive to Melzer's reagent and the insides of spores are sometimes differentiated. Group 3, Glomus sp.: similar to groups 1 and 2 but with a shell-like rigid outer wall detachable from the inner wall. Group 4, G. clarum: distinguished by a three-layered spore wall made of a semipermanent outer wall and two superposed laminated walls. Group 5, G. mosseae: yellow spores often embedded in a peridium, characterized by septate and funnel-shaped subtending hyphae. See also Fig. S3 in the supplemental material.
Morphological limits and molecular boundaries.
A morphological threshold is defined as the point at which it is impossible to distinguish a strain as a different species. Such species boundaries were characterized here using the partial sequences of an inorganic phosphate transporter gene. Morphologically distinct species share 97.0% or fewer of their nucleotides and amino acids, giving a clear delimitation between species. Amino acid sequence identities within species range between 99% and 100%, which is a comfortable and acceptable level of intraspecific variation. The PT gene sequences of the Glomus phylogroup Aa (29), represented by G. mosseae and G. coronatum, and Glomus group Ab, represented by G. irregulare and G. intraradices, show only 73% similarity (Table 2). Below 80% nucleotide and amino acid similarity, subgenera or new genera should be defined, as has been recently done for the polyphyletic Scutellospora species (21).
Since the phosphate transporter gene sequences allow clear differentiation between the morphologically defined Glomus species, we propose that they be used as a routine tool for species identification and particularly to discriminate between closely related species.
Nucleotide sequence accession numbers.
Sequences were deposited in GenBank; accession numbers are given in Table 1.
ACKNOWLEDGMENTS
This work was supported by the FQRNT (Fonds québécois de la recherche sur la nature et les technologies).
We thank Sylvie Séguin for her technical help in strain maintenance and DNA extraction and J. Blaszkowski for providing G. irregulare type species material. We are also grateful to F. Stéfani for his help in the MODELTEST analysis and to Andrew P. Coughlan for reviewing the English.
FOOTNOTES
- Received 15 April 2010.
- Accepted 20 December 2010.
- Accepted manuscript posted online 30 December 2010.
§ Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.00919-10.
- Copyright © 2011, American Society for Microbiology
