Technical Comments

Comment on “Global diversity and geography of soil fungi”

See allHide authors and affiliations

Science  26 Jun 2015:
Vol. 348, Issue 6242, pp. 1438
DOI: 10.1126/science.aaa4269


Tedersoo et al. (Research Article, 28 November 2014, p. 1078) present a compelling study regarding patterns of biodiversity of fungi, carried out at a scale unprecedented to date for fungal biogeographical studies. The study demonstrates strong global biogeographic patterns in richness and community composition of soil fungi. What concerns us with the study is what we do not see. Unfortunately, this study underestimates the fungal diversity of one key group of soil fungi due to reliance on a single primer with known flaws.

The internal transcribed spacer (ITS) region of the ribosomal RNA (rRNA) operon has been a de facto barcode for fungal taxonomy and ecology for 25 years since seminal publications by White et al. (1) and Gardes et al. (2). Although these primers have been valuable for describing fungal diversity both taxonomically and ecologically, unfortunately like almost any single primer set, they cannot capture all diversity in the fungal kingdom in an unbiased fashion. Tedersoo et al. (3) correctly acknowledge in the main text that the Tulasnellaceae and Microsporidia will not be amplified with their chosen primer set. They have also taken considerable care to include variants of the forward ITS3 primer to increase the representativeness of amplicon pools and describe these variants in their supplementary materials. However, it was previously shown by Rosling et al. (4), in the paper describing this new class (Archaeorhizomycetes) and cited by Tedersoo et al., that the ITS4 two mismatches to all known species in the class. These mismatches have been shown to result in at least a 10-fold underrepresentation of Archeaorhizomycetes in artificial communities using known DNA template amounts from isolates (5), and as a consequence other taxa in the artificial community also appeared to be more abundant than they were.

The Archaeorhizomycetes are a widespread, highly diverse, and ancient class of the Ascomycota documented in more than 100 studies of soil fungal communities (6). These fungi were first identified from 28S rRNA gene sequences from tundra soils in 2003 (7) and subsequently shown to be globally distributed in soil rRNA gene studies by Porter et al. (8) and expanded upon by Rosling et al. (4) and Menkis et al. (6). Using perfectly matched large subunit primers, Archaeorhizomycetes are shown to be dominant in many soils, typically comprising >10% of rRNA gene sequences recovered (8) and in one case >90% (9). In another specific example, Taylor and co-workers (10) performed extensive studies of soil fungal communities in Alaska using primers unbiased against amplification of Archaeorhizomycetes. In Alaska, at least 53 putative species-level operational taxonomic units (OTUs) of Archaeorhizomycetes could be identified, with diversity ranging from 2 to 25 OTUs per sample. Several of these putative species-level groups were among the most abundant OTUs across Alaska (11) and comprised at least 5% of the species richness of Alaskan soil fungal communities. In their current study, Tedersoo and colleagues identify only two OTUs belonging to the Archaeorhizomycetes in one out of two Alaskan sites studied, and thus Archaeorhizomycetes are reported in their paper to represent only ~0.3% of the fungal diversity in arctic tundra samples. In contrast, previous reports of alpine and arctic tundra samples revealed >10% overall relative abundance of the Archaeorhizomycetes class when suitable primers were used (7, 11). It should also be noted that as all studies of rRNA gene diversity are inherently semiquantitative (based on relative abundances), a likely 10-fold underestimation of diversity and abundance of the class Archaeorhizomycetes would affect the ratios of other functional and phylogenetic classes. Tedersoo et al. specifically discuss the patterns of Archaeorhizomycete diversity and abundance at the global scale in several instances within the paper but do not acknowledge the likely bias created by mismatched polymerase chain reaction (PCR) primer. Thus, Archaeorhizomycetes are not low-diversity, low-abundance soil fungi as described in (3), but in fact quite the opposite.

Although it is apparent that the authors used a single ITS4 primer variant to facilitate inclusion of the barcode tag sequences for multiplexing samples, this is no longer necessary because methods now exist for adding barcode tags after amplification (12). With the rapid advancement in parallelization and throughput of samples enabled by current sequencing technologies, the use of multiple primer sets or primer variants is achievable and advisable. The biases of PCR primers for studies of “unknown” organisms in bacterial and fungal ecology have been widely reported, and the continued reliance on single primer sets and variants for such studies compromises our understanding of the diversity and biogeography of microorganisms. It is also incumbent upon microbial ecologists to interpret patterns conservatively, especially given known biases.


View Abstract

Navigate This Article