Diversity and phylogenetic affinities of foliar fungal endophytes in loblolly pine inferred by culturing and environmental PCR

We examined endophytic fungi in asymptomatic foliage of loblolly pine (Pinus taeda) in North Carolina, U.S.A., with four goals: (i) to evaluate morphotaxa, BLAST matches and groups based on sequence similarity as functional taxonomic units; (ii) to explore methods to maximize phylogenetic signal for environmental datasets, which typically contain many taxa but few characters; (iii) to compare culturing vs. culture-free methods (environmental PCR of surface sterilized foliage) for estimating endophyte diversity and species composition; and (iv) to investigate the relationships between traditional ecological indices (e.g. Shannon index) and phylogenetic diversity (PD) in estimating endophyte diversity and spatial heterogeneity. Endophytes were recovered in culture from 87 of 90 P. taeda leaves sampled, yielding 439 isolates that represented 24 morphotaxa. Sequence data from the nuclear ribosomal internal transcribed spacer (ITS) for 150 isolates revealed 59 distinct ITS genotypes that represented 24 and 37 unique groups based on 90% and 95% sequence similarity, respectively. By recoding ambiguously aligned regions to extract phylogenetic signal and implementing a conservative phylogenetic backbone constraint, we recovered well supported phylogenies based on ca. 600 bp of the nuclear ribosomal large subunit (LSUrDNA) for 72 Ascomycota and Basidiomycota, 145 cultured endophytes and 33 environmental PCR samples. Comparisons with LSUrDNA-delimited species showed that morphotaxa adequately estimated total species richness but rarely corresponded to biologically meaningful groups. ITS BLAST results were variable in their utility, but ITS genotype groups based on 90% sequence similarity were concordant with LSUrDNA-delimited species. Environmental PCR yielded more genotypes per sampling effort and recovered several distinct clades relative to culturing, but some commonly cultured clades were never found (Sordariomycetes) or were rare relative to their high frequency among cultures (Leotiomycetes). In contrast to traditional indices, PD demonstrated spatial heterogeneity in endophyte assemblages among P. taeda trees and study plots. Our results highlight the need for caution in designating taxonomic units based on gross cultural morphology or ITS BLAST matches, the utility of phylogenetic tools for extracting robust phylogenies from environmental samples, the complementarity of culturing and environmental PCR, the utility of PD relative to traditional ecological indices, and the remarkably high diversity of foliar fungal endophytes in this simplified temperate ecosystem.