Diversity of conidia of aquatic hyphomycetes assessed by microscopy and by DGGE

Water samples from a Canadian stream were passed through membrane filters between 22 July 2002 and 19 May 2003. Filters with trapped conidia of aquatic hyphomycetes were cut in half. One half was examined under a light microscope, and conidia were counted and identified. From the second half, DNA was extracted and amplified with fungal primers. The number of different ITS sequences (phylotypes) in the amplified DNA was assessed with DGGE. On average, the number of visually identified species per sample (12.4) was higher than the number of phylotypes (11.7), but the difference was nonsignificant (P = 0.36). However, the difference between species and phylotype numbers increased significantly with the number of species or conidia present on the filter, indicating that the sensitivity of DGGE decreases with sample size. When few conidia were present, phylotype numbers often exceeded species numbers, suggesting insufficient resolution of visual identification or the presence of DNA from nonconidial sources. A modification of the described method may be useful to check the accuracy of taxonomy and identification based on conidial morphology.