Comparison of epifluorescence and transmission electron microscopy for counting viruses in natural marine waters

Transmission electron microscopy (TEM) and epifluorescence microscopy of DAPI and Yo-Pro-1 stained samples were used to estimate viral abundance in natural communities along a transect from the oligotrophic central Gulf of Mexico to the productive near-shore waters at Post Aransas, Texas (USA). Estimates of viral abundance based on TEM averaged only 66% (range 26 to 108%) of those made using epifluorescence microscopy and the cyanine-based dye, Yo-Pro-1. DAPI staining provided estimates that were much closer and averaged 86% (range 72 to 109%) of those made using Yo-Pro. However, all 3 methods provided similar estimates at viral abundances <10(6) ml(-1). The precision of the Yo-Pro and DAPI methods (coefficient of variation 8 and 11%, respectively) was much greater than for the TEM method (25%). Experiments with cultures indicated that grazing by flagellates was unlikely to be a significant source of viral-size particles that could interfere with the DAPI or Yo-Pro method. Estimates of viral abundance made using the Yo-Pro method ranged from 0.3 x 10(6) to 39 x 10(6) ml(-1) in surface water along the transect. Across the investigated environments viral and bacterial abundances were well correlated (r = 0.929), although the slope of the relationship was significantly greater than 1, indicating that viral abundance increased more rapidly than that of bacteria. These results extend previous observations by showing that epifluorescence microscopy is suitable for counting viruses in very oligotrophic waters, that DAPI and Yo-Pro stained samples provide similar estimates of viral abundance and that grazing by flagellates is not a significant source of particles that could interfere with the epifluorescence method. The study supports the use of epifluorescence microscopy over TEM for obtaining accurate estimates of viral abundances in natural waters.