Where are the viruses? (Flaws of virus count methods)

Marine viruses represent a major driving force in plankton ecology and, due to perhaps false assumptions that the majority of marine bacteriophages were DNA-containing viruses, most study in the field over the last 20 years has been on such viruses. Steward et al set out to quantify eukaryote infecting RNA viruses by comparing the total mass of viral RNA and DNA from seawater samples and estimating the abundances of each virus type. They found that RNA virus abundance was comparable to, and possibly exceeding, that of DNA viruses. Picorna-like viruses were found to dominate the RNA viral fraction of their samples, which was in agreement with previous metagenomic surveys of the marine RNA viral fraction that suggested that positive-sense, single-strand picornavirads were dominant. However, as these picornavirads posses small genomes that are often below the detection threshold of most fluorescence-based counting methods, the authors concluded that virus counts may be vastly underestimating true numbers.

Direct quantification of RNA virus abundance is difficult and unreliable due to a number of limiting factors such as stain sensitivity when using stains to differentiate DNA and RNA viruses for counts and small genome size preventing detection by such stains. The researchers took a novel approach and measured the relative masses of RNA and DNA viruses from tropical seawater samples collected from Kane’ohe Bay, O’ahu, Hawaii, and paired their data with estimates of the mass of nucleic acid per RNA and DNA virion to estimate marine RNA virus abundance.

After the seawater samples were collected, they were filtered and processed into subsamples, from which the viruses were concentrated and purified into fractions of 0.5ml each. The RNA and DNA content of these fractions were measured via fluorometric assays in parallel tests and nucleic acid masses were predicted from the assay results to a ±95% confidence interval. The RNA concentration for each sample was plotted as a function of the given RNA and DNA concentrations and statistics including regression and a two-tailed t-test performed to test for differences.

The researchers created and analysed metagenomes of one subsample of each fraction. RNA treated with DNase from these subsamples was pooled and amplified. They used pyrosequencing to generate sequence libraries from the amplified samples (now DNA). The community composition of each sample was analysed and taxonomic assignments applied. Sequences were compared with the SILVA database to rule out contamination from cellular RNA and the initial RNA mass estimates were adjusted accordingly. Abundances of RNA viruses were estimated by calculating the number of genomes from nucleic acid masses, done by using the relative representation of different taxa in the metagenomic library to calculate an average RNA mass per virion. DNA virus abundances were calculated from the DNA content of the fractions. It should be noted that they based their assumption of DNA content per virion on an average of data from a wide range of environments in their calculations and admitted overestimation may have occurred.

Their results showed that 50-57% of the RNA viral genomes likely originated from eukaryote-infecting RNA viruses, mostly positive-sense ssRNA viruses belonging to Picornavirales. An implication of this is that protists, while of low abundance, possibly contribute more to marine viral dynamics than expected via RNA virus infection. Around 1% of the sequences matched known double-strand RNA viruses when compared with the metagenomic library. Around 42% of the sequences had no match in the library and remain unidentified, although they exhibited virus-like qualities. They estimated the contribution of RNA viruses to the total number of viruses at 38-63%, taking cellular RNA into consideration. These numbers were minimum estimates to account for processing losses. DNA virus abundances varied throughout the year, from 17-33% in June to 5-9% in August. This variability suggests uncertainty which may relate to laboratory practices and to unreliability of the counting methods. For example, some physically larger viruses (above 0.2mm) may have been lost during filtration when preparing the samples for analysis. Since all larger viruses are double stranded DNA types, the results are biased against DNA viruses. However the magnitude of this bias is unknown and may only be slight if the contribution of larger viruses to the fraction mirrors electron microscopy data suggesting that these virus are low in number.

In their closing paragraph, they suggest a review of RNA virus and single-strand DNA virus count methods to improve reliability and account for smaller and undetectable genomes that may skew data, and push for the development of new infallible methods.

Reference

Steward, G.F., Culley, A.I., Mueller, J.A., Wood-Charlson, E.M., Belcaid, M. and Poisson, G. (2013) Are we missing half the viruses in the ocean? J ISME 7: 672-679.