DNase treatment improves viral enrichment in agricultural soil viromes

Abstract

The small genomes of most viruses make it difficult to fully capture viral diversity in metagenomes dominated by DNA from cellular organisms. Viral size-fraction metagenomics (viromics) protocols facilitate enrichment of viral DNA from environmental samples, and these protocols typically include a DNase treatment of the post-0.2 μm viromic fraction to remove contaminating free DNA prior to virion lysis. However, DNase may also remove desirable viral genomic DNA (e.g., contained in virions compromised due to frozen storage or laboratory processing), suggesting that DNase-untreated viromes might be useful in some cases. In order to understand how virome preparation with and without DNase treatment influences the resultant data, here we compared 15 soil viromes (7 DNase-treated, 8 untreated) from 8 samples collected from agricultural fields prior to tomato planting. DNase-treated viromes yielded significantly more assembled viral contigs, contained significantly less non-viral microbial DNA, and recovered more viral populations (vOTUs) through read mapping. However, DNase-treated and untreated viromes were statistically indistinguishable, in terms of ecological patterns across viral communities. Although results suggest that DNase treatment is preferable where possible, in comparison to previously reported total metagenomes from the same samples, both DNase-treated and untreated viromes were significantly enriched in viral signatures by all metrics compared, including a ~225 times greater proportion of viral reads in untreated viromes compared to total metagenomes. Thus, even without DNase treatment, viromics was preferable to total metagenomics for capturing viral diversity in these soils, suggesting that preparation of DNase-untreated viromes can be worthwhile when DNase treatment is not possible

Substantial differences in soil viral community composition within and among four Northern California habitats

“…1D). The percentage of 16S rRNA gene sequences in each virome ranged from 0.01-0.044% (consistent with prior reports of 0.028% bacterial 16S rRNA gene content in similarly prepared viromes from agricultural soils [18]) and did not differ significantly by habitat (ANOVA, P = 0.595). Thus, viral diversity estimates did not seem to be disproportionately skewed by sampling effort or the presence of non-viral sequences in viromes.…”

Spatial turnover of soil viral populations and genotypes overlain by cohesive responses to moisture in grasslands

“…Supernatants were removed, and pellets were resuspended in 100 µl of ultrapure water. As previously shown 46,52 , the DNase treatment step that serves to remove free DNA at this stage is not compatible with samples stored frozen (we suspect that this is because freezing compromises virions), so we were unable to perform a DNase treatment. We have previously shown that non-DNase-treated soil viromes still successfully enrich the viral signal relative to total metagenomes and capture the same ecological trends as DNasetreated viromes from the same samples 46 .…”

“…To consider the relic DNA pool in our samples more directly, we recovered reads classified as 16S rRNA gene fragments from virome profiles. Given that viral enrichment in viromes was achieved via 0.22-µm filtration prior to DNA extraction, any bacterial and archaeal sequences present in these libraries likely originated from relic DNA or small (< 0.22 µm) microbial cells 46, 47 . Interestingly, the relative abundance of Actinobacteria 16S rRNA gene reads recovered from T2-50 viromes was significantly higher than in any other group of samples ( Supplementary Figure 7c-d ).…”

Diversity in the soil virosphere: to infinity and beyond?

“…An important distinction can be made between the study of virus sequences identified in: (i) 'total' metagenomes (i.e., shotgun sequencing of all of the DNA in a sample), which tend to be dominated by sequences from cellular organisms, and (ii) 'virus-targeted' metagenomes (viromes), from which cells are depleted and viral particles enriched, typically by passing the sample through a 0.1-0.45 μm filter prior to shotgun sequencing of the viral fraction. Direct comparisons between soil viromes and total metagenomes are relatively rare, but so far, they suggest a much higher recovery of viral genomes from viromes (about one to two orders of magnitude) and no obvious biases in the types of viruses recovered through viromics, provided that multiple-displacement amplification (MDA) or similar whole-genome amplification methods were not used [30,80]. Based purely on the number of virus sequences recovered, a cost-benefit analysis should thus usually swing in favor of viromics for soil viral community ecology.…”

“…For example, stable isotope probing (SIP) through 13 C or H 2 18 O incorporation identified abundant active viral populations in grassland rhizospheres, agricultural soils, and peatlands [56,64,79], and 58% of viral populations identified in thawing permafrost peat metagenomes were classified as active via detection in metatranscriptomes [57]. Additionally, high viral diversity often remains in soil viromes treated with DNase prior to virion lysis, suggesting that a large portion of the soil viral community can be contained in intact and potentially still infective virions [30,80]. We expect that widespread viral activity will be revealed across many soil types and habitats, such that high soil viral diversity is not simply an accumulation of relic, inert virions, but rather represents dynamic viruses with substantial impacts on microbial processes and biogeochemical cycling.…”