Global diversity of putative functional viral auxiliary metabolic genes and their relationship with soil organic carbon stocks

Abstract

Soil viruses are increasingly recognized as components of microbial communities with the potential to alter ecosystem functioning, yet the frequency and functional distribution of virus-encoded metabolic genes in soils remain poorly understood. Here, the Global Soil Virus (GSV) Atlas gene catalog, comprising 1,432,147 viral genes from 1,223 samples across 12 ecosystem types, was searched against KEGG, Pfam, and CAZy for functions related to carbon cycling, nitrogen cycling, and antimicrobial resistance. Annotations associated with host entry or cell-wall lysis were separated from putative auxiliary metabolic functions. Of 5,760 initial gene-level functional annotations across the three categories, 3,902 (67.7%) were classified as peptidoglycanases. Their exclusion left 1,858 genes (0.13% of the catalog), and carbon-cycling functions represented 96.7% of the enzyme-resolved annotations. Chitinase-associated genes formed the largest carbon category, but 78.2% belonged to GH19, a family containing both chitinases and phage endolysins. Among GH19 genes on contigs encoding detectable holins or spanins, 86.8% occurred within three genes of these lysis components, a distribution comparable to that of canonical phage lysozymes. Across six matched metagenomic studies, estimated viral contributions to the targeted functions ranged from 0.10% to 9.86% (mean, 2.60%). Two of the three estimates exceeding 2% were GH19-associated chitinase estimates from rhizosphere soils and were therefore treated as upper estimates. Thus, the putative soil viral metabolic inventory was sparse and concentrated in carbon-degradation functions, whereas much of the initial carbohydrate-active enzyme signal reflected core viral lysis functions rather than auxiliary metabolism.