A metagenomic analysis of urban river samples reveals high numbers of sequences related to mycoviruses

Mycoviruses have been classified into 35 virus families so far. In addition to numerous mycoviruses with proven fungal or oomycetes hosts, many newly discovered viruses from environmental water and soil samples and various animal or plant specimens exhibit significant similarity to classified mycoviruses, thereby expanding the known sequence space of fungal and related viruses. In this study, we searched for mycoviruses in two environmental water samples that had been collected from the Teltow Canal and the Havel River in Berlin, Germany. Viral sequences with similarity to members of 16 virus families were identified. The most common viruses in our samples were botourmia-like viruses with moderate similarity to members of the genus Ourmiavirus. Notably, 58 of the ourmia-like sequences from the Teltow Canal and Havel River and 10 ourmia-like sequences from other sources exhibited a dicistronic genome layout. The second open reading frame (ORF) of these viruses encoded a putative capsid protein with an S domain that showed little similarity to the structural proteins of the classified ourmiaviruses. The second-largest virus group (59 sequences) was assigned to the order Ghabrivirales, and 13 of these sequences exhibited similarity to members of the suborder Alphatotivirineae (families Orthototiviridae, Pseudototiviridae, Botybirnaviridae, and Chrysoviridae). Thirty-three sequences clustered with members of the suborder Betatotivirineae - three of them with members of the family Artiviridae. Fifteen highly divergent toti-like sequences suggest the need to establish a new higher-order taxon within the order Ghabrivirales. Other virus sequences were assigned to the families Mitoviridae (three unuamitoviruses and 10 highly divergent mitovirus-like sequences), Narnaviridae (five "alphanarnavirus"-like sequences), Amalgaviridae (two zybavirus-like sequences), Hypoviridae (one partial RdRP sequence), and Mymonaviridae (one partial RdRP sequence), and one was not classified (Sclerophthora macrospora B-like virus). Notable results include a clade of highly divergent mitovirus-like sequences with a standard translation code, three narnavirus-like sequences with a reverse-frame ORF, and a clade of four Ghabrivirales-like replicase sequences that were found to have numerous termination codons regardless of which translation table was used.