A hidden gene in astroviruses encodes a cell-permeabilizing protein involved in virus release

Valeria Lulla; Andrew E. Firth

We grouped astrovirus sequences into clades (Fig. 1B) , generated codon-based alignments of concatenated ORF1a-ORF1b-ORF2 coding sequences within each clade, and analyzed the alignments with synplot2 which tests for regions where synonymous substitutions occur less often than average for the sequence alignment 9 . Regions with significantly enhanced synonymous site conservation typically harbour overlapping functional elements which constraint sequence evolution 9 . The analysis of subgroup "Ia" astroviruses ( Fig. 1B ; classical HAstVs besides some feline and sea lion astroviruses) revealed conserved elements at the junction of ORF1a and ORF1b (representing the ribosomal frameshifting signal and ORF1a/ORF1b overlap), upstream of ORF2 (presumed to represent elements involved in sgRNA synthesis), towards the 3′ end of ORF2 (possible replication elements), and overlapping the 5′ end of ORF2 (the putative overlapping ORFX) (Fig. 1C) . We applied the same analysis to 13 other astrovirus genogroup or sub-genogroup clades, revealing conserved regions overlapping the 5′ end of ORF2 in genogroup I, III and IV astroviruses but, generally (see below for exceptions), not in genogroup II or VI astroviruses (Fig. S3) .