Phage spanins: diversity, topological dynamics and gene convergence

Abstract

BackgroundSpanins are phage lysis proteins required to disrupt the outer membrane. Phages employ either two-component spanins or unimolecular spanins in this final step of Gram-negative host lysis. Two-component spanins like Rz-Rz1 from phage lambda consist of an integral inner membrane protein: i-spanin, and an outer membrane lipoprotein: o-spanin, that form a complex spanning the periplasm. Two-component spanins exist in three different genetic architectures; embedded, overlapped and separated. In contrast, the unimolecular spanins, like gp11 from phage T1, have an N-terminal lipoylation signal sequence and a C-terminal transmembrane domain to account for the topology requirements. Our proposed model for spanin function, for both spanin types, follows a common theme of the outer membrane getting fused with the inner membrane, effecting the release of progeny virions.ResultsHere we present a SpaninDataBase which consists of 528 two-component spanins and 58 unimolecular spanins identified in this analysis. Primary analysis revealed significant differences in the secondary structure predictions for the periplasmic domains of the two-component and unimolecular spanin types, as well as within the three different genetic architectures of the two-component spanins. Using a threshold of 40% sequence identity over 40% sequence length, we were able to group the spanins into 143 i-spanin, 125 o-spanin and 13 u-spanin families. More than 40% of these families from each type were singletons, underlining the extreme diversity of this class of lysis proteins. Multiple sequence alignments of periplasmic domains demonstrated conserved secondary structure patterns and domain organization within family members. Furthermore, analysis of families with members from different architecture allowed us to interpret the evolutionary dynamics of spanin gene arrangement. Also, the potential universal role of intermolecular disulfide bonds in two-component spanin function was substantiated through bioinformatic and genetic approaches. Additionally, a novel lipobox motif, AWAC, was identified and experimentally verified.ConclusionsThe findings from this bioinformatic approach gave us instructive insights into spanin function, evolution, domain organization and provide a platform for future spanin annotation, as well as biochemical and genetic experiments. They also establish that spanins, like viral membrane fusion proteins, adopt different strategies to achieve fusion of the inner and outer membranes.Electronic supplementary materialThe online version of this article (10.1186/s12859-018-2342-8) contains supplementary material, which is available to authorized users.

Phage S144, a New Polyvalent Phage Infecting Salmonella spp. and Cronobacter sakazakii

“…As in other Type II holins, ORF34 has two transmembrane domains, while ORF35 is recognizable as an endolysin by an O-glycoside hydrolase domain, showing protein to lysins of other phages infecting Enterobacteriaceae strains ( Table 2 and Supplementary material). Since ORF36 has a transmembrane domain at the N-terminal, we classified it as a putative integral inner membrane protein (i-spanin) and we found the outer membrane lipoprotein (o-spanin) at the +1 reading frame ( Table 2 and Supplementary Material), as previously suggested [39].…”

Phage-encoded cationic antimicrobial peptide used for outer membrane disruption in lysis

“…Based on its proximity to the endolysin gene and its multiple predicted transmembrane domain architecture, gene 29 was annotated as the holin: 127 aa, 13.2 kDa, with 4 predicted transmembrane domains. A rigorous analysis showed that there were no possible lipobox sequences in any of the remaining open reading frames (10). This rules out the presence of a putative spanin of either the unimolecular or two-component type, since both involve a lipoprotein.…”

“…Bioinformatic analysis of 677 genomes of phages that infect Gram-negative hosts found that 586 encode identifiable spanins: 528 have two-component spanins systems encoding separate i-spanin and o-spanin subunits like λ; 58 are of a different type, the unimolecular spanins, in which a single polypeptide has both OM lipoprotein and IM transmembrane domain determinants (7,10). The u-spanins are proposed as analogs of Class II viral fusion proteins ( Fig.…”

Galaxy and Apollo as a biologist-friendly interface for high-quality cooperative phage genome annotation

“…From there, the naïve gene sequences are translated and analyzed in three ways. 1) Protein similarity to spanins in a curated database [23] is determined by BLASTp with a 0.001 expectation value cutoff. 2) Potential lipobox motifs expected for outer membrane spanins are identified using the tools LipoP [27] and Identify Lipoboxes, a less stringent regex amino acid motif search for the four-residue motif 3) Inner-membrane spanin candidates are identified through TMHMM, which predicts transmembrane regions [28].…”