The structure of the membrane-containing bacteriophage PRD1 has been determined by X-ray crystallography at about 4 A resolution. Here we describe the structure and location of proteins P3, P16, P30 and P31. Different structural proteins seem to have specialist roles in controlling virus assembly. The linearly extended P30 appears to nucleate the formation of the icosahedral facets (composed of trimers of the major capsid protein, P3) and acts as a molecular tape-measure, defining the size of the virus and cementing the facets together. Pentamers of P31 form the vertex base, interlocking with subunits of P3 and interacting with the membrane protein P16. The architectural similarities with adenovirus and one of the largest known virus particles PBCV-1 support the notion that the mechanism of assembly of PRD1 is scaleable and applies across the major viral lineage formed by these viruses.

The unusual bacteriophage PRD1 features a membrane beneath its icosahedral protein coat. The crystal structure of the major coat protein, P3, at 1.85 A resolution reveals a molecule with three interlocking subunits, each with two eight-stranded viral jelly rolls normal to the viral capsid, and putative membrane-interacting regions. Surprisingly, the P3 molecule closely resembles hexon, the equivalent protein in human adenovirus. Both viruses also have similar overall architecture, with identical capsid lattices and attachment proteins at their vertices. Although these two dsDNA viruses infect hosts from very different kingdoms, their striking similarities, from major coat protein through capsid architecture, strongly suggest their evolutionary relationship.

Our discovery that the major coat protein of bacteriophage PRD1 resembles that of human adenovirus raised the unexpected possibility that viruses infecting bacteria could be related by evolution to those infecting animal hosts. We first review the development of this idea. We then describe how we have used structure-based modeling to show that several other viruses with no detectable sequence similarity are likely to have coats constructed from similar proteins-the "double-barrel trimer." There is evidence that the group includes a diversity of viruses infecting very different hosts in all three domains of life: Eukarya; Bacteria; and Archaea that diverged billions of years ago. The current classification of viruses obscures such similarities. We propose that the occurrence of a double-barrel trimer coat protein in an icosahedral dsDNA virus with large facets, irrespective of its host, is a very strong indicator of its membership in a lineage of viruses with a common ancestor.

Parasites provide a selective pressure during the evolution of their hosts, and mediate a range of effects on ecological communities. Due to their short generation time, host-parasite interactions may also drive the virulence of opportunistic bacteria. This is especially relevant in systems where high densities of hosts and parasites on different trophic levels (e.g. vertebrate hosts, their bacterial pathogens, and virus parasitizing bacteria) co-exist. In farmed salmonid fingerlings, Flavobacterium columnare is an emerging pathogen, and phage that infect F. columnare have been isolated. However, the impact of these phage on their host bacterium is not well understood. To study this, four strains of F. columnare were exposed to three isolates of lytic phage and the development of phage resistance and changes in colony morphology were monitored. Using zebrafish (Danio rerio) as a model system, the ancestral rhizoid morphotypes were associated with a 25–100% mortality rate, whereas phage-resistant rough morphotypes that lost their virulence and gliding motility (which are key characteristics of the ancestral types), did not affect zebrafish survival. Both morphotypes maintained their colony morphologies over ten serial passages in liquid culture, except for the low-virulence strain, Os06, which changed morphology with each passage. To our knowledge, this is the first report of the effects of phage-host interactions in a commercially important fish pathogen where phage resistance directly correlates with a decline in bacterial virulence. These results suggest that phage can cause phenotypic changes in F. columnare outside the fish host, and antagonistic interactions between bacterial pathogens and their parasitic phage can favor low bacterial virulence under natural conditions. Furthermore, these results suggest that phage-based therapies can provide a disease management strategy for columnaris disease in aquaculture.