by Daniel P. Haeusser
A few years ago I attended an ASM Branch meeting where an investigator gave a talk about a metagenomic survey of oceanic bacteriophages. In typical fashion for this type of study, one slide listed dozens genes of note identified as being encoded in phage genomes. With surprise I noticed that one of these was ftsZ, which encodes the prokaryotic tubulin homolog responsible for cell division in most bacteria and several phyla of archaea. I wondered: Why on Earth would a phage contain a cytoskeletal protein? The ftsZ gene is even found in chloroplasts and the mitochondria of certain protists, but at least these organelles have the evolutionary history of having been independent membrane-bound cells. But with a virus there is no ‘cyto’ in which to place any ‘skeleton’. An obvious hypothesis is that the phage might make use of a tubulin-like protein within the host for it nefarious phage reproduction cycle – but in what way?
Enter two recent studies (Kraemer, et al. and Oliva, et al.) reporting phage-encoded tubulin homologs and the start of their characterization. A general model emerges for these proteins in directing phage DNA replication within host cells and maximizing phage proliferation. Intriguingly, the phages encoding these tubulin homologs have particularly large genomes, suggesting a possible correlation between giant genome size and the need for a phage-encoded cytoskeleton.
The newly published phage tubulin homologs fall into two phylogenetic clusters (Fig. 2), each unique in the tubulin super-family. One cluster contains representatives in different chromosome, plasmid, or phage of the genus Clostridium. The tubulin homolog in the study by Oliva et al. comes from C. botulinum phage c-st, a phage already known for its production of the botulism neurotoxin. The crystal structure of this tubulin homolog (Fig. 3) reveals strong similarity to the TubZ family, which is involved in low-copy number plasmid segregation in several Bacillus species. In this system, the C-terminal tail of TubZ binds to TubR, an adaptor protein that also binds tubS centromeric sites on plasmid DNA.
Like canonical TubZ, c-st phage TubZ assembles into two-stranded helical filaments in the presence of GTP. In addition to encoding TubZ and a TubR homolog, C. botulinum c-st phage also contains a novel factor called TubY that Oliva et al. characterize as a robust in vitro modulator of TubZ assembly, capable of disassembling and reshaping TubZ polymers. Because the c-st phage replicates as a plasmid it is not surprising that a TubZRS system could be involved. However, it still remains to be determined whether c-st phage TubZ can form filaments in vivo or whether the system has significant benefits for phage reproduction.
The second phylogenetic cluster of phage-encoded tubulin homologs has been named PhuZ for “Phage tubulin/FtsZ”. Representatives of this family hail from phages of the genus Pseudomonas, the PhuZ of P. chlororaphis phage 201ɸ2-1 being the subject of the study by Kraemer et al. Like TubZ, PhuZ assembles in the presence of GTP into two-stranded helical filaments that look more generally like f-actin than a tubulin. However, there are some notable differences between PhuZ and TubZ (Fig. 3). PhuZ lacks a conserved interdomain helix that is important for polymerization of other tubulin homologs. Instead, to achieve polymerization, PhuZ uses a unique acidic patch at the end of its extended C-terminus. In the model proposed by the authors, six acidic amino acids of the thirteen extreme C-terminal residues of one PhuZ monomer form a ‘knuckle’ that nestles into a basic patch formed by helices H3-5 on the adjacent monomer, creating an overlapping pattern of polymerization. A deletion of the knuckle and strategic point mutations abolished PhuZ assembly in vitro, supporting this structural model. Thus, instead of using its C-terminus to interact with other factors as TubZ and FtsZ do, or to enhance lateral interactions, PhuZ appears to employ its tail for the self-interactions of filament assembly.
Admirably, Kraemer et al. addressed the in vivo role of PhuZ. At high expression levels, GFP-PhuZ formed dynamic filaments across the length of P. chlororaphis cells (Fig. 4). The authors then lowered GFP-PhuZ expression below the threshold level where filaments were observed and infected cells with phage 201ɸ2-1. As a result, fluorescent PhuZ polymers would only form when additional PhuZ was synthesized by the native phage. Time-lapse microscopy revealed that PhuZ filaments formed and persisted until the host lyses. Staining the DNA revealed a high concentration of phage-encapsidated DNA that formed a single rosette structure at midcell with frequent contacts to the ends of the PhuZ filaments (Fig. 5). Expression of a PhuZ mutant that forms non-dynamic filaments displayed mislocalized phage DNA at the cell poles, frequently scattered the DNA into smaller nucleoids, and dramatically reduced phage burst size. Although it is not yet clear how this compares to burst size under normal PhuZ expression levels, it suggests that PhuZ filaments help to increase phage yield.
So, how exactly does PhuZ facilitate phage DNA replication and phage proliferation? One possibility is that it directly interacts with phage DNA, or indirectly through an adapter protein like TubZ does with TubR. Or the system may be even more complex. Studies on ɸ29 phage of Bacillus have identified a small coiled-coil protein, p1, which self-assembles into filaments and sheets that also appears to play a similar role in organizing phage DNA replication. Apart from the p1 protein, ɸ29 phage expresses additional proteins involved in this process that bind to phage DNA and which depend on the host cell’s MreB cytoskeleton for their localization and activity. Thus, PhuZ may be a single part of a larger apparatus that includes both other phage proteins and host cell factors.
The correlation between phage genome size and these phage-encoded cytoskeletal proteins is enticing, but the meaning is not yet clear. Is there some inherent need to better-organize a large genome within the host compared to a smaller one? If so, why carry your own cytoskeleton rather than simply make use of the one already provided by the host as many eukaryotic viruses do? Perhaps using the host’s cytoskeleton would be too prone to host-driven assembly regulation that could thwart the phage’s goal of lysis. Rather than fight to control assembly of the host’s natural cytoskeleton, using your own protein when and where you want could be easier, particularly when your genome affords you the space to include it. As these and other questions are addressed, the cytoskeletal family will surely only continue to grow, particularly with the amazing field of bacteriophage open for exploration.
NOTE: A separate and more detailed version of this story is published as a Dispatch with William Margolin.
Daniel is a postdoctoral fellow in the Margolin Lab in the Department of Microbiology & Molecular Genetics at the University of Texas, Houston Medical School. He also teaches as an adjunct professor at the University of Houston-Downtown in the Department of Natural Sciences.
Kraemer JA, Erb ML, Waddling CA, Montabana EA, Zehr EA, Wang H, Nguyen K, Pham DS, Agard DA, & Pogliano J (2012). A phage tubulin assembles dynamic filaments by an atypical mechanism to center viral DNA within the host cell. Cell, 149 (7), 1488-99 PMID: 22726436
Oliva MA, Martin-Galiano AJ, Sakaguchi Y, & Andreu JM (2012). Tubulin homolog TubZ in a phage-encoded partition system. Proceedings of the National Academy of Sciences of the United States of America, 109 (20), 7711-6 PMID: 22538818