Themes and Variations of viral small terminase proteins

Abstract

Viruses are the most populous ‘biological entities’ on Earth (Krupovic and Bamford, 2011). The most common of these are the tailed, double-stranded DNA (dsDNA) containing bacteriophages, which outnumber bacteria by 10 fold. Since they do number so many, bacteriphages greatly affect global processes (Suttle, 2007). Related to dsDNA phages are the eukaryotic dsDNA Herpesviruses, which cause a host of diseases including chicken pox, shingles, Karposi’s sarcoma and mononucleosis (Cardone et al., 2012). Thus, understanding the assembly of this class of viruses is important for both the global climate and human health.

The complex choreography of dsDNA virus assembly

The tailed dsDNA bacteriophages and Herpesviruses package their genomes into preassembled precursor capsids, called procapsids, which have icosahedral symmetry (Casjens and Molineux, 2012). In general, procapsids are the product of co-assembly of multiple copies of a major capsid protein with a scaffolding protein, which acts as an assembly chaperone. Scaffolding proteins serve to position the major capsid proteins such that the procapsids attain the correct geometry. Icosahedral viruses have their capsid proteins arranged in 11 pentons and variable numbers of hexons to form a closed sphere with 20 triangular faces (Prasad and Schmid, 2012). The number of hexons determines the size of the capsid. A portal protein complex, a dodecamer that functions as a channel during DNA packaging, is incorporated during assembly into the twelth penton position. The DNA is replicated as concatamers of several genomes (Casjens and Molineux, 2012). The small terminase protein complex (TerS), described further below and in this issue (Roy et al., 2012), recognizes the concatameric viral DNA and interacts with large terminase proteins (TerL) as well as the portal protein complex to actively package the DNA into the procapsids. The TerL proteins are ATPases that use one ATP for approximately every two base pairs of DNA packaged. TerL proteins also possess a nuclease that cleaves a single genome from the concatameric DNA (Feiss and Rao, 2012). The DNA is pumped into the procapsids at a maximum rate of about 2000 base pairs/sec. During DNA packaging, the procapsid undergoes a maturation event that includes removal of the scaffolding proteins and changes to the structure of the major capsid protein (Johnson, 2010; Prasad and Schmid, 2012). The packaged DNA is stabilized within the mature capsid by the addition of portal closure proteins and tail proteins (Casjens and Molineux, 2012).

Capsid assembly has evolved so each step involves conformational transitions critical for the next step, thereby ensuring fidelity during assembly. Procapsids contain only viral proteins and exclude host proteins, unless specifically required for the next round of infection. Portal complexes are incorporated into procapsids at only one penton position, and the terminase proteins recognize portals only when incorporated into procapsids. The small terminase proteins interact with only viral DNA. An exception is in the transducing viruses, which infrequently package host DNA. The portal closure proteins recognize the portal complex only after the DNA is packaged. Failure of specificity at any step would result in a non-infectious particle.

Structural homology—or not.

As increasing numbers of capsid protein structures have been solved, there is evident structural homology even in the complete absence of sequence homology. The major capsid proteins of dsDNA viruses seem to fall into one of four distinct classes, irrespective of the origin of the virus such that each class includes phages and viruses (Krupovic and Bamford, 2011). Portal proteins of dsDNA viruses are also structurally related, as are the large terminase proteins, which have an RNAse H1 fold in one domain and a Walker box ATPase motif in the second motor domain (Feiss and Rao, 2012; Roy and Cingolani, 2012). It is intriguing that all of these proteins are structurally homologous—but the small terminase proteins are distinctly less so (Roy et al., 2012).

TerS proteins are surprisingly non-homologous in structure; however, all have three domains. There is a central oligomerization domain, although the oligomeric state varies from dimeric for phage λ (in a complex with a single TerL that assembled into a tetramer in vivo) to nonameric for P22 and Sf6 (note that Sf6 and SF6 are different phages) to 11- or 12-mers for phage 44RR (Casjens and Molineux, 2012). The diameter of the central channel formed from the different oligomers of TerS proteins varies from 11Å at the narrowest for SF6 phage to 37Å at the widest for 44RR phage. There are also DNA and TerL recognition domains. However, of the five partial or complete TerS structures currently determined, only the subunits of Sf6 and P22 TerS proteins are easily superimposable (Roy et al., 2012).

The TerS proteins have two important roles in DNA packaging (Casjens and Molineux, 2012). First, they distinguish viral DNA from host DNA. Second, TerS proteins interact with the TerL protein and direct it to the correct position on the DNA, where the TerL nuclease activity cuts the DNA at the appropriate site. Some experiments, including those presented in this issue, have suggested that TerS proteins also regulate the ATPase activity of TerL proteins (Roy et al., 2012).

Two models have been proposed to describe the interaction of TerS proteins with DNA, with its 23Å diameter. There are data from different systems to support each model. First, a ‘wrap around’ model suggests that the DNA wraps around the outside of the TerS protein and is directed to the TerL and portal protein complex. Second, the DNA is proposed to thread through the central channel. Support for the ‘wrap around’ model comes from observation that the DNA binding domain of Sf6 TerS is on the outside of the octamer (Zhao et al., 2010). Also, the SF6 TerS oligomers have a central channel that is too small to accommodate the DNA and, here too, the DNA-binding domains are found on the surface of the complex (Buttner et al., 2012). In this model, how the DNA is threaded to the TerL and into the central channel of portal complex has not been well described.

Conversely, the phage P22 TerS complex is suggested to interact with DNA through its channel (Roy et al., 2012). In this complex, the channel is large enough to accommodate the DNA and unlike the other phages’ TerS oligomers, the DNA binding regions, which are at the C-termini, are modeled to make an ‘extended barrel’ lined with basic amino acids. These structural data indicate that DNA movement through the channel may be feasible and provides compelling evidence to support this model. On the other hand, the P22 TerS protein complex was shown to be quite stable, so how the concatemeric DNA could be threaded though the protein is unclear since it is unlikely to disassemble and reassemble around the DNA.

Regardless of which model—or both—is correct, the observation that the TerS proteins are functionally, but not structurally, homologous seems to be unique among these viruses and raises questions about the evolution of the TerS proteins. Resolution of the conundrums presented by the data and models will need further structural and biochemical analysis of the DNA/TerS/TerL/portal protein complex.