STIV (Sulfolobus turreted icosahedral virus) has been the subject of detailed structural, genetic, transcriptomic, proteomic and biochemical studies. STIV arguably has been investigated in more detail than any other archaeal virus. As a result, we know more about STIV than other viruses infecting members of the Archaea domain. Like most viruses isolated from crenarchaeal hosts, STIV has little in common with viruses that infect eukaryotic and bacterial hosts and should be considered the founding member of a new virus family. However, despite this lack of obvious homology with other viruses, STIV has components of gene content, replication strategy and particle structure reminiscent of viruses of the Eukarya and Bacteria domains, suggesting an evolutionary relationship between viruses from all domains of life. The present mini-review describes the current knowledge of this virus and insights it has given us into viral and cellular evolution, as well as newly developed tools for the further study of STIV–host interactions.
STIV (Sulfolobus turreted icosahedral virus) was originally isolated from an acidic (pH 2.2) high-temperature hot spring (82°C) in Yellowstone National Park . The virus was isolated from an enrichment culture of an unsequenced Sulfolobus isolate, YNPRC179. This enrichment culture was screened by TEM (transmission electron microscopy) for the presence of virus-like particles which were subsequently purified and characterized .
STIV has a limited host range. The virus has been shown to infect the Italian isolates Sulfolobus solfataricus strains P1, P2 and a highly susceptible strain derived from S. solfataricus P2, termed strain P2-2-2-12. Infection of other species, including Sulfolobus islandicus, Sulfolobus acidocaldarius and Sulfolobus tokodaii, as well as other uncharacterized Sulfolobus isolates from Yellowstone National Park has been unproductive [1,2]. The full host range of the virus is not known.
STIV is the first lytic virus known to infect a member of the genus Sulfolobus [3,4]. Infection of S. solfataricus strain P2-2-2-12 produces plaques and high virus titres in liquid cultures 36 hpi (hours post-infection). More than 90% of cells viewed by electron microscopy show evidence of virus infection (Figure 1). TEM analysis of a time course of the infection demonstrates lysis of infected cells. Lysis appears to be an active process directed by virus gene expression, with alteration of the host S-layer preceding cell membrane disruption. TEM and SEM (scanning electron microscopy) analyses of these cells show pyramid-like structures on the cell surface just before lysis. The mechanism by which these structures form, the host and virus genes that are involved in the formation of these structures, and the biochemical details of the lysis process have yet to be elucidated.
Transmission electron micrographs of S. solfataricus cells infected with STIV
The virus encapsidates a circular 17663 bp dsDNA (double-stranded DNA) genome which replicates episomally with no evidence of integration into the host genome  (Figure 2). The current annotation of the genome includes 37 ORFs (open reading frames) [1,5]. Overall, there is little similarity to ORFs encoded in the STIV genome with other genes in the databases. However, bioinformatics approaches have found similarity for three STIV ORFs to proteins in the public databases (C557, B116 and C92), but each of these predicted viral proteins is of unknown function . Another ORF, B164, was predicted by homology modelling to be an ATPase . The other 33 ORFs have no homology with proteins in the public databases. The lack of homology at the genetic and amino acid level to other known proteins suggests that function must be determined through further genetic, biochemical and/or structural analysis.
The virion is ∼74 nm in diameter with an outer protein shell built on a pseudo T=31 lattice . This is the first virus with this symmetry to be described. At each of the 12 5-fold axes, there are turret-like appendages projecting from the virion surface from which the virus derives its name. These 5-fold turrets project 13 nm from the virus particle surface and are approx. 24 nm in diameter at their broadest point . The function of these turrets is unknown, but one can speculate that they may be involved in receptor-mediated cell-surface attachment, injection of DNA into the cell and/or packaging of DNA into the capsid. MS has identified 11 proteins in the virus particle, nine of which are viral and two that are host-derived . ORF B345 was identified by MALDI (matrix-assisted laser-desorption ionization) MS as the MCP (major capsid protein) of the virion. The MCP is glycosylated, but the nature of the carbohydrates and the chemical linkage to the coat protein subunit are unknown . Four ORFs are thought to code for proteins involved with the turret structure formation, including an ATPase (ORF B164). The functions of the remaining four virus-coded proteins are unknown. Two host-derived proteins are found within the virion. The first is a small DNA-binding protein (Sso7D), which is likely to be involved with packaging the viral genome. The second host-derived protein (SSo0881) is a hypothetical protein with a potential VPS (vacuolar protein sorting) 24 domain that is found in both Saccharomyces cerevisiae and mammals and is conserved across the Sulfolobus family . The virion also contains a host-derived internal lipid layer surrounding the nucleocapsid genomic core. This internal membrane comprises a subset of host lipids that are highly enriched for acidic lipids . The MCP forming the exterior capsid shell is thought to anchor to the internal lipid through insertion of its C-terminus into the lipid layer. The mechanism by which this membrane is acquired and its function are both unknown, but it is tempting to speculate that the function of this lipid envelope is to help protect the viral genome from acid hydrolysis and/or to help stabilize the virion in its high-temperature environment.
Comparative structural analysis has revealed interesting insights into virus evolution . The cryo-EM reconstruction of the virion shows an overall architecture that is similar to viruses of the other two domains of life (Figure 3). The crystal structure of the MCP also reflects a common overall structure that is shared with viruses that infect eukaryotic hosts (adenovirus) and bacterial hosts (PRD1). This was the first example of a virus morphology found in all three domains of life. Structure predictions for several other proteins from STIV support further this relationship among viruses from all domains of life. C381 and A223 both share predicted structural similarity, as determined by fold-recognition algorithms, to the P5 protein of PRD1. The virion architecture, along with the tertiary structure of the capsid protein itself, suggests a virus lineage that spans all domains of life, despite the inability to detect this homology at the sequence level .
Cryo-electron microscopy reconstruction of STIV
A number of additional crystal structures exist for non-structural STIV proteins, which have provided insights into their function. A197 is a putative glycosyltransferase , and indeed the virus capsid has been demonstrated to be glycosylated . B116 is a DNA-binding protein , and F93 is a winged-helix DNA-binding protein , suggesting that both may be involved in controlling viral and/or host gene expression.
Microarray experiments following an infected culture and comparing it with an uninfected culture have provided insights into the timing of events during viral infection as well as the host's response to the virus . In a near-synchronous single replication cycle, there is little evidence for distinct temporal regulation of viral gene expression. Transcription of viral genes was first detected 8 hpi, with a peak at 24 hpi. Lysis of cells was first detected at 32 hpi. The majority of the viral genes are co-transcribed between 16 and 24 hpi. This is in contrast with many DNA viruses of bacteria and eukarya, which demonstrate distinct temporal patterns of early and late viral gene expression. A number of putative operons could be identified based on levels of transcription and timing. One gene, C67, was transcribed in co-ordination with the viral structural genes, but has not been identified as part of the virion. During the infection, 177 host genes were determined to be differentially expressed, with 124 genes up-regulated and 53 genes down-regulated. The up-regulated genes were dominated by genes associated with DNA replication and repair and those of unknown function, whereas the down-regulated genes, mostly detected at 32 hpi, were associated with energy production and metabolism. The observed patterns of transcription suggest that up-regulated genes are likely to be used by the virus to reprogramme the cell for virus replication, whereas the down-regulated genes signal the imminent lysis of the cells. The majority (94%) of the host genes, however, showed no differential regulation between infected and uninfected cultures, suggesting that the virus life cycle is tailored to avoid a host stress response. This is corroborated further by the absence of typical heat-shock- or UV-induction stress-response-type genes. And, in fact, some genes that are up-regulated under other stresses are found to be down-regulated in response to STIV infection.
The S. solfataricus P2 genome has six combined CRISPR (cluster of regularly interspaced palindromic repeats)–Cas loci which are thought to a form a new viral immunity system . Several of these CRISPR loci have spacer sequences that are exact matches to STIV. Despite these matches, STIV infects S. solfataricus to high levels and no new STIV-specific spacers were detected post-infection. In fact, the CRISPR–Cas system was not differentially regulated in response to STIV infection compared with uninfected cells. The nature of this system in this host and its interaction with STIV appears to be modified in comparison with the systems described in many bacterial species. Unravelling the function of this system in S. solfataricus should prove to be an interesting area of research.
STIV genetic system
Further investigation of gene function of the virus depends on the ability to create site-specific mutations in the viral genome and analyse the resultant phenotypes. To this end, we have constructed an infectious clone of the STIV genome. This clone was constructed by cloning PCR-amplified segments of the genome into an Invitrogen TA vector. We have demonstrated that the wild-type cloned genome is infectious upon transformation into host cells. Particles purified from primary transformants can be used as an inoculum for infecting naïve cells, demonstrating infectivity of the cloned virus. Strains and species of Sulfolobus that are competent for transfection with the clone closely follow the host range of the native virus.
Despite the large insert size for this vector, the construct appears to be stable and replicates well in Escherichia coli. However, the full-length vector does not replicate or produce particles in S. solfataricus. Removal of the E. coli vector and re-ligation before electroporation is required for productive transformation. Additionally, linearized cloned DNA with or without the E. coli vector does not produce an infection. This is in contrast with the SSV1 (Sulfolobus spindle-shaped virus 1)-based shuttle vector that can be transformed as a linear piece of DNA and transfect the Sulfolobus culture, albeit at lower efficiency than circular DNA .
A number of gene disruption mutant genomes have been constructed. Initial targets for gene disruptions were guided by previous studies of the virus. The putative glycosyltransferase (A197), a DNA-binding protein (B116), for which structures exist, and C92, which was highly up-regulated in the transcriptomics experiments, as well as the MCP and the putative turret protein (C557), were targeted for gene disruption by either PCR-based mutagenesis or digestion and end repair to create frame-shifted mutants. Not surprisingly, none of the mutants tested to date has produced viable virus particles or can be detected by qPCR (quantitative PCR) following transformation into S. solfataricus. However, a microarray analysis of a time-course transfection with ΔA197 shows that disruption of the glycosyltransferase gene still results in viral transcription, indicating that the deficiency in this virus occurs at a downstream point in the viral replication cycle. We have also constructed a random library of gene disruptions utilizing a modified Tn5-based transposon. We are in the process of screening this library for altered viral phenotypes. Finally, the MCP and C557 are being targeted by PCR-based mutagenesis to look for regions that might tolerate short peptide insertions.
Recent advances in tool development for both the virus and host are allowing for a more in-depth understanding of STIV and its interactions with its host. In particular, the recent development of genetic tools for both STIV and Sulfolobus [11–13] should greatly accelerate our understanding of this unique virus system. The ability to manipulate genetically the viral genome is an invaluable tool in determining gene function and elucidating viral life cycle. Detailed genetic analysis in combination with continued structural, proteomic, transcriptomic and environmental metagenomic studies will continue to provide insights into the life cycle of STIV. Despite recent advances, many details of STIV's replication cycle, including mechanisms and timing of entry, disassembly, control of viral transcription, genome replication, assembly and cellular lysis, remain unknown. However, with these new tools in hand, rapid progress should be forthcoming.
Molecular Biology of Archaea: Biochemical Society Focused Meeting held at University of St Andrews, U.K., 19–21 August 2008. Organized and Edited by Stephen Bell (Oxford, U.K.) and Malcolm White (St Andrews, U.K.).
This work was supported by the National Science Foundation [grant number EF-0802200] and the Air Force Office of Scientific Research [grant number FA 9550-07-1-0533].