|Home | About | Journals | Submit | Contact Us | Français|
Archaea often live in extreme, harsh environments such as acidic hot springs and hypersaline waters. To date, only two icosahedrally symmetric, membrane-containing archaeal viruses, SH1 and Sulfolobus turreted icosahedral virus (STIV), have been described in detail. We report the sequence and three-dimensional structure of a third such virus isolated from a hyperthermoacidophilic crenarchaeon, Sulfolobus strain G4ST-2. Characterization of this new isolate revealed it to be similar to STIV on the levels of genome and structural organization. The genome organization indicates that these two viruses have diverged from a common ancestor. Interestingly, the prominent surface turrets of the two viruses are strikingly different. By sequencing and mass spectrometry, we mapped several large insertions and deletions in the known structural proteins that could account for these differences and showed that both viruses can infect the same host. A combination of genomic and proteomic analyses revealed important new insights into the structural organization of these viruses and added to our limited knowledge of archaeal virus life cycles and host-cell interactions.
The members of the Archaea are the least understood of the three domains of life. Known archaeal viruses (~40) constitute less than 1% of the ~8,000 species of viruses that have been described (13, 59). So far, virtually all the archaeal viruses have double-stranded DNA (dsDNA) genomes, with the first viral single-stranded DNA (ssDNA) genome being the most recently described (46, 61). Thus, our knowledge of archaeal viruses is currently heavily biased due to this underrepresentation. Nevertheless, when viruses of the Archaea are isolated, it has been observed that they exhibit an exceptional diversity of morphotypes, most of which are not encountered among dsDNA viruses of the Bacteria and Eukarya (10, 49).
Hot acidic springs, in which the temperatures rise above 75°C and the pH drops below 4.0, are typical environments for the archaeon Sulfolobus, a member of the Crenarchaeota. Four different types of viruses with different morphologies have been found associated with Sulfolobus; they are members of the Fuselloviridae, Rudiviridae, Lipothrixviridae, and Guttaviridae (10, 50, 55). Members of the Fuselloviridae (Sulfolobus spindle-shaped virus 1 [SSV1], SSV2, SSV4, SSV5, SSV6, SSV7, SSV Kamchatka 1 [SSV-K1], and SSV Yellow stone 1 [SSV-Y1]) have circular, ~15-kbp-long dsDNA genomes. They are spindle shaped (100 by 60 nm), with two coat proteins enclosing the genome and a DNA-binding protein (VP2) (52, 53, 65). Members of the Rudiviridae (Sulfolobus islandicus rod-shaped virus 1 [SIRV1] and SIRV2) are stiff, rod-shaped (830- to 900- by 23-nm) viruses with linear, ~35-kbp dsDNA genomes (44, 48). The flexible, filamentous (2,000- by 24-nm) Sulfolobus islandicus filamentous virus (SIFV) (Lipothrixviridae) has a 41-kbp linear dsDNA genome, with two structural proteins. Finally, the droplet-shaped (100- to 185- by 70- to 95-nm) Sulfolobus newzealandicus droplet-shaped virus (SNDV) (Guttaviridae) has a 20-kbp circular dsDNA genome. In addition to the viruses that belong to the four characterized families, an unassigned virus, Sulfolobus turreted icosahedral virus (STIV), also infects Sulfolobus (55). So far, only two viruses infecting Sulfolobus (STIV and SIRV2) are known to be lytic (9, 12, 27, 36, 42, 55).
STIV was isolated from a hyperthermoacidophilic archaeon, Sulfolobus solfataricus, originating from a hot spring in Yellowstone National Park (55). The circular genome of STIV is 17.6 kbp long, with 37 predicted open reading frames (ORFs) (36, 55). Cryo-electron microscopy (cryo-EM) and image reconstruction of STIV to 27-Å resolution (55) revealed a virus capsid with a pseudo-triangulation number (T) of 31. The vertices are decorated by large turrets, which are thought to be involved in host recognition and/or attachment. There is a membrane underlying the capsid (55). So far, four STIV protein structures have been determined: the 37-kDa major capsid protein B345 (27) and three proteins with putative functions: a glycosyltransferase, A197 (34), and the DNA-binding proteins F93 (32) and B116 (33). The double-β-barrel fold of the major capsid protein is conserved within viruses from very diverse hosts, like the human adenovirus, Paramecium bursaria Chlorella virus type 1 (PBCV-1) and bacteriophage PRD1 (6, 40, 56). These viruses, of the so-called “PRD1-adeno lineage,” are suggested to have a common ancestor that precedes the division of the three domains of life (5).
Here we report the isolation, genome sequence, identification of the major virion proteins, and three-dimensional structural characterization of a novel, icosahedrally symmetric archaeal virus. We demonstrate that, despite overall architectural similarity to STIV, the turreted host attachment structures of these two viruses differ significantly. Despite these differences, the viruses can infect the same host, indicating that the host cell recognition site lies in conserved regions of the capsid structure.
The Sulfolobus growth medium was 25 mM (NH4)2SO4, 3 mM K2SO4, 1.5 mM KCl, 20 mM glycine, 4.0 μM MnCl2, 10.4 μM Na2B4O7, 0.38 μM ZnSO4, 0.13 μM CuSO4, 62 nM Na2MoO4, 59 nM VOSO4, 18 nM CoSO4, 19 nM NiSO4, 0.1 mM HCl, 1 mM MgCl2, 0.3 mM Ca(NO3)2 adjusted to pH 3.5 with H2SO4, 0.2% tryptone (T medium). For ST medium, a small amount of elemental sulfur was added to the same medium, and for CT medium, the final Ca2+ concentration was adjusted to 0.1% with supplemental CaCl2. Plates were made in T medium with 0.8% Gelzan (Sigma); soft agar contained 0.4% Gelzan (12).
Samples were taken from the hot spring IceG4 (88.3°C, pH 3.5) in the Hverakjalki valley, Hveragerdi, Iceland, in April 2006. A 1-ml sample of both liquid and silt was diluted in 50 ml of ST medium and shaken for 1 week at 78°C with aeration. A 1-ml aliquot was transferred to 50 ml of fresh ST medium and incubated for a further 4 days. This enrichment culture was designated G4ST. One round of colony purification was carried out, and 50 colonies were grown in 5 ml of ST medium. STIV2 was observed in the supernatant of strain G4ST-2 in transmission electron microscopy (TEM) by negative staining of the sample (Fig. (Fig.1).1). In order to identify the host, its 16S rRNA gene was amplified by PCR and sequenced (primers 8aF, TCYGGTTGATCCTGCC, and 1512uR, ACGGHTACCTTGTTACGACTT).
Sulfolobus islandicus Hve10/4 (43) and S. solfataricus P2 DSM1617 and P1 DSM1616 were tested as hosts using spot tests, but we were unable to detect any infection. However, very recently, we found that plaques did form on lawns of a new isolate of S. solfataricus 2-2-12 (a kind gift of Mark Young), essentially as described previously (12). Briefly, 2-2-12 was grown in 25 ml of T medium, pH 3.5, at 80°C and 130 rpm for 48 h to an optical density at 600 nm (OD600) of 0.3. The culture was diluted 1:25 in 50 ml of CT medium, pH 3.5, and grown for an additional 48 h at 80°C and 130 rpm to an OD600 of ~0.3 (2 × 108 CFU/ml). For the plaque assay, 500 μl of cells was mixed with 100 μl of virus and incubated at 80°C for 15 min. Five milliliters of prewarmed 0.4% Gelzan in CT medium was mixed with the cells, and the mixture was poured on 0.8% Gelzan plates in CT medium. The presence of STIV2 in the plaques was tested by PCR using primers for the gene a345 (MCP-F, GCCGCCGCTAGCATGGGAAGTATATATACAGAAA, and MCP-R, GCCGCCCTCGAGTTATCTCTTAATGCTCTTTCTC).
In order to determine how large a fraction of G4ST-2 cells were infected by STIV2, the strain was plated, and 59 colonies were picked and restreaked. Colony PCR was used to detect the presence of STIV2 in all of the colonies. Two of these colonies, G4ST-2-41 and G4ST-2-42, were transferred to liquid culture, and after 2 days, cells from 1 ml of each were harvested by centrifugation (5 min at 6,000 × g), washed three times in nutrient-free medium, and then plated. Twelve colonies each of G4ST-2-41 and G4ST-2-42 were picked, restreaked, and then shown by PCR to continuously carry STIV2. Two of the restreaked colonies from G4ST-2-42, G4ST-2-42-1, and G4ST-2-42-2 were also transferred to liquid culture, where after 2 days of growth, STIV2 particles could be observed in the supernatant by TEM.
To produce STIV2, a fresh culture of G4ST-2 was grown for 3 to 5 days, and the medium was cleared by low-speed centrifugation. The resulting supernatant was concentrated 12,000 times by ultrafiltration (Vivaflow 200 casette; 100-kDa cutoff; Sartorius AG) and subsequently filtered using 0.45-μm and 0.2-μm filters (Whatman) to remove residual cells. The cleared supernatant was further concentrated by ultrafiltration (Millipore Amicon Ultra; 100-kDa molecular-mass cutoff). The concentrated supernatant was purified by rate-zonal centrifugation on a linear 15 to 50% sucrose gradient in 25 or 50 mM sodium citrate, pH 3.5 or 4.5 (40,000 rpm; 25 min; 18°C; Beckman SW50.1 rotor). The resulting fractions were diluted 1:5, and the viruses were collected by differential ultracentrifugation (SW50.1 rotor; 40,000 rpm; 40 min; 18°C). The virus pellets were resuspended in 25 or 50 mM sodium citrate, pH 3.5 or 4.5, and used immediately for preparation of vitrified specimens and protein analysis.
To confirm the presence of virus, we used colony PCR in addition to electron microscopy. Cells were resuspended in distilled water. Each PCR mixture (25 μl) contained 1 μl of the template. The PCRs were done according to the manufacturer's instructions (Phusion High-fidelity DNA polymerase; Finnzymes). The primers used to detect gene a345 were STIV2-F4, ATCCACGACCTACTTCT, and STIV2-R4, CTTTCTCATTCTTCTTCACCTC, or FSUC345, AGATTGGTGGCATGGGAAGTATATATACA, and RSUC345, GAGGAGAGTTTAGACTTATCTCTTAATGCTCTT. The primers used to detect gene a259 were FSUB259, AGATTGGTGGCATGGGATTAGGAACAGAG, and RSUB259, GAGGAGAGTTTAGACTCACCCGTGCTTATGTTT. Plaques were picked and resuspended in distilled water prior to PCR, as described above.
STIV2 was digested with protease K and phenol-chloroform extracted, and the DNA was subsequently precipitated with ethanol and redissolved in 10 mM Tris-HCl, pH 8. The preparation was used for random DNA amplification with the GenomiPhi amplification kit (GE Healthcare) and then digested with EcoRI. Two bands from an agarose gel were cloned into pUC19 and sequenced in MegaBACE 1000 DNA Sequencers (GE Healthcare). Based on the sequences of the two fragments, primers were designed for long-range PCR (LA TaKaRa; Takara Bio Inc.). Two divergent long-range PCR products were obtained. These PCR products were then sonicated and made into a library based on the Linker Amplified Shotgun Library method (http://www.sci.sdsu.edu/PHAGE/LASL/). The plasmid DNA of library clones was purified using a Model 8000 Biorobot (Qiagen) and sequenced. Sequences from the shotgun library and from the EcoRI fragments were assembled using Sequencher 4.5 (Gene Codes Corporation). Ambiguities and low-coverage regions were checked by PCR, using the original STIV2 DNA preparation as a template. Gene recognition, annotation, and sequence comparisons were done using MUTAGEN (11) and NCBI BLAST (2).
Homology modeling of structural STIV2 proteins and the putative ATPase was done using the I-TASSER server (66, 67). Only models with a TM value of >0.5 were considered to have the correct topology (66, 67). A trimer of A345 was built from the homology model by superimposing the monomer on each of the three monomers in the atomic model of the PRD1 trimer (Protein Data Bank identifier [PDB ID], 1HX6) to create a capsomer model in COOT (7, 17). Visualization of the protein models was done with UCSF Chimera (45).
The proteins in the virus preparations were resolved using SDS-polyacrylamide gels (31, 41). A Fermentas PageRuler or PageRuler Plus Prestained Protein Ladder (size range, 11 to 250 kDa) was used as a standard. For mass spectrometric analyses, the proteins were resolved using a 20% SDS-polyacrylamide gel (31), the lane containing the virus sample was divided into 10 pieces, and each gel piece was analyzed separately. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis was performed using an Ultimate 3000 nano-LC (Dionex) and a QStar Elite hybrid quadrupole time-of-flight (TOF) mass spectrometer (Applied Biosystems/MDS Sciex) with nano-electrospray (nano-ES) ionization. The LC-MS/MS samples were first loaded on a ProteCol C18 trap column (3 μm; 10 mm by 150 μm; 12 nm) (SGE), followed by peptide separation on a PepMap100 C18 analytical column (5 μm; 15 cm by 75 μm; 10 nm) (LC Packings/Dionex) at 200 nl/min. The separation gradient consisted of 0 to 50% B in 20 min, 50% B for 3 min, 50 to 100% B in 2 min, and 100% B for 3 min (buffer A, 0.1% formic acid; buffer B, 0.08% formic acid in 80% acetonitrile). MS data were acquired using Analyst QS 2.0 software. The LC-MS/MS data were searched with the in-house Mascot version 2.2 against the sequence of STIV2.
For negative staining, purified virus particles were applied to carbon- and Formvar-coated grids (Ted Pella Inc.) and stained with 2% acidic uranyl acetate. Micrographs were taken with a Jeol 1200 EX II electron microscope operating at 80 kV (Electron Microscopy Platform at the Institute Pasteur, Paris, France). Virus samples were vitrified for cryo-electron microscopy as previously described (1). Digital image collection on a Gatan UltraScan 4000 charge-coupled device (CCD) camera under low-dose conditions using 200 kV on an FEI Tecnai F20 (Electron Microscopy Unit, Institute of Biotechnology, University of Helsinki) has been described previously (24).
Images were processed as described previously (24), except that the images were binned to a pixel size of 4.42 Å pixel−1. The three-dimensional model of the archaeal virus STIV (55) was scaled to the appropriate size and used as a starting model to determine the orientations and origins of the virus particles in a model-based approach (3). PFT2 and EM3DR2 (3) were used in the initial rounds of refinement and POR and P3DR (25, 37) for subsequent rounds. The resolution of the three-dimensional reconstruction was assessed using a Fourier shell correlation cutoff of 0.5 (19). To visualize the turrets of STIV2, the homology model trimer of the major capsid protein A345 was fitted into the capsid shell in UCSF Chimera (45), and a pseudo-atomic model of the capsid was generated in Bsoft (20). The pseudo-atomic model was further filtered to correspond to the resolution of the three-dimensional reconstruction prior to subtraction of the densities. The turret was subsequently manually segmented using the EMAN program QSEGMENT (35). The density connecting the turret to the membranes was segmented from the whole virion map. The difference imaging of the STIV turrets was done similarly, using the atomic model of STIV B345 (PDB ID, 2BBD) (27).
The genome sequence has been deposited with GenBank (accession number GU080336). The 16S rRNA sequence of G4ST-2 has been deposited with GenBank (accession number GQ495670). The STIV2 reconstruction has been deposited in the EMDB and at the PDB with the accession code EMD-1679.
An icosahedral virus resembling STIV (55) was observed in a mixed enrichment culture while examining the viral diversity of Icelandic terrestrial acidic hot springs. The host strain, G4ST-2, that produced this icosahedral virus was purified by single-colony isolation. Phylogenetic analysis of the 16S rRNA sequence of the G4ST-2 isolate (GenBank accession no. GQ495670) identified it as most closely related to Sulfolobus islandicus strains Y.G.57.14 and M.16.27. The virions were observed in the liquid culture medium by TEM (Fig. (Fig.1),1), and due to the similarity of its appearance to that of STIV, the new isolate was named STIV2 (Sulfolobus Turreted Icosahedral Virus 2). Once the genome sequence was completed, colony PCR was used to show that 100% of the colonies tested (n = 59) were infected with STIV2 and that this infection persisted through two further rounds of culturing. Virions were detected in the supernatant of these cultures by TEM, suggesting that the detected DNA belonged to a replicating form of STIV2. Thin sections have so far not revealed STIV2 virions inside G4ST-2 (data not shown).
The host range of STIV2 was screened with three different S. solfataricus isolates. We were able to determine the titer of STIV2 produced from G4ST-2 on an S. solfataricus isolate (2-2-12) that is also susceptible to STIV (12), but not on the other tested isolates. Plaques were detected after 2 to 3 days of incubation at 80°C, indicating that STIV2 retards the growth of S. solfataricus 2-2-12, possibly due to host cell lysis (12). The titer of STIV2 in the G4ST-2 culture supernatant could be as low as 102 PFU/ml. The presence of STIV2 in the plaques was confirmed by PCR using primers derived from the STIV2 sequence. The virus was still infectious after purification.
Virion proteins were identified from a concentrated virus preparation separated by SDS-PAGE (see Fig. S1 in the supplemental material), digested with trypsin within gel slices, and analyzed by LC-MS/MS. Peptides were found that belonged to the STIV2 proteins A259, C141, A55, B72, A103, E132, B631, A345, and C510 (see Table S1 in the supplemental material).
Additionally, two host-derived proteins were detected, Sso0881 and Sso0909 (UniProtKB Q97ZL3 and Q97ZJ7, respectively) (57). However, these two proteins are probably not virion proteins but components of host-derived vesicles that copurify with STIV2. Sso0909 is a p60 katanin ATPase (57), and Sso0881 has been shown to contaminate STIV preparations, as well (36).
Long-range PCR products from the STIV2 genome showed that it was circular. The 16,622-bp dsDNA genome, illustrated in Fig. Fig.2,2, includes 64 ORFs >50 amino acids long (allowing for the use of GTG and TTG start codons, which are common in Archaea) (15, 63). However, our analysis suggests that only 34 of these are genes, based on the presence of archaeal TATA box and Shine-Dalgarno sequences (Fig. (Fig.2;2; see Table S2 in the supplemental material). The proteins encoded by these ORFs range in predicted size from 6 to 67 kDa. The genes are named according to their reading frames (a to f) and the numbers of amino acid residues of the gene products coded for, e.g., a259, in accordance with the established naming convention for some other archaeal viruses.
Of the 34 suggested genes, 2 ORFs, coding for the viral structural proteins A103 and E132, are unique to STIV2, with no homologues in the sequence databases. Overall, the genome organization is highly similar to that of STIV, and some regions exhibit over 70% nucleotide identity. This allowed us not only to annotate the STIV2 genome, but also to reannotate STIV. Twenty-five ORFs have homologues in STIV using the latest available STIV annotation (accession number NC_005892.1) (55). We have annotated another 7 STIV ORFs through their similarity to STIV2 (STIV ORFs e93, f60, c55, c88, d161, f65a, and b111) (see Table S2 in the supplemental material). Additionally, two STIV ORFs were extended by considering their homology to STIV2 and the occurrence of TTG as a start codon, as it is used in 10 to 15% of all genes in sequenced Sulfolobus strains (63). Thus, a106 becomes a213, making it equivalent to STIV2 c221, and c67 becomes c133, which, together with b109, forms a full-length equivalent to STIV2 a259. The last is most likely a gene fusion in STIV2, since b109 and c133 are similar to the 5′ and 3′ sequences of a259, respectively. Furthermore, there is a large insert in STIV2 (b631 compared to STIV a223), a 5′ insertion in STIV2 (b204 compared to STIV b164), and a 5′ deletion in STIV2 (b510 compared to STIV c557). All of these changes occur in genes coding for confirmed virion structural proteins. It is noteworthy that STIV2 lacks a structural protein corresponding to STIV C381.
STIV2 also has a few limited similarities to other virus genes. One STIV2 protein (B72) is found in STIV and is similar to the SSV1 (Fusellovirus) DNA-binding protein VP2 (54). Furthermore, there are two ORFs homologous to common crenarchaeal ORFs. B60 is in a family of potential RHH transcriptional regulators (8, 44, 51, 64), and b116 is in a family coding for DNA-binding proteins (26, 33).
It was possible to reliably model four STIV2 proteins (see Table S3 in the supplemental material), all of which have STIV counterparts whose structures have been solved by X-ray crystallography (27, 32-34). Importantly, these models include the conserved major coat protein A345 as a double-β-barrel protein shown in Fig. S2A in the supplemental material (27). Hence, it is highly likely that A345 (sequence identity to STIV, 88.4%), B116 (sequence identity to STIV, 60.2%), F98 (sequence identity to STIV, 78.6%), and B197 (sequence identity to STIV, 80.2%) all have folds similar to those of their STIV counterparts (see Fig. S2 and Tables S2 and S3 in the supplemental material). A reliable model was generated for the putative ATPase B204 (see Fig. S2 and Tables S2 and S3 in the supplemental material) as a P-loop ATPase that would belong to a dsDNA virus-specific clade that has been predicted from sequence analysis (23, 36, 62). It is predicted to have a fold similar to that of the Pseudomonas aeruginosa FstK DNA translocase (39).
We analyzed enriched preparations of STIV2 from 48 liters of G4ST-2 using cryo-EM and three-dimensional image reconstruction (14, 16, 18). Due to the very sparse occurrence of the virus (often less than 5 per EM grid) and the large number of host-derived vesicles, the images were collected on a CCD camera, where they could be immediately analyzed for the presence or absence of particles. The majority of the STIV2 particles were DNA-filled icosahedral particles (Fig. (Fig.1)1) with large, protruding turrets on the vertices (Fig. (Fig.1,1, inset A, white arrow). Occasionally, empty, DNA-lacking, and often damaged particles were observed in which both the capsid and what appeared to be an internal membrane were evident (Fig. (Fig.1,1, inset B).
Image reconstruction to 20 Å resolution of 713 particles from 358 micrographs with a defocus range of 0.5 to 5.2 μm (Fig. (Fig.3A)3A) showed an outer protein capsid and an underlying lipid membrane following the icosahedral shape of the capsid and enclosing the dsDNA genome (Fig. (Fig.3B).3B). Based on the length of the genome and the internal volume of STIV2 (enclosed by the membrane), the packaging density of the genome is 0.58 bp nm−3. The dimensions of the icosahedral capsid are 71 nm facet to facet and 93 nm turret to turret. The average thickness of the capsid is 9 nm, and that of the membrane is 7 nm (calculated from a radial profile [not shown]). The capsomers of STIV2 are arranged on a pseudo-T=31d lattice. Each capsomer has local 3-fold symmetry and readily accommodates a trimer of the major capsid protein A345, with the two β-barrels per monomer giving the trimeric capsomers pseudo-6-fold symmetry. Thus, the copy number of A345 is 300. These trimers are packed together with approximate local p3 plane group symmetry (58).
Strikingly, large 5-fold symmetric turrets protrude from the vertices (Fig. (Fig.3C).3C). These turrets extend 11 nm outside the capsid shell, and the length of the turret is 24 nm (Fig. (Fig.3D).3D). The turrets are approximately 10 nm across at the top and 7.5 nm at the capsid level, with a clear 1.5-nm-wide tunnel in the middle (Fig. (Fig.3D).3D). The calculated mass for each STIV2 turret is ~520 kDa, assuming a protein concentration of 1.35 g/ml and a contour level of 1 σ above the mean in EMAN (35).
When the reconstruction of STIV is compared to that of STIV2, the most prominent difference is the overall shape of the turrets (Fig. 3C and E). For comparison, the turrets were segmented from the reconstructions by subtracting all 300 copies of the major capsid protein from the density, leaving the turrets and the membrane. The superposition and difference imaging of the two turrets showed that there are similar barrels common to the two turrets, but STIV has in addition ear-like appendages and a cap on top of the barrel (Fig. 3D and F). The base plate that has been described previously in STIV appears to be the outer layer of the membrane in STIV2 (Fig. (Fig.3G)3G) (27).
We introduce a new member of the growing pool of characterized hyperthermophilic icosahedral archaeal viruses. STIV2 was isolated by enrichment from a Sulfolobus sp. host growing in a hot terrestrial spring in Iceland. As for many other archaeal viruses described to date, analysis of the genome and its included genes is difficult due to a lack of similarity to other sequences deposited in public databases. However, comparison with the related crenarchaeal virus STIV allowed a deeper analysis of the genomes and structures of both viruses, leading to new hypotheses about the proteins involved in host cell lysis (27, 36, 55).
The sequence comparison of STIV2 and STIV clarifies our view of their genome organizations, confirms predicted ORFs in both viruses, and highlights genome rearrangements. The comparison illustrates that genome rearrangements have occurred, as well as the loss and gain of some genes in either virus, e.g., the presence of the unique genes e132 and a103 in STIV2. As STIV2 also shares genes with other archaeal viruses, such as the Fuselloviridae, and harbors a conserved gene (b116) for a crenarchaeal DNA-binding protein, it seems evident that horizontal gene transfer takes place between archaeal viruses (28, 29). This could be promoted by virus movement between communities, as suggested by earlier population studies on the globally distributed metacommunity (60).
We observed two apparently different levels of virus production, depending on the host used. In the natural isolate S. islandicus, we found that successive single-colony isolates were positive for the virus DNA and virus particles, and the latter were produced at a very low frequency. We did not manage to cure the host, so controlled studies of the effect of the virus on the natural host were not possible, and we turned to the literature on other archaeal viruses for potential explanations. Such behavior has been observed before for archaeal viruses, such as SSV1 (formerly called SAV 1) (38). In SSV1, the viral DNA was found as independently replicating genetic elements and also integrated into the genome of the host. When SSV1-infected cells were induced with UV light, many more viral particles were produced, apparently without killing the host cells. Recently, the halophilic euryarchaeal virus Halorubrum pleomorphic virus 1 (HRPV-1) has also been shown to be secreted from Halorubrum without host cell lysis or integration into the genome. It retards the growth of the host so that plaques can be detected (46). STIV apparently also produced only very few viruses on its original host, S. solfataricus YNPRC179, although the reason for this has not been reported (55). When a different susceptible Sulfolobus strain was isolated for STIV (S. solfataricus 2-2-12), virus production was much higher, and lysis of the host cells was noticed, along with release of virus through pyramidal protrusions of the cell envelope (12). We hypothesize that this explains the second phenotype that we noticed with STIV2, that is, plaque formation on S. solfataricus 2-2-12, and we are working to verify this. Interestingly, SIRV2 has recently been reported to erupt from pyramidal protrusions causing cell lysis, as well (9). As no lysis genes have yet been identified in these archaeal viruses, we used our genome analysis to look for common genes between SIRV2, STIV, and STIV2 and came up with only two candidate genes common to all three viruses: STIV2 b60 and b116 (see Table S2 in the supplemental material). Further experiments can now be planned to test whether these genes are involved in host cell lysis in Sulfolobus. In addition, one can also look for similar host cell lysis for other viruses containing the same gene pair, in order to elucidate the mechanism. Candidate viruses that we have identified through sequence analysis are Acidianus filamentous virus 1 (AFV1), Acidianus rod-shaped virus (ARV1), and Acidianus two-tailed virus 1 (ATV) (8, 51, 64). Notably, the euryarchaeal virus SH1, which is also clearly lytic, apparently without resorting to pyramids, does not have genes homologous to those of these Sulfolobus viruses (4, 47).
Newly released virions need to recognize and attach to new host cells to initiate fresh infections. In STIV, it has been proposed that the turrets are responsible for recognition/attachment (55). As we noticed that the major difference between STIV and STIV2 structures was the presence of the ears on the STIV turrets, we tested whether STIV2 could infect the same host as STIV. Indeed S. solfataricus 2-2-12 was susceptible to STIV2, so we conclude that the STIV ears are probably not required for infection in that host. In addition, by comparing the known structural-protein contents of the two viruses from mass spectrometry (see Fig. S1 and Tables S1 and S2 in the supplemental material), it was obvious that only one protein (STIV C381) is missing from STIV2 (36, 55). Maaty et al. (36) suggested that the STIV turrets are formed from three proteins (STIV C381, A223, and C557), with C557 forming the ears due to its proposed ankyrin-like fold in the N terminus. Since C557 and A223 homologues are present in STIV2 (see Table S2 in the supplemental material) but C381 is found only in STIV, we propose that the STIV turret ears are most likely formed from C381. The turrets thus remind us of the T-even phage tail fibers and the spike protein, P5, of phage PRD1, which are important for increasing the efficiency of infection by helping in host cell recognition but are not essential for irreversible attachment to the host (21, 22, 61).
Based on protein masses, segmented volumes, and structure prediction, we propose a model for the organization of the nine known structural proteins of STIV2 (Fig. (Fig.3H;3H; see Table S1 in the supplemental material). We assign B631 and C510 to the turret as pentamers, with the C terminus of C510 containing the predicted transmembrane helices facing the membrane. This would give a mass of 580 kDa for the turret, which agrees reasonably well with our turret volume calculation, which gave a mass of 520 kDa. A345 is the double-β-barrel-containing major capsid protein that forms the majority of the pseudo-T=31 lattice, interacting directly with the turret proteins. We place 5 additional proteins with predicted transmembrane helices into the membrane (A259, C141, E132, A55, and A103), including the two that are apparently unique to STIV2 (E132 and A103). B72 is highly basic and similar to known DNA-binding proteins of other archaeal viruses and is thus suggested to interact with the circular dsDNA genome (53). A204 appears to be an ATPase, so it could be used, for instance, in DNA packaging and thus be attached to the capsid either transiently or as a structural protein. We saw a few empty particles in the preparations that could be procapsids, but so far, we have not detected the ATPase by mass spectrometry of the virion, so it has been omitted from our current model.
Through sequence, structure, infection, and protein analyses of STIV2 and comparison to other known archaeal viruses, we propose a model for the organization of STIV2 and suggest the existence of a possible common lysis mechanism reliant on two candidate genes that is used by viruses infecting Sulfolobus species.
We thank Pasi Laurinmäki, Benita Löflund, and Annunziata Pennino for excellent technical support, Mark Young for kindly providing the Sulfolobus solfataricus 2-2-12 isolate, Kim Brügger for organizing the MUTAGEN database used for annotation and analyses of the virus genomes, Jack Johnson for providing the STIV reconstruction, Esko Oksanen for useful discussions, Gunilla Rönnholm and Elina Ahola-Iivarinen for mass spectrometry analyses, and the Electron Microscopy Unit and Protein Chemistry Core Facility, Institute of Biotechnology, Helsinki University, and the Electron Microscopy Platform at the Institut Pasteur, Paris, France, for providing facilities.
The work was supported by the Academy of Finland Centre of Excellence Programme in Virus Research (2006 to 2011; 1129684 [S.J.B.]) and Biocentrum Helsinki (S.J.B.) and by the Agence Nationale de la Recherche, France (NT05-2_41674 [D.P. and P.R.]), the Danish Research Council for Natural Sciences (X.P.), the Norwegian Research Council (grant 172206 to L.J.R.), and the Viikki Graduate School in Molecular Biosciences (L.J.H.).
Published ahead of print on 17 February 2010.
†Supplemental material for this article may be found at http://jvi.asm.org/.