Particle Assembly and Ultrastructural Features Associated with Replication of the Lytic Archaeal Virus Sulfolobus Turreted Icosahedral Virus

Abstract

Little is known about the replication cycle of archaeal viruses. We have investigated the ultrastructural changes of Sulfolobus solfataricus P2 associated with infection by Sulfolobus turreted icosahedral virus (STIV). A time course of a near synchronous STIV infection was analyzed using both scanning and transmission electron microscopy. Assembly of STIV particles, including particles lacking DNA, was observed within cells, and fully assembled STIV particles were visible by 30 h postinfection (hpi). STIV was determined to be a lytic virus, causing cell disruption beginning at 30 hpi. Prior to cell lysis, virus infection resulted in the formation of pyramid-like projections from the cell surface. These projections, which have not been documented in any other host-virus system, appeared to be caused by the protrusion of the cell membrane beyond the bordering S-layer. These structures are thought to be sites at which progeny virus particles are released from infected cells. Based on these observations of lysis, a plaque assay was developed for STIV. From these studies we propose an overall assembly model for STIV.

Crenarchaeal viruses are characterized by morphologies and genes not detected in viruses infecting other organisms (37, 41). Although we have genome sequences for most of the crenarchaeal viruses isolated, we have only a rudimentary understanding of archaeal virus assembly and release from cells. Among archaeal viruses, there are some that lyse their host cells (5, 11, 16, 32, 35, 36, 38-40, 43, 50), but most of these infect Euryarchaeota. Bacteriophages that are architecturally similar to the crenarchaeal virus Sulfolobus turreted icosahedral virus (STIV) lyse their host cells (3, 9, 12, 22), but most viruses that infect crenarchaeal and some euryarchaeal hosts have been observed to extrude from the cell without causing cell death (13, 30, 41, 57).

Most of the information about cell lysis is based on observations of a decrease in optical density (OD) in an infected culture, plaque assays, or plate growth inhibition assays, but little is known about the actual mechanism of archaeal lysis. Three exceptions are the ψM1 and ψM2 phages and the prophage ψM100, which infect members of the Euryarchaeota and encode the lytic enzyme pseudomurein endoisopeptidase (28, 39). This enzyme hydrolyzes the pseudomurein protein in the cell wall, causing host lysis (19). All known double-stranded DNA (dsDNA) bacteriophage encode a holin-endolysin system that creates holes in the bacterial cell membrane, followed by disruption of the cell wall by a murein-degrading enzyme (50, 55, 58-60). The single-stranded RNA and ssDNA phages use an amurin type of protein that prevents cell wall synthesis (7, 8, 34), leaving the cell membrane unprotected and subject to lysis. Unlike Bacteria, Crenarchaeota lack muramic acid but have an S-layer composed of a single glycoprotein directly associated with the cytoplasmic membrane (21). A small protein (15) may anchor the S-layer to the cell membrane, as in Sulfolobus acidocaldarius, while narrow bridges between three globular domains of the S-layer protein may allow curved surfaces to form around the cell (6, 10, 44, 53).

STIV was isolated from a boiling acid hot spring found in Yellowstone National Park, WY (47, 48). Sequence analysis of the 17.3-kb circular dsDNA genome predicts 37 open reading frames (ORFs) with little or no similarity to other genes in the databases except for ORFs encoded by other viruses infecting Sulfolobus spp. (29, 37, 48). Structural analysis of individual viral proteins has identified possible functions for four of the ORFs, including the major capsid protein (MCP; B345) (18, 48), two DNA binding proteins (ORFs B116 and F93) (23, 24), and a glycosyl transferase (A197) (25). Proteomic analysis of purified STIV particles revealed 11 proteins, with two host proteins (Sso7d and SSO0881) and nine virus proteins (29). Structural analysis of the 72-nm icosahedral virion revealed particles built upon a pseudo-T=31 lattice with turret-like features extending from the fivefold axis of the particle and an internal lipid membrane derived from the host (18, 48). The overall architecture of STIV and the structure of the MCP have been shown to be similar to bacteriophage PRD1, adenoviruses, the algal virus PbCV-1, the euryarchaeal virus SH1, and, more recently, the marine phage PM2, suggesting an ancient evolutionary link (1, 2, 4, 18, 20, 42, 45, 48). Beyond the overall architecture of the particles, bioinformatics analysis of structural proteins has suggested that the STIV protein B164 may be a P-loop ATPase, while A223 and C381 from STIV are predicted to have similarities to the P5 vertex protein of PRD1 (29). The combination of genetic and structural studies has revealed a great deal about the virion itself, but the replication cycle of STIV including virus capsid assembly and DNA packaging remains unknown.

Analysis of Sulfolobus solfataricus strain 2-2-12 infection by STIV using microarrays has provided information about virus transcription and differential expression of host genes (36). The infection cycle of STIV was shown to be nearly synchronous, with a single round of viral transcription first detected at 8 h postinfection (hpi), a peak in transcription at 24 hpi, and cell lysis beginning at 32 hpi. All of the structural genes on the microarray were first detected at 16 hpi. In response to infection, 124 host genes were upregulated, and 53 genes were downregulated. The upregulated genes appeared to be used by the virus for replication while the downregulated genes, detected at 32 hpi, were likely a response to the approaching cell lysis. The nearly synchronous single cycle replication of STIV in S. solfataricus strain 2-2-12 led us to examine the cellular ultrastructural changes associated with STIV infection. In this study, we have used scanning and transmission electron microscopy (SEM and TEM, respectively) to follow the assembly of mature virus particles within the cytoplasm of infected cells, formation of membrane-associated structures, and subsequent lysis of cells. Detection of cell lysis suggested that a plaque assay could be developed to further study host-virus interactions in this system. As a result of these studies, a general assembly and release model for STIV is proposed.

MATERIALS AND METHODS

Time course of STIV infection in S. solfataricus.

S. solfataricus strain 2-2-12 (36) was isolated from S. solfataricus P2 (ATCC isolate DSM 1617) and found to be highly susceptible to STIV infection. Cells were grown from glycerol stocks in medium 182 (M182; http://www.dsmz.de/microorganisms/html/media/medium000182.html) at 80°C, pH 3.5. The culture was grown for ∼48 h, reaching late log phase, and was diluted to 500 ml in M182 medium at pH 2.5. At an OD at 650 nm (OD₆₅₀) of ∼0.3, the culture was further diluted to 2 liters and split into two separate 1-liter cultures, which were grown to an OD₆₅₀ of ∼0.3. One culture was infected with purified STIV particles at a multiplicity of infection of ∼1.5 to 2 while the other culture was used as an uninfected control. Aliquots from both cultures were collected every 8 to 56 hpi for analysis of cell density and STIV genome abundance and for microarray analysis, as previously described (36). Samples across the time course were also taken for SEM analysis (described below). Additional samples were also collected on an hourly basis from 23 to 33 hpi for TEM analysis.

TEM analysis.

At each time point, cells were collected from 1.0-ml samples by low-speed centrifugation. Cells were fixed for 48 h in 3% glutaraldehyde, collected by centrifugation, and resuspended in a small volume of cool 2% noble agar. Once solidified, the cell pellet was removed from the microcentrifuge tube, cut into small pieces, and fixed overnight with 3% glutaraldehyde in 0.05 M potassium sodium phosphate buffer (PSPB), pH 7.2. Agar pieces were subsequently washed three times with PSPB for 10 min each, followed by postfixation in 2% osmium tetroxide at room temperature for 4 h. Samples were dehydrated using an ethanol series (50 to 100%), followed by addition of the transitional solvent propylene oxide. Following dehydration, cell pieces were gradually infiltrated with Spurr's resin (49) and baked overnight at 70°C. Thin sections, 60 to 90 nm, were cut with a Diatome diamond knife on a Reichert OM-U2 ultramicrotome, floated onto 300-mesh copper grids, and stained with uranyl acetate and Reynold's lead citrate (46). All sections were viewed with a LEO 912AB TEM and photographed with a Proscan 2,048- by 2,048-pixel charge-coupled-device camera.

TEM image analysis.

Thin sections of uninfected and infected cells prepared for TEM were viewed under low magnification (×8,000), and cell types were counted at each time point. Only thin sections from different blocks or nonsequential sections were used for counts. At least 700 cells were counted at each time point and assigned to the following categories: (i) intact cells without virus, (ii) intact cells with virus, (iii) cells devoid of cellular content (empty cells), and (iv) broken cells.

SEM analysis.

Aliquots (1 ml) of cell culture from times 0, 32, and 48 hpi were fixed with 3% glutaraldehyde in PSPB, pH 7.2. After fixation, cells were gently centrifuged, washed two times for 10 min each in PSPB, and resuspended in 2% osmium tetroxide in PSPB for 4 h at room temperature. Cells were subsequently collected by centrifugation and washed two times with PSPB for 10 min each. After resuspension, cells were adsorbed to polylysine-coated silicon wafers and treated as previously described (52). Some of the cells, adsorbed onto silicon wafers, were also subjected to critical-point drying using a SAMDRI-795 (Tousimis) critical-point drier. Samples were examined using a Zeiss Supra 55 VP field emission SEM.

Plaque assay development.

S. solfataricus strain 2-2-12 was grown in 25 ml of M182 medium at pH 3.5 and 80°C for ∼48 h. The culture was diluted to 100 ml using M182 medium, pH 2.5, and grown for 18 h at 80°C. Cells were prepared for the plaque assay by a second 1:25 dilution followed by 18 h of growth to a final OD₆₅₀ of ∼0.3. For the plaque assay, 500 μl of cell culture (∼2 × 10⁸ cells) was incubated for 15 min with a range of virus from 2 ×10⁷ to 2 × 10¹¹ purified STIV particles at 80°C without shaking. After incubation, 5 ml of top-layer medium (M182 medium, pH 2.5, with 0.3% Gelrite) was mixed with the cells. The top layer was poured on prewarmed plates (M182 medium, pH 2.5, with 0.8% Gelrite) and allowed to solidify prior to incubation in sealed plastic containers at 80°C. After 3 to 4 days, clear plaques were observed. Plaques were tested for the presence of STIV by performing PCR using primers specific for the major coat protein of STIV on material directly from the plaque region itself or from the surrounding cellular lawn. Samples from the plaque regions were also directly inspected by TEM for the presence of STIV particles. The presence of infectious virus particles associated with the plaques was assayed by adding material from the center of a plaque to a fresh culture of S. solfataricus strain 2-2-12 at pH 2.5 and monitoring the culture OD₆₅₀. Subsamples of the culture were analyzed by immune dot blot detection with polyclonal antibodies against the MCP of STIV.

RESULTS

Nearly synchronous, single-cycle virus replication observed.

As has been previously described (36), a highly susceptible isolate of S. solfataricus P2 was used for these studies. The majority of cells were infected with virus, allowing for TEM- and SEM-based analysis to follow a time course of infection (Fig. 1). Detection of virus genomes by quantitative PCR confirmed that a single wave of virus genome production occurred, which reached a maximum at 40 hpi (36). By 40 hpi, 91% of the cells appeared to be lysed, based on evidence that cells were broken or devoid of cellular content (Fig. 1 and 2). Analysis of cultures in the absence of virus suggests that 7 to 10% of the cells are nonviable (results not shown), indicating that nearly 100% of the viable cells become infected by STIV.

FIG. 1.

Percentage of four different cell forms (cells without virus, cells with virus, cells devoid of cellular content [empty], and broken cells, as shown in TEM images) were determined over 56 hpi in a culture infected with STIV. Cells are approximately l μm in diameter.

FIG. 2.

Representative TEM sections taken during a time course of the population of S. solfataricus cells at 8 to 56 h hpi. (A to D) Noninfected control cells. (E to H) STIV-infected cells. Bar, 500 nm.

Assembly of virus particles within infected cells.

No significant differences between uninfected control cells and STIV-infected cells were observed by TEM until 24 hpi. Between 24 to 32 hpi, virus-like particles were beginning to appear within the interior of infected cells (Fig. 1, 2, and 3). Initially there was a general increase in cellular membranes not seen in uninfected cells (Fig. 3A). Abundant circular crystalline structures were evident in less dense areas of the cells (Fig. 3B). We estimate that 55% of the population had virus-like particles by 32 hpi, with as many as 50 virus-like particles detected in a single cell section. The first virus-like particles detected appeared to be immature assembly products that could be distinguished by their more spherical shape. These immature particles appear to be made up of the internal lipid coated by the major coat protein (Fig. 3C). These particles also have incorporated the exterior projecting virus protein spikes we refer to as the turret-like structures (Fig. 3C and D,). Many of the immature particles appear to have thicker protein shells (Fig. 3E). Starting at 24 hpi, there is a mixed population of immature virus particles lacking dense cores (Fig. 3C to E) and mature particles with dense cores (Fig. 3E). We assume that particles with dense cores have been filled with the viral genomic DNA. These particles are fully mature: they have an angular shape, possess the distinct turret-like projections from the particle surface, and have enhanced core density (Fig. 3E). These features are similar to those seen previously in three-dimensional TEM cryo-image reconstructions of purified STIV particles (48). STIV particle assembly is not completely synchronized as mature virus particles were found along with immature particles up until the time of cell lysis (Fig. 3B to E). By 32 hpi, infected cells were nearly completely packed with a dense array of virus particles, which were released into the medium following cell lysis.

FIG. 3.

STIV assembly products within infected S. solfataricus cells are shown from different time points, as indicated. (A) Increase in membranes present within the cell before virus development (arrows). (B) Presence of circular crystalline-like areas (arrowheads and insets 1 and 2). (C) Immature virus particles forming with MCP on the exterior lipid surface (thick arrow). Some immature particles and mature particles show the presence of exterior surface-projecting turret-like structures (thin arrows). (D and E) Mixed populations of immature virus particles lacking dense interior cores (thick arrows) and mature virus particles with dense interior cores (arrowheads). Turret-like projections are present in both mature and immature particles (thin arrows). Scale bars are indicated.

Cellular ultrastructural changes associated with STIV infection.

STIV infection results in the lysis of its host cell. There appears to be a strict temporal regulation of cell lysis. Prior to 32 hpi, little or no cell lysis is observed (Fig. 1). By 32 hpi, the cell membrane could be seen extruding through the exterior S-layer (Fig. 4A2 and B2) and ultimately bursting (Fig. 4A3 and B3), leaving empty cellular remains held together by the remaining S-layers (Fig. 4A4 and B4). The protrusion of the cellular membrane was particularly striking when viewed by SEM, appearing as star points or pyramid-like projections extending out from the cell surface (Fig. 4A2 and enlargements C1, C2, and C3). Individual protrusions and clusters of pyramid-like protrusions were often seen (Fig. 4A2, B2, and C2) around the cell. The distribution of these protrusions does not suggest any type of spatial control. The pyramid-like protrusions appeared to expand through breaches in the protective S-layer. Perhaps internal osmotic pressure ultimately leads to disruption of this unprotected membrane (Fig. 4A3 and B3). Folds seen on the pyramid-like membrane (Fig. 4A3, C1, and C2) could be artifacts produced by the dehydration required for TEM/SEM sample preparation although no differences were seen between cells prepared with or without critical-point drying during SEM analysis. The empty cells with an S-layer boundary and the broken membrane could be observed in TEM and SEM images (Fig. 4A3, B3, and D1). We could not determine if the cell membrane of the pyramid-like structure was actually continuous with the regular cell membrane. In some cases it appeared to be a separate membrane (Fig. 4D1), but this could also be an artifact caused by display of the cell thin section onto a two-dimensional surface.

FIG. 4.

SEM images (row A) and corresponding TEM images (row B) of S. solfataricus cells show different stages of infection. (A1 and B1) Noninfected cells. (A2 and B2) Cells infected with STIV displaying membrane protrusions (thin arrows). (A3 and B3) Lysing cells releasing virus (thin arrows) and cell contents. (A4 and B4) Empty cells showing S-layer and broken membrane fragments (thin arrows). Pyramid-like structures from STIV-infected cells observed by SEM (C1 and C2) and TEM (C3) are also shown.(D1) TEM image of broken membrane and S-layer after cell lysis. Scale bars are indicated.

STIV forms plaques.

Infection of S. solfataricus strain 2-2-12 with STIV consistently results in plaques with an approximate diameter of 0.3 to 0.8 cm (Fig. 5 A) 3 to 4 days after infection. PCR-based detection of the major coat protein gene (Fig. 5B) and visual inspection with TEM demonstrate the presence of STIV particles in the plaques. STIV particles sampled from the interior of these plaques were used to successfully infect a new culture of S. solfataricus strain 2-2-12 (Fig. 5C). The stock of purified virus was determined to have 2 × 10⁵ PFU/ml (independently repeated four times). We estimated the total number of virus-like particles in this purified STIV stock using an epifluorescence microscopy counting assay to be 3.22 × 10¹¹ virus-like particles/ml (56). Therefore, the absolute efficiency of plating (the plaque titer divided by the number of virus particles in the experimental sample) is 4.2 ×10⁻⁴.

FIG. 5.

(A) STIV plaques on a lawn of S. solfataricus cells. (B) PCR-based assay for the STIV MCP gene from independent plaques (lanes 2 to 5), the nonplaque region in the lawn (lanes 6 and 7), a no-template PCR control (lane 8), and PCR control of cloned STIV MCP (lane 9). Lane 1, DNA molecular weight standards. (C) Immune dot blot detection of STIV MCP from a time course of S. solfataricus cells inoculated directly with plaque material.

DISCUSSION

The analysis of the ultrastructural changes associated with infection of S. solfataricus strain 2-2-12 by STIV demonstrates that it is a lytic virus that follows an orchestrated program of virion assembly and cell lysis. After an initial wave of viral transcription, intermediates in viral assembly are observed within the cell. These intermediates include immature lipid particles associated with both the MCP and turrets on the exterior surface and complete virus particles lacking viral DNA. These particles subsequently package their viral DNA, converting the particles to their mature form within the cell. Soon afterwards, the infected cells undergo programmed lysis due to the disruption of the cellular S-layer and subsequent break in the cell membrane. The lytic nature of STIV allowed the development of a plaque assay for this virus, which should greatly facilitate future genetic analysis of the virus-host relationships.

Based on the results of this study, we propose the following model for STIV particle assembly. Mature STIV particles have an internal lipid envelope, and we speculate that the round areas in the cytoplasm, which can be seen along with the appearance of the virus protein shell, are lipid material. It appears that the membrane for each particle forms as a round crystalline area, and the protein shell either assembles very quickly around it or coassembles with the lipids. We presume that this protein shell is composed of the previously identified 38-kDa MCP, which forms the icosahedral particle based on a T=31 lattice (48). PRD1 is known to have a protein that acts as an assembly factor and helps form the lipid membrane. This protein, P10, does not occur in the mature particle (33). It is possible that assembly proteins are present during assembly of STIV particles, but none has been identified at this time although several ORFs have yet to have functions assigned. There does not appear to be a connection between the areas where particles are formed and the cellular membrane. Based on previous studies (29), we know that the composition of lipids in the STIV particles is basically an enriched subset of the host membrane, suggesting that the lipids are selected specifically by the virus. This appears to be similar for other viruses such as SH-1, φ6, PM2, and PRD1, where the phage membrane is enriched in phosphatidylglycerol compared to the host (3, 17, 26, 27). As in STIV, assembly of PRD1 involves the formation of empty particles that contain lipid and proteins found in mature virions (33). Soon after formation of this lipid and protein vesicle, the viral turret structures of STIV are assembled and inserted into the lipid membrane. The early assembly product is devoid of the viral genome as empty, thick-shelled, rounded particles with a circular inner membrane were observed in thin sections of STIV-infected cells. These empty particles likely represent an intermediate stage as is observed for PRD1 and Bam35 (1, 12, 31, 54). However, creating a prohead does not appear to be a universally conserved feature of the PRD1-adenovirus lineage since other members such as PM2 appear to assemble their membrane around the viral DNA genome (1). Packaging of the prohead is followed by the insertion of the turrets. As in PRD1 infections, a single turret likely facilitates the entry and packaging of the viral DNA into the STIV particles (14, 51). Although the occurrence of a single differentiated turret has not been shown for STIV, the available structural information for STIV particles does not preclude its existence. Based on the similarities between STIV and PRD1, we propose a similar model for packaging of the genome of STIV. After packaging of the genomic DNA, the particles appear more angular and have the thinner shells associated with the mature particles.

Lysis of S. solfataricus cells due to STIV infection appears to be quite different from lysis seen in bacterial cells due to phage infection. From microarray studies, there is no evidence that STIV interferes with the synthesis of the cell membrane or S-layer (36); however, this is based on transcription of the genes, and viral control could occur at a later step. There are no visible holes, damage, or weak areas in the Sulfolobus S-layer or cell membrane prior to pyramid formation, suggesting that a holin/endolysin-type system is not present in this virus. Likewise, there is no evidence that the STIV genome codes for holin- or endolysin-like proteins. We can only speculate as to the actual mechanism and components involved in the lysis. The pyramid structure does not have any S-layer associated with it, suggesting that the virus may disrupt the anchoring glycoprotein, preventing the S-layer from attaching to the cell membrane. Several of the unidentified proteins in the STIV genome appear to have transmembrane domains and could act to prevent S-layer attachment without disrupting the cell membrane. The development of a reliable plaque assay provides a valuable tool for further investigation of viral protein involvement in cell lysis. Regardless of the exact mechanism, it appears that STIV, like other lytic viruses, precisely orchestrates cell lysis.