A Novel Type of Polyhedral Viruses Infecting Hyperthermophilic Archaea

ABSTRACT

Encapsidation of genetic material into polyhedral particles is one of the most common structural solutions employed by viruses infecting hosts in all three domains of life. Here, we describe a new virus of hyperthermophilic archaea, Sulfolobus polyhedral virus 1 (SPV1), which condenses its circular double-stranded DNA genome in a manner not previously observed for other known viruses. The genome complexed with virion proteins is wound up sinusoidally into a spherical coil which is surrounded by an envelope and further encased by an outer polyhedral capsid apparently composed of the 20-kDa virion protein. Lipids selectively acquired from the pool of host lipids are integral constituents of the virion. None of the major virion proteins of SPV1 show similarity to structural proteins of known viruses. However, minor structural proteins, which are predicted to mediate host recognition, are shared with other hyperthermophilic archaeal viruses infecting members of the order Sulfolobales. The SPV1 genome consists of 20,222 bp and contains 45 open reading frames, only one-fifth of which could be functionally annotated.

IMPORTANCE Viruses infecting hyperthermophilic archaea display a remarkable morphological diversity, often presenting architectural solutions not employed by known viruses of bacteria and eukaryotes. Here we present the isolation and characterization of Sulfolobus polyhedral virus 1, which condenses its genome into a unique spherical coil. Due to the original genomic and architectural features of SPV1, the virus should be considered a representative of a new viral family, “Portogloboviridae.”

INTRODUCTION

One of the most unexpected results of recent studies on viral diversity is the unveiling of astounding morphological variety of DNA viruses in geothermal environments with temperatures exceeding 80°C (1). Hardly over two dozen viruses isolated from such environments, all parasitizing hyperthermophilic members of the domain Archaea, display diverse virion shapes, many of which have not been observed among viruses from the two other domains of life, Bacteria and Eukarya. Due to distinct genome compositions and peculiar morphological properties of these viruses, 11 new virus families were established by the International Committee on Taxonomy of Viruses (ICTV) for the classification of these viruses (1–4).

The remarkable morphological diversity of hyperthermophilic archaeal viruses radically contrasts the limited structural diversity of double-stranded DNA (dsDNA) viruses of Bacteria, nearly all of which, with the exception of several pleomorphic species, have icosahedral particles, with or without helical tails (5). The existing picture seems to accurately reflect the actual differences between bacterial and archaeal viruses. The morphological landscape of bacterial viruses is formed by more than six thousand species, whereas that of hyperthermophilic archaeal viruses is formed by only a couple of dozen known species. Moreover, while no new morphotype of bacterial dsDNA virus has been identified for half a century, despite the isolation of thousands of new species (5), the variety of archaeal viral morphotypes continues to expand with the description of viruses from new environments and hosts (6–15). There are reasons to believe that the known variety of archaeal viruses represents no more than the tip of an iceberg and that comprehensive information on them may shed light on the problems of origin and evolution of viruses and virus-host interactions (16–28).

In attempts to further contribute to the knowledge on the diversity of archaeal viruses, we have analyzed virus-host systems in the hot, acidic springs in Beppu, Japan. Here, we report on the isolation and characterization of a polyhedral hyperthermophilic archaeal virus, Sulfolobus polyhedral virus 1 (SPV1), with a type of virion organization not previously observed for other known icosahedral viruses.

RESULTS

Virus isolation and virus-host relationships.

The aliquots of an environmental sample collected from the hot, acidic spring Umi Jigoku in Beppu, Japan, were used to inoculate the medium favorable for the growth of members of the genus Sulfolobus, which are known to dominate in acid thermal environments. After incubation at 75°C for 10 days, cell growth was detected, and the presence of polyhedral virus-like particles (VLPs) was observed in the enrichment culture by transmission electron microscopy (TEM). The particles appeared to be isometric, likely icosahedral, uniform in overall appearance and size (not shown). From the same culture, 50 isolates, numbered S1 to S50, were colony purified. None of the isolates was capable of VLP production. They were further analyzed for the ability to replicate the observed VLPs. Aliquots of the cell-free enrichment culture (5 μl) were applied onto Phytagel plates that contained cells of each of the 50 isolates prior to the lawn development. Cells of 40 isolates grew as lawns on the plates after incubation at 75°C for 3 days. Turbid zones of growth inhibition were formed around drops applied on the lawn formed by the isolate S38. The 16S rRNA gene sequence of the isolate revealed that it represents a strain of Sulfolobus shibatae, with the highest identity (99%) to the sequence of S. shibatae B12. From the lawn of Sulfolobus sp. strain S38, the zone of growth inhibition was excised and placed in the growing cell culture of the same species. As a result, the production of polyhedral VLPs was observed in the culture. For the verification of the infectious nature of these particles, they were collected from the cell-free culture supernatant and added to the growing culture of Sulfolobus sp. S38. A dramatic increase of the VLP concentration, as observed by TEM, indicated effective replication of the particles and confirmed that they represent infectious virions of a Sulfolobus virus.

Sulfolobus sp. S38 was subjected to an additional round of colony purification, and the isolated strain, designated Sulfolobus sp. S38A, was selected as a standard virus host for all following experiments. The polyhedral virus that replicated in Sulfolobus sp. S38A was named Sulfolobus polyhedral virus 1 (SPV1). The virus SPV1 was purified by 5 to 20% sucrose rate-zonal centrifugation followed by isopycnic centrifugation in the gradient of CsCl. The morphology and size of SPV1 virions were identical to those of the VLPs observed in the enrichment culture (Fig. 1).

FIG 1.

Electron micrographs of SPV1 virions. (A) Sample negative stained with 2% (wt/vol) uranyl acetate. (B) Sample embedded in vitreous ice. The open arrowhead points to the projection in the axial view, and the filled arrowhead points to the projection in the side view. Bars, 100 nm.

To study the host range of SPV1, purified virus particles were added to exponentially growing cultures of potential hosts, and virus propagation was monitored by TEM. Besides Sulfolobus sp. S38A, the virus could replicate in Sulfolobus islandicus strain REY15A (29). S. islandicus strains HVE10/4 and LAL14, S. solfataricus strains P1, P2, and 98/2, and Sulfolobus acidocaldarius did not serve as hosts for the SPV1 virus.

Infection of exponentially growing cells of Sulfolobus sp. 38A with SPV1 at a multiplicity of infection (MOI) of about 7 caused only slight retardation of host growth and did not lead to host cell death and lysis, as validated by the optical density at 600 nm (OD₆₀₀) measurements and enumeration of viable cells in the cell culture (Fig. 2). In line with these observations, application of 5 μl of concentrated virus preparation onto freshly prepared lawns of Sulfolobus sp. 38A did not cause, after development of the lawns, the appearance of zones of cell lysis (not shown). Consequently, we failed to establish a plaque assay on the lawn of Sulfolobus sp. S38A cells. Thus, virus titer was estimated by TEM, based on the comparison of SPV1 particle numbers with the numbers of virions in a standard preparation of Sulfolobus rod-shaped virus 2 with the known titer. A significant increase in the SPV1 titer in the culture of infected cells could be observed about 24 h postinfection and did not cause a decrease in the turbidity of the cell culture (Fig. 2). All these observations suggest that SPV1 may be a nonlytic virus.

FIG 2.

Impact of SPV1 infection on the growth kinetics of Sulfolobus sp. strain S38A culture. Growth curves (OD₆₀₀) of noninfected and infected cell cultures are shown by a broken line (white circles) and continuous black line (black circles), respectively. The bars show the numbers of CFU at indicated time intervals in noninfected (white bars) and infected (gray bars) cultures. The vertical arrow corresponds to the time of infection (5 h).

Virion structure.

The virions of SPV1, as observed by TEM and cryo-electron microscopy (cryo-EM), represent polyhedra with a diameter of about 87 nm from vertex to vertex and with a diameter of 83 nm from facet to facet (Fig. 1A and B). The virions carry an outer shell that can be removed partially (Fig. 3A and B) or completely (Fig. 3C and D) from the intact virions by one cycle of freezing and thawing or prolonged storage. The shell seems to be responsible for the polyhedral shape of the virions. The cores remaining after removal of the shell have slightly pleomorphic, spherical shape (Fig. 3C and D). Protrusions were observed on the surface of the core and can represent either remains of the outer shell or specific structures mediating the contact between the core and the outer shell (Fig. 3C and D, white filled arrowheads). The cores consist of the envelope (Fig. 3D, black arrow) and viral DNA packaged in the form of condensed nucleoprotein filament (Fig. 3D). The nucleoprotein filament remained condensed even after partial detachment of the core envelope (Fig. 3C, inset). In such cases, unwinding and release of the nucleoprotein filament could be observed (Fig. 3C, white open arrowheads).

FIG 3.

Electron micrographs of partially disrupted SPV1 virions. (A and B) Particles partially devoid of the outer shell. (C and D) Particles completely devoid of the outer shell. The solid white arrowheads point to protrusions on the surface of the inner core. The small black arrow points to the envelope of the inner core. Open arrowheads point to the nucleoprotein filament released from the core. (A and C) Samples negatively stained with 2% (wt/vol) uranyl acetate. (B and D) Samples embedded in vitreous ice. Bars, 100 nm.

Analysis of the cryo-EM micrographs of intact virions taken in two different projections provided information on the organization of the nucleoprotein filament in the virion (Fig. 1B). The patterns projected in the axial views depict seven concentric rings (Fig. 1B, black open arrowhead), whereas those projected in the side views depict 14 linear striations (Fig. 1B, black filled arrowhead). The observations suggest that the nucleoprotein filament, about 3 nm in width, is wound up sinusoidally into a spherical coil. This spherical coil is about 60 nm in diameter and comprises 14 loops of the nucleoprotein filament with various diameters and even spacing between the loops (Fig. 1B).

Virus genome.

The genome of SPV1 was extracted from highly purified virions and sequenced using the Illumina platform. Assembly of the sequencing reads resulted in a single circular contig of 20,222 bp. The viral genome was sensitive to various type II restriction endonucleases; digestion of the genome with the single-cutters XbaI and BglII resulted in a linearized product, which migrated in the agarose gel as a single sharp band, consistent with the genome being a circular dsDNA molecule. The SPV1 genome has a GC content of 38.3%, which is similar to that of various Sulfolobus genomes (e.g., 35% for S. islandicus [30]). The SPV1 genome contains 45 open reading frames (ORFs), which are tightly arranged and occupy 89.1% of the genome (Fig. 4). The ORFs are generally short, with median length of 103 codons.

FIG 4.

Genome map of SPV1. The ORFs are represented with arrows that indicate the direction of transcription. Genes encoding structural proteins (gray), homologs to other ORFs of Sulfolobales viruses (hatching), and ORFs encoding predicted membrane proteins (asterisks) are indicated.

Homology searches using the BLASTP program (31) revealed that less than one-third (∼27%) of SPV1 gene products are significantly similar (E-value cutoff of 1e−03) to sequences in the nonredundant protein database (Table 1). Notably, 10 ORFs had homologs in various viruses infecting hosts of the order Sulfolobales, including members of the families Rudiviridae, Lipothrixviridae, Fuselloviridae, and Turriviridae. Homologs of two additional ORFs were encoded in the genomes of members of the order Sulfolobales, rather than their viruses. A combination of BLASTP analysis and a more sensitive hidden Markov model (HMM)-based HHpred (32) analysis allowed the assignment of putative functions to just one-fifth of all SPV1 ORFs (nine ORFs, 20%). Seven ORFs encoded putative proteins containing various DNA-binding domains, including zinc finger (ORF02-234, ORF07-40, and ORF15-65), (winged) helix-turn-helix (ORF16-96, ORF22-147, and ORF35-111), and ribbon-helix-helix (ORF23-115) domains, which are frequently encoded by crenarchaeal viruses (33–35). In addition, ORF29-310 and ORF32-168 encode glycosyltransferase with the GT-B fold and the S-adenosyl-l-methionine-dependent methyltransferase, respectively. The closest homologs of the two proteins are encoded by Acidianus-infecting viruses of the order Ligamenvirales. In addition, ORF25-400 and ORF26-357 are homologous to each other and to the minor structural proteins conserved in all members of the family Rudiviridae (36) as well as in some members of the families Lipothrixviridae and Bicaudaviridae, all infecting members of the archaeal order Sulfolobales. Notably, whereas the N-terminal regions of ORF25-400 and ORF26-357 products showed highest similarity to the viral homologs, the central regions of the corresponding proteins were more similar to homologs encoded by “Candidatus Nanopusillus acidilobi,” a symbiotic archaeon of the order Nanoarchaeota inhabiting terrestrial geothermal environments (37).

TABLE 1.

Summary of predicted ORFs in the SPV1 genome

ORFa	Coordinatesb	Length (aa)c	No. of TMDsd	Annotatione	HHpred hit	Probability (%)	BLAST hit	Identity (%), E valuef
ORF01-171	77–592	171
ORF02-234	599–1303	234		Zn finger DNA-binding protein	2n25	99.2
ORF03-65	1300–1497	65
ORF04-110	1494–1826	110
ORF05-39	1813–1932	39
ORF06-83	1929–2180	83
ORF07-40	c(2305–2427)	40		C2H2 Zn-binding protein	2dmd	99.4	Sulfolobus islandicus rudivirus 3 (YP_009272958)	14/39 (36), 1e−03
ORF08-107	c(2414–2737)	107
ORF09-63	c(2771–2962)	63
ORF10-36	c(2946–3056)	36
ORF11-131	c(3064–3459)	131	1	GepA-like uncharacterized conserved protein	COG3600	99.5	Acidianus hospitalis (WP_048054695)	37/120 (31), 2e−06
							Acidianus filamentous virus 3 (YP_001604369)	33/113 (29), 1.7
ORF12-57	c(3590–3763)	57	2
ORF13-64	c(3750–3944)	64	1
ORF14-62	c(4186–4374)	62
ORF15-65	c(4364–4561)	65		SWIM Zn finger protein	PF04434g	97.1	Sulfolobus islandicus (WP_012953064)	28/66 (42), 5e−10
							Sulfolobus monocaudavirus SMV1 (YP_009008084)	24/61 (39), 4e−09
ORF16-96	c(4548–4838)	96		DNA-binding protein, wHTH	PF14947	97.6	Sulfolobus spindle-shaped virus 1 (NP_039783)	31/90 (34), 4e−10
ORF17-112	c(4838–5176)	112
ORF18-62	5475–5663	62	2
ORF19-208	5660–6286	208					Sulfolobus spindle-shaped virus 6 (YP_003331465)	62/180 (34), 1e−19
ORF20-34	6283–6387	34
ORF21-84	6380–6634	84
ORF22-147	6781–7224	147		DNA-binding protein, HTH; conserved in Sulfolobales	3m8j	97.6	Acidianus hospitalis (WP_048054614)	43/138 (31), 4e−06
ORF23-115	7205–7552	115		Transcriptional regulator, CopG-like RHH	PF12441	96.7
ORF24-43	c(7810–7941)	43
ORF25-400	c(7967–9169)	400		Structural protein	PF10102	98.7	Acidianus rod-shaped virus 1 (YP_001542644)	47/130 (36), 4e−10
							“Candidatus Nanopusillus acidilobi” (AMD30014)	72/259 (28), 3e−06
ORF26-357	c(9211–10284)	357		Structural protein	PF10102	98.7	Sulfolobus islandicus rudivirus 3 (YP_009272982)	38/106 (36), 1e−08
							“Candidatus Nanopusillus acidilobi” (AMD29619)	66/193 (34), 9e−03
ORF27-187	c(10289–10852)	187
ORF28-64	c(10846–11040)	64		Leucine-rich repeat protein	PF14580	98.7
ORF29-310	11083–12015	310		Glycosyltransferase, GT-B fold	1rzu	99.9	Acidianus filamentous virus 2 (YP_001496947)	93/354 (26), 9e−29
ORF30-103	c(12010–12321)	103	2				Sulfolobus turreted icosahedral virus 2 (YP_003591085)	59/103 (57), 3e−25
ORF31-70	c(12371–12583)	70
ORF32-168	12849–13355	168		SAM-dependent methyltransferase, FkbM-like	2py6	99.5	Acidianus rod-shaped virus 2 (YP_009230239)	74/144 (51), 3e−45
ORF33-75	c(13357–13584)	75
ORF34-84	c(13655–13909)	84
ORF35-111	c(13884–14219)	111		α-Helical protein, transcriptional regulator, HTH	PF04297	98.2	Sulfolobales archaeon AZ1 (EWG08170)	25/78 (32), 1e−03
ORF36-72	c(14216–14434)	72
ORF37-69	c(14424–14633)	69
ORF38-274	c(14639–15463)	274
ORF39-112	c(15456–15794)	112
ORF40-337	c(15807–16820)	337	1	Mainly β-stranded
ORF41-305	c(17106–18023)	305	1	Mainly β-stranded
ORF42-304	c(18023–18937)	304	3	Mainly α-helical
ORF43-181	c(18979–19524)	181		Mainly β-stranded
ORF44-119	c(19536–19895)	119	2	Mainly α-helical
ORF45-78	c(19906–20142)	78		α-Helical protein

^aGenes encoding structural proteins are shown in boldface type.

^bThe coordinates on the complementary strand are shown in parentheses after c (for complementary strand).

^cLength is shown by the number of amino acids (aa).

^dTMD, transmembrane domain.

^ewHTH, winged helix-turn-helix; RHH, ribbon-helix-helix; SAM, S-adenosyl-l-methionine.

^fThe number of identical to total number of aligned amino acids and percent identity shown in parentheses are shown before the commas. The E values are shown after the commas.

^gAccession numbers beginning with “PF” may be found in the Pfam database, http://pfam.xfam.org/.

Virion proteins.

Highly purified virions of SPV1 were analyzed by SDS-PAGE. Three prominent bands (bands B1, B2, and B3), and four weaker bands (B4, B5, B6, and B7), were observed on the gel stained with Coomassie brilliant blue (Fig. 5A). The protein content in each separate band was analyzed by liquid chromatography and tandem mass spectrometry (LC-MS/MS) (Fig. 5A and C). The results of the analysis have revealed that the protein in band B1 with an apparent molecular mass of about 8 kDa is encoded by SPV1 ORF45-78 (Fig. 5C). We denote it as viral protein 1 (VP1). The prominent band B2 contained two proteins—both with an apparent molecular mass of 12 kDa—encoded by SPV1 ORF35-111 and ORF44-119, which we denote as VP2 and VP3, correspondingly (Fig. 5A and C). The third prominent band, B3, covers a wide area corresponding to proteins with the apparent molecular masses ranging from 20 to 32 kDa. It contains a single protein encoded by SPV1 ORF43, denoted as VP4 (Fig. 5A and C). The traces of protein VP4 were also detected across the whole length of the gel above B3. The diffuse appearance of the B3 band could be caused by different degrees of putative posttranslational modifications of the VP4 as well as formation of protein multimers. Along with the genes of the four major virion proteins, the genes of the minor virion proteins were also identified: ORF42-304 (band B4 [VP5]), ORF41-305 (band B5 [VP6]), ORF40-337 (band B6 [VP7]), ORF25-400, and ORF26-357 (band B7 [VP8 and VP9]) (Fig. 5A and C).

FIG 5.

Characterization of SPV1 proteins. (A) SDS-PAGE of proteins in intact virions; (B) SDS-PAGE of proteins in virions with detached outer shell; (C) identification of the genes encoding proteins in intact virions by LC-MS/MS. B1 to B7, protein bands of proteins with identified genes; M, molecular mass standards.

Sequence analysis of the virion components showed that the products of ORF40-337, ORF41-305, ORF42-304, and ORF44-119 are predicted to be integral membrane proteins with one to three membrane-spanning α-helical segments (Table 1). This result suggests that the four minor structural proteins might be embedded within the envelope surrounding the virion nucleoprotein and located beneath the outer protein shell. The five remaining virion proteins are predicted to be soluble. Among these proteins, VP1 encoded by ORF45-78 is a highly basic (pI = 9.53), α-helical protein of 78 amino acids, which due to its positive charge is likely to bind to the viral genome. Interestingly, VP2 is also α-helical and is homologous to helix-turn-helix domain-containing transcription factors ubiquitous in archaea (but not viruses), suggesting that it is also a DNA-binding protein. VP4, the abundant protein constituent of the SPV1 virion, shares no similarity to other cellular or viral proteins. Importantly, partially disassembled virions in which the outer shell was largely removed by freezing and thawing contained a dramatically reduced amount of VP4, whereas the amounts of VP1 to VP3 remained rather constant (Fig. 5B). This result suggests that VP4 is the main component of the external polyhedral protein shell. Analysis of the predicted secondary structure suggests that VP4 is rich in β-strands. Finally, given that homologs of virion proteins encoded by ORF25-400 and ORF26-357 are found in morphologically different viruses infecting archaea of the order Sulfolobales, the two proteins are likely to mediate certain aspects of virus-host interactions. Consistently, both proteins contain the functionally uncharacterized domain DUF2341, which is widespread in bacterial and archaeal proteins and is usually fused to various lectin, immunoglobulin, and CARDB-like (cell adhesion related domain found in bacteria) domains (PF10102 [http://pfam.xfam.org/family/PF10102]), suggesting that products of ORF25-400 and ORF26-357 are also likely to participate in host recognition.

Virion lipids.

Lipids were extracted from purified SPV1 virions and analyzed by thin-layer chromatography, as described in Materials and Methods. One major band and several minor bands were observed on the chromatogram (Fig. 6). The pattern was dramatically different from that of the host lipids (Fig. 6) in the number of lipid types and in their relative abundance. This result suggests selective acquisition of viral lipids from the pool of host lipids. Most likely, the lipids are principal components of the envelope of the virion core. Notably, similar selectivity toward particular lipid species has been observed in several other archaeal viruses (38–40).

FIG 6.

Thin-layer chromatography of lipids extracted from purified SPV1 virions and uninfected cells of Sulfolobus sp. S38A. The solid black arrowhead points to the main lipid type of the SPV1 virion, and the open arrowheads point to the main lipid types of the host cell.

DISCUSSION

All viruses with polyhedral capsids known thus far obey icosahedral symmetry. Thus, SPV1 is also likely to have an icosahedral capsid, but this supposition remains to be verified experimentally. Viruses with icosahedral virions represent more than half of all recognized virus families (41). These viruses display remarkable diversity of virion sizes and complexity, from small capsids composed of a single protein, as in the case of simple single-stranded RNA (ssRNA) and ssDNA viruses (42, 43), to extravagantly large, multilayered virions, such as those of mimiviruses (44). There are three major architectural classes of viruses with icosahedral capsids and dsDNA genomes, two of which are featured by viruses infecting hosts in all three cellular domains, testifying to their antiquity (41). The first of these architectural classes includes viruses that build their capsids using the major capsid proteins (MCPs) with the so-called HK97-like fold (45). Bacterial and archaeal viruses with the HK97-like MCPs are classified into the order Caudovirales, whereas eukaryotic members belong to the order Herpesvirales. Besides the MCPs, these viruses employ similar virion assembly and maturation as well as genome packaging mechanisms, which are not found in viruses outside this virus assemblage (46). The second architectural class encompasses highly diverse dsDNA viruses with the vertical double-jelly-roll (DJR) MCPs, including members of the bacterial virus families Corticoviridae and Tectiviridae, archaeal viruses of the family Turriviridae, and eukaryotic viruses of the families Adenoviridae and Lavidaviridae as well as the proposed order “Megavirales” that comprises most of the large and giant eukaryotic viruses (47–49). Most viruses within this class contain an internal membrane vesicle located between the protein capsid and the dsDNA genome and carry genes that encode homologous A32-like genome packaging ATPases not found in other virus groups. Furthermore, internal membrane-containing bacterial and archaeal viruses of the family Sphaerolipoviridae encode two paralogous MCPs with single-jelly-roll folds which form capsomers similar to those formed from the DJR MCPs (50, 51). Thus, the latter group of viruses is considered to be evolutionarily related to, and possibly predates, viruses with the DJR MCPs (52). The third architectural class of viruses includes eukaryotic dsDNA viruses with small genomes (∼5 to 8 kb), namely, members of the families Polyomaviridae and Papillomaviridae. Viruses of the latter group encode single-jelly-roll MCPs and are thought to have evolved from eukaryotic ssDNA viruses (42, 53). SPV1 does not fall into any of the three architectural classes and displays several marked differences compared to the members of these classes, as detailed below.

The VP4 protein, which apparently forms the outer shell of the SPV1 virion and is equivalent to the MCPs of other polyhedral viruses, is not recognizably similar to any other viral or cellular protein. The secondary structure prediction suggests that 181-amino-acid (aa)-long VP4 contains 10 β-strands and 2 α-helixes. Thus, based on the secondary structure prediction and its sheer size, the protein is highly unlikely to adopt either the DJR or HK97-like fold. Although the predicted secondary structure is consistent with the single-jelly-roll fold, SPV1 does not seem to be related to sphaerolipoviruses, all of which employ two MCPs for the formation of external icosahedral capsid. Similarly, polyomaviruses and papillomaviruses are architecturally very different from SPV1 in that their genomes are considerably smaller (4.7 to 8.4 kb) and they do not contain the internal enveloped nucleoprotein core observed in SPV1. Furthermore, unlike dsDNA viruses with the DJR and HK97-like MCPs, which encode distinct types of genome packaging motors, SPV1 lacks recognizable genes for putative ATPases, a class of proteins that is typically among the easiest to recognize by sequence analysis due to the presence of highly conserved Walker A and Walker B motifs. Instead, SPV1 condenses its genome in the form of a unique spherical core, consisting of the highly ordered nucleoprotein complex.

The VP2 protein, one of the major structural proteins of SPV1, is homologous to ubiquitous archaeal transcription regulators with the helix-turn-helix motif (Table 1). Notably, some archaeal viruses, such as spindle-shaped virus SSV1, incorporate host-encoded chromatin proteins into their virions (38). It is likely that in the course of evolution, the cellular gene encoding the DNA-binding protein has been horizontally acquired by the ancestor of SPV1 and adapted to function as a structural component of the virion. Indeed, recruitment of cellular proteins for virion structure appears to be a recurrent theme in the evolution of viruses (41).

Whereas the proteins involved in virion formation, with the notable exception of the putative receptor-binding proteins encoded by ORF25-400 and ORF26-357, are specific to SPV1, a considerable fraction of SPV1 genes are shared with other hyperthermophilic viruses infecting members of the Sulfolobales. The latter category of genes includes several DNA-binding proteins, a glycosyltransferase and a methyltransferase. This observation is in agreement with the recent bipartite network analysis of the known archaeal virosphere which revealed that the archaeal virus network consists of 10 modules which are linked via connector genes encoding a small set of widespread proteins, most notably the ribbon-helix-helix domain-containing transcription factors and glycosyltransferases (33), neither of which is a viral hallmark protein (54). However, such a lack of strong connectivity among the modules indicates that most of the viral groups within the archaeal virosphere are evolutionarily distinct. SPV1 integrates into the global archaeal virus network on the account of the same connector genes but is otherwise unrelated to other archaeal viruses.

Due to the unique genomic and architectural features of SPV1, we propose that the virus be considered the first representative of a new viral family, which we tentatively name “Portogloboviridae” (from Latin porto [to bear, carry] and globus [a ball]).

MATERIALS AND METHODS

Enrichment culture and isolation of SPV1 and host strain.

The environmental sample of translucent liquid mixed with red sand was collected from the hot acidic spring Umi Jigoku (80°C, pH 3.7) in Beppu, Japan. An aliquot (10 ml) of the sample was used to inoculate 40 ml of Sulfolobus growth medium (55), and the culture was incubated for 10 days at 75°C under aerobic conditions without shaking. Cell-free culture supernatant containing VLPs was collected by centrifugation, followed by filtration through a membrane with a pore size of 0.22 μm (Merck Millipore). The single strains were colony purified from the enrichment culture by plating on Phytalgel (Sigma-Aldrich) plates and incubated for 5 days at 75°C as described previously (55), except that Gelzan CM Gelrite was substituted with Phytalgel. The oligonucleotide primers used for 16S rRNA gene amplification are A21F (F stands for forward) (5′-TTCCGGTTGATCCYGCCGGA-3′) and U1525R (R stands for reverse) (5′-AAGGAGGTGATCCAGCC-3′). The primers S38A_399F (5′-CCGGAGACCAGGATACCAACTAG-3′) and S38A_1065R (5′-TCTACAAAGGCGGGGGAATAAG-3′) were designed based on the obtained sequences. The 16S rRNA gene sequence of Sulfolobus sp. S38A was assembled.

The following collection strains were analyzed for the ability to replicate the SPV1 virus: S. islandicus LAL14/1 (30), REY15A (29), and HVE 10/4 (29), S. solfataricus P1 (GenBank accession no. NZ_LT549890), P2 (56), and 98/2 (57), and S. acidocaldarius DSM 639 (58). Aliquots of virus preparation were added to the exponentially growing cultures of these strains, and after incubation for 2 days at 78°C, the cultures were diluted 1:50 and further grown for 2 days. The presence of virus particles in the cultures was monitored by TEM.

Infection studies.

We attempted to establish plaque assay on Sulfolobus sp. strain S38A lawns in different media, as described earlier (55). The plates were incubated for up to 3 days at 75°C.

To test the effect of SPV1 infection on host cell growth, the cells of Sulfolobus sp. S38A at the early logarithmic growth stage were infected with SPV1 at an MOI of about 7 and incubated at 78°C with shaking. The cell density (OD₆₀₀) and the number of viable cells (CFU) were measured at appropriate time intervals. The CFU counting was performed as described previously (27).

Due to the inability of SPV1 to form plaques on the host lawn, the virus titer was estimated by counting the numbers of virions on meshes on copper grids by TEM. As a standard, we used a preparation of Sulfolobus rod-shaped virus 2 (59) with known titer.

SPV1 production and purification.

For virus production, 250 ml of exponentially growing Sulfolobus sp. S38A cell culture (OD₆₀₀, ∼0.5) was infected by SPV1 at an MOI of 0.1 to 0.5. The infected culture was incubated for about 24 h at 78°C with shaking and diluted with fresh medium 1:5 and incubated further for 36 to 48 h. The cells were removed by centrifugation (Sorvall SLA 3000 rotor, 9,000 rpm, 30 min), and the supernatant was collected and filtered through a series of membranes (Merck Millipore) with pore sizes of 1.2 μm, 0.65 μm, and 0.45 μm for complete removal of cells and cell debris. The virus particles were collected from the cell-free fraction either by precipitation with ammonium sulfate (60% [wt/vol] saturation) at 4°C (60) or by flip-flow filtration through Vivaflow-200 cassettes (Sartorius Stedim Biotech, France). The concentrated SPV1 particles were resuspended in buffer containing 50 mM citric acid, 100 mM Na₂HPO₄, and 500 mM NaCl (pH 3.6).

The concentrated SPV1 particles were purified by two rounds of centrifugation. First, they were purified by 5% to 20% sucrose rate zonal centrifugation (25,000 rpm, 20 min, 10°C, Beckman rotor SW40Ti). The fraction containing the light-scattering zone was collected, and the presence of virus particles was verified by TEM. This fraction was further purified by isopycnic gradient centrifugation in CsCl as described previously (38).

The outer shell was removed from the SPV1 virions by freezing at −80°C, followed by thawing at room temperature. The partially disassembled virions were subjected to CsCl gradient centrifugation as described above.

Electron microscopy.

For TEM and cryo-EM, the samples were prepared and analyzed as described previously (13). The pictures were recorded using a Falcon II direct electron detector (FEI, USA).

Extraction and analysis of SPV1 DNA.

Nucleic acid was isolated from purified viral particles as described previously (11). SPV1 genome libraries were prepared with the Nextflex PCRFree kit (Bioo Scientific), and samples were sequenced by Illumina Miseq with paired-end 250-bp read lengths (for both ends) (Genomics Platform, Institut Pasteur, France). An average coverage of 3,000 was obtained. The sequence was assembled using CLC Genomics Workbench software package. ORFs were predicted using GeneMark.hmm v3.25 (61) and RAST v2.0 (62). The in silico-translated protein sequences were used as queries to search for sequence homologs in the nonredundant protein database at the National Center for Biotechnology Information using BLASTP (31) with an upper threshold E value of 1e−3. Searches for distant homologs were performed using HHpred (32) against different protein databases, including PFAM (Database of Protein Families), PDB (Protein Data Bank), CDD (Conserved Domains Database), and COG (Clusters of Orthologous Groups), which are accessible via the HHpred website. Transmembrane domains were predicted using TMHMM (63), whereas the secondary structure was predicted using Jpred (64) and PsiPred (65).

Analysis of SPV1 structural proteins.

The highly purified SPV1 virions were analyzed for protein content by using 4 to 12% gradient NuPAGE Bis-Tris precast gels (Thermo Fisher Scientific). Protein bands were stained with Coomassie blue using InstantBlue (Expedeon). Stained protein bands were excised, and digested “in gel” with trypsin. The digested peptides were analyzed at the Proteomics Platform of the Institut Pasteur by nano-LC-MS/MS using an Ultimate 3000 system (Dionex) as described previously (11). The peptide masses were searched against annotated SPV1 proteins using Andromeda (66) with MaxQuant software, version 1.3.0.5 (67).

Analysis of SPV1 lipids.

Lipids were extracted from highly purified SPV1 preparation and from uninfected cells of Sulfolobus sp. S38A as described previously (68). The lipid extracts were dissolved in chloroform-methanol-H₂O (65:24:4 [vol-vol-vol]) and separated by thin-layer chromatography on silica gel 60 plates (Merck) using chloroform-methanol-H₂O (65:24:4 [vol-vol-vol]) as the solvent. The lipids were visualized by molybdate (68).

Accession number(s).

The SPV1 genome sequence has been deposited in the GenBank database (accession no. KY780159). The Sulfolobus sp. S38A 16S rRNA gene partial sequence has been submitted to the GenBank database (accession no. KY927925).