ORF4 of the Temperate Archaeal Virus SNJ1 Governs the Lysis-Lysogeny Switch and Superinfection Immunity

Archaeal viruses are important parts of the virosphere. Understanding how they regulate their life cycles and interact with host cells provide crucial insights into their biological functions and the evolutionary histories of viruses. However, mechanistic studies of the life cycle of archaeal viruses are scarce due to a lack of genetic tools and demanding cultivation conditions. Here, we discover that the temperate haloarchaeal virus SNJ1, which infects Natrinema sp. strain J7, employs a lysis-lysogeny switch and establishes superinfection immunity like bacteriophages. We show that its ORF4 is critical for both processes and acts as a repressor of the replication of SNJ1. These results establish ORF4 as a master regulator of SNJ1 life cycle and provides novel insights on the regulation of life cycles by temperate archaeal viruses and on their interactions with host cells.

ABSTRACT

Recent environmental and metagenomic studies have considerably increased the repertoire of archaeal viruses and suggested that they play important roles in nutrient cycling in the biosphere. However, very little is known about how they regulate their life cycles and interact with their hosts. Here, we report that the life cycle of the temperate haloarchaeal virus SNJ1 is controlled by the product ORF4, a small protein belonging to the antitoxin MazE superfamily. We show that ORF4 controls the lysis-lysogeny switch of SNJ1 and mediates superinfection immunity by repression of genomic DNA replication of the superinfecting viruses. Bioinformatic analysis shows that ORF4 is highly conserved in two SNJ1-like proviruses, suggesting that the mechanisms for lysis-lysogeny switch and superinfection immunity are conserved in this group of viruses. As the lysis-lysogeny switch and superinfection immunity of archaeal viruses have been poorly studied, we suggest that SNJ1 could serve as a model system to study these processes.

IMPORTANCE Archaeal viruses are important parts of the virosphere. Understanding how they regulate their life cycles and interact with host cells provide crucial insights into their biological functions and the evolutionary histories of viruses. However, mechanistic studies of the life cycle of archaeal viruses are scarce due to a lack of genetic tools and demanding cultivation conditions. Here, we discover that the temperate haloarchaeal virus SNJ1, which infects Natrinema sp. strain J7, employs a lysis-lysogeny switch and establishes superinfection immunity like bacteriophages. We show that its ORF4 is critical for both processes and acts as a repressor of the replication of SNJ1. These results establish ORF4 as a master regulator of SNJ1 life cycle and provides novel insights on the regulation of life cycles by temperate archaeal viruses and on their interactions with host cells.

INTRODUCTION

The study of archaeal viruses is still in its infancy compared to the widely reported bacteriophages (about 8,500 species) (1). Only about 120 species of archaeal viruses have been sequenced and investigated to date; however, they display remarkable diversity in virion morphologies and gene contents (2–4). Temperate life cycles are common among viruses of bacteria and archaea, whereby viral genomes are stably maintained within the cell either as extrachromosomal episomes or are integrated into the host chromosomes. This mode of propagation provides ample opportunities for the virus and its host to coevolve (5, 6). Therefore, research on the viral life cycle and the interaction between temperate viruses and their hosts can provide crucial insights into the evolutionary histories of viruses and novel modes of virus-host interactions (7, 8). However, due to the lack of suitable genetic tools and demanding cultivation conditions, only a few archaeal temperate viruses have been studied experimentally so far (9–17).

Unlike temperate archaeal viruses, temperate bacteriophages are ubiquitous, and the regulation of their life cycle has been extensively studied (18–20). A logical parallel may be drawn from these studies: temperate viruses have a dual life cycle, which can terminate either in host cell lysis or lysogeny. If the lytic pathway is chosen, the virus actively replicates its genome, produces progeny virions, and lyses the host cell. In contrast, if the lysogenic pathway ensues, the viral genome will be present in a dormant state either as an extrachromosomal element or as a provirus integrated into the host chromosome, and the expression of the genes responsible for the lytic pathway is repressed. Temperate viruses often confer immunity to their host cells against subsequent infection by the same and closely related viruses, a phenomenon known as superinfection immunity (21, 22). When host cell fitness decreases, such as upon exposure to DNA-damaging agents, the provirus switches to the lytic pathway (23). The switch between the lytic and lysogenic states of temperate bacteriophages is very sophisticated and is normally controlled by a master regulator which senses the fitness of the host cell and represses the expression of the lytic pathway genes. The cI protein of the Escherichia coli phage λ is a prototypical master regulator which represses the expression of the lytic pathway genes and is cleaved under conditions of DNA damage (23–26). Among archaeal viruses, the switch between lytic and temperate life cycles has been explored in the case of the spindle-shaped virus SSV1 infecting hyperthermophilic archaea of the genus Sulfolobus (15–17, 27). It has been suggested that SSV1 protein F55 binds to host protein RadA, and this complex represses the genes responsible for active virus replication; dissociation of the F55-RadA complex upon DNA damage leads to transcriptional derepression and active replication of SSV1 (17). However, whether such a mechanism is conserved in other archaeal viruses, especially in viruses infecting distantly related archaea, such as halophiles, remains unknown.

The temperate haloarchaeal virus SNJ1 isolated from Natrinema sp. J7-1 resides in the cytoplasm as a circular plasmid called pHH205 (9, 28). A small number of virions (with a titer of about 10⁶ PFU ml⁻¹) are released during cultivation, presumably due to spontaneous induction in some cells (see Fig. S1 in the supplemental material). Upon treatment of J7-1 with the DNA-damaging agent mitomycin C (MMC), SNJ1 is triggered to undergo the lytic life cycle, and up to 10¹⁰ PFU ml⁻¹ of SNJ1 virions are produced. Released SNJ1 viruses can infect a strain cured from SNJ1, named Natrinema sp. CJ7 (10), and forms turbid plaques on lawns of CJ7. Here, we investigated how SNJ1 controls its life cycle and identified its ORF4 as a critical regulator.

RESULTS

Discovery of clear plaque mutants of SNJ1.

Our lab previously used SNJ1’s proviral genome, plasmid pHH205, to construct a series of E. coli-Natrinema shuttle vectors (29). One of these vectors, pYC-S, was obtained by insertion of the E. coli vector pUC19-mev at the SacI site of the SNJ1 genome (Fig. S2). Since pYC-S contains the whole genome of SNJ1, we tested whether infectious viruses could be produced from a SNJ1-cured strain carrying pYC-S (CJ7/pYC-S) upon MMC treatment using a double-layer plaque assay. As shown in Fig. 1a, many plaques formed on the lawn of CJ7 using supernatants from an MMC-treated CJ7/pYC-S culture, suggesting that SNJ1 progeny viruses were generated. However, unlike the homogenous turbid plaques formed by wild-type SNJ1 viruses, viruses produced using pYC-S formed two kinds of plaques, one was similar to the turbid plaques formed by wild-type SNJ1, while the other was much bigger and clearer (Fig. 1a). This suggested that viruses produced using pYC-S were a mixture of viruses. Moreover, viruses purified from the turbid or clear plaques maintained their plaque morphotypes on the lawns of CJ7 (Fig. 1a, bottom). In liquid culture of CJ7, the addition of the turbid-plaque-forming viruses slowed down the growth of CJ7 similarly to wild-type SNJ1, while clear-plaque-forming viruses completely blocked its growth (Fig. 1b). Consistent with this, we were unable to obtain stable lysogen of the clear-plaque SNJ1 mutant viruses, whereas lysogens of the turbid-plaque SNJ1 mutant viruses were obtained at the same frequency as wild-type SNJ1, suggesting that the clear-plaque SNJ1 mutant viruses could only undergo the lytic pathway. Since disruption of the lysis-lysogeny switch of temperate viruses frequently resulted in the appearance of clear plaques, we suspected that insertion of the pUC19-mev fragment at the SacI site of SNJ1 disrupted the lysis-lysogeny life cycle of SNJ1.

FIG 1.

Discovery of clear-plaque mutants of SNJ1. (a) Viruses obtained from CJ7/pYC-S culture formed both clear and turbid plaques on lawns of CJ7. A late-exponential-phase culture of CJ7 was infected with SNJ1 viruses or viruses generated using CJ7/pYC-S. SNJ1 formed turbid plaques, while viruses generated using pYC-S formed two kinds of plaques. Arrows represent turbid plaques, and triangles represent clear plaques. Viruses purified from the turbid or clear plaques were propagated and used to infect CJ7 as above, the morphotypes of the plaques were maintained for either virus. (b) Ability of different viruses to lyse cells. The same amount of viruses was added into CJ7 culture (OD₆₀₀ ≈ 0.1). Growth of the culture was monitored for 49 h postinfection by measuring OD₆₀₀. SNJ1-Ca/Cb and SNJ1-Ta/Tb represented two clear-plaque viruses and two turbid-plaque viruses isolated form lawns of CJ7 infected with viruses from CJ7/pYC-S culture, respectively. CJ7 without virus infection was used as a control. The experiment was repeated twice, and error bars indicate the standard deviations.

ORF4 is critical for the lysis-lysogeny switch of SNJ1.

To determine how insertion of foreign DNA disrupted the life cycle of SNJ1, we first used PCR to check the genomic DNA (gDNA) integrity of the mutant viruses generated from pYC-S. Interestingly, we found that a primer pair designed to amplify the pUC19-mev region of pYC-S always generated products smaller than its expected size using mutant SNJ1 viruses as the templates. This suggested that parts of the pUC19-mev fragment were lost in the SNJ1 mutant viruses. To confirm this, the pUC19-mev region and its flanking sequences from 20 clear or turbid-plaque viruses were amplified by PCR (HJ-test-F/R primers listed in Table S2) and sequenced. As shown in Fig. 2, all of the sequenced viral gDNA had various degrees of deletion of the pUC19-mev fragment; some also contained deletions in the flanking sequences from SNJ1. The total length of pYC-S (21,425 bp) was 30% longer than the length of wild-type SNJ1 genome (16,492 bp, Fig. S2), and it had been reported that phage packaging systems never pack exceed about 10% over the parental gDNA (30, 31). Hence, it was unsurprising that all the gDNA of mutant viruses generated from pYC-S contained deletions. Examination of the lengths of the mutant virus genomes showed that they ranged from 16,010 to 18,054 bp (97 to 109% of the genome length), suggesting that the capsid of SNJ1 could tolerate a maximum of ∼1.5 kb of foreign DNA.

FIG 2.

Schematic diagrams of the genomic deletion of 20 turbid and clear plaque viral genomes. (a) Scheme of SNJ1 genome from 1 to 1,134 bp and locations of genomic deletion of turbid/clear plaque viruses of SNJ1. The putative ORFs of SNJ1 were indicated by numbered arrows, while the end site of genomic deletion in turbid/clear plaque viruses were marked by T1-T20 and C1-C20. The red dotted arrow represents the region of genomic deletion. Detailed deletion locations are shown in panels b (turbid plaque viruses) and c (clear plaque viruses). The pUC19-mev fragment and flanking sequences of 20 turbid and clear-plaque viruses were amplified by the HJ-F/R primer pair and sequenced. Black solid lines represent genomes packaged in viral particles, while black dotted lines represent deletions. The putative ORFs were indicated by numbered arrows and the SacI site, where foreigner DNA was inserted, was indicated. The numbers on the right side represent the start (x) and end (y) locations of the genomic deletion. “z” represents the genome size of mutant SNJ1 viruses, and the smaller genomes are indicated in boldface compared to the wild-type SNJ1 (16,492 bp).

Strikingly, the orf4 gene of SNJ1 (nucleotide sequence 499 to 293) was disrupted in all clear-plaque SNJ1 viral genomes, whereas it was intact in every turbid-plaque mutant virus (Fig. 2), suggesting that disruption of orf4 resulted in the appearance of clear plaque forming viruses. To test this, the plaque morphologies of ORF4-disrupted viruses were determined using viruses induced from CJ7/pYC-S-4M (start codon mutation of orf4, SNJ1^orf4 mut) and CJ7/pYC-SΔ1-575 (deletion mutation, SNJ1^Δorf4). As shown in Fig. 3a, SNJ1^orf4 mut and SNJ1^Δorf4 viruses formed only clear plaques on lawns of CJ7, suggesting that ORF4 controlled the lysis-lysogeny switch of SNJ1. To further test this, we checked whether expression of ORF4 in trans would restore turbid plaque formation of the clear-plaque SNJ1 mutant viruses. For this test, we used a plasmid containing orf4 and its cotranscribed downstream gene orf3, as well as flanking sequences of this operon (sequence 1 to 656), to preserve its promoter and elements critical for its expression. As shown in Fig. 3b, plaque formation of the SNJ1^orf4 mut and SNJ1^Δorf4 viruses was significantly inhibited on lawns of CJ7 expressing orf4 (with plasmid pFJ6-1-656), although some turbid plaques could form at the 10⁰ to 10⁻³ dilutions. On lawns of CJ7 containing the empty vector, SNJ1^orf4 mut and SNJ1^Δorf4 viruses formed clear plaques up to the 10⁻⁷ dilutions. Notably, the complementation of ORF4 in CJ7 also blocked the plaque formation of SNJ1 completely. These results demonstrated that ORF4 was critical for the lysis-lysogeny switch of SNJ1 and suggested that it also mediated superinfection immunity of the temperate virus (see below).

FIG 3.

Identification of orf4 as a critical factor for SNJ1 lysis-lysogeny switch. (a) orf4 disrupted viruses formed only clear plaques on lawns of CJ7. A late-exponential-phase culture of CJ7 was infected with SNJ1^{orf4 mut} and SNJ1^Δorf4 viruses (generated from CJ7/pYC-S-4M and CJ7/pYC-SΔ1-575 cultures, respectively), and only clear plaques were observed. (b) Expression of ORF4 in trans restored turbid plaque formation of clear-plaque SNJ1 mutant viruses and inhibited plaque formation dramatically. Tenfold serial dilutions of SNJ1 viruses or its clear-plaque mutants were spotted onto the lawns of CJ7-F/pFJ6-MCS, or CJ7-F/pFJ6-1-656 (with orf4). Plates were incubated at 37°C for 48 h and photographed. The maximum dilution for observed plaques is highlighted by asterisk.

We previously found that inactivation of ORF4 affected the stability and copy number of SNJ1 based E. coli-Natrinema shuttle vectors (29), suggesting that it was likely also critical for the maintenance of the SNJ1 lysogenic state. Consistent with previous observations, we found that in the presence of ORF4, the shuttle vector pYC-SHS, which contained the replication and regulatory elements of SNJ1, was stably maintained at about two copies per chromosome. However, in the absence of ORF4 (pYC-SHS-4M), the copy number of this plasmid increased to 20 copies per chromosome and was extremely unstable (Fig. 4). Expression of ORF4 in trans (pFJ6-1-656) restored stability and decreased the copy number of pYC-SHS-4M, confirming the role of ORF4 in plasmid stability and copy number control. More strikingly, we found that the copy number of pYC-SHS (with orf4) increased from 2 copies to 25 copies per chromosome upon MMC treatment, whereas the copy number of pYC-SHS-4M (without orf4) was only slightly affected by MMC (35 copies comparing to 20 copies per chromosome). Since pYC-SHS contained the gene clusters required for SNJ1 replication and regulation, the mechanism responsible for the induction of SNJ1 during MMC treatment of J7-1 was likely the same. Thus, these results suggested that ORF4 acted as a repressor of replication of SNJ1 to maintain it in the lysogenic state. Since the lack of ORF4 and MMC induction both lead to activation of SNJ1 lytic cycle, we speculated that ORF4 may be inactivated upon DNA damage, resulting in the active replication of SNJ1.

FIG 4.

ORF4 is critical for stability and copy number control of SNJ1-based plasmids. (a) Relative copy numbers of SNJ1-based plasmids pYC-SHS (with orf4) and pYC-SHS-4M (without orf4) in CJ7 cells with or without MMC treatment. “WT” represents wild-type pYC-SHS plasmid, “Δorf4” represents pYC-SHS-4M (start codon mutation of orf4), and “Δorf4+orf4” represents pFJ6-1-656 complemented pYC-SHS-4M. These plasmids were transformed to CJ7 strain and cultured to late exponential phase (24 h) in 18% MGM+Mev and treated with MMC (1 μg ml⁻¹) for 30 min, while cultures not treated with MMC were set as controls. Cells were collected by centrifugation and washed twice in the same volume of 18% MGM to remove MMC. Cell pellets were resuspended in 18% MGM+Mev and cultured for 24 h. Samples were taken for qPCR analysis using the primer pairs vector-F/vector-R, orf14-F/orf14-R, and radA-F/radA-R to determine the plasmid copy numbers relative to the chromosome. Three independent experiments were performed, and the error bars indicate the standard deviations. (b) Plasmid stability of pYC-SHS and pYC-SHS-4M during passage. Portions (100 μl) of stationary-phase cultures of CJ7/pYC-SHS, CJ7/pYC-SHS-4M, and CJ7/pYC-SHS-4M+pFJ6-1-656 were inoculated into 5 ml of Halo-2 medium every day. Samples were taken and measured by qPCR using the primer pairs vector-F/vector-R, orf14-F/orf14-R, and radA-F/radA-R for 5 days. Three independent experiments were performed, and error bars indicate the standard deviations.

ORF4 mediates superinfection immunity of SNJ1.

The finding discussed above (Fig. 3b) that the presence of ORF4 in CJ7 cells inhibited plaque formation of SNJ1 viruses suggested that SNJ1 had immunity against a second infection by itself and that ORF4 played a critical role. To confirm this, we infected CJ7 or strain J7-1 with SNJ1 viruses. As shown in Fig. 5a, while SNJ1 formed plaques on CJ7, J7-1 was completely immune to the infection of SNJ1, confirming that SNJ1 establishes superinfection immunity. Moreover, the presence of a plasmid carrying the orf3-4 operon in CJ7 completely blocked infection by SNJ1 and inactivation of ORF4, but not ORF3, by a frameshift mutation abolished the ability to block SNJ1 infection. This confirmed ORF4 is the critical regulator of superinfection immunity.

FIG 5.

ORF4 mediates superinfection immunity of SNJ1. (a) Expression of ORF4 confers resistance to SNJ1 infection in CJ7. To test the immunity to against SNJ1, 10-fold serial dilutions of SNJ1 virus stocks were spotted onto lawns of CJ7, J7-1, and CJ7-F strains carrying pFJ6 plasmids with or without orf3 and orf4. The maximum dilution for observed plaques was highlighted by asterisk. (b) Sequence alignment of ORF4 with other MazE/SpoVT family members. C68 protein from hybrid virus-plasmid pSSVx (residues 1 to 68, PDB code 3O27) (32), the N-terminal domains of MazE (residues 1 to 53, PDB code 1UB4) (33), AbrB (residues 1 to 51, PDB code 1YSF) (34), and SpoVT (residues 1 to 55, PDB code 2W1T) (35) were aligned with ORF4 using Clustal Omega. The secondary structure of ORF4 was predicted by PSIPRED in MPI Bioinformatics Toolkit. ORF4 contains five β-strands (light-gray arrows) and one α-helix (dark-gray cylinder). The same color is used for other proteins. (c) The N-terminal 33 amino acids of SNJ1 were necessary and sufficient for immunity against SNJ1. Tenfold serial dilutions of SNJ1 were spotted onto lawns of CJ7-F harboring plasmid pFJ6-MCS or its derivatives carrying different portions of ORF4; superscript denotes the amino acid residues of ORF4. Hpro stands for the promoter of heat shock protein 70 from Haloferax volcanii DS52. The plates were incubated at 37°C for 48 h and photographed.

orf4 was predicted to encode a 7.7-kDa protein possessing 68 residues. Western blot of a green fluorescent protein (GFP) fusion of ORF4 confirmed that it encoded a protein, although the fusion protein was not functional in superinfection immunity (Fig. S3). Bioinformatic analysis of ORF4 revealed that it belonged to the MazE antitoxin superfamily and showed homology with the Bacillus subtilis sporulation regulatory protein SpoVT and the C68 protein of the hybrid virus-plasmid pSSVx (Fig. 5b). These proteins usually form dimers through a swapped hairpin domain and act as transcriptional regulators, which is a distinctive structural characteristic of this family (32–35). Protein secondary-structure prediction suggested that ORF4 contained 5 β-strands and one α-helix. To determine which part of ORF4 was important for superinfection immunity, it was truncated and expressed under the promoter of heat shock protein 70 from Haloferax volcanii D52 in plasmid pFJ6. Remarkably, a fragment containing only the N-terminal 33 amino acids was sufficient to block SNJ1 infection (Fig. 5c). Further truncation of the 33 amino acids abolished its ability to restrict SNJ1 infection, demonstrating that the N-terminal 33 amino acids of ORF4 were necessary and sufficient for superinfection immunity.

ORF4 represses DNA replication of SNJ1.

Generally, there are five stages in a viral life cycle: adsorption, gDNA ejection, macromolecular synthesis, packaging, and release. ORF4 may inhibit any step of the SNJ1 viral life cycle to prevent its replication. We first checked whether adsorption of SNJ1 to the host cells was blocked by ORF4. To do this, SNJ1 was incubated with cells with or without SNJ1 (CJ7 or J7-1, respectively) or cells with or without ORF4 expression (CJ7-F/pFJ6-MCS or CJ7-F/pFJ6-Hpro-orf4) at a multiplicity of infection (MOI) of 0.1, 1, and 10 for 1 h. After incubation, the titer of the unabsorbed viruses was quantified by a double-layer plaque assay and compared to the titer of the viruses before the adsorption assay. The difference of the titers before and after the adsorption assay was considered the adsorption efficiency. As shown in Fig. 6a, there was no obvious difference in adsorption efficiency in the presence or in the absence of SNJ1, nor with or without ORF4 expression at a low or high MOI. Therefore, we concluded that ORF4 did not act by preventing viral adsorption.

FIG 6.

ORF4 blocks SNJ1 infection by inhibiting viral gDNA replication. (a) ORF4 did not affect SNJ1 adsorption to host cells CJ7. SNJ1 was incubated with early-exponential-phase cultures of CJ7, J7-1, CJ7-F/pFJ6-MCS (− orf4), and CJ7-F/pFJ6-Hpro-orf4 (+ orf4) at different MOIs (0.1, 1, and 10) for about 1 h at 45°C. After absorption, the culture was centrifuged, and the titers of unbound viruses in the supernatant were measured by the double-layer method. The adsorption efficiency was determined by comparing the titer before and after virus adsorption. Three independent experiments were performed, and error bars indicate the standard deviations. (b) Relative copy number of viral gDNA intercellular 1 h postinfection. Indicated strains were infected with SNJ1 as in panel a. After 1 h of adsorption, the cells were centrifuged, and the relative copy numbers of viral gDNA intercellular were measured by qPCR using the primer pairs orf14-F/orf14-R and radA-F/radA-R, which represented proviral genome and host chromosome, respectively. The bars represent the means and standard deviations of three independent experiments. Significance testing against CJ7-F/pFJ6-Hpro-orf4 was performed using a one-sample t test (***, P < 0.001; **, P < 0.01). (c) ORF4 repressed SNJ1 genome replication. SNJ1 was incubated with CJ7 or CJ7-F/pFJ6-MCS at an MOI of 0.1 and CJ7-F/pFJ6-Hpro-orf4 at an MOI of 0.5 for 1 h. The cells were collected and washed in fresh Halo-2 medium twice to eliminate free viruses. The cells were then cultivated for 8 h, and samples were taken every hour postinfection. qPCR analyses were performed using the same primer pairs as in panel b. The bars represent the means and standard deviations of three independent experiments. Significance testing against CJ7-F/pFJ6-Hpro-orf4 was performed using a one-sample t test (***, P < 0.001; **, P < 0.01; *, P < 0.05). (d) The genomic integrity of provirus was not affected by ORF4. Experiments were performed as in panel c expect that SNJ1 was incubated at an MOI of 5, with all strains and samples for Southern blot analysis taken every 2 h after 1 h of infection. The DNA samples were electrophoresed on agarose gels and transferred onto positively charged nylon membranes with alkaline transfer buffer. A specific probe that recognized nucleotides 1483 to 2537 of SNJ1 was used for Southern blot analysis. The DNA marker were shown on the left. Lanes: C, CJ7 strain; –, CJ7-F/pFJ6-MCS strain; +, CJ7-F/pFJ6-Hpro-orf4 strain.

Next, we tested whether viral gDNA ejection was inhibited by ORF4. The relative copy number of viral gDNA in the host cells with or without ORF4 expression were quantified 1 h postinfection by qPCR. The single-copy gene radA located on the chromosome was used as a reference, while orf14 of SNJ1 was used for viral gDNA. As shown in Fig. 6b, at an MOI of 0.1, copies of viral gDNA in cells without ORF4 (0.0065 in CJ7 and 0.0068 in CJ7/pFJ6-MCS) was about 3-fold higher than cells with ORF4 (0.0020 in CJ7-F/pFJ6-Hpro-orf4). The difference increased to about 6-fold at an MOI of 1 (0.0691 and 0.0601 versus 0.0106) and 9-fold at an MOI of 10 (0.8877 and 0.8100 versus 0.1080). These results suggested that ORF4 may prevent gDNA from ejecting into the host cells. However, we could not exclude the possibility that the ejected gDNA had been degraded in the presence of ORF4 given that this was examined 1 h postinfection. It was also possible that ejection of gDNA was not affected, but replication of the ejected gDNA was suppressed by ORF4 (see below).

To test whether replication of SNJ1 was repressed by ORF4, we quantified the copy number of gDNA of SNJ1 in the infected cells every hour postinfection for 9 h. For a better comparison, we infected cells without ORF4 (CJ7 and CJ7/pFJ6-MCS) at an MOI of 0.1, while cells with ORF4 (CJ7-F/pFJ6-Hpro-orf4) at an MOI of 0.5 to ensure similar gDNA copy number of SNJ1 at 1 h postinfection (0.0046, 0.0039, and 0.0038 copies per chromosome, respectively). As shown in Fig. 6c, in the absence of ORF4 (strain CJ7 and CJ7/pFJ6-MCS), the viral gDNA duplicated rapidly over time and reached up to 2.5280 or 2.0998 copies per chromosome 9 h postinfection. However, the viral gDNA was barely replicated in the strain CJ7-F/pFJ6-Hpro-orf4 (0.0582 copies per chromosome). These results showed that gDNA replication in the host cell was significantly repressed in the presence of ORF4.

To ensure that ORF4 repressed replication of the SNJ1 gDNA instead of degrading it, we repeated the experiment by Southern blot analysis using probes that specifically recognized SNJ1. Note that restriction digestion of the gDNA was not carried out in this assay to preserve the integrity of the sample and that, as a consequence, multiple bands of SNJ1 gDNA could show up as the DNA molecule could be supercoiled. Consistent with the above results, gDNA bands were detected in strain CJ7 and CJ7/pFJ6-MCS 1 h postinfection and accumulated extensively as time increased in the absence of ORF4 (Fig. 6d). However, only a weak signal of SNJ1 genome was observed in the ORF4-expressing strain CJ7-F/pFJ6-Hpro-orf4 even 9 h postinfection. More importantly, SNJ1 genome remained intact during the whole cultivation process, suggesting that ORF4 did not prevent replication by degrading viral DNA. In conclusion, these results suggested that ORF4 precluded SNJ1 infection through inhibition of viral gDNA replication.

ORF4 is highly conserved in SNJ1-like proviruses.

Because ORF4 played such an important role in the SNJ1 life cycle, its presence in a genome may indicate the existence of a SNJ1-like provirus. Thus, we searched for ORF4-like proteins using a protein-protein BLAST algorithm. We found that an ORF in Natrinema versiforme strain BOL5-4 plasmid pNVE19 and an ORF in Haloterrigena jeotgali strain A29 plasmid unnamed5 shared high homologies to ORF4 (94 and 83% identities, respectively). When the genomes of these two plasmids were aligned with the genome of SNJ1, we found that they were highly similar, indicating that they could be SNJ1-like proviruses (Fig. 7). First, their genomes were about the same size and encoded roughly similar number of ORFs. BOL5-4 plasmid pNVE19 was 18,925 bp and contained 28 putative genes (GenBank accession no. NZ_CP040333), while A29 plasmid unnamed5 was 17,189 bp and contained 31 putative genes (GenBank accession no. CP031302) (36). Second, most of the ORFs on these two plasmids were homologous to ORFs of SNJ1 and arranged into clusters similarly as SNJ1. Most importantly, genes homologous to the ORFs important for SNJ1 replication shared an identity of >40%, and homologs of the core structural elements for Sphaerolipoviridae were also present in these two plasmids (37). These included homologs of the two major capsid proteins and the putative packaging ATPase of SNJ1 (indicated in blue and pink, respectively, in Fig. 7). Based on these features, these two plasmids were likely SNJ1-like proviruses and ORF4-like protein may play a critical role in the regulation of their life cycles.

FIG 7.

Alignment of SNJ1 with two SNJ1-like plasmids. Genomic alignment of SNJ1, Natrinema versiforme strain BOL5-4 plasmid pNVE19 (GenBank accession no. NZ_CP040333, starting from 8,613 bp) and Haloterrigena jeotgali strain A29 plasmid unnamed5 (GenBank accession no. CP031302, starting from 9,455 bp). ORFs of SNJ1 are noted in the arrows. Proteins predicted to be regulators or related to DNA replication are indicated by red or yellow arrows. Capsid proteins defined by mass spectrometric analysis previously (9) are colored orange. Characteristic conserved proteins in family Sphaerolipoviridae, including the packaging ATPase, the small major capsid protein, and the large major capsid protein, are marked in pink, light blue, and dark blue (bottom legend). Homologous ORFs in the other two plasmids are shown in the same color, and the percentages of protein identities were shown.

DISCUSSION

Overall, our results showed that ORF4 is a master regulator of the life cycle of SNJ1 (Fig. 8). It not only regulates the lysis-lysogeny switch of SNJ1 but also mediates superinfection immunity, two phenomena which are presumably widely present but have rarely been studied for archaeal viruses. Sequence analysis of ORF4 showed that it contains the SpoVT/AbrB-like domain and shares homology with proteins belonging to the antitoxin MazE superfamily. Proteins of the SpoVT/AbrB family normally act as transcriptional regulators (38–40). MazE is the antidote to the toxin MazF, and MazE-MazF in E. coli is a regulated prokaryotic chromosomal addiction module. MazE is a labile protein that is degraded by ClpAP serine protease (41, 42). We suspect that ORF4 works as a transcriptional repressor of the genes responsible for the lytic pathway.

FIG 8.

Schematic view of the life cycle of SNJ1. (i) SNJ1 proviral genome resides in J7-1 cytoplasm as plasmid pHH205 with relative copy numbers 1 to 3. ORF4 represses the expression of the lytic pathway genes, presumably by binding to the viral DNA as a dimer, thus maintains SNJ1 in the lysogenic state. (ii to v) Upon MMC treatment, ORF4 was inactivated by an unknown mechanism, resulting in expression of the lytic pathway genes and replication of viral genome by the RepA protein. Host cells are lysed, and assembled progeny viruses are released into supernatant. (vi to x) Released SNJ1 viruses infect CJ7 either by entering the lysogenic state as a plasmid or by replicating actively in a process controlled by ORF4. (xi) A lysogen of SNJ1 (J7-1) was immune to superinfection of SNJ1 because ORF4 represses replication of the ejected gDNA.

The expression level of ORF4 in the host cell appears to be critical for determining the route of the SNJ1 infection. In the absence of ORF4, SNJ1 formed clear plaques, whereas in its presence, SNJ1 formed turbid plaques, and with its ectopic expression, SNJ1 infection was blocked. This suggests that the lysis-lysogeny switch is controlled by a ratio of ORF4 with a yet-to-be-discovered SNJ1-encoded protein(s). The observations that MMC treatment and disruption of ORF4 both resulted in the active replication of SNJ1 and increased the copy number of SNJ1-based plasmids indicate that DNA damage in the host cell may lead to the inactivation of ORF4. Presumably, ORF4 is cleaved and/or degraded by host proteases in a manner similar to the cleavage/degradation of the cI repressor of λ phage (23). It is also possible that ORF4 dissociates from its target sites like the F55 repressor of SSV1 (43). Future studies to test the ability of ORF4 to bind DNA and regulate gene expression, as well as its inactivation upon DNA damage in the host cell, will undoubtedly provide critical missing information on the mechanism of the lysis-lysogeny switch of SNJ1.

Superinfection immunity has been well documented for many bacteriophages and eukaryotic viruses (44–46). However, only a couple of cases have been reported for archaeal viruses. It was found that the adsorption of Sulfolobus islandicus rod-shaped virus 2 (SIRV2) to its host cells was significantly reduced by a previous infection, indicating that SIRV2 established superinfection exclusion (47). However, the exact mechanism underlying this phenomenon has not yet been elucidated. In addition, hyperthermophilic archaeal viruses SPV1 and SPV2 carry mini-CRISPR arrays containing spacers against each other and, consistently, seem to restrict each other’s replication. Thus, it has been suggested that virus-borne mini-CRISPR arrays might represents a distinct mechanism of heterotypic superinfection exclusion (48). Here, we present evidences that lysogenic SNJ1 confers immunity against subsequent infections and identify ORF4 as a critical factor in this process by repressing SNJ1 genome replication. However, we cannot exclude the possibility that it may also block gDNA ejection into the host cells because our results showed that at early stage of infection (1 h postinfection) gDNA in cells without ORF4 were remarkably higher than cells with ORF4. Strikingly, we found that only the N-terminal 33 amino acids of ORF4 are necessary and sufficient for blocking SNJ1 infection. It is likely that the N-terminal 33 amino acids of ORF4 is sufficient to bind DNA and repress transcription of the lytic pathway genes. However, biochemical characterization of ORF4 is necessary to confirm this and reveal its mechanism of action.

In conclusion, we showed that the temperate haloarchaeal virus SNJ1 employs a lysis-lysogeny switch and establishes superinfection immunity. We also identified ORF4 as a key factor for both of these processes and found that it acts as a repressor of SNJ1 replication. We suggest that SNJ1 could be used as a model system to understand the regulation of archaeal viral life cycle and virus-host interactions. Using the SNJ1 mutant viruses and E. coli-Natrinema shuttle vectors generated in this study, many other viral proteins of SNJ1 and host factors necessary for the regulation of SNJ1 life cycle could be discovered. Characterization and elucidation of how these proteins and ORF4 work should greatly increase our understanding of the enigmatic archaeal virosphere.

MATERIALS AND METHODS

Strains, culture conditions, and transformation methods.

All strains, plasmids, and primers used in this study were listed in Tables S1 to S3 in the supplemental material. Different derivatives of Natrinema sp. J7 were cultured in Halo-2 medium at 45°C as described previously (10). Halo-2 medium contained 250 g of NaCl, 30 g of MgCl₂·6H₂O, 2.5 g of lactalbumin hydrolysate (Difco Laboratories), and 2 g of Bacto yeast extract (Difco Laboratories) per liter of water. Casamino Acids medium (Hv-Ca) or 18% modified growth medium (MGM) were prepared as described earlier (29, 49). Hv-Ca contained 144 g of NaCl, 18 g of MgCl₂·6H₂O, 21 g of MgSO₄·7H₂O, 4.2 g of KCl, 5 g of Amicase (Sigma), 0.5 g of CaCl₂, and 30 ml of 1 M Tris-HCl (pH 7.5) per liter of water. 18% MGM contained 144 g of NaCl, 18 g of MgCl₂·6H₂O, 21 g of MgSO₄·7H₂O, 4.2 g of KCl, 5 g of peptone (Difco Laboratories), 3 g of Bacto yeast extract (Difco Laboratories), 0.5 g of CaCl₂, and 30 ml of 1 M Tris-HCl (pH 7.5) per liter of water.

Transformation of Natrinema sp. CJ7 and CJ7-F was performed using the modified polyethylene glycol method as described previously (50, 51). CJ7 transformed with pYC vectors were plated on 18% MGM plates with 5 μg ml⁻¹ mevinolin (Mev), while CJ7-F transformed with pFJ6 vectors were selected on Hv-Ca medium. E. coli were cultured in Luria-Bertani medium at 37°C, supplemented with ampicillin (0.1 mg ml⁻¹) when necessary. E. coli were transformed according to the CaCl₂ method (52). Solid and soft agar media were prepared by adding 12 and 5 g liter⁻¹ agar, respectively.

Plasmid construction.

To construct pYC-S-4M, overlap extension PCR was performed by primer pairs pYC-S-AflII-F/4M-R and 4M-F/pYC-S-NcoI-R using pYC-S as a template. PCR products were ligated into AflII-NcoI digested pYC-S by Hieff Clone Plus One-Step cloning kit (Yeasen, China). To construct pYC-SΔ1-575, the SacI-NcoI fragment on the pYC-S was replaced with the SacI-NcoI fragment on the pYC-1 by T4 DNA Ligase (TaKaRa). To construct E. coli-Natrinema sp. J7 shuttle vectors containing full-length SNJ1, SNJ1 proviral genomes pHH205 was extracted from strain J7-1 using TIANpure Mini plasmid kit (Tiangen). Purified pHH205 was digested with FastDigest SacI restriction enzyme (Thermo Fisher Scientific) and ligated with SacI-digested pUC-mev.

Derivatives of pFJ6 plasmids were all constructed by the same approach. For pFJ6-1-656, a DNA fragment containing nucleotides 1 to 656 of SNJ1 viral genome was amplified with the primer pair SphI-1-F/MunI-656-ORF4-R from J7-1. The product was digested with MunI-SphI and ligated into MunI-SphI-digested pFJ6-MCS by T4 DNA Ligase (TaKaRa). Frameshift mutations were introduced by overlap extension PCR using the primer pairs SphI-1-F/(ORF3-FSM-R, ORF4 FSM-R) and (ORF3-FSM-F, ORF4 FSM-F)/MunI-656-ORF4-R. For pFJ6-Hpro, target genes were amplified by using the primer pair MunI-Hpro-F/NotI-Hpro-R from pYCJ-H, and the product was digested with MunI-NotI and ligated into pFJ6-MCS. For pFJ6-Hpro-orf4, pFJ6-Hpro-orf4-4FSM, and pFJ6-Hpro-orf4^x, target genes were amplified using NotI-ORF4-499-F/(SphI-ORF4-293-R, SphI-ORF4-150-R, SphI-ORF4-99-R, SphI-ORF4-72-R, SphI-ORF4-48-R, SphI-ORF4-24-R) and (NotI-ORF4-FSM-F, NotI-ORF4-400-F)/SphI-ORF4-293-R, in succession, from J7-1, and then the products were digested with NotI-SphI and ligated into pFJ6-Hpro.

For pFJ6-PphaRP-GFP, PCR was performed by the primer pair ORF4-GFP-1-F/ORF4-GFP-3-R using plasmid pRF as a template (53). PCR products were ligated into AflII-SphI-digested pFJ6-MCS by using a Hieff Clone Plus One-Step cloning kit. For pFJ6-PphaRP-ORF4-GFP-1 and pFJ6-PphaRP-ORF4-GFP-2, target genes were amplified by the primer pairs ORF4-GFP-1-F/ORF4-GFP-1-R, ORF4-GFP-2-F/(ORF4-GFP-2-R, ORF4-GFP-4-R), and (ORF4-GFP-3-F, ORF4-GFP-4-F)/ORF4-GFP-3-R, in succession, and the PCR products were ligated into AflII-SphI-digested pFJ6-MCS as well.

All of the plasmids were transformed into E. coli DH5α and JM110 successively to get demethylated plasmids. Finally, the plasmids were transformed into Natrinema sp. CJ7 or CJ7-F accordingly.

SNJ1 induction, propagation, and infection procedures.

SNJ1 was induced from J7-1 strain by MMC (Roche). J7-1 was cultivated in Halo-2 medium at 45°C for about 24 h to late exponential phase (optical density at 600 nm [OD₆₀₀] ≈ 0.6) and treated with MMC (1 μg ml⁻¹) for 30 min at 37°C with aeration. Cells were collected by centrifugation (10,000 rpm, 5 min) and resuspended in the same volume of Halo-2 medium. After 24 h of cultivation at 45°C and 200 rpm, the culture was centrifuged at 10,000 rpm for 20 min, and the supernatant was collected and passed through a filter (0.22 μm) to remove cell debris. The titer of the virus stock was measured by double-layer plaque assay as follows.

SNJ1 titers were calculated as PFU ml⁻¹ as described previously (54), with minor modifications. Then, 100 μl of SNJ1 virus stock dilutions and 400 μl of CJ7 culture (late exponential phase, OD₆₀₀ ≈ 0.6) were mixed with 4 ml of soft melted Halo-2 medium (top layer) and then poured onto Halo-2 solid plates (bottom layer) immediately. The plates were incubated at 37°C for 48 h and counted. If necessary, CJ7 cells at early-exponential-growth phase (OD₆₀₀ ≈ 0.3) were collected and infected with SNJ1 (MOI = 10) to promote virus titer.

The mutant virus stocks were obtained from CJ7/pYC-S, CJ7/pYC-S-4M, and CJ7/pYC-SΔ1-575 by using MMC treated as described above. A late-exponential-phase culture of CJ7 was infected with mutant viruses by double-layer plaque assay; the plates were then incubated at 37°C for 48 h and photographed.

Analysis of the lytic ability of different viruses to CJ7.

CJ7 was cultured in Halo-2 medium at 45°C to early-exponential growth phase (OD₆₀₀ ≈ 0.2). The same amount of viruses (SNJ1, two turbid and clear-plaque viruses from CJ7/pYC-S culture) was added into CJ7 culture and incubated at 45°C for an hour. Cells were collected by centrifugation (10,000 rpm, 5 min) and resuspended in the same volume of Halo-2 medium. The cultures were then cultivated at 45°C and 200 rpm, and the growth curve was monitored for 49 h postinfection by measuring the OD₆₀₀. CJ7 without virus infection was set as a control.

Superinfection immunity assay.

Superinfection immunity assay was performed by the double-layer plaque assay with minor modifications. First, 400 μl of late-exponential-phase cultures of the indicated strains were mixed with soft melted Halo-2 agar (4 ml) and poured onto Halo-2 agar plates immediately. After 10 min, when the soft agar was solidified, 2 μl of the 10-fold serial dilutions of SNJ1 virus was spotted onto the plates carefully. The plates were incubated at 37°C for 48 h and photographed.

Western blot assay.

CJ7/pFJ6-PphaRP-GFP and CJ7/pFJ6-PphaRP-ORF4-GFP were cultivated in Halo-2 medium to late exponential phase, and crude extracts were taken and loaded for SDS-PAGE. Anti-GFP antibodies (TransGen Biotech) were used for Western blot assay of GFP and ORF4-GFP fusion protein.

Adsorption assay.

SNJ1 was incubated with 20 ml of early-exponential-phase (OD₆₀₀ ≈ 0.3) cultures of the indicated strains at different MOIs (0.1, 1, and 10) for about 1 h at 45°C. After absorption, each culture was centrifuged (10,000 rpm, 5 min), the supernatant was collected, and the titers of unbound viruses in supernatant were measured by the double-layer method. The adsorption efficiency was determined by comparing the titer before and after virus adsorption. Three independent experiments were performed, and error bars indicated the standard deviations.

Quantification of viral gDNA.

The relative copy number of viral gDNA intercellular at 1 h postinfection was detected by the qPCR method using cells collected from an adsorption assay as the templates. The templates from different cultures were prepared as described previously (29, 55). Briefly, a 1-ml aliquot of the cell culture was harvested by centrifugation and washed twice with Halo-2 medium to eliminate the free viruses in the supernatant. The cells were resuspended in 1 ml of 18% (wt/vol) NaCl solution. To avoid the negative effect of a high salt concentration on qPCR, 10 μl of the cell suspension was added to 490 μl of distilled water, which also resulted in rapid cell lysis. A single-copy radA gene located on the chromosome of host strain was used as a reference, while orf14 of the SNJ1 gnome was used as target gene. The specific primers used for qPCR are listed in Table S3. For the reactions, 20-μl mixtures were prepared containing 5 μl of template, 10 μl of iTaq Universal SYBR green Supermix (Bio-Rad), 1 μl of primer pairs (10 μM), and 4 μl of distilled water. Amplification was performed according to the manufacturer’s instructions. Finally, the qPCR data were analyzed by the 2^−ΔCT method (56). Three independent experiments were performed, and error bars indicated the standard deviations.

SNJ1 replication assay.

SNJ1 was incubated with 100 ml of early-exponential-phase (OD₆₀₀ ≈ 0.3) cultures of CJ7 or CJ7-F/pFJ6-MCS at an MOI of 0.1 and CJ7-F/pFJ6-Hpro-orf4 at an MOI of 0.5 for 1 h at 45°C. After 1 h of absorption, the culture was centrifuged (10,000 rpm, 20 min) to end the infection procedure. The cell pellets were washed twice with fresh Halo-2 medium to remove the free viruses, resuspended in the same volume of Halo-2 medium, and incubated at 45°C for 9 h. Samples were taken every hour postinfection. qPCR analyses were performed using the same primer pairs and procedure as described above in “Quantification of viral gDNA.” The qPCR data were analyzed according to the 2^−ΔCT method (56).

Genomic integrity of SNJ1 by Southern blotting.

CJ7, CJ7-F/pFJ6-MCS, and CJ7-F/pFJ6-Hpro-orf4 cells were infected by SNJ1 as described in “SNJ1 replication assay” except that SNJ1 was incubated at an MOI of 5 with all strains. Multiple aliquots of 8 ml of culture were collected every 2 h after 1 h of infection. To remove the unbound viruses, cells were collected by centrifugation (10,000 rpm, 5 min) and washed twice by the same volume of Halo-2 medium. The total genomic DNA of infected cells was extracted according to the online protocol (http://www.haloarchaea.com/resources/halohandbook) with modifications. Briefly, cell pellets were resuspended gently in 2 ml of distilled water to completely lyse the cells. Next, 2 ml of phenol-chloroform solution was mixed the cell lysate to extract proteins (and probably some carbohydrate). Finally, the genomic DNA was precipitated by adding 2 volumes of ethanol and dissolved in 150 μl of Tris-EDTA buffer. Then, 5 μl of DNA preparations was electrophoresed on a 0.8% agarose gel. The DNA bands were then transferred onto positively charged nylon membranes with alkaline transfer buffer according to the online protocol (57). The primers ORF7-F and ORF11-R were designed to synthesize a probe corresponding to the region from 1,483 to 2,537 bp of the SNJ1 genome (Table S3). Southern blotting was performed as described previously (58).

The DNA probe preparation, hybridization, and detection were performed using DIG High Prime DNA labeling and detection starter kit I (Roche) according to the manufacturer’s instructions.

Determination of plasmid copy number and stability.

CJ7/pYC-SHS and CJ7/pYC-SHS-4M were cultured to late-exponential phase (24 h) in 18% MGM plus Mev medium (MGM+Mev) and treated with MMC (1 μg ml⁻¹) for 30 min, while cultures without MMC were set as controls. Cells were collected by centrifugation and washed twice in same volume of 18% MGM to remove MMC. Cell pellets were resuspended in 18% MGM+Mev and cultured for 24 h. Samples were taken for qPCR analysis by using the primer pairs vector-F/vector-R, orf14-F/orf14-R, and radA-F/radA-R to determine the relative plasmid copy number to chromosome. The qPCR data were analyzed by the 2^–ΔCT method (56). Three independent experiments were performed, and error bars indicate the standard deviations.

For plasmid stability assay of pYC-SHS and pYC-SHS-4M during passage: 100 μl of stationary-phase culture of CJ7/pYC-SHS and CJ7/pYC-SHS-4M in Halo-2 medium was inoculated into 5 ml of Halo-2 medium every day. Samples were taken and measured by qPCR using the primer pairs vector-F/vector-R, orf14-F/orf14-R, and radA-F/radA-R. Three independent experiments were performed, and error bars indicate the standard deviations. The qPCR data were analyzed by the 2^−ΔCT method (56).

Bioinformatic and statistical analysis.

Homologous proteins were searched by using BLASTP and SyntTax (59), and conserved domains were detected by CD-search at the National Center for Biotechnology Information (NCBI). Clustal Omega and Constraint-based Multiple Alignment Tool at NCBI (COBALT) was used for multiple protein alignments (https://www.ebi.ac.uk/Tools/msa/clustalo and https://www.ncbi.nlm.nih.gov/tools/cobalt/). The secondary structure of ORF4 was predicted by PSIPRED in the MPI Bioinformatics Toolkit (https://toolkit.tuebingen.mpg.de) (60, 61). The global alignment of two proteins were performed by EMBOSS Needle toll at EMBI-EBI (https://www.ebi.ac.uk/Tools/psa/emboss_needle/).