Tolerance of Sulfolobus SMV1 virus to the immunity of I-A and III-B CRISPR-Cas systems in Sulfolobus islandicus

ABSTRACT

Sulfolobus islandicus Rey15A encodes one Type I-A and two Type III-B systems, all of which are active in mediating nucleic acids interference. However, the effectiveness of each CRISPR system against virus infection was not tested in this archaeon. Here we constructed S. islandicus strains that constitutively express the antiviral immunity from either I-A, or III-B, or I-A plus III-B systems against SMV1 and tested the response of each host to SMV1 infection. We found that, although both CRISPR immunities showed a strong inhibition to viral DNA replication at an early stage of incubation, the host I-A CRISPR immunity gradually lost the control on virus proliferation, allowing accumulation of cellular viral DNA and release of a large number of viral particles. In contrast, the III-B CRISPR immunity showed a tight control on both viral DNA replication and virus particle formation. Furthermore, the SMV1 tolerance to the I-A CRISPR immunity did not result from the occurrence of escape mutations, suggesting the virus probably encodes an anti-CRISPR protein (Acr) to compromise the host I-A CRISPR immunity. Together, this suggests that the interplay between viral Acrs and CRISPR-Cas systems in thermophilic archaea could have shaped the stable virus-host relationship that is observed for many archaeal viruses.

1. Introduction

Clustered regularly interspaced short palindromic repeats (CRISPR) and the CRISPR associated (cas) genes constitute an inheritable adaptive immune system that defends against infection of viruses and plasmids in most archaea and many bacteria. Their expression yields two components: mature CRISPR RNAs (crRNAs) that are generated from CRISPR arrays and Cas proteins that are expressed from cas gene cassettes. These CRISPR components form ribonucleoprotein complexes that recognize invading genetic elements by sequence complementarity between crRNA and invading nucleic acids and specifically target the foreign genetic elements for destruction [1–4]. Currently, CRISPR-Cas systems are classified into six different types (Type I-VI) according to their cas gene content [5,6], among which Type I, III and IV systems rely on a ribonucleoprotein complex of multiple protein subunits to conduct nucleic acid interference (Class 1 CRISPR-Cas systems), whereas Type II, V and VI systems employ effector complexes of a single Cas protein to perform the same function (Class 2 systems) [1–4].

In the past decade, CRISPR research has yielded a rudimentary understanding of the interference mechanisms for the main types of CRISPR-Cas systems. For Type I, II and V systems, short DNA motifs called protospacer adjacent motif (PAM) have been identified on invading DNAs by sequence analyses [7]. Upon the PAM recognition, the sequence complementarity between crRNA and the target DNA then triggers the DNA interference activity from the CRISPR endonuclease for target DNA destruction [1–3]. Nucleic acid interference by Type III systems involves more complex mechanisms: (a) the systems exhibit dual DNA/RNA interference in vivo and their DNA interference is transcription-dependent, (b) their effector complexes show three distinct activities in vitro, i.e. backbone target RNA cleavage, RNA UA cleavage and target RNA-activated DNA cleavage activity [8–11]. More recently, two III-A effector complexes were shown to catalyze the synthesis of cyclic oligoadenylate (cOA), and the cOA second messenger couples III-A systems to Csm6, a Cas accessory protein containing a CRISPR-associated Rossman Fold (CARF) and a High Eukaryotes and Prokaryotes Nucleotide-binding (HEPN) domain, and it has been demonstrated that cOA binds to the CARF domain of Csm6 and allosterically regulate the RNase activity from its HEPN domain [12,13].

S. islandicus Rey15A is an extensively studied genetic model in the Archaeal domain for which very versatile genetic toolbox has been developed [14]. The archaeon contains a complete I-A system (including an interference module and an adaptation module) that has two CRISPR arrays of identical repeats and two Type III-B modules (denoted Cmr-α and Cmr-β) (Fig. S1) [15], all of which are active in nucleic acid interference as tested by invader plasmid assay [16–18]. The I-A interference module mediates DNA interference in a PAM-dependent fashion (I-A CRISPR immunity) [16,19,20]. Cmr-α was the first type III system that has been shown to mediate transcription-dependent DNA interference and dual DNA/RNA interference activity in vivo (III-B CRISPR immunity) [17,18]. Its effector complex has the backbone RNA cleavage and target RNA-activated DNA cleavage activities [21,22], in analogy to the III-B systems of Pyrococcus furiosus and Thermotoga maritima [23,24]. This archaeon is also a host for several viruses, such as Sulfolobus monocaudavirus1 (SMV1), Sulfolobus tengchongensis spindle-shaped virus 2, and the fusellovirus SSV2 [25–29]. Interestingly, SMV1 virus is tolerant to active CRISPR-Cas systems of S. islandicus, despite these presence of multiple CRISPR-Cas systems in the same microbial host and a matching spacer in the host CRISPR arrays [25,29]. Furthermore, the virus is capable of escaping the active spacer acquisition by CRISPR-Cas in the host [25].

We were interested in investigation of the interactions between SMV1 and the CRISPR-Cas systems of S. islandicus Rey15A. Archaeal strains carrying plasmid-borne mini-CRISPR arrays targeting SMV1 genome were constructed and used as hosts for SMV1 infection. Both infected and uninfected cultures were investigated for the interplay between host cells and the SMV1 virus. We show that type III-B systems provide robust immunity against SMV1, while the type I-A CRISPR immunity allows viral proliferation, and the virus tolerance does not result from any escape mutation to the I-A CRISPR immunity.

2. Results

2.1. Design and overview of experiments

S. islandicus Rey15A contains one I-A and two III-B CRISPR-Cas systems, all of which confer CRISPR interference against invading nucleic acids [16–18]. Nevertheless, their interference activities have not been tested by virus challenge in this archaeon. Previous investigation of spacer acquisition in S. islandicus with SMV1, a Sulfolobus virus that was isolated from Yellowstone National Park, showed that the virus can infect this organism but the infection does not trigger spacer acquisition. [25] Here we exploited the incapability of spacer acquisition to design experiments to evaluate the effectiveness of immunities of I-A and/or III-B CRISPR systems on virus infection and the arms race between the virus and the antiviral systems during virus-host interaction.

The rationale of the experimental design was as the following: (a) S. islandicus strains were constructed for constitutively expressing one or more CRISPR-Cas effector complexes against SMV1 virus, i.e. the strains should be pre-immune to SMV1 infection, and (b) the sensitivity/tolerance of the virus to the host CRISPR-Cas immunity was then to be investigated by following the infected cultures.

Strain construction was based on the finding that crRNAs generated from spacers on plasmid-borne mini-CRISPR arrays direct either I-A or III-B CRISPR immunity or both activities against target nucleic acids in S. islandicus [30]. Different types of CRISPR immunity were designed by careful selection of spacers based on the following principles: (a) The I-A CRISPR immunity works on both strands and is stringently dependent on the ‘CCN’ or ‘TCN’ PAM, and (b) the III-B CRISPR immunity only targets the template strand due to the nature of transcription-dependent targeting, and the target site selection relies on the mismatches between the target site and the 5ʹ-handle of the crRNA.

We designed three spacers (Sgp11-1, Sgp11-2 and Sgp11-3) with the sequence of SMV1 gp11, a gene that codes a coat protein of the virus (Fig. S2). As shown in Fig. 1, Sgp11-1 has a ‘CCA’ motif on the corresponding protospacer that is located on the coding strand of the target gene, therefore, it is only to be targeted by the Type I-A CRISPR immunity; the protospacer of Sgp11-2 is on the template strand and has a ‘TTT’ motif that does not license I-A CRISPR immunity, meaning that the second spacer only guides the III-B DNA interference; the target of Sgp11-3 is also on the template strand, and it has a ‘CCT’ motif, and therefore, the third spacer mediates both I-A and III-B immunities in S. islandicus.

Figure 1.

Overview of the experimental design. (A) Design of three spacers targeting the gp11 gene of SMV1 (Sgp11-1, Sgp11-2 and Sgp11-3). The sequences of the spacers (blue letters) were selected from the gp11 gene, with Sgp11-1 from the coding strand and the other two from the template strand (marked in red boxes). The 3-nt protospacer flanking sequences (PFS) are highlighted in red. (B) The crRNA expressed from the CRISPR array in the plasmid and the Type I-A and III-B Cas proteins form corresponding interference effectors. (C) The I-A effectors carrying crRNA expressed from Sgp11-1 and Sgp11-3 are able to confer immunity due to the presence of a PAM (CCN) upstream of the protospacer while the III-B effectors carrying crRNA from Sgp11-2 and Sgp11-3 are functional because the crRNA can target the transcripts of the protospacer. The PFS, either from the DNA or the transcript, is shown in red letters, and the region of the spacer sequences of gp11 gene is shown in orange.

The sequences of the three spacers were used for construction of plasmid-borne artificial mini-CRISPR arrays and subsequently corresponding CRISPR plasmids. These plasmids, pSgp11-1, pSgp11-2 and pSgp11-3 were introduced into the S. islandicus E233S1, to yield archaeal strains that carried each of the CRISPR plasmids. These strains were suitable for testing the antiviral immunity of I-A and III-B systems against the SMV1 virus.

At the start of the experiment, all constructed strains were inoculated in SCV medium and grown to A₆₀₀ (absorbance at 600 nm) = 0.2. Then, these included cultures of the E233S1 containing pSeSD as a control or reference (Ref) and those containing either pSgp11-1, -2 or -3 that were programmed to be pre-immune to SMV1 infection due to the constitutive expression of either I-A or III-B CRISPR immunity or both immunities (denoted Pre-IA, Pre-IIIB and IA-IIIB). Each culture was infected with SMV1 at an MOI of 3, yielding four different SMV1-infected S. islandicus cultures (SMV1/Ref, SMV1/Pre-IA, SMV1/Pre-IIIB and SMV1/IA-IIIB). All infected and uninfected cultures were cultured for an extended period during which cell samples were then taken for the following analyses: (a) growth by determination of A₆₀₀, (b) relative viral DNA content in the cells, (c) plaque formation analysis of the production of viral particles, and (d) flow cytometry of DNA content distribution. By 46 hpi, the A₆₀₀ values of cultures were between 0.7 and 1.0, and they were then diluted to A₆₀₀ ≈ 0.1 and incubated further. More cell samples were taken during incubation and analyzed as described above.

2.2. The III-B CRISPR immunity effectively relieved growth retardation by SMV1

The growth data of the eight strains were summarized in Fig. 2, which could be divided into two stages. In the first stage (0–46 hpi), all infected strains grew in a similar rate as their corresponding uninfected strains except that a marginal difference was observed during 20–46 hpi, suggesting that SMV1 infection did not yield much growth retardation within a short incubation. In the second stage (46–140 hpi), growth retardation was observed at 92 hpi for all SMV1-infected cultures: The most severe growth inhibition was observed for the SMV1/Ref culture, which was followed by the SMV1/Pre-IA culture. Further, the infected strains that constitutively expressed the III-B interference activity (SMV1/Pre-IIIB, SMV1/IA-IIIB) only grew slightly slowly than the uninfected references (Fig. 2). Nevertheless, continuous incubation showed that, while the SMV1/Pre-IA strain grew as slowly as infected reference strain (SMV1/Ref), the remaining two infected archaeal strains grew faster but still grew more slowly than their uninfected reference strains (Fig. 2). These data indicated that, while the programmed III-B CRISPR immunity could moderately relieve the growth inhibition on S. islandicus by SMV1, the programmed I-A CRISPR immunity could not.

Figure 2.

Growth curves of uninfected and SMV1-infected S. islandicus cultures. The absorbance at 600 nm (A₆₀₀) was measured over a 140 h window. At 46 h, the cultures were diluted to A₆₀₀ ≈ 0.1 and further grown at the same conditions.

Error bar indicates the standard error (SD) of three repeats for each treatment.

2.3. Programmed I-A and III-B immunities showed differential effects on replication of viral genome and production of virus particles

Next, total DNAs were extracted from the cell samples and analyzed for the relative abundance of viral DNA in the infected cells during incubation. The amount of viral DNA was estimated by quantitative PCR of the viral gp11 gene in the total DNA samples whereas the amount of host chromosomal DNA was determined by analysis of the sul7 gene.

To simplify the data for comparison, the relative abundance value of gp11 and sul7 in SMV1/Ref cells at 30 hpi was arbitrarily set to 1, with which relative ratios of viral and genomic DNA were calculated for all other cell samples. At this time point, the amount of SMV1 DNA in SMV1/Ref cells was 50–500 fold higher than that in any of the infected cells harboring one or two programmed CRISPR immunity (Fig. 3A), indicating that each of the tested antiviral activities efficiently destroyed viral DNA at the early stage of virus infection. In addition, the viral DNA content in SMV1/Pre-IIIB cells was ca. 10-fole lower than that in SMV1/Pre-IA cells. Then, cellular viral DNA content showed certain variations at this time point in different strains: whereas the viral DNA topped at 92 hpi in SMV1/Ref, SMV1/Pre-IA and SMV1/IA-IIIB cells, with an increase of 15, 179 and 57-fold, respectively, the viral DNA level was maintained at a constantly low level in SMV1/Pre-IIIB (Fig. 3A). These data suggested that (a) the programmed I-A CRISPR immunity became less effective in inhibiting viral DNA replication during incubation, (b) the programmed III-B CRISPR immunity was constantly effective on the inhibition of viral DNA replication, and (c) the occurrence of both I-A CRISPR immunity and III-B CRISPR immunity compromised the efficiency of the antiviral activity in the infected cells. Interestingly, while the vial DNA in SMV1/IA-IIIB cells was moderately accumulated from 30 to 92 hpi, further incubation led to dramatic decrease of the viral DNA level, suggesting that a I-A CRISPR immunity and III-B CRISPR immunity could more effectively cooperate to efficiently degrade the viral DNA at a late stage of incubation.

Figure 3.

Effects of I-A, III-B or both I-A and III-B immunities on viral DNA replication and product of viral particles. (A) qPCR analysis of relative viral DNA abundance in different SMV1-infected cultures. Total DNA was extracted at indicated time points from the infected cells and qPCR was performed to estimate the abundance of gp11 gene with the abundance of a host gene, sul7, for normalization. The normalized value of the measurement at 30 hpi in the SMV1/Ref culture was set to 1. (B) Plaque formation unit (/µl) of SMV1 particles at 140 hpi in different SMV1-infected cultures.

Error bar indicates the standard error (SD) of three repeats for each treatment.

The effects of the programmed I-A and III-B immunities on virus propagation were further examined by determination of virus titers in their culture media (Fig. 3B). The S. islandicus ΔC1C2 was chosen as the indicator strain for plague assay since it is highly susceptible to SMV1 infection [29]. We found that, the plaque formation units (PFU/µl) was 4.8 × 10⁵ in the SMV1/Ref culture at 108 hpi, whereas the number was dropped by 16 folds in the SMV1/Pre-IA culture. Furthermore, virus titer in SMV1/III-B was further reduced to a level below 100 plaques/µl whereas virus particle produced in the SMV1/IA-IIIB culture was estimated to 7.5 × 10³ CFU/µl. These data are in agreement with the copy number of viral DNA determined for the corresponding host cells.

Previously we showed that invader plasmids are effectively targeted by the I-A CRISPR immunity in S. islandicus and only host cells carrying mutations that inactivate the I-A CRISPR immunity can circumvent the immunity [16]. To examine whether the high virus titer present in the SMV1/Pre-IA could be due to the inactivation of the I-A CRISPR immunity, we analyzed viral genomes in 26 individual plaques by PCR amplification of the target locus. Agarose gel electrophoresis of the PCR products did not detect any deletion or insertion and sequencing of the PCR products did not reveal any mutation in the PAM or protospacer sequences of the viral genome (Fig. S3). Together, our results indicated that only pre-expressed III-B CRISPR immunity effectively prevents SMV1 proliferation.

2.4. The I-A CRISPR immunity facilitated DNA-less cell formation in SMV1-infected cultures

Next, we analyzed DNA content distributions in the cells of uninfected and infected cultures over the 140 h window of cultivation as shown in Fig. 2. Uninfected cultures of all four stains contained a majority of cells containing 2 chromosomes (Fig. 4A), a typical DNA content distribution observed for Sulfolobus species [31]. This indicated that the artificial CRISPR plasmids did not yield any impact on the cell cycle of the archaeal host. The flow cytometry profile of the infected cultures changed during the experiment: in the early stage of cultivation (by 67 hpi), infected cells displayed DNA content distributions that were similar to their uninfected references; at 92 hpi, a population of cells containing a DNA content of >2 chromosomes appeared in the SMV1/Ref culture (Fig. 4A). Their DNA content further increased at 108 hpi, and kept more or less constant for the following 32 h. During the entire incubation period, minor DNA-less cells were observed (Fig. 4B, Fig. S4), as reported in previous observations that SMV1 infection did not induce cell lysis [25,29].

Figure 4.

Type I-A CRISPR immunity induced DNA-less cell formation in SMV1-infected cells. (A) DNA content distributions of uninfected (left panels) and SMV1-infected (right panels) different Sulfolobus cultures were analyzed by flow cytometry at indicated time points. Horizontal axis (DNA content) and vertical axis (cell count) are shown in liner scale. The DNA-less cells and cells containing one or two copies of genomic DNA are indicated with ‘L’, ‘1’ and ‘2’, respectively. The cells containing DNA content more than 2 chromosomes are highlighted in red boxes. The data are the representatives of three independent experiments. (B) Quantification of DNA-less cells in the SMV1-infected Sulfolobus cultures.

Error bar indicates the standard error (SD) of three repeats for each treatment.

In the SMV1/Pre-IA culture, cells of DNA content >2 chromosomes also appeared at 92 hpi, which was followed by the formation of DNA-less cells for a relatively large cell population at 108 hpi. The population of DNA-less cells then increased gradually until 140 hpi, the last sampling time of the experiment (Fig. 4B, Fig. S4). The data suggested that type I-A CRISPR immunity facilitated chromosomal DNA degradation in SMV1-infected cells.

In the SMV1/Pre-IIIB and SMV1/IA-IIIB cultures, the cell population containing a higher DNA content was not detected during the incubation time (Fig. 4A), which was in agreement with the effective suppression of viral DNA replication by the III-B antiviral immunity. In addition, we quantified the subpopulation of DNA-less cells in the cell samples of SMV1/Pre-IIIB and SMV1/IA-IIIB and found that the latter, which carried the programmed I-A and III-B immunities, accumulated more DNA-less cells than the former where only the III-B CRISPR immunity is present (Fig. 4B, Fig. S4). Since several DNA damage agents induce DNA-less cell formation [32], a plausible explanation to this phenomenon is that the interaction between the I-A antiviral immunity and SMV1 virus in the host cells mimics DNA damage treatment when virus copy number is high, inducing DNA-less cell formation to the SMV1-infected cells.

3. Discussion

In this study, we have investigated the response of S. islandicus Rey15A to SMV1 infection. Three strains were tested, each of which contains a specific type of constitutively expressed CRISPR-Cas immunity against SMV1 virus, including the I-A CRISPR immunity, or the III-B CRISPR immunity, or both I-A and III-B immunities. We found that, the III-B CRISPR immunity strongly inhibits the proliferation of SMV1 in the entire experiment period, while the I-A CRISPR immunity only effectively inhibits the virus proliferation at the early stage of cultivation and a much elevated level of cellular viral DNA and high virus titers were observed after prolonged cultivation (Fig. 3). In fact, the wild-type SMV1 virus persists in the infected cultures during the entire cultivation period since no mutant viruses have been detected by the end of growth experiments. This reveals that the SMV1 virus has evolved a mechanism to tolerate the CRISPR immunity in S. islandicus.

The tolerance of viruses to CRISPR immunity has not been addressed in archaeal CRISPR research previously. In early experiments where CRISPR immunity has been demonstrated in archaea, CRISPR interference activity was tested in S. islandicus and Haloferax volcanii using the invader plasmid assay [16,20,33] and in Sulfolobus solfataricus by transformation of recombinant virus constructs [34]. Escape mutations have been reported in all these researches. In addition, invader plasmid assay has also revealed escape mutations for the III-B CRISPR immunity in different archaea [17,35]. Nevertheless, these genetic approaches do not allow the dissection of the interplay between the host CRISPR immunity and its targeting genetic elements because the experiments are based on colony formation of mutant cells on selective plates.

In contrast, virus challenge experiments reported in this work directly address the arms race between the host antiviral immunity and the targeting genetic element in an archaeon. To this end, we have made two interesting discoveries: (a) the III-B CRISPR immunity has the capability to greatly reduce the cellular virus content and virus particle production in S. islandicus. These results are consistent with the mechanisms of nucleic acid interference revealed from investigation of type III CRISPR-Cas systems, i.e. the system is to be activated by viral transcript and degrades both invading DNA and RNA [36–38]. (b) The persistence of the wild-type SMV1 virus in the SMV1/Pre-IA culture suggests that SMV1 may encode an anti-IA system and the following evidence supports the conclusion: If the I-A CRISPR immunity effectively prevents the proliferation of the targeting virus, the interference would provide a strong selection for escape mutant viruses such that mutant viruses would dominate the virus population at the late stage of the experiment, as reported for the type II immunity in Staphylococcus epidermidis and the type I-F immunity in Marinomonas mediterranea [38,39]. The persistence and fluctuation of virus titers during cultivation are indicative of a fierce competition between the I-A CRISPR immunity and the anti-CRISPR activity as demonstrated for the presence of anti-CRISPR proteins in Pseudomonas aurogenosa bacteriophages [40]. Nevertheless, further experiments are required to identify the SMV1-encoded anti-IA CRISPR protein as have shown for a number of very diverse anti-CRISPR genes present in different bacteriophages [41].

Another important feature of the archaeal host-SMV1 virus interaction is the formation of DNA-less cells, a unique programmed cell death mechanism that is induced by DNA damage in S. islandicus and involves a series of cellular events that leads to DNA-less cell formation [32]. We show that the pre-expressed I-A CRISPR immunity in S. islandicus facilitates the cell death process either when the I-A CRISPR immunity is present alone or when it co-exists with the III-B CRISPR immunity (Fig. 4B). Considering I-A CRISPR system mediates interference to dsDNA, we reason that, when virus DNA reaches a high level, the arms race between the I-A CRISPR immunity and the SMV1 anti-CRISPR system could produce a signal that mimics the signal produced from DNA damage stress in this organism. The signal would then trigger the programmed cell death to elimination of cells with an elevated number of SMV1 viruses and thereby reducing the production of virus particle in the population. Indeed, cell death induced by genetic elements such as the SIRV2 virus and pKEP9 conjugative plasmid has also been documented [42,43]. Possibly this represents an important measure to increase the ecological fitness of the archaeal organism.

Most recently, III-A CRISPR systems have been shown to synthesize cOAs as a second messenger to activate Csm6, a general RNase that has the potential to produce dormant cells upon virus infection [12,13]. Since the III-B systems function under the same principle and Csx1, the Csm6 homolog, is activated by tetraadenylates [44], the III-B CRISPR immunity should have the same capability to induce general degradation of both viral and host transcripts during SMV1 infection, and this is probably why the system efficiently inhibits SMV1 proliferation in this archaeon. Furthermore, Sulfolobus SSV9 virus is known to induce cell dormancy in S. islandicus RJW002 strain and the host CRISPR-Cas system is able to release the dormant cells by inhibiting viral proliferation [45]. Apparently, the interplay between the host CRISPR immunity and their targeting viruses involves complex interactions that shape the evolution of both hosts and viruses in order to avoid exposure of viruses to the harsh conditions of the host habitat [46–48]. To this end, the occurrence of CRISPR-Cas systems in archaea and anti-CRISPR systems in their viruses have probably played very important roles in the adaptation of these biological entities to extremely thermophilic environments as evidenced by the prevalence of CRISPR-Cas systems and viruses for these organisms.

4. Materials and methods

4.1. Sulfolobus strains and transformation

S. islandicus E233S1 was grown in SCVU medium (0.2% sucrose, 0.2% Casamino acids, 10 ml/l of vitamin mixture solution and 20 µg/ml uracil) at 78°C while the strains carrying a plasmid were grown in SCV medium (lacking uracil). Plasmid transformation was performed as previously described [49].

4.2. Construction of artificial mini-CRISPR plasmid (pAC)

Plasmids carrying an artificial CRISPR array were constructed using pSe-Rp as the vector previously described [18]. Briefly, two DNA oligonucleotides were designed for each spacer fragment (listed in Supplementary Table 1). DNA fragments of spacers were obtained by annealing of each pair of oligos, which were then inserted into Sap I-digested pSe-Rp vector to yield the artificial CRISPR plasmids (pAC). The expression of CRISPR array is under the control of a strong promoter derived from the araS gene coding for an arabinose-binding protein [50].

All the DNA oligoes to be used for DNA cloning were synthesized from Integrated DNA Technology (IDT, USA). Sequences of all pAC plasmids were verified by DNA sequencing at GATC Biotech (Germany).

4.3. Plague formation assay

Virus titers present in the media of infected cultures were determined by plaque formation assay at 140 hpi. Cell mass in each culture was removed by centrifugation at 8000 rpm for 10 min, giving supernatants containing virus particles. Then, the supernatants were serially diluted, and 10 μl of each diluted sample was mixed with 4 ml preheated fresh S. islandicus ΔC1C2 cells (A₆₀₀ = 0.2 ~ 0.3), a mutant strain that is highly susceptible to SMV1 [16,29]. The mixture was further mixed with 4 ml preheated 0.4% Gelrite (Roth, Germany) and layered onto a 0.7% solid Gelrite SCVU plate. Plaques formed after incubation for 3 days at 78°C. If applicable, the plaques were directly used as DNA template for the PCR analysis of gp11 gene with the primer set Test-gp11-F and Test-gp11-R (Supplementary Table 1) and the PCR product was sequenced by GATC Biotech (Germany).

4.4. SMV1 infection

SMV1 was prepared as described previously [29] and the titer was calculated by plaque formation assay. Before SMV1 infection, the cultures were grown at exponential phase for at least 72 h, and then early exponential phase cultures (the absorbance at a wavelength of 600 nm (A₆₀₀) is about 0.2) were supplemented with SMV1 at a MO1 ≈ 3. Aliquots without SMV1 were set as controls. At 46 h during the incubation, the cultures were diluted with fresh medium to A₆₀₀ ≈ 0.1 and further grown for about 100 h. During the incubation, samples were taken at indicated time points for analysis.

4.5. Flow cytometry

The flow cytometry analysis was performed as described previously [32]. Briefly, cells for flow cytometry were fixed with 70% ethanol, washed with the wash buffer (10 mM Tris-NaCl, pH 7.5 and 10 mM MgCl2), and stained with 140 μl staining solution (containing 100 μg/mL mithramycin A (Apollo Chemical, Tamworth, UK) and 40 μg/ml ethidium bromide (Sigma–Aldrich, St. Louis, USA). Flow cytometry was conducted with an Apogee Flow A-40 flow cytometer (Apogee Flow Systems, Hemel Hempstead, U.K.) with a 405-nm laser. A dataset of at least 60,000 cells was collected for each sample. Quantification of DNA-less cells was carried out in DNA content-side scattered light (SSC) cytograms with Apogee Flow Histogram as described previously [32].

4.6. Total DNA extraction and qPCR

Cells samples were collected from infected cultures at 30, 46, 67, 92, 108 and 132 hpi, respectively and used for total DNA extraction. Cell suspensions (in TE buffer) were treated with 0.2 mg/ml Protease K (Sigma-Aldrich), and the resulting genomic DNA was extracted with equal volume of phenol/chloroform, followed by chloroform extraction. Then, the DNA was precipitated with isopropanol and washed by 70% ethanol. Dried DNA samples were resuspended with water and diluted to 10 ng/µl for qPCR analysis.

qPCR was conducted with a CFX96 TouchTM real-time PCR detection system (Bio-Rad), using Maxima SYBR Green/ROX qPCR Master Mix (2×, Thermo Scientific), and the PCR condition was as following: denaturing at 95°C for 5 min, 40 cycles of 95°C 15 s, 52°C 15 s, 72°C 20 s. Relative amounts of viral DNAs compared with corresponding host were calculated using the comparative Ct method [51].

Funding Statement

This work was supported by the Natur og Univers, Det Frie Forskningsråd [DFF-4181-00274].

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Ling Deng for conducting some preliminary experiments at an early stage of this work and other members of the Danish Archaea Centre for their support and insightful discussions. Tong Guo is a recipient of a PhD studentship from the China Scholarship Council.

Supplementary material

Supplemental data for this article can be accessed here.