Nairovirus polymerase mutations associated with the establishment of persistent infection in human cells

ABSTRACT

Crimean-Congo hemorrhagic fever virus (CCHFV), a tick-borne virus of the Orthonairovirus genus, persistently infects tick cells. It has been reported to establish persistent infection in non-human primates, but virological analysis has not yet been performed in human cells. Here, we investigated whether and how nairoviruses persistently infect human cells using Hazara orthonairovirus (HAZV), a surrogate model for CCHFV. We established a human cell line that was persistently infected with HAZV. Surprisingly, virions of persistently infected HAZV (HAZVpi) were not observed in the culture supernatants. There were five mutations (mut1, mut2, mut3, mut4, and mut5) in L protein of HAZVpi. Mutations in L protein of HAZVpi contribute to non-detection of virion in the supernatants. Lmut4 was found to cause low viral growth rate, despite its high polymerase activity. The low growth rate was restored by Lmut2, Lmut3, and Lmut5. The polymerase activity of Lmut1 was extremely low, and recombinant HAZV carrying Lmut1 (rHAZV/Lmut1) was not released into the supernatants. However, genomes of rHAZV/Lmut1 were retained in the infected cells. All mutations (Lmut1-5) found in L protein of HAZVpi were required for experimental reproduction of HAZVpi, and only Lmut1 and Lmut4 were insufficient. We demonstrated that point mutations in viral polymerase contribute to the establishment of persistent HAZV infection. Furthermore, innate immunity was found to be suppressed in HAZVpi-infected cells, which also potentially contributes to viral persistence. This is the first presentation of a possible mechanism behind how nairoviruses establish persistent infection in human cells.

IMPORTANCE

We investigated whether and how nairoviruses persistently infect human cells, using Hazara orthonairovirus (HAZV), a surrogate model for Crimean-Congo hemorrhagic fever virus. We established a human cell line that was persistently infected with HAZV. Five mutations were found in L protein of persistently infected HAZV (HAZVpi): mut1, mut2, mut3, mut4, and mut5. Among them, Lmut1 and Lmut4 restricted viral growth by low polymerase activity and low growth rate, respectively, leading to inhibition of viral overgrowth. The restriction of viral growth caused by Lmut1 and Lmut4 was compensated by other mutations, including Lmut2, Lmut3, and Lmut5. Each of the mutations found in L protein of HAZVpi was concluded to cooperatively modulate viral growth, which facilitates the establishment of persistent infection. Suppression of innate immunity also potentially contributes to virus persistence. This is the first presentation of a possible mechanism behind how nairoviruses establish persistent infection in human cells.

INTRODUCTION

Crimean-Congo hemorrhagic fever virus (CCHFV) is a tick-borne virus (1) that belongs to the Orthonairovirus genus of the Nairoviridae in the order Bunyavirales (https://ictv.global/taxonomy). CCHFV is a cause of Crimean-Congo hemorrhagic fever that is highly lethal (approximately 30%) and is classified as a biosafety level (BSL) 4 agent. CCHFV genome has recently been detected in the testis of non-human primates 1 month after infection (2), indicating persistent CCHFV infection. CCHFV is enveloped, and its genome consists of three segments of single-stranded, negative-sense RNA (1). The three segments, S (small), M (medium), and L (large), encode nucleoprotein (N), glycoprotein (Gn and Gc), and L protein, respectively (1).

Hazara orthonairovirus (HAZV), originally isolated from Ixodes ticks collected in Pakistan, is closely related to CCHFV (3). HAZV can be handled in a BSL 2 facility because it is non-pathogenic to humans. HAZV and CCHFV show the same pathogenicity in type I interferon receptor knockout mice (4, 5). Furthermore, HAZV infection is lethal to embryonated chicken eggs (6), and its pathological condition is similar to CCHFV infection (7). HAZV is thus thought to be a good surrogate model for the study of virological characteristics of CCHFV.

The N protein primarily functions to encapsidate genomic RNA, and oligomerization of N is necessary for effective binding with RNA (8). N proteins of CCHFV and HAZV act to inhibit apoptosis (9, 10). HAZV N protein was recently found to function as a type I interferon antagonist (11). The glycoprotein precursor of nairoviruses is a typical class I membrane protein, which has an exterior amino-terminal domain and its carboxy-terminus anchored in the membrane, and is cleaved into two major structural proteins (Gn and Gc) by host proteases (12, 13). Unlike other negative-stranded RNA viruses, budding of nairoviruses is known to occur at the smooth membranes of the Golgi (1). Gn and Gc proteins are targeted to the Golgi (14, 15), so these proteins seem to be important for viral assembly. Nairoviruses do not have a matrix protein (1), suggesting a direct interaction between ribonucleocapsids and viral envelope proteins (Gn and Gc) through the Golgi membranes. L proteins of nairoviruses are much larger than those of other members in the Bunyavirales (1). L protein is an RNA-dependent RNA polymerase and interacts with the nucleocapsid or ribonucleoprotein complexes. L proteins of nairoviruses have right hand-like structures, similar to other negative-strand viruses (16, 17). They contain three subdomains (finger, palm, and thumb) where six functional conserved motifs (motifs A-F) reside (18). Nairovirus L proteins possess ovarian tumor (OTU) domains in their N-termini. The OTU domain of nairoviruses such as CCHFV and HAZV interfere with the host innate immune system via its ubiquitin protease activity (19, 20).

Virus rescue systems are powerful tools for studying viral pathogenesis, viral replication mechanisms, and functions of viral proteins. There have been reports of rescue systems of bunyaviruses including Bunyamwera orthobunyavirus (Peribunyaviridae) (21), La Crosse orthobunyavirus (LACV) (Peribunyaviridae) (22), Rift Valley fever phlebovirus (Phenuiviridae) (23), Argentinian mammarenavirus (Arenaviridae) (24), Lujo mammarenavirus (Arenaviridae) (25), CCHFV (26), and HAZV (27). All were T7 polymerase-based rescue systems, and in the case of CCHFV and HAZV rescue, high expression levels of T7 RNA polymerase were required for efficient virus rescue (26, 27).

Both CCHFV and HAZV established persistent infection in tick cells (28, 29). CCHFV has also been reported to persistently infect non-human primates (2). The animals persistently infected with CCHFV were clinically normal, so the pathogenicity of persistently infected CCHFV (CCHFVpi) seems to be different from that of wild-type (wt) CCHFV. However, there has not yet been virological analysis comparing CCHFVpi with wt CCHFV. Furthermore, mechanisms of establishment of persistent CCHFV infection in human cells remain unknown. Sendai virus, a member of the family Paramyxoviridae, also has the potential to establish persistent infection (30). Persistently infected SeV has been shown to have low cytotoxicity and temperature sensitivity. Our series of experiments demonstrated that these phenotype alterations were derived from mutations of viral proteins, such as M and L proteins (31–33).

In this study, we generated a human HEK293 cell line of persistently infected HAZV (HEK/HAZVpi). We analyzed HAZVpi growth and its genome sequences and investigated whether HAZVpi-encoded proteins contribute to the unique characteristics of HAZVpi using a reverse genetics system.

RESULTS

Persistent HAZV infection was established in HEK293 cells

SW13 cells and HEK293 cells were infected with HAZV at a multiplicity of infection (MOI) of 0.1 for 96 h. Under this condition, SW13 cells were predominantly detached from the culture dishes (Fig. 1A). Amounts of actin in lysates of HAZV-infected SW13 cells at 96 hours post-infection (hpi) markedly decreased (Fig. 1B), also indicating that few cells were attached. These cells could not be further passaged. In contrast, HAZV-infected HEK293 cells were still attached to the culture dish at 96 hpi (Fig. 1A). These cells had similar appearance to mock cells for 96 h culture. Thus, there seemed to be no virus-induced cytopathic effects (CPEs) (Fig. 1A). The cells were passaged every 3 to 4 days until passage 15. We confirmed whether these cells were infected with HAZV by immunoblot and microscopic observation using anti-N mAb. These cells retained the expression of N protein up to passage 15 (Fig. 2A), and N protein was detected in all cells (Fig. 2B), indicating that HEK293 cells were persistently infected with HAZV (HEK/HAZVpi).

Fig 1.

Time course of HAZV-infected SW13 cells and HEK293 cells. (A) SW13 cells and HEK293 cells were infected with HAZV at an MOI of 0.1 for the indicated hours and were subjected to light microscopy. Scale bars: 100 µm. (B) The lysates of the indicated cells infected with HAZV, as described in A, were subjected to immunoblot using anti-N mAb (911-1). Actin was used as a loading control.

Fig 2.

Establishment of persistent HAZV infection in HEK293 cells. (A) Immunoblot analysis of passaged HAZV-infected cells was performed as described in Fig. 1B. (B) Mock-infected HEK293 cells and HEK293 cells persistently infected with HAZV (HEK/HAZVpi) were fixed, permeabilized, and stained with anti-N mAb (3102–1) (green). Nuclei were stained with DAPI (blue). The cells were analyzed by Nikon Eclipse Ts2. Scale bar: 50 µm.

Virions of HAZVpi were not detected in the culture supernatant

We analyzed the subcellular localization of N protein in HEK/HAZVpi cells by immunofluorescence assay. N protein in HEK/HAZVpi cells showed a homogenous distribution within cytoplasm, while N protein in SW13 cells and HE293 cells with acute infection forms granules in the cytoplasm (Fig. 3A). To analyze the growth of HAZV in HEK/HAZVpi cells, the amount of viruses in culture supernatant of HEK/HAZVpi cells was quantified by plaque assay. Surprisingly, virions were not detected in culture supernatant of HEK/HAZVpi cells, despite N protein expression (Fig. 3B and C). To investigate the reason for non-detection of HAZVpi virion in the supernatants, total RNA was isolated from HEK/HAZVpi cells, and their HAZV genome was analyzed. No mutations were found in either the protein-coding or non-coding regions of the S segment. One mutation, C3738T, was observed in the protein-coding region of M segment, corresponding with A1224V in G protein (Gmut). Five mutations were found in the protein-coding region of L segment, all of which were missense mutations: Q1190R (Lmut1), G1683S (Lmut2), N2189D (Lmut3), I2685T (Lmut4), and E2852G (Lmut5) in L protein. No mutations were found in the non-coding regions of M and L segments.

Fig 3.

Analysis of HEK293 cell line persistently infected with HAZV. (A) HEK293 cells and SW13 cells were mock infected or infected with HAZV at an MOI of 0.1 for 24 h, and HEK/HAZVpi cells were fixed, permeabilized, and stained with anti-N mAb (3102–1) (green). Nuclei were stained with DAPI (blue). The cells were analyzed by Zeiss LSM900. Scale bar: 10 µm. (B) HEK293 cells were infected with HAZV at an MOI of 0.1 for the indicated hours. Immunoblot analysis of HAZV-infected HEK293 cells and HEK/HAZVpi cells was performed as described in Fig. 1B. (C) The amount of virus in the supernatants of HEK/HAZVpi cells and HAZV-infected HEK293 cells prepared as described in (B) was measured by plaque assay. PFU/mL values are shown as the means from three independent experiments. Error bars indicate standard deviations.

Mutations in L protein, but not G protein of HAZVpi, were involved in virion production

To examine the effect of Gmut on virion formation, we generated recombinant HAZV (rHAZV) using reverse genetics (Fig. 4A) and analyzed the growth of rHAZV/Gmut. The amounts of virion in the supernatants of rHAZV/Gmut-infected cells were similar to those of wt rHAZV- or HAZV JC280 strain-infected cells (Fig. 4B). The G_A1224V mutation does not seem to contribute to non-detection of HAZVpi virion in the supernatants.

Fig 4.

Effects of G protein on HAZVpi virion production. (A) Schematic workflow for rHAZV rescue is shown. BSR T7/5 cells were transfected with three pMK-RQ plasmids encoding HAZV S, M, or L genome and incubated for 120 h. A pMK-RQ plasmid encoding C3738T M genome was used for generation of rHAZV/Gmut. The supernatants were then inoculated into SW13 cells, which are highly susceptible to HAZV infection. After 48 h, the supernatants were collected as rHAZV stock. (B) SW13 cells were infected with the indicated HAZVs at an MOI of 0.01, and the supernatants were collected at 48 hpi. Viruses in these supernatants were quantified by plaque assay.

Next, we investigated the effects of L mutations on virion production. HEK/HAZVpi cells were transfected with pCAGGS encoding wt L protein, and virions in the supernatants were collected. Immunoblot using anti-N mAb revealed that virions in the supernatants of HEK/HAZVpi cells could be detected by overexpression of wt L protein (Fig. 5A), a clear indication that L protein of HAZVpi is involved in virion production.

Fig 5.

Effects of L protein on HAZVpi virion production. (A) HEK/HAZVpi cells were transfected with pCAGGS-L. After 48 h, the cell lysates and virions in the supernatants (collected as described in the Materials and Methods section) were subjected to immunoblotting, as described in Fig. 1B. HEK293 cells with mock and HAZV infection at an MOI of 0.1 for 48 h were used as negative and positive controls, respectively. (B and C) Virus rescue was performed as described in Fig. 4A. Transfected BSR T7/5 cells at 120 hours post-transfection (B) and rHAZV-infected SW13 cells at 48 hpi (C) were subjected to immunoblot, as described in Fig. 1B. “-” indicates the results from non-transfected cells. “S-,” “M-,” and “L-” indicate the results without the indicated HAZV genome. (D) The viruses in the supernatant of SW13 cells infected with the indicated rHAZVs for 48 h were quantified by plaque assay, as described in Fig. 3C. (E) Transfected BSR T7/5 cells were passaged three times every 3 days. Their supernatants were then infected with SW13 cells for 48 h. The lysates of BSR T7/5 cells and SW13 cells were subjected to immunoblot, as described in Fig. 1B. (F) Virus stocks were prepared from the supernatants of the passaged cells in E, and plaque assay was performed as described in D.

L_Q1190R (Lmut1) exhibited extremely low polymerase activity

We generated rHAZV carrying L_{Q1190R/G1683S/N2189D/I2685T/E2852G} (Lmut1–5) (rHAZV/Lmut1–5), which contains all mutations found in HAZVpi, as shown in Fig. 4A. Surprisingly, little or no expression of N protein was observed in either BSR T7/5 cells transfected with Lmut1–5 genome or infected SW13 cells (Fig. 5B and C, lane 11). To investigate which mutation is responsible for non-expression of N protein, we generated rHAZVs carrying mutated L proteins with single mutations, including Lmut1, Lmut2, Lmut3, Lmut4, or Lmut5. rHAZV/Lmut1, similar to rHAZV/Lmut1–5, did not express N protein in transfected BSR T7/5 cells or in infected SW13 cells (Fig. 5B and C, lane 6). Lmut1 is therefore suggested to exhibit extremely low polymerase activity, which consequently results in impaired virion production.

Interestingly, rHAZV/Lmut4-infected cells showed little expression of N protein (Fig. 5C, lane 9), despite sufficient amounts of N protein in the transfected BSR T7/5 cells (Fig. 5B, lane 9). Virions of rHAZV/Lmut4 could not be detected in the supernatants (Fig. 5D). We therefore expected the establishment of persistent infection of HAZV without virion production, similar to HAZVpi. To confirm whether HAZVpi was generated, transfected BSR T7/5 cells were passaged three times every 3 days before infecting SW13 cells with the supernatants. N protein was retained in the passaged BSR T7/5 cells (Fig. 5E). Unfortunately, N protein was also detected in the infected SW13 cells (Fig. 5E), and virions of rHAZV/Lmut4 virus stock could be detected in the supernatants (Fig. 5F). Lmut4 only is therefore thought to be insufficient for generation of HAZVpi.

Lmut2, Lmut3, and Lmut5 affected N protein expression levels in neither transfected BSR T7/5 cells nor infected SW13 cells (Fig. 5B and C, lanes 7, 8, and 10). Viral growth was also unaffected by these mutations (Fig. 5D).

I2685T mutation in L protein (Lmut4) decreased the virus growth rate

To further examine Lmut4 mutation, we generated rHAZVs containing additional mutations [G1683S/I2685T (Lmut2+4), N2189D/I2685T (Lmut3+4), G1683S/N2189D/I2685T (Lmut2–4), and G1683S/N2189D/I2685T/E2852G (Lmut2–5)], as well as rHAZV/Lmut4, and prepared virus stocks by our standard rescue system (Fig. 4A). Lmut4-containing mutations did not affect N protein expression levels in transfected cells (Fig. 6A). N protein in SW13 cells infected with rHAZV/Lmut2+4, rHAZV/Lmut3+4, rHAZV/Lmut2–4, and rHAZV/Lmut2–5 appeared earlier than in rHAZV/Lmut4 (Fig. 6B). The growth peak of rHAZVs carrying Lmut4-containing mutations was delayed compared with that of wt rHAZV, although their maximum titers were similar (data not shown). Lmut4 was suggested to result in a remarkable low growth rate, which was restored by Lmut2, Lmut3, and Lmut5.

Fig 6.

Analysis of rHAZVs carrying L mutants. (A and B) BSR T7/5 cells were transfected with HAZV genomes, as described in Fig. 4A. The supernatants of the transfected cells were transferred to SW13 cells and incubated for the indicated hours. The cell lysates of transfected BSR T7/5 cells (A) and infected SW13 cells (B) were subjected to immunoblot as described in Fig. 1B. “L-” indicates the results without L genome. (C) Transfected BSR T7/5 cells (passage 0) were passaged every 2–3 days until passage 5. Immunoblot was performed as described in A. (D) Total RNA of BSR T7/5 cells at passage 5 was isolated, reverse-transcribed, and subjected to qRT-PCR using primers amplifying HAZV genomes, as described in the Materials and Methods section.

Q1190R mutation in L protein (Lmut1) is a key mutation for the establishment of HAZV persistent infection

rHAZV/Lmut1 could not be rescued (Fig. 5B and C, lane 6). Transfected BSR T7/5 cells were therefore passaged every 3 days until passage 5. N protein was not observed in transfected BSR T7/5 cells (Fig. 6C, passage 0). Lmut1–5 only partially restored N protein expression (Fig. 6C, passage 0), and N protein expression could not be retained in the passaged cells (Fig. 6C, passage 5). However, each genome of rHAZV/Lmut1 and rHAZV/Lmut1–5 was retained, even after five times of passage (Fig. 6D). Lmut1 therefore seems to be a key mutation for the establishment of persistent infection.

All five mutations (Lmut1–5) found in L protein of HAZVpi are required for experimental reproduction of HAZVpi

Both rHAZV/Lmut1 and rHAZV/Lmut1–5 showed little or no N protein expression (Fig. 6C), which is critically different from HAZVpi. HAZVpi was speculated to be generated in HEK293 cells where HAZV genomes were heterogeneous, including wt ones at early passage. To mimic this condition, rescue of rHAZV/Lmut1 was performed in the presence of wt L protein (Fig. 7A). BSR T7/5 cells were transfected with pTM1-wt L, together with HAZV genomes, and passaged every 3 days until passage 5. N protein of rHAZV/Lmut1 and rHAZV/Lmut1+4 disappeared during cell passage (Fig. 7B). In contrast, N protein of rHAZV/Lmut1–5 was retained in passaged BSR T7/5 cells. Nevertheless, N protein was not detected in SW13 cells infected with supernatants of BSR T7/5 cells (Fig. 7C), indicating the generation of rHAZV expressing N protein in the infected cells without virion formation. All five mutations (Lmut1–5) found in L protein of HAZVpi and wt L protein are therefore required for the experimental reproduction of HAZVpi.

Fig 7.

Reproduction of HAZVpi by modified virus rescue of rHAZV. (A) Schematic workflow for modified rescue system is shown. BSR T7/5 cells were transfected with pTM1-wt L, together with HAZV genomes, and passaged every 3 days until passage 5. The supernatants of passaged cells were transferred to SW13 cells. (B and C) The lysates of the cells in (A) were subjected to immunoblot, as described in Fig. 1B.

Q1190R mutation in L protein (Lmut1) decreased its polymerase activity

We next examined the effect of these L mutations on polymerase activity using our HAZV minigenome system (34). Luciferase activity of Lmut1–5 was approximately fivefold higher than that of wt L (Fig. 8), despite only a little expression of N protein in rHAZV/Lmut1–5-infected cells (Fig. 5B, lane 11, and 6C). We then analyzed the effects of each mutation. Only Lmut1 resulted in a decrease of luciferase activity, consistent with the results in Fig. 5B (lane 6). In contrast, luciferase activity of three single L mutants (Lmut2, Lmut3, and Lmut4) was threefold to fivefold higher than that of wt L. Lmut5 did not affect luciferase activity. Lmut1 was the only mutant showing lower luciferase activity, so we analyzed double mutations containing mut1. Luciferase activity of the mutants Q1190R/G1683S (Lmut1+2), Q1190R/N2189D (Lmut1+3), and Q1190R/I2685T (Lmut1+4) was lower than that of single mutants of Lmut2, Lmut3, or Lmut4, suggesting that low luciferase activity of Lmut1 is compensated by Lmut2, Lmut3, and Lmut4. Q1190R/E2852G (Lmut1+5) resulted in almost no luciferase activity. We next examined combinations of mut2, mut3, and mut4, all of which showed higher luciferase activity. Luciferase activity of double mutants G1683S/N2189D (Lmut2+3), Lmut3+4, and Lmut2+4; triple mutant Lmut2–4; and quadruple mutant Lmut2–5 was similar to or moderately higher than that of their single mutant.

Fig 8.

Polymerase activity of L mutants. BSR T7/5 cells were transfected with plasmids carrying HAZV M minigenome expressing Rluc, together with pTM1-N and pTM1-wt L or pTM1-mutant L. pTM1-Fluc was used as an internal control. Two days post-transfection, luciferase activity was measured. Rluc expression from minigenomes was normalized by Fluc expression, and relative values are shown (wt L = 1). “L-” indicates the results from minigenomes without pTM1-L plasmid. Mutations are represented as white characters in black boxes. Data represent means from three independent experiments. Error bars indicate standard deviations.

Innate immunity was suppressed in HEK/HAZVpi

To investigate whether innate immunity affects virus persistence, we quantified the expression levels of representative innate immune genes, including pro-inflammatory cytokines/chemokines (IL-6, IL-8, IL-1β, and TNF-α) and anti-viral cytokine (IFN-β). Acute HAZV infection in SW13 cells caused a stronger induction of innate immune genes (especially IL-6 and IFN-β) than HEK293 cells with acute HAZV infection (Fig. 9). mRNAs of IL-6, IL-8, IFN-β, and TNF-α were upregulated by acute HAZV infection in HEK293 cells (Fig. 9). The amounts of these mRNAs in HEK/HAZVpi were lower than those in HEK293 cells with acute infection (Fig. 9). Innate immunity is therefore suggested to be suppressed in HEK/HAZVpi.

Fig 9.

Effects of HAZVpi on innate immunity. Total RNA of the indicated cells was isolated and reverse transcribed, and the expression of the indicated mRNAs was measured by qRT-PCR, as described in the Materials and Methods section. Data are the means from three independent experiments and are shown as the relative value (mock = 1). Error bars indicate standard deviations.

DISCUSSION

HAZV has recently been reported to establish persistent infection in tick cells without causing CPEs (29). CCHFV also established persistent infection in the testis, but lesions in the testis were mild (2). The low CPE is suggested to allow facilitation of the establishment of persistent infection. This is also supported by our result that persistent HAZV infection was established in HEK293 cells with almost no CPE by HAZV infection, but not in SW13 cells with severe CPE by HAZV infection (Fig. 1).

We found five mutations in L protein of HAZVpi (Q1190R, G1683S, N2189D, I2685T, and E2852G). None of these mutated residues were located in the conserved structural motifs (motif A–F) of L protein that are essential for polymerase functions. rHAZV carrying L_I2685T (Lmut4) showed extremely low viral growth rate (Fig. 6B), despite the high polymerase activity of L_I2685T (Fig. 8). The poor growth rate of rHAZV/Lmut4 was restored by other mutations, L_G1683S (Lmut2) and L_N2189D (Lmut3) (Fig. 6B). Mutation of L_Q1190R (Lmut1) decreased its polymerase activity (Fig. 8), which led to failure in the rescue of rHAZV/Lmut1 (Fig. 5B and C, lane 6). Lmut1 seems to abolish the high polymerase activity of L_G1683S (Lmut2), L_N2189D (Lmut3), and Lmut4, because L_{Q1190R/G1683S} (Lmut1+2), L_{Q1190R/N2189D} (Lmut1+3), and L_{Q1190R/I2685T} (Lmut1+4) showed lower polymerase activity than single mutants of Lmut2, Lmut3, and Lmut4, respectively (Fig. 8). rHAZVs carrying these double mutants containing Lmut1 could not be generated (data not shown). Lmut1 is therefore suggested to be fatal to virus growth. Notably, however, genomes of rHAZV/Lmut1 were retained in passaged cells (Fig. 6D), indicating the establishment of persistent infection. L_{Q1190R/G1683S/N2189D/I2685T/E2852G} (Lmut1–5), which is identical to L protein of HAZVpi, was the only Lmut1-containing mutant with high polymerase activity (Fig. 8). It was therefore unexpected that rHAZV/Lmut1–5 could not restore N protein expression and genome replication (Fig. 5B and C, lane 11; Fig. 6C and D). rHAZV carrying Lmut1–5 and Gmut, whose genomes are completely identical to those of HAZVpi, also did not affect N protein expression or genome replication (data not shown). Higher polymerase activity may not therefore be viable for infectious virus cycle. A similar pattern was observed in our previous report that a point mutation of N protein of human parainfluenza virus type 2 enhanced polymerase activity, but a virus carrying this mutation could not be rescued (35). Enhanced polymerase activity might disturb the appropriate balance between transcription and replication. By virus rescue in the presence of wt L protein as described in Fig. 7A, we generated rHAZV/Lmut1–5 expressing N protein in the infected cells without virion production (Fig. 7B). Conditions of the transfection experiments using BSR T7/5 cells were different from those of persistent HAZV infection in HEK293 cells, but we could generate a virus whose characteristics were similar to those of HAZVpi. wt L protein seems to be essential for establishment of persistent infection at an early stage. rHAZV/Lmut1 and rHAZV/Lmut1+4 could not restore N protein expression by this modified rescue system either (Fig. 7B). All five mutations found in HAZVpi L protein are likely to contribute to generation of HAZVpi.

Q1190 is critical for polymerase activity of HAZV (Fig. 8). This might also be true for other closely related viruses, because Q1190 is highly conserved among Nairoviridae (Table 1). I2685T mutation drastically the decreased growth rate of HAZV (Fig. 6B), regardless of its high polymerase activity (Fig. 8). I2685 therefore seems to be important for virus growth, but amino acid sequences around I2685 are diverse among nairoviruses (Table 1). G1683S and N2189D restored the virus growth rate of rHAZV/Lmut4 (Fig. 6B). N2189 is completely conserved among Nairoviridae (Table 1), and the region around this residue is also highly conserved. This residue therefore seems to be important for the polymerase activity of nairovirus L protein. G1683 in HAZV corresponds to S in CCHFV, so L protein of HAZVpi is a “CCHFV-like” L protein. However, G1683S mutation is not sufficient to explain the difference between CCHFV and HAZV because there are no reports that polymerase activity of CCHFV is higher than that of other nairoviruses. E2852 is a relatively highly conserved residue (Table 1). E2852 in HAZV L protein corresponds to G in Sakhalin orthonairovirus L protein, indicating that E2852G mutation (Lmut5) is a minor change for nairoviruses. E2852G mutant actually affects neither polymerase activity nor viral growth (Fig. 5D and 8).

TABLE 1.

Comparison of the mutated residues of HAZVpi L protein with those of L proteins of other members of the genus Orthonairovirus^a

Virus	Position of amino acid residue in HAZV L protein
	1190	1683	2189	2685	2852
HAZV	Q	G	N	I	E
HAZVpi	R	S	D	T	G
CCHFV	Q	S	N	L	E
Dugbe orthonairovirus	Q	A	N	L	E
Nairobi sheep disease orthonairovirus	Q	C	N	L	E
Thiafora orthonairovirus	Q	S	N	V	E
Sakhalin orthonairovirus	Q	G	N	I	G
Artashat orthonairovirus	Q	M	N	L	E
Hughes orthonairovirus	Q	-	N	E	E
Dera Ghazi Khan orthonairovirus	Q	S	N	R	E
Keterah orthonairovirus	N	T	N	-	A
Qalyub orthonairovirus	S	-	N	P	A
Chim orthonairovirus	C	T	N	P	A
Tamdy orthonairovirus	Q	S	N	E	T

^a The accession numbers of nucleotide sequences encoding orthonairovirus L proteins used are BBF98753 (HAZV), AEO72046 (CCHFV), NP_690576 (Dugbe orthonairovirus), YP_009361832 (Nairobi sheep disease orthonairovirus), YP_009513191 (Thiafora orthonairovirus), AKC89334 (Sakhalin orthonairovirus), YP_009666119 (Artashat orthonairovirus), QLA46953 (Hughes orthonairovirus), AMT75389 (Dera Ghazi Khan orthonairovirus), YP_009361838 (Keterah orthonairovirus), AKC89319 (Qalyub orthonairovirus), YP_009666113 https://www.ncbi.nlm.nih.gov/protein/YP_009666113(Chim orthonairovirus), and QFU19352 (Tamdy orthonairovirus).

Surprisingly, virions of HAZVpi were not observed in the culture supernatants. We initially expected that the virus budding would be impaired by mutation in Gn or Gc protein. One mutation (C3738T) was found in M segment of HAZVpi, which corresponds with A1224V in G protein. Based on the cleavage sites of HAZV G protein proposed by Shimada et al., this residue is located in the extracellular domain of Gc protein at its C-terminus (13). No mutations were found in S segment of HAZVpi, so the interaction between Gc and N proteins is not thought to affect the defect of virion formation of HAZVpi. The release of virions from HEK/HAZVpi cells was not recovered by transient expression of G protein (data not shown). Furthermore, rHAZV carrying A1224V mutation was detected in the supernatants, similar to wt rHAZV (Fig. 4B). The residue A1224 does not seem to contribute to virion production.https://www.ncbi.nlm.nih.gov/protein/AKC89319https://www.ncbi.nlm.nih.gov/protein/AKC89319

Expression of wt L protein in HEK/HAZVpi cells led to the detection of virions in the supernatants (Fig. 5A), a clear indication that mutations in L protein of HAZVpi are involved in virion production. Nairovirus L protein is much larger than L proteins of other members of the Bunyavirales (1), and there are regions in which the function remains unknown. Nairovirus L protein might have important roles, like matrix protein, in regulating virion formation. Mutations in HAZVpi L protein are likely to result in defects for virus assembly rather than budding, because no infectious virions of HAZVpi were detected intracellularly (data not shown). Interestingly, virions of rHAZV/Lmut4 generated as described in Fig. 4A could not be detected in the supernatants (Fig. 5D), despite sufficient amounts of N protein in the infected cells (Fig. 5B, lane 9); this is similar to HAZVpi. However, virions of rHAZV/Lmut4 could be detected in the supernatants of infected cells after three passages (Fig. 5E). There were equal amounts of N protein in wt rHAZV-infected and rHAZV/Lmut4-infected cells (Fig. 5E), but the amounts of rHAZV/Lmut4 virion in the supernatants were approximately 100-fold lower than those of wt rHAZV virion (Fig. 5F). Lmut4 by itself might therefore directly interfere with the virion assembly. Lmut1 also appears to affect virion production because virions of rHAZV/Lmut1 were not detected in the culture supernatants (Fig. 5C, lane 6). Lmut1 might simply remarkably delay virion formation due to its extremely low polymerase activity. Detailed investigations into how these mutations affect virion assembly are currently underway. Measles virus variants causing subacute sclerosing panencephalitis are defective for cell-free virus production and establish persistent infection (36). Impaired virion formation might be an advantage in the establishment of persistent infection by facilitating the escape from the host immune system.

The amounts of innate immunity-related genes in HEK/HAZVpi cells were lower than those in HEK293 cells with acute HAZV infection (Fig. 9). Innate immunity therefore appears to be suppressed in HEK/HAZVpi, probably because of some effects caused by long-term HAZV infection. This might contribute to persistent infection. Acute HAZV infection in SW13 cells exhibited a much stronger induction of innate immune genes such as IL-6 and IFN-β than HEK293 cells with acute infection (Fig. 9). The lower responsiveness to innate immune stimuli in HEK293 cells might also facilitate persistent infection. Retroviruses incorporate themselves into host cell chromosomes and their genomes are maintained by cell division (37). However, cell division does not seem to be involved in the HAZV persistence because HAZV genomes were not detected in host cell genomic DNA (data not shown).

Both CCHFV and HAZV have been reported to persistently infect tick cells (28, 29). This is not surprising, because ticks are reservoir hosts of CCHFV and HAZV (1). Salvati et al. demonstrated the presence of viral-derived DNA forms (vDNAs) after HAZV infection in tick cells, which is involved in the establishment of persistent infection (29). However, this might be specific to tick cells, because vDNAs were not detected in mammalian cells (29). Furthermore, unlike our results, virions were detected in the supernatants of tick cells persistently infected with HAZV (29). The mechanisms of establishment of persistent infection in tick cells are likely to be different from those in mammalian cells.

In summary, we found that two mutations (Q1190R and I2685T) in L protein of HAZVpi restrict viral growth by low polymerase activity and low growth rate, respectively, and this inhibits viral overgrowth. The restriction of viral growth caused by Q1190R and I2685T was compensated for by other mutations, including G1683S, N2189D, and E2852G. All mutations found in L protein of HAZVpi cooperatively modulate viral growth, which facilitates the establishment of persistent infection. We demonstrated that point mutations in viral polymerase contribute to the establishment of persistent HAZV infection. This is the first presentation of a possible mechanism behind how nairoviruses establish persistent infection in human cells. HAZV L protein has not yet been fully structurally and functionally characterized. Mutations found in HAZVpi L protein might affect unknown functions of L protein other than polymerase activity, leading to the establishment of persistent infection. Further study is required for better understanding of how nairoviruses establish persistent infection.

MATERIALS AND METHODS

Cells and virus

SW13 cells and HEK293 cells were grown in Dulbecco’s modified Eagle’s minimal essential medium (DMEM) containing 10% and 5% fetal calf serum (FCS), respectively. BSR T7/5 cells that constitutively express T7 RNA polymerases (38) were grown in DMEM containing 2.5% FCS. All cells were maintained in a humidified incubator at 37°C with 5% CO₂. HAZV (strain JC280) was used in this study (3).

Antibodies

MAbs against HAZV N protein (911-1 used for immunoblot and 3102-1 used for immunofluorescence) have been previously described (10, 11). Anti-actin mAb was purchased from Wako (Osaka, Japan).

Plasmids

HAZV L cDNA was cloned into pCAGGS vector (39). pMK-RQ-S, pMK-RQ-M, and pMK-RQ-L used for HAZV reverse genetics were kindly gifted by Dr. John N. Barr (University of Leeds, UK) (27). Rluc-expressing HAZV M minigenome, pTM1-N, pTM1-wt L, and pTM1-Fluc were previously described (34). Construction of plasmids M and L mutants was prepared by a standard PCR mutagenesis method, and cDNAs encoding M and L mutants were cloned into pMK vector. cDNAs containing L mutants were also cloned into pTM1 vector.

Immunoblot assay

Cells were harvested, sonicated for 30 seconds three times in lysis buffer containing 50 mM Tris-HCl (pH7.4), 150 mM NaCl, and 0.6% NP-40, and then centrifuged. Cell lysates were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and analyzed by western blot technique.

Immunofluorescence assay

Cells grown on cover glasses in 24-well plates were fixed with 4% paraformaldehyde for 20 min and permeabilized with phosphate-buffered saline containing 0.2% Triton X-100 for 15 min. The cells were then incubated for 60 min with anti-N mAb. The secondary antibody used was Alexa Fluor 488 goat anti-mouse IgG (Invitrogen, Carlsbad, CA, USA). The cells were mounted with Fluoromount-G (Southern Biotech, Birmingham, AL, USA) and analyzed by a Nikon Eclipse Ts2 microscope (Nikon, Tokyo, Japan) or Zeiss LSM900 confocal microscope (Zeiss, Oberkochen, Germany).

Plaque assay

SW13 cells grown in 12-well plates were infected with HAZV diluted serially 10-fold in DMEM containing 2% FCS, 0.4% SeaKem ME agarose, and 0.4% SeaPlaque agarose (FMC Bioproducts, Rockland, ME, USA) until plaques were visible. The cells were then stained with 0.3% amide black.

HAZV genome sequence

The total RNA was isolated from HEK/HAZVpi cells by Isogen (Nippon Gene, Tokyo, Japan) according to the manufacturer’s instructions. cDNA was synthesized using the PrimeScript RT Reagent Kit (TaKaRa, Shiga, Japan) with random hexamer oligo as the primer for reverse transcription and was subjected to PCR. PCR products were cloned into pDrive cloning vector using the QIAGEN PCR Cloning Kit (QIAGEN, Hilden, Germany). The 1,677-nt length of HAZV S genome was amplified using primers F: 5′-tctcaaagacaaacgtgccgcagacgcccc-3′ and R: 5′-tctcaaagatatcgttgccgcacagcccca-3′. The 4,575-nt length of M genome was divided into two regions, and primers used for amplifying each region were as follows: Region 1 (1–2,383 nt; F: 5′-ctcaaagacagacttgcggcacacacaaaa-3′, R: 5′-ccgcgggccctctagacttgtcattgactggccatcaggac-3′) and Region 2 (2,365–4,575 nt; F: 5′-gatggccagtcaatgacac-3′, R: 5′-tctcaaagatatcgtggcggcacaccctaa-3′). The 11,980-nt length of L genome was divided into six regions, and primers used for amplifying each region were as follows: Region 1 (1–2,654 nt; F: 5′-tcaaagacatcatcccccttatccccaa-3′, R: 5′-ctgaagagatggcttatggaattc-3′), Region 2 (2,359–4,928 nt; F: 5′-cagcttgactacaatagcc-3′, R: 5′-gtgtttgggtttagggtttcaagcagattgag-3′), Region 3 (4,102–6,655 nt; F: 5′-aaagagaatatgctgtgctg-3′, R: 5′-cgcttctaggtagtctctcgcctg-3′), Region 4 (6,433–8,550 nt; F: 5′-tcaggatccgagccattttccctcagtct-3′, R: 5′-aatctcgaggggatcttctatgactctgt-3′), Region 5 (7,991–10,150 nt; F: 5′-tatggatccgttctgctgtggctctaga-3′, R: 5′-tatctcgaggccagcttagcacacag-3′), and Region 6 (9,969-11,980 nt; F: 5′-ctacttaggaacaacggcaa-3′, R: 5′-tctcaaagatatcgttcccc-3′).

Virus rescue

BSR T7/5 cells grown in six-well plates were transfected with 1.5 µg of pMK-RQ-S, 1.0 µg of pMK-RQ-M, and 1.5 µg of pMK-RQ-L using Mirus TransIT-LT1 Transfection Reagent (Mirus Bio, Madison, WI, USA). For generation of mutant viruses, the wt plasmid was replaced with the corresponding mutant plasmid. At 120 h post-transfection, the supernatants were transferred to SW13 cells. For the modified rescue system, BSR T7/5 cells were transfected with 0.05 µg of pTM1-wt L, together with HAZV genomes, and then passaged until passage 5. Virus production was confirmed by detection of N protein in SW13 cells using immunoblot and by quantification of viruses in culture supernatants using plaque assay.

Purification of virus particles

The supernatants of the HAZV-infected cells were centrifuged at 1,000 × g for 5 min to remove the cell debris. They were loaded onto TNE buffer [Tris-HCl (pH 8.0), 100 mM NaCl, and 1 mM EDTA] containing 25% glycerol and then centrifuged at 20,000 × g for 20 min. Pellets were suspended in sample buffer, and N proteins were detected by immunoblot using anti-N mAb.

HAZV minigenome assay

BSR T7/5 cells grown in 12-well plates were transfected with plasmids (HAZV M minigenome, 0.5 µg; pTM1-N, 0.4 µg; and pTM1-L or empty vector, 0.2 µg; pTM1-Fluc, 0.1 µg) using XtremeGene HP (Roche, Basel, Switzerland). At 2 days post-transfection, the Rluc and Fluc activities were measured by a dual-luciferase assay kit (Promega) according to the manufacturer’s instructions.

Quantitative real-time reverse transcription PCR (qRT-PCR)

Total RNA was isolated from HAZV-infected cells by Isogen, and cDNA synthesis was carried out using the PrimeScript RT Reagent Kit with the specific primers for HAZV S (5′-gccaccaacatcaacatcat-3′), M (5′-actctatagcttcgcccgc-3′), or L genome (5′-cggtgtggagtgttgctga-3′). cDNAs were subjected to qRT-PCR using Brilliant III Ultra-Fast SYBR Green Master Mix (Agilent Technologies, Santa Clara, CA, USA). The primers used for qRT-PCR were as follows: HAZV S (F: 5′-gattgcggctaggttcactg-3′, R: 5′-aaagatatcgttgccgcacag-3′), M (F: 5′-gaccatagccagctagagca-3′, R: 5′-ctcaaagatatcgtggcggc-3′), and L genomes (F: 5′-tgtgtctccaaacccacaact-3′, R: 5′-cccacaccccaattttaatcttct-3′). A standard curve was generated from dilutions whose copy numbers were known, and the RNA per well of samples was quantified based on this standard curve. To amplify the host cell genes, reverse transcription was performed using oligo-dT primer. The primers used for qRT-PCR were as follows: IL-6 (F: 5′-actggtcttttggagtttgaggt-3′, R: 5′-gggtcaggggtggttattgc-3′), IL-8 (F: 5′-tactccaaacctttccacccc-3′, R: 5′-caaccctctgcacccagttt-3′), IL-1β (F: 5′-gtacctgtcctgcgtgttga-3′, R: 5′-gggaactgggcagactcaaa-3′), IFN-β (F: 5′-gcacaacaggtagtaggcga-3′, R: 5′-tggaaagagctgtcgtggag-3′), TNF-α (F: 5′-atggcgtggagctgagagat-3′, R: 5′-tctggtaggagacggcgatg-3′), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (F: 5′-gaaggtcggagtcaacggattt-3′, R: 5′-atcttgaggctgttgtcatacttct-3′). GAPDH was used as an internal control.