The
Flaviviridae family comprises the genera
Flavivirus,
Pestivirus, and
Hepacivirus; the latter includes the human pathogen
Hepatitis C virus (HCV) (
13). With an estimated 200 million cases of chronic infections worldwide, HCV is a major cause of liver cirrhosis and hepatocellular carcinoma. Due to their close relationship, pestiviruses, especially bovine viral diarrhea virus (BVDV), represent a widely used surrogate system for HCV.
The single-stranded RNA genome of BVDV is of positive polarity and has a length of 12.3 kb. Gene expression occurs via translation of one polyprotein which is processed by cellular and viral proteases giving rise to 12 mature proteins (
18). Processing of nonstructural protein 2-3 (NS2-3) into NS2 and NS3 is exerted by a recently characterized vital cysteine autoprotease located in NS2 (
17). Interestingly, this enzyme is distantly related to the HCV NS2-3 protease which mediates the analogous cleavage in the HCV polyprotein (
11,
14,
18). The NS4A-dependent chymotrypsin-like serine protease in NS3 catalyzes four processing events in the viral polyprotein (
29,
32,
33). Moreover, NS3 has helicase and NTPase activity (
28,
31). The enzymatic functions of NS3 are essential for viral RNA replication which is accomplished by a replication complex (replicase) containing NS3 and four other NS proteins including NS5B, the viral RNA-dependent RNA polymerase (
34), as essential constituents (
3,
12). The NS2 protease-mediated cleavage of NS2-3 is essential for replication of BVDV, since its cleavage product, NS3, cannot be functionally replaced by NS2-3 in the viral replicase (
17). Uncleaved NS2-3 plays a crucial but so-far-undefined role in the generation of infectious progeny virus (
1).
In tissue culture cells, BVDV and other pestiviruses appear in two biotypes: the replication of noncytopathogenic (noncp) strains of BVDV does not induce obvious changes in cell morphology and viability, whereas cytopathogenic (cp) isolates cause the death of the infected cell (
21). Only the noncp BVDV strains have the remarkable capability to establish lifelong persistent infections, which is of crucial importance for the maintenance of BVDV in cattle populations worldwide. In this context, it is noteworthy that specifically in noncp BVDV-infected cells, a unique temporal regulation of NS2-3 processing was observed (
17). While cleavage is almost complete in the first hours of infection, its efficiency decreases below the detection level as early as 9 h after infection. The resulting decrease in NS3 release correlates with a severe downregulation of viral RNA replication. cp BVDV mutants emerge from noncp ancestors in the course of persistent infection and cause progression to lethal disease. These cp BVDV strains have lost the ability to downregulate the generation of NS3 during infection and are characterized by a dramatically upregulated RNA synthesis and their inability to establish persistent infections (
17-
19,
21). These findings emphasize the biological significance of NS2-3 cleavage regulation in biotype control and viral persistence. The subject of this study was the molecular basis for the unique temporal regulation of the NS2 autoprotease in noncp BVDV-infected cells. One possible mechanism was a limited amount of a cellular factor required for NS2-3 cleavage. With respect to this hypothesis, it was intriguing that overexpression of a cellular chaperone of the J-domain family strongly stimulated NS2-3 processing in noncp BVDV-infected cells, leading to a change in the viral biotype from noncp to cp (
22,
25). Moreover, in BHK-21 cells, noncp BVDV-derived NS2-3 displayed significant cleavage only upon coexpression of this chaperone (
25). Since NS2 and the chaperone form a complex, the latter was termed “J-domain protein interacting with viral protein” (Jiv); its human ortholog, HDJ3, has been recently described (
6). For interaction with NS2 as well as for NS2-3 cleavage induction, a 90-amino-acid (aa) fragment of Jiv, termed Jiv90, was found to be sufficient (
25). The mechanism by which Jiv or Jiv90 stimulates NS2-3 cleavage as well as the significance of the intracellular Jiv level for viral replication were studied in this report. A central question was the role of Jiv/Jiv90 in NS2-3 cleavage.
The data obtained strongly suggest that cellular Jiv acts as an activating cofactor of the viral NS2 autoprotease which in turn regulates viral RNA replication. A direct correlation between alterations in the intracellular Jiv level and the efficiency of NS2-3 cleavage as well as viral RNA replication could be demonstrated. The observed limitation of viral RNA replication by the level of a cellular chaperone represents a unique regulatory mechanism which is crucial for the ability of this virus to establish lifelong persistent infections.
MATERIALS AND METHODS
Cells and viruses.
Madin-Darby bovine kidney (MDBK) cells, bovine kidney fibroblast cell line PT, and BHK-21 cells were grown in Dulbecco's modified Eagle's medium with 10% fetal calf serum. Cells were maintained at 37°C and 5% CO
2. PT cells were kindly provided by R. Riebe (Friedrich-Loeffler-Institut, Riems, Germany). Cell line MDBK
tet-onJiv, which allows an inducible overexpression of Jiv, has been published previously (
25). Vaccinia virus modified virus Ankara-T7pol (
27) was generously provided by G. Sutter (GSF, Oberschleißheim, Germany). BVDV strains NCP8 and NCP7 were described previously (
8).
Antibodies and antisera.
The anti-Flag tag and anti-glutathione-
S-transferase (GST) monoclonal antibody (MAb) were purchased from Sigma-Aldrich (Taufkirchen, Germany). For the detection of NS3, mouse MAb 8.12.7 (
7) was used. Secondary species-specific antibodies were purchased from Dianova (Hamburg, Germany).
Expression plasmids.
pCITE (Novagen, Madison, Wis.) encompasses the internal ribosomal entry site of encephalomyocarditis virus downstream of the T7 RNA polymerase promoter. The following constructs are based on pCITE and code for the indicated amino acids (the amino acid positions refer to BVDV strain SD-1 [
9]): pflag-NS2-4A (MDYKDDDDKL followed by aa 1137 to 2336 of BVDV NCP7), pflag-NS2 (MDYKDDDKL followed by aa 1137 to 1589 of BVDV NCP7), and pGST-Jiv90 (based on pCITE and codes for GST followed by aa 533 to 622 of bovine Jiv [Jiv90]). Truncations of Jiv90 were generated by PCR and code for GST and the indicated amino acids of the bovine Jiv protein: pGSTJiv90ΔN10 (aa 543 to 622), pGSTJiv90ΔN20 (aa 553 to 622), pGSTJiv90ΔN30 (aa 563 to 622), pGSTJiv90ΔN40 (aa 573 to 622), pGSTJiv90ΔN50 (aa 583 to 622), pGSTJiv90ΔN60 (aa 593 to 622), pGSTJiv90ΔC5 (aa 543 to 617), pGSTJiv90ΔC10 (aa 543 to 612), pGSTJiv90ΔC20 (aa 543 to 602), pGSTJiv90ΔC30 (aa 543 to 592), pGSTJiv90ΔC40 (aa 543 to 582), pGSTJiv90ΔC50 (aa 543 to 572), pGSTJiv90ΔC60 (aa 543 to 562), and pGSTJiv90ΔN10/ΔC5 (aa 543 to 617). Mutations within Jiv or NS2 were introduced by PCR or the QuikChange method (Stratagene, Heidelberg, Germany). All constructs were verified by DNA sequencing.
Metabolic labeling of transiently expressed proteins.
For transient expression of proteins, the T7-vaccinia virus system was used. A total of 106 BHK-21 cells per 3.5-cm tissue culture dish were infected with vaccinia virus modified virus Ankara-T7pol at a multiplicity of infection of 5 in 1 ml culture medium lacking fetal calf serum for 1 h at 37°C. For transfection of plasmid DNA (3 μg), Superfect was applied (QIAGEN, Hilden, Germany). Two hours posttransfection of plasmid DNA, cells were incubated for 30 min with Dulbecco's modified Eagle's medium without cysteine and methionine (label medium) at 37°C prior to the addition of 1 ml of label medium containing 70 μCi of [35S]methionine-[35S]cysteine (35S-ProMix; Amersham Biosciences, Freiburg, Germany) per 3.5-cm tissue culture dish; protein expression was allowed to proceed for 2 h at 37°C. Cells were lysed in radioimmunoprecipitation (RIP) assay (RIPA) buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% [vol/vol] NP-40, 1% [wt/vol] deoxycholate, 0.1% [wt/vol] sodium dodecyl sulfate [SDS], and 0.5 mM PefablocSC [Merck, Darmstadt, Germany]).
Metabolic labeling of proteins in BVDV-infected cells.
BVDV infection with noncp BVDV strain NCP7 was performed in a 3.5-cm tissue culture dish at a multiplicity of infection of 10 for 1 h at 37°C. Prior to metabolic labeling, cells were incubated for 1 h with label medium at 37°C prior to the addition of 1 ml of label medium containing 300 μCi of [35S]methionine-[35S]cysteine (35S-ProMix; Amersham Biosciences, Freiburg, Germany). Protein expression was allowed to proceed for 1 h at 37°C. Cells were lysed in RIPA buffer.
RIP and SDS-polyacrylamide gel electrophoresis (PAGE).
Protein A-Sepharose (Sigma-Aldrich, Taufkirchen, Germany) and RIPA buffer were used for RIP. In order to apply the antibodies in excess amounts, they were titrated against the highest amount of antigen used throughout the study. Proteins were separated in polyacrylamide-Tricine gels with 8 or 10% polyacrylamide (also see reference
25).
Quantification of NS2-3 cleavage efficiencies.
Radioactivity of proteins in the dried SDS-PAGE gels was determined by phosphorimaging. Protein amounts were calculated under consideration of the numbers of methionine and cysteine residues in the proteins. Cleavage efficiency was calculated as the quotient of the signals of an NS2-3 cleavage product and this signal plus the signal of the precursor [product/(product + substrate)].
Generation of Jiv knockdown cells.
Bovine kidney fibroblast cell line PT
tet-on, which stably expresses the reverse tetracycline transcriptional transactivator protein of the Tet-on system (
10), was established as described previously for cell line MDBK
tet-on (
25). Due to the low endogenous expression level of Jiv in all bovine cell lines tested so far, this protein could be detected by neither Western blot nor RIP (
25); accordingly, the effectiveness of Jiv-specific hairpin RNAs which were constitutively expressed by a pSUPER (
5) derivative and further processed into siRNAs targeting the Jiv mRNA were tested by Western blot in BHK-21 cells overexpressing Jiv (data not shown). Based on these tests, pSUPER-Jiv1 was selected and cotransfected with pEF-Pac (
25) into PT
tet-on cells; cell clones were selected in the presence of 5 μg/ml puromycin. In addition, pSUPER-Jiv1 encoded the green fluorescent protein under the control of the tetracycline-regulated promoter to ease the identification of stably transfected cells. The sequence inserted downstream of the H1 promoter between sites for BglII and HindIII was GATCCCC
GTGGCTCGACTCTTGACCATTCAAGAGA
TGGTCAAGAGTCGAGCCACTTTTTGGAAA (Jiv-specific sense and reverse small interfering RNA [siRNA] sequences are in boldface type). The relative amount of Jiv mRNA in PT “Jiv knockdown” cell clones and parental cells was compared by quantitative real-time reverse transcription-PCR (RT-PCR) analysis using an ABI Prism 7000 apparatus (Applied Biosystems). Cell line PT-Jiv-kd#8.7, which was used in this study, is termed PT-Jiv-kd and displayed a reduction of the Jiv mRNA amount by about 75 to 85% in several independent tests and stably maintained this status over several passages (data not shown).
Generation of PT-Jiv-kd-rescue cells.
Silent mutations were introduced into pTRE-Jiv (
25) coding for Jiv under the control of a tetracycline-inducible promoter at positions 1477 (T→G), 1480 (A→C), and 1483 (C→G) of the published Jiv sequence (
25) to render the transcribed RNAs resistant against the Jiv-specific siRNA expressed in cell line PT-Jiv-kd#8.7. When compared to authentic Jiv mRNA, the recombinant Jiv mRNA lacked a short upstream open reading frame in the 5′ nontranslated region; accordingly, the recombinant Jiv mRNA should be translated with a significantly higher efficiency (
25). Two micrograms of this plasmid was cotransfected with 50 ng of plasmid pTK-Hyg (Clontech, Palo Alto, Calif.) to confer hygromycin resistance to the transfected cells. Stable cell clones were selected and grown in medium containing 0.5 mg/ml hygromycin B. Further passaging of the cells was performed in the absence of hygromycin. Jiv expression was verified in the obtained cell clones upon induction with 10 μM doxycycline (Dox) by indirect immunofluorescence using a Jiv-specific rabbit antiserum directed against recombinant Jiv90 expressed in
Escherichia coli and a secondary cyanogen-3-labeled antibody (data not shown).
Quantitative real-time RT-PCR.
Total cellular RNA was prepared using the Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany). For comparative quantification of the cellular Jiv mRNA, 500 ng of the RNA from different cell lines was subjected to real-time RT-PCR analysis. After reverse transcription using the reverse primer Jiv02R (GCCAAGAGAAGATCCAGGTGG) and Superscript II reverse transcriptase (Invitrogen, Karlsruhe, Germany), the quantitative real-time PCR was run using the AbiPrism7000 Sequence Detection system (Applied Biosystems, Branchburg, N.J.) in the “Absolute Quantification” modus with TaqMan Universal PCR Master mix (Applied Biosystems, Branchburg, N.J.) without MgCl2 in combination with the Jiv-specific probe JivTaq01 (VIC-TGACCGGCTAGGCTGGAGGGATAAA-6-carboxytetramethylrhodamine) and the primer pair Jiv01 (GGCGGTTTCTGGTAGGATTG)-Jiv02R. Cycling conditions were 1 cycle (2 min at 50°C and 10 min at 95°C) followed by 40 cycles (15 s at 95°C and 1 min at 60°C).
For comparative quantification of viral genomic RNA in BVDV-infected cells, the BVDV-1-specific probe pvtaq01 (6-carboxyfluorescein-ACAGTCTGATAGGATGCTGCAGAGGCCC-6-carboxytetramethylrhodamine) and the primer pair pv02 (GTGGACGAGGGCATGCC)-pv03R (TCCATGTGCCATGTACAGCAG) were used under the conditions described above. The relative mRNA amounts in comparison to control cells were in both cases calculated using the following formula: % mRNA = 1/(2ΔCt) × 100.
BVDV growth curves and virus titration.
Infection with noncp BVDV strain NCP8 was carried out as described previously for 1 h at 37°C (
25). End-point titration was performed in four replicates on MDBK cells. Virus detection occurred after 72 h at 37°C by indirect immunofluorescence analysis using MAb 8.12.7 directed against NS3 of BVDV (
7) and a secondary cyanogen-3-labeled antibody as described previously (
25).
DISCUSSION
Gene expression of the large group of viruses with positive-strand RNA genomes involves translation of polyproteins. Proteolytic processing of these precursors is subtly regulated and highly critical for viral replication. Perturbations of this process as a rule interfere with viral replication (
16). Along this line, the pestiviral NS2 autoprotease is vital for pestiviral RNA replication by generating NS3, an essential component of the viral replicase (
17). Interestingly, the cellular chaperone Jiv forms a complex with pestiviral NS2 and stimulates NS2-3 cleavage (
25); the role of Jiv in NS2-3 processing was one topic of this study. Several lines of evidence argue against a proteolytic activity of Jiv itself. First, the minimal fragment of Jiv able to induce NS2-3 cleavage is only 80 amino acids long (amino acids 11 to 90 of Jiv90) and would therefore represent the shortest protease active in
trans. Second, in this peptide, not even the identity of a single amino acid is absolutely essential for the induction of cleavage; only an aromatic residue has to reside at position 39. This finding excludes for this Jiv peptide the existence of an amino acid motif essential for cleavage induction, a typical characteristic of all known classes of proteases. Third, Jiv can be efficiently coprecipitated with NS2, and much more than just catalytic amounts of Jiv are required for optimal induction of NS2-3 cleavage. The fact that proteases usually do not form stable complexes with their substrates together with the required amounts of Jiv strongly favor another role of this protein, namely, to function as a protease cofactor. In agreement with this hypothesis, mutations that have been shown to abolish the activity of the NS2 autoprotease (
17) were also detrimental for the cleavage of noncp BVDV-derived NS2-3 in the presence of Jiv. Taken together, the results obtained strongly imply that the cellular chaperone Jiv is not a protease itself but represents a cofactor required for the activation of the pestiviral NS2 autoprotease. Future detailed studies on the interaction between these proteins together with structural data are required to elucidate the molecular basis of NS2 protease activation by Jiv.
Jiv is a member of the J-domain protein family which represents a heterogeneous group of chaperones characterized by a 70-amino-acid-long so-called J domain. Binding of substrate as well as of the J-domain protein to Hsp70 dramatically stimulates the protein folding activity of the latter. Cellular J-domain proteins have already been implicated in the replication of viruses. For example, the assembly of the replication complex of hepatitis B virus was found to require the J-domain protein Hdj1 as well as Hsp70 (
2,
15). The positive-strand RNA virus brome mosaic virus depends on the assistance of cellular chaperones including the J-domain protein Ydj1 for the formation of its replication complex (
30). In contrast to these systems utilizing authentic chaperones, we observed that a small fragment of Jiv (Jiv90) was sufficient to assist in pestiviral replication. Since Jiv90 does not encompass the J-domain, a crucial determinant for Hsp70 binding, it is questionable whether members of the Hsp70 family of chaperones are essentially involved in the activation of the NS2 protease.
Probably the most intriguing aspect of the pestiviral NS2 protease is its temporal downregulation leading to a severe restriction of viral RNA replication late in noncp BVDV infection. The data obtained in this study fit into a model whereby each NS2-3 molecule recruits one Jiv molecule to activate its NS2 protease for autocleavage. This assumption is supported by the amounts of Jiv required for optimal NS2-3 cleavage and the fact that the released NS2 and Jiv could be coprecipitated in a complex in which both proteins were detected at a ratio of 1:1.
In noncp BVDV-infected cells, the synthesis rate of viral NS2-3 is much higher than the one of Jiv (
25). Therefore, the pool of Jiv molecules present at the time of infection most likely is the major determinant for the number of NS2-3 molecules that will undergo autoprocessing in a noncp BVDV-infected cell. Moreover, a Jiv molecule which has induced one NS2-3 cleavage reaction seems to remain bound to NS2 (Fig.
1) and is most likely not available for further reactions. This is indicated by the results shown in Fig.
6A (left). In noncp BVDV-infected bovine fibroblasts, NS2-3 is translated in the presence of endogenous Jiv in the first 8 h postinfection. When cells were labeled between 8 and 9 h (or later), new (labeled) NS2-3 is produced in those cells; in this time window, only NS2-3 but not NS3 is detected by RIP. This suggests that Jiv molecules present in those cells are not capable of inducing further cleavages; accordingly, recycling of Jiv for cleavage induction does not take place. At the moment, we cannot distinguish whether this is due to the stable binding of Jiv to NS2 (strongly favored by the data presented in Fig.
1 and previously published work on Jiv/NS2 interaction [
25]) or a shortened half-life of Jiv bound to NS2.
In naive bovine fibroblasts, detectable NS2-3 cleavage is restricted to the first 9 h postinfection (
17; this study). We demonstrate here that in cells with a decreased endogenous Jiv level, NS2-3 cleavage disappears as soon as 7 h postinfection. Overexpression of Jiv led to an impairment of cleavage regulation, and a stable steady-state level of NS2-3 cleavage was observed as soon as saturating amounts of NS2-3 were available in the infected cell. Accordingly, alterations in the intracellular Jiv level strictly correlated with changes in viral RNA replication efficiency, underlining the biological significance of NS2 protease regulation by Jiv. Our data strongly suggest that the intracellular Jiv level restricts the replication of noncp BVDV by limiting the generation of NS3 and thereby the number of active replicase complexes formed in the infected cell.
Several viral proteases, like the NS3 proteases of HCV, pestiviruses, and flaviviruses, depend in their activity on cofactors, namely, NS4A of HCV and pestiviruses or NS2B of flaviviruses (
18). However, in contrast to Jiv, these cofactors are virus-encoded peptides. An advantage of a cellular cofactor could be a tight coupling of viral replication to its host cell environment and a possible determination of viral tissue tropism. Our data suggest that Jiv is an important and possibly essential host factor for noncp BVDV and other pestiviruses (this report; A. Müller and N. Tautz, unpublished data). Downregulation of the Jiv level in host cells of the virus led to a significant decrease in viral RNA replication and progeny virus production. It appears likely that replication of noncp BVDV will not occur in cells with very low or no Jiv expression. Preliminary real-time RT-PCR experiments using MDKB and PT cells did not reveal a correlation between noncp BVDV infection and changes in the cellular Jiv mRNA level. Only very limited data are presently available with respect to the expression of Jiv in different bovine cell types (
23); accordingly, the significance of Jiv expression for tissue tropism of the virus in the host has yet to be established. However, our experiments clearly show that the low endogenous amount of Jiv present in bovine cells limits RNA replication of noncp BVDV; when this limitation is disturbed, e.g., in cell lines overexpressing Jiv, viral replication is strongly upregulated (this report), which correlates with a switch of the viral phenotype from noncp to cp (
25). Another intriguing aspect in this context is the presence of Jiv-coding sequences in the genomes of several cp pestiviruses (
21,
22). These viruses have acquired cell-derived Jiv-coding sequences by RNA recombination. Expression of the cell-derived protease cofactor Jiv from the viral genome deregulates NS3 generation, renders these mutant viruses cytopathogenic, and correlates with the progression from persistent infection to lethal disease in cattle (
19,
21,
22). In the BVDV system, the maintenance of the noncp biotype is of high significance, since only noncp strains can establish lifelong persistent infections; such animals continuously shed the virus and represent its major reservoir. In contrast, cp BVDV strains have lost the ability to persist in the animal and from the epidemiological point of view may be regarded as dead-end products (
20). Accumulating evidence suggests that the regulation of NS2-3 processing is the key factor for the biotype control of BVDV. While NS2-3 cleavage is strictly downregulated early after infection in noncp BVDV-infected cells, all cp BVDV isolates have lost this capacity (
17,
18,
20,
21). Only one cp BVDV-derived mutant selected in cell culture by Qu and coworkers (
24) displayed a deregulated NS2-3 cleavage in spite of its noncp biotype, but it is unknown yet whether this virus is capable of establishing persistent infection in the animal.
According to the data presented in this study, limiting amounts of the cellular cofactor Jiv represent the molecular basis for the temporal downregulation of NS2-3 cleavage and viral RNA replication in noncp BVDV-infected cells. This regulation is crucial for maintenance of the noncp biotype of this virus which in turn is essential for its capacity to establish persistent infections. The unique limitation of viral RNA replication by a low abundant cellular protein serving as a cofactor of a viral protease is thus part of a strategy which allows the virus to establish lifelong persistent infections. The results obtained in this study provide an entirely novel aspect to the understanding of the molecular basis of viral persistence in the host.