|Home | About | Journals | Submit | Contact Us | Français|
The functions of the alphavirus-encoded nonstructural protein nsP3 during infection are poorly understood. In contrast, nsP1, nsP2, and nsP4 have known enzymatic activities and functions. A functional analysis of the C-terminal region of nsP3 of Semliki Forest virus revealed the presence of a degradation signal that overlaps with a sequence element located between nsP3 and nsP4 that is required for proteolytic processing. This element was responsible for the short half-life (1 h) of individually expressed nsP3, and it also was functionally transferable to other proteins. Inducible cell lines were used to express native nsP3 or truncated mutants. The removal of 10 C-terminal amino acid (aa) residues from nsP3 increased the half-life of the protein approximately 8-fold. While the deletion of 30 C-terminal aa residues resulted in a similar stabilization, this deletion also changed the cellular localization of nsP3. This truncated mutant no longer exhibited a punctate localization in the cytoplasm, but instead filamentous stretches could be formed around the nuclei of induced cells, suggesting the existence of an additional functional element upstream of the degradation signal. C-terminally truncated uncleavable polyprotein P12CA3del30 was localized diffusely, which is in contrast to P12CA3, which is known to be associated with vesicle membranes. The induction of nsP3 or its truncated forms reduced the efficiency of virus multiplication in corresponding cells by affecting different steps of the infection cycle. The expression of nsP3 or a mutant lacking the 10 C-terminal aa residues repressed the establishment of infection, while the expression of nsP3 lacking 30 C-terminal aa residues led to the reduced synthesis of subgenomic RNA.
Semliki Forest virus (SFV) is a member of the genus Alphavirus in the family Togaviridae. The alphavirus genome is a positive-stranded RNA molecule that is approximately 11.5 kb in length. Four nonstructural (ns) proteins (nsP1 to nsP4) are encoded by the 5′ region, and these proteins are translated directly from the genomic RNA. Genes that encode the structural proteins are located in the 3′ region of the genome, and their corresponding proteins are translated from a subgenomic (SG) mRNA generated from an internal promoter located on the complementary minus-strand template (43). In the case of SFV, the ns proteins initially are produced as a single, large P1234 polyprotein, which is processed into its final products in a well-controlled manner (33, 54). The polyprotein-processing intermediate, P123 together with nsP4, is responsible for the synthesis of minus-strand RNA, which occurs during an early stage of infection and is halted approximately 3 to 4 h postinfection. Completely processed ns proteins form mature replication complexes and are active only in synthesizing positive-strand genomic RNAs and SG-mRNAs from the minus-strand RNA template (27, 42, 43, 57). Not all synthesized ns proteins are included in replicase complexes. The fraction of ns proteins that fail to be included are distributed to different locations within the infected cells (22, 41). Although their functions are not related to the replication of viral RNA and are less understood, they also are thought to play important roles in viral infection. For example, individual nsP1 has been shown to induce the formation of filopodium-like structures on the plasma membrane of infected cells (24, 41). In its free form, nsP2 is responsible for inducing cytotoxic effects, such as the shutdown of cellular transcription (8, 13, 15), and also is important for the suppression of antiviral responses (10, 16).
The replication complexes of alphaviruses are attached to the membranes of modified endosomes and lysosomes. The membranes of these virus-modified organelles contain invaginations (or spherules) with diameters of approximately 50 nm (see reference 22 and references therein). Each spherule represents a site of viral RNA synthesis. nsP1 is the sole membrane anchor of the replication complex (2, 11, 41). nsP2 is a papain-like protease that is responsible for P1234 processing (33, 43, 54) and possesses NTPase (39) and RNA helicase activity (14). Both nsP1 and nsP2 are involved in capping viral genomic and SG-mRNAs (1, 2, 23, 53). nsP4 represents the catalytic subunit of the viral RNA-dependent RNA polymerase (21, 40, 48). Although purified nsP4 is capable of synthesizing the minus strands of the viral genome, the formation of its active conformation requires the presence of P123 (40, 48).
The alphavirus nsP3 is composed of three domains. The first domain, which has been designated the macrodomain or X domain, is conserved among alphaviruses, coronaviruses, the hepatitis E virus, and the rubella virus, and it also is found in the proteins of different cellular organisms (21, 37). The three-dimensional structure of this domain has been solved for coronaviruses (7) and for several alphaviruses (32). The only known enzymatic activity of alphaviral nsP3, which involves the activity of an adenosine di-phosphoribose 1′-phosphate phosphatase, is associated with a macrodomain (32). Whether or not the nsP3s of all alphaviruses exhibit this activity remains unknown. Whereas the activity was clearly demonstrated for the macrodomains of the Chikungunya virus and the Venezuelan equine encephalitis virus (32), the enzymatic activity of the SFV macrodomain, if any, was below the limit of detection (7). The second domain is conserved only among the nsP3s of alphaviruses (43 and our unpublished analysis). The third domain, which is located at the C terminus, is hypervariable in both length and sequence and is dispensable for replication (12, 25, 43). nsP3 is required for the correct formation and localization of replication complexes (41). Its involvement in RNA synthesis has been deduced from experiments utilizing temperature-sensitive mutants. nsP3 participates in SG-mRNA and minus-strand RNA synthesis, either independently or as part of the polyprotein P23 (18, 25, 26, 27, 42, 57); however, its exact role in these processes is not well understood. SFV also might harbor some virulence determinants in the conserved and nonconserved regions of nsP3 (47, 49, 50). nsP3 is phosphorylated at multiple serine and threonine residues that are located mainly in the hypervariable domain (55, 56). The importance of such modifications is not fully understood, but in the case of Sindbis virus (SIN; the prototype member of genus Alphavirus), the phosphorylation of nsP3 potentially modulates the efficiency of minus-strand RNA synthesis (4).
Several alphaviruses express two different isoforms of nsP3. In SIN, a shorter isoform of nsP3 (545 amino acid [aa] residues) is produced due to termination at an in-frame stop codon located proximally to the end of the nsP3 coding sequence. A longer isoform (551 aa residues) is produced by the readthrough of a termination codon and the processing of the resulting polyprotein between nsP3 and nsP4 (43). In contrast to SIN, most SFV isolates, including the laboratory strain SFV4, possess only one form of nsP3 (482 aa residues for SFV4), whose counterpart is the longer isoform of SIN nsP3. To date, the possibility that different isoforms of nsP3 have different properties or functions has not been studied.
Although the exact stoichiometry of ns proteins in an alphavirus replication complex is unknown, information about the cellular factors involved in replication and/or interactions with the replication complex is just beginning to emerge. A number of cellular proteins that bind to SIN nsP3 have been identified, including the 14-3-3 proteins, G3BP, and different hnRNP proteins. However, the functional importance of these proteins in SIN infection generally is uncharacterized (3, 9). The macrodomain of nsP3 binds RNA and ADP-ribose (7, 32), and the variable C-terminal domain of nsP3 has been shown to bind poly(ADP-ribose) polymerase-1 in SIN-infected cells (36). The significance of these functions for alphavirus replication also is poorly understood.
Different types of nsP3-containing complexes in SIN-infected cells recently have been described (17). Complexes of one type interacted with the plasma membrane and endosomes, while complexes of another type were tightly bound to nuclei. In addition to their different localizations, these complexes differed from each other with regard to their cellular protein content and, most importantly, the presence of double-stranded viral RNA, which indicates their different roles in SIN infection. Interestingly, complexes formed by nsP3 from SFV4 also have been detected in cells that transiently express the protein (38). In contrast to the structures formed by P123 or those observed in SFV4-infected cells, these inclusions do not contain membranes and represent amorphic aggregates of nsP3 proteins (41).
We previously employed a tetracycline-inducible HEK293 T-REx cell line-based system to study the properties of nsP1 from SFV4 (19). Using a similar approach, we found that nsP3 from SFV4 is unstable when expressed alone. The element responsible for its short half-life was mapped to the C-terminal region of nsP3 and overlaps with the nsP3 portion of the processing site between nsP3 and nsP4 (hereafter designated the P side of the 3/4 processing site). This site also contains the major determinants of 3/4 site processing (31). The removal of this element increased the stability of nsP3 approximately 10-fold. Conversely, the fusion of the C terminus of nsP3 from SFV4 to heterologous proteins resulted in their increased degradation. The C terminus of the longer isoform of nsP3 from SIN also displayed a similar property. The analysis of the functional significance of nsP3 in the context of SFV4 infection revealed that the presence of nsP3 and its more stable deletion mutants affected different stages of SFV4 multiplication.
A ubiquitin fusion technique (52) was used to obtain nsP3 with a native N-terminal alanine residue. The pcDNA4/TO (Invitrogen) vector was used for foreign protein expression. Details of the construction and sequences of all plasmids are available from the authors upon request. The inducible cell lines used for the expression of nsP3, nsP3del10, and nsP3del30 (designated T-REx-nsP3, T-REx-nsP3del10, and T-Rex-nsP3del30, respectively) were constructed and propagated as described previously (19). Cells were grown at 37°C in 5% CO2 in Iscove's modified Dulbecco's medium (IMDM; Gibco) supplemented with 10% fetal bovine serum, blasticidin (5 μg/ml), and zeocin (30 μg/ml). Viral protein expression was induced by the addition of tetracycline, at a final concentration of 1 μg/ml, to the growth medium.
SFV4 was generated from an infectious cDNA clone (30), and SFV4 mutants were constructed and produced from the infectious plasmid pCMV-SFV4 (51). BHK-21 cells were used for the plaque titration of the virus stocks and to generate VRPs (virus replicon particles) using an SFV-based replicon vector that expresses enhanced green fluorescent protein (EGFP) under the control of the SG promoter (58). The susceptibilities of T-REx-nsP3, T-REx-nsP3del10, and T-Rex-nsP3del30 cells to SFV infection were analyzed by infecting noninduced cells with titrated VRP stocks. To analyze virus multiplication, T-REx-nsP3, T-REx-nsP3del10, and T-Rex-nsP3del30 cells were grown in 35-mm dishes and mock induced or induced with tetracycline 18 h prior to infection with SFV4 at an MOI (multiplicity of infection) of 0.1. At selected time points, the supernatants were collected and the titers of released virus determined. To analyze the efficiency of established SFV infection, induced or mock-induced T-REx-nsP3, T-REx-nsP3del10, and T-Rex-nsP3del30 cells were infected with SFV VRPs at an MOI of 0.3, and the percentage of EGFP-positive cells was analyzed 18 h postinfection using an LSR II flow cytometer (BD Biosciences). The multiplication of SFV4 and its mutants was analyzed in BHK-21 cells. These cells were grown in 100-mm dishes and infected with the corresponding virus at an MOI of 1, and the virus stocks were collected 48 h later. Viral titers were obtained by plaque assay.
For immunofluorescence microscopy, HeLa cells were grown on coverslips in 35-mm dishes until 50% confluence and then transfected with 5 μg pcDNA4/TO vector encoding viral polyprotein P123, P12CA3, P123del10, P12CA3del10, P123del30, or P12CA3del30 using Lipofectamine 2000 (Invitrogen) reagent. CA indicates a cysteine-to-alanine mutation in the catalytic center of nsP2, which results in the inactivation of its protease activity and, consequently, in nonprocessed polyproteins. The control cells were mock transfected. To determine the half-life of nsP3 that has been produced through the processing of P123, T-REx cells were grown in 35-mm dishes until 75% confluence and then transfected with 5 μg plasmid encoding P123 using Lipofectamine LTX (Invitrogen) reagent. Tetracycline, at a final concentration of 1 μg/ml, also was added to the growth medium.
T-REx-nsP3, T-REx-nsP3del10, and T-REx-nsPdel30 cells were grown in 35-mm dishes to 75% confluence and induced with tetracycline for 18 h. T-REx cells transfected with plasmid encoding P123 of SFV were analyzed at 18 h posttransfection. The induced or transfected cells subsequently were incubated for 45 min in methionine- and cysteine-free medium, followed by a pulse with 50 μCi (1.85 MBq) [35S]methionine and [35S]cysteine (PerkinElmer) for 45 min. If needed, the proteasome inhibitor MG-132 (Calbiochem), at a final concentration of 10 μg/ml, was added to the growth medium. The half-life of nsP3 was evaluated in SFV4-infected T-REx cells 2 h postinfection by incubating the cells for 30 min in methionine- and cysteine-free medium, followed by a pulse with 50 μCi (1.85 MBq) [35S]methionine and [35S]cysteine for 30 min. Pulse-chase experiments and the immunoprecipitation of the labeled samples were carried out as described previously (19, 31). X-ray film or a Typhoon Trio (Amersham Biosciences) instrument was utilized for the detection of the labeled nsP3.
The procedure employed has been described previously (19). Briefly, for indirect immunofluorescence microscopy, cells were grown on coverslips in 35-mm dishes. At 50% confluence, the cells were induced with tetracycline to analyze the localization pattern of nsP3 or its mutants in the cell lines. To analyze the localization of viral polyproteins or their separate components, cells were transfected as described above. At selected time points, cells were fixed, permeabilized, and treated with rabbit polyclonal nsP3-specific (and with guinea pig polyclonal nsP1-specific for P123-expressing cells) antiserum. The primary antibody incubation was followed by treatment with anti-rabbit antibodies conjugated to Alexa Fluor 488 and with anti-guinea pig antibodies conjugated with tetramethyl rhodamine isocyanate (TRITC). Nuclei were stained with 4′,6′-diamidino-2-phenylindole (DAPI). Samples were analyzed using a Zeiss Axiovert 200 M or Olympus FV1000 microscope.
Lysates prepared from equal amounts (~105 cells per lane) of induced or mock-induced cells and were separated by SDS-PAGE. Proteins were transferred to nitrocellulose membranes and detected using antisera against nsP3 (in-house) and enhanced chemiluminescence reagents, as described previously (46). An antibody against β-actin was used as a control.
T-REx-nsP3, T-REx-nsP3del10, and T-REx-nsP3del30 cells were seeded in 96-well plates (9 × 104 cells per well) and allowed to grow for 24 h. Recombinant protein expression was induced for 18 or 42 h, and control cells remained uninduced. Subsequently, MTT reagent [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrasodium bromide; final concentration, 0.5 mg/ml; Calbiochem] was added to each well, and the plate was incubated for an additional 2 h under cell culture growth conditions. The growth medium then was removed, and 200 μl of dimethylsulfoxide (DMSO) was added. The A540 was obtained using a microplate reader.
T-REx-nsP3, T-REx-nsP3del10, and T-REx-nsP3del30 cells grown in 35-mm dishes were induced with tetracycline or mock induced for 18 h prior to infection with SFV4 at an MOI of 0.2. At selected times postinfection, total RNA from infected cells was isolated using TRIzol reagent (Invitrogen) and analyzed by Northern blotting, as described previously (46). An RNA probe complementary to β-actin mRNA was used as a control.
pcDNA4/TO vectors encoding EGFP or variants of EGFP fused to C-terminal fragments of nsP3 from SFV4 were electroporated into BHK-21 cells (5 μg of plasmid per 5 × 106 cells). Electroporated cells were split equally between four 35-mm plates. After 18 h, cellular proteins were pulse labeled with 50 μCi (1.85 MBq) [35S]methionine and [35S]cysteine for 1 h, followed by 2-, 6-, and 24-h chases. Immunoprecipitation of the labeled samples was carried out as described previously (31, 46). pcDNA4/TO plasmid encoding firefly luciferase (Luc) or its chimeric variants containing C-terminal fragments of nsP3 was electroporated into BHK-21 cells (200 ng per 2 × 106 cells). Electroporated cells were split equally between 6 wells of a 24-well plate. After 18 h, the translation inhibitor cycloheximide, at a final concentration of 100 μg/ml, was added. At selected time points, the cells were harvested and analyzed using the Dual-Luciferase Reporter Assay kit (Promega). Renilla luciferase activity produced via coelectroporation of pRL-TK (Promega) was used to monitor the efficiency of transfection.
The nsP3 sequences of the different alphaviruses were obtained from the National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov). Multiple sequence analysis (MSA) was performed by using the t-coffee (34) web server (http://www.tcoffee.org). For predicting nsP3 secondary structures, different programs listed on the ExPASy proteomics server (http://www.expasy.ch/tools/) were used.
The HEK293 T-REx-based stable cell line T-REx-nsP3, which can be induced to express SFV4 nsP3 with a native N-terminal alanine residue, was used to analyze the stability of nsP3. nsP3 that was expressed by itself was rapidly degraded (Fig. 1A and E), and the quantification of radiolabeled nsP3 in pulse-chase experiments revealed that the half-life of the protein is approximately 1 h. The analysis of the stability of nsP3 in the presence of the proteasome inhibitor MG-132 revealed that even though the inhibitor did not prevent the degradation of nsP3 completely, its presence considerably increased the stability of nsP3, which indicated that proteasomes are involved in its rapid degradation (Fig. 1B and E). These results were surprising since, in contrast to unstable nsP4 (6), nsP3 is considered to be a stable protein in SFV4-infected cells. We hypothesized that, during SFV4 infection, the instability of nsP3 went unnoticed due to its specific subcellular localization and/or stabilization by other viral and/or cellular factors. To test this hypothesis, HEK293 T-REx cells were infected with SFV4 at an MOI of 20 and metabolic labeling was initiated 2 h postinfection, followed by a pulse-chase experiment. The greatest amount of nsP3 was detected after a 1-h chase. The obvious source of such a peak is polyprotein processing, which produces individual nsP3 from its polyprotein precursors (Fig. 1C and E). Importantly, a rapid decrease in the amount of radiolabeled nsP3 was observed in samples chased for 1 and 3 h, which corresponds to 4 and 6 h postinfection, respectively (Fig. 1C and E). These data are concordant with the hypothesis that a significant fraction of nsP3, which is produced by the processing of polyprotein precursors, is rapidly degraded in infected cells. The fraction of nsP3 that escapes such degradation, however, becomes relatively stable, which suggests the existence of factors that are responsible for the stabilization of nsP3 in infected cells (Fig. 1C and E).
In SFV4-infected cells, a significant portion of nsP3 molecules that are synthesized most likely are incorporated into replication or other complexes (17, 41), which leads to their stabilization. A similar phenomenon has been described for nsP4 of alphaviruses (6). To determine whether the stabilization of nsP3 resulted from expression in the form of polyprotein and/or from viral replication, nsP3 levels from the P123 polyprotein were analyzed. Our repeated attempts to generate cell lines with stably expressed or inducible P123 were unsuccessful, and this most likely resulted from the toxic effects of nsP2, which also was released from expressed P123. A small leakage of the Tet-inducible promoter, which resulted in P123 expression, was detected even in T-REx cells. Therefore, T-REx cells transfected with an expression plasmid encoding the P123 polyprotein were pulse-chased 18 h posttransfection. Results show that nsP3 was indeed remarkably stabilized, possibly by the interactions with other products of P123 processing, nsP1 and/or nsP2. This experiment also revealed that virus replication was unnecessary for nsP3 stabilization (Fig. 1D and E).
The N-terminal alanine of nsP3 from SFV4 is a destabilizing residue (52) and may represent one factor that is involved in the rapid degradation of nsP3. This assumption is supported by the fact that nsP3 degradation is reduced in the presence of a proteasome inhibitor. The situation with the degradation of nsP3 is, at least partially, similar to that observed for nsP4, in which a destabilizing N-terminal tyrosine residue is thought to be responsible for its instability (6). The addition of a peptide corresponding to the last 30 residues of nsP3 from SFV4 to the C terminus of EGFP has been reported to result in the reduced half-life of the recombinant protein. This reduction also occurs when the protein has a stabilizing N-terminal glycine residue, which suggests that the C-terminal region contains a key element that is responsible for the rapid degradation of nsP3 (45). To verify this hypothesis, fragments corresponding to the last 6, 10, or 30 aa residues of SFV4 nsP3 were used to construct chimeric luciferase reporters, designated Luc6, Luc10, and Luc30, respectively (Fig. (Fig.2A).2A). Since no information about the three-dimensional structure of the nonconserved region of nsP3 was available and different secondary structure prediction programs provided contradictory results, the sizes of the C-terminal fragments were selected based on data from different functional studies (31, 45). Following the cycloheximide-induced inhibition of protein synthesis, cells containing chimeric reporters displayed a more rapid decrease in luciferase activity than cells containing the native Luc (Fig. (Fig.2B).2B). The half-life of Luc was reduced from approximately 5 to 2 h by all three C-terminal fragments of nsP3. The rate of reporter degradation was not increased further if Luc was fused to longer C-terminal fragments of SFV4 nsP3 (data not shown). Thus, 6 to 10 C-terminal aa residues of nsP3 from SFV4 were sufficient to accelerate the degradation of a Luc reporter.
To verify these data, a peptide entitled 83del10, corresponding to residues 400 to 473 of SFV4 nsP3, was fused to the C terminus of Luc (Fig. (Fig.2A).2A). Subsequent analysis revealed that the cycloheximide treatment of Luc83del10-expressing cells did not result in a significantly faster decrease of luciferase activity than that observed in cells expressing native Luc (Fig. (Fig.2B).2B). This result confirmed that the most important functional part of the degradation element was indeed located at the very end of the SFV4 nsP3 C terminus.
Since there is a stretch of sequence in the C-terminal one-third of alphavirus nsP3 that is not conserved, whether the degradation element is a unique property of SFV4 nsP3 or is common to alphaviruses cannot be predicted. To further investigate this degradation element, chimeric reporters containing C-terminal fragments of two different nsP3 isoforms from SIN were constructed (Fig. (Fig.2A)2A) and used for the analysis (designated LucShort30 and LucLong36; in the latter, the leaky termination codon was replaced with a cysteine codon). Only the fusion of the Long36 element from SIN nsP3 with Luc caused a more pronounced reduction in luciferase activity (Fig. (Fig.2C).2C). The fusion of the Short30 element produced a far less prominent effect and was detectable only 6 to 8 h after the addition of cycloheximide (Fig. (Fig.2C).2C). Thus, nsP3 of SIN also contained a degradation element, which was similar to that of SFV4, and the residues immediately upstream of the 3/4 processing site were crucial for its activity.
To assess if the C-terminal fragments of nsP3 from SFV4 and SIN can enhance the degradation of other proteins, chimeric EGFP reporters were constructed and analyzed using pulse-chase experiments. This analysis (Fig. 3A and B) revealed that the fusion of 30 C-terminal residues from SFV4 nsP3 to the C terminus of EGFP indeed reduced the stability of EGFP. Elevated degradation rates also were observed for the EGFP6 and EGFP10 (containing 6 and 10 C-terminal aa residues of SFV4 nsP3, respectively) reporters, although the effect was less prominent (Fig. (Fig.3B).3B). Similarly to these and other results, analyses of the stabilities of EGFPShort30 and EGFPLong36, which are analogous to LucShort30 and LucLong36 (Fig. (Fig.2A),2A), revealed that the fusion of the Long36 peptide to EGFP reduced the stability of the EGFP molecule, whereas the fusion of the Short30 element had a lesser effect (Fig. 3A and B).
The overlap of the mapped degradation signal with sequences that are required for the first event in P1234 processing did not allow for an analysis of the effects of nsP3 C-terminal deletions using recombinant SFV4 genomes; therefore, the HEK293 T-REx-based cell lines T-REx-nsP3del10 and T-REx-nsP3del30 (expressing nsP3 without the last 10 or 30 C-terminal aa residues, respectively) were constructed and utilized. T-REx cells, infected with SFV4 at an MOI of 5 and lysed 6 h postinfection, were used to obtain control lysates for the comparison of tetracycline-induced expression levels of nsP3, nsP3del10, and nsP3del30 with those in infected cells. The nsP3 deletion mutants were present in induced cells at levels higher than those of wild-type nsP3 (Fig. (Fig.4A),4A), and similar results were obtained by using a number of independent cell lines (not shown). Therefore, the difference in the expression levels of recombinant protein could not be attributed to clonal differences. Instead, data obtained in previous experiments suggested that smaller quantities of native nsP3 result from a more rapid degradation of the protein. To verify this possibility, the stabilities of nsP3del10 and nsPdel30 were determined by pulse-chase experiments. The removal of 10 or 30 C-terminal aa residues from nsP3 indeed resulted in a significant stabilization of the expressed protein (compare Fig. 4B and C to Fig. 1A and E). The quantification of the corresponding radiolabeled proteins revealed that the half-life of nsP3del10 and nsP3del30 was approximately 8 h, which contrasts with the 1-h half-life for native nsP3.
The expression of nsP3 or the C-terminal mutants had no significant effect on cell viability, as determined by an MTT assay (data not shown), indicating that none of these proteins were cytotoxic. The analysis of the subcellular localization of the induced proteins revealed that both nsP3 and nsP3del10 exhibited a punctate localization, with a fraction of the induced protein localized diffusely throughout the cytoplasm (Fig. 5A and B). This distribution was very similar to that observed in cells transiently expressing SFV4 nsP3 (41) and indicates the presence of cytoplasmic aggregates formed by induced proteins. In contrast, nsP3del30 formed filamentous stretches surrounding the cell nucleus. These stretches were detected in all induced cells, but their visibility depended on the focal plane (Fig. (Fig.5C).5C). Neither nsP3, nsP3del10, nor nsP3del30 colocalized with actin or tubulin fibers (data not shown). The dots of nsP3 and nsP3del10 as well as stretches of nsP3del30, however, did colocalize with G3BP (data not shown), which is a protein previously found in complexes formed by nsP3 from SIN (3, 9, 17). Thus, the punctate localization and aggregate formation of individual nsP3 requires the presence of residues 453 to 473.
Given that nsP3 is expressed in infected cells as part of an ns polyprotein, the effects of its C-terminal deletions also were assayed in the context of the P123 polyprotein. Since we were unable to obtain corresponding inducible cell lines, transiently transfected HeLa cells, which have more cytoplasmic volume than HEK293 T-REx cells, were used for this analysis. Expression plasmids encoding P123, P123del10, and P123del30 were used to produce solitary nsP3, nsP3del10, or nsP3del30 from their polyprotein precursors, respectively. In all cases, nsP1 that was not colocalizing with nsP3 was found at the plasma membrane of transfected cells, where it induced the formation of filopodium-like structures (Fig. 6A to C). nsP3 and nsP3del10 both were localized diffusively, but they also could form aggregates (compare Fig. 6A and B to Fig. 5A and B), as in the case of individually expressed proteins. On the other hand, nsP3del30 was scattered throughout the cytoplasm and, although no filamentous stretches were found, nsP3del30 still could be found in the perinuclear space. Thus, the context of nsP3, nsP3del10, or nsP3del30 production did not significantly affect their subcellular localization. The localization patterns of the recombinant proteins were further studied by the transient expression of uncleavable polyproteins P12CA3, P12CA3del10, and P12CAdel30 in HeLa cells. Polyprotein P12CA3 is known to localize to the membranes of vesicles (41), mimicking the localization of the viral replication complex, which was found to be the case in our study (Fig. (Fig.6D).6D). A similar localization also was characteristic of P12CA3del10 (Fig. (Fig.6E).6E). In contrast, P12CA3del30 did not form any discrete structures, but rather it was dispersed randomly in the cells (Fig. (Fig.6F).6F). Taken together, these data suggest the presence of at least one additional functional element in the C-terminal region of SFV4 nsP3.
The lack of cytotoxic effects following nsP3, nsP3del10, or nsP3del30 expression permitted the analysis of the effects of cellular nsP3 on SFV4 infection. The comparison of SFV4 growth curves in induced or mock-induced T-REx-nsP3 and T-REx-nsP3del10 cells revealed that the induction of nsP3 expression had only moderate and temporary effects on the multiplication of SFV4 (Fig. (Fig.7A).7A). The induction of nsP3del10 expression resulted in an approximately 10-fold reduction of released virus titer at all assessed time points (Fig. (Fig.7B).7B). The more prominent and persistent effect of nsP3del10 expression could result from its longer half-life. This factor may be important, since the SFV4-induced shutdown of cellular transcription and translation prevented the expression of cellular nsP3s during late stages of infection. Initial high levels of nsP3del10 (Fig. (Fig.4A)4A) and/or its possible dominant-negative effects on SFV4 replication represent alternative explanations. A similar analysis revealed that the expression of nsP3del30 caused an approximately 5-fold reduction of released virus at all time points (Fig. (Fig.7C).7C). Different effects resulting from the expression of nsP3del10 and nsP3del30, which possessed similar stabilities (Fig. 5B and C), could reflect different influences on SFV4 multiplication.
The inhibition of SFV4 multiplication in induced T-REx-nsP3, T-REx-nsP3del10, and T-REx-nsP3del30 cells could result from multiple factors, such as a reduced number of infected cells, reduced levels of viral RNA, and/or reduced protein synthesis. The EGFP-expressing SFV replicon vector, which is incapable of spreading in infected cell cultures, was used to investigate whether the expression of nsP3 or nsP3 C-terminal deletion mutants could affect the number of successfully infected cells. The percentage of infected (EGFP-positive) cells was reduced 1.5-fold following the induction of nsP3 expression. The expression of nsP3del10 caused an even more prominent, 3-fold decrease in the number of infected cells (Fig. (Fig.7D).7D). In contrast, the expression of nsP3del30 only had a subtle effect. Thus, the inhibition of the establishment of SFV4 infection depended on the presence of the nsP3 region located between residues 453 and 473 as well as on the level of expression and/or the stability of the recombinant protein.
The reduction in the number of primarily infected cells may represent the main reason for the modest, temporary inhibition of SFV4 multiplication in induced T-REx-nsP3 cells. However, as was especially evident for T-REx-nsP3del30 cells, this hardly could represent the sole reason for the inhibition of SFV4 multiplication by nsP3 deletion variants. Thus, other effects of these proteins might have contributed to the observed interference with SFV4 multiplication. Therefore, the possible role of cellular nsP3 on viral RNA synthesis was analyzed by Northern blotting. This assay revealed that the induction of native nsP3 slightly reduced the accumulation of viral genomic and SG-mRNAs. In the case of nsP3del10 expression, the accumulation of genomic and SG-mRNAs was both delayed and strongly reduced (Fig. (Fig.8A).8A). It is not clear whether the increased effect of nsP3del10 was due to functional interference with SFV4 replication or merely represented a consequence of its greater stability and/or increased levels. Again, the effect caused by nsP3del30 expression was different. The synthesis of SG-mRNA was specifically reduced, resulting in an approximately 2-fold change in the genomic RNA:SG-mRNA ratio at all evaluated time points (Fig. (Fig.8B).8B). The decreased SG-mRNA level could result in the less efficient multiplication of SFV4 in the presence of nsP3del30, since this would decrease the expression of the structural protein and, accordingly, decrease the production and release of viral particles.
The analysis of regions overlapping with identified functional elements was carried out using the MSA software t-coffee (34). Initial analysis revealed that nsP3 of Venezuelan, Eastern, and Western equine encephalitis viruses can be separated from the rest of the group and were excluded from any further analysis (data not shown). For the rest of the viruses it was confirmed that although most of this C-terminal one-third of nsP3 is highly divergent, the conserved structural elements do exist at the very end of nsP3 (Fig. (Fig.9A).9A). Concerning the nsP3 of SFV4, two highly similar elements, LTFGDFD and ITFGDFD, were found; these are separated by a 10-aa-long nonconserved stretch. It also was found that other alphaviruses possess similar elements at the end of nsP3 (Fig. (Fig.9A).9A). The found elements were used as a query against different protein motif databases, but no significant hits were obtained (data not shown).
In the case of nsP3del10, both elements still were present, but only a short fragment (LFTG) of the first element was preserved in nsP3del30 (Fig. (Fig.9B).9B). To investigate the putative role of these elements during SFV4 infection, mutant virus SFVdel30-11 was constructed (Fig. (Fig.9B).9B). A second mutant, in which only the segment between these two elements was removed, also was engineered and named SFVdel8 (Fig. (Fig.9B).9B). Both mutant viruses were viable but differed in their multiplication efficiency. The final titer of SFVdel8, obtained with a single-cycle infection experiment, was approximately 8 × 109 PFU/ml, which is almost identical to what is obtained with SFV4 (approximately 7 × 109 PFU/ml) (Fig. (Fig.9C).9C). In contrast, the titer for SFVdel30-11 was significantly lower, approximately 9 × 107 PFU/ml (Fig. (Fig.9C).9C). This result indicates that the removal of the conserved sequence elements drastically reduces the multiplication of SFV4, thus confirming their functional importance.
A significant fraction of ns proteins expressed during the early stages of alphavirus infection are not assembled into functional replicase complexes (22) and exist in infected cells as solitary proteins. Several functions have been assigned to nsP2 (8, 10, 13, 15, 16) and nsP1 (2, 19, 41), which are known to be stable proteins (19). In contrast, the nsP3 C-terminal element with a maximum length of 10 aa residues caused the instability of independently expressed SFV4 nsP3. An element with a similar function was present in the longer isoform of SIN nsP3. Thus, similarly to nsP4, nsP3 represents a conditionally unstable replicase protein in alphaviruses.
The fact that SIN, and possibly several other alphaviruses, express nsP3 isoforms with different stabilities indicates that these two isoforms have different functions during viral infection. Presumably, complexes engaged in SIN replication should contain the longer nsP3 isoform. The role, if any, of the shorter nsP3 isoform is less obvious, especially given that the presence of a short nsP3 isoform is not crucial for SIN infection (17, 20, 44). SFV also tolerates the leaky terminator at the analogous position of the nsP3 coding region (20, 50). In the case of both viruses, however, these mutations do result in changes in RNA synthesis. These changes could be attributed to the drastic differences in nsP4 expression levels, but we cannot exclude the possibility that the altered amount of nsP3 containing a degradation signal also represents an important, and controlling, factor that is required for optimal replication efficiency.
In contrast to the known instability of nsP4, the rapid degradation of nsP3 has not been described previously, most likely because the instability of nsP3 is more difficult to detect in alphavirus-infected cells. In this study, however, we were able to detect a rapid initial decline of the amounts of labeled nsP3 in SFV4-infected cells (Fig. (Fig.1C,1C, ,1E).1E). Only after a chase time of >3 h did the remaining amount of nsP3 become relatively stable, while no such stabilization was observed for independently expressed nsP3 in T-REx-nsP3 cells. This finding indicates that SFV nsP3 could be stabilized by components of the viral replicase machinery or, alternatively, that alphavirus replication can inhibit nsP3 degradation. Our finding that nsP3 expressed in the form of P123 is much more stable clearly favors the first possibility. In addition, EGFP fused to the 30 C-terminal aa residues of nsP3 was rapidly degraded in SFV4-infected cells. This accelerated degradation was observed only when the chimeric protein was separated from the replicase complex of SFV4 (45). Although replication does not block the function of the degradation signal, nsP3 must be either released from or not incorporated into the replicase complexes in order to be rapidly degraded. Recently, a study was carried out in neurons to investigate the functional importance of the SIN nsP3 macrodomain. Specific mutations in the macrodomain affected the neurovirulence and, curiously, the stability of nsP3, which further supports our claims that nsP3 could be degraded rapidly in different cells (35).
Our findings gave rise to an additional question: what functional significance might this phenomenon have? The rapid degradation of nsP3 synthesized at early (3 h postinfection) stages of replication was observed (Fig. 1C and E), but no enhanced degradation was observed during the later stages of infection (8 h postinfection; data not shown). Similarly, in the case of SIN, free nsP3 containing a C-terminal degradation signal can be produced exclusively during an early stage of infection, since after the switch of polyprotein processing to the P12 + P34 pathway, the P34 polyprotein is not further processed (5). Thus, unless the degradation signal also is functional when located in an inner position of P34 of SIN, it should function only during early infection. In the case of SFV4, nsP3 containing a degradation signal could be produced until the shutdown of ns protein expression, since P34 of SFV4 can be processed in trans by nsP2 (31, 33). Currently, the specific functions of rapidly degraded free nsP3 can only be proposed; however, the possibility that unstable nsP3 interacts with some host components and accelerates their degradation or inhibits their function represents an attractive hypothesis.
The nsP3 C-terminal degradation signal overlaps with the P side of the 3/4 processing site. However, there is no indication that the P sides of the other two processing sites in the ns polyprotein also function as degradation signals. Furthermore, nsP1, the C terminus of which corresponds to the P side of the 1/2 processing site, is a stable protein (19). The use of a proteasome inhibitor did prevent, but did not completely abolish, the rapid degradation of SFV4 nsP3. Therefore, additional mechanisms could be responsible for facilitating nsP3 degradation. At the beginning of the study, we hypothesized that the degradation signal of nsP3 is responsible for the formation of aggregates of nsP3, which subsequently are degraded in cells. This, however, was not supported by our data that were obtained via the analysis of the subcellular localization of recombinant proteins. nsP3del10 formed structures very similar to those formed by nsP3 (Fig. 5A and B), but its stability was remarkably increased (Fig. (Fig.4B).4B). In contrast, similarly stable nsP3del30 mostly was localized diffusely and formed filamentous stretches surrounding the cell nucleus (Fig. (Fig.5C).5C). SIN nsP3 present in infected cells can be isolated from two different types of complexes. One complex is associated with the plasma membrane and endosomes, and the second type is found in the perinuclear region (17). Whether or not the different localization patterns of nsP3 and nsP3del30 reflect the existence of similar types of complexes in SFV4-infected cells remains to be determined.
The exact composition and nature of structures formed by individual nsP3 of SFV4 and its mutant forms requires better understanding. However, when nsP3 was expressed as a part of uncleaved P12CA3 polyprotein, it formed structures that were very similar to those detected in a previous study (41). The removal of the final 10 aa did not alter the localization pattern of the corresponding polyprotein (Fig. 6D and E). In contrast, this localization was lost in the case of P12CA3del30, which was scattered in the cytoplasm (Fig. (Fig.6F).6F). This observation confirmed the existence of an additional element at the C terminus of SFV nsP3. These findings further support the suggestion that nsP3 plays an important role in the formation of the SFV4 replication complex (41).
The expression of cellular nsP3 or nsP3del10 decreased the number of successfully infected cells, indicating that, in contrast to nsP1 (19), the presence of these proteins in cells reduced the efficiency of the establishment of incoming SFV4 replication. Interestingly, the deletion of an additional 20 residues from the nsP3 C terminus almost completely abolished this effect (Fig. (Fig.8).8). Instead, the expression of nsP3del30 specifically reduced the synthesis of SG-mRNAs, supporting the conclusion that the nsP3 C-terminal domain contains another functional element located just upstream of the degradation signal, which is missing in nsP3del30. Indeed, experiments using temperature-sensitive SIN mutants have suggested a role for nsP3 in SG-mRNA synthesis (25). In addition, the binding of SIN nsP4 to the promoters that are responsible for synthesizing genomic and SG-mRNAs requires the presence of other ns proteins (28, 29). Therefore, nsP3 could be one of the mediators responsible for nsP4 binding to the genomic and/or SG promoter, and the C terminus of the protein may participate in this process. Congruently with this hypothesis, an SFV4 mutant in which residues 431 to 473 of nsP3 were deleted displayed reduced levels of viral RNA synthesis (12). The MSA indicated that the C-terminal tail of nsP3 for several alphaviruses contains conserved sequence elements and that the deletion of the last 30 aa residues removes most of them (Fig. 9A and B). These elements probably are important for the localization of nsP3, and their removal from the infectious virus genome resulted in nearly a hundredfold reduction of the multiplication efficiency (Fig. (Fig.9C).9C). Further analysis of the C-terminal region of nsP3 is required to better understand its functions.
Recent findings regarding the composition of nsP3-formed complexes in SIN-infected cells (17) emphasize the importance of nsP3 in alphavirus-host interactions. Since alphaviruses possess a wide host range and the infection varies from noncytopathogenic in insect cells to highly cytopathogenic in vertebrate cells, the possibility that nsP3 functions in general and the nsP3 C-terminal region in particular may be distinct in different cell types. Indeed, the ability of SFV to replicate in the neurons of adult mice depends on the properties of nsP3 (50, 55). Recently, the C-terminal region of nsP3 has been shown to be involved in the interaction with host proteins in neuronal cells. This interaction may influence the outcome of infection (35, 36). Therefore, the further functional analysis of nsP3 and the nsP3 C-terminal region in different cell types and different hosts represents an important topic for further studies.
We thank Heiti Paves, Martin Pook, and Kaja Kiiver for their help with confocal microscopy, Aleksei Lulla for useful experimental suggestions and comments, Ingrid Tagen for contributing to the construction of the cell lines, Nele Tamberg for contributing to the initiation of the project, and Sirle Saul and Liane Ülper for technical assistance.
This research was supported by grant 067575 from the Wellcome Trust, grant 7407 from ESF, target financing project SF0180087s08, and by the European Union through the European Regional Development Fund via the Center of Excellence in Chemical Biology.
Published ahead of print on 16 December 2009.