|Home | About | Journals | Submit | Contact Us | Français|
The alphaviruses induce membrane invaginations known as spherules as their RNA replication sites. Here, we show that inactivation of any function (polymerase, helicase, protease, or membrane association) essential for RNA synthesis also prevents the generation of spherule structures in a Semliki Forest virus trans-replication system. Mutants capable of negative-strand synthesis, including those defective in RNA capping, gave rise to spherules. Recruitment of RNA to membranes in the absence of spherule formation was not detected.
Alphaviruses are a well-studied group of enveloped viruses with a single-stranded, positive-sense RNA genome and include several human and animal pathogens (1). The replication of all positive-strand RNA viruses takes place in association with cellular membranes which are heavily modified or reorganized to act as efficient platforms for genome amplification (2, 3). It seems that there are two main classes of such membrane modifications: the formation of small invaginations (termed spherules) on membranes and the generation of double-membrane vesicles (4). The alphaviruses and related alphavirus-like plant viruses, as well as nodaviruses, flaviviruses, and tombusviruses, induce the formation of spherule-type structures (4).
Alphavirus replication is catalyzed by four virus-encoded nonstructural proteins, nsP1 to nsP4 (reviewed in reference 5), which arise from a precursor polyprotein termed P1234. nsP1 is responsible for the capping of the viral genome through its methyltransferase and guanylyltransferase activities. nsP1 also attaches the replication complexes to membranes via an amphipathic alpha helix and covalent palmitoylation of the protein. nsP2 contains an N-terminal helicase domain, which also acts as RNA triphosphatase, and a C-terminal protease domain responsible for polyprotein processing. The role of nsP3 is poorly understood compared to the roles of the other nsPs, but it is heavily phosphorylated and has important roles in virus-host interactions. nsP4 is the catalytic RNA-dependent RNA polymerase (RdRp) subunit acting in RNA synthesis.
We have recently constructed a plasmid-based trans-replication system for Semliki Forest virus (SFV) which allows high-level expression and analysis of replicase protein mutants that would be lethal in the context of virus infection (6, 7). We previously showed that the presence of both an RNA template and a functional polymerase is required for the formation of spherules (7). We have now introduced to the trans-replication system multiple point mutations of nonstructural proteins which destroy enzymatic activities or posttranslational modifications in order to study the requirements for RNA replication and for the formation of membranous spherule structures. The mutations (Table 1) were transferred from prior constructs using standard methods, followed by sequence verification.
The polyprotein, together with a short template (Fig. 1A) which yields the highest luciferase levels, was transfected to T7 polymerase-expressing BSR T7/5 cells using Lipofectamine LTX (Invitrogen) as previously described (6, 7). Initially, Renilla luciferase expression from the template was used to assess the efficiency of replication at 16 h posttransfection. Since most of the mutations were completely inactivating for basic functions of the proteins, it was not surprising that only two showed a low-level increase of luciferase activity above the background (Table 1): the Δ50 mutant, in which the phosphorylated region of nsP3 was deleted, and the P1^2^3Z4 mutant, in which the cleavage sites between nsP1, nsP2, and nsP3 were inactivated.
The expression levels of nsP1 and nsP3-ZsG fusion were verified by Western blotting (6) for those mutants that had normal polyprotein processing (Fig. 1B), and the presence or absence of nsP4 was verified for the relevant mutants (Fig. 1C). The levels mostly corresponded to those of the wild-type protein, although some exceptions were noted; e.g., nsP1 was specifically present at lower levels when its ability to interact with membranes was compromised (mutants R253E and W259A). Mutants perturbing the processing of the polyprotein (CA and P1^2^3Z4) gave rise to P123Z4 and P123Z (arrows in Fig. 1D), respectively, but also to anomalous smaller fragments.
The levels of positive-strand and negative-strand RNAs present were analyzed by Northern blotting (6) (Fig. 1E). Only the wild-type polyprotein catalyzed a large excess of positive-strand synthesis compared to the template-only control. Interestingly, the P1^2^3Z4 processing-defective mutant made as much negative-strand RNA as the wild-type protein. Additionally, the nsP1 mutants H38A, D64A, and Y249A each catalyzed significant (30% to 40%) negative-strand synthesis (all quantifications were made with the program ImageJ 1.47v).
The association of RNA and proteins with membranes was studied by flotation in a discontinuous sucrose gradient prepared in 150 mM NaCl–50 mM Tris-HCl (pH 7.5) in 5-ml tubes, as previously described (11), with a slightly modified sucrose composition (Fig. 2). The template Tmed (7) was used to achieve exactly comparable conditions in subsequent experiments. In the presence of wild-type P123Z4 (full replication), the negative strands were predominantly localized with membranes in floating fractions 2 and 3. Significant amounts of the positive strands were also membrane associated (Fig. 2A to toC).C). In contrast, when the RNA was expressed alone in the absence of replicase, it did not associate with membranes. This situation did not change when the polymerase-defective mutant GAA was coexpressed with the RNA (Fig. 2C, upper panels). Thus, we did not detect significant recruitment of positive-strand RNA to the membrane fraction in the absence of replication. When the fractions were treated with 100 μg/ml RNase A for 1 h at room temperature, only the membrane fraction prepared under full-replication conditions showed protection of both RNA strands (Fig. 2C, lower panels), indicating the presence of protective structures. There were significant quantities of negative strands in the soluble fraction, and these were vulnerable to RNase treatment, suggesting that some spherules might break down during the experiment. The membrane-anchoring protein nsP1 was always quantitatively associated with membranes, irrespective of the presence or absence of RNA or the polymerase mutation within the replicase (Fig. 2D). For a positive control, we were able to induce ~50% RNA recruitment to membranes artificially by expression of a strong RNA binding protein modified to bind membranes and by inclusion of the corresponding RNA binding sites in the template (Fig. 2E to toGG).
Since many of the mutants did not replicate, it was not possible to visualize the presence of the template using the previous constructs, which had a red fluorescent marker under the control of the viral subgenomic promoter (7). Therefore, a new version of Tmed, Tmed_Vis, which contains an additional reading frame for the marker, independently driven by a second T7 promoter, was created (Fig. 3A). The sequence encoding fluorescent protein mCherry with a nuclear localization signal (PKKKRKV) added to its C terminus was first cloned in vector pTM1 (22), followed by PCR amplification of the entire new cassette from the T7 promoter to the terminator, which was then cloned to the SacI site of Tmed (7). The derived Tmed_Vis template replicated as well as the original Tmed, based on luciferase activity (Fig. 3B). It also showed strong red fluorescence in transfected cells that was mostly concentrated in the nucleus (Fig. 3C). The nsP3-ZsG fusion was localized in cytoplasmic granules/aggregates (Fig. 3C), which are prominent for nsP3 both during SFV replication and when the protein is expressed alone (23, 24). Thus, cells containing both the replicase and the template could easily be visualized for the replication-defective mutants.
Cells cotransfected with the mutant replicases and Tmed_Vis were then subjected to correlative light and electron microscopy (CLEM) analysis as described previously (7). In each case, only cells expressing both markers were selected for visualization at the EM level. Spherule structures were clearly found in cells with four mutants: H38A, D64A, Y249A, and P1^2^3Z4 (Fig. 3D). The results seen with the Δ50 mutant were inconclusive (see next section), and in spite of an extensive search, spherules were not found for any of the other mutants. Thus, there was an absolute correlation with the ability to synthesize RNA in the generation of spherules. The helicase and protease activities of nsP2 (14, 25), the polymerase activity of nsP4 (19), and the membrane binding ability of nsP1 (9) were each essential for all replication steps, since neither negative- nor positive-strand RNA was detected (Fig. 1E). The P1^2^3Z4 mutant synthesized negative strands at wild-type levels, but few positive strands, fully recapitulating previous results indicating a specific role in negative-strand synthesis for the partially processed polyprotein (26,–29).
RNA capping activity was not needed for negative-strand synthesis or for spherule biogenesis (Fig. 3D). We think it is likely that negative-strand synthesis is sufficient for spherule formation. However, similarly to the P1^2^3Z4 mutant, the capping mutants may be able to make some positive-sense RNA, which for the nsP1 mutants would be uncapped and therefore unstable and poorly translated. In Saccharomyces cerevisiae expressing similar Brome mosaic virus (BMV) RNA capping mutants, newly made positive-strand RNAs could also be detected when the major RNA degradation pathway in the yeast was disrupted (30). The current results also indicate that the main function of residue Y249, which is highly conserved in the alphavirus-like capping enzymes (31, 32), crucially contributes to the enzymatic activity of nsP1, since the mutant phenotype is similar to those of the other active site mutants and not to those of the membrane association mutants, with which Y249A has also been grouped (9). The Δ50 mutant gave increased luciferase activity and therefore should have produced some new negative strands, which we failed to detect, and it also produced few if any spherules. However, it may be that the actual level of RNA synthesis is very low. Luciferase activity greatly amplifies the replication signal, probably also due to the capping and efficient translation of RNAs made by the SFV replicase compared to the uncapped RNAs made by T7 polymerase. Thus, it seems that the Δ50 mutant has a specific defect in trans-replication, since Δ50 mutant viruses replicate quite well (18), producing prominent spherules (33).
For the polymerase mutant (Fig. 2C), or for P123Z in the absence of nsP4 (data not shown), we did not find evidence for the recruitment of positive-strand RNA to membrane fractions. Also, we did not observe significantly reduced luciferase translation from the template RNAs when any of the replication-defective polyproteins were coexpressed (data not shown), which could serve as another indication for viral RNA-protein interactions prior to replication. Viral RNA recruitment to membranes has been studied and prominently detected in the presence of BMV and Flock House nodavirus (FHV) replicase proteins (34,–36). For BMV, spherules are formed by protein 1a also in the absence of replication, and it seems that RNA is mainly stabilized inside the spherules (34). In contrast, FHV, similarly to SFV, requires RNA replication for spherule biogenesis (37). Nevertheless, a polymerase-inactive FHV mutant can still fully recruit viral RNA to a membrane-associated state (35). Several domains of the FHV replicase protein A, including the polymerase domain itself, are required for membrane recruitment (35), but no other point mutants beside polymerase inactivation have been studied to date.
Although SFV trans-replication is efficient overall (6), it may be that the RNA recruitment was inefficient or readily reversible under the conditions used, and in future work we will explore alternative methods, such as RNA-protein cross-linking to detect the potential recruitment intermediate, which should exist at least transiently. In summary, we have established a tight link between RNA replication and spherule biogenesis for multiple mutants of SFV replicase and have developed Tmed_Vis as a tool to detect the presence of template RNA in individual cells.
We thank Ilkka Fagerlund for performing some Northern blotting experiments and Mervi Lindman and Arja Strandell for excellent technical assistance in EM.
This study was supported by the Academy of Finland (grant 265997). K.K. was supported in part by a fellowship from the Helsinki Graduate Program in Biotechnology and Molecular Biology and K.H. by an Academy of Finland postdoctoral fellowship.
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.