Recombinant Semliki Forest virus vectors encoding hepatitis B virus small surface and pre-S1 antigens induce broadly reactive neutralizing antibodies.

Most hepatitis B virus (HBV) vaccines consist of viral small surface (S) protein subtype adw2 expressed in yeast cells. In spite of good efficacy, HBV-genotype and subtype differences, escape mutants and insufficient Th1 activation remain potential problems. To address these problems, we generated recombinant Semliki Forest virus (rSFV) vectors encoding S protein, subtype adw2 or ayw2, or a fragment of the large surface protein, amino acids 1-48 of the pre-S1 domain, fused to S (pre-S1.1-48/S). The antigen loop in S protein and the selected pre-S1 sequences are known targets of neutralizing antibodies. BALB/c mice were immunized intravenously with 10(7) rSFV particles and 10(8) rSFV particles 3 weeks later. Antibodies induced by rSFV encoding S proteins reacted preferentially with subtype determinants of yeast-derived S antigen but equally well with patient-derived S antigen. Immunization with rSFV encoding pre-S1.1-48/S resulted in formation of pre-S1- and S-specific immunoglobulin G (IgG), while immunization with the isogenic mutant without S start codon induced pre-S1 antibodies only. Neutralizing antibodies were determined by mixing with plasma-derived HBV/ayw2 and subsequent inoculation of susceptible primary hepatocyte cultures from Tupaia belangeri. S/adw2 antisera neutralized HBV/ayw2 as effectively as antisera raised with S/ayw2. The pre-S1 antibodies also completely neutralized HBV infectivity. The IgG1/IgG2a ratios ranged from 0.28 to 0.88 in the four immunized groups and were lowest for the pre-S1.1-48/S vector, indicating the strongest Th1 response. This vector type may induce subtype-independent and S-escape-resistant neutralizing antibodies against HBV.

Semliki Forest virus biodistribution in tumor-free and 4T1 mammary tumor-bearing mice: a comparison of transgene delivery by recombinant virus particles and naked RNA replicon.

Semliki Forest virus (SFV) vectors are promising tools for cancer gene therapy because they ensure a high level of transgene expression and a rapid and strong cytopathic effect. However, broad tissue tropism and transient expression make it more difficult to develop an optimal cancer treatment strategy. In this study, we have compared the distribution of recombinant SFV particles (recSFV) and naked viral RNA replicon (recRNA) in tumor-free and 4T1 mammary tumor-bearing mice as a consequence of different vector administration strategies. The high potential of SFV recRNA as a biosafe approach for the development of therapeutic treatment was demonstrated. Intravenous (i.v.) inoculation of recRNA provided primary brain targeting in both tumor-free and 4T1 tumor mouse models, but local intratumoral inoculation revealed a high expression level in tumors. Moreover, we observed the predominant tumor targeting of recSFV at a reduced viral dose on i.v. and intraperitoneal (i.p.) virus inoculation, whereas the dose increase led to a broad virus distribution in mice. To prolong transgene expression, we have tested several i.v. and i.p. reinoculation strategies. A detailed evaluation of vector distribution and readministration properties could have an impact on cancer gene therapy clinical trial safety and efficacy.

The missing link between envelope formation and fusion in alphaviruses.

Recent structural analyses of the Semliki Forest virus envelope suggest that the spike subunit E1, which is responsible for virus membrane fusion, also maintains the organization of the spike protein shell that encompasses the enveloped virus. This gives E1 a unique opportunity to control membrane stability during the membrane fusion reaction. Here, we present a model for this control mechanism.

Expression of proteins using Semliki Forest virus vectors.

Semliki Forest virus (SFV) vectors have been developed to provide a convenient system to express protein-encoding sequences in virtually any animal cell. This unit presents two strategies for protein expression using SFV vectors. In both cases the protein-coding sequence of interest is cloned into a plasmid vector, which is subsequently used to produce recombinant RNA in vitro. This RNA, which is of positive polarity, is transfected into cells and there is amplified by virtue of its self-encoded RNA replicase. The same replicase also produces a shorter RNA species that encodes the protein of interest. In the first protocol, cells are transfected (either by electroporation or liposome-mediated transfection) and directly analyzed for expression of the heterologous protein. Accompanying support protocols provide methods for checking expression and transfection through galactosidase assays of transfected cells and cell lysates. The other strategy employs in vivo packaging of the RNA into SFV particles; recombinant RNA is cotransfected with a special helper RNA that codes for the structural proteins needed for virus assembly. SFV particles carrying only recombinant RNA are formed and are used to infect cells for analysis of protein expression. Accompanying support protocols describe methods for titrating and purifying recombinant virus stocks. Although the protocols presented here are designed for use with BHK (baby hamster kidney) cells, the virus has a very broad host range and can be used with many different cell types.

Membrane proteins organize a symmetrical virus.

Alphaviruses are enveloped icosahedral viruses that mature by budding at the plasma membrane. According to a prevailing model maturation is driven by binding of membrane protein spikes to a preformed nucleocapsid (NC). The T = 4 geometry of the membrane is thought to be imposed by the NC through one-to-one interactions between spike protomers and capsid proteins (CPs). This model is challenged here by a Semliki Forest virus capsid gene mutant. Its CPs cannot assemble into NCs, or its intermediate structures, due to defective CP-CP interactions. Nevertheless, it can use its horizontal spike-spike interactions on membrane surface and vertical spike-CP interactions to make a particle with correct geometry and protein stoichiometry. Thus, our results highlight the direct role of membrane proteins in organizing the icosahedral conformation of alphaviruses.

Minimal exclusion of plasma membrane proteins during retroviral envelope formation.

The retrovirus forms its envelope by budding at the plasma membrane (PM). This process is primarily driven by its cytoplasmic core-precursor protein, Gag, as shown by the efficient formation of virus-like Gag particles in the absence of its envelope protein, Env. Most interestingly, several studies have demonstrated incorporation of various PM proteins into retrovirus, but the underlying mechanism of this phenomenon has remained elusive. We have purified Moloney murine leukemia virus Gag particles by sedimentation in an iodixanol gradient and donor PMs by flotation in a sucrose gradient and compared their protein compositions at equal lipid basis. We found that most PM proteins are present at similar density in both membranes. The inclusion of PM proteins was unaffected by incorporation of Env protein into the envelope of the Gag particles and whether these were produced at high or low level in the cells. These findings indicate that most PM proteins become incorporated into the retrovirus envelope without significant sorting. This feature of retrovirus assembly should be considered when studying retrovirus functions and developing retrovirus vectors.

Perturbation of the synaptic release machinery in hippocampal neurons by overexpression of SNAP-25 with the Semliki Forest virus vector.

We have examined whether the Semliki Forest virus (SFV) expression vector can be used to manipulate the exocytotic machinery in cultured hippocampal neurons. Autaptic responses were recorded in individually identified neurons which overexpressed either a non-synaptic protein, the transferrin receptor, or the synaptic SNARE protein SNAP-25 (synaptosomal-associated protein of 25 kDA). In neurons overexpressing the transferrin receptor, autaptic responses occurred in a similar proportion and had similar amplitudes (12-18 h postinfection) as in uninfected control neurons. With increasing time after the infection, an increasing proportion of the transferrin receptor-overexpressing neurons showed changes in the shape of the cell body, but the autaptic responses appeared normal as long as recordings could be performed (up to 30 h postinfection). In contrast, in SNAP-25-overexpressing neurons, the proportion of responding cells was reduced 12-18 h after the infection, and the amplitude of the autaptic current in responding neurons was also reduced. The sensitivity to exogenously applied glutamate was, however, unchanged. Biochemical analysis showed that 50% of the overexpressed SNAP-25 was palmitoylated. The levels of two other SNAREs, syntaxin and synaptobrevin (also called vesicle-associated membrane protein), were not affected. Our results indicate that the SFV vector can provide an effective tool to study the function of proteins participating in neurotransmitter release.

Virus maturation by budding.

Enveloped viruses mature by budding at cellular membranes. It has been generally thought that this process is driven by interactions between the viral transmembrane proteins and the internal virion components (core, capsid, or nucleocapsid). This model was particularly applicable to alphaviruses, which require both spike proteins and a nucleocapsid for budding. However, genetic studies have clearly shown that the retrovirus core protein, i.e., the Gag protein, is able to form enveloped particles by itself. Also, budding of negative-strand RNA viruses (rhabdoviruses, orthomyxoviruses, and paramyxoviruses) seems to be accomplished mainly by internal components, most probably the matrix protein, since the spike proteins are not absolutely required for budding of these viruses either. In contrast, budding of coronavirus particles can occur in the absence of the nucleocapsid and appears to require two membrane proteins only. Biochemical and structural data suggest that the proteins, which play a key role in budding, drive this process by forming a three-dimensional (cage-like) protein lattice at the surface of or within the membrane. Similarly, recent electron microscopic studies revealed that the alphavirus spike proteins are also engaged in extensive lateral interactions, forming a dense protein shell at the outer surface of the viral envelope. On the basis of these data, we propose that the budding of enveloped viruses in general is governed by lateral interactions between peripheral or integral membrane proteins. This new concept also provides answers to the question of how viral and cellular membrane proteins are sorted during budding. In addition, it has implications for the mechanism by which the virion is uncoated during virus entry.

Recent advances in gene expression using alphavirus vectors.

Alphavirus vectors use RNA replication in the cell cytoplasm to direct gene expression. New developments of vectors put persistency of expression and infection of specific cells in focus. Furthermore, a new application shows that the system can be used for production of retrovirus vectors carrying genes with introns and control/regulatory regions.

The M1 and NP proteins of influenza A virus form homo- but not heterooligomeric complexes when coexpressed in BHK-21 cells.

The nucleoprotein (NP) and matrix protein (M1) are the most abundant structural proteins of influenza A virus. M1 forms a protein layer beneath the viral envelope and NP constitutes the protein backbone of the ribonucleoproteins (RNPs). In order to elucidate the functions of these proteins in virus assembly we have expressed NP and M1 in BHK-21 cells using Semliki Forest virus replicons and analysed their molecular interactions. We found that both M1 and NP engaged in extensive homooligomerization reactions soon after synthesis. However, there was no detectable heterooligomerization taking place between the two viral proteins, nor between these and host proteins. One interpretation of these results is that homooligomers, and not monomers, of NP and M1 are used as building blocks during RNP assembly and formation of the submembranous M1 layer, respectively. The complete absence of M1-NP heterooligomers suggests, on the other hand, that these two major viral proteins do not interact directly with each other during virus assembly. We also found that a fraction of M1 associated with cellular membranes. This did not, however, result in membrane budding or vesicularization as was the case with the matrix protein of vesicular stomatitis virus when expressed separately (P. A. Justice and others, Journal of Virology 69, 3156-3160, 1995).

The first step: activation of the Semliki Forest virus spike protein precursor causes a localized conformational change in the trimeric spike.

The structure of the particle formed by the SFVmSQL mutant of Semliki Forest virus (SFV) has been defined by cryo-electron microscopy and image reconstruction to a resolution of 21 A. The SQL mutation blocks the cleavage of p62, the precursor of the spike proteins E2 and E3, which normally occurs in the trans-Golgi. The uncleaved spike protein is insensitive to the low pH treatment that triggers membrane fusion during entry of the wild-type virus. The conformation of the spike in the SFVmSQL particle should correspond to that of the inactive precursor found in the early stages of the secretory pathway. Comparison of this "precursor" structure with that of the mature, wild-type, virus allows visualization of the changes that lead to activation, the first step in the pathway toward fusion. We find that the conformational change in the spike is dramatic but localized. The projecting domains of the spikes are completely separated in the precursor and close to generate a cavity in the mature spike. E1, the fusion peptide-bearing protein, interacts only with the p62 in its own third of the trimer before cleavage and then collapses to form a trimer of heterotrimers (E1E2E3)3 surrounding the cavity, poised for the pH-induced conformational change that leads to fusion. The capsid, transmembrane regions and the spike skirts (thin layers of protein that link spikes above the membrane) remain unchanged by cleavage. Similarly, the interactions of the spikes with the nucleocapsid through the transmembrane domains remain constant. Hence, the interactions that lead to virus assembly are unaffected by the SFVmSQL mutation.

Moloney murine leukemia virus envelope protein subunits, gp70 and Pr15E, form a stable disulfide-linked complex.

The nature and stability of the interactions between the gp70 and Pr15E/p15E molecules of murine leukemia virus (MLV) have been disputed extensively. To resolve this controversy, we have performed quantitative biochemical analyses on gp70-Pr15E complexes formed after independent expression of the amphotropic and ecotropic Moloney MLV env genes in BHK-21 cells. We found that all cell-associated gp70 molecules are disulfide linked to Pr15E whereas only a small amount of free gp70 is released by the cells. The complexes were resistant to treatment with reducing agents in vivo, indicating that the presence and stability of the disulfide interaction between gp70 and Pr15E are not dependent on the cellular redox state. However, disulfide-bonded Env complexes were disrupted in lysates of nonalkylated cells in a time-, temperature-, and pH-dependent fashion. Disruption seemed not to be caused by a cellular factor but is probably due to a thiol-disulfide exchange reaction occurring within the Env complex after solubilization. The possibility that alkylating agents induce the formation of the intersubunit disulfide linkage was excluded by showing that disulfide-linked gp70-Pr15E complexes exist in freshly made lysates of nonalkylated cells and that disruption of the complexes can be prevented by lowering the pH. Together, these data establish that gp70 and Pr15E form a stable disulfide-linked complex in vivo.

Packaging of intron-containing genes into retrovirus vectors by alphavirus vectors.

Efficient and controllable expression of a transgene usually requires the presence of intron sequences and much efforts have been made to produce retrovirus vectors that can transduce and integrate genes with introns. However, this has proven difficult because the viral RNA is spliced when it is synthesized in the nucleus of a producer cell. We describe a novel approach to avoid this problem. In our system the retroviral RNA is synthesized in the cytoplasm of the cell, not in the nucleus, in a reaction driven by the Semliki Forest virus (SFV) expression system. The approach was tested with a recombinant Moloney murine leukemia virus genome containing the chloramphenicol acetyltransferase (CAT) gene in association with an intron. This was inserted into a SFV transcription plasmid and the corresponding SFV vector RNA was transcribed in vitro. BHK-21 cells were then transfected with this vector RNA together with two additional SFV vectors that encode the Moloney murine leukemia virus packaging proteins. Retrovirus vectors containing intron-CAT sequences were produced at titers up to 1.3 x 10(6) infectious particles per ml during a 5-hr incubation period. The vectors faithfully transduced the intron-containing CAT gene into NIH 3T3 cells, where the intron-CAT RNA was subjected to efficient splicing and used for high level enzyme expression. Thus, the results show that intron containing genes can be efficiently packaged into retrovirus vectors by the SFV expression system.

Specific interactions between retrovirus Env and Gag proteins in rat neurons.

In this work we have studied the intracellular localization properties of the Gag and Env proteins of Moloney murine leukemia virus (MLV) and human immunodeficiency virus (HIV) in dorsal root ganglion (DRG) neurons of rat. These neurons form thick bundles of axons, which facilitates protein localization studies by immunofluorescence analyses. When such neuron cultures were infected with recombinant Semliki Forest virus particles carrying the gag genes of either retrovirus, the expressed Gag proteins were localized to both the somatic and the axonal regions of the DRG neurons. In contrast, the Env proteins were confined only to the somatic region. When the Gag and Env proteins were coexpressed, the Gag proteins were also excluded from the axons. This effect of the Env proteins was shown to be dependent on the concentration of the Gag proteins in the neuron and also to be specific for homologous pairs of retrovirus proteins. Therefore, the results suggest that there are specific interactions between the Env and the Gag proteins of MLV and HIV in the DRG neurons.

Oligomerization-dependent folding of the membrane fusion protein of Semliki Forest virus.

The spikes of alphaviruses are composed of three copies of an E2-E1 heterodimer. The E1 protein possesses membrane fusion activity, and the E2 protein, or its precursor form, p62 (sometimes called PE2), controls this function. Both proteins are, together with the viral capsid protein, translated from a common C-p62-E1 coding unit. In an earlier study, we showed that the p62 protein of Semliki Forest virus (SFV) dimerizes rapidly and efficiently in the endoplasmic reticulum (ER) with the E1 protein originating from the same translation product (so-called heterodimerization in cis) (B.-U. Barth, J. M. Wahlberg, and H. Garoff, J. Cell Biol. 128:283-291, 1995). In the present work, we analyzed the ER translocation and folding efficiencies of the p62 and E1 proteins of SFV expressed from separate coding units versus a common one. We found that the separately expressed p62 protein translocated and folded almost as efficiently as when it was expressed from a common coding unit, whereas the independently expressed E1 protein was inefficient in both processes. In particular, we found that the majority of the translocated E1 chains were engaged in disulfide-linked aggregates. This result suggests that the E1 protein needs to form a complex with p62 to avoid aggregation. Further analyses of the E1 aggregation showed that it occurred very rapidly after E1 synthesis and could not be avoided significantly by the coexpression of an excess of p62 from a separate coding unit. These latter results suggest that the p62-E1 heterodimerization has to occur very soon after E1 synthesis and that this is possible only in a cis-directed reaction which follows the synthesis of p62 and E1 from a common coding unit. We propose that the p62 protein, whose synthesis precedes that of the E1 protein, remains in the translocon of the ER and awaits the completion of E1. This strategy enables the p62 protein to complex with the E1 protein immediately after the latter has been made and thereby to control (suppress) its fusion activity.

The nucleocapsid-binding spike subunit E2 of Semliki Forest virus requires complex formation with the E1 subunit for activity.

Alphaviruses, such as Semliki Forest virus (SFV), mature by budding at the plasma membrane (PM) of infected cells and enter uninfected ones by a membrane fusion process in the endosomes. Both processes are directed by the p62/E2-E1 membrane protein heterodimer of the virus. The p62 protein, or its mature form E2, provides a cytoplasmic protein domain for interaction with the nucleocapsid (NC) of the virus, and the E1 protein functions as a membrane fusogen. We have previously shown that the p62/E2 protein of SFV controls the membrane fusion activity of E1 through its complex formation with the latter (A. Salminen, J. M. Wahlberg, M. Lobigs, P. Liljeström, and H. Garoff, J. Cell Biol. 116:349-357, 1992). In the present work, we show that the E1 protein controls the NC-binding activity of p62/E2. We have studied E1 expression-deficient SFV variants and shown that although the p62/E2 proteins can be transported to the PM they cannot establish stable NC associations.

Multiple mechanisms for the inhibition of entry and uncoating of superinfecting Semliki Forest virus.

Recombinant Semliki Forest viruses (SFV) that express one or none of the viral structural proteins were used to infect cells and to analyze the fate of incoming superinfecting wild-type viruses. It was found that in addition to the previously described block in replication that superinfecting viruses encounter within 15 min of infection, other mechanisms of superinfection inhibition occurred at later times. Over a 6-hr infection period, inhibition was seen in binding of virus to the cell surface, in acid-activated penetration into the cytoplasm, and in uncoating of nucleocapsids. For each of these processes, the inhibitory mechanism was investigated. In summary, we found that infection evoked several independent mechanisms for blocking the entry and uncoating of superinfecting viruses. The results also offered new insights into the normal processes of penetration and uncoating of SFV.

Efficient multiplication of a Semliki Forest virus chimera containing Sindbis virus spikes.

Using the Semliki Forest virus (SFV) and Sindbis virus (SIN) cDNAs we have constructed recombinants in which the spike genes were exchanged. Analyses of expression showed that the SFV/SIN(spike) RNA directed efficient assembly of infectious virus, whereas the reciprocal SIN/SFV(spike) RNA was completely unable to assemble virus. This was apparently due to a defective capsid-spike interaction.

Targeting of Moloney murine leukemia virus gag precursor to the site of virus budding.

Retrovirus Moloney murine leukemia virus (M-MuLV) matures by budding at the cell surface. Central to the budding process is the myristoylated viral core protein precursor Gag which, even in the absence of all other viral components, is capable of associating with the cytoplasmic leaflet of the plasma membrane and assembling into extracellular virus-like particles. In this paper we have used heterologous, Semliki Forest virus-driven, expression of M-MuLV Gag to study the mechanism by which this protein is targeted to the cell surface. In pulse-chase experiments, BFA, monensin, and 20 degrees C block did not affect incorporation of Gag into extracellular particles thereby indicating that the secretory pathway is not involved in targeting of Gag to the cell surface. Subcellular fractionation studies demonstrated that newly synthesized Gag became rapidly and efficiently associated with membranes which had a density similar to that of plasma membrane-derived vesicles. Protease-protection studies confirmed that the Gag-containing membranes were of plasma membrane origin, since in crude cell homogenates, the bulk of newly synthesized Gag was protease-resistant as expected of a protein that binds to the cytoplasmic leaflet of the plasma membrane. Taken together these data indicate that targeting of M-MuLV Gag to the cell surface proceeds via direct insertion of the protein to the cytoplasmic side of the plasma membrane. Furthermore, since the membrane insertion reaction is highly efficient and specific, this suggests that the reaction is dependent on as-yet-unidentified cellular factors.

Preformed cytoplasmic nucleocapsids are not necessary for alphavirus budding.

According to the present model for assembly of alphaviruses, e.g. Semliki Forest virus (SFV), the viral genome is first encapsidated into a nucleocapsid (NC) in cytoplasm and this is then used for budding at plasma membrane (PM). The preformed NC is thought to act as a template on which the viral envelope can be organized. In the present work we have characterized two SFV deletion mutants which did not assemble NCs in the cytoplasm but which instead appeared to form NCs at the PM simultaneously with virus budding. The deletions were introduced in a conserved 14 residue long linker peptide that joins the amino-terminal RNA-binding domain with the carboxy-terminal serine-protease domain of the capsid protein. Despite the deletions and the change in morphogenesis, wild-type (wt)-like particles were produced with almost wt efficiency. It is suggested that the NC assembly defect of the mutants is rescued through spike-capsid interactions at PM. The results show that the preassembly of NCs in the cytoplasm is not a prerequisite for alphavirus budding. The apparent similarities of the morphogenesis pathways of wt and mutant SFV with those of type D and type C retroviruses are discussed.