Oligomerization, secretion, and biological function of an anchor-free parainfluenza virus type 2 (PI2) fusion protein

A number of studies indicate that the transmembrane domain, the cytoplasmic domain, or both regions of viral surface glycoproteins are involved in quaternary structure formation. In this report, the transmembrane domain and cytoplasmic tail coding sequence of the fusion (F) glycoprotein gene from parainfluenza type 2 virus was truncated by PCR and the resulting gene (PI2F') was expressed in HeLa-T4 cells by using the vaccinia virus-T7 transient expression system. Pulse-chase experiments indicated that the anchor-free PI2F' was expressed and processed into F(1) and F(2) subunits. Both the processed and the unprocessed anchor-free PI2F' proteins were found to be efficiently secreted into the culture medium. Examination of the oligomeric form of the anchor-free PI2F' by chemical cross-linking demonstrated that it assembles posttranslationally into dimers and trimers with a pattern similar to that of the wild-type PI2F protein. In an effort to better understand the biological properties of the truncated form of PI2F', we anchored PI2F' by a glycosyl-phosphatidylinositol (GPI) linkage. The GPI-anchored PI2F' protein, when coexpressed with PI2HN, did not induce cell fusion seen as syncytium formation, but was found to initiate lipid mixing (hemifusion) as observed by transfer of R-18 rhodamine from red blood cells to the GPI-PI2F'/PI2HN cotransfected cells. The results therefore indicate that the extracellular domain of the PI2 fusion protein contains not only the structural information sufficient to direct assembly into higher oligomers, but also is competent to initiate membrane fusion, suggesting that the anchor-free PI2F' may be useful for further structural studies.     

Assembly of Sendai virus: M protein interacts with F and HN proteins and with the cytoplasmic tail and transmembrane domain of F protein

Sendai virus matrix protein (M protein) is critically important for virus assembly and budding and is presumed to interact with viral glycoproteins on the outer side and viral nucleocapsid on the inner side. However, since M protein alone binds to lipid membranes, it has been difficult to demonstrate the specific interaction of M protein with HN or F protein, the Sendai viral glycoproteins. Using Triton X-100 (TX-100) detergent treatment of membrane fractions and flotation in sucrose gradients, we report that the membrane-bound M protein expressed alone or coexpressed with heterologous glycoprotein (influenza virus HA) was totally TX-100 soluble but the membrane-bound M protein coexpressed with HN or F protein either individually or together was predominantly detergent-resistant and floated to the top of the density gradient. Furthermore, both the cytoplasmic tail and the transmembrane domain of F protein facilitated binding of M protein to detergent-resistant membranes. Analysis of the membrane association of M protein in the early and late phases of the Sendai virus infectious cycle revealed that the interaction of M protein with mature glycoproteins that associated with the detergent-resistant lipid rafts was responsible for the detergent resistance of the membrane-bound M protein. Immunofluorescence analysis by confocal microscopy also demonstrated that in Sendai virus-infected cells, a fraction of M protein colocalized with F and HN proteins and that some M protein also became associated with the F and HN proteins while they were in transit to the plasma membrane via the exocytic pathway. These studies indicate that F and HN interact with M protein in the absence of any other viral proteins and that F associates with M protein via its cytoplasmic tail and transmembrane domain.     

Influenza virus M2 integral membrane protein is a homotetramer stabilized by formation of disulfide bonds

The oligomeric structure of the influenza A virus M2 integral membrane protein was determined. On SDS-polyacrylamide gels under nonreducing conditions, the influenza A/Udorn/72 virus M2 forms disulfide-linked dimers (30 kDa) and tetramers (60 kDa). Sucrose gradient analysis and chemical cross-linking analysis indicated that the oligomeric form of M2 is a tetramer consisting of either a pair of disulfide-linked dimers or disulfide-linked tetramers. In addition, a small amount of a cross-linked species of 150-180,000 kDa, which the available data suggest contains only M2 polypeptides, was observed. The role of M2 cysteine residues in disulfide bond formation and their role in forming oligomers were examined by converting each of the two extracellular and single cytoplasmic cysteine residues to serine residues and expressing the altered M2 proteins in eukaryotic cells. Removal of either one of the N-terminal cysteines at residues 17 or 19 indicated that tetramers formed that consisted of a pair of noncovalently associated disulfide-linked dimers, suggesting that each of the cysteine residues is equally competent for forming disulfide bonds. When both cysteine residues were removed from the M2 N-terminal domain, no disulfide-linked forms were observed. When solubilized in detergent this double-cysteine mutant lost reactivity with a M2-specific mAb and exhibited an altered sedimentation pattern on sucrose gradients. However, chemical cross-linking of this double-cysteine mutant in membranes indicated that it can form tetramers. Taken together, these data suggest that disulfide bond formation, although not essential for oligomeric assembly, stabilizes the M2 tetramer from disruption by detergent solubilization.     

Isolation of the influenza C virus glycoprotein in a soluble form by bromelain digestion

The spike glycoprotein of influenza C/Johannesburg/1/66 was isolated in a soluble form by digestion of MDCK cell-grown virions with bromelain. The whole ectodomain of the glycoprotein could be recovered with an apparent molecular weight of 75,000 daltons determined in SDS-PAGE. Comparison to Triton X-100-isolated glycoprotein revealed that a C-terminal peptide of 3000-4500 daltons must have remained in the viral membrane. When purified by sucrose density gradient centrifugation the glycoprotein sedimented with a sedimentation coefficient of 10 S, indicating a molecular weight of 206,000 daltons, which is consistent with a trimeric structure of the spike molecule. The trimeric form was stabilized in sucrose gradients by Ca2+ ions. Bromelain digestion of virions with uncleaved glycoprotein, grown in MDCK cells without trypsin, produced two disulphide-linked subunits with similar electrophoretic mobilities in SDS-PAGE to the biologically active glycoprotein. The smaller subunit differed from the product cleaved in vivo (gp 30) by the presence of an additional arginine residue at the N-terminus. The soluble glycoprotein appears to possess both receptor-binding and receptor-destroying enzyme activities, as isolated glycoprotein inhibited hemagglutination of intact influenza C virions and showed RDE activity in an in vitro test. Glycoprotein exposed to low pH, which was sensitive to trypsin digestion, also demonstrated both these biological activities. Glycoprotein-mediated hemolysis could not be observed.     

Biosynthesis of glycophospholipid bound and secreted murine class I Qa-2 polypeptides

Murine T cells synthesize and express a cell-surface glycophospholipid anchored 40 kDa and a secreted water-soluble 39 kDa Qa-2 polypeptide. We have examined the biosynthetic pathways which lead to the production of the membrane-bound and water-soluble isoforms of the Qa-2 molecule. Using the detergent TX-114, both detergent (membrane)-bound and soluble Qa-2 polypeptides can be identified in cell lysates and can be distinguished by charge and molecular weight. Two membrane-bound forms, a 40-kDa Endo H resistant cell-surface form and a 38 kDa-Endo H sensitive form can be identified, both of which can be biosynthetically labeled with 3H-ethanolamine and can be converted to water soluble forms by digestion with a phosphatidylinositol specific phospholipase C. In addition, several water soluble polypeptides at 39, 37, 35 kDa, and a minor species at 33 kDa were identified, none of which radiolabel with 3H-ethanolamine. While the 39-kDa polypeptide was Endo H resistant, the other isoforms were sensitive to Endo H digestion. Pulse chase experiments and molecular weights of the deglycosylated core polypeptides suggest a precursor to product relationship between the intracellular water-soluble species and the mature 39-kDa secreted Qa-2 molecule. This relationship is supported by the observation that murine L cells transfected with the Qa-2 encoding class I gene Q7 fail to express membrane-bound Qa-2 molecules yet synthesize both intracellular water-soluble and secreted Qa-2 molecules. These findings argue for a pathway in which secreted soluble Qa-2 molecules are derived from intracellular precursors.     

Reducing agent-sensitive dimerization of the hemagglutinin-neuraminidase glycoprotein of Newcastle disease virus correlates with the presence of cysteine at residue 123

Viruses within the Newcastle disease virus (NDV) serotype induce a wide array of disease manifestations ranging from an almost apathogenic pattern to the high mortality caused by avirulent or virulent isolates, respectively. A disulfide-linked dimer form of the NDV hemagglutinin-neuraminidase (HN) glycoprotein can be demonstrated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under nonreducing conditions for only some of these isolates. For others, indeed the majority of those we have studied, no such reducing agent-sensitive dimeric form of HN is demonstrable. Apparently, there is no causal relationship between disulfide-linked dimeric HN and virulence. Using the deduced amino acid sequence of the dimeric HN of isolate AV as a basis for selection of oligonucleotide primers, we sequenced three additional reducing agent-sensitive dimeric HN glycoproteins and eight for which a disulfide-linked dimer has not been identified, using primer extension and dideoxy sequencing. The deduced amino acid sequences reveal a strict correlation between the presence of cysteine at residue 123 and reducing agent-sensitive dimerization of HN.     

Intracellular processing of the vesicular stomatitis virus glycoprotein and the Newcastle disease virus hemagglutinin-neuraminidase glycoprotein

The kinetics of intracellular transport of the vesicular stomatitis virus (VSV) glycoprotein (G) and the Newcastle disease virus (NDV) hemagglutinin-neuraminidase (HN) glycoprotein in chicken embryo cells were compared. To assay for the appearance of pulse-labelled glycoprotein at the cell surface, an antibody-binding assay was developed which allowed the precipitation of only those molecules on the outside surfaces of infected cells. Using this assay, it was found that pulse-labelled VSV G protein appeared at the cell surface with a half-time of approximately 27 min, while pulse-labelled NDV HN glycoprotein reached the cell surface with a half-time of approximately 78 min. To determine the transit time of these glycoproteins to trans-Golgi membranes, the kinetics of the acquisition of endoglycosidase H resistance was analyzed. The half-time of the transit of the G protein to the trans-Golgi membranes was found to be approximately 13 min while that of the HN glycoprotein was found to be approximately 60 min. Since the G protein migrates to the trans-Golgi membranes with a half-time of 13 min, and the cell surface with a half-time of 27 min, the half-time for the transit between the trans-Golgi membrane and the plasma membrane must be approximately 14 min. In a similar analysis, the half-time for the transit of the HN glycoprotein from the trans-Golgi membrane to the plasma membrane must be approximately 18 min, a time not significantly different from that of the G protein. Thus the difference in the kinetics of the intracellular transport of these two glycoproteins resides primarily in the transit from the rough endoplasmic reticulum to the trans-Golgi membranes. These results argue against a non-selective mechanism for the transport of plasma membrane glycoproteins to the cell surface.     

Reconstitution of membranes with individual paramyxovirus glycoproteins and phospholipid in cholate solution.

A method has been developed for reconstitution of biologically active membranes with individual Sendai virus glycoproteins (HN and F) and phosphatidylcholine. The glycoproteins were isolated in the presence of Triton X-100 and transferred into cholate solution by sedimentation into a sucrose gradient containing 2% cholate. Membranes were then reconstituted with HN or F and phosphatidylcholine by removal of cholate by dialysis. Experiments with radioactively labeled detergents showed that the removal of Triton X-100 by sedimentation into cholate and the removal of cholate by dialysis were essentially complete. With the F protein, vesicles 400–800 Å in diameter and filaments 600–800 Å in length were formed, depending on the proportions of lipid and protein in the initial mixture. Both structures were covered with glycoprotein spikes. With the HN protein, only vesicles were formed, and the densities of the spikes on the surface was dependent on the initial lipid to protein ratio. Membranes reconstituted from HN protein exhibited hemagglutinating and neuraminidase activities. Reconstituted particles containing the F protein exhibited hemolytic activity when a mechanism was provided to attach the F protein-lipid complex to the cell, i.e., by the addition of wheat germ agglutinin. These results have confirmed the role of the F protein in membrane fusion and have shown that the requirements for F protein activity are the previously demonstrated proteolytic processing of F, the insertion of the F protein into lipid, and the presence of an attachment mechanism.     

Molecular domains involved in oligomerization of the Friend murine leukemia virus envelope glycoprotein

The oligomeric structure of the Friend murine leukemia virus envelope glycoprotein has been investigated using crosslinking reagents and sucrose density gradient centrifugation. The results obtained provide evidence that both the precursor and the processed molecules are oligomeric and probably form tetramers. Pulse-chase analyses indicate that assembly occurs sequentially, within 30 min of protein synthesis and prior to cleavage of the precursor. Studies using chimeric envelope glycoproteins and deletion mutants indicate that the transmembrane and cytoplasmic domains are not essential for the formation of oligomers. Evidence is also presented that the SU subunit remains in an oligomeric form following disassociation from the TM subunit. Oligomeric envelope glycoprotein complexes linked by intermolecular disulfide bonds were also observed under certain conditions. Mink cell focus-forming virus envelope glycoprotein constructs lacking the transmembrane domain or both the transmembrane and the cytoplasmic domains formed intermolecular disulfide bonds more readily than the full-length molecule, suggesting that these regions are likely to make a contribution to the conformation of the glycoprotein. These data indicate that there are several points of interaction between retrovirus envelope glycoprotein monomers which contribute to assembly of the oligomer and that contacts within the ectodomain appear to be of critical importance.     

Formation of concatemeric DNA as an intermediate in the replication of bacteriophage T1 DNA molecules.

The structure of intracellular DNA extracted from phage T1 infected cells was analysed by sedimentation through sucrose gradients. DNA labelled with 3H-dThd during a short pulse given at any time during T1 DNA synthesis sedimented in neutral gradients as a broad heterogeneous band with a large fraction of the label sedimenting more rapidly than mature T1 DNA molecules. Rapidly-sedimenting label was also observed when pulse-labelled DNA was denatured and analysed on alkaline sucrose gradients. Electron microscopy of intracellular T1 DNA revealed linear molecules of variable length the longest of which were three to four times the mature genome length. The distribution of lengths derived from electron microscopy are consistent with the molecular length distributions calculated from the sedimentation coefficients. We conclude that the rapidly-sedimenting DNA is in the form of concatemers consisting of linear tandem repeats of the T1 genome. The concatemeric form of replicating T1 DNA is a precursor of progeny T1 genomes since in pulse-chase experiments it was converted efficiently into mature, infectious T1 phage particles. The identification of this concatemeric form of T1 DNA provides supporting evidence for the model proposed by Gill & MacHattie (1976) to account for the formation of the very limited number of cyclic permutations of gene sequence found for mature T1 DNA molecules.     

Secreted Reelin molecules form homodimers

During mammalian brain development, neurons are generated along the ventricle, migrate radially, and become aligned in defined patterns. These precise patterns of neuronal alignment are regulated by an extracellular matrix protein Reelin, and binding of Reelin to its receptors induces tyrosine phosphorylation of the intracellular adaptor protein disabled 1 (Dab1). We recently reported that Reelin molecules assemble to form a homomeric protein complex. Although the number of molecules in the full-length complex is unknown, recombinant N-terminal fragments, which contain the epitope for the function-blocking CR-50 antibody, assembled to form a complex of more than 40 monomers. When the N-terminus was deleted from Reelin, the truncated protein did not form a stable complex. To further characterize the Reelin assembly, we performed biochemical analysis of the full-length Reelin assembly in this study. Here, we report that a full-length Reelin forms a disulfide-linked homodimer. A chemical crosslinking experiment on secreted Reelin confirmed that only dimers are formed by the full-length protein. However, interestingly, chemical crosslinking of the N-terminus-truncated Reelin resulted in the formation of larger complexes, in addition to dimers, suggesting that the tertiary structure required for the proper and stable assembly/dimerization was altered by the truncation. The truncated protein did not induce efficient tyrosine phosphorylation of Dab1, although it bound well to the receptors. These findings demonstrate the functional importance of the N-terminal region of Reelin for proper dimerization and signaling. Proper but not simple extracellular crosslinking of the receptors by these dimers may be important for Reelin signaling to occur.     

The Nonstructural Small Glycoprotein sGP of Ebola Virus Is Secreted as an Antiparallel-Orientated Homodimer

The nonstructural small glycoprotein sGP, which unlike the transmembrane GP is synthesized from primary nonedited mRNA species, is secreted from infected cells as a disulfide-linked homodimer. Site-directed mutagenesis of all cysteine residues revealed that dimerization is due to an intermolecular disulfide linkage between cysteine residues at positions 53 and 306. Formic acid hydrolysis of sGP demonstrated that sGP dimers consist of monomers in antiparallel orientation. Another editing product of the GP gene of Ebola virus (ssGP), which shares 295 amino-terminal amino acid residues with sGP, is secreted from cells in a monomeric form due to the lack of the carboxyl-terminal part (present in sGP), including cysteine at position 306.     

Characterization of different oligomeric species of the Yersinia enterocolitica outer membrane protein YadA

The oligomeric structure of the plasmid-encoded outer membrane protein YadA of Yersinia enterocolitica was studied by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and sucrose gradient sedimentation, respectively. The apparent molecular weight (M _r) of the oligomeric 200-kDa YadA species detected by SDS-PAGE varied from 152,000 to 240,000 depending on the respective acrylamide concentration. The atypical electrophoretic behavior of the 200-kDa YadA species results from an exceptionally high relative free mobility as revealed by the Ferguson plot. In contrast, the apparent M _r of 53,000 of the YadA monomer was independent of the acrylamide concentration. An additional oligomeric 116-kDa YadA species was detected by SDS-PAGE when membrane preparations of Y. enterocolitica were solubilized in SDS at 37 °C. The gel-purified 116-k Da YadA species was completely converted to the 200-kDa species by heating at 100 °C and to the monomeric form (M _r 53,000) by heating in the presence of 10 M urea without reducing agents, respectively. This suggests that the 116-kDa YadA species represents the native oligomeric form of YadA, whereas the 200-kDa species is only generated from native YadA during denaturation in SDS. The significance of the 116-kDa YadA species is also supported by the rather slow sedimentation at about 6 S of detergent-solubilized YadA in sucrose gradients, which probably contains only two or three monomers.     

Biosynthesis of J-chain in human lymphoid cells producing immunoglobulins of various isotypes

The relationship between synthesis, secretion, and subcellular localization of J-chain, IgM, IgA, and IgG was investigated in cultures of PWM-stimulated human PBL and in lymphoblastoid cell lines. Cells were examined for surface, cytoplasmic, and secreted immunoglobulins (Igs) and J-chain by immunofluorescence and radioimmunoassay (RIA). By these techniques, J-chain was detected in cells that produce polymeric or monomeric Igs. In PWM-stimulated PBL the synthesis of J-chain paralleled the production of Igs. In both PWM-stimulated (for 2 days) and unstimulated PBL, equal proportions of free and disulfide-linked J-chain were found. Increased amounts of intracellular J-chain were produced at later stages in PWM-stimulated PBL and J-chain occurred mostly in a free form. In tissue culture fluids, J-chain was not secreted in a free form but was always disulfide-linked to polymeric Igs. In lymphoblastoid cell lines, J-chain was present in a disulfide-linked form in IgM and IGA producers, but in IgG cells and in an IgM cell line (DAUDI) that did not secrete IgM but expressed it on the cell membrane, intracellular J-chain was present in free form. Although various proportions of polymeric and monomeric IgA were seen in culture fluids from IgA-secreting cell lines, intracellular IgA occurred mostly in a monomeric form. Further studies revealed that the ability to produce polymers was not equally distributed among all cells and might vary according to their content of J-chain and stage of maturation. Subcellular fractionation and subsequent analyses for J-chain and Ig in PWM-stimulated PBL and in IgM or IgG-producing cell lines revealed that these proteins were associated with fractions that contained ribosomes, cell sap, and low molecular weight RNA. In lysates of IgG and J-chain producing cells grown in the presence of 3H-labeled amino acids, intracellular J-chain was not disulfide-linked to IgG.     

整合素及其信号传导通路

整合素是位于细胞表面的重要黏附分子,通过其双向信号传导通路,介导细胞与细胞外基质及细胞与细胞间的黏附。整合素由胞外域、跨膜域和胞内域3部分组成。胞内域与细胞内信号分子结合,启动胞内-胞外信号传导激活整合素,提高与相应配体亲合力。而胞外域与相应配体结合后,通过胞外-胞内信号传导,调节细胞生存、增殖、黏附、分化功能。近年研究显示,整合素结构功能及信号传导通路异常与多种疾病有关。     

Studies on the assembly of the envelope of Newcastle disease virus.

The association of the envelope proteins of Newcastle disease virus with membranes of infected BHK 21-F cells and their incorporation into mature envelopes has been investigated in a study employing cell fractionation. The principal fractions obtained by sucrose density gradient centrifugation of cytoplasmic extracts were rough endoplasmic reticulum and smooth membranes derived predominantly from smooth endoplasmic reticulum and Golgi apparatus. Furthermore, by adsorption to red blood cells it was possible to isolate virions and a hemadsorptive fraction of smooth membranes believed to be immediate precursors of mature envelopes. In addition to the cytoplasmic fractions, plasma membranes obtained as cell ghosts have been analyzed. Each fraction showed a distinct pattern of virus-specific proteins. Pulse-chase experiments indicated that glycoprotein HN and F_o were synthesized on the rough endoplasmic reticulum and transferred from there via smooth intracellular membranes to the plasma membrane and into virions. In the course of migration, F_o is converted to F. In contrast to the glycoproteins, protein M was found to be incorporated into the plasma membrane immediately after synthesis. Pulse-chase experiments also demonstrated that this protein appears in the hemagglutinating fraction of smooth membranes and in mature virions more rapidly than the glycoproteins. These results suggest that M is incorporated into membranes that contain already viral glycoproteins and that this process is one of the last steps in envelope assembly.     

The signal for clathrin-mediated endocytosis of the paramyxovirus SV5 HN protein resides at the transmembrane domain-ectodomain boundary region.

The hemagglutinin-neuraminidase (HN) glycoprotein of the paramyxovirus SV5 is internalized from the cell surface via clathrin-coated pits. However, the cytoplasmic domain of SV5 HN does not contain a previously characterized internalization motif. A cell-surface-expressed chimeric protein (APK), consisting of the cytoplasmic tail, transmembrane (TM) domain, and 12 residues of the ectodomain of HN joined to the cytoplasmic protein pyruvate kinase is internalized, indicating that the N-terminal region of HN contains an internalization signal. Although SV5 HN is internalized at a rate similar to that of influenza virus hemagglutinin (HA) mutant Y543, which contains a degenerate tyrosine-based signal in its cytoplasmic tail, the elimination of the majority of the HN cytoplasmic tail, or substitution of the HN TM domain with leucine residues, did not affect the rate of HN internalization. The HN protein of the closely related virus, Newcastle disease virus (NDV), is not internalized from the cell surface. Working under the usual convention that the TM domain consists of the hydrophobic residues bounded by two charged residues, analysis of internalization of mutant and chimeric NDV HN molecules indicates that the first seven SV5 HN ectodomain residues are critical for internalization of HN. A glutamic acid residue (E37) that abuts this presumptive HN TM domain/ectodomain boundary is important for SV5 HN internalization.     

The effects of monoclonal antibodies against the hemagglutinin-neuraminidase and fusion protein on the release of Sendai virus from infected cells

Summary Vero cell cultures in Leighton tubes were infected with egg-grown Sendai virus at high multiplicity of infection. Four hours after infection, the cultures were labelled with³⁵S-methionine, after which various concentrations of fourteen and five mouse monoclonal antibodies directed against different antigenic determinants of the hemagglutinin-neuraminidase (HN) and fusion (F) protein, respectively, were added to the medium. Fourty-eight hours after infection radiolabelled virions released into the medium were collected and purified by discontinuous sucrose gradient centrifugations. The amount of virus-bound radioactivity obtained in the various extracellular materials allowed an estimation of the capacity of the different monoclonal antibodies to inhibit the release of Sendai virus. In addition, the release of virions from infected cells was studied ultrastructurally.Based on their serological reactivity the fourteen anti-HN monoclonal antibodies could be divided into four groups. The first group of clones could not inhibit any biological activity of the virus. These clones were binding proximally, near the base of the HN glycoprotein and could not inhibit the release of the virus. The second group blocked hemolysis, but did not block hemagglutination (HA) or neuraminidase (NA) activity. The third group of clones blocked all biological activities of the HN glycoprotein. The fourth group could only block NA activity. With the exception of one of five monoclonal antibodies belonging to the second group, antibodies of the second, third and fourth group were found to bind more distally on the HN glycoprotein. Except for two monoclonal antibodies of the second group they could all effectively inhibit release of the virus from infected cells. Ultrastructurally, these antibodies caused aggregation of virions in contact with the plasma membrane.The five monoclonal antibodies directed against the F protein reacted with four different antigenic sites. These antibodies could not prevent the release of Sendai virus.With 5 Figures     

Membrane fusion activity, oligomerization, and assembly of the rabies virus glycoprotein.

The spike glycoprotein (G protein) of rabies virus (CVS strain) expressed in HeLa cells from cloned cDNA mediated membrane fusion after exposure to pHs of 6.1 or below. Chemical crosslinking showed that the rabies G protein, like the vesicular stomatitis virus (VSV) G protein, could be crosslinked to dimers and trimers, indicating that rabies G protein is a trimer. However, unlike the VSV G protein, rabies G protein trimers were not stable to sedimentation in sucrose gradients, even at a mildly acidic pH which stabilizes the VSV G protein trimers. In addition, we report that the expressed rabies virus G protein was functional because it could assemble into VSV particles (tsO45) lacking VSV G protein and rescue infectivity. These VSV (rabies) pseudotypes were neutralized only by an antibody to the rabies G protein. We also examined the properties of a hybrid protein containing the extracellular domain of the rabies virus glycoprotein and the transmembrane and cytoplasmic domains of the VSV G protein. This protein was transported to the cell surface and could be crosslinked to form dimers and trimers, but had little or no detectable membrane fusion activity. The lack of fusion activity was paradoxical because the hybrid protein could rescue VSV infectivity, although the titers were lower than those obtained with the wild-type rabies G protein.     

Expression and characterization of soluble human parainfluenza virus type 1 hemagglutinin-neuraminidase glycoprotein.

Human parainfluenza virus types 1 (hPIV-1), 2, and 3 represent significant respiratory pathogens for which no antiviral treatment is currently available. To characterize the biochemical functions of the hPIV-1 hemagglutinin-neuraminidase (HN) glycoprotein, a potential target for antiviral therapy, we cloned and expressed a soluble portion of hPIV-1 HN (amino acid residues 137-575), lacking the N-terminal hydrophobic membrane anchorage region, in insect cells using the baculovirus secretion expression system. The expressed HN protein was purified through cation-exchange chromatography followed by metal affinity chromatography, using the 6xHis epitope introduced at the carboxyl terminus of the recombinant protein. N-terminal amino acid sequence analysis of purified HN indicated that the honeybee melittin secretion signal peptide was correctly removed during post-translational processing. Further characterization revealed that the purified HN protein was N-glycosylated and exhibited neuraminidase activity whose characteristics resembled those of the native HN protein of hPIV-1 virions. The establishment of this expression and purification system has allowed us to further explore the biochemical characteristics of paramyxovirus HN and to obtain material that could be suitable for X-ray crystallography studies.