The respiratory syncytial virus small hydrophobic protein is phosphorylated via a mitogen-activated protein kinase p38-dependent tyrosine kinase activity during virus infection.

The phosphorylation status of the small hydrophobic (SH) protein of respiratory syncytial virus (RSV) was examined in virus-infected Vero cells. The SH protein was isolated from [35S]methionine- and [33P]orthophosphate-labelled RSV-infected cells and analysed by SDS-PAGE. In each case, a protein product of the expected size for the SH protein was observed. Phosphoamino acid analysis and reactivity with the phosphotyrosine specific antibody PY20 showed that the SH protein was modified by tyrosine phosphorylation. The role of tyrosine kinase activity in SH protein phosphorylation was confirmed by the use of genistein, a broad-spectrum tyrosine kinase inhibitor, to inhibit SH protein phosphorylation. Further analysis showed that the different glycosylated forms of the SH protein were phosphorylated, as was the oligomeric form of the protein. Phosphorylation of the SH protein was specifically inhibited by the mitogen-activated protein kinase (MAPK) p38 inhibitor SB203580, suggesting that SH protein phosphorylation occurs via a MAPK p38-dependent pathway. Analysis of virus-infected cells using fluorescence microscopy showed that, although the SH protein was distributed throughout the cytoplasm, it appeared to accumulate, at low levels, in the endoplasmic reticulum/Golgi complex, confirming recent observations. However, in the presence of SB203580, an increased accumulation of the SH protein in the Golgi complex was observed, although other virus structures, such as virus filaments and inclusion bodies, remained largely unaffected. These results showed that during RSV infection, the SH protein is modified by an MAPK p38-dependent tyrosine kinase activity and that this modification influences its cellular distribution.

Recombinant dengue virus type 1 NS3 protein exhibits specific viral RNA binding and NTPase activity regulated by the NS5 protein.

The full-length dengue virus NS3 protein has been successfully expressed as a 94-kDa GST fusion protein in Escherichia coli. Treatment of the purified fusion protein with thrombin released a 68-kDa protein which is the expected molecular mass for the DEN1 NS3 protein. The identity of this protein was confirmed by Western blotting using dengue virus antisera. Two related activities of the recombinant NS3 protein were characterized, which were the binding of the protein to the 3'-noncoding region of the dengue virus RNA genome and NTPase activity. We demonstrated using a band shift assay that the DEN1 NS3 protein could form a complex with the stem-loop structure in the 3'-noncoding region (3'-NCR), although sites outside the stem-loop may also participate in binding. Using various unlabeled homopolymeric and heteropolymeric RNAs as competitors for binding, it was further shown that the DEN1 NS3 protein exhibits preferential binding to a 94-nt RNA transcript from the 3'-NCR of the dengue virus. The NTPase activity of the recombinant DEN1 NS3 protein was characterized using a thin-layer chromatography assay. We found that the DEN1 NS3 protein possesses some aspects of NTPase activity, which are distinct from those found in other flaviviruses. Although the NS3 protein was able to utilize all four ribonucleoside triphosphates as its substrates, the NS3 protein showed a distinct preference for purine triphosphates (i.e., ATP and GTP). The addition of poly(U) did not stimulate NTPase activity in DEN1 NS3 protein, which contrasts with the reports for other flaviviral NS3 proteins. However, NTPase activity was specifically stimulated by the viral NS5 protein, which was manifested by a more than twofold increase in the rate of ATP hydrolysis and a 25% increase in the yield of ADP at the end of a 120-min reaction. These data suggest that the NTPase activity of the NS3 protein may be regulated by the viral NS5 protein during virus replication.

Expression of the dengue virus structural proteins in Pichia pastoris leads to the generation of virus-like particles.

We have expressed cDNA encoding the dengue virus structural proteins in Pichia pastoris by chromosomal integration of an expression cassette containing the dengue virus structural genes (CprME). The yeast recombinant E protein migrated during SDS-PAGE as a 65 kDa protein when analysed by Western blotting and radioimmunoprecipitation, which is the expected molecular mass for correctly processed and glycosylated E protein. Treatment with endoglycosidases showed that the recombinant E protein was modified by the addition of short mannose chains. The E protein migrated with a buoyant density of 1.13 g/cm3 when analysed using sucrose density gradient centrifugation. Spherical structures with an average diameter of 30 nm, whose morphology resembles dengue virions, were observed in the purified fractions using transmission electron microscopy. Furthermore, the virus-like particles were immunogenic in animals and were able to induce neutralizing antibodies. This is the first report that expression of the structural genes of a flavivirus in yeast is able to generate particulate structures that resemble virions.

The production of recombinant dengue virus E protein using Escherichia coli and Pichia pastoris.

The dengue virus envelope protein was expressed as a GST fusion protein using E. coli and P. pastoris as expression hosts. In E. coli the recombinant E protein is expressed initially as a soluble 81 kDa GST fusion protein. Treatment of the fusion protein with thrombin released a 55 kDa protein, which is the expected size for correctly processed, non-glycosylated recombinant E protein. The antiserum from animals immunised with this recombinant E protein was found to specifically recognise the dengue virus E protein in virus-infected cells, thus demonstrating the immunogenic nature of the recombinant E protein. This expression system allowed production of up to 2 mg of purified recombinant E protein from a 1 1 bacterial culture. In contrast, expression of this GST fusion protein in P. pastoris is associated with extensive proteolytic degradation of the recombinant E protein. However, this proteolytic degradation was not observed in the truncated E protein sequences which were expressed. One of these recombinant fusion proteins, GST E401 was secreted into the culture medium at levels of up to 100 microg/l of growth medium.

Recombinant dengue type 1 virus NS5 protein expressed in Escherichia coli exhibits RNA-dependent RNA polymerase activity.

The complete nonstructural NS5 gene of dengue type 1 virus, Singapore strain S275/90 (D1-S275/90) was expressed in Escherichia coli as a glutathione S-transferase (GST) fusion protein (126 kDa). The GST-NS5 fusion protein was purified and the recombinant NS5 protein released from the fusion protein by thrombin cleavage. The recombinant NS5 had a predicted molecular weight of 100 kDa and reacted with antiserum against D1-S275/90 virus in Western blot analysis. The purified recombinant NS5 protein possessed RNA-dependent RNA polymerase activity which was inhibited (>99%) by antibodies against the recombinant NS5 protein. The polymerase product was shown to be a negative-stranded RNA molecule, of template size, which forms a double-stranded complex with the template RNA.

Alternative native flap conformation revealed by 2.3 A resolution structure of SIV proteinase.

A large conformational change is observed between HIV-1 proteinase in the ligand-free state and in complexes with transition-state inhibitors. Crystal structures of this enzyme have either the flaps open for the native or ligand-free enzyme or the flaps closed for peptidomimetic ligand-bound enzyme. We describe the structure of native recombinant SIV proteinase which like other retroviral proteinases crystallizes as a perfect 2-fold symmetric dimer but in a different crystal packing arrangement. In contrast to HIV-1 PR we show that SIV proteinase in the ligand-free state adopts the closed flaps conformation, demonstrating that ligand binding is not a prerequisite for the closed flaps conformation. The catalytic water was clearly observed between the two aspartates which were not perfectly co-planar, and in this structure the active site cleft is more restricted than for either inhibitor bound or ligand-free HIV-1 proteinase. Accommodation of two bulkier side-chains in the simian enzyme core has resulted in a more exposed N terminus than for HIV-1 PR which we predict could enhance autocatalytic cleavage at the N terminus.

Purification of crystallizable recombinant SIVmac251-32H proteinase.

We have cloned a simian immunodeficiency virus (SIV) proteinase gene directly from proviral DNA of the infectious viral stock SIVmac251-32H (11/88 pool). The deduced amino acid sequence from this proteinase gene is similar to that for the published SIVmac239 molecular clone. SIVmac251-32H proteinase (SIV PR) and its flanking pol sequences were expressed in Escherichia coli as a fusion protein with most of the T7 bacteriophage gene 10 protein. The expressed protein formed cytoplasmic inclusion bodies which were solubilized in 8 M urea, and the recombinant SIV PR was refolded, yielding active, self-processed enzyme. The SIV PR was purified to homogeneity using a single pepstatin A affinity chromatography step, and had a specific peptidolytic activity of 20 mumol/min/mg. Enzymatic characteristics similar to those previously documented for other immunodeficiency virus proteinases (EC 3.4.23) were observed. These include an acidic pH optimum (pH 5.3), sensitivity to sodium chloride concentration, and complete inhibition by pepstatin A. In addition to these properties we have observed quantitative crystallization from low protein concentrations. We describe the first crystal habit for the proteinase from the HIV-2/SIV class of immunodeficiency virus, which is distinctly different from that for HIV-1 proteinase crystals.

Crystallization and preliminary X-ray investigation of recombinant simian immunodeficiency virus proteinase.

Simian immunodeficiency virus (SIV) proteinase has been crystallized from sodium acetate buffer with sodium chloride as precipitant. The crystals are orthorhombic and the space group is C222(1) with unit cell dimensions a = 32.18 A, b = 62.52 A, c = 95.76 A, alpha = beta = gamma = 90 degrees, indicating a single monomer of 10 kDa in the asymmetric unit. The crystals grow to dimensions of 0.2 mm x 0.2 mm x 0.07 mm within a week and are stable in the X-ray beam for at least 50 hours. A different crystal lattice was observed for SIV proteinase crystallized in the presence of pepstatin. The space group was P2(1)2(1)2(1) with cell dimensions of a = 35.26 A, b = 58.59 A, c = 93.95 A, alpha = beta = gamma = 90 degrees. Diffraction beyond 1.7 A was observed, indicating that a high resolution structure analysis is feasible.

Evidence that the amantadine-induced, M2-mediated conversion of influenza A virus hemagglutinin to the low pH conformation occurs in an acidic trans Golgi compartment.

Amantadine treatment of cells infected with H7 strains of influenza A viruses causes an M2 protein-mediated conversion of hemagglutinin (HA) from its native to its low pH conformation. Immunofluorescence and electron microscopic observations showed that the structural alteration and hence drug action occur shortly after HA exits from the Golgi complex during its passage through the strans Golgi region. Using the DAMP/anti-DNP pH probe it is evident that virus infection causes increased acidity of the trans Golgi region and that vesicles containing low pH HA in amantadine-treated virus-infected cells are particularly acidic. These results indicate therefore that the alteration in HA is the direct consequence of exposure to an adverse low pH and provide further support for the conclusion that the M2 protein, the target of amantadine action, is involved in regulating vesicular pH, a function important for the correct maturation of the HA glycoprotein.

Structural characteristics of the M2 protein of influenza A viruses: evidence that it forms a tetrameric channel.

The evidence presented shows that the M2 protein of influenza A viruses exists in infected cells as a homotetramer composed of two disulfide-linked dimers held together by noncovalent interactions. The amphiphilic nature of the transmembrane alpha-helical domain is consistent with the protein forming a transmembrane channel with which amantadine, the specific anti-influenza A drug, interacts. Together these features provide a structural basis for the hypothesis that M2 has a proton translocation function capable of regulating the pH of vesicles of the trans-Golgi network, a role important in promoting the correct maturation of the hemagglutinin glycoprotein.

Specific structural alteration of the influenza haemagglutinin by amantadine.

Amantadine hydrochloride specifically blocks the release of virus particles from H7 influenza virus infected cells. This appears to be the direct consequence of an amantadine induced change in the haemagglutinin (HA) to its low pH conformation. The effect is indirect and mediated via interaction of the drug with the M2 protein since mutants altered in this component alone are insensitive to amantadine. The timing of drug action, some 15-20 min after synthesis, and its coincidence with proteolytic cleavage indicates that the modifications to HA occur late during transport but prior to insertion into the plasma membrane. Reversal by mM concentrations of amines and 0.1 microM monensin indicates that amantadine action causes a reduction in intravesicular pH which triggers the conformational change in HA. We conclude, therefore, that the function of M2 inhibited by amantadine is involved in counteracting the acidity of vesicular compartments of the exocytic pathway in infected cells and is important in protecting the structural integrity of the acid-sensitive glycoprotein.

Palmitoylation of the influenza A virus M2 protein.

The M2 proteins of a variety of influenza A viruses of different subtypes were shown to possess associated palmitate. Susceptibility to removal by reduction or treatment with hydroxylamine is consistent with attachment via a thioester linkage to cysteine. The absence of the acyl group from the M2 proteins of several equine viruses of the H3N8 subtype correlates with the replacement of cysteine 50 with phenylalanine and points to this as the site of palmitate attachment.