The m protein of vesicular stomatitis virus: variability in lipid-protein interaction compatible with function

Solubility of vesicular stomatitis virus M protein in the cytosol of infected cells or isolated from virions

The peripheral membrane M protein of vesicular stomatitis virus purified by detergent extraction of virions and ion-exchange chromatography was determined to be a monomer in the absence of detergent at high salt concentrations. Reduction of the ionic strength below 0.2 M resulted in a rapid aggregation of M protein. This self-association was reversible by the detergent Triton X-100 even in low salt. However, aggregation was not reversible by high salt concentration alone. M protein is initially synthesized as a soluble protein in the cytosol of infected cells, thus raising the question of how the solubility of M protein is maintained at physiological ionic strength. Addition of radiolabeled M protein purified from virions to unlabeled cytosol from either infected or uninfected cells inhibited the self-association reaction. Cytosolic fractions from infected or uninfected cells were equally effective at preventing the self-association of M protein. Self-association could also be prevented by an irrelevant protein such as bovine serum albumin. Sedimentation velocity analysis indicated that most of the newly synthesized M protein is monomeric, suggesting that the solubility of M protein in the cytosol is maintained by either low-affinity interaction with macromolecules in the cytosol or interaction of a small population of M-protein molecules with cytosolic components.

Distribution of M protein and nucleocapsid protein of vesicular stomatitis virus in infected cell plasma membranes

The association of M protein and the nucleocapsid (N) protein of vesicular stomatitis virus (VSV) with the cytoplasmic surface of plasma membranes prepared from infected cells was examined by double label immunofluorescence. M protein in association with the cytoplasmic surface of the plasma membrane was distributed in two distinct labeling patterns. Punctate labeling of M protein in the plasma membrane was observed in association with corresponding labeling for the nucleocapsid protein. Diffusely labeled M protein was distributed in areas of the plasma membrane that were devoid of any detectable labeling for the nucleocapsid protein. Similar results were obtained with two different cell types at 4 h and later times postinfection. The diffuse label for M protein was present in membranes prepared from cells infected with a temperature-sensitive M protein mutant at the nonpermissive temperature, but neither the punctate label for M protein nor labeling for the nucleocapsid protein was observed. Upon shift to permissive temperature, both the punctate label for M protein and labeling for the nucleocapsid protein began to reappear in membranes prepared from cells infected with the M protein mutant. These results indicate that M protein can associate with the plasma membrane without prior binding to nucleocapsids and that association of functional M protein with the plasma membrane is required for the stable association of nucleocapsids with the membrane during the process of viral budding.

Sequence alterations in temperature-sensitive M-protein mutants (complementation group III) of vesicular stomatitis virus

Sequences were determined of the coding regions of the M-protein genes of the Glasgow and Orsay strains of vesicular stomatitis virus (Indiana serotype) and of two group III (M-protein) mutants derived from each wild type. Synthetic primers were annealed with viral genomic RNA and extended with reverse transcriptase. The resulting high-molecular-weight cDNA was sequenced directly. Both Glasgow and Orsay wild types differed in 13 bases from a clone of the San Juan strain sequenced by J. K. Rose and C. J. Gallione (J. Virol. 39:519-528, 1981). Six of these base changes caused amino acid changes in each wild type, whereas seven were degenerate. The Orsay and Glasgow sequences resembled each other more closely than either resembled that of Rose and Gallione, differing in eight nucleotides and four amino acids. Each of the four mutants, however, differed from its parent wild type in only one or two point mutations. Every mutation caused a change either from or to a charged amino acid; the change for tsG31 was Lys (position 215) to Glu, the change for tsO23 was Gly (position 21) to Glu, the change for tsO89 was Ala (position 133) to Asp, the changes for tsG33 were Lys (position 204) to Thr and Glu (position 214) to Lys. The charge differences predicted from these amino acid changes was confirmed by nonequilibrium pH gradient electrophoresis for tsG31, tsG33, tsO23, and the two wild types. These mutations affect residues spanning nearly 85% of the linear sequence, although the mutants possess nearly identical phenotypic properties.