Langat Virus M Protein Is Structurally Homologous to prM

Abstract

Langat (LGT) virus M protein has been generated in a recombinant system. Antiserum raised against the LGT virus M protein neutralizes tick-borne encephalitis serocomplex flaviviruses but not mosquito-borne flaviviruses, indicating that the M protein is exposed on the surface of virions. The antiserum recognizes intracellular LGT virus prM/M and binds to prM and M in Western blots of whole-cell lysates and purified virus, respectively. These data suggest that the prM and M proteins are structurally similar under native conditions and support the hypothesis that the “pr” portion of prM facilitates proper folding of the M protein for expression on the virion surface.

The flavivirus genome contains a single open reading frame that encodes three structural proteins and seven nonstructural proteins. The structural proteins consist of the capsid protein and the membrane (M) and envelope (E) proteins, which are on the surface of virions. The M protein is initially expressed as a preprotein (prM) and is cleaved by the enzyme furin to generate the mature virus particle (8). The M and E proteins form a heteromeric complex on the surface of virions (4). The E glycoprotein is the major surface protein for the flaviviruses, and the crystal structure for the soluble ectodomain of the E protein has been determined (7). The E protein is highly antigenic and is thought to contain the receptor binding domain and fusion peptide required for receptor-mediated entry into host cells. The role of the M protein on the surface of virions is unknown. It has been suggested that the prM protein plays a role in the folding of the E protein by facilitating a conformation that is unstable at the low pH of endosomes (1, 2, 6). The cleavage of the prM to form M may serve as the priming reaction for the metastable state of the E protein. In order to examine the biological properties of the M protein, Langat (LGT) virus M protein has been expressed in Escherichia coli and an antiserum has been generated to study the expression patterns of prM/M in LGT virus-infected cells.

LGT virus M protein was expressed in E. coli as a glutathione S-transferase fusion protein (Pharmacia) consisting of the 40 residues between the furin cleavage site and the first of the putative transmembrane domains. The M protein–glutathione S-transferase fusion was affinity purified on a glutathione-agarose column (Sigma), cleaved with thrombin, and further purified on a size exclusion column. Purified M protein runs as a diffuse band of about 5 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels (results not shown).

Purified M protein was used to generate M protein-specific antiserum in rabbits (Alpha Diagnostic, San Antonio, Tex.). In order to examine the protective capacity of the M protein-specific antiserum, plaque reduction neutralization tests (PRNT) (3) were carried out against the tick-borne LGT and Powassan (POW) flaviviruses and the mosquito-borne dengue-4 (DEN-4), Japanese encephalitis (JE), and yellow fever (YF) viruses. The M protein-specific serum neutralized both of the tick-borne flaviviruses (LGT titer, 1:80; POW titer, 1:160) but had no activity against the mosquito-borne viruses (titers of <1:20) (Table 1). These data were a little surprising in that very few prM/M antibodies have been found to be neutralizing (5, 9), but even more interesting was the apparent serogroup-specific neutralization elicited by the M antiserum. To further characterize the antiserum, hemagglutination inhibition (HI) assays (3) were performed against virus and purified viral antigen from the viruses tested in the PRNT assays. The M antiserum was unable to inhibit hemagglutination (titer, ≤10) (Table 1), which was consistent with current thinking that the E protein is the viral hemagglutinin for the flaviviruses.

TABLE 1.

Serological examination of M-specific antiserum^a

Virus strain	PRNT titerb	HI titer with virus	HI titer with viral antigen
LGT TP21	80	10	<10
POW L8	160	<10	<10
DEN-4 1007	<20	NDc	<10
JE P3	<20	ND	10
YF 17D	<20	ND	10

^aPRNT and HI assays were performed as previously described (3). Viruses and antigens were obtained from the WHO World Arbovirus Reference Collection. 

^bPRNT titers reflect an 80% plaque reduction in a 100-PFU test. 

^cND, not determined. 

To determine if M antiserum recognized biologically relevant prM/M protein epitopes, the antiserum was examined for the ability to recognize prM/M viral protein in virus-infected cells. Vero cells were grown on coverslips and infected with virus at a multiplicity of infection of 0.1. The virus was allowed to propagate for 3 days. The cells were washed thrice with phosphate-buffered saline and fixed with 4% paraformaldehyde prior to two phosphate-buffered saline washes and acetone extraction of the cell membranes. The cells were probed with M antiserum followed by a rhodamine-tagged secondary antibody. Indirect immunofluorescence demonstrates that the M antiserum recognizes intracellular forms of the prM/M protein in LGT virus-infected cells (Fig. 1). Intracellular prM/M can be seen clearly in the perinuclear region of the cell, indicating the presence of prM/M in the endoplasmic reticulum and the Golgi network. In some cells, staining was found to be more widespread within the cytoplasm. M antiserum also detected prM/M in POW virus-infected cells but not in YF virus-infected cells (results not shown).

FIG. 1.

M protein antiserum recognizes viral protein within the cytoplasm of LGT virus-infected Vero cells. (A) LGT virus-infected Vero cells fixed and probed at 3 days postinfection. (B) Mock-infected Vero cells. Cells probed with only the secondary antibody had no background fluorescence.

To determine the form of M protein recognized by M antiserum, Western blot analysis was carried out on purified wild-type LGT virus and on cell lysates of cells infected with either wild-type or monoclonal antibody escape variants with potential secondary structure changes in the M protein (unpublished data). M antiserum recognized the M protein in the purified virus preparations when the virus was assayed under either reducing or nonreducing conditions (Fig. 2). When whole-cell lysates were examined, the prM protein of all three viruses was recognized by the M antiserum under nonreducing conditions, but the M protein was not. Neither protein was recognized when run under reducing conditions (results not shown). The fact that M protein was not seen in whole-cell lysates is probably due to a smaller proportion of M protein versus prM protein in the infected cells. These data clearly demonstrate that the conformation of the M protein in mature virions is determined by that of prM and that epitope recognition is very much conformationally dependent. The prM protein contains three disulfide bridges within the cleaved (pr) portion of the molecule, whereas the M protein has no disulfide bridges. The data shown here also suggest that the conformation of the M protein is somewhat stable even under the denaturing conditions of the SDS-PAGE gel.

FIG. 2.

Western blot analysis of LGT virus and virus-infected cell lysates. Wild-type LGT virus (TP21) was purified through a 20% sucrose gradient and run in 4 to 20% acrylamide gradient SDS-PAGE gels under both reducing (Red) and nonreducing (NR) conditions. Virus-infected cell lysates were run under denaturing nonreducing conditions in a 12% gel. Blots were probed with M protein antiserum. 9F9I and 9F9II are variant LGT viruses with mutations near the prM/E cleavage site that may alter the prM secondary structure (unpublished data).

Given that M antiserum can neutralize LGT and POW virus infections, it is also evident that at least a portion of the M protein is solvent exposed and that the conformation of the M protein following cleavage is the same as prior to furin cleavage. Sequence analysis of the extracellular domain of the M protein indicates a relatively high degree of identity between the tick-borne flaviviruses (65%) (Fig. 3), whereas the mosquito-borne flavivirus sequences are less conserved. Among the viruses examined, 5 of 40 (12.5%) amino acids were identical (His-7, Trp-19, Glu-33, Trp-35, and Asn-39). Each of these residues is biochemically unique and is probably required to retain the structural integrity of the M protein. Three other positions are conserved in retaining a positive charge (residues 15, 31, and 38). These may also play a significant role in maintaining the conformation of the M protein or its interaction with the E protein. There are also a number of residues that are unique to the tick-borne flaviviruses and may confer the serospecificity seen in the M antiserum PRNT assays. Interestingly, there is a five-amino-acid sequence (residues 23 to 27) that shares no identity and little homology among the viruses examined. The role of this region of the M protein may be vital to the individuality of each virus or may be a hypervariable region of the M protein.

FIG. 3.

Sequence alignment of flavivirus M proteins. The viruses used for this analysis were LGT strain TP21 (GenBank protein sequence accession no. A41704), POW strain LB (Q04538), Louping Ill (LI) (NP044677), tick-borne encephalitis virus (TBE) strain HYPR (Q01299), DEN-4 (P09866), JE strain SA-14 (P27395), and YF 17D (GNWVY). Sequences were aligned using the AlignX alignment program in the VectorNTI sequence analysis package (InforMax).

This study has demonstrated the feasibility of using recombinant proteins to generate neutralizing antiserum against the flavivirus M protein. The M antiserum binds to the M protein in purified virus and recognizes intracellular prM/M in virus-infected cells. These data represent the first direct experimental evidence that the conformation of the M protein is similar to that of both the prM protein and the cleaved M protein as found on the virion surface. A further examination of the interactions between the E and M proteins will undoubtedly prove that there is a functional link between the two proteins and that the M protein is vital to the proper interaction between the virus and host cells.