Progeny rabies virions that bud from infected cells are able to spread from cell to cell in cell or tissue culture (in vitro) presumably as they do in vivo. They have the option of spreading to contiguous cells (direct cell-to-cell spread) or to noncontiguous cells, which are surrounded by interstitial space. In the case of direct cell-to-cell spread, rabies virus spreads despite a continuous presence of serum virus-neutralizing antibody (VNA). Alternatively, virus that buds from an infected cell into the surrounding interstitial space must find another cell to infect. In this case, the spread of virus is limited by the presence, in vitro and in vivo, of VNA that blocks virus attachment to cellular receptors and subsequent virus entry into a susceptible cell (Dietzschold et al., 1985; Flamand et al., 1993). In vivo, rabies virus also can spread in the cell, particularly cells of peripheral nerves and neuronal cells of the CNS, through intraaxonal transport in a microtubule network-dependent process. Virus that moves intraaxonally can cover great distances, particularly in bipolar neurons, before reaching and crossing the synapse of one dendritic process into another (Tsiang, 1979; Kucera et al., 1985; Gillet et al., 1986; Ceccaldi et al., 1989; Coulon et al., 1989; Lafay et al., 1991). It was even postulated that naked viral nucleocapsids (RNPs) might be transported in the axoplasmic flow along axons and through the synapse into the postsynaptic neuron, particularly because this provides an alternative mechanism of virus spread along neuronal networks when mature virions cannot be detected at synapses (Gosztonyi, 1994). This hypothesis of transsynaptic transfer of naked nucleocapsids, however, was weakened considerably, if not negated, by recent studies that demonstrate an absolute requirement for the virion G in transsynaptic spread of rabies virus both in vivo and in vitro (Etessami et al., 2000). Other significant studies of rabies virus spread also point to the nature of the viral tropism and its relationship to the G of the virion. In particular, viruses that are pathogenic or virulent in vivo differ in their ability to invade the CNS and spread within the brain in comparison with the apathogenic or avirulent virus phenotype (Coulon et al., 1989; Lafay et al., 1991). This phenotypic difference, which is determined by the surface G (Coulon et al., 1983; Dietzschold et al., 1983b), is discussed further in Sect. VI.E.
VI. MOLECULAR BIOLOGY OF RABIES VIRUS PROTEINS A. Nucleocapsid Protein (N)
The N, P, and L proteins, which form the RNP core of the virus, will be discussed first, and then M and G proteins will be discussed as they interact with the RNP to form the assembled virion. The N contains 450 amino acids and has a molecular weight of ~57,000. It is a major component of the virus and the major protein of the internal helical NC (RNP core). It represents the virus group-specific core antigen (Schneider et al., 1973). The N is the most conserved of the viral components in terms of amino acid sequence similarity within genotypes, despite a relatively high degree of genetic diversity within short regions of the N gene between the genotypes (Conzelmann et al., 1990; Bourhy et al, 1993a; Kissi et al., 1995). The highest degree (98-99%) of N amino acid sequence similarity is shared by the different rabies virus "fixed" laboratory strains (genotype 1) (Wunner et al., 1988; Conzelmann et al., 1990). Based on present knowledge, it appears that rabies and rabies-related virus isolates that share less than 80% of nucleotide similarity in the N gene (including noncoding regions) and less than 92% of N amino acid similarity belong to different genotypes (Bourhy et al., 1993; Kissi et al., 1995). These often represent populations of stable, optimally adapted variant virus phenotypes. One of the reasons for the high level of conservation, particularly within specific regions in the N, is that key functions of the N must be retained. For example, the N has specific and absolute requirements for N-specific encapsidation of genome RNA and protection of the RNA template from ribonuclease activity (Sokol et al., 1969; Wunner, 1991). Another example is its role in regulating RNA transcription and modulating viral RNA transcription and replication by promoting read-through of the termination signals (Patton et al., 1984; Yang et al., 1998, 1999). Some of the amino acid differences produce unique, genotype-specific epitopes in the N that can be used to assign viruses to the different serotypes and genotypes on the basis of their reactivity patterns with a panel of anti-N monoclonal antibodies (MAbs) (Flamand et al., 1980a; Dietzschold et al., 1987a; Smith, 1989). This qualitative diversity built into the N that appears to reliably characterize the different lyssaviruses antigenically also has been exploited at the nucleotide level by the polymerase chain reaction (PCR) technology (Sacramento et al., 1991; Bourhy et al., 1993). The so-called molecular or genetic typing technique, which employs reverse transcriptase (RT)-PCR technology and the viral genome RNA as template, is targeted either to a selected short (~200 nucleotide) region of the N gene or to the entire N gene. First used as a fairly simple diagnostic technique, the RT-PCR approach has been used recently to elucidate the epidemiologic and evolutionary relationships between rabies and rabies-related viruses (Smith et al., 1992; Nadine-Davis et al., 1994; Kissi et al., 1995) (see Chap. 3).
Nascent N can be found by immunostaining technique distributed diffusely in the cytoplasm of infected cells. The N is quickly consumed in the formation of homologous (N-N) and heterologous (N-P) complexes and assembled into nascent viral nucleocapsids (RNP complexes). The N also may be concentrated, usually in complexes with P, in one or more cytoplasmic inclusion bodies in the cell (in vitro) or in Negri bodies in vivo (Chenik et al., 1994; Kawai et al., 1999). The immediate interaction of nascent N with P serves to prevent self-aggregation of N and provide the specificity required for N-viral RNA encapsidation (Masters and Banerjee, 1988; Chenik et al., 1994; Fu et al., 1994; Yang et al., 1998). Interestingly, N interaction with P occurs either at the N or C terminus of N, both of which modulate and enhance N encapsidation of viral RNA. Moreover, the function of P in supporting RNA encapsidation by N depends on the de novo interaction between P and N (Yang et al., 1998). The RNA binding site in N appears to lie within the N-terminal domain (first 376 amino acids), possibly between amino acid residues 298 and 352 (Kouznetzoff et al., 1998). After N binds to viral RNA, it undergoes conformational change, acquiring a number of conformation-dependent epitopes, and is phosphorylated at serine 389 (Dietzschold et al., 1987a; Kawai et al., 1999). It is possible that during RNA encapsidation, an encapsidation-associated conformational change exposes serine at position 389, making it accessible for phosphorylation (Kawai et al., 1999). Phosphorylation of N is unique to rabies virus, in contrast to N of VSV and other rhabdoviruses (Sokol et al., 1974; Sokol and Koprowski, 1975), but not to NC proteins of other negative-strand RNA viruses (Lamb and Choppin, 1977; Robbins and Bussell, 1979; Hsu and Kingsbury, 1982). The phosphorylated N of rabies virus raises the interesting question of whether phosphorylation of N in rabies virus causes the N to function differently in viral RNA transcription and replication compared with the unphosphorylated N in VSV. One study that attempts to address this question has expressed unphosphorylated N (one that has substituted serine at position 389 with arginine) in a rabies virus-like particle instead of the wild-type N that is phosphorylated. The result was that the rates of transcription and replication of the rabies-like viral RNA were significantly lower with unphosphorylated N than with expression of phosphorylated N even though the unphosphorylated N bound less strongly to the leader RNA (Yang et al., 1999). Perhaps this, at least in part, explains the slower growth of rabies virus compared with VSV.
When recombinant rabies virus N produced in insect cells was bound to viral RNA, ring structures were generated that were suggestive of the helical structure of rabies virus nucleocapsids (Schoehn et al., 2001). Importantly, the recombinant N that is expressed in insect cells is phosphorylated (Prehaud et al., 1990). These recombinant ring structures are biochemically and biophysically indistinguishable from authentic rabies virus N-RNA complexes, and when observed in the EM, they appear to be identical to rabies virus nucleocapsids (Iseni et al., 1998; Schoehn et al., 2001). In the EM images, the N molecule of the recombinant ring structures bends slightly, giving it a bilobed shape and permitting one N molecule to contact the next N monomer on the viral RNA strand at two nonidentical sites. Such images also reveal that the N subunits are packed closely, every 9 to 11 nucleotides, along the RNA. This is sufficient to protect the viral RNA against ribonuclease activity and to give the recombinant structure the same density as that of authentic viral N-RNA complex (Iseni et al., 1998). As a model for authentic viral nucleocapsids, it is interesting to see how the P binds to the recombinant N in the N-RNA ring. The P, which binds to the unique trypsin-sensi-tive cleavage site close to the C-terminal end of N (Dietzschold et al., 1987a; Kouznetzoff et al., 1998), ends up in different positions, one bending toward the inside and the other bending toward the outside of the ring (Schoehn et al., 2001). In either position, the interaction of N with P probably causes a further conformational change in N similar to the conformational change that is believed to occur when N first associates with P after N encapsidates the viral RNA (Kawai et al., 1999).
The N is the second most extensively analyzed of the rabies virus proteins (after the G) with respect to its antigenic and immunogenic structure and function. The immunologic interest in N stems from the observation that the RNP core of rabies virus induces protective immunity against a peripheral challenge of lethal rabies virus in animals (Dietzschold et al., 1987b; Tollis et al., 1991). Two distinct immunogenic features associated with the rabies virus N relate to the immunoprotection induced in animals by the internal viral nucleocapsid. First, N is able to protect animals against a peripheral challenge with rabies virus in the absence of detectable VNA (Lodmell et al., 1991; Sumner et al., 1991; Tollis et al., 1991). Second, the N is able to prime the immune system and enhance the production of VNAs following subsequent inoculation of animals with inactivated rabies virus vaccine (Dietzschold etal., 1987b; Tollis et al., 1991; Fu et al., 1991). B- and T-cell-specific epitopes in N were defined initially using MAbs and synthetic peptides in competitive binding assays. These reagents delineated the topography of functional antigenic sites common to rabies and rabies-related viruses (Flamand et al., 1980a; Lafon and Wiktor, 1985; Dietzschold et al., 1987a; Ertl et al., 1989). Subsequently, several linear B- and T-cell epitopes were physically mapped on N (reviewed in Fu et al., 1994; Goto et al., 1995, 2000). Three linear antigenic epitopes (antibody binding sites) on N were mapped to amino acids 358-367 (antigenic site I), and three linear epitopes (antigenic site IV) were mapped to two independent regions, amino acids 359-366 and 375-383 (Minamoto et al., 1994; Goto et al., 2000). Although the region between residues 359 and 366 is shared by the two independent antigenic sites (I and IV), the MAbs that recognize epitopes within these sites do not compete with each other for binding to the N antigen. Thus it would appear that the respective epitopes are detected on different forms of N, one that represents N that is diffuse in the cytoplasm and the other that is associated with cytoplasmic inclusion bodies. The two forms of N might even have different degrees of protein folding or maturation (Goto et al., 2000). The fact that the N that is associated with inclusion bodies may be the more mature form, having gained a greater degree of folding, is suggested by another MAb that is specific for antigenic site II, which only recognizes a conformation-specific epitope on the inclusion body-associated N antigen. The linear epitope (amino acids 373-383) that appears to overlap but may not be identical to another epitope (375-383) in site IV is recognized by a different anti-N MAb (Dietzschold et al., 1987a). Three of the five epitopes from sites I and IV in rabies virus N, which are also shared by rabies-related viruses, represent cross-reactive determinants (Goto et al., 2000). Conformation-dependent epitopes are present in antigenic sites II and III (Minamoto et al., 1994).
The N protein is a major target antigen for T-helper (Th) cells that cross-react among rabies and rabies-related viruses (Celis et al., 1988a,b; Ertl et al., 1989). Several Th cell epitopes in the rabies virus N were identified and mapped using a series of overlapping synthetic peptides corresponding to N sequences of approximately 15 amino acids in length (Ertl et al., 1989). Antigenic peptides (bearing a specific epitope for each subset of Th cells) that were capable of stimulating rabies virus-specific Th cells in vitro were selected and subsequently tested for stimulation of rabies virus-specific Th cells in vivo (Ertl et al., 1989, 1991). One such peptide, designated 31D, which corresponds to N residues 404-418, was found to be an immunodominant epitope capable of stimulating production of rabies virus-specific Th cells, at least in vitro. The same peptide also induced a significant Th cell response in vivo and an accelerated VNA response after a booster immunization with inactivated rabies virus in vivo (Ertl et al., 1989). Neither the peptide-induced Th cell response nor the increase in VNA titer on vaccine boost induced by this peptide epitope, regardless of the immun-odominance, however, was sufficient to protect against a lethal virus challenge dose in mice.
Finally, certain other properties and immune responses that the rabies viral nucleocapsid specifically elicits in humans suggest that the rabies virus N functions as an exogenous superantigen (Lafon et al., 1992). It is perhaps the only viral superantigen that has been identified in humans (Lafon, 1997). Some of the properties and responses found not only in humans but in mice that are attributable to the rabies virus N in the role of superantigen include (1) its potent activation of peripheral blood lymphocytes in human vaccinees (Herzog et al., 1992), (2) its ability to produce a more rapid and heightened VNA response on injection of inactivated rabies vaccines (Dietzschold et al., 1987b; Fu et al., 1991), (3) its induction of early T-cell activation steps and expansion and mobilization of CD4+ Vß8 T cells to trigger and support production of VNA (Lafon et al., 1992; Martinez-Arends et al., 1995), and (4) its ability to bind to HLA class II antigens expressed on the surface of cells (Lafon et al., 1992).
The P of lyssaviruses, like the P in other Mononegavirales viruses, is a multifunctional protein in its interaction with N and a key component of the virion-associated RNA polymerase complex as a regulatory protein in viral genome replication. The P has been identified by other designations in the literature, which may be confusing to anyone entering the field for the first time. The Ml designation described the earliest interpretative view of the protein as a membrane-associated protein, similar to the M2 (now designated M) (Gyorgy et al., 1971). The NS designation, referring to nonstructural protein, was adopted from the corresponding protein in VSV, so designated because large amounts were found in virus-infected cell extracts and none was detected in purified virions (Kang and Prevec, 1971). The P designation, which refers to the protein as the nominal phosphoprotein in rabies virus, is becoming the accepted designation because it conforms to the currently accepted designation for the corresponding protein in VSV and other Mononegavirales viruses. It is not, however, the only phosphorylated protein in rabies virus because it shares this posttranslational modification with the N of rabies virus.
The rabies virus P contains 297 amino acids (MOKV P has 303 amino acids) (38-41 kDa), is well conserved (>97%) among genotype 1 lyssaviruses and is found in a variety of phosphorylated forms (Gupta et al., 2000). Among its many functions, the P acts as a chaperone of soluble nascent N, preventing its polymerization (self-assembly) and nonspecific binding to cellular RNA. The P in N-P complexes specifically directs N encapsidation of the viral RNA (Chenik et al., 1994; Fu et al., 1994; Gigant et al., 2000). As a subunit of the RNA polymerase (P-L) complex, the P plays a pivotal role as a cofactor in transcription and replication of the viral genome (Chenik et al., 1994; Fu et al., 1994; Chenik et al., 1998). Essentially, the P serves both to stabilize the L protein (Curran et al., 1994) and to place the polymerase complex on the RNA template, which the L protein alone is unable to do (Mellon and Emerson, 1978).
The P in both rabies virions and virus-infected cells is present in two prominent forms, one that is hyperphosphorylated and the other that is hypophospho-rylated. These two forms migrate with different mobilities in SDS-PAGE (Wunner et al., 1985; Gupta et al., 2000). Other forms of P, also detected in SDS-PAGE, may exist in rabies virus-infected cells. These will have fewer amino acids than P, the result of N-terminally truncated products due to translation initiation at internal AUG codons located in-frame within the P mRNA (Conzelmann et al., 1990; Fu et al., 1994; Chenik et al., 1995). The most hydrophilic and conserved region of P is located in the center of the sequence between amino acids 139 and 170, which is where one might predict most of the phosphate acceptor amino acids to be located (Conzelmann et al., 1990). This is in contrast, however, to VSV P, where the highest degree of hydrophilicity and most of the phosphate residues are found in the N-terminal portion (between amino acids 35 and 106) of the protein (Bell and Prevec, 1985; Hsu and Kingsbury, 1985). Interestingly, the center region of the rabies virus P is not the region where most residues are phosphorylated. Recent studies investigating different protein kinases present in a cell extract from rat brain have shown that rabies virus P is phosphorylated in the N-terminal portion by two distinct types of protein kinases, one of which is a unique heparin-sensitive protein kinase (Gupta et al., 2000). This unique 71-kDa kinase, designated rabies virus protein kinase (RVPK), phosphorylates recombinant P (36 kDa, expressed in Escherichia coli) at serine 63 and serine 64 (CVS strain) and alters its mobility in SDS-PAGE to migrate more slowly, as a protein of 40 kDa. The other phosphorylating enzyme is protein kinase C, which has several isomers (PKCa, (J, x, and 8). In contrast to the RVPK, phosphorylation of P by the PKC isoforms, dominated by PKC7 activity, did not alter the migration of P in SDS-PAGE (Gupta et al., 2000). On analyzing the protein kinase activity in rabies virions for the presence of these two types of enzymes, it was concluded the RVPK is selectively packaged in mature rabies virions along with a smaller amount of the predominant PKC-y isoform as the rabies virion-associated protein kinases (Gupta et al., 2000).
In accordance with the multifunctional role of rabies virus P, like the P molecules in other Monegavirales viruses, its cofunction with the N and L is mediated by domains of P that specifically interact with each protein. As mentioned previously, nascent rabies virus P binds to nascent soluble N and maintains Nina competent form for RNA encapsidation. Recent findings indicate that P binds to the C-terminal part of N in the RNA-N complex (Schoehn et al., 2001). Using the method of deletion mutant analysis to map the region(s) of the rabies virus P that binds to N, at least two independent N-binding sites were found on P (Fu et al., 1994; Chenik et al., 1994). One site is located within the C-terminal 30 amino acids of P and another in the N-terminal portion of the protein between amino acids 69 and 177 (Chenik et al., 1994). Depending on whether the two proteins are synthesized simultaneously, mimicking the in vivo situation, or synthesized separately and then mixed together, the two proteins interact in a manner that is mutually independent (Fu et al., 1994). For example, when the two proteins are synthesized simultaneously, P mutants with C-terminal deletions of up to 166 amino acids still form complexes with N, indicating that the N-terminal region of P is involved in the interaction. In a contrasting manner, when the two proteins are synthesized individually and then mixed together, deletions of more than 47 amino acids from the C terminus of the P fail to bind to the N (Fu et al., 1994). This demonstrates two points. The first point is that the P molecule has a C-terminal binding site for N (between amino acids 267 and 297) that is used when the two proteins are presynthesized and then mixed together. The second point is that the initial interaction of the two proteins, involving the N-terminal region of P binding to N, is linked directly and immediately to the simultaneous synthesis of the two proteins in vitro, mimicking the in vivo situation. That is, when P molecules with N-terminal deletions of up to 68 amino acids are synthesized simultaneously with N, binding to N occurs (Chenik et al., 1994), but with less efficiency compared with full-length P binding to N (Fu et al., 1994). When the two proteins are synthesized separately and then mixed, P molecules with N-ter-minal deletions bind to N with equal efficiency (Fu et al., 1994). This suggests that P molecules with N-terminal deletions are able to bind to N as well as full-length P via the C-terminal binding site on P or via a binding site for N that is located further downstream from amino acid 69, or both. One study has mapped the N-terminal binding site on P between amino acids 69 and 177 (Chenik et al., 1994). Together these findings suggest that P molecules use the two binding sites to interact with N, but at different times and perhaps for different purposes. The interaction involving the N-terminal binding site of P requires that P interacts with N soon after the two proteins are synthesized in vivo but that it may compete with another molecule that interacts with N independently. The other interacting molecular species that competes for the binding site on N might be endogenous RNA. This dual-binding-site model that depicts the P associating with N reflects how the P may play a regulatory role in viral RNA transcription and replication.
Once bound to the RNA-N template in progeny RNP formation, the P is required to bind L to produce a virus-encoded RNA-polymerase complex that is fully active (Chenik et al., 1998). The P subunit in the P-L complex has a major binding site for L protein within the first 19 amino acids of P (Chenik et al., 1998). This is in agreement with the model that suggests that the L binding site resides in the negatively charged N terminus of the VSV P, although another region also has been shown to contribute to L binding (Takacs and Banerjee,
1995). Unlike the P of VSV, the rabies virus P does not appear to be required for L stabilization. The P is able to oligomerize (form complexes) with itself, although it is not entirely clear whether trimers or tetramers, or both, are formed and which of these oligomers is necessary for binding to L (Gao et al., 1996; Spadafora et al., 1996; Gigant et al., 2000). The oligomeric forms of rabies virus P coexist in the cytoplasm in equilibrium with the monomer species. Like the P of Sendai virus, a paramyxovirus, oligomerization of rabies virus P does not require phosphorylation nor is the N-terminal domain (first 52 amino acids) necessary for oligomerization or binding to the N-RNA template (Curran et al., 1995; Tarbouriech et al., 2000; Gigant et al., 2000). This is in contrast to the P of VSV, which requires phosphorylation for oligomer formation to be fully active and necessary for binding both to L protein and to the RNA template (Gao et al.,
Mention of other protein-protein interactions that involve the rabies virus P begins to address the question of whether rabies virus proteins specifically interact with cellular factors (i.e., proteins) beside the host cell receptor to influence or help regulate virus tropism and cell-to-cell spread. Using the yeast two-hybrid approach to identify interactive cellular factors, the cytoplasmic dynein light chain (LC8), an 89-amino-acid protein (10 kDa), was found to interact strongly with the P of rabies virus and Mokola virus (Jacob et al., 2000; Raux et al., 2000). In both studies, the P domain that interacts with dynein LC8 was mapped to the N-terminal half of the P; one of them mapped the interactive site to within amino acids 138-172 in the P (Raux et al., 2000). Of particular interest, with regard to the manner in which rabies virus spreads along neurons in vivo, often over long distances, is the manner in which dynein LC8 might facilitate this movement. Dynein
LC8, as a part of cytoplasmic dynein and myosin V, participates in the myosin V complex, a microtubule-associated motor protein complex that is implicated in the actin-based motor transport of ER vesicles in brain neurons (Jacob et al., 2000). Thus the retrograde axonal transport of uncoated nucleocapsids along axons to the perikaryon and transport of nascent RNP from the perikaryon along dendrites to the next neuron might be mediated by the P protein-dynein LC8 complex in transporting the viral RNP along the microtubule network (Ceccaldi et al., 1989).
C. Virion-Associated RNA Polymerase or Large Protein (L)
The L in rabies virus is encoded in the fifth gene, which comprises more than half (54%) of the coding potential of the rabies virus genome. The L contains 2142 and 2127 amino acids in the PV and SAD-B19 strains of rabies virus, respectively, and 2127 amino acids in MOKV (244 kDa). The L is the catalytic component of the polymerase complex and along with the noncatalytic cofactor P is responsible for the majority of enzymatic activities involved in viral RNA transcription and replication. Many of the activities of this multifunctional enzyme have been demonstrated in genetic and biochemical studies with VSV, the prototype virus and model for studying the virion-associated RNA polymerase of negative-strand RNA viruses (Banerjee and Chattopadhyay, 1990). The viral RNA-polymerase plays a unique role at the start of infection by initiating the primary transcription of the genome RNA once the NC core is released into the cytoplasm of the infected cell. The enzymatic steps of transcription include initiation and elongation of the Le+ RNA and mRNA transcripts as well as cotranscriptional modifications of the mRNAs that include 5' capping, methyla-tion, and 3' polyadenylation. Comparisons of L sequences from different Mononegavirales viruses have helped to map the functionally homologous and unique sequences in attempts to locate the ascribed enzyme activities (Tordo et al, 1988; Poch et al., 1990; Barik et al., 1990). One of the main features of L that comes out of the sequence comparison is that the domains of sequence homology are not distributed randomly along the protein. Some domains are highly conserved, with the high proportion of amino acids either strictly or conservatively maintained in identical positions, whereas other domains are more variable, consistent with the multifunctional nature of L (Tordo et al., 1988; Poch et al., 1990; Banerjee and Chattopadhyay, 1990). Four motifs, labeled A through
D, in the central part of the rabies virus L, between residues 530 and 1177 and between residues 532 and 1201 in the VSV L, represent regions of highest similarity (Tordo et al., 1988; Poch et al., 1989; Barik et al., 1990). These motifs, which are thought to constitute the polymerase module of L, maintain the same linear arrangement and location in all viral RNA-dependent RNA and DNA
polymerases (Poch et al., 1990; Bank et al., 1990; Delarue et al., 1990). Among the conserved sequences in these four motifs is the tri-amino acid core sequence GDN (standing for glycine, aspartic acid, and asparagine) in motif C, which is extensively conserved in all nonsegmented negative-strand RNA viruses (Poch et al., 1989). A recent study has shown that not only the GDN core sequence but also specific amino acids downstream from the core sequence are crucial for the maintenance of polymerase activity that catalyzes the polymerization of nucleotides (Schnell and Conzelmann, 1995). In addition, at least two other sequences between amino acid residues 754-778 and 1332-1351 in the VSV L have been identified as consensus sites for binding and utilization of ATP, similar to those found in cellular kinases (Barik et al., 1990; Canter et al., 1993). Three essential activities encoded by L are involved in the binding and utilization of ATP. These are (1) the transcriptional activity that requires binding to substrate ribonucleoside triphosphates (rNTPs), (2) polyadenylation, and (3) protein kinase activity for specific phosphorylation of the P in transcriptional activation (Sanchez et al., 1985; Chattopadhyay and Banerjee, 1987). Many of the putative functions of this multifunctional protein, including mRNA capping, methylation, and polyadenylation, remain to be delineated and mapped within the L of rabies virus. The process of mapping active sites in the rabies virus L using the mutational and deletion approach will be helped considerably now that it is feasible to apply the powerful technique of reverse genetics (Conzelmann and Schnell, 1994; Schnell et al., 1994; Schnell and Conzelmann, 1995). With reverse genetics, it is possible to express the rabies virus L with amino acid deletions and point mutations introduced into the cDNA of the L gene to map the locations in L that have functional activities. Similarly, the role of the cofactor P in the RNA-polymerase complex can be better defined using reverse genetics. Since the L protein relies exclusively on its interaction with the phosphorylated P to be fully active, the major question, does the P complement any of the specific enzymatic functions of L or does the P function solely as a regulatory protein in the RNA transcription and replication process? For example, does the cofactor P unwind the RNA-N complex of the NC to facilitate entry as well as movement of L on the genome template (De and Banerjee, 1985)? The cooperative function of the non-catalytic cofactor P and catalytic L in the polymerase complex is clearly intriguing and of critical importance to warrant further examination.
D. Matrix Protein (M)
The M of rabies virus and of MOKV is the smallest of the virion proteins. It contains 202 amino acids (25 kDa) (Rayssiguier et al., 1986; Todo et al., 1986; Conzelmann et al., 1990; Hiramatsu et al., 1993; Bourhy et al., 1993a; Gould et al., 1998). The M forms a sheath around the RNP core in virion assembly, producing the skeleton structure of the virion. It, too, is a multifunctional protein that interacts with viral proteins and protein components of cellular membranes. The functional properties of M include downregulation of viral RNA transcription, condensation of helical NC cores into tight coils, association with membrane bilayers, and involvement in the cytopathogenesis of virus-infected cells (see references in Ito et al., 1996) The N-terminal region of the rabies virus M has a high content of charged amino acids and proline residues (Poch et al., 1988), similar to the M of VSV (Rose and Gallione, 1981) and paramyxoviruses (Chambers et al., 1986). It would appear from anti-M MAb blocking studies that the N terminus of rabies virus M plays a critical role in the regulation of RNA transcription (Ito et al., 1996). The M of VSV is a potent inhibitor of RNA transcription, shutting down transcription of both viral genes and independently cellular genes of the infected cell by inhibiting host RNA polymerase (Ferran and Lucas-Lenard, 1997; Ahmed and Lyles, 1998). The ability of the VSV M to inhibit host transcription correlates with the cell rounding cytopathic effect observed in VSV-infected cells in culture (Blondel et al., 1990; Simon et al., 1990). However, a similar cytopathic effect is not as prominent in rabies virus-infected cultures, suggesting that the rabies virus M may not play the same role or perhaps have the same specificity for cellular factors to inhibit cellular RNA transcription (Lyles and McKenzie, 1997). The central portion contains a hydrophobic domain (residues 89-107) that is presumed to interact with membrane lipids (Capone and Ghosh, 1984; Tordo et al., 1986b). The M of rabies virus is also palmitoylated, although the site(s) of the presumed cysteine residue(s) for palmitoylation has not been identified (Gaudin et al., 1991).
The M binds to and condenses the nascent NC core into a tightly coiled, helical ribonucleocapsid-M protein (the skeleton) complex. Approximately 1200 to 1500 copies of M molecules bind to rabies virus RNP core. At the same time M binds to the RNP structure, it mediates binding of the viral core structure to the host membrane at the marginal region of the cytoplasm, where it initiates rabies virus budding from the cell plasma membrane (Mebatsion et al., 1999). The M gives the virion its characteristic bullet-like shape, regardless of whether its location is within the RNP core or on the external surface of the core (Barge et al., 1993; Lyles et al., 1996). The mechanism by which M mediates the budding of virus off the cell membrane appears to be associated with a proline-rich (PPPY, PP*Y or PY) domain located at residues 35-38 within the highly conserved 14-amino-acid sequence near the N terminus of the rabies virus M (Harty et al., 1999). A corresponding proline-rich motif (in the single-letter code, P is proline and Y is tyrosine, and x is any amino acid) is found in the M of VSV (Gill and Banerjee, 1986), as well as in the M of Ebola and Marburg viruses (Sanchez et al., 1993). The PY motif is very similar to the late budding domain identified in viral proteins such as the Gag protein p2b of Rous sarcoma virus (Wills et al., 1994) and the p6 Gag protein in human immunodeficiency virus (Gottlinger et al., 1991), both of which are associated with virus budding. The unique function of the PY motif is that it interacts with a WW domain, 38-40 amino acids long with two highly conserved tryptophans (in single-letter code, tryptophan is W) spaced 20-22 amino acids apart, found in a wide range of cellular proteins. Some of the WW domain-containing proteins are involved in cytoskeletal formation, whereas others are involved in signal transduction and gene regulation (Sudol, 1996). It is therefore likely that the rabies virus M involves cellular proteins in the release (exocytosis) of rabies virions from the cell (Harty et al., 1999; Craven et al., 1999; Jayakar et al., 2000). Although the exocytotic release of virus particles requires the M in RNP-M skeletons, the efficiency of virus budding is enhanced greatly by the interaction of the RNP-M complex with the envelope G (Mebatsion et al., 1996). Increased virion production as a result of direct interaction of the cytoplasmic domain of the transmembrane spike G and the viral RNP-M core suggests that a concerted action of both core and spike proteins is necessary for efficient recovery of virions. However, the interaction of M with the cytoplasmic domain of G does not need to be optimal; i.e., the interaction is sufficient if the G of different viruses are substituted for the homologous G in budding virions (Mebatsion et al., 1995; Morimoto et al., 2000).
E. Glycoprotein (G)
The mature G of all rabies virus strains examined is a 505-amino-acid (~65 kDa) type I membrane glycoprotein (there are 503 amino acids in MOKV G) translated from a G-mRNA transcript that encodes 524 amino acids (522 amino acids from MOKV G-mRNA) (Benmansour et al., 1992; Bourhy et al., 1993). The first 19 amino acids represent the signal peptide (SP) that provides the membrane insertion signal, which transports the nascent protein into the membranes of the rough ER-Golgi-plasma membrane pathway before it is cleaved from the N terminus of the G molecule in the Golgi apparatus. The G, which forms the trimeric spikes that extend 8.3 nm from the virus surface, is the only surface protein of the virion. Each G of the spike is anchored in the viral envelope by a 22-amino-acid transmembrane (TM) domain located between residues 439 and 461 (Gaudin et al., 1992). The C-terminal portion of G, the cytoplasmic domain (CD), extends from under the viral envelope to the cytoplasm of the infected cell, where it interacts with M of the skeleton particle to complete the virion assembly. The ectodomain of G (residues 1-439), that portion which extends outward on the virion surface, is the business end of the molecule. It is responsible for virus interaction of rabies virus with its cellular binding sites (receptors) and therefore is important in viral pathogenesis. It is critical to the host immune response to rabies virus infection because it is responsible for the induction of VNA as well as being the target of VNA, and it is a target for virus-specific helper and cytotoxic T cells.
Rabies virus G is a fusion protein that mediates virus entry into host cells. Following binding to its receptor(s) on target host cells, the virus is internalized, and the G spike fuses in a low-pH-dependent process with the endosomal membranes as it enters the endosome. In this process, the G goes through significant and critical conformational change whereby it assumes at least three structurally distinct conformational states (Gaudin etal., 1991b, 1993, 1995a, 1999). Prior to virus binding to the cellular receptor, the G on the virion surface is in its native state. After the virus attaches to the receptor and the virus is internalized, the G is activated to a hydrophobic state, enabling it to interact with the hydrophobic endosomal membrane. On entering the endosomal compartment and low-pH environment of the cellular compartment, the fusion capacity of the G is activated via a major structural change in the G that exposes the fusion domain, which interacts with and destabilizes one or both of the participating membranes (Gaudin et al., 1995a). The low-pH-induced fusion domain, which is thought to lie between amino acids 102 and 179 (Gaudin et al., 1995a), is not to be confused with the proposed fusogenic domain on the rabies virus G that appears to be involved in pH-independent (neutral pH) cell fusion (Morimoto et al., 1992). The pH for endosomal membrane fusion in the viral entry process is 6.2-6.3. After fusion, the G assumes a reversible fusion-inactive conformation, which makes the G monomer appear longer than the native conformation and assume selective antigenic distinctions (Gaudin et al., 1993). The fusion-inactivated G, which is no longer relevant to the fusion process, is highly sensitive to cellular proteases and appears to be in a dynamic equilibrium with the native G that is regulated by lowering and raising the pH (Gaudin et al., 1991b, 1996). Interestingly, the fusion-inactive conformation serves the G in another capacity. During nascent viral protein synthesis, the G assumes an inactive state-like conformation, which protects the G posttranslationally from fusing with the acid nature of Golgi vesicles while it is transported through the Golgi stacks to the cell surface, where it acquires its native conformation and structure (Gaudin et al., 1995b).
Mutations in the rabies virus G play a critical role in viral pathogenesis. Amino acid Arg-333 (or Lys-333) in the wild-type (normal) G is responsible for the virulence phenotype of rabies virus. Virus variants that have a glutamine (Gin), isoleucine (lie), glycine (Gly), methionine (Met), or serine (Ser) substituted for Arg-333 in the G express a phenotype that is either less pathogenic or avirulent in comparison with the parental wild-type virus when inoculated intracerebrally into adult immunocompetent mice (Dietzschold et al., 1983b; Seif et al., 1985; Tuffereau et al., 1989). It is remarkable that this single amino acid substitution (e.g., Gin for Arg-333) can affect the rate of virus spread from cell to cell (Dietzschold et al., 1985) as well as the neuronal pathway(s) that the virus takes to reach the CNS (Kucera et al., 1985). Interestingly, even the transsynaptic spread of the virus requires the envelope spike G, again showing the critical role of the G in viral pathogenesis (Etessami et al., 2000).
It is to the advantage of rabies virus that it is capable of spreading from cell to cell in tissue culture without budding into the culture medium, where it would be neutralized in the presence of antirabies VNA (Dietzschold et al., 1985). While the precise mechanism of direct cell-to-cell spread of virus is not clear, the observation that virus is internalized by a cell without being compromised, i.e., prevented from attaching to the cell surface receptor, points to importance of the fusion function of the G in virus spread. When rabies virus of the virulent phe-notype was used to infect cultures of neuroblastoma (NA) cells and baby hamster kidney (BHK) cells in the presence of antirabies VNA, the virus spread throughout the NA cell culture, whereas the avirulent virus failed to spread. The two viruses spread cell to cell equally well in the BHK cell culture (Dietzschold et al., 1985). In many ways, the pathogenic and avirulent viruses behaved in vitro in a manner that reflects their ability to spread in vivo after direct inoculation into the brain of the mouse (Dietzschold et al., 1985). In vivo, the pathogenic virus spreads more rapidly in the CNS and infects more neurons than the avirulent virus. Could it be that the fusion function of the G was altered due to the amino acid substitution at position 333? Others have suggested, based on observations in NA and BHK cells, both of which constitutively expressed the G of rabies virus, that only the pathogenic type G (with arginine in position 333) demonstrated an ability to induce syncytium formation (cell-cell fusion) at neutral pH (pH-independent fusion) in the NA cell culture. Thus the G with arginine in position 333 without other viral proteins, as in the pathogenic virus, and not the G with glutamine in position 333 (as in the avirulent virus) has the ability to mediate virus spread among neuronal cells. Moreover, since some Gln-333 variants can kill adult immunocompetent mice when infected by stereotaxic inoculation (Yang and Jackson, 1992), it appears that the G Arg-333 is essential not only for the neuropathogenicity but also for the axonal/transsynaptic spread of the virus in vivo. It is apparent also that the pH-independent cell fusion induced by the rabies virus G may involve the interaction of one or more neuronal cell-specific host cell factors, which are expressed in the NA cells but not in BHK-21 cells. Another difference between the pathogenic and avirulent viruses is the ability of fixed rabies virus strains of the pathogenic phenotype to invade the CNS from a peripheral site (Kucera et al., 1985; Etassami et al., 2000). In these studies it was suggested that the selection of different neuronal pathways to the brain from a peripheral site of inoculation might account for the pathogenic virus reaching the brain faster than the avirulent phenotype. Yet another study shows that the difference is not in the rate and pathway of virus spread to the brain after peripheral inoculation but rather that the pathogenic virus infected many more neurons than did the avirulent virus (Jackson, 1991). In such cases, the pathogenic virus may use different sites for entry to the CNS, and the avirulent virus will fail to penetrate some of the same sites because they involve use of different receptors that the avirulent virus can no longer recognize. Since it is possible in the research efforts so far that all the rabies virus-specific receptors have not been identified or proven biologically, one can only speculate on the reasons why virus of the avirulent phenotype is sometimes unsuccessful or inefficient in spreading to the CNS.
The G of rabies virus is also of major importance immunologically for the induction of the host immune response against virus infection, and because of this, it is probably the most extensively studied rabies virus antigen (Dietzschold et al., 1988; Benmansour et al., 1991). The G induces conformational and linear epitope-specific VNA and stimulates helper as well as cytotoxic T-cell activity. Consequently, studies to functionally map and then physically locate specific epitopes within antigenic domains for binding antibody and for binding T cells to the rabies virus G have been an ongoing and extensive process. At least eight antigenic sites (I—VI, "a", and Gl) have been located on the ectodomain (amino acid residues 1-439) of the G of different virus strains (Lafon et al., 1984; Prehaud et al., 1988; Dietzschold et al., 1988, 1990; Benmansour et al., 1991). Sites I, III, VI, and "a" involve the amino acids located at positions 231, 330-338, 264, and 342, respectively. Site II is a discontinuous antigenic site that involves two separate stretches of amino acids in position 34-42 and 198-200 that presumably are linked by a disulfide bridge (Prehaud et al., 1988). Sites VI and Gl are defined as linear or nonconformational, whereas the others are conformational and readily destroyed on denaturation. Epitopes recognized by T cells have been mapped on the G using chemically cleaved and synthetic peptides or T-cell lines and clones derived from individuals immunized with rabies virus vaccine (Macfarlan et al., 1984, 1986; Celis et al., 1988a,b). These fine mapping studies have given some limited insight into the structue of the various functional domains on the G of rabies virus. To gain a fuller understanding of the function of the rabies virus G, it is necessary to determine its overall conformation and, ultimately its three-dimensional structure. However, crystallization of viral membrane glycoproteins is difficult, due to the TM domain and to oligosaccharide microheterogeneity, and attempts at crystallization of the rabies virus G have so far not been successful.
Was this article helpful?
All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.