Mapping the domains on the phosphoprotein of bovine respiratory syncytial virus required for N-P and P-L interactions using a minigenome system

Bovine respiratory syncytial virus (BRSV) is a major cause of respiratory disease in calves, resulting in substantial economic losses to the cattle industry (Van der Poel et al., 1994 ; Collins et al., 1988 ; Stott & Taylor, 1985 ). BRSV is an enveloped RNA virus which, along with human respiratory syncytial virus (HRSV) and pneumonia virus of mice, belongs to the genus Pneumovirus of the family Paramyxoviridae. BRSV and HRSV are similar in gene and protein compositions. The genome of BRSV is a single-stranded negative-sense RNA of 15140 nt (Buchholz et al., 1999 ). The BRSV genome encodes 11 proteins (Lerch et al., 1989 ; Mallipeddi et al., 1990 ). Four proteins are associated with the genomic RNA, namely nucleocapsid (N) protein, phosphoprotein (P), the large polymerase (L) protein and the transcription anti-termination factor M2-1. Three proteins are transmembrane components of the envelope; namely, fusion (F), attachment glycoprotein (G) and the small hydrophobic (SH) proteins. The matrix (M) protein is an inner virion protein. NS1 and NS2 are nonstructural proteins.

The P protein is a component of the ribonucleoprotein (RNP) complex, which contains the N-RNA nucleocapsid associated with N, P, M2-1 and L proteins. The N protein encapsidates the genomic RNA, and is, therefore, present in large quantities. The P protein is a major phosphorylated protein and is present in smaller quantities. The L protein, which is thought to form the enzymatic component of virus RNA-dependent RNA polymerase, is also present in smaller quantities. The RNP complex is required for transcription and replication of viral genome (Collins et al., 1996 ). During transcription, gene-start and gene-end signals on the viral genome are recognized, and 10 capped and polyadenylated mRNAs are produced. Replication, on the other hand, results in an encapsidated, full-length, antigenomic RNA. It is not completely understood how the switch between transcription and replication occurs. Studies with Sendai and vesicular stomatitis viruses suggest that the polymerase complex (PL) is required for both transcription and replication; whereas, replication requires an additional complex composed of unassembled N and P proteins (Curran et al., 1991 ; Horikami et al., 1992 ). However, a complete understanding of the transcription and replication will require knowledge about how different proteins of the RNP complex interact with each other.

We have previously used an in vitro protein-blotting protein overlay technique (Samal et al., 1993 ) and a yeast two-hybrid system (Mallipeddi et al., 1996 ) to study the interaction of P and N proteins. However, these systems might not reflect the true proteinprotein interactions occurring in virus-infected cells. Recently, we developed a BRSV minigenome system to map the domains of N protein interacting with P protein (Khattar et al., 2000 ). In this report, we used this system to map the domains of P protein necessary for interacting with N and L proteins.

In order to identify the regions of P protein required for interaction with N and L proteins, a total of 12 deletions of 20 amino acids each were made throughout the 241 amino acid P protein (mutants PΔ1-20, PΔ21-40, PΔ41-60, PΔ61-80, PΔ81-100, PΔ101-120, PΔ121-140, PΔ141-160, PΔ161-180, PΔ181-200, PΔ201-220 and PΔ221-241, shown schematically in Fig. 1). Plasmid pTM1-P, containing the BRSV P gene (A51908 strain) cloned downstream of the T7 RNA polymerase promoter of plasmid pTM1, has been used to generate these mutants. All the internal deletions of the P gene of BRSV were generated by a single round of PCR using the 5' phosphorylated internal primer pair (Byrappa et al., 1995 ). All the P gene mutants were sequenced in their entirety and confirmed for correct P protein expression by in vitro transcription and translation, using the rabbit reticulocyte lysate system (Promega). To determine the amount of expression of these mutant proteins, HEp-2 cells were transfected with either wild-type or mutant pTM1-P plasmids, along with other minigenome system components as described later, and infected with modified vaccinia virus Ankara (MVA) expressing T7 RNA polymerase (a gift from B. Moss, NIH, USA). Expression of the wild-type or mutant P proteins was determined by Western blot assays using P-specific antiserum. As shown in Fig. 2, all the deletion mutant P proteins accumulated intracellularly to a level comparable to that of wild-type P protein; thus, the defect in the mutant P proteins to interact with N or L protein, described later, was not due to differences in expression or stability of various mutant P proteins.

(12K): Fig. 1. Schematic representation of all the mutant BRSV P proteins that were used for NP and PL interaction studies. The names of the mutant P proteins indicate the deleted amino acids and the remaining sequences are shown as solid bars.

(46K): Fig. 2. Expression of wild-type or mutant BRSV P proteins in transfected HEp-2 cells. HEp-2 cells were infected with vaccinia virus (MVA-T7) and transfected with plasmids encoding the BRSVCAT minigenome (0·4 µg), wild-type (WT) or mutant BRSV P protein as indicated (0·2 µg), N protein (0·4 µg) and L protein (0·1 µg). At 48 h post-transfection, cells were harvested and lysates were electrophoresed on a 420% gradient gel under reducing conditions. The separated proteins were probed in Western blot with P-specific antiserum. BRSV-specific P polypeptide is indicated to the right.

We have utilized a BRSV minigenome system to identify the regions of BRSV P protein that interact with N and L proteins. The advantage of this system is that proteinprotein interactions occur under conditions of reconstituted BRSV transcription and replication, and thus, it is probably a close facsimile of the authentic situation. In this system, a plasmid-synthesized negative-sense BRSV minigenome containing the CAT gene is transcribed and replicated by BRSV N, P and L proteins supplied in trans from plasmids, using the vaccinia virus/T7 RNA polymerase expression system (Yunus et al., 1998 ). Construction of cDNA encoding the BRSVCAT minigenome and construction of support plasmids pTM1-P, pTM1-N and pTM1-L containing P, N and L genes of the BRSV A51908 strain, respectively, cloned downstream of the T7 RNA polymerase promoter of plasmid pTM1 have been described previously (Yunus et al., 1998 ). Specifically, HEp-2 cells were infected with the MVA strain of vaccinia virus and transfected with a mixture of plasmids encoding the BRSV minigenome (0·4 µg), wild-type or mutant BRSV P protein (0·2 µg), N protein (0·4 µg) and L protein (0·1 µg) by using LipoFECTAMINE (Life Technologies). At 24 h post-transfection, cells were incubated in the presence of [³⁵S]methionine for 5 h, and lysates were prepared in a buffer containing 50 mM TrisHCl (pH 7·5), 150 mM NaCl and 0·7% Nonidet P-40. The cell lysates were subjected to immunoprecipitation with polyclonal P-specific antiserum (Fig. 3). Preparation of P-specific antiserum was described earlier (Khattar et al., 2000 ). Using P-specific antiserum, N protein (Fig. 3A, lane 3) and L protein (Fig. 3B, lane 3), coprecipitated with the P protein, indicating NP and PL complex formation, respectively, in transfected cells. These coimmunoprecipitations were not due to nonspecific precipitation of either N or L protein, since P-specific antiserum did not precipitate N protein (Fig. 3A, lane 1) or L protein (Fig. 3B, lane 1) in the absence of P protein.

(46K): Fig. 3. Identification of BRSV P protein regions involved in interaction with N and L proteins. HEp-2 cells were transfected with the minigenome system components, as described in the legend to Fig. 2, and with plasmids encoding wild-type or mutant P proteins as indicated. The cells were pulse-labelled at 24 h post-infection or -transfection for 5 h with [35S]methionine and lysates were processed and subjected to immunoprecipitation with P-specific serum. Under these conditions, N protein (A) or L protein (B) coprecipitated due to its interaction with P protein. Immunoprecipitated proteins were separated by 10% SDSPAGE under reducing conditions. As negative controls, the cells in lanes 1 and 2 in (A) did not receive plasmids encoding WT P or N proteins, respectively, and the cells in lanes 1 and 2 in (B) did not receive plasmids encoding WT P or L proteins, respectively. The molecular mass markers are indicated to the left (in kDa). BRSV-specific P, N or L polypeptides are indicated to the right.

To identify domains on the BRSV P protein required for binding to N protein, all the P deletion mutants generated above were tested for their ability to bind N protein in a coimmunoprecipitation assay with P-specific antiserum (Fig. 3A). Whereas the amount of labelled coimmunoprecipitated P protein varied in these mutants due to deletion of a variable number of methionine residues, the amount of N protein brought down by P-specific antiserum was essentially the same for all the P mutants. Deletion of 21 amino acids at the C terminus of the P protein (mutant PΔ221-241) completely abolished the interaction with the N protein (Fig. 3A, lane 15). On the other hand, a deletion of 20 amino acids at the N terminus (mutant PΔ1-20) did not abolish the interaction with the N protein (Fig. 3A, lane 4). Each of the internal deletion mutants formed a complex with the N protein (Fig. 3A, lanes 511, 13 and 14), except for PΔ161-180 (Fig. 3A, lane 12). Thus, an internal region of 161180 amino acids is also important for interaction with P protein. Taken together, these results indicated that two independent N-binding regions exist on P protein: one is located between amino acids 161 and 180, the other is located in the C-terminal part between amino acids 221 and 241. A similar study of NP interactions in various other non-segmented RNA viruses has suggested that the C terminus of P protein is particularly important in interaction with the nucleocapsid protein N or NP (Ryan & Portner, 1990 ; Takacs et al., 1993 ; Chenik et al., 1994 ; Zhao & Banerjee, 1995 ; Barr & Easton, 1995 ; Garcia-Barreno et al., 1996 ; Mallipeddi et al., 1996 ). Studies in rabies virus (Chenik et al., 1994 ) and Sendai virus (Ryan & Portner, 1990 ) indicated the existence of an internal region, in addition to a C-terminal region, important for binding to P protein. Using a yeast two-hybrid system for BRSV (Mallipeddi et al., 1996 ), and protein overlay blot system for pneumonia virus of mice (Barr & Easton, 1995 ), the N terminus of P protein has been shown to be important for NP interactions. Our data did not indicate the involvement of the N terminus in interactions with P protein. This may be due to the fact that these interaction studies were performed in different systems. Moreover, in the yeast two-hybrid system, the N terminus of the protein may not be available to interact efficiently with the target protein.

To identify domains on the BRSV P protein required for binding to L protein, all the P mutants were tested for their ability to bind L protein in a coimmunoprecipitation assay with P-specific antiserum. We found that the PL interaction occurred in the presence or absence of N support plasmid; however, the PL interaction was stronger in the absence of N support plasmid. Therefore, we omitted N support plasmid from the transfection mixture for PL interaction studies. Deletion of 20 amino acids from the N-terminal end (mutant PΔ1-20) and 21 amino acids from the C-terminal end (mutant PΔ221-241) had no effect on complex formation between P and L proteins (Fig. 3B, lanes 4 and 16). Each of the internal deletion mutants strongly interacted with L protein (Fig. 3B, lanes 59 and 1315) except for PΔ121-140 (lane 10) and PΔ141-160 (lane 11). The interaction of mutants PΔ121-140 and PΔ141-160 with L protein was weak. This indicated that these two regions are important for interaction with L protein. To further confirm our findings, we constructed deletion mutant PΔ121-160, in which amino acids 121140 and 141160 were deleted. As shown in Fig. 3(B, lane 12), interaction of mutant PΔ121-160 with L protein was completely abolished, which confirms and further delineates the importance of this domain for interaction with L protein. The additional bands immunoprecipitated with anti-P antibody, and detected between P and L protein bands in Fig. 3(B, lanes 311 and 1316) may be either due to degradation of L protein or due to coimmunoprecipitation of some cellular proteins associated with PL complexes. Our results are in agreement with the previous findings reported for rabies virus (Chenik et al., 1998 ) and Sendai virus (Smallwood et al., 1994 ) that N and L protein binding sites on P protein do not overlap. This correlates well with the concept that P protein can interact simultaneously with L and N proteins to act as transcription factor when complexed with L protein, and as a replication factor when complexed with N protein.

In conclusion, we have identified amino acids at position 121160 and amino acids at positions 161180 and 221241 of the P protein which are important for binding to L and N proteins, respectively. Although our results probably represent identification of authentic sites of interaction between P and L proteins, it must be considered that N protein was not included in our study due to the reasons mentioned above. Sequence comparison of the P protein of BRSV with other pneumovirus P proteins indicated a high degree of amino acid similarity from residues 128 to 184 (Ling et al., 1995 ). Computer-assisted analysis (Lupas et al., 1991 ) of amino acids 100241 of the P protein indicated the formation of coiled coils at the immediate C terminus of the protein. The introduction of two proline residues at positions 220 and 236, which would be expected to prevent the formation of α-helices, has been shown to result in reduced ability of HRSV P protein to interact with N protein (Slack & Easton, 1998 ). It appears, thus, that the α-helices and coiled coils present in this region may play an important role in NP and PL interactions. Further, the coiled coil structure has been found to be important for oligomerization in several paramyxoviruses. Various reports in paramyxoviruses indicated the role of oligomerization in P protein functionality. In HRSV, domains for oligomerization have been mapped in the region between 99 and 165 amino acids (Asenjo & Villanueva, 2000 ). Whether the same region is important for oligomerization in BRSV is unknown. If it is so, then it is possible to speculate that disruption of oligomerization in this region of BRSV may prevent interaction of P protein with L protein. Further studies are needed to map the region of BRSV P protein important for oligomerization. Finally, it is important to note that the alterations in the P protein could cause conformational changes. Thus, the actual interactive site(s) may not be present within the deleted domains. The change of conformation of the mutant proteins may directly affect the interacting site(s) present elsewhere in the protein. Detailed studies by the use of peptides would certainly help us to understand the interacting domains of the P protein.

We thank Daniel Rockemann for excellent technical assistance. This work was supported by USDA grant 9702414.