The effect of foreign gene insertion in position 1 or 2 of the b/h PIV3 genome on in vitro and in vivo virus replication and foreign antigen expression was evaluated in this study. The three antigens chosen were the RSV F, RSV G, and hMPV F surface glycoproteins, all of which have the potential to be incorporated into the b/h PIV3 virion. We wanted to study whether all foreign membrane glycoproteins expressed by b/h PIV3 would produce viruses with similar phenotypes or whether the type of protein inserted would yield viruses with different replication characteristics. PIV3 genome positions 1 and 2 were expected to yield the highest levels of foreign protein expression since paramyxoviruses transcribe the 3′-most proximal genes with the highest frequency. Foreign gene insertions in position 1 could interfere with virus replication by delaying and/or reducing bPIV3 N protein expression, which now takes place from genome position 2. Insertion in the first position also has a greater potential for effecting initiation of transcription and replication of the antigenome. In contrast, gene insertions in position 2 should alter the N-P mRNA ratio, which may impair virus replication as well. Therefore, it was of interest to test which of the genome positions would better tolerate foreign gene insertions without compromising virus replication.
Our studies showed that b/h PIV3 expressed the RSV F, RSV G, or hMPV F proteins efficiently from genome positions 1 and 2. Curiously, the 3′ proximal location of the foreign antigens did not result in enhanced protein expression compared to that of wild-type RSV, as expected from the transcription gradient known to be generated by nonsegmented negative-strand RNA viruses. Chimeric b/h PIV3/RSV G expressed slightly less RSV G protein from position 1 than from position 2; however, this effect was not observed for b/h PIV3/RSV F1 or b/h PIV3/hMPV F1. The RSV F0
protein expressed by PIV3 was processed into F1
, like the RSV F protein derived from wild-type RSV. A smaller proteolytic fragment accumulated to large amounts in lysates derived from b/h PIV3/RSV F1 and F2, which was not observed in wild-type RSV lysates. The precise origin of this peptide is not known for certain, but the antibodies used to probe the Western blot recognized epitopes in the N-terminal half of the glycoprotein. While the precursor RSV F0
and the processed RSV F1
were observed for b/h PIV3/RSV F1 and F2, F protein processing was not observed for hMPV F expressed by b/h PIV3 from position 1 or 2. b/h PIV3/hMPV F1 or F2 displayed only protein bands corresponding in size to the uncleaved F0
protein precursor. Analysis of the F protein cleavage site revealed that the hMPV F protein cleavage site consisted of multiple charged amino acid residues (RQS
R) but differed from the conserved furin protease cleavage site seen in related viruses like RSV or APV (RKR
R and RRR
R, respectively). Underlining indicates altered amino acids that are different in hMPV. Sendai virus, a paramyxovirus with a nonconsensus furin protease cleavage site needs exogenous protease for multiple rounds of replication in tissue culture (19
). The “weak” cleavage site of the hMPV F protein may therefore be responsible for the presence of only the F0
protein detected by immunoprecipitation in this study, since the F1
fragments would be present only at low levels. Inefficient F protein cleavage may hinder the cell-to-cell spread of hMPV, accounting for slow replication in tissue culture, and may explain the trypsin requirement of some hMPV strains (25
PIV3 harboring foreign genes in position 1 was more difficult to recover by reverse genetics than viruses harboring gene insertions in position 2. This may be due to an impairment of bPIV3 N protein levels, since the N gene was moved to genome position 2 in b/h PIV3 harboring foreign genes in position 1. An analysis of the kinetics of virus replication showed that viruses harboring foreign gene insertions in position 2 replicated to peak titers observed for b/h PIV3, the virus vector. In contrast, viruses containing the gene insertion in position 1 of the PIV3 genome displayed a delayed onset of peak virus replication and reduced virus peak titers compared to b/h PIV3. This was most pronounced for b/h PIV3/hMPV F1 and b/h PIV3/RSV G1, which suggested that the type of foreign gene inserted into the PIV3 genome also has an effect on virus replication, even though all of the antigens are surface glycoproteins with similar functions. All chimeric PIV3 viruses with gene insertions in position 2 replicated to high titers of 107 to 108 PFU/ml in Vero cells, and thus position 2 is a better position for maximum level of foreign gene expression.
To study whether foreign gene insertions had an effect on in vivo virus replication, hamsters were infected intranasally with the chimeric b/h PIV3 expressing RSV or hMPV proteins. Only b/h PIV3 expressing the RSV F or RSV G protein in position 1 displayed restricted replication in the lower respiratory tract of hamsters. However, protection from challenge with hPIV3 or RSV was not affected by the lower levels of replication in the lungs of hamsters observed for b/h PIV3/RSV G1 or F1. Notably, b/h PIV3/RSV F1 displayed lower titers of RSV neutralizing and HAI serum antibodies than b/h PIV3/RSV F2. High levels of replication were observed for b/h PIV3/hMPV F1 and F2, but b/h PIV3/hMPV F1 protected immunized hamsters only partially in the upper respiratory tract. Additional studies are necessary to determine the reason for partial protection displayed by b/h PIV3/hMPV F1. The levels of RSV neutralizing and PIV3 HAI antibody titers of b/h PIV3/hMPV F1 and F2 were not affected by genome position, because both viruses elicited antibody titers equivalent to those observed for wild-type hMPV.
b/h PIV3 had been used previously to express RSV proteins and to demonstrate immune protection in hamsters and primates (20
). In these studies, the genes encoding the RSV F or G protein were inserted singly in the 3′ proximal position of the PIV3 genome or inserted in tandem as a combination of RSV G and F. Characterization of RSV protein expression did not reveal a good correlation between genome position and the amount of RSV protein expressed. The data showed that RSV F protein was expressed at smaller amounts from position 1 in the PIV3 genome (b/h PIV3-FA
) than from the RSV F gene located in position 2 in b/h PIV3 GA
, which is not expected in the presence of a transcriptional gradient. Similarly, RSV G protein expression for the single RSV G gene inserted at the 3′ proximal position was higher than that observed for b/h PIV3 RSV GA
, in which the RSV G gene is also located at the 3′-most proximal end of the viral genome (21
). These results suggested that PIV3 genome location alone may not be the sole determinant for levels of foreign gene expression. The gene start or gene end sequences and/or RNA elements that direct mRNA stability or translation efficiency may also play a role in determining efficiency of viral mRNA transcription and viral protein expression. Interestingly, the RSV neutralizing antibody titers present in sera derived from hamsters infected with b/h PIV3/RSV F1 or G1 in the study by Schmidt et al. (20
) were higher than those in our data, although the neutralization antibody titers for wild-type RSV sera from both studies were identical. This difference in antibody titers may be due to higher replication levels of b/h PIV3/RSV F1 or G1 generated by Schmidt et al. in the respiratory tract of hamsters. b/h PIV3/RSV F1 and G1 displayed titers 0.5 and 1.7 log10
higher, respectively, in the upper respiratory tract and 1.1 and 2.2 log10
higher, respectively, in the lower respiratory tract of hamsters than the replication titers observed in this study (20
). The increased levels of viral replication in the respiratory tract of hamsters may result in higher RSV neutralizing antibody titers. The origins of the hPIV3 F and HN genes, the Texas strain for the b/h PIV3 generated in this study and the JS strain for the b/h PIV3 generated by Schmidt et al. (20
), may direct the in vivo hamster replication efficiencies observed for the chimeric viruses. The neutralizing antibody titers of b/h PIV3/RSV F1 and G1 hamster sera were also higher than wild-type RSV serum titers (20
). In contrast, the neutralizing antibody titers of wild-type RSV-infected rhesus monkeys were higher than those observed for b/h PIV3/RSV F1 or G1 primate sera when the same viruses were evaluated in rhesus monkeys, albeit wild-type RSV replicated to lower titers in rhesus monkeys than chimeric PIV3/RSV (21
). In order to resolve these discrepancies, it would be necessary to compare the b/h PIV3/RSV F1 and G1 viruses generated in the two different laboratories directly in the same study to determine whether the origin of the hPIV3 surface glycoprotein genes and/or other genetic components, such as the types of gene start and stop sequences, can influence virus replication and immunogenicity.
In this study, the chimeric b/h PIV3 viruses expressing the RSV or hMPV surface glycoproteins were shown to function as bivalent vaccines in a small animal model. These viruses have characteristics that make them suitable for evaluation as live attenuated virus vaccines for RSV, hMPV, and hPIV3. The chimeric viruses grow to high titers in tissue culture. The inserted antigens of the chimeric b/h PIV3 were genetically stable and maintained up to passage 10 in Vero cells. In vaccinated hamsters, the viruses elicited a protective immune response upon challenge with wild-type hPIV3, RSV, and hMPV. The challenges were carried out with homotypic viruses; however, the degree of amino acid identity is very high for the two subgroups of RSV A and B (89%), as well as the two subgroups of hMPV A and B (95%), and therefore immunological cross protection is expected to occur. Although this remains to be tested directly, RSV F MAbs such as Synagis can neutralize RSV originating from subgroup A as well as subgroup B. The chimeric viruses induced an effective neutralizing antibody response to the vectored non-PIV3 glycoproteins despite not being incorporated into the virion. The latter finding should allay concerns that a change in viral tropism is not likely as a consequence of new virus-cell receptor interactions of the b/h PIV3 expressing glycoproteins from RSV or hMPV. The b/h PIV3/RSV F2 and b/h PIV3/hMPV F2 viruses will be further evaluated for safety and efficacy in a primate model.
This study demonstrated the utility and versatility of b/hPIV3 as a virus vector for expression and delivery of three different foreign viral antigens: RSV F, RSV G, and hMPV F. The ease, speed, and effectiveness of generating potential vaccine candidates for newly discovered viral pathogens for which infectious cDNAs and recombinant virus recovery systems are not available further underscore the importance of developing b/h PIV3 as a virus vaccine vector.