INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Human parainfluenza virus type 3 (PIV3), a member of the genus Paramyxovirus of the family Paramyxoviridae, has a single-stranded, negative-sense RNA genome that is 15,462 nucleotides (nt) in length. PIV3 is a major cause of serious lower respiratory illness requiring hospitalization of infants and young children (4). A vaccine is needed to prevent the severe disease caused by this virus, and two live attenuated candidate PIV3 vaccines are currently being evaluated in humans (21, 22). One of these is a bovine strain of PIV3 that is discussed elsewhere (21). The other was produced by passaging the human PIV3 wild type (wt), JS strain, at low temperature for 45 passages to yield the PIV3 cold-passaged 45 (cp45) candidate vaccine virus (1). The cp45 vaccine virus possesses temperature-sensitive (ts), cold adaptation (ca), and attenuation (att) phenotypes (1, 6). The att phenotype is manifested by attenuation of replication in the upper and lower respiratory tracts of rodents and nonhuman primates (6, 15, 16). In addition, the virus appears to be satisfactorily attenuated, phenotypically stable, and immunogenic in seronegative infants and children (22) and therefore is a promising vaccine candidate. Comparison of the complete nucleotide sequences of the cp45 and wt (JS strain) viruses indicated that cp45 possesses multiple point mutations in coding and noncoding regions of the genome, including three point mutations in the L polymerase gene that each encode an amino acid substitution (34).

Previously, we recovered recombinant PIV3 from a full-length antigenomic cDNA clone of wt PIV3 (JS strain), the parent of cp45, and demonstrated that the recovered virus was not temperature sensitive and replicated to a level in the respiratory tract of rodents comparable to that of the biologically derived wt (JS strain) virus (9). This meant that it was now possible to systematically examine the genetic basis of the att phenotype of PIV3 candidate vaccines such as cp45. Since the polymerase genes are the sites of many att and ts mutations for influenza virus and respiratory syncytial virus (RSV) (8, 11, 20, 24, 32) and since preliminary data suggested that the L gene of cp45 possesses a ts mutation (30), we initiated our studies to examine the genetic basis of attenuation of cp45 by introducing the mutations yielding the seven possible combinations of the three amino acid substitutions present in the L gene of cp45 into the cDNA clone of its wt (JS strain) parent. Seven recombinant viruses (three single, three double, and one triple mutant) were isolated and analyzed for their ts and att phenotypes. Analysis of these mutants indicated that each of the three mutations in the L protein is a major separate contributor to the ts and att phenotypes of this promising vaccine candidate. Furthermore, this study illustrates the usefulness of the newly developed reverse-genetics systems for characterizing and manipulating a nonsegmented negative-strand virus.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Viruses and cells. The PIV3 wt (JS strain) and cp45 viruses were grown in simian LLC-MK2 cells as described previously (16). The vTF7-3 recombinant vaccinia virus (12) and the modified vaccinia virus Ankara (MVA-T7) (36), which each express the T7 polymerase, were kindly provided by Linda Wyatt and Bernard Moss. HEp-2 (ATCC CCL 23) and LLC-MK2 (ATCC CCL 7.1) cells were maintained in OptiMEM (Life Technologies, Gaithersburg, Md.) supplemented with 2% fetal bovine serum (FBS) and gentamicin sulfate (50 µg/ml). L-132 cells (ATCC CCL 5) were grown in Earle's minimal essential medium (Life Technologies) supplemented with 10% FBS, 2 mM glutamine, 20 mM HEPES, 1 mM nonessential amino acids, and 100 U of streptomycin-neomycin/ml.

Construction of point mutations in the L gene of PIV3. pUC19 was modified to accept a fragment of the wt (JS strain) PIV3 L gene to introduce point mutations into the L gene by site-directed mutagenesis. First, a unique NheI restriction site in pUC19 was introduced, and the naturally occurring BamHI site was ablated by ligating a pair of complementary oligonucleotides (5' GATCGATGCTAGCCC 3' and 5' GATCGGGCTAGCATC 3') containing an NheI restriction site into the BamHI site of pUC19 to create pUC19(N). We previously described the construction and functional testing of pTM(L), which is a T7 expression plasmid that contains a complete copy of the L translational open reading frame (9, 10). The SphI-to-NheI (PIV3 nt 11317 to 14087) fragment of pTM(L) (see Fig. 1), which includes the positions where the three coding changes in cp45 occur and which can be directly introduced into the full-length PIV3 cDNA (see below), was cloned into the SphI and NheI sites of pUC19(N) to create pUCL(N-S). Point mutations were introduced into pUCL(N-S) by using mutagenic oligonucleotides with the Transformer Mutagenesis kit (Clontech, Palo Alto, Calif.) for the purpose of (i) creating the desired amino acid substitutions at L protein positions 942, 992, and 1558, individually and in combination, and (ii) ablating one specific naturally occurring restriction enzyme site proximal to each codon substitution as a marker (see Table 1). Mutations introduced in pUCL(N-S) derivatives were verified by dideoxynucleotide sequencing of plasmid DNA. The NheI-to-PinAI (PIV3 nt 14087 to 12485) fragment of pUCL(N-S) encoding the cp45 mutation at position 1558 was subcloned into the NheI-to-PinAI sites of pUCL(N-S)-942, -992, and -942/992 to give pUCL(N-S)-942/1558, -992/1558, and -942/992/1558. The SphI-to-BamHI (PIV3 nt 11317 to 13733) fragment containing all seven point mutations in pUCL(N-S) was subcloned into the SphI-to-BamHI window of pTM(L) (Fig. 1). The nucleotide sequence of the 2.7-kb SphI-to-NheI fragment of pTML(L)-942/992/1558 was determined and was found to differ from the wild-type sequence only in the intended positions. Seven pTM(L) plasmids containing each of the possible configurations of the cp45 L gene mutations were tested at a permissive temperature (32°C) for their ability to direct the expression of the chloramphenicol acetyl transferase marker gene in the previously described minireplicon system consisting of plasmid-encoded minigenome RNA and N, P, and L proteins (10). The various mutant L plasmids supported marker gene expression at 75 to 106% of the level of wt L (data not shown), indicating that each engineered cDNA was free of major spurious mutations. The SphI-to-NheI fragments of each of the mutant pTM(L) plasmids were then subcloned into the SphI-to-NheI window of the full-length PIV3 JS antigenomic cDNA plasmid p3/7(131)2G+ (9) to create the seven full-length PIV3 cDNA clones, which represent every possible combination of the three amino acid substitutions.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Map of plasmid pTM(L)-942/992/1558, which contains the PIV3 L gene cDNA coding for amino acid substitutions at positions 942, 992, and 1558. The relative position of each coding change is indicated, together with the amino acid difference and the naturally occurring restriction site which was deliberately ablated as a marker. Restriction sites used for cloning (SphI, PinAI, BamHI, and NheI) are indicated. The arrow shows the direction of the L open reading frame (ORF); corresponding amino acid positions are indicated.

Recovery of recombinant PIV3 (rPIV3) bearing one, two, or three cp45 L protein amino acid substitutions. Each full-length antigenomic cDNA bearing one or more cp45 L gene mutations, together with the three support plasmids pTM(N), pTM(P), and pTM(L), was transfected into HEp-2 cells on six-well plates (Costar, Cambridge, Mass.) by using LipofectACE (Life Technologies) and MVA-T7 as described previously (9). After incubation at 32°C for 4 days, the transfection harvest was passaged onto HEp-2 cells on six-well plates, which were incubated at 32°C for 4 days. Each passage-1 supernatant was harvested and transferred to a T-25 flask of LLC-MK2 cells, which was incubated at 32°C for 5 to 6 days. The passage-2 supernatant was harvested, and the presence of recombinant virus was initially confirmed by immunoperoxidase staining of virus plaques (28) with anti-hemagglutinin-neuraminidase (anti-HN) monoclonal antibody (MAb) 77/5, which binds to both biologically derived and recombinant PIV3s, and MAb 423/6, which does not bind to cDNA-derived virus because its epitope had been deliberately ablated to serve as a marker (9). Virus present in the passage-1 supernatant was subjected to two or three rounds of plaque purification on LLC-MK2 cells as described previously (16). Each biologically cloned recombinant virus was amplified twice in LLC-MK2 cells at 32°C to produce virus for further characterization. Virus was concentrated from clarified medium by polyethylene glycol precipitation (26), and viral RNA (vRNA) was extracted with Trizol reagent (Life Technologies). Reverse transcription (RT) was performed on vRNA by use of the Superscript II preamplification system (Life Technologies) with random hexamer primers. The Advantage cDNA PCR kit (Clontech) and sense (5'-GCATTATCTAGATGTGTCTTCTGGTCAGAG-3'; nt 11190 to 11219) and antisense (5'-CCTGAATTATAATAATTAACTGCAGGTCCT-3'; nt 14140 to 14111) primers specific for the PIV3 L gene were used to amplify the region spanning the SphI-to-NheI fragment. The PCR fragments were analyzed by digestion with each of the restriction enzymes whose recognition sites had been ablated during insertion of the three cp45 amino acid substitution mutations in L (see Table 1 and Fig. 3).

Efficiency of plaque formation at permissive and restrictive temperatures of rPIV3 bearing one, two, or three cp45 L protein amino acid substitutions. The levels of temperature sensitivity of plaque formation in vitro of control and recombinant viruses were determined at 32, 37, 38, 39, 40, and 41°C in LLC-MK2 monolayer cultures as described previously (16), and plaques were enumerated by hemadsorption with guinea pig erythrocytes following removal of the methylcellulose overlay. Alternatively, the viral plaques present in the monolayer were identified by immunoperoxidase staining with a mixture of two PIV3-specific anti-HN murine MAbs, 101/1 and 454/11, diluted 1:500 (28).

Evaluation of rPIV3 mutant viruses for the ca phenotype. Growth of mutant and wt rPIV3 viruses was determined at 32 and 20°C on confluent L-132 cell monolayers prepared in 24-well tissue culture plates. Duplicate wells of each of two plates were inoculated with 0.2 ml of each mutant or wt rPIV3 virus at a multiplicity of infection of 0.01. After 1 h of adsorption at room temperature, the inoculum was aspirated and the monolayers were washed with 1 ml of Dulbecco phosphate-buffered saline (Life Technologies) per well. The inoculated cultures were overlaid with 0.5 ml of Earle's minimal essential medium supplemented with 10% FBS, 2 mM glutamine, 20 mM HEPES, 1 mM nonessential amino acids, and 100 U of streptomycin-neomycin/ml. One plate was sealed in a waterproof pouch (Kapak) and then submerged in a 20°C bath for 13 days. The duplicate plate was placed at 32°C in a CO₂ incubator for 3 days. At the end of the incubation period, virus was harvested by freeze-thawing. The titer of virus recovered from each well was determined by plaque assay in LLC-MK2 cells at 32°C by using hemadsorption with guinea pig erythrocytes to visualize the plaques. Two wt and two cp45 reference stocks were used as controls.

Hamster studies. Four- to 16-week-old golden Syrian hamsters in groups of six were inoculated intranasally with 0.1 ml of OptiMEM containing 10^5.5 PFU of wt rPIV3 (JS strain), PIV3 cp45, or one of the rPIV3s containing one or more cp45 L protein amino acid substitution(s). On day 4 postinfection, the hamsters were sacrificed, the lungs and nasal turbinates were harvested, and the virus was quantified as described previously (9). The mean log₁₀ 50% tissue culture infectious dose per gram was calculated for each group of six hamsters.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Introduction of the PIV3 cp45 L protein amino acid substitution mutations into wt rPIV3 (JS strain). Mutations yielding the three amino acid substitutions present in the L protein of cp45 (Table 1) were introduced individually or in combinations into the antigenomic cDNA that encodes its wt parent, PIV3 JS strain. Each introduced mutation was engineered to be marked with a silent mutation that ablated a proximal naturally occurring restriction enzyme site to facilitate monitoring of the mutation in recovered rPIV3 (Table 1). In the engineered virus r1558, the coding change resulting in a substitution at amino acid (aa) 1558 was designed to contain two nucleotide changes, compared to the 1-nt substitution in cp45, to reduce the chance of reversion at this site during in vitro or in vivo replication (Table 1). It was not possible to do the same for the recombinant viruses with mutations coding for changes at positions 942 and 992 (i.e., r942 and r992). Seven rPIV3s bearing one, two, or all three of the amino acid substitutions were recovered in tissue culture by transfection of each antigenomic cDNA together with the pTM(N), pTM(P), and pTM(L) support plasmids and coinfection with the vaccinia virus MVA-T7 recombinant. Each rPIV3 possessed the MAb resistance marker that had been deliberately introduced into the HN gene by engineering the antigenomic cDNA (9) (data not shown). The rPIV3s were biologically cloned by two or three cycles of plaque-to-plaque passage to ensure that each virus preparation was genetically homogeneous. This precaution was taken because vaccinia virus can mediate recombination between the antigenomic cDNA and the support plasmids, as has been demonstrated in a comparable system (14).

View larger version (32K): [in this window] [in a new window]   FIG. 2.   RT-PCR product made from rPIV3 vRNA. vRNA was isolated from virions grown in LLC-MK2 cells infected with either wt rPIV3 or the indicated mutant rPIV3. The RNA was subjected to reverse transcription (+) or was mock treated without enzyme (), followed by PCR with L-specific primers that flank the SphI and NheI sites and yield a 2,951-bp cDNA. M, marker DNA (a mixture of lambda DNA digested with HindIII and X174 DNA digested with HaeIII). Nucleotide lengths are indicated for several marker bands (in base pairs).

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Restriction enzyme digestion of RT-PCR products prepared from the L gene of wt and mutant rPIV3s. RT-PCR product (2,951 bp) flanking the SphI and NheI sites in the L gene prepared from vRNA as described in the legend to Fig. 2 was digested individually with each of the restriction enzymes described below. The sources of the vRNA template were r942 (lanes 1), r992 (lanes 2), r1558 (lanes 3), r942/992 (lanes 4), r942/1558 (lanes 5), r992/1558 (lanes 6), r942/992/1558 (lanes 7), and wt rPIV3 (lanes 8), as indicated. (A) Analysis with EaeI. EaeI cuts the wt RT-PCR product (lane 8) once and produces two fragments (2,667 and 284 bp; the latter species was not retained on the gel shown). Mutant DNAs encoding the Tyr-to-His mutation at position 942 (lanes 1, 4, 5, and 7) also have this site ablated and are therefore resistant to cleavage by EaeI. (B) Analysis with BsrI. BsrI cuts the wt RT-PCR product (lane 8) two times, producing three fragments (1,590, 922, and 439 bp; the latter species was not retained on the gel shown). Mutant DNA encoding the Leu-to-Phe mutation at position 992 (lanes 2, 4, 6, and 7) also has a proximal BsrI site ablated and produces two DNA fragments of 1,590 and 1,361 bp. (C) Analysis with AvaII. AvaII cuts the wt RT-PCR product (lane 8) three times, resulting in four fragments (2,119, 594, 210, and 28 bp; the latter two species were not retained on the gel shown). Mutant DNA encoding the Thr-to-Ile mutation at position 1558 (lanes 3, 5, 6, and 7) is also missing a proximal AvaII site and produces three fragments of 2,329, 594, and 28 bp. M, marker DNA (a mixture of lambda DNA digested with HindIII and X174 DNA digested with HaeIII). Nucleotide lengths are indicated for several marker bands (in base pairs).

Efficiency of plaque formation in LLC-MK2 cells of rPIV3s bearing the cp45 L gene mutations at permissive and restrictive temperatures. The seven rPIV3s bearing the various combinations of cp45 L protein amino acid substitutions were assayed for their ability to form plaques on LLC-MK2 monolayers at various permissive and restrictive temperatures (Table 2). As shown in Table 2, each rPIV3 bearing a cp45 L protein amino acid substitution was temperature sensitive, whereas the wt rPIV3 parent (JS strain) was not restricted in plaque formation at any temperature tested. The shutoff temperature of plaque formation for r942 and r992 was 40°C. At 40°C, r942 manifested a 300-fold reduction of plaque formation and formed plaques that were small in size, indicating that its replication was reduced at this restrictive temperature; at 41°C, it failed to form plaques. r992 demonstrated greatly reduced plaque formation (more than a 10⁵-fold reduction) at 40°C. The shutoff temperature of plaque formation for r1558 was 39°C. These results indicate that each of the three cp45 amino acid substitution mutations individually specifies the ts phenotype, although the mutation yielding the substitution at aa 942 specifies the lowest level of temperature sensitivity. The double-mutant virus r942/1558 had a shutoff temperature of 39°C, while that of r942/992 and cp45 was 38°C. Thus, for these two double mutants, a combination of cp45 amino acid substitutions provides a greater degree of temperature sensitivity than individual amino acid substitutions do. The third double mutant, r992/1558, exhibited only a 250-fold reduction in titer at 40°C and thus is less temperature sensitive than the r992 and r1558 mutants bearing individual mutations. This shows that combining these two particular cp45 mutations reduces rather than increases the level of temperature sensitivity. The finding that the double-mutant virus r942/992 is more temperature sensitive than r942/992/1558 is a second indication that certain combinations of mutations result in a reduced rather than increased level of temperature sensitivity. These observations suggest that the temperature sensitivity specified by the L gene of cp45 is a function of an interaction between the three amino acid substitution mutations rather than simply the sum of their individual effects. Also, the temperature sensitivity of r942/992/1558 is comparable to that of cp45, suggesting that the three mutations at positions 942, 992, and 1558 are major determinants of its temperature sensitivity.

Growth in hamsters. Groups of six hamsters were inoculated intranasally with wt rPIV3 (JS strain), biologically derived cp45, or an rPIV3 containing one or more cp45 L protein amino acid substitutions, and the level of virus replication in the lungs and nasal turbinates was determined 4 days later. The peak virus titer of PIV3- and cp45-infected hamsters was previously demonstrated to occur on day 4 postinfection (6). Each of the rPIV3s bearing a single amino acid substitution was restricted in replication in the upper and lower respiratory tracts (Table 3). Although r942, the least temperature-sensitive virus, was only marginally suppressed in replication, the mutation yielding a substitution at aa 942 clearly contributed to attenuation when present in a double- or triple-mutant recombinant. These data indicate that each of the three amino acid substitutions in the L protein of cp45 contributes to the att phenotype. The triple mutant r942/992/1558 was as restricted in replication in the upper and the lower respiratory tracts as was cp45, indicating that these three L protein mutations combine to specify a level of attenuation similar to that of cp45 and thus are the major contributors to the att phenotype.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Reverse-genetics systems are providing powerful new tools for the characterization of attenuating mutations present in existing vaccines and for the development of new vaccine viruses (27). Recently, such systems have been developed for several members of the virus order Mononegavirales, and this new capability is changing the manner in which new vaccine viruses are being developed (5, 9, 14, 17, 23, 25, 29, 31). An excellent example of this new capability involves the characterization and generation of live attenuated vaccines for RSV (7). One approach involved the generation of a series of live attenuated RSV strains by the conventional techniques of passage in tissue culture or chemical mutagenesis, followed by nucleotide sequence analysis of promising vaccine candidates to identify putative attenuating mutations. The putative attenuating mutations were introduced singly and in combination into full-length antigenomic cDNA encoding the RSV wt parent virus, followed by the characterization of recovered viruses to identify phenotypic changes associated with each mutation alone and in combination (7, 20). Another approach is to create de novo attenuating mutations which had not been found previously in nature. The successful application of this is exemplified by the identification of the attenuating effect of the deletion of the small hydrophobic gene of RSV (3). It is now possible to assemble a menu of attenuating mutations by using these two approaches and to combine the mutations from this menu into a live attenuated recombinant RSV virus via cDNA intermediates. The findings in the present study represent our initial results from applying these principles to the development of cDNA-based vaccines for the human parainfluenza viruses.

The recovery of a recombinant version of the JS strain of PIV3, the parent of cp45, has allowed us to begin to identify the mutations in cp45 that specify the ts and att phenotypes (9). The first step was to completely sequence the wt parent and cp45, the most promising of several cp derivatives (33, 34). Previous studies also suggested that cp45 contains ts and non-ts attenuating mutations and indicated that the att phenotype was stable after replication in vivo (15, 16, 22). The present study demonstrates that each of the three amino acid substitutions in the L protein of cp45 independently specifies both the ts and att phenotypes. However, none specifies the ca phenotype. Most importantly, a recombinant virus possessing all three L protein amino acid substitutions was as attenuated in the upper and lower respiratory tracts of hamsters as cp45 and was almost as temperature sensitive as cp45. These findings therefore identify the three L protein mutations as major contributors to the ts and the att phenotypes of cp45. The finding that three independent mutations in the L protein contribute to the ts and att phenotypes is a partial explanation for the observed stability of the ts and att phenotypes of this virus following replication in vivo (15). This is the first attenuating mutation or set of mutations that has been identified for PIV3 and thus begins our assembly of a menu of attenuating mutations for PIV3. The level of temperature sensitivity exhibited by each rPIV3 was consistent with its level of attenuation in vivo, showing that this in vitro marker is a useful predictor of the level of replication in vivo specified by these polymerase mutations.

It might have been predicted that combining the individual ts mutations would yield increased levels of temperature sensitivity and attenuation, as has been observed for influenza A virus (35). Indeed, this was observed in several instances in the present study. For example, the double mutants r942/992 and r942/1558 and the triple mutant r942/992/1558 were more temperature sensitive and attenuated than viruses bearing the individual mutations. Unexpectedly, the third double mutant, r992/1558, was less temperature sensitive and attenuated than either of the single mutants r992 or r1558. In this instance, the effect of combining the mutations was to reduce, rather than increase, the level of temperature sensitivity of the virus. Similarly, the triple mutant was less temperature sensitive and attenuated than r942/992 and marginally less attenuated than r942/1558. Thus, the level of temperature sensitivity and attenuation achieved by the stepwise combination of mutations was cumulative except in two cases, r992/1558 and r942/992/1558, where the effect of combining mutations was more complex than the effect of the sum of the individual mutations. This is suggestive of an interaction between the mutations.

The mechanisms by which these three mutations contribute to the ts and att phenotypes remain unknown. A ts mutation typically is thought to act through destabilization of protein folding, an effect which is aggravated by increased temperature. The finding that the level of temperature sensitivity was a good predictor of attenuation implies that perturbation of protein folding is involved in both phenotypes. This also provides a reasonable basis for interpreting the interaction between mutations. For example, the finding that each of the mutations at positions 992 and 1558 alone resulted in greater temperature sensitivity and attenuation than they did together could mean that each mutation partially suppresses the destabilizing effect of the other. Interestingly, the tyrosine-to-histidine mutation at position 942, arguably the most conservative substitution of the three mutations, provided the least temperature sensitivity. The L polymerase of PIV3 is a large polypeptide, 2,233 aa in length, and is thought to be a multifunctional protein that consists of multiple domains, including those required for association with the P protein, RNA binding, RNA polyadenylation, RNA transcription, and RNA replication (4). The amino acid substitutions in L at positions 942 and 992 are located near regions that are well conserved among other members of the Paramyxoviridae family (2, 13). The mutation at position 1558 is in a region of the polymerase that appears to have less sequence identity with other L polymerases. Although the mechanism by which the ts phenotype is conferred by the triple amino acid substitution in L is not known, it may be that multiple activities are affected or that the interplay between the various activities may be affected, as has been suggested for a vesicular stomatitis virus L gene mutant (19). In fact, substitution mutations in the Sendai virus L polymerase at aa 1571, a region analogous to that of the cp45 Thr-to-Ile substitution at position 1558, affected multiple polymerase activities (18).

The L gene containing mutations yielding all three attenuating amino acid substitutions can now be used alone or in conjunction with other attenuating mutations as they are identified or created in a cDNA-derived vaccine for PIV3. Furthermore, the attenuating L gene can now be transferred to the recently isolated PIV1-PIV3 chimeric recombinant virus bearing the protective antigens, the HN and fusion glycoproteins, of PIV1 on a background of PIV3 genes (35a). In this way, it should be possible to rapidly generate a live attenuated PIV1 candidate vaccine.