Location of antigenic sites recognized by monoclonal antibodies in the influenza A virus nucleoprotein molecule

The influenza A virus genome comprises eight segments of negative-sense viral RNA encoding 11 peptides. RNA genome segments are associated with multiple copies of nucleoprotein (NP), the major internal component of the virion. NP also acts as a multifunctional molecule during the virus reproduction cycle, interacting with several viral and cellular proteins. The functional domains of NP have been mapped in the primary structure of the molecule (Portela & Digard, 2002). NP is a target of cytotoxic T lymphocytes (CTL) and specific antibodies. There is conclusive evidence that the CTL response against NP provides immune protection, and the epitopes recognized by CTL in the NP molecule have been analysed in several studies (Fu et al., 1997; Kreijtz et al., 2008; Voeten et al., 2000). Data on the protective effect of anti-NP antibodies in mice have also been presented (Carragher et al., 2008). However, information about the antigenic sites reacting with antibodies is scarce. The sites have been operationally defined (van Wyke et al., 1980) and shown to exhibit partial overlapping (van Wyke et al., 1981). Several amino acid positions have been proposed to be relevant to the antigenic variation of NP on the basis of variations in the primary structure of NP proteins among influenza A virus strains differing in their reactions with anti-NP monoclonal antibodies (mAbs) (Herlocher et al., 1992; Varich & Kaverin, 2004). However, the location of the antigenic sites has not been conclusively determined. Amino acid position Ala85 was found to be important for the reaction with a mAb, using NP peptides (Yang et al., 2008). Recently, X-ray crystallographic structures of the NP molecule have been presented for human (Ye et al., 2006) and avian (Ng et al., 2008) influenza virus strains, thus enabling the mapping of antigenic sites in the 3D model of NP.

In the present studies, we determined the antigenically relevant amino acid positions of NP by introducing amino acid changes by site-specific mutagenesis in a prokaryotic expression system and subsequently determined the reactivity of the expressed protein with a set of anti-NP mAbs. The sites for mutations were chosen by comparing the reactivity of influenza virus strains with anti-NP mAbs and the strain variation of the NP amino acid sequence.

Influenza viruses A/WSN/33 (H1N1), A/Puerto Rico/8/34 (H1N1) (Mount Sinai), A/Puerto/Rico/8/34 (H1N1) (Cambridge), A/USSR/90/77 (H1N1), A/Brasil/78 (H1N1) and A/Udorn/72 (H3N2) were obtained from the virus collection of the D. I. Ivanovsky Institute of Virology, Moscow, Russia. mAbs 3/1, 5/1, 7/3, 150/4 and 469/6 were produced against A/WSN/33 (H1N1) virus (van Wyke et al., 1980) and have been used in several previous studies (Herlocher et al., 1992; van Wyke et al., 1980, 1981). ELISA was performed as described by Philpott et al. (1989), and the binding percentage was calculated according to the equation: % binding=100×(B_xv/B_pv)/(B_xw/B_pw), where B_xv is the binding of a mAb to the test virus, B_pv is the binding of pooled mAbs to the test virus, B_xw is the binding of a mAb to the wild-type virus, and B_pw is the binding of pooled mAbs to the wild-type virus (Philpott et al., 1989). In experiments with Escherichia coli lysates each lysate was titrated in ELISA against the mixture of mAbs to determine the saturation curve, and the saturating concentration of the antigen was used as a working dose in the reactions with individual mAbs.

The plasmid pET32b (Novagen) was chosen as a vector for cloning and expressing the NP gene. A cDNA copy of the NP gene was transcribed with RT primer Uni from the genomic RNA of A/Puerto Rico/8/34 (H1N1) (Mount Sinai), and then amplified with the cloning primer pair NP(NdeI)F/Np(stKpn)R. The PCR fragment was cloned into pET32b digested with restriction endonucleases NdeI and KpnI. Site-directed mutagenesis of the plasmid pET32b containing the wild-type NP gene was performed with a QuikChange Multi Site-Directed Mutagenesis kit (Stratagene) using specific oligonucleotide primers. Sequences of primers used for reverse transcription, cloning, site-directed mutagenesis and sequencing are shown in Supplementary Table S1, available in JGV Online.

Constructions containing wild-type and mutant NP sequences were expressed overnight in E. coli strain B834 (DE3) co-transformed with pLysS. The T7 promoter was induced at 20 °C with 0.5 mM IPTG when the OD₆₀₀ of the culture reached 0.6. Cells from a 200 ml overnight culture were resuspended in 10 ml PBS and lysed by sonication. The supernatant obtained from centrifuging the cell lysate was used in the ELISA.

In the preliminary stage of the studies, we performed ELISA with five anti-NP mAbs and several human influenza A virus strains. Each mAb was titrated against A/WSN/33 (H1N1) virus and used in a saturating concentration for further determinations. The results (Table 1⇓) confirmed the data reported in earlier studies (Herlocher et al., 1992; van Wyke et al., 1980). Comparative sequence analysis revealed that, among the amino acid positions exposed on the surface of the NP molecule (Ye et al., 2006), three amino acid residues (positions 146, 372 and 455) differed between the viruses recognized and those not recognized by mAb 150/4. Two amino acid residues (98 and 305) differed between the viruses recognized and not recognized by mAb 469/6. One residue (470) differed between the strains that reacted and those that failed to react with mAb 3/1. The strains A/Puerto Rico/8/34 (H1N1) (Mount Sinai) and A/WSN/33 (H1N1) were differentiated by mAb 7/3, which reacted with A/WSN/33 (H1N1) and failed to react with A/Puerto Rico/8/34 (H1N1). The strains differed in four amino acid positions (194, 236, 348 and 353) exposed on the surface of the NP molecule (Ye et al., 2006). Overall, eight amino acid positions on the surface of the NP molecule varied in correlation with the antigenic specificity changes revealed by the mAbs (Table 1⇓).

In our previous comparative studies (Herlocher et al., 1992), the same approach was used, and several amino acid residues differing in the NP of influenza virus strains were identified. However, due to an error in deducing the amino acid positions from the nucleotide sequence, the positions were shifted downstream by 15 amino acids.

Data from the comparative analysis were used to choose the mutations to be introduced into the plasmid expressing the NP protein of A/Puerto Rico/8/34 (H1N1) (Mount Sinai). Individual amino acid changes R98K, A146T, R305K, E372D, D455E and K470R were introduced, and the mutant proteins were expressed and analysed by ELISA. The results (Table 2⇓) revealed that the amino acid substitution E372D abolished the reaction with mAb 150/4, the substitution R305K abolished the reaction with mAb 469/6, and the amino acid change K470R abolished the reaction with mAb 3/1.

Because NP of A/Puerto Rico/8/34 (H1N1) failed to react with mAb 7/3, we attempted to restore the ability of NP to react with this anti-WSN mAb by sequentially introducing amino acid changes at positions at which the NPs of A/WSN/33 (H1N1) and A/Puerto Rico/8/34 (H1N1) differed. We started with the substitution L353V, and then introduced sequentially the substitutions V194I, K236R and K348R, thus producing a series containing single (L353V), double (L353V/V194I), triple (L353V/V194I/K236R), and quadruple (L353V/V194I/K236R/K348R) mutants. Introduction of the L353V mutation failed to restore the reactivity with mAb 7/3. This was expected because the variant A/Puerto Rico/8/34 (H1N1) (Cambridge) has the same residue as A/WSN/33 (H1N1) in this position and did not react with mAb 7/3 (Table 1⇑). Nor did the introduction of two substitutions (L353V/V194I) restore the reaction with mAb 7/3. However, both the triple and the quadruple mutants reacted with mAb 7/3. Since the triple mutant did not contain the substitution K348R, the results indicated that this substitution was not necessary for the restoration of reactivity. The results also indicated that substitution K236R was indispensable for restoration of the reaction with mAb 7/3. To elucidate the role of substitutions L353V and V194I, further analysis was required. They were obviously not sufficient to restore reactivity, because neither the single mutant (L353V) nor the double mutant (L353V/V194I) reacted with mAb 7/3. However, this lack of reaction did not exclude a possible role of these substitutions in the restoration of reactivity. Both substitutions were present in the triple and the quadruple mutant, so they could be necessary (even if not sufficient) for the restoration of reactivity. To elucidate the possible role of the residues in positions 353 and 194, we produced double mutants L353V/K236R and V194I/K236R and single mutant K236R. Both double mutants and the single mutant reacted with mAb 7/3. The data indicated that the substitutions in positions 353 and 194 were not necessary for the binding of mAb 7/3, whereas substitution K236R was both necessary and sufficient for the restoration of the reaction with mAb 7/3.

Interestingly, A/USSR/90/77 and A/Brasil/78 did not react with mAb 7/3 despite having Arg in position 236, as A/WSN/33 does. It is likely that mAb 7/3 recognizes other residues besides that at position 236, some of which are different in A/WSN/33 and the other two strains. However, since the reaction of A/Puerto Rico/8/34 with mAb 7/3 was completely restored by substitution K236R, it is obvious that for this strain the lack of reaction can be adequately explained by the presence of Lys in this position.

The results of site-specific mutagenesis allowed us to identify four amino acid residues recognized by individual anti-NP mAbs. mAbs 3/1, 7/3, 150/4 and 469/6 recognize residues at positions 470, 236, 372 and 305, respectively. In our earlier studies (Varich & Kaverin, 2004), we found by immunoblotting that mAb 3/1 reacts with a linear epitope, whereas mAbs 7/3, 150/4 and 469/6 recognize conformational epitopes. When mapped in the 3D model of NP (Ye et al., 2006), the amino acid residues were found to be located in separate parts of the molecule (Fig. 1⇓). Most likely, they represent the non-overlapping parts of four different antigenic sites, but not the areas where epitopes of mAbs 3/1, 150/4 and 469/6 were shown by competition ELISA to partially overlap (van Wyke et al., 1981). Amino acid positions 305 and 372 are located in the domains of the NP molecule presumed to participate in the binding of PB2 protein, and position 470 is located in the C-terminal acidic part acting as a repressor of PB2 and NP binding (Portela & Digard, 2002). Noteworthy, these amino acid substitutions (R305K, E372D and K470R) are conservative, which may reflect the necessity for preservation of the function of the NP domains.

The amino acid positions suggested as antigenically relevant in our previous studies of closely related influenza virus NPs (Varich & Kaverin, 2004) did not coincide with the positions identified in the present studies. The discrepancies might be due to the difference in the reactions in ELISA and in radioimmunoprecipitation. For example, A/USSR/90/77 (H1N1) virus reacted with mAb 150/4 in radioimmunoprecipitation (Varich & Kaverin, 2004), but not in ELISA (Table 1⇑). Such differences may arise from the incomplete folding of labelled NP in cell lysates. An association of the masking and unmasking of the antigenic epitopes with the intracellular processing of NP has recently been demonstrated (Prokudina et al., 2008). It is possible that the epitopes that are masked in the mature NP in the virions may be accessible for a mAb in the immature NP in cell lysate.

The results of this study provide the first direct data about the positions of several antigenically relevant amino acid residues in the NP of the influenza virus. Our results provide information only about the location of the antigenic sites on the NP molecule and not about the detailed structure of the antigenic epitopes. The data indicate that site-specific mutagenesis is an appropriate tool for analysing the antigenic structure of NP and that it can be used in future studies for a more detailed analysis of NP epitopes.