Antigenic drift in the influenza A virus (H3N2) nucleoprotein and escape from recognition by cytotoxic T lymphocytes

Abstract

Viruses exploit different strategies to escape immune surveillance, including the introduction of mutations in cytotoxic T-lymphocyte (CTL) epitopes. The sequence of these epitopes is critical for their binding to major histocompatibility complex (MHC) class I molecules and recognition by specific CTLs, both of which interactions may be lost by mutation. Sequence analysis of the nucleoprotein gene of influenza A viruses (H3N2) isolated in The Netherlands from 1989 to 1999 revealed two independent amino acid mutations at the anchor residue of the HLA-B27-specific CTL epitope SRYWAIRTR (383 to 391). A R384K mutation was found in influenza A viruses isolated during the influenza season 1989-1990 but not in subsequent seasons. In the influenza season 1993-1994, a novel mutation in the same CTL epitope at the same position was introduced. This R384G mutation proved to be conserved in all influenza A viruses isolated from 1993 onwards. Both mutations R384K and R384G abrogated MHC class I presentation and allowed escape from recognition by specific CTLs.

FIG. 1

Maximum likelihood tree based on nucleotide sequences of the NP gene. Part of the NP genes (nt 720 to 1175) of 162 influenza A (H3N2) viruses isolated from 1989 to 1999 was sequenced and subjected to phylogenetic analysis. Included in the figure are the influenza viruses A/Texas/1/77, A/Memphis/5/80, and A/Beijing/353/89. The numbers shown in the figure correspond to the numbers shown in Table 1. The insert represents a protein distance tree (Protdist, Fitch) based on the NP sequence of the representative influenza virus strains. The letter code used for the respective amino acid sequences corresponds to that of Fig. 2.

FIG. 2

Amino acid sequences of the NP (aa 240 to 391) of influenza A (H3N2) viruses isolated from 1989 to 1999. The consensus sequence of each season is shown in the upper rows, whereas variant sequences are shown in the lower rows. CTL epitopes are underlined and shown in boldface. Amino acid differences between seasons are shown in boldface, while differences within a season are shown in boldface italic. All mutations are marked with an arrow. The number in parentheses refers to the number of isolates showing that sequence.

FIG. 3

Confirmation of specificity and HLA restriction of CTL clones. HLA-A3- and -B27-positive (A and B) and HLA-A3- and -B27-negative (C and D) B-LCL cells were incubated with the HLA-A3-specific peptide ILRGSVAHK (solid circles) or the HLA-B27-specific peptide SRYWAIRTR (solid triangles) or left untreated (open circles and open triangles, respectively) followed by incubation with the NP/A3 (A and C) or NP/B27 (B and D) CTL clone. CTL assays were performed in triplicate at three E:T ratios. Mean percentages of specific lysis are shown.

FIG. 4

Effect of mutations in the HLA-B27 epitope on CTL-mediated lysis of target cells. (A and B) HLA-A3- and -B27-positive B-LCL cells of one donor were incubated with the HLA-A3-specific peptide ILRGSVAHK (solid circles), the HLA-B27-specific peptide SRYWAIRTR (solid squares), or the HLA-B27 mutant peptide SGYWAIRTR (open squares) or SKYWAIRTR (open hexagons) or left untreated (open circles) and used as targets in CTL assays with the NP/A3 (A) or NP/B27 (B) CTL clone as effector. (C and D) The same B-LCL cells were infected with a control vaccinia virus (open inverted triangles), a vaccinia virus expressing the NP of A/PR/8/34 (solid triangles), or a vaccinia virus expressing NP of A/Neth/18/94 (open triangles) followed by incubation with the NP/A3 (C) or NP/B27 (D) CTL clone. (E and F) Also, B-LCL cells were infected with influenza virus A/PR/8/34 (solid diamonds), A/Neth/18/94 (open diamonds), or A/Neth/651/89 (open hexagons) or left untreated (open circles) followed by incubation with the NP/A3 (E) or NP/B27 (F) CTL clone. CTL assays were performed in triplicate at three E:T ratios. Mean percentages of specific lysis are shown.