From immunological and phylogenetic analyses of H3 influenza viruses i
solated from pigs and ducks in the People's Republic of China (China),
Hong Kong, Taiwan and Japan, between 1968 and 1982, we arrived at the
following conclusions. The H3 haemagglutinin and N2 neuraminidase gen
es from swine isolates can be segregated into four mammalian lineages,
including: (i) the earliest human strains; (ii) early swine strains i
ncluding Hong Kong isolates from 1976-1977; (iii) an intermediate stra
in between the early swine and recent human strains; and (iv) recent h
uman strains. In this study we found an unusual swine strain (sw/Hong
Kong/127/82) belonging to the third lineage which behaved like those o
f the early swine-like lineage in the haemagglutination inhibition tes
t; but neuraminidase inhibition profiles with monoclonal antibodies in
dicated that this virus is related to late human strains. On the basis
of pairwise comparisons of complete or partial nucleotide sequences t
he genes encoding the three polymerase proteins (PB2, PB1, PA), the nu
cleoprotein, the membrane protein and possibly the nonstructural prote
ins of sw/Hong Kong/127/82 are of the swine H1N1 lineage, whereas gene
s encoding the two surface glycoproteins belong to the human H3N2 line
age. In contrast, all RNA segments of one swine isolate (sw/Hong Kong/
81/78) are similar to those of recent human H3N2 viruses. This study i
ndicated that frequent interspecies infections between human and swine
hosts appeared to occur during 1976-82. Although the evolutionary rat
es of human (0.0122/site/year), swine (0.0127/site/year) and avian (0.
0193/site/year) virus genes are similar when based upon synonymous sub
stitutions, nonsynonymous substitutions indicated that viral genes der
ived from human and swine viruses evolved about three times faster (0.
0026-00027/site/ year) than those of avian viruses (0.0008/site/year).
Furthermore, the evolutionary mechanism by which human and swine H3 h
aemagglutinin genes evolve at a similar rate, based on nonsynonymous s
ubstitutions, appeared to be quite different from previous evidence wh
ich showed that human H1 haemagglutinin genes evolved three times fast
er than those of swine viruses. However, comparison of the number of n
onsynonymous substitutions in the antigenic sites (A-E) of haemaggluti
nin molecules demonstrated that swine viruses evolve at a rate that is
about one fifth to one tenth that of human viruses, reflecting the co
nservative nature of the antigenic structure in the former.