A total of 391 influenza A (H1N1) virus isolates obtained from January 2006 to February 2008 were tested in a fluorescence-based NA inhibition assay to determine sensitivity to the NAIs. Of these influenza A (H1N1) viruses, nine strains (2.3%) had increased zanamivir IC50s compared to the mean IC50 of NAI-susceptible influenza A (H1N1) strains collected between 2006 and 2008 (n = 367) (Table ), and two other viruses (0.5%) had significantly increased oseltamivir IC50s (due to a H274Y NA mutation) (data not shown). No influenza A (H3N2) viruses (n = 450) or influenza B viruses (n = 275) had significantly increased oseltamivir or zanamivir IC50s. Sequence analysis of the NA gene from the viruses with reduced zanamivir susceptibility revealed an amino acid mutation of glutamine (Q) to lysine (K) at the highly conserved residue 136 of the NA gene (N2 NA numbering) (Table ) (sequences available in the GenBank database; accession numbers are listed in Table ). The viral isolates A/Brisbane/297/2006, A/Brisbane/308/2007, A/Philippines/604/2006, A/Philippines/1279/2006, and A/Philippines/905/2006, which had zanamivir IC50s between 6.0 and 47.4 nM, demonstrated evidence of a mixed population of viruses with either Q or K at residue 136, while sequence chromatograms for the isolates with higher zanamivir IC50s (ranging from 97.5 to 238.8 nM) revealed a homogeneous population of viruses with the K136 mutation (data not shown). The isolates with reduced zanamivir susceptibility were recovered from individuals who had not been undergoing either zanamivir or oseltamivir treatment, and there was no epidemiological information to suggest any link between the cases.
Characteristics of viruses containing a Q136K mutation
To confirm the effect of the Q136K NA mutation on NAI sensitivity in a homogenous population, the A/Philippines/1279/2006 isolate was plaque purified, and its NA gene was cloned for reverse genetics experiments. The zanamivir IC50s for the plaque-purified and reverse genetics-derived Q136K mutant viruses were 98.1 and 124.2 nM, respectively, equivalent to a 248- and 327-fold reduction in zanamivir susceptibility compared to the wild-type strain (Table ). The plaque-purified and recombinant Q136K mutant viruses also demonstrated 64- and 75-fold reductions in peramivir susceptibility, but they remained susceptible to oseltamivir (Table ), thus confirming the role of the Q136K mutation.
Impact of Q136K NA mutation on the NAI susceptibility of plaque-purified and reverse genetics-derived viruses
As some of the isolates were mixed populations of Q136 wild-type and K136 mutant viruses, it was important to determine the presence of the mutation in the clinical specimens from which the isolates were grown. RT-PCR products for the NA gene were successfully amplified from only three of the primary original specimens that were available for analysis (viruses A/Christchurch/62/2007, A/Macau/229/2008, and A/Brisbane/308/2007), while the other specimens were RT-PCR negative, presumably due to nucleic acid degradation. Conventional sequencing analysis of the viral population in the primary clinical specimens revealed only the presence of the Q136 wild-type genotype. To achieve greater sensitivity in detecting minor viral proportions, the PCR products from these original specimens were cloned, and the colonies were screened for the presence or absence of the Q136K mutation. Surprisingly, even though over 100 colonies were screened for each specimen, only Q136 wild-type NA clones were detected (Table ). This suggested either that the Q136K mutant virus was not present in the original specimen or that it was present at a frequency of less than 1%. In comparison with the cloning results from the original specimens, when the MDCK-grown isolate of A/Christchurch/62/2007 (MDCK passage 4 [MDCK4]) was cloned, 34 out of 43 colonies were found to contain the Q136K mutation (Table ), confirming the ability of these methods to identify mutant clones.
Molecular clonal analysis of viruses to determine the presence or absence of the Q136K NA mutation
To determine how the proportion of the Q136K mutant may have changed upon MDCK culture, the A/Christchurch/62/2007 and A/Macau/229/2008 strains were serially passaged, and the viral populations were analyzed (Fig. ). Based on pyrosequencing quantitation data, the proportion of the Q136K mutant virus in the A/Christchurch/62/2007 viral population increased from undetectable (<1%) in the original specimen and 3.9% in MDCK1 to 51.1% by MDCK2. The proportion of mutant virus continued to increase until passage 7, after which it remained between 80% and 90% until the final passage (Fig. ). A similar trend was observed with the A/Macau/229/2008 virus, where the proportion of mutant started at a very low level at early passage (MDCK2, 0.6%), increased dramatically in the next few passages (MDCK3, 29.8%; MDCK4, 45.3%) and then plateaued between 90% and 95% in the isolates from passage 8 to the passage 11 (Fig. ). In contrast, pyrosequencing analysis of eight NAI-sensitive influenza A (H1N1) isolates that had been passaged five to six times in MDCK cells revealed either undetectable levels of the Q136K mutant (0.0% for six strains) or very low levels (0.2% and 0.5%) that were undetectable in repeat assays.
FIG. 1. Proportion of Q136K mutant in viral isolates following serial passage in MDCK cells. Means and standard deviations of the proportion of Q136K mutant virus in populations were determined following pyrosequencing analysis of amplified PCR product from three (more ...)
The growth rates of the plaque-purified A/Philippines/1279/2006 viruses and the recombinant strains containing the Q136K mutation (Table ) were tested in MDCK cells (Fig. ). The plaque-purified PP-2 (K136 mutant) virus with reduced zanamivir susceptibility grew to a significantly higher titer and at a faster rate in MDCK cells than the PP-1 (Q136 wild-type) zanamivir-susceptible virus (Fig. ). The superior growth characteristics of the mutant strain compared to the wild-type virus in MDCK cells provides support for the serial passaging data of the A/Christchurch/62/2007 and A/Macau/229/2008 viruses, where the mutant virus outgrew the wild-type within the mixed viral population. In contrast, the infectivity and transmissibility of the same two strains in a ferret model were similar. Both strains, even at the lowest inoculum (102 TCID50), infected all inoculated ferrets and was transmitted readily to naïve ferrets (Table ). Titers of the two viruses in nasal washes of ferrets were also not significantly different; the mean titer (± standard deviation) of nasal washes from days 4 and 6 was 2.1 ± 1.0 log10 TCID50/0.1 ml for the PP-1 (Q136 wild-type) zanamivir-susceptible strain compared with 2.5 ± 1.1 log10 TCID50/0.1 ml for the PP-2 (K136 mutant) strain. Throughout the experiment both viruses appeared stable, with no hemagglutinin or NA sequence differences detected between the viruses used to infect the donor ferrets and those collected following nasal washing. Assessments of wellness and measurements of weight loss and temperature changes were also not significantly different between ferrets infected with the zanamivir-susceptible and -resistant strains (data not shown).
FIG. 2. Virus growth curves of plaque-purified viruses and recombinant viruses in MDCK cells. (A) Growth curves of A/Philippines/1279/2006 plaque-purified viruses PP-1 (Q136 wild-type) (•) and PP-2 (K136 mutant) (○) in MDCK cells over 72 h. (B) (more ...)
Ferret infectivity and transmission of plaque-purified and recombinant viruses
A comparison of the viral fitness of the two recombinant viruses RG-1 (Q136 wild-type) and RG-2 (K136 mutant), which contained the NA from either the PP-1 (Q136 wild-type) or the PP-2 (K136 mutant) virus and the remaining seven genes from A/Puerto Rico/8/1934 virus, revealed different results from those obtained with the plaque-purified viruses. The RG-2 recombinant strain with reduced zanamivir susceptibility grew more poorly in MDCK cells (Fig. ) and was less infectious and transmissible in ferrets (Table ) than the zanamivir-susceptible RG-1 virus. This suggested that one or more compensatory amino acid mutations, in addition to the Q136K NA mutation, may be responsible for the plaque-purified mutant's having greater viral fitness than the recombinant virus. Although no NA sequence differences apart from the Q136K were detected between the plaque-purified wild-type viruses and the mutant strains from the A/Philippines/1279/2006, A/Christchurch/62/2007, and A/Macau/229/2008 isolates, sequence analysis of the HA gene from the same strains did identify some amino acid differences (I131T in A/Philippines/1279/2006, H47P and N48H in A/Christchurch/62/2007, and K75E in A/Macau/229/2008). No sequence differences were observed between the matrix genes from the wild-type strains compared with the mutant strains.