Selection of H5N1 Influenza Virus PB2 during Replication in Humans

To investigate the significance of PB2-627Lys and PB2-701Asn in H5N1 viruses for replication in humans, we extracted viral RNAs from six original specimens (once per specimen) using Isogen (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. RNAs were reverse transcribed with SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA) and an oligonucleotide complementary to the 12-nucleotide sequence at the 3′ end of the viral RNA and amplified by PCR with Pwo DNA polymerase (Roche, Basel, Switzerland) and primers specific for the PB2 segment of the H5N1 influenza virus (sequences available upon request). The PCR products were gel purified with a QIAquick gel extraction kit (Qiagen, Hilden, Germany) and cloned into pCR-Blunt II-TOPO (Invitrogen) by TOPO cloning technology (Invitrogen) and transformed into TOP10 cells (Invitrogen). Escherichia coli clones were picked, and the nucleotide sequences of the viral genes that cover the regions that include PB2-627 and -701 were analyzed. All experiments with infectious H5N1 viruses were performed under biosafety level 3 containment.

We first examined two different specimens obtained from the same patient (HN3040), believed to be directly infected with H5N1 avian viruses. One was a pharyngeal swab collected from the upper respiratory tract; the other was a tracheal aspirate from the lower respiratory tract (Table 1). Maines et al. (12) previously isolated H5N1 viruses with PB2-627Lys (A/Vietnam/1203/04 [H5N1]) from sample HN3040II and PB2-627Glu (A/Vietnam/1204/04 [H5N1]) from HN3040I (both possess PB2-701Asp) (Table 1). A direct analysis of the viral RNAs in these two original specimens revealed avian-type (no PB2-627Lys or PB2-701Asn) and human-type (PB2-627Lys or PB2-701Asn) viruses in each specimen (Fig. 1). Mixed PB2 populations in two other specimens from patient HN3030 were also observed (Fig. 1). These results suggested that although it is possible that the viruses possessing the human-type PB2 genes were already circulating in birds, human-adapted viruses likely emerged during replication in humans.

Most of the virus in HN3040I possessed human-type PB2 (11/12 clones), compared to 7/12 clones in HN3040II (Fig. 1). In the other two specimens collected from the upper respiratory tract, HN3047III and HN3062, we found only human-type PB2 amino acids (Fig. 1). On the other hand, 47% and 20% of the specimens from the lower respiratory tract, HN3030I and HN3030II, respectively, contained avian-type PB2 (Fig. 1). These results demonstrate that the human-type PB2-627Lys and PB2-701Asn are selected during virus replication in the upper respiratory tract of humans.

Generally, MDCK cells are used for virus isolation from mammals (except horses [8]), and embryonated chicken eggs are used for virus isolation from birds. Human viruses with hemagglutinin mutations are selected during propagation in eggs (10, 14). Thus, these procedures may misrepresent virus populations in humans. To test this possibility, we inoculated specimens (Table 1) into MDCK cells and chicken eggs, incubated them at 37°C for 2 to 5 days, and compared the PB2 gene in the propagated viruses with the original specimens (Fig. 1). We found no viruses with avian-type PB2 in any of the specimens, indicating that under these conditions, H5N1 viruses with avian-type PB2 do not propagate in MDCK cells or chicken eggs.

To reveal the factor(s) that caused this discordance between the original specimens and the viral stocks, we inoculated HN3040I and HN3040II into MDCK cells (Fig. 2) and chicken eggs (Fig. 3), incubated them at 33°C and 37°C, harvested the cultured supernatants and allantoic fluids at various time points, and analyzed the PB2 gene in the propagated viruses.

Although most of the viruses grown in MDCK cells possessed human-type PB2, all of the viruses grown in these cells with HN3040II at a 10-fold dilution at 37°C, but not at 33°C, possessed the avian-type PB2 (Fig. 2). These results show that avian-type viruses are selected at higher temperatures in MDCK cells. The inconsistent results with HN3040II-inoculated MDCK cells at 37°C (Fig. 1 and 2B) may reflect low virus titer in the original specimen. A limited number of virions containing only human-type PB2 may have been inoculated into the cells. This also explains the lack of virus with the 100-fold-diluted HN3040II (Fig. 2B, right column).

Viruses grown in eggs were more heterogeneous than those grown in MDCK cells with HN3040 samples (Fig. 2 and 3). Since the viruses with PB2-701Asn were not detected in the original specimen, the detection of such viruses at 33°C but not at 37°C in eggs suggests that PB2-701 may promote their growth under this condition.

Our direct analysis of the viral RNAs from the original patient specimens showed that PB2-627Lys and PB2-701Asn of the H5N1 avian influenza viruses were independently selected during replication in the human respiratory tract. The direct sequencing of the samples identified a mixed population with avian-type and human-type PB2, suggesting that viruses with human-type amino acids arose under selective pressures in humans. We also demonstrated that a PB2 gene population grown in either cell culture or eggs does not always reflect that in patients. A direct analysis of the original specimens is required for accurate information on the viral genome and to monitor for prepandemic viruses.

We thank Susan Watson for editing the manuscript.

This work was supported by ERATO (Japan Science and Technology Agency); by a grant-in-aid for specially promoted research; by a contract research fund for the Program of Founding Research Centers for Emerging and Reemerging Infectious Diseases from the Ministry of Education, Culture, Sports, Science, and Technology; by grants-in-aid from the Ministry of Health, Labor, and Welfare of Japan; and by National Institute of Allergy and Infectious Disease Public Health Service research grants.