|Home | About | Journals | Submit | Contact Us | Français|
Reassortment, which is the rearrangement of viral gene segments in a host cell infected with two different viruses, is an important mechanism for the evolution of influenza viruses. Mixed infections with multiple virus types could lead to reassortment. To better understand the occurrence of quasispecies in a single host, we investigated mixed infections in individual isolates of seasonal influenza A viruses using amantadine sensitivity as a marker. We cultured viruses with amantadine and performed sequencing, restriction fragment length polymorphism analysis, cloning, and quantitative PCR to detect mixed populations. Culturing with amantadine showed evidence of a high number of mixed populations, while the other assays could hardly detect mixed populations. The existence of quasispecies in each isolate was common. However, the proportion of these, which can be less than 1%, is too low to be detected by conventional methods. Such mixed populations in which reassortment occurs may have a significant role in the evolution of viruses.
The influenza A virus has negative single-stranded RNA with eight gene segments. The virus causes acute respiratory infection. Currently, two subtypes of influenza A virus, namely, H1N1 and H3N2, are cocirculating in the human population. Influenza A viruses cause epidemics and pandemics by antigenic drift and shift, respectively (24). Antigenic drift is due to the accumulation of point mutations that lead to minor and gradual antigenic changes. Antigenic shift involves major antigenic changes through the introduction of a new hemagglutinin (HA) and/or neuraminidase (NA) subtype into the human population (24).
The transmembrane domain of M2 has ion channel activity, which is involved in the process of uncoating of the virus in the cell (18). Amantadine is an antiviral that inhibits virus replication by blocking the flow of H+ ions from the acidified endosome into the interior of the virion (23). Amino acid substitutions in the transmembrane domain of M2 are responsible for the acquisition of amantadine resistance (23). Amantadine-resistant influenza A viruses are commonly isolated from clinical samples (12, 21), and they can easily be generated in vitro by culturing in the presence of amantadine (2, 10). Resistant viruses can replicate as efficiently as sensitive ones and can also be efficiently transmitted (1, 11). Recently, substantial increases in the incidence of amantadine-resistant influenza A viruses of both the H1N1 and the H3N2 subtypes have been reported, and most of these viruses have an amino acid substitution, serine to asparagine (S31N), at position 31 in the M2 domain (3, 5).
Reassortment, which is the rearrangement of viral gene segments in a host cell infected with two or more different influenza viruses, is an important mechanism for the evolution of influenza viruses (24). An important aspect of reassortment is the generation of novel influenza virus strains that can cause influenza pandemics. Pandemic strains can be generated by reassortment, in which the human virus acquires a novel HA and/or NA subtype from the influenza virus of a nonhuman species (14, 20, 24). Reassortment also facilitates the emergence of more virulent phenotypes by allowing the virus to acquire segments with markers for virulence. Seasonal influenza A viruses, for example, have acquired by reassortment a gene with a drug resistance mutation (8, 22). Nelson et al. also showed the frequent intrasubtype reassortment of influenza viruses (17).
Mixed infections with multiple virus types could lead to reassortment. Ghedin et al. have shown that individuals can harbor influenza viruses that differ in major phenotypic properties, including those that are antigenically distinct and those that differ in their sensitivities to antiviral agents (9). However, it is still not known if such a coinfection with two or more viruses is common. To better understand the occurrence of quasispecies in one host, we investigated mixed infections in individual isolates of seasonal influenza A viruses using amantadine sensitivity as a marker.
Nasopharyngeal swab specimens were collected from patients with influenza-like illnesses who visited pediatric clinics in the city of Sendai, Japan, from 2005 to 2008.
Clinical specimens were inoculated into MDCK cells with 3.5 μl/ml of trypsin. After 2 to 3 days, the supernatant was harvested and stored at −80°C. The original clinical specimens or isolates obtained after the first passage in MDCK cells were used in the following experiments. A total of 159 samples were positive for influenza A virus. Viral cloning (purification) was twice performed by plaque purification. To induce amantadine-resistant amino acid substitutions in sensitive viruses, isolates were passaged in MDCK cells with 8 μg/ml of amantadine up to 10 times.
Viral RNA was extracted by using a PureLink viral RNA/DNA kit (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. The viral RNA was then reverse transcribed to cDNA by using an influenza A virus-specific generic primer, primer Uni12, as reported elsewhere (13). Conventional PCR was performed to amplify each segment. The thermocycling protocol and primer designs for PCR were kindly provided by Reiko Saito of Niigata University. The PCR products were sequenced to analyze the phylogenetics and detect the presence of amino acid substitutions that result in amantadine resistance. The templates were labeled by use of a cycle sequencing reaction with BigDye Terminator (version 1.1; Applied Biosystems, Foster City, CA), and the products were analyzed with an automatic sequencer (model 3130 genetic analyzer; Applied Biosystems), according to the manufacturer's instructions.
We modified a method developed by Saito et al. (19) in order to detect single amino acid changes at position 31 in the M2 gene by PCR-restriction fragment length polymorphism (RFLP) analysis. In that analysis, each aliquot of PCR product was treated with a specific endonuclease. The product amplified by the forward primer (5′-ATGATCTYCTTGAWAATTTRCAG-3′) and the reverse primer (5′-TATCARGTGCAMRATCCCAAYA-3′) was 114 nucleotides. The PCR conditions were as follows: 94°C for 7 min, followed by 35 cycles of 94°C for 30 s, 53°C for 30 s, and 72°C for 30 s; and the final extension was run at 72°C for 7 min. A 5-μl aliquot was treated with 0.6 μl of buffer (NEB4; Biolabs, Beverly, MA) and 0.5 μl of SspI (Biolabs) for 2 h at 37°C. The amplicons were analyzed with a model 2100 Bioanalyzer (Agilent, Santa Clara, CA). PCR primers were designed to amplify the M genes of both H1N1 and H3N2. By RFLP analysis, the sequence coding S31N can be digested into fragments of 90 bp and 24 bp, and the sequence coding S31 cannot be digested (Fig. (Fig.11).
A total of 159 clinical samples, including 86 samples containing H1N1 virus and 73 samples containing H3N2 virus, were tested to validate the assay. Of these, the viruses in 64 samples were amantadine sensitive and the viruses in 95 samples showed resistance through the S31N amino acid substitution, as confirmed by sequencing. Compared to the results obtained by sequencing, the PCR amplicons of all viruses with the S31N amino acid substitution were digested, while those lacking S31N were intact. That is, the sensitivity and the specificity were each 100%.
For the selected samples, the PCR product was cloned into the pGEM T-Easy vector (Promega, Madison, WI). Plural clones were selected and tested by PCR-RFLP analysis.
The quantitative PCRs were carried out with TaqMan Universal PCR master mix reagents (Applied Biosystems), according to the manufacturer's protocol. The following TaqMan Minor Groove Binder probes were used: one probe labeled with 6-carboxyfluorescein (FAM) was designed to detect the sequence coding S31N, while the other probe labeled with VIC was designed to detect the sequence coding S31. The PCRs were cycled 40 times under the following conditions: 95°C for 15 s and 60°C for 1 min. The PCR mixture contained 5 μl cDNA sample, 1 μM forward primer, 1 μM reverse primer, 200 nM VIC probe, 200 nM FAM probe, and 25 μl TaqMan universal PCR master mix reagents in a total volume of 50 μl. The sequences of the primers and probes used are as follows: for H1N1, the forward primer sequence was 5′-CAGGCCTATCARAAACGAATGG-3′, the reverse primer sequence was 5′-CACAATATYARGTGCACAATCCCAAT-3′, the FAM probe sequence was 5′-TTTGCGGCAACAAC-3′, and the VIC probe sequence was 5′-CTTGCGGCAACAAC-3′; and for H3N2, the forward primer sequence was 5′-AACGRATGGGRGTGCAGA-3′, the reverse primer sequence was 5′-AGCATTCTGCTGYTCCTTTCGATA-3′, the FAM probe sequence was 5′-CAAGATCCCAATGATAT-3′, and the VIC probe sequence was 5′-CAAGATCCCAATGATAC-3′.
The M genes of A/Sendai-H/F361/2007, which has S31 in the M2 domain, and A/Sendai-H/F093/2007, which has the S31N amino acid substitution in the M2 domain, were cloned into the pGEM T-Easy vector. These were used as controls to validate the sequencing, PCR-RFLP assay, and quantitative PCR results, as follows: SC, sensitive control containing only A/Sendai-H/F361/2007; RC, resistant control containing only A/Sendai-H/F093/2007; 1:1C, a mixed-population control containing A/Sendai-H/F361/2007 and A/Sendai-H/F093/2007 at a 1:1 ratio; and 10:1C, another mixed-population control containing A/Sendai-H/F361/2007 and A/Sendai-H/F093/2007 at a 10:1 ratio. For the quantitative PCR for H1N1, A/Sendai/TU17/2008 (S31) and A/Sendai/TU13/2008 (S31N) were used as controls in the same way.
Phylogenetic trees were constructed with bootstrap values (from 500 iterations) by the neighbor-joining method with the Mega program (version 4) (15). Some data from online databases were added to the phylogenetic trees.
RNA extraction was performed with MDCK cells from a well that had not been inoculated for use as a negative extraction control in order to detect cross-contamination events that may have occurred during culturing. In addition, no-template controls were included to allow the detection of cross-contamination events by pipetting for the conventional and the quantitative PCRs. None of these negative controls gave positive results, excluding the possibility of cross-contamination.
The accession numbers of all sequences used in the study, including our newly determined gene sequences, are given in list SA1 in the supplemental material.
We analyzed 11 randomly selected amantadine-sensitive isolates. The M genes of these isolates did not have the S31N amino acid substitution (i.e., they had S31), as confirmed by Sanger sequencing. The profiles of the isolates are listed in Table Table11.
We passaged these 11 viruses in MDCK cells with amantadine in the culture medium in order to induce amino acid substitutions for amantadine resistance in M2. After multiple passages, we extracted the viral RNA and analyzed the sequences from the isolates. The amino acid substitutions that induced amantadine resistance were V27A, A30V, S31N, and G34E (Table (Table2),2), which, according to earlier data, are mutations known to induce resistance to amantadine in influenza viruses (1, 10).
We then constructed phylogenetic trees with the sequences of our original virus isolates, passaged virus isolates, and H3N2 and H1N1 viruses reported by others (Fig. (Fig.2).2). Amantadine-sensitive and -resistant viruses reported by others formed distinct lineages in each tree (we defined the lineages as the S lineage and the R lineage, respectively). All of our original amantadine-sensitive isolates (green in Fig. Fig.2)2) were grouped into the S lineage.
Six of 11 isolates that acquired amino acid substitutions for amantadine resistance after passage with amantadine were grouped into the S lineage (red in Fig. Fig.3;3; Table Table1),1), indicating that these viruses are close to the original viruses that did not have an amino acid substitution for amantadine resistance. The other five virus isolates, which had amino acid substitutions for amantadine resistance (S31N) after passage with amantadine, were grouped into the R lineage (red in Fig. Fig.3;3; Table Table2),2), indicating that these are similar to viruses with the amino acid substitution for amantadine resistance, which were circulating at the same time.
We then performed plaque purification with the initial isolates with a course of only one passage in MDCK cells. After purification, the isolates were also passaged in medium with amantadine, and the sequences of their M genes were analyzed. Sequencing showed that plaque-purified viruses acquired the V27A, A30V, A30T, S31N, or G34E amino acid substitution in M2 (Table (Table2).2). The phylogenetic trees showed that all viruses of both the H3N2 and the H1N1 subtypes passaged in amantadine after plaque purification were grouped in the S lineage (blue in Fig. Fig.3;3; Table Table22).
When viruses without S31N were mixed with viruses with S31N in equal proportions (1:1C), sequencing (Sanger) and PCR-RFLP analysis could detect the mixed population (Fig. (Fig.11 and and4).4). The assay could not identify any S31N amino acid substitution in 11 samples of the mixed populations tested. However, the sensitivity of the assay is not high. When sensitive and resistant viruses were mixed at a ratio of 10:1 (10:1C), it was not possible to detect the sequences of the minor population by either sequencing or the PCR-RFLP assay (Fig. (Fig.11 and and44).
We then analyzed the amplification products by TA cloning and PCR-RFLP analysis. The assay could identify 1:1C and 10:1C as mixed populations (Table (Table2).2). For the clinical samples, the S31N amino acid substitution was found in 1 of 70 clones in one sample (Table (Table2,2, sample 7). For the other 10 samples, none of the clones had the S31N amino acid substitution.
We also performed quantitative (real-time) PCR analysis (Fig. (Fig.5).5). We used TaqMan MGB probes: one probe was labeled with FAM and was designed to detect the sequence encoding the S31N amino acid substitution, while the other probe was labeled with VIC and was designed to detect the sequence coding S31. In theory, the presence of two probe pairs in each reaction mixture allows genotyping of the two possible variants at a single nucleotide polymorphism site. However, the assay could not distinguish the two sequences encoding S31 and S31N. The reason for this is not known. The probe for the sequence encoding S31N annealed and reacted with the sequence encoding S31 (SC), but with an efficiency lower than that of the reaction with the sequence encoding S31N (RC) (Fig. (Fig.5).5). Therefore, the assay could not be used to detect or determine the proportions in a mixed population.
In the present study, we found a mixed population within an influenza virus isolate using amantadine sensitivity as a marker. When sensitive isolates whose M genes were in the S lineage were cultured in medium with amantadine, the viruses with M genes from a different lineage (R lineage) appeared after several passages. The results suggest that mixed populations of S-lineage and R-lineage viruses (of the same subtype) existed in some individual samples. Therefore, the M gene in the R lineage with the S31N amino acid substitution was selected through passages with amantadine. However, Sanger sequencing or PCR-RFLP analysis could not identify mixed populations in the samples. Even the cloning of 36 to 94 clones per sample identified the S31N amino acid substitution in only 1 clone. The proportion of minor viruses in the mixed population may be too small to be identified by these assays. The results of cloning assays suggest that the proportion of resistant viruses in a mixed population dominated by sensitive viruses should be less than 1%.
Lackenby et al. suggested the existence of quasispecies in one host (16). They indicated that the influenza virus could mutate rapidly in the infected host and that new variants (i.e., quasispecies) continuously arise from the predominantly infecting virus during viral replication, resulting in the existence of heterologous viruses in one host. It is also possible that the M gene of the R lineage was generated by the accumulation of mutations in the M gene of the S lineage during passages in cell lines. Mutations can occur during isolation or passage in cell lines (7, 25); however, when we passaged isolates in amantadine after plaque purification, the M gene in the S lineage acquired a point mutation that led to amantadine resistance. The results showed that the de novo accumulation of mutations in the S lineage did not generate the M gene in the R lineage. Therefore, the mixed populations in the present study were not caused by the de novo accumulation of mutations in each host or passages in cell lines. The mixed population was probably constructed by the coinfection of two (or more) genetically different viruses. The appearance of the M gene in the R lineage in the assay is probably the consequence of selection.
The results also reject the possibility of cross-contamination. If cross-contamination had occurred during the assay, viruses in the R lineage with the M gene would have appeared in plaque-purified viruses; however, that did not occur. Besides, some samples should contain a mixed population, but the M gene in the R lineage did not appear after passages with amantadine. Viruses passaged with amantadine and those passaged after plaque purification were clearly distinct (e.g., sample 1; see the arrows in Fig. Fig.22 and and3).3). These findings suggest that mixed populations of diverse viruses existed even in the same lineage.
A mixed population in an individual host could be caused by simultaneous infections with two genetically different viruses. Ghedin et al. showed that individuals can harbor influenza viruses that differ in their major phenotypic properties and confirmed the presence of a mixed population by cloning (9). They showed an intrasubtype mixed population in only two samples, with approximately 3% of the samples having some evidence of mixed infection. However, this might be an underestimation of the rate, since the sensitivity of the assay that they used might be low. In the present study, we showed a considerably higher proportion (at least 6/11) of mixed populations in the samples analyzed by culturing viruses with amantadine or cloning. Such a high proportion of mixed populations suggests that populations of distinct viruses are transmitted from human to human as a mixture. It was beyond the scope of the present study to investigate other segments, because we could not determine whether other segments that appeared after passage were originally possessed by the minor or the major viruses, as reassortment might occur.
Different lineages of the same subtype of influenza viruses cocirculate during a single season in the same location (6, 8, 17), and reassortment between viruses in distinct lineages generates new variants (8, 17, 22). Reassortment requires coinfection with two viruses. Here, we propose that intrasubtype mixed infections with distinct influenza viruses can commonly occur with seasonal influenza viruses. Mixed populations of diverse viruses may be transmitted from human to human while maintaining their diversity. This mechanism has evolutionary importance. At present (2009), pandemic and seasonal H1N1 influenza viruses are cocirculating in human populations (4). Mixed infections with the pandemic and seasonal viruses can occur, and the mixed population would be maintained through transmission. It is possible that new variants will be generated by reassortment. It is important to consider mixed infections, as well as the emergence of new variants, to better understand the evolution of influenza viruses.
In conclusion, our analysis, in which we used amantadine sensitivity as a marker, indicated that quasispecies in single isolates are common. However, the proportion of the minor population appears to be too low to be detected by conventional methods (e.g., Sanger sequencing). Such minor populations may be significant if reassortments between strains in the major and the minor populations create viruses with significant phenotypic changes, such as antiviral resistance.
We thank Hidekazu Nishimura and all other staff members of the Virus Research Center of the Sendai Medical Center. We are indebted to the medical practitioners in Sendai, Japan (Makoto Shoji, Kazuhisa Kawamura, Jun Kayaba, Syunzo Hayamizu, Masataka Itano, and Kikuya Metoki), who collected clinical specimens. We also thank the Biomedical Research Core of Tohoku University Graduate School of Medicine for technical support.
This work was supported by JSPS KAKENHI (grant 21406014).
Published ahead of print on 25 November 2009.
†Supplemental material for this article may be found at http://jcm.asm.org/.