Learn more: PMC Disclaimer | PMC Copyright Notice
Influenza A Virus M2 Ion Channel Activity Is Essential for Efficient Replication in Tissue Culture
Abstract
The amantadine-sensitive ion channel activity of influenza A virus M2 protein was discovered through understanding the two steps in the virus life cycle that are inhibited by the antiviral drug amantadine: virus uncoating in endosomes and M2 protein-mediated equilibration of the intralumenal pH of the trans Golgi network. Recently it was reported that influenza virus can undergo multiple cycles of replication without M2 ion channel activity (T. Watanabe, S. Watanabe, H. Ito, H. Kida, and Y. Kawaoka, J. Virol. 75:5656–5662, 2001). An M2 protein containing a deletion in the transmembrane (TM) domain (M2-del29–31) has no detectable ion channel activity, yet a mutant virus was obtained containing this deletion. Watanabe and colleagues reported that the M2-del29–31 virus replicated as efficiently as wild-type (wt) virus. We have investigated the effect of amantadine on the growth of four influenza viruses: A/WSN/33; N31S-M2WSN, a mutant in which an asparagine residue at position 31 in the M2 TM domain was replaced with a serine residue; MUd/WSN, which possesses seven RNA segments from WSN plus the RNA segment 7 derived from A/Udorn/72; and A/Udorn/72. N31S-M2WSN was amantadine sensitive, whereas A/WSN/33 was amantadine resistant, indicating that the M2 residue N31 is the sole determinant of resistance of A/WSN/33 to amantadine. The growth of influenza viruses inhibited by amantadine was compared to the growth of an M2-del29–31 virus. We found that the M2-del29–31 virus was debilitated in growth to an extent similar to that of influenza virus grown in the presence of amantadine. Furthermore, in a test of biological fitness, it was found that wt virus almost completely outgrew M2-del29–31 virus in 4 days after cocultivation of a 100:1 ratio of M2-del29–31 virus to wt virus, respectively. We conclude that the M2 ion channel protein, which is conserved in all known strains of influenza virus, evolved its function because it contributes to the efficient replication of the virus in a single cycle.
The discovery that the influenza A virus M2 protein has proton-selective ion channel activity stemmed from an understanding of the life cycle of influenza virus and the two steps in the life cycle that are inhibited by the antiviral drug amantadine (reviewed in reference 14). Direct evidence that the M2 protein has a low-pH-activated, proton-selective conductance was obtained by expressing the M2 protein in oocytes of Xenopus laevis (20, 21, 47, 53, 64, 65, 67) or mammalian cells (6, 40, 41, 66). M2-specific cell surface currents were measured, and they were found to be specifically blocked by amantadine. Furthermore, when either peptides corresponding to the M2 transmembrane (TM) domain or purified M2 protein was incorporated into planar bilayers, an amantadine-sensitive current was measured (11, 36, 52, 63).
Amantadine inhibits the early step of uncoating of influenza virus in endosomes (reviewed in references 14, 31, 32, and 34). When a virion has entered the cell by receptor-mediated endocytosis and the virus particle is in the acidic environment of the endosomal lumen, the M2 ion channel is activated and conducts protons across the viral membrane. The lowered internal virion pH is thought to weaken protein-protein interactions between the viral matrix protein (M1) and the ribonucleoprotein (RNP) core (4, 5, 38, 72, 73; reviewed in reference 18). In the presence of amantadine, influenza virus uncoating is incomplete because the M1 protein is not released from the RNPs and the RNPs fail to enter the nucleus: normally, influenza virus RNPs are transcribed and replicated in the nucleus (reviewed in reference 32). For some influenza virus subtypes, amantadine inhibits a late step in virus replication. The M2 ion channel activity is activated during transport of the M2 protein through the exocytic pathway, and this ion channel activity causes the equilibration of the acidic pH of the lumen of the trans Golgi network with the cytoplasm (7, 8, 12, 13, 44, 50, 58, 60). Thus, the intralumenal pH of the trans Golgi network is kept above the threshold at which the hemagglutinin (HA) conformational change to the low-pH-induced form occurs.
The use of amantadine to inhibit the growth of influenza virus is not restricted to tissue culture cells in the laboratory. Amantadine and its derivative rimantadine have proven to be effective drugs in humans (and mice) for the prevention of influenza, and the drugs are therapeutic for a more rapid resolution of clinical signs and symptoms (3, 10, 17, 39, 49, 62, 69, 70). Rimantadine (Flumadine; Forest Pharmaceuticals, Inc.) is licensed by the U.S. Food and Drug Administration for prophylaxis in adults and children and for treatment in adults.
Recently it was reported that influenza A virus can undergo multiple cycles of replication without M2 ion channel activity (68). Mutant viruses were constructed and recovered from cloned DNA, with all the influenza virus RNA segments being derived from strain A/WSN/33 except for RNA segment 7 (encoding the M1 and M2 proteins), which was derived from strain A/Udorn/72. The recovered viruses contained mutations in the M2 protein TM domain (M2-A30P and M2-del29–31). It had been observed previously that M2 proteins containing these mutations, when expressed in oocytes of X. laevis, did not exhibit a detectable ion channel activity. However, M2-A30P and M2-del29–31 were found to replicate at least as efficiently as the wild-type (wt) virus in cell culture (68). Furthermore, viruses containing M2 proteins in which the M2 TM domain was supplanted with that from HA (M2HATM) or neuraminidase also replicated efficiently in tissue culture. When the replication of the M2 proteins lacking detectable ion channel activity (M2-A30P, M2-del29–31, and M2HATM) was tested with mice, none of these viruses were found to be recovered in nasal turbinates; however, M2-A30P and M2-del29–31 but not M2HATM could be recovered from lungs (68).
Analyses of the effect of amantadine on influenza A virus replication in tissue culture indicated that influenza virus is sensitive to the drug but that the sensitivity of virus growth depends somewhat on the strain of influenza A virus (50% inhibitory concentration, ∼1 to 10 μM amantadine) (1, 16, 57). At optimal drug concentrations, virus yield was found to be inhibited 90 to 99% depending on the subtype tested. However, it is quite clear from the early studies, provided amantadine was added to cells either before infection or at the time of infection, that although virus binding and virus penetration occurred, there was no detectable primary transcription of the input genome by the virion-associated RNA transcriptase (16, 25, 57). Thus, the finding that influenza A viruses that lack an M2 ion channel activity grow normally in tissue culture (68) is seemingly inconsistent with the effect of amantadine on virus replication in tissue culture. However, the caveat has to be added that the studies examining amantadine inhibition of primary transcription (but not those studies examining virus yield) were done with high drug concentrations, ∼1 mM. At this concentration, amantadine, in addition to M2 ion channel inhibition, also causes a lysosomaltropic effect, raising intralumenal pH (43, 48). This causes inhibition of uncoating for several viruses that utilize low-pH steps (9; reviewed in reference 37). Therefore, given the experiments described in the literature, there is a compelling need to reexamine the effect of amantadine at low concentrations on the influenza virus life cycle.
The M2 ion channel protein is a homotetrameric integral membrane protein, with each chain of the mature protein containing 96 amino acid residues (19, 28, 33, 35, 45, 51, 59, 61, 71). The TM domain consists of 19 residues, and a considerable body of experimental evidence indicates that the M2 protein TM domain constitutes the proteinaceous core (the channel pore) that allows a flux of protons across the membrane: M2 TM domain histidine residue 37 is associated with proton selectivity and conduction (6, 11, 36, 40, 41, 47, 55, 65). Influenza viruses that are resistant to amantadine contain mutations in the M2 protein TM domain (15, reviewed in reference 14), and when these M2 mutants were expressed in oocytes the ion channel activity was insensitive to amantadine (47). The coding regions for the M2 protein have been conserved in all known strains of avian, swine, equine, and human influenza A viruses, and the amino acid sequence of the M2 protein TM domain has been conserved to a greater extent than the remainder of the protein (24). Thus, we have reexamined the extent of amantadine inhibition of influenza virus growth and characterized the biology of a recovered influenza virus containing a deletion of residues 29 to 31 in the M2 protein TM domain.
MATERIALS AND METHODS
Cells and viruses.
293T cells (a human embryonic fibroblast cell line transfected with the E1 region of adenovirus and the simian virus 40 T antigen) and Madin-Darby canine kidney (MDCK) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco-BRL, Gaithersburg, Md.) supplemented with 10% fetal bovine serum. Influenza viruses were plaque purified and propagated in MDCK cells in DMEM supplemented with 1.0 μg of N-acetyl-trypsin (NAT) (Sigma-Aldrich, St. Louis, Mo.)/ml.
Construction of plasmids.
Influenza A/Udorn/72 virus was grown in embryonated chicken eggs, and virions were purified by sucrose gradient centrifugation (23). Viral RNA was isolated as described previously (29). First-strand cDNA to viral RNAs (vRNAs) was made using avian myeloblastosis virus Super Reverse Transcriptase (Molecular Genetic Resources, Tampa, Fla.) and a vRNA 3′-end-specific primer, 5′-AGC AAA AGC AGG-3′, and second-strand synthesis was done using a Klenow fragment of DNA polymerase and a specific vRNA 5′-end-specific primer, 5′-AGT AGA AAC AAG G-3′, using the previously described protocol (27). cDNAs were cloned into the vector pGEM3 (Promega, Madison, Wis.), and full-length cDNAs to each of the influenza virus RNA segments were identified and their complete nucleotide sequences were determined. To transfer the cDNAs to pHH21 such that the cDNA inserts were flanked by the human RNA polymerase I promoter and the mouse RNA polymerase I terminator as reported previously (42), the cDNAs in pGEM3 were amplified by a PCR using Vent DNA polymerase (New England Biolabs, Beverly, Mass.) and eight pairs of gene-specific primers flanked by BsmBI restriction enzyme sites. The PCR DNA product was digested with BsmBI and cloned into the BsmBI sites of pHH21. The cDNAs encoding PB2, PB1, PA, and NP of the A/Udorn/72 strain were also cloned into the expression vector pcDNA (Invitrogen, Carlsbad, Calif.) to generate four support plasmids (42). Plasmids containing cDNAs specific for A/WSN/33 (pHH21, pCAGGS, and pcDNA), required for the generation of WSN virus from cloned DNA (42), were kindly provided by Yoshihiro Kawaoka. Conversion of WSN M2 asparagine residue 31 to serine (N31S-M2WSN) and the deletion of nucleotides encoding amino acid residues 29 to 31 of the Udorn M2 protein (M2-del29–31Udorn) were performed by using two rounds of PCR and specific oligonucleotide primers (2), and the DNA was recloned into pHH21. The complete nucleotide sequence of the cDNAs containing N31S-M2WSN and M2-del29–31Udorn was determined using an ABI Prime 310 genetic analyzer and EditView software (PE Biosystems Inc., Foster City, Calif.).
Recovery of infectious virus from cloned DNA.
Viruses containing the WSN genetic background, i.e., wt WSN, N31S-M2WSN, and MUd/WSN, were generated as reported previously (42). To generate viruses containing the Udorn genetic background, i.e., wt Udorn and M2-del29–31Udorn, the procedure of Neumann and colleagues (42) was modified slightly: at 9 h posttransfection (p.t.), Opti-MEM culture media (Gibco-BRL) were replaced with Opti-MEM containing 3.0 μg of NAT/ml, and at 19 h p.t., the transfected 293T cells were scraped into the culture media and overlaid onto subconfluent monolayers of MDCK cells. At various times p.t., culture supernatants were collected.
Plaque assays and immunostaining.
Confluent monolayers of MDCK cells were incubated with 10-fold-serially-diluted virus samples in DMEM-1% bovine serum albumin for 1 h. The inoculum was removed, and the cells were washed with phosphate-buffered saline (PBS). The cells were then overlaid with DMEM containing 1% agarose and NAT (1.0 μg/ml). To examine the effect of amantadine on plaque formation, monolayers were preincubated with DMEM supplemented with amantadine (5 μM) at 37°C, and virus samples were preincubated with DMEM-1% bovine serum albumin with amantadine (5 μM) at 4°C for 30 min before infection. At 2 to 3 days after infection, the monolayers were fixed and stained with naphthalene black dye solution (0.1% naphthalene black, 6% glacial acetic acid, 1.36% anhydrous sodium acetate). All the plaques detected were counted. For plaques formed in the presence of amantadine, this total includes larger plaques of presumptive amantadine-resistant mutants that arise at a frequency of approximately 1 in 104 of the pinpoint plaques. To stain the plaques for immunological specificity, the monolayers were fixed with 1% glutaraldehyde and incubated for 1 h with a goat anti-A/Udorn/72 virus serum in a blocking solution consisting of PBS with 3% egg albumin (Sigma Aldrich, Rockville, Md.). The monolayers were then washed with PBS and incubated for 1 h with peroxidase-conjugated donkey anti-goat immunoglobulin G (Jackson Immuno Research Laboratories Inc., West Grove, Pa.) in the blocking solution with 3% egg albumin. After the monolayers were washed with PBS, the peroxidase-conjugated antibody was reacted with a DAB with metal solution (ImmunoPure Metal Enhanced DAB Substrate Kit) (Pierce Chemical Co., Rockford, Ill.).
Multiple-cycle replication kinetics.
Confluent monolayers of MDCK cells (6.0-cm-diameter dishes) were incubated with wt Udorn or the M2-del29–31Udorn virus at a multiplicity of infection (MOI) of 0.001 PFU per cell for 1 h at 37°C, washed with PBS, and then incubated with DMEM supplemented with NAT (1 μg/ml) at 37°C. At various times postinfection (p.i.), infectivity (PFU) of the culture supernatants was determined. Experiments were performed in triplicate.
Fitness assay.
To construct a nucleotide sequence electropherogram standard for a mixtures of two plasmids, aliquots of pHH21-MUd and pHH21-M2-del29–31 were mixed in various ratios and subjected to direct sequencing using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (PE Biosystems) and a gene-specific primer, Mf2 (5′-GAG GCC ATG GAG GTT GCT AGT CAG G-3′). Nucleotide sequencing and analysis were performed as described above. To examine the relative biological fitnesses of wt Udorn and M2-del29–31Udorn viruses, 70 PFU (MOI per cell, 0.0001) of wt Udorn were cocultured with 7,000 PFU (MOI, 0.01) of M2-del29–31Udorn virus in MDCK cells. At various times p.i., vRNAs of the viruses in the culture supernatant were purified using the QIAamp Viral RNA Mini Kit (QIAGEN, Valencia, Calif.) and reverse transcribed into cDNAs with primer Mf1 (5′-TAA CAT GGA CAG AGC AGT TAA ACT G-3′). The cDNAs were amplified by PCR using two primers. The PCR products were subjected to direct DNA sequencing, and electropherograms were compared to the standards.
Protein labeling and quantification.
Monolayers of MDCK cells in 24-well plates were infected with various viruses at an MOI of 3 to 5 PFU/cell. For comparison of the growth of wt Udorn and M2-del29–31Udorn virus, equivalent numbers of virus particles were used to infect cells as determined by the HA titer. At various times after infection, the infected cells were metabolically labeled for 30 min with 35S-Promix (100 μCi/ml) (Amersham Pharmacia Biotech, Piscataway, N.J.) in DMEM deficient in methionine and cysteine (DMEM met-, cys-) and then lysed and boiled in protein lysis buffer (1% sodium dodecyl sulfate [SDS], 31 mM Tris-HCl [pH 6.8], 2.5% dithiothreitol). To examine the effect of amantadine, cells and viruses were preincubated with amantadine (5 μM) as described above, followed by infection and incubation in the presence of amantadine (5 μM). Polypeptides were separated by SDS-polyacrylamide gel electrophoresis (PAGE). The gels were fixed and dried, and radioactivity was analyzed and quantified using a Fuji BioImager 1000 and Mac Bas software (Fuji Medical Systems, Stamford, Conn.).
RESULTS
Sensitivity of influenza viruses containing the A/WSN/33 or A/Udorn/72 M gene to amantadine.
The pore of the influenza virus M2 ion channel protein is its 19-residue TM domain, and its amino acid sequence confers sensitivity to the antiviral drug amantadine: influenza viruses that are resistant to amantadine contain mutations in the M2 protein TM domain (15; reviewed in reference 14). Furthermore, when these M2 mutants were expressed in oocytes, or in mammalian cells, the ion channel activity was found to be insensitive to amantadine, in contrast to the amantadine-sensitive wt M2 ion channel conductance (6, 47). Thus, the observation that influenza virus can undergo efficiently multiple cycles of replication in tissue culture without M2 ion channel activity (68) was unexpected and considered worthy of further investigation.
Watanabe and colleagues (68) used a virus and derivative mutants that had a genetic background derived from A/WSN/33 except for RNA segment 7, encoding the M1 and M2 proteins, which was derived from A/Udorn/72. Thus, it seemed possible that the gene constellation of this virus had unexpected properties. We tested four related viruses for their sensitivity to amantadine (5 μM) in a plaque assay (Fig. (Fig.1).1). To make the viruses comparable, A/WSN/33 and A/Udorn/72 were recovered from cloned DNAs and considered wild types. A/WSN/33 was found to be resistant to amantadine, as found previously (16). To investigate whether the presence of an asparagine at M2 TM domain residue 31 (a mutation known to confer amantadine resistance [15]) was the sole determinant of the resistance of A/WSN/33 to amantadine, a mutant virus (N31S-M2WSN) was recovered from cloned DNA such that the asparagine residue at position 31 in the WSN M2 protein TM domain was replaced with a serine residue. As shown in Fig. Fig.1,1, N31S-M2WSN virus was amantadine sensitive. The MUd/WSN virus, which possesses seven RNA segments from WSN plus the M segment (RNA segment 7) derived from the A/Udorn/72 strain, was found to be sensitive to amantadine as expected, and A/Udorn/72 virus recovered from cloned Udorn-specific DNAs was amantadine sensitive as expected (Fig. (Fig.1).1). For all amantadine-sensitive viruses, pinpoint plaques could be detected after three days p.i. Thus, genetic background differences, except for the nature of the M2 protein residue 31 between WSN and Udorn strains, did not affect sensitivity to 5 μM amantadine in a plaque assay. Nonetheless, it is emphasized that amantadine is not a particularly potent inhibitor of plaque titer, since for all three amantadine-sensitive viruses, the plaque titer was reduced only 46- to 50-fold after 2 days’ incubation with 5 μM amantadine. Similar conclusions were reached previously when virus yield in milligrams of viral protein was measured (16).
Effect of amantadine (aman) on influenza A virus plaque formation. Influenza viruses were generated from cloned DNAs to the eight RNA segments using a reverse-genetics procedure (42). WSN virus was recovered from cDNAs as previously reported (42). N31S-M2WSN virus differs from amantadine-resistant WSN virus in that an asparagine residue at position 31 in the M2 protein TM domain was replaced with a serine residue to convert the virus to one that was amantadine sensitive. MUd/WSN virus possesses seven RNA segments from WSN plus the M segment (RNA segment 7) derived from the A/Udorn/72 strain. Udorn virus was recovered from cloned DNAs corresponding to the eight Udorn strain RNA segments. Monolayers of MDCK cells were preincubated with or without amantadine (5 μM) for 30 min and infected with WSN, N31S-M2WSN, MUd/WSN, and Udorn viruses. After 1 h of incubation, monolayers were washed with PBS (with or without amantadine) and overlaid with 1% agarose-DMEM supplemented with 1.0 μg of trypsin/ml, with or without amantadine (5 μM). After 2 to 3 days of incubation, monolayers were fixed and stained with naphthalene black dye solution. The equivalent dilution is shown for each matched pair of monolayers with and without amantadine (+ and − aman). All the plaques detected at 10-fold plaque dilutions were counted where possible. For plaques formed in the presence of amantadine, the total number includes larger plaques that are presumptive amantadine-resistant mutants that arose spontaneously at a frequency of approximately 1 in 104 of the pinpoint plaques formed.
Inhibition of protein synthesis by amantadine.
Considerable experimental evidence indicates that amantadine blocks dissociation of the viral M1 protein from the RNP in the endosome (reviewed in references 14 and 18), and hence viral mRNA transcription and subsequent viral mRNA translation are blocked. To examine the extent to which inhibition of viral protein was incomplete at a concentration of amantadine (5 μM) that lowers the titer of virus ∼50-fold yet does not raise intracellular pH nonspecifically, a time course of viral protein synthesis for WSN, N31S-M2WSN, MUd/WSN, and Udorn viruses was performed. As shown in Fig. Fig.2,2, in the absence of amantadine, virus-specific protein synthesis can be observed as early as 1 h p.i., as observed previously (30, 56). For the amantadine-sensitive strains, in the presence of the drug much less protein synthesis occurred at each time point, and by 4 h p.i. the amount of NP protein synthesized was reduced by >80% from that observed in the absence of the drug. Amantadine had no effect on virus protein synthesis in WSN-infected cells, indicating that amantadine at the concentration used was not acting nonspecifically and affecting factors such as endosomal intralumenal pH, which in turn would affect influenza virus disassembly.
Effect of amantadine on the time course of influenza virus protein synthesis. Monolayers of MDCK cells were preincubated with DMEM with (+ aman) or without (− aman) amantadine (5 μM) for 30 min and then infected with WSN, N31S-M2WSN, MUd/WSN, or Udorn viruses at an MOI of 3.0 PFU per cell in the presence or absence of amantadine. After 1 h the virus inoculum was removed, and cells were incubated in DMEM with or without amantadine. At various intervals, cultures were labeled metabolically for 15 min with 35S-Promix (50 μCi/ml) in DMEM met-, cys-, cells lysed in protein lysis buffer (30), and the polypeptides were analyzed by SDS-PAGE. Radioactivity was analyzed and quantified with a Fuji BioImager 1000 and Mac Bas software. Graphs show the quantified results of the relative amount of NP (% NP) normalized to the amount of NP at 4 h p.i. in the non-drug-treated samples. Squares and circles show percent NP in the absence and in the presence of 5 μM amantadine, respectively.
Inhibition of primary translation by amantadine.
In influenza virus-infected cells, input genome vRNA is transcribed in the nucleus to mRNA (primary transcription), and in the presence of a protein synthesis inhibitor further RNA species are not synthesized (reviewed in reference 32). An indirect measure of mRNAs that have accumulated through primary transcription can be taken by examining protein synthesis in vivo immediately after washout of the protein synthesis inhibitor (30). If amantadine is added to cells both before influenza virus infection and throughout the protein synthesis block, the extent of the block to primary transcription and translation can be estimated. To block protein synthesis, MDCK cells were incubated with cycloheximide (100 μg/ml) before, during, and after virus infection. When required, amantadine (5 μM) was added before, during, and after infection. The cells were infected with N31S-M2WSN at a MOI of 5 PFU/cell and incubated in the drugs to allow mRNA synthesis. At 4 h p.i. the medium was removed and the cultures were washed five times with DMEM containing amantadine (5 μM). Cells were then metabolically labeled for 10 min with 35S-Promix (100 μCi/ml). Either polypeptides were analyzed directly by SDS-PAGE or NP was immunoprecipitated using an anti-NP monoclonal antibody. As shown in Fig. Fig.3,3, amantadine treatment of cells reduced the amount of NP produced by 84%. Addition of actinomycin D, to block further influenza virus transcript accumulation (26), made little difference to the amount of NP synthesized, indicating that significant amounts of new mRNAs were not synthesized and translated after cycloheximide washout. Although mRNAs synthesized during primary transcription were not analyzed directly, the data do indicate that amantadine causes a significant decrease in primary translation, consistent with the notion that the drug blocks virus uncoating. Previous analysis of primary transcript mRNAs synthesized by fowl plague virus in the presence of 1 mM amantadine indicated that mRNA transcript accumulation was nearly undetectable (57). Taken together, the time courses of the protein synthesis experiment (Fig. (Fig.2)2) and the primary translation experiment (Fig. (Fig.3)3) both indicate that amantadine significantly reduces viral protein synthesis, but there is a basal level of virus-specific protein synthesis (15 to 20% of no-drug control level), which we designate the amantadine leak level.
Primary translation of influenza virus proteins. Monolayers of MDCK cells (3.5-cm diameter dishes) were washed with PBS, and the medium was replaced with DMEM containing cycloheximide (100 μg/ml) and amantadine (5 μM) for 30 min. The cells were then infected with N31S-M2WSN at a MOI of 5 PFU/cell in the presence of cycloheximide (100 μg/ml) and amantadine (5 μM). After an adsorption period of 1 h, the inoculum was removed and replaced with DMEM containing cycloheximide (100 μg/ml) and amantadine (5 μM). When used, actinomycin D (act. D) (final concentration, 5 μg/ml) was added to cultures at 3.5 h p.i. At 4 h p.i., the medium was removed and the cultures were washed five times with 2 ml of DMEM containing amantadine (5 μM). This washing procedure did not exceed 5 min. Cells were then metabolically labeled for 10 min with 35S-Promix (100 μCi/ml) in DMEM met−, cys− containing amantadine (5 μM), and cells were lysed in cell lysis buffer. Polypeptides were analyzed directly by SDS-PAGE (A), or NP was immunoprecipitated using an anti-NP monoclonal antibody using the SDS immunoprecipitation protocol (B) (46). Lane M, marker polypeptides. Radioactivity was analyzed and quantified as described in the legend to Fig. Fig.2.2. The relative expression (Rel. exp.) levels of NP are indicated for the different treatments.
Generation and growth of an influenza virus with an M2 protein lacking ion channel activity.
We have previously reported on two mutations in the M2 TM domain, M2-A30P and M2del29–31, that caused a loss of ion channel activity when measured in oocytes of X. laevis (20). However, Watanabe and colleagues (68) found that when these mutations were incorporated into MUd/WSN, these viruses could be rescued from cloned DNA as efficiently as wt virus and that these mutant viruses showed growth kinetics indistinguishable from that of wt virus. Watanabe and colleagues (68) also found that viruses in which the M2 TM domain was replaced with the HA TM domain grew as efficiently as wt virus in tissue culture. We reinvestigated the ion channel properties of M2-del29–31, expressed in oocytes of X. laevis, with a two-electrode voltage clamp using the methods we reported previously (47). There was no detectable ion channel activity induced by low pH. Furthermore, there was no amantadine-sensitive ion channel activity. The currents were indistinguishable from control, uninjected oocytes that have a small (50 μA at −100 mV) amantadine-sensitive (100 μM) leakage current.
As a representative M2 protein that lacks detectable ion channel activity, we selected M2del29–31 and incorporated this deletion into the Udorn genetic background to create M2-del29–31Udorn virus from cloned DNA. A virus was recovered, on six independent occasions, in the same time frame as wt Udorn virus, but when plaqued in MDCK cells it formed only pinpoint plaques (Fig. (Fig.4A).4A). Analysis of the growth properties of M2-del29–31Udorn in comparison to wt Udorn virus indicated that virus yield was delayed, and the titer between 40 and 60 h p.i. was 1 to 1.5 logs lower for M2-del29–31Udorn than wt Udorn (Fig. (Fig.4B).4B). Amantadine (5 μM) had no effect on the size of the plaques of M2-del29–31Udorn, and the drug had no effect on the growth rate of M2-del29–31Udorn, indicating that the virus was resistant to amantadine (data not shown). Analysis of the complete nucleotide sequence for all eight RNA segments of M2-del29–31Udorn did not show any changes from the sequence of wt A/Udorn/72 except for the M2del29–31 mutation.
Plaque assay, immunostaining, and multiple-step replication kinetics of the influenza M2-del29–31Udorn virus. (A) M2-del29–31Udorn (Δ29–31) virus was recovered from cloned DNAs corresponding to the eight Udorn strain RNA segments such that it contained a deletion of three amino acids (residues 29 to 31) in the M2 protein TM domain as described in Materials and Methods. Wt Udorn (wt Ud) and M2-del29–31Udorn viruses were plaqued in MDCK cells and stained or immunostained as described in Materials and Methods. Plaque size varied for both wt Ud and M2-del29–31 within a given population. As is often found with influenza virus, this diversity in plaque size still occurred after replaquing of individual plaques. (B) Monolayers of MDCK cells were infected with wild-type Udorn (wt Ud) or M2-del29–31Udorn viruses at an MOI of 0.001 PFU per cell and incubated with DMEM supplemented with 1.0 μg of trypsin/ml. At various times the plaque titers of the culture supernatants were determined. Errors bars indicate ± standard error of the mean.
To compare the level of protein synthesis between wt Udorn and M2-del29–31Udorn, approximately equivalent numbers of physical virus particles were used, as determined by hemagglutinating units (HAU). Comparison of the time course of protein synthesis using equivalent PFU of the two viruses is not an informative measurement because if infectivity depends solely on the presence or absence of a natural or surrogate form of ion channel activity, then the time course of viral protein synthesis would be similar. MDCK cells were infected with 16 HAU of wt Udorn or M2-del29–31Udorn, and the time course of protein synthesis was examined. The extent of NP synthesis was calculated and normalized to the amount of wt Udorn NP at 6 h. As shown in Fig. Fig.5,5, for M2-del29–31Udorn there was much less protein synthesis occurring at each time point than with wt Udorn, and by 6 h p.i. the amount of NP protein was reduced by ∼90%. The amount of protein synthesis at each time point for M2-del29–31Udorn is comparable to that found for wt Udorn in the presence of amantadine (Fig. (Fig.22).
Comparison of the time course of influenza virus protein synthesis in M2-del29–31Udorn virus- and wt Udorn virus-infected cells. Monolayers of MDCK cells were infected with approximately equivalent numbers of physical virus particles, based on the HA titer (16 HAU). At various times p.i., the infected cells were labeled metabolically with 35S-Promix (100 μCi/ml) for 30 min, cells were lysed in protein lysis buffer, and the polypeptides were subjected to SDS-PAGE analysis. Radioactivity was analyzed and quantified as described in the legend to Fig. Fig.2.2. Open and filled squares show the quantified results of the relative amounts of NP (percent NP) of wt Ud and M2-del29–31Udorn, respectively, normalized to the 6-h wt Ud NP amount.
Taken together, the data for M2-del29–31Udorn on plaque size, growth rate, and time course of protein synthesis are all very similar to those for wt Udorn virus in the presence of amantadine. Thus, these data support the view that M2-del29–31Udorn grows equivalently to the amantadine leak level.
Biological fitness assay of M2-del29–31Udorn in comparison to wt Udorn.
As an alternative approach to comparing the efficiency of growth of M2-del29–31Udorn virus, which lacks detectable ion channel activity, to the growth of wt Udorn virus, we performed a biological fitness assay. We cocultivated in MDCK cells vastly disproportional amounts of M2-del29–31Udorn (7,000 PFU; MOI/cell = 0.01) with wt Udorn (70 PFU; MOI/cell = 0.0001). At 1, 2, and 4 days the nucleotide sequence of the region encoding the M2 protein TM domain from virus released into the cell medium was determined by reverse transcription-PCR. After 1 day the virus released from the cells had the sequence of M2-del29–31Udorn (Fig. (Fig.6),6), and after 2 days the virus released from the cells was a mixed population of M2-del29–31Udorn and wt Udorn (an approximately 5:1 ratio as estimated from the mixture standards). After 4 days, however, predominantly wt Udorn M2 sequence was detected, indicating that wt Udorn outgrew M2-del29–31Udorn. Thus, the preponderance of data lead us to conclude that the presence of ion channel activity has a selective advantage for the replication of influenza virus in tissue culture.
Fitness assay for M2-del29–31Udorn virus in comparison to wt Udorn virus. (A) Nucleotide sequence standardization electropherograms for known mixtures of wt Udorn M segment and M2-del29–31 M segment DNAs. Triangles indicate the first peak of the codon encoding amino acid 29 in the wt sequence. The green, black, blue, and red peaks indicate A, G, C, and T bases, respectively. (B) Seven thousand PFU (MOI per cell = 0.01) of the M2-del29–31Udorn (Δ29–31) virus was cocultured with 70 PFU (MOI per cell = 0.0001) of the wt Udorn virus in MDCK cells. At various times p.i., the resulting virus population was amplified by reverse transcription-PCR over the M2-encoding region, and the products were subjected to direct sequencing. By 2 days p.i., a mixture of M2-del29–31Udorn and wt Udorn sequence was present, but at 4 days p.i., predominantly wt sequence could be detected.
DISCUSSION
Recently Watanabe and colleagues (68) concluded that influenza A virus can undergo multiple cycles of replication without M2 ion channel activity. Although this conclusion may be correct if narrowly defined, it could be popularly construed to imply that influenza virus does not require the M2 ion channel activity. We chose to examine events, including those within a single replicative cycle (6 to 8 h) for influenza A virus-infected cells treated with the M2 ion channel inhibitor amantadine and for cells infected with M2-del29–31Udorn, a virus containing an M2 protein that lacks detectable ion channel activity. In contrast, Watanabe and colleagues (68) mostly examined growth over a much longer time scale (12 to 72 h p.i.), which precludes direct analysis of early events. Our data on determining the degree of inhibition of influenza virus by amantadine using a concentration that does not raise intracellular pH and characterization of the amantadine block are consistent with the observed growth properties of M2-del29–31Udorn: a small-plaque phenotype, reduced growth kinetics, and delayed replication in a single cycle. M2-del29–31Udorn plaque size and growth rate were unaffected by amantadine (5 μM). An additional factor that might influence the data of Watanabe and colleagues (68) is the possibility that the recovered viruses acquired second site mutations that overcame the lack of an M2 ion channel activity, e.g., a change in M1 protein or NP that altered the pH sensitivity of their interaction. The complete nucleotide sequence for all eight RNA segments of our M2-del29–31Udorn virus was determined, and it did not show additional mutations compared to the known wt Udorn sequence.
Amantadine results in an essentially complete blocking of the ion channel activity measured in oocytes, and for all practical purposes the block is irreversible (67). However, amantadine only blocks influenza virus yield in tissue culture about 90 to 99%, depending on the virus strain (16). Most likely the amantadine leak level does not reflect a failure to inhibit the ion channel activity. More likely, the amantadine leak level occurs because the requirement for protons to permeate the interior of the virus particle is not absolute or because for tissue-culture-grown virus and egg-grown virus, the M2 ion channel is not the only way for protons to gain entry into the virion. If uncoating relies on protonation of an ionizable group present inside the virion, then some protonation, and thus uncoating, will occur in the absence of proton entry because of the statistical nature of protonation. It has also been suggested that at the time of fusion pore formation, mediated by the low-pH-induced form of HA, protons could permeate virions (68). Alternatively, virions that have remained at 37°C in tissue culture medium or in egg allantoic fluid may become membrane leaky, a phenomenon described previously for the paramyxovirus, Sendai virus (22, 54).
The rapid single-step growth curve of influenza virus (6 to 8 h) most likely serves an important purpose in countering host defenses in an animal infection. Thus, factors that delay rapid replication are likely to be deleterious to survival of the virus. Indeed, these factors may contribute to the efficacy of rimantadine in the prophylaxis and treatment of influenza A virus infection in humans. Watanabe and colleagues (68) could not recover WSN/M Udorn M2-del29–31 from nasal turbinates in mice, and this observation may be related to a delay in the early phase of virus replication. Even in tissue culture cells, wt Udorn had an enormous selective advantage, as shown in our biological fitness assay of cocultivation of 100 times the PFU of M2-del29–31Udorn with wt Udorn.
In summary, we find that influenza virus which contains an M2 protein which lacks a detectable ion channel activity is debilitated in growth to an extent similar to that of influenza virus grown in the presence of amantadine, the M2 ion channel blocker. Thus, taking all the data together, we believe it is reasonable to conclude that the M2 ion channel protein, which is conserved in all known strains of influenza virus, evolved its function because it contributes to the efficient replication of the virus in a single cycle.
Acknowledgments
We are grateful to Yoshihiro Kawaoka for making available to us the complete plasmid set of A/WSN/33 virus necessary for establishing the reverse-genetics system for influenza viruses in our initial experiments. We thank all members of the Lamb laboratory for helpful discussions.
This research was supported by research grants R37 AI-20201 (R.A.L.) and AI-31882 (L.H.P.) and fellowship F32 AI-10382 (A.P.) from the National Institute of Allergy and Infectious Diseases. K.S. was supported by the National Institutes of Health Training Program in Cellular and Molecular Basis of Disease (GM-08061). M.T. is an Associate and R.A.L. is an Investigator of the Howard Hughes Medical Institute.