Influenza A virus, which causes flu, is known as one of the most prevalent pathogenic viruses. Seasonal flu results in more than 50,000 deaths worldwide every year [1]. Pandemic flu caused over 50 million deaths in 1918 and about 22 million people infected in 2009 [2]. At present, the basic principles for defense of Influenza A virus are development of vaccines and antiviral drugs [3]. However, timely development of effective vaccines against new variants for each year is highly difficult and challenging, as influenza A virus is readily able to develop new variants resistant to new vaccines by simple genomic reassortment of hemaglutanin (HA) and neuroaminidase (NA). As a result, development of vaccines always lags behind the rate that a new subtype of Influenza A virus emerges. Although antiviral drugs is an alternative for effective treatment of influenza A virus infection, influenza A virus again readily develops drug resistance to these drugs by acquiring new mutations.

The most commonly used targets for antiviral drugs are M2 ion channel or neuraminidase encoded by the viral genome [4], [5]. M2 protein, which is crucial for viral infection, forms a tetrameric proton channel in the host cell membrane. Amantadine and rimantadine are well-known inhibitors to effectively block the wild type M2 proton channel and were used widely in the past for flu treatment [5]. But due to emergence of numerous M2 mutants, influenza A virus now is almost 100% resistant to these drugs. Here, we developed a novel high throughput screening system for agents to effectively block the M2 proton channel, thus to prevent viral infection.

The drug screening system was developed using the single protein production (SPP) system in E. coli [6], [7]. In the SPP system, the synthesis of all cellular proteins is inhibited by the induction of MazF, an ACA-specific mRNA interferase from Escherichia coli, except for the synthesis of a designated protein, of which mRNA is designed to have no ACA sequences [7]. Thus, an ACA-less gene coding for M2 protein (from residue 2 to 49 for M2 channel) was synthesized and cloned into pColdII(sp-4) [7]. Expression of M2 protein was toxic, and hence hardly detected (Figure 1A, lane 2), presumably because it disrupted the proton gradient across the membrane [8]. As a result, cells are unable to produce ATP and protein production is temporarily stopped. However, it was sufficiently well expressed to be easily detectable by SDS-PAGE in the presence of amantadine (Figure 1A, lanes 3, 4, and 5). This result clearly indicates that M2 protein is capable of forming an active proton channel, which is sensitive to amantadine. To validate this conclusion, we further tested seven other drugs for their inhibitory effect on M2 proton channel. Five of them were found to function as effective M2 channel inhibitors (Figure 1B). Similar results were observed with an N-terminal AcGFP-M2 fusion protein (Figure 1C). AcGFP is a green fluorescent protein from Aequorea coerulescens, which contains one residue substitution in its chromophore to make it much brighter than the native green fluorescent protein [9]. From these results, compounds 10, 34, 35, 282 and 293 were identified to be potential inhibitors of influenza A virus infection.

In the SPP system, cell growth is completely inhibited, while cells are metabolically active as they are at the quasi-dormancy state so that side effects of drugs on cell growth can be excluded. Surprisingly, although AcGFP fails to be as a reporter since it is not fluorescent in the AcGFP-M2 fusion protein, we found in the present manuscript that cell growth is resumed as measured by the increase of cell density using O.D.₆₀₀, when AcGFP-M2 fusion protein is expressed in the presence of inhibitors for the M2 channel activity (Figure 1D). Significant increases of the cell density were observed in the cultures, in which compounds 10, 34, 35, 282, 293 and amantadine were added, while compounds 15 and 314 were unable to resume cell growth as the control experiment without addition of any drugs. In a time-course experiment (Figure 1E), M2 protein or AcGFP-M2 fusion protein was induced together with MazF at 5 hrs by the addition of 1 mM IPTG in the presence or absence of 50 μM amantadine. To our surprise, cell growth was resumed 3 hrs after induction only in the culture containing amantadine to produce AcGFP-M2 (shown by arrows). The amount of AcGFP-M2 further increased for the next 16 hrs in the culture with amantadine, while no significant changes in protein patterns were observed in the other culture without the addition of amantadine (Figure 2A), Expression of M2 protein was observed only in the presence of amantadine and accumulated gradually during the 19-hr induction time (Figure 2B).

We further tested the inhibitory effect of other eight compounds on the production of AcGFP-M2 and found that out of these compounds, eight compounds (10, 34, 35, 180, 189, 206, 282 and 293) were able to resume the synthesis of the fusion protein (Figure 3A). Consistently with the result with amantadine, cell growth was resumed after overnight incubation with all these eight compounds as in the case of amantadine (Figure 3B).

Next, we tested if the same compounds are able to inhibit the M2 (V27A) mutant channel, which is known to be resistant to amantadine [10], [11]. The V27A mutation was introduced to pColdIIAcGFP-M2 (sp-4) by site-directed PCR mutagenesis. It was found that Ac-GFP-M2 (V27A) was well expressed only in the presence of compound 180 or 206 (Figure 3C). Subsequent O.D.₆₀₀ measurements confirmed that only the SPP expression systems that contain compounds 180 and 206 displayed significant cell growth resumption (Figure 3D).

Similar results were obtained by electrophysiology assay in Xenopus laevis oocyte. As summarized in tabe1, Compound 180 and 206 appear to be the best two inhibitors for M2 (V27A) mutant ion channel, which exhibit 82.7% and 94.5% inhibition of channel conductivity, respectively (Tabel 1). The eight compounds (10, 34, 35, 180, 189, 206, 282 and 293), which was identified above as potential inhibitors of M2 channel, show around 90% or more inhibition of the wild type M2 ion channel conductivity in the electrophysiology assay. A few compounds were also tested in plaque reduction assay as reported earlier [12], [13]. Compound 34 and 35 significantly inhibit wild type A/M2 virus replication at 5 µM, comparable to the potency of amantadine at the same concentration. In consistent with the electrophysiology assay, compound 34 also inhibits the replication of virus harboring A/M2-V27A mutation when tested at 50 µM [12]. Another two compounds 180 and 206, which are more potent than 34 and 35 in inhibiting V27A in electrophysiology assay, show higher viral inhibition activities at 10 µM. Remarkably, the most potent compound 206 completely inhibits V27A virus replication at 5 µM concentration [13]. Compound 34 and 35 correspond to compound 8 and 9, respectively, in Balannik V. et al., 2009. Compound 180 and 206 correspond to compound 1 and 9, respectively, in Wang J. et al., 2011.

Subsequently, screening was carried out by using 96-well plates. A 100-ml culture of the SPP system harboring pColdII AcGFP-M2 (sp-4) was grown to O.D.₆₀₀ = 0.5∼0.6. After cold-shock at 15°C for 60 min, 1 mM of IPTG was added to the culture for induction of AcGFP-M2 fusion protein. A hundred μl of the culture was distributed to each well of 96-well plates and individual compounds were added to the wells. After overnight incubation at 15°C with vigorous shaking, O.D.₆₀₀ was measured using micro plate reader (Molecular devices spectramax 384 plus). A similar set of compounds as previously isolated were observed to be potential inhibitors for M2 proton channel since cell growth in these wells were detected (Figure 3E).

It was previously reported that expression of M2 in bacteria would result in retarded cell growth because the host cell membrane became permeable by the M2 channel [14]. However, our results demonstrate for the first time a high correlation between the rate of growth resumption and the expression of AcGFP-M2 protein in the SPP system. The reason that causes growth resumption in the SPP system is still unknown. It has been confirmed that cells after growth resumption still contain the plasmid which produces MazF and MazF retains its toxicity (data not shown). It is suspected that some fragments from AcGFP-M2 fusion protein may play a role in neutralizing the toxicity of MazF. Nevertheless, this screening method using E. coli provides an extremely simple, easy and fast assay system to screen anti-influenza drugs in comparison with tedious patch clamping assay systems to measure the channel conductivity [15]. Using this method we identified a number of channel inhibitors, which can be potentially used as anti-influenza drugs. This high throughput screening system may be extended for screening of drugs against ion channel proteins from other RNA and DNA viruses such as Vpu protein form HIV-1 and p7 from HCV [16], [17].