Two research teams have captured snapshots of the influenza virus's membrane-bound hydrogen-ion channel, which is essential for infection and virulence. Their findings agree on the basics, but differ in details.
The paperstorm of research directives blowing out of our biomedical policy offices sometimes drives me to take emotional shelter in a precept of one of my old professors: “Applied research, applied well, soon becomes basic research.” The common flu virus illustrates the broad wisdom of this maxim. Influenza and its human vector, playing a cat-and-mouse game for millennia, have come to a rather civilized understanding: you spread me around and I'll spare you to cough another day. Now that the bird-flu media hysterics have died down, we can once again view influenza dispassionately, as a prevalent disease carried by a low-virulence virus with a nasty penchant for — just to keep the game interesting — infrequent but lethal tantrums. This familiar malady has motivated decades of applied (now coyly termed 'translational') research that, instead of a magic bullet, has yielded deep insights across a broad swath of molecular mechanisms in biology, from RNA-based information-processing to weapons of immunological destruction to membrane fusion.
Two papers in this issue further exemplify the link between clinical research into influenza and basic discovery. Stouffer et al.1 (page 596) and Schnell and Chou2 (page 591) unveil high-resolution structures of M2, an ion-channel protein whose proton (H+)-conducting activity in the membrane of the influenza virion is necessary for infection. The structures portray our first views of a H+-specific channel, and they suggest how the virus has outfoxed amantadine, an anti-flu drug effective ten years ago but now well-nigh useless.
An influenza virion engulfed by a lung epithelial cell initially finds itself caged within an intracellular compartment, the endosome. When the virion's membrane fuses with that of the endosome — a process triggered by the acidic endosomal milieu — its RNA genome escapes into the cell to replicate and wreak collateral damage. But for this to work properly, acid must enter the virion through a 'leak' pathway in its membrane — the M2 channel3,4.
Proton currents mediated by M2 have been described through electrophysiological studies4. Earlier work also showed that M2 assembles into a membrane-spanning tetramer5 to form the channel, which opens and closes at low and neutral pH, respectively. Each of the four protein subunits of the channel is seemingly simple, with only 97 amino-acid residues and a single transmembrane helix.
The M2 structures emerge from complementary high-resolution techniques applied to differently truncated versions of this channel. To solve the structure of the transmembrane peptide of M2, Stouffer and colleagues1 used X-ray crystallography, and for capturing the image of a longer peptide that includes 15 residues following the transmembrane region, Schnell and Chou2 used nuclear magnetic resonance (NMR) spectroscopy. Casually viewed, the two structures, which were determined in the presence of amantadine-like inhibitors, agree well. Each is a four-helix, cone-shaped bundle with a polar, proton-friendly pore running along a central axis that is topped by a constriction too narrow for any other type of ion to pass (Fig. 1). Moreover, two functionally 'hot' residues — the gate (Trp 41), which opens when the proton sensor (His 37) experiences low pH — occur at locations that make sense4. The NMR structure is apparently a closed state, as the four inward-pointing Trp 41 side chains occlude the pore, whereas the X-ray structure seems to be open, with the helices splaying out on the cytoplasmic side to widen the Trp 41 gate.
Despite this general agreement, however, these papers are going to generate sharp controversy, as is intimated by the gentle whiffs of tendentiousness appearing sporadically in each. The controversy stems not so much from incongruities in structural detail — the channels, after all, are thought to reflect different gating conformations — but from the incompatible amantadine-inhibition mechanisms inferred. The disagreement is not subtle. The X-ray structure shows a single amantadine molecule plugging the open pore. The NMR structure, with no room in the closed pore, surprises us with four drug molecules bound at the channel's lipid-exposed outer surface, one at each helix–helix interface. Each picture suggests its own inhibitory mechanism. Stouffer et al. envision the drug physically blocking the H+ pathway in the open pore. Schnell and Chou, in contrast, propose that the drug binds preferentially to, and thereby stabilizes, the closed state. (I avoid employing the once-precise word 'allosteric', now so carelessly overused as to have become almost meaningless.) Both mechanisms are known to be used by various inhibitors of other ion channels.
Each team cites evidence in favour of its own proposed mechanism. In the now common amantadine-resistant flu strains6, M2 mutations at any one of six positions in the transmembrane region impair drug inhibition of the channel4. In the open structure, four of these residues project side chains into the pore near the amantadine-blocking site, where substitutions could plausibly disrupt drug binding. The closed structure, by contrast, shows three of these side chains projecting sideways to suture the helix–helix interfaces together, far away from the bound amantadine; mutations here could plausibly destabilize the closed state and hence disfavour amantadine binding.
But the correspondence is not perfect, as several of these six side chains project in the wrong direction for each proposed mechanism. I don't consider this a serious weakness, however, as psychoanalysis of protein mutations, even with computational assistance, is an uncertain undertaking. More directly relevant to the issue are several details of amantadine action on full-length M2, but even here neither candidate gets the vote. A stoichiometry of one amantadine molecule per tetramer, inferred indirectly from earlier studies7,8, naturally fits the X-ray but not the NMR picture. But amantadine action is faster at neutral pH, where the channel is mostly closed, than at low pH, where the open state is favoured7, in accord with closed-state stabilization but not with open-pore block.
These and other ambiguities abound, not least of which is the embarrassment that neither truncated construct used for the structures has been certified for proper H+-channel function. It had been convincingly shown9 that the minimal transmembrane peptide forms amantadine-binding tetramers in detergent solution. But the literature is littered with disquieting variability in this peptide's ion-transport function in membranes; single-channel turnover rates, for instance, range from 10 to 107 ions per second (in one case10, this millionfold discrepancy appearing in the same figure). So although the new structures1,2 give us consistent glimpses of this fascinating proton channel's overall architecture, they clash in their mechanistic inferences. As for representations of the M2 channel in a biological membrane, the discrepant views cry out for resolution, which will require further structural work combined with its essential companion, close functional scrutiny.
Stouffer, A. L. et al. Nature 451, 596–599 (2008).
Schnell, J. R. & Chou, J. J. Nature 451, 591–595 (2008).
Helenius, A. Cell 69, 577–578 (1992).
Pinto, L. H. & Lamb, R. A. J. Biol. Chem. 281, 8997–9000 (2006).
Sakaguchi, T., Tu, Q., Pinto, L. H. & Lamb, R. A. Proc. Natl Acad. Sci. USA 94, 5000–5016 (1997).
Bright, R. A. et al. Lancet 366, 1175–1181 (2005).
Wang, C., Takeuchi, K., Pinto, L. H. & Lamb, R. A. J. Virol. 67, 5585–5594 (1993).
Czabotar, P. E., Martin, S. R. & Hay, A. J. Virus Res. 99, 57–61 (2004).
Salom, D., Hill, B. R., Lear, J. D. & DeGrado, W. F. Biochemistry 39, 14160–14170 (2000).
Hu, J. et al. Proc. Natl Acad. Sci. USA 103, 6865–6870 (2006).
About this article
Arylation of adamantanamines: VIII. Optimization of the catalytic system for copper-catalyzed arylation of adamantane-containing amines
Russian Journal of Organic Chemistry (2017)
Fast Atomic Charge Calculation for Implementation into a Polarizable Force Field and Application to an Ion Channel Protein
Journal of Chemistry (2015)
Chemical Reviews (2013)
Computational study of drug binding to the membrane-bound tetrameric M2 peptide bundle from influenza A virus
Biochimica et Biophysica Acta (BBA) - Biomembranes (2011)
Heterodimers of Histidine and Amantadine as Inhibitors for Wild Type and Mutant M2 Channels of Influenza A
Chinese Journal of Chemistry (2010)