No job name

J. Med. Chem. 2004, 47, 4975-4978
Interaction between an Amantadine Analogue and the Transmembrane Portion
of the Influenza A M2 Protein in Liposomes Probed by 1H NMR Spectroscopy of
the Ligand
Antonios Kolocouris,† Raino K. Hansen,‡ and R. William Broadhurst*,§ Department of Pharmaceutical Chemistry, School of Pharmacy, University of Athens, Panepistimioupolis, Zografou,Athens 15 771, Greece, CBR, IMBG, The Panum Institute, University of Copenhagen, 2200 Copenhagen N, Denmark, andDepartment of Biochemistry, University of Cambridge, 80 Tennis Court Road, Cambridge CB2 1GA, U.K. 1H NMR spectroscopy of a fluoroamantadine ligand was used to probe the pH dependence ofbinding to the transmembrane peptide fragment of the influenza A M2 proton channel (M2TM)incorporated into 1,2-dimyristoyl-sn-glycero-3-phosphocholine liposomes. Above pH 7.5, whenM2TM bound the ligand, fluoroamantadine resonances became too broad to be detected.
Fluoroamantadine interacted weakly with the liposomes, indicating it may first bind to thebilayer and then block target channels after diffusion across the membrane surface.
In Europe in the 20th century pandemics of influenza and Udorn strains: SSDPLVVAASIIGILHLILWILDRL) A caused more fatalities than any other infectious forms tetrameric amantadine-sensitive proton channels disease. During less severe epidemics, g10% of the in planar lipid bilayers.4,5 Insights from FT-IR,6 solid- population can be infected, resulting in temporary state NMR,7 and molecular dynamics8 have yielded a debilitation, with significant economic consequences.1a,b high-resolution structure of the monomer backbone of The likelihood of future pandemics highlights the need M2TM in 1,2-dimyristoyl-sn-glycero-3-phosphocholine to develop more effective therapies.
(DMPC) bilayers. Along with the measurement of a few Amantadine (Am) 1 (1-aminoadamantane hydrochlo-
intermolecular distance constraints,7 these techniques ride), a drug in the aminoadamantane series,2 is li- have led to models of the M2TM tetrameric assemblyas a left-handed parallel bundle of R-helices. Thesemodels account for cysteine mutagenesis data,9 indicatethat residues V6, A9, G13, H16, and W20 of M2TM linethe pore, and are consistent with the imidazole sidechain of H16 both acting as a proton shuttle andinteracting with the indole side chain of W20 to occludethe pore. 7b Am 1 blocks the proton channel activity of M2. A
censed for use in the prophylaxis and therapy of
influenza A virus infections. Am 1 inhibits the activity
current hypothesis is that this is achieved via interac- of the influenza A M2 protein, which is critical for virus tions of the drug with the mouth of the M2 pore.
replication. M2 is a small 97-residue integral membrane Molecular modeling suggests that the luminal space protein with an essential role in the acid-induced between L26 and H37 is complementary in its shape, process of uncoating the viral RNA during infection. It hydrophobicity, and polarity to Am 1. Binding of the
comprises an extracellular N-terminal domain (21 resi- drug is expected to block proton channel activity by dues), a transmembrane (TM) domain (25 residues), and displacing water molecules that are essential for proton an intracellular C-terminal domain (51 residues). M2 self-assembles as a homotetramer to form an ion chan- A variety of biophysical techniques have been used nel. Each subunit of the tetramer contributes a single to study the monomer/tetramer equilibrium of M2TM transmembrane R-helix (M2TM) that together consti- and its interaction with Am 1 over a range of pH values
tute a proton selective pore that is activated by low pH in dodecylphosphocholine (DPC) micelles.4a In this paper environments, such as those found in endosomes.3 The we use the line shape of ligand resonances in 1H NMR M2 channel allows protons to enter the interior of the spectra to report on the interactions between amanta- viral particle, and this acidification dissociates the dine analogue 2 and the M2TM peptide in DMPC
matrix protein M1 that coats the viral RNA genome.1c liposomes, a more realistic lipid bilayer system.4b NMR It has been shown that the 25-residue M2TM peptide spectroscopy has been widely used to study interactions (the transmembrane portion of the wild-type M2 pro- between ligands and receptors by observing changes in tein, corresponding to residues S22-L46 of the Singapore the properties of either the macromolecule or the smallmolecule. Several different NMR parameters can be * To whom correspondence should be addressed. Phone: 44-1223- monitored for small molecules, including changes in 766035. Fax: 44-1223-766002. E-mail: r.w.broadhurst@bioc.cam.ac.uk.
chemical shift, relaxation rates, translational diffusion, and intermolecular or intramolecular magnetization transfer.11 An advantage of monitoring the ligand is that Journal of Medicinal Chemistry, 2004, Vol. 47, No. 20 there is no upper limit to the size of the receptor thatcan be investigated. This is an important considerationfor studies of integral membrane protein systems bysolution NMR techniques.
Many important pharmacological agents are first solvated in the lipid bilayer prior to their interaction
with their protein target.12 Since the adamantane core
of the aminoadamantane drugs is hydrophobic, we also
undertook a 1H NMR study of the interaction between
aminoadamantane 2 and DMPC lipid bilayers. The
isosteric replacement of a bridgehead hydrogen atom of
Am 1 with fluorine, resulting in F-Am 2,13 provides a
useful probe molecule for studying drug-M2TM and
drug-lipid bilayer interactions by both 1H and 19F
NMR.
The effects of the binding of Am 1 on the monomer/
tetramer equilibrium of M2TM over pH 5-8 have Figure 1. 1H NMR spectra of samples containing (a) 0.6 mM
recently been investigated by a variety of biophysical F-Am 2 in phosphate buffer at 298 K, pH 8, (b-e) 0.6 mM
techniques. Fluorescence spectroscopy, circular dichro- F-Am 2 in phosphate buffer containing 30 mM DMPC at 298
K for pH range 5-8, and (f) F-Am 2.
ism (CD), and analytical ultracentrifugation (AUC) inDPC micelles have shown that tetramerization of M2TM weak unassigned signal appeared at 2.45 ppm, close to and binding of Am 1 are both favored at pH 7.5-8. A
the H-5,7 resonance of F-Am 2 in the absence of DMPC.
pKa of 6.8 was determined for the histidine side chain The broadening of the signals of aminoadamantane 2
of M2TM in the monomeric state in DPC micelles by in the presence of DMPC vesicles suggests that free drug 1H NMR.4a It was suggested that the M2TM channel molecules bind to and are released from the large conducts protons via one or two protonated His side phospholipid assemblies at a rate that is fast compared chains that line the pore of the tetramer and that Am to the changes in chemical shift between the bound and 1 binds to the neutral tetrameric closed state of the
unbound states.11 Although the unbound drug is likely channel at elevated pH.4a Electrophysiological studies to be in excess, the NMR spectrum is strongly affected also showed that Am 1 possesses its highest binding
by the rapid relaxation rate of the fraction that interacts affinity for the wild-type M2 protein at pH 814 and that with the liposomes. The line broadening effect could also M2TM forms proton-selective channels.15 Thus, studies be caused by the rate of exchange between the bound of the M2TM peptide have implications for the function and free states being in the intermediate regime or by large changes in magnetic susceptibility at the bilayer In this work we describe a 1H NMR study of the interface. Increasing the acidity of the sample from pH M2TM/aminoadamantane drug system in DMPC lipo- 8 to pH 5 made little difference to the appearance of somes over the pH range 5-8. The signals of F-Am 2
the 1H spectrum, although the signal of unbound acetate were used as a probe to follow the binding of the drug at 2.0 ppm shifted slightly downfield toward pH 5 to M2TM, in contrast to previous studies, which focused (Figure 1b-e). When Am 1 was added to DMPC vesicles
instead on the spectral characteristics of the peptide at the same concentration, the signals from the drug itself in DPC micelles.4a The DMPC liposomes used in were too broad to be observed (not shown), suggesting the present study mimic native lipid bilayers more that Am 1 binds to the liposomes with higher affinity
closely than detergent micelles,4b which have notable than F-Am 2.
differences in curvature, lateral packing pressure, and Between pH 5 and 8 the presence of M2TM enhanced the stability of DMPC liposomes, which remained col-loidal for several weeks at room temperature.17 The M2TM/Aminoadamantane Drug System in
dimensions of these vesicles (diameter of ∼70 nm by DMPC Liposomes
DLS and EM) are similar to the natural membranes in 1H NMR spectra of Am 1 and F-Am 2 in aqueous
which M2 is known to function, the virion surface, and solution are shown in Supporting Information. Between the trans Golgi network.1 When F-Am 2 was added to
pH 5 and pH 8 no significant changes in the spectra of DMPC/M2TM proteoliposomes at pH 8, the line broad- Am 1 or F-Am 2 were observed. This is consistent with
ening of ligand signals in the 1H NMR spectrum was both molecules possessing a high pK 16 more dramatic (Figure 2a). Resonances from F-Am 2
the same protonated state throughout the pH range could no longer be detected, with only signals from residual molecules of DMPC in solution and unbound Dynamic light scattering (DLS) and electron micros- acetate remaining. This result indicates that F-Am 2
copy (EM) studies indicated that between pH 5 and 8 interacts more strongly with lipid vesicles in the pres- DMPC liposomes prepared by dialysis of mixed lipid/ ence of the M2TM peptide than with DMPC liposomes detergent micelles remained colloidal at room temper- alone, reducing the off rate so that the contribution from ature for > 5 days and possessed diameters between the transverse relaxation rate of the bound state 50 and 75 nm (not shown). In the presence of DMPC dominates. After acidification of the sample to pH 5, liposomes at pH 8, the 1H NMR signals of F-Am 2
resonances from F-Am 2 reappeared (Figure 2b), with
broadened considerably, but no changes in chemical line widths similar to those found in the presence of shift were observed, as illustrated in Figure 1a,b. A DMPC vesicles alone (Figure 1b). At low pH in the Journal of Medicinal Chemistry, 2004, Vol. 47, No. 20 able to bind M2TM under acidic conditions; the alteredstructure of the tetramer has lower affinity for the drugso that any residual interactions cannot be distin-guished from those between the drug and DMPC.
At pH 7.5 the resonances of F-Am 2 were significantly
broadened in the presence of M2TM and DMPC lipo-
somes; by pH 8 they were no longer detectable (Figure
2d,e). Since M2TM exists mainly as a neutral tetramer
above pH 7.5 in DPC micelles, this behavior can be
interpreted in terms of the recovery of a peptide tet-
ramer conformation that is capable of binding to the
drug in a reversible manner. In agreement with this
view, at pH g 7.5 in DMPC liposomes Am 1 has been
shown to insert into the pore of the M2TM channel,
where it is stabilized through favorable hydrophobic and
polar interactions.10 It is worth noting that simple 1H
NMR spectroscopy was successful at following the
binding of F-Am 2 to M2TM only because the interaction
Figure 2. 1H NMR spectra of samples containing 0.6 mM
between this drug and DMPC was relatively weak. The F-Am 2 in phosphate buffer with 30 mM DMPC and 0.6 mM
M2TM monomer at 298 K: (a) pH 8; (b) pH 5; (c) pH 7; (d) pH
disappearance of resonances due to Am 1 in the pres-
7.5; (e) pH 8. Annotations used are the following: D, DMPC ence of DMPC leaves no way of discriminating further peaks; F, F-Am 2 peaks; /, free acetate.
interactions with peptides incorporated into the bilayer.
Although the chemical shift degeneracy of Am 1 gives
presence of M2TM, the drug appeared to lose the it the potential to act as a high-sensitivity probe of increased affinity for the bilayer environment that it had interactions with the closed state of the M2TM channel, displayed at pH 8. Increasing the pH to 7 had little our results highlight the utility of the fluorinated further effect (Figure 2c), but after a further increment derivative. Our ongoing work with this system aims to to pH 7.5 the resonances of F-Am 2 showed signs of
exploit the high sensitivity and low background signal additional broadening (Figure 2d). When the sample of 19F NMR by introducing fluorine labels into the was returned to pH 8, all resonances from F-Am 2
M2TM peptide as well as the F-Am 2 ligand.
The DMPC/M2TM molar ratio used in this study (25) Conclusions
favors the formation of the tetrameric state of M2TMat pH 8.4b This pH value is optimal for the binding of The work presented in this paper is an extension of Am 1 to both full length M2 and M2TM.4,14 The
previously reported studies of the pH dependence of the disappearance of the signals of F-Am 2 at pH 8 in the
monomer/tetramer equilibrium of and the binding of Am presence of the peptide is therefore consistent with the 1 to the M2TM peptide in DPC micelles. Our investiga-
drug binding tightly to intact M2TM channels that are tion used simple 1H NMR spectroscopy to focus on a embedded in the bilayers of slowly tumbling lipid fluorinated aminoadamantane analogue and its interac- tions with M2TM incorporated into DMPC liposomes.
The 1H NMR spectra obtained at pH 5-7 (Figure 2b,c) Both approaches demonstrate that M2TM is not able were similar to the spectra of samples containing only to bind the aminoadamantane ligand between pH 5 and F-Am 2 and DMPC (Figure 1b-d), indicating that under
pH 7, either because of the formation of a non-native these conditions there is no special interaction between monomeric state or a change in the conformation of the the drug and the M2TM channel. This result was not tetrameric M2TM assembly. Above pH 7.5, M2TM unexpected because in DPC micelles M2TM exists adopts a neutral tetrameric form that binds the drug mainly in a monomeric state that does not bind Am 1
with high affinity. Under these conditions NMR signals between pH 5 and pH 7.4a The oligomeric state of M2TM from the drug cannot be detected. A broadening of drug in DMPC vesicles has not yet been established, although resonances in the presence of DMPC suggests that the tetramers have recently been observed in studies of a drug interacts with the lipid bilayers first and then possibly diffuses across the membrane until it encoun- of intermolecular cysteine bridges in the extracellular ters a tetrameric M2TM assembly in the closed state, domain of the full length M2 protein indicates that the whereupon it binds into the channel pore. 12 channel is tetrameric under all conditions in native Since we focused on inspection of the aminoadaman- membranes. The monomeric state observed for M2TM tane ligand rather than its peptide M2TM receptor, we in DPC micelles at low pH is therefore probably an anticipate that this study will have implications for artifact of this non-native membrane-mimetic environ- future work on the medicinal chemistry of this series ment. The dissociation constant of the tetrameric state of drugs. In principle, similar ligand focused experi- of M2TM19-46 in DMPC liposomes has been estimated ments could be extended to investigate the interaction to be smaller than that in DPC micelles by a factor of of aminoadamantane drugs with full length receptors >102.4b Hence, although M2TM probably remains tet- on intact virus particles.18 Future 19F NMR experiments rameric between pH 5 and 7 in DMPC liposomes, it is using F-Am 2 will aim to elucidate further aspects of
likely that the conformation of the tetramer is different how aminoadamantane drugs exert their anti-influenza from that at pH 8 because of the opening of the proton A activity. Since M2TM is also a minimal model for the channel. This could explain why F-Am 2 is no longer
study of proton channel proteins, such results would Journal of Medicinal Chemistry, 2004, Vol. 47, No. 20 Scheme 1a
(4) (a) Salom, D.; Hill, B. R.; Lear, J. D.; DeGrado, W. F. pH- dependent tetramerization and amantadine binding of the
transmembrane helix of M2 from the influenza A virus. Bio-
chemistry 2000, 39, 14160-14170. (b) Cristian, L.; Lear, J. D.;
DeGrado, W. F. Use of thiol-disulfide equilibria to measure the
energetics of assembly of transmembrane helices in phospholipid
bilayers. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 14772-14777.
(5) Duff, K. C.; Kelly, S. M.; Price, N. C.; Bradshaw, J. P. The secondary structure of influenza A M2 transmembrane domain.
A circular dichroism study. FEBS Lett. 1992, 311, 256-258.
(6) Kukol, A.; Adams, P. D.; Rice, L. M.; Brunger, A. T.; Arkin, I. T.
Experimentally based orientational refinement of membrane
protein models: a structure for the influenza A M2 H+ channel.
J. Mol. Biol. 1999, 286, 951-962.
(7) (a) Wang, J.; Kim, S.; Kovacs, F.; Cross, T. A. Structure of the transmembrane region of the M2 protein H+ channel. Protein
Sci. 2001, 10, 2241-2250. (b) Nishimura, K.; Kims, S.; Zhang,
L.; Cross, T. A. The closed state of a H+ channel helical bundle Reagents and conditions: (a) KMnO4, KOH, 3 h, 60 °C (yield combining precise orientational and distance restraints from 70%); (b) TBAHSO4, NaHCO3, CH3I, acetone, 48 h, room temp solid state NMR. Biochemistry 2002 41, 13170-13177. (c) Tian,
(99%); (c) DAST, CH2Cl2, 3 h, -80 °C f 25 °C (48%); (d) NaOH, C.; Gao, P. F.; Pinto, L. H.; Lamb, R. A.; Cross, T. A. Initial MeOH, THF, H2O, 12 h, room temp (85%); (e) DPPA, TEA, 45 min, structural and dynamic characterization of the M2 protein reflux, then BnOH, benzene, 72 h, reflux (quantitative); (f) H2, transmembrane and amphipathic helices in lipid bilayers.
10% Pd/C, AcOH, 50 psi, 6 h, room temp (84%).
Protein Sci. 2003, 12, 2597-2605.
(8) Forrest, L. R.; Kukol, A.; Arkin, I. T.; Tieleman, D. P.; Sansom, M. S. P. Exploring models of the influenza A M2 channel: MD open new avenues for discovering how ion channel- simulations in a phospholipid bilayer. Biophys. J. 2000, 78, 55-
blocker interactions can be described.
(9) Pinto, L. H.; Dieckmann, G. R.; Gandhi, C. S.; Shaughnessy, M.
A.; Papworth, C. G.; Braman, J.; Lear, J. D.; Lamb, R. A.; Experimental Section
DeGrado, W. F. A functionally defined model for the M2 proton The synthesis of F-Am 2 is illustrated in Scheme 1. Details
channel of influenza A virus suggests a mechanism for its ion
selectivity. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 11301-11306.
of the preparation of liposomes and proteoliposomes and the (10) Gandhi, C. S.; Shuck, K.; Lear, J. D.; Dieckmann, G. R.; 1H NMR spectroscopy are reported in Supporting Information.
DeGrado, W. F.; Lamb, R. A.; Pinto, L. H. Cu(II) inhibition ofthe proton translocation machinery of the influenza A virus M2 Acknowledgment. A.K. gratefully acknowledges
protein. J. Biol. Chem. 1999, 274, 5474-5482.
(11) Stockman, B. J.; Dalvit, C. NMR screening techniques in drug the Royal Society for the award of a travel grant to discovery and drug design. Prog. Nucl. Magn. Reson. Spectrosc. perform NMR experiments in Cambridge.
2002, 41, 187-231.
(12) Mason, R. P.; Rhodes, D. G.; Herbette, L. G. Reevaluating Supporting Information Available: Experimental de-
equilibrium and kinetic binding parameters for lipophilic drugsbased on a structural model for drug interaction with biological tails. This material is available free of charge via the Internet membranes. J. Med. Chem. 1991, 34, 869-877.
(13) Jasys, V. J.; Lombardo, F.; Appleton, T. A.; Bordner, J.; Ziliox, M.; Volkmann, R. A. Preparation of fluoroadamantane acids and References
amines: impact of bridgehead fluorine substitution on thesolution- and solid-state properties of functionalized adaman- (1) (a) Cox, N. J.; Subbarao, K. Global epidemiology of influenza: tanes. J. Am. Chem. Soc. 2000, 122, 466-473 and references
past and present. Annu. Rev. Med. 2000, 51, 407-421. (b) Reid,
A. H.; Taubenberger, J. K. The origin of the 1918 pandemic (14) Wang, C.; Takeuchi, K.; Pinto, L. H.; Lamb, R. A. Ion channel influenza virus: a continuing enigma. J. Gen. Virol. 2003, 84,
activity of influenza A virus M2 protein: characterization of the 2285-2292. (c) Helenius, A. Unpacking the incoming influenza amantadine block. J. Virol. 1993, 67, 5585-5594.
virus. Cell 1992, 69, 577-578.
(15) Duff, K. C.; Ashley, R. H. The transmembrane domain of (2) (a) Kolocouris, N.; Kolocouris, A.; Foscolos, G. B.; Fytas, G.; influenza A M2 protein forms amantadine-sensitive proton Neyts, J.; Padalko, E.; De Clercq, E. Synthesis and antiviral channels in planar lipid bilayers. Virology 1992, 190, 485-489.
activity evaluation of some new aminoadamantane derivatives.
(16) F-Am 2 has a pKa ≈ 9.5. See ref 13.
2. J. Med. Chem. 1996, 39, 3307-3318. (b) Zoidis, G.; Kolocouris,
(17) Hansen, R. K.; Broadhurst, R. W.; Skelton, P. C.; Arkin, I. T.
N.; Foscolos, G. B.; Kolocouris, A.; Fytas, G.; Karayannis, P.; Hydrogen/deuterium exchange of hydrophobic peptides in model Padalko, E.; Neyts, J.; De Clercq, E. Are the 2-isomers of the membranes by electrospray ionization mass spectrometry. J. Am. drug rimantadine active anti-influenza A agents? Antiviral Soc. Mass Spectrom. 2002, 13, 1376-1387.
Chem. Chemother. 2003, 14, 153-164.
(18) Meyer, B.; Peters, T. NMR spectroscopy techniques for screening (3) Sakaguchi, T.; Tu, Q.; Pinto, L. H.; Lamb, R. A. The active and identifying ligand binding to protein receptors. Angew. oligomeric state of the minimalistic influenza virus M2 ion Chem., Int. Ed. 2003, 42, 864-890.
channel is a tetramer. Proc. Natl. Acad. Sci. U.S.A. 1997, 94,
5000-5005.

Source: http://www.raino.dk/ak040225.pdf