Investigation of the free energy profiles of amantadine and rimantadine in the AM2 binding pocket

Abstract The purpose of this work was to study the mechanism of drug resistance of M2 channel proteins by analyzing the interactions between the drugs amantadine and rimantadine and M2

DOI 10.1007/s00249-015-1077-y

ORIGINAL PAPER

Investigation of the free energy profiles of amantadine

and rimantadine in the AM2 binding pocket

Hung Van Nguyen 1 · Hieu Thanh Nguyen 1 · Ly Thi Le 1,2

Received: 10 June 2015 / Revised: 20 August 2015 / Accepted: 30 August 2015

© European Biophysical Societies’ Association 2015

membrane for control of proton conductance when the virus penetrates cells (Pinto et al 1992; Pielak and Chou

2010; Lin and Schroeder 2001) Acidification weakens electrostatic interactions between matrix proteins and ribo-nucleoprotein (RNP) complexes, causing the disintegration

of the viral membrane and the movement of the uncoated RNP from the cytosol to the nucleus Because of its crucial involvement in influenza viral pathogenesis, a variety of M2 channel structures have been solved for structure-based drug development (Tran et al 2013) by use of different techniques, for example site-directed infrared dichroism (Kukol et al 1999), UV resonance Raman spectroscopy (Okada et al 2001), electron spin resonance, and solid-state nuclear magnetic resonance (NMR) spectroscopy (Nishimura et al 2002; Kovacs et al 2000; Tian et al 2003; Schnell and Chou 2008)

The M2 channel has four identical subunits and contains 97 residues per monomer (Lamb et al 1985) in three main segments:

an extracellular N-terminal segment (residues 1–23), a trans-membrane (TM) segment (residues 24–46), and an intracellular C-terminal segment (residues 47–97) (Pielak and Chou 2010) The transmembrane (TM) segment is the main region responsible for proton conduction and inhibition of the channel In investiga-tions of the proton conductance of M2 channel proteins for drug development, this segment has been studied in depth (Nishimura

et al 2002; Hu et al 2007; Stouffer et al 2008; Cady and Hong

2008; Cady et al 2010; Acharya et al 2010) Residues 24–46 of the four monomers form a TM helix bundle lined by the polar residues Val27, Ser31, Gly34, His37, Trp41, Asp44, and Arg45 The tetrameric His37–Trp41 cluster is the center of acid activa-tion and proton conductance (Tang et al 2002; Venkataraman

et al 2005; Hu et al 2006) The ionizable His37 is essential for proton selectivity, and acts as a channel sensor (Wang et al 1995), whereas Trp41 is important for unidirectional conductance and acts as a proton gate (Pielak and Chou 2010; Tang et al 2002)

Abstract The purpose of this work was to study the

mechanism of drug resistance of M2 channel proteins by

analyzing the interactions between the drugs amantadine

and rimantadine and M2 channel proteins (including the

wild type and the three mutants V27A, S31N, and G34A)

and the drug binding pathways, by use of a computational

approach Our results showed that multiple drug-binding

sites were present in the M2 channel, and the trajectory of

the drugs through the M2 channel was determined A novel

method was developed to investigate of free energy profiles

of the ligand–protein complexes Our work provides a new

explanation of the large amount of experimental data on

drug efficacy

Keywords AM2 virus · Pathway docking · Amantadine ·

Rimantadine · M2 pocket

Introduction

The influenza A M2 channel protein is a target in

anti-influenza drug design because of its importance in viral

infection (Holsinger and Lamb 1991; Sugrue and Hay

1991; Takeda et al 2002) The tetrameric structure of the

M2 protein forms a pH-dependent channel across the viral

H Van Nguyen and H T Nguyen contributed equally to this

work.

* Ly Thi Le

ly.le@hcmiu.edu.vn

1 Life Science Laboratory, Institute for Computational Science

and Technology, Ho Chi Minh City, Vietnam

2 School of Biotechnology of International University, Vietnam

National University, Ho Chi Minh City, Vietnam

In addition to the Trp41 gate, it has been recently been proposed

that Val27, together with His37, forms a secondary gate (Yi et al

2008; Nguyen and Le 2015)

Several M2 inhibitors have been approved by the FDA

In particular, 1-aminoadamantane hydrochloride, known

as amantadine, is the first efficient drug in influenza

thera-peutics Amantadine indirectly frustrates virus activity by

using a hydrophobic cage to prevent proton conductance by

the ion-channel The other drug approved for treating

influ-enza A, rimantadine (α-methyl-1-adamantane methylamine

hydrochloride), has efficacy comparable with that of

aman-tadine but with a greater risk of adverse side effects

(Ste-phenson and Nicholson 2001; Jefferson et al 2004) The

WHO has, however, restricted the application of

amanta-dine and rimantaamanta-dine for treatment of influenza A because

of the rapidly increasing occurrence of resistant strains

(Tran et al 2011; Le and Leluk 2011; Tran and Le 2014)

The sites of binding of amantadine and rimantadine on

the M2 channel have been a controversial issue for more

than 20 years, because of the location of drug-resistant

mutations at residue 26, 27, 30, 31, 34, and 38 (Hay et al

1985; Wang et al 1993) Interestingly, the side chains of

amino acids 27, 31, and 34 are predicted to face the

chan-nel interior, leading to a hypothesis that the drugs bind to

the inside of the channel (Pielak and Chou 2010) Schnell

and Chou (2008), however, by use of nuclear Overhauser

effect (NOE) experiments, detected four equivalent

bind-ing sites of rimantadine outside the channel Rimantadine

acts as a link between the two adjoining helixes and

indi-rectly keeps the channel gate closed Although recent study

has suggested that binding positions of amantadine and its

derivatives inside the M2 channel are more energetically

favorable, the mechanism of the binding process remains

unclear (Jing et al 2008; Ohigashi et al 2009) For this

rea-son, study of the molecular mechanism of binding of M2

inhibitors to their wild type and mutant targets will provide

important insight into the mechanism of M2 drug resistance

In this research, we used a novel method, called

“path-way docking”, to investigate the interaction between the

M2 channel and amantadine and rimantadine during their

entry into the channel pore of the wild type (WT) and three

drug-resistant mutants (V27A, S31 N and G34A), to gain

insight into the effects of mutations

Materials and methods

Materials

The 3D structure of the M2 channel was taken from Protein

Data Bank (PDB; entry 2L0J, strain A/Udon/307/1972)

It was derived from a complex embedded in DMPC

liposomes (Sharma et al 2010) Amantadine was extracted

from complex 3C9J and rimantadine was constructed on the basis of the amantadine structure by use of Gauss-View 5.0; the geometry of the drugs was then optimized

by use of Gaussian 09 (Table 1) (Hada et al 2004) The V27A, S31N, and G34A mutant models were generated by use of the mutagenesis tool of Visual Molecular Dynamics (VMD) (Humphrey et al 1996) The free energies of bind-ing between the inhibitors and the M2 channels were pre-dicted by AutoDock Vina software (Trott and Olson 2010) Finally, the conformation having the lowest energy and smallest RMSD from docking output was chosen for sub-sequent analysis

Free energy scanning method

The basis of this method is movement of the grid box for docking along the M2 channel from serine 22 to histidine

37 while the box size is modified to ensure the inhibitors bind to the inside of the channel only (Fig 1) The bind-ing energies correspondbind-ing to each step were then assem-bled to produce the energy profile In particular, the center

of mass of serine 22 and histidine 37 were chosen to form

a symmetric axis This axis was then parallelized with the

z-axis by pivoting the protein, the ligands were docked into

it after the center of box had been moved 0.05 Å along the

z-axis from Ser22 to His37, or approximately 25 Å At each

step, the x and y dimensions of the box were determined by

the difference between the respective maximum and

mini-mum of the x and y coordinates of the protein’s atoms; their

z coordinates were the interval of z box size The largest length of the ligands was used to define the z size of the box and only one z box size corresponded to each inhibitor

(Table 1)

Analytical methods

The free energies predicted for the interactions between the inhibitors and the M2 channels were represented as a

function of the z-coordinate of the center of the box when

the docking process was complete On the basis of these

Table 1 Adamantane-based inhibitors The colored branches are

functional groups and their z box sizes are determined by the

maxi-mum length they can reach

Chemical structure

free energy values, geometric features, as functions of z,

were displayed as significant when compounds “moved”

from outside to inside the pore In the docking method,

the ligands were separated into two parts, i.e “root” and

“branches” Specifically, the geometric center of the

ada-mantane cage represents the “root”, and that of each

func-tional group represents a “branch” (Fig 2) Besides the root

and branches, the center of the inhibitor was associated

with a relative binding position in the M2 pore The

ori-entation of each compound was represented the angle, θ, between the rotatable bond and the z-axis, and the angle,

φ , between the rotatable bond and the x-axis To determine their differences, however, the Δθ and Δφ values were a

better choice for characterization of the change of state of the ligands; these were defined as:

Results and discussion Image of the penetration of amantadine and rimantadine: the coexistence of two binding sites

of different energy inside the M2 channel and their motion along the M2 pore

As shown in Figs 3 and 4, the images of the penetration of amantadine and rimantadine were characterized by profiles

of the binding free energies and by Δθ and Δφ Here, the

free energy values, which were indicative of strong inter-action of the inhibitors binding along the pore, and the dif-ferences between the angles indicated the stable structure

of M2 channels This means that the positions at the bound states were stable when the difference between the angles

(1)

∆θi = θi+1 − θi

(2)

∆φi=φi+1−φi

Fig 1 The pathway docking

method moves the grid box

0.05 Å every docking step from

serine 22 to histidine 37 along a

symmetric axis (paralleled with

the z axis) of the M2 pore (a),

the z box size is determined by

the largest size of the

inhibi-tor (b) and is kept constant

throughout the molecular

dock-ing process

x

y

z

Branch

Root

Center 0

o

Fig 2 Simplified rimantadine geometric features represented by

the centers of the root (red), branch (yellow), and whole compound

(green), and the angles between the rotatable bond (blue) and the z

axis, θ, and the x axis, φ

was close to 0 To clarify the relationship between the

posi-tion of the box and the posiposi-tions of amantadine and

rimanta-dine, the z coordinates of the root, the branches, and whole

molecule were represented as a function of the z coordinate

for the center of the box

As is apparent from Figs 3 and 4, the pathway along

which amantadine enters the pore (above His37) from outside

(Ser22) was described as follows The drug was first isolated

by water molecules; it then interacted with residues 22–26

Next, it stayed in front of the Val27 barrier (14–17 Å) until it

crossed the barrier and went deeper into the pore There would

be a region between Val27 and Ser31 (8–10 Å) where the

binding energy gradually declined from the maximum close

to Val27 to a lower value at Ser31 When moving through

this area, the order of the root, branches, and the center of the drug were unchanged (Fig 4a) Here, the result shows that the region between Val27 and Ser31 has stronger binding affinity for amantadine After passing Val27, there were two positions that amantadine had bound to, which ranged from −3 to 0 Å and from 3 to 7 Å; they were separated by the Gly34 residue Between these two positions, the bound state at the Ser31 pore was more stable and the free energy (−6.1 kCal/mol) was higher than at the other (−6.0 kCal/mol) Finally, it stopped in front of the sensor His37 and Trp41 gate These findings con-firmed previous results:

1 the existence of the secondary gate Val27 permeable not only to water molecules but also to amantadine (Yi

et al 2008);

-6

-4

-2

0

2

-120

0

120

∆θ

-6 -3 0 3 6 9 12 15 18

Z (Å) -120

0

120

∆o

B

Val27 Gly34 Ser31

-6

-4

-2

0

-120

0

120

-6 -3 0 3 6 9 12 15 18

Z (Å) -120

0

120

Val27

Gly34

His37

A

Ser31

21 24

21 24

∆E

∆θ

∆o

Fig 3 Free energy profile (top) and differences between the z angles

θ (middle) and x angles φ (bottom) along the channel from the C to

N-terminus (blue left–right arrow) were predicted by pathway

dock-ing for amantadine (a) and rimantadine (b)

-9 -6 -3 0 3 6 9 12 15 18

Z (Å) -9

-6 -3 0 3 6 9 12 15 18 21

CEN Root NH2 CH3

B

Val27 Ser31

Gly34

-9 -6 -3 0 3 6 9 12 15 18

Z (Å) -9

-6 -3 0 3 6 9 12 15 18 21

CEN Root Branch

A

Val27

Ser31 Gly34

21

21

Fig 4 The z position of the center of the roots, branches, and the

whole amantadine (a) and rimantadine (b) molecules were

repre-sented as a function of the z position in alignment with the center of

the docking box

2 the region between Val27 and Ser31 could be a

provi-sional binding position before amantadine reached the

most stable or the real bound state (Sansom and Kerr

1993); and

3 two positions, Ser31–Gly34 (SG) (Cady et al 2009)

and Gly34–His37 (GH) (Gandhi et al 1999), which

were potential binding sites of amantadine

New information from this representation were also

val-uable: although the plots of energies and angles indicated

that amantadine should be at pore SG, a minor difference

in energies suggests that it was also binding to the GH site

In other words, there could be two significant binding sites,

resulting in a high possibility that amantadine could move

from one binding site to the other as a result of thermal

motion

When amantadine penetrated the M2 channel via the

path-way with the lowest energy, its orientation along the path

had rotated the hydrophobic cage and the hydrophilic head

(Fig 4a) In front of residue Val27, its root—the adamantane

cage and its branch—the primary amine group, was horizontal

to the z-axis or at the same z position (12–18 Å) After

cross-ing the barrier, they had turned vertically, the branch pointed

to Ser31 and the root pointed to Val27 (9–12 Å) This

orien-tation was maintained for a short time before the root turned

back and pointed toward the C-terminal (7–9 Å) Next, its

branch turned back and pointed toward Gly34 whereas the

root pointed at Ser31 (0–7 Å) In addition, the primary amine

group bound at that position until the root crossed Gly34 and

pointed to His37 (−2 to 0 Å) The sequential orientation above

was considered reasonable only if the root and the branch of

amantadine exchanged their role as a “hook” The cage would

hook around Val27 (11 Å) and the primary amine group

rotated around it It then held at Ser31 (6.3 Å) and the cage

rotated around it Finally, the primary amine group hooked to

Gly34 (0.7 Å) and the Adamantane cage could stay above it

in the most stable bound state or rotated around it to move to

the second stable state This explains why experimental

struc-tures of the complex of amantadine and the M2 channel have

different positions (Stouffer et al 2008; Yi et al 2008; Cady

et al 2009; Gandhi et al 1999) and orientations (Stouffer et al

2008; Cady et al 2009; Gandhi et al 1999) They were

sim-ply positions in the “walking” process of the inhibitors The

“walking” process of amantadine is shown in Fig 5 These

results explain why the resistant mutations span more than

three helical turns, whereas amantadine has a diameter of only

5 Å and cannot interact with the entire N-terminal half of the

M2 channel (Pielak and Chou 2010)

For rimantadine, the Val27, Ser31, and Gly34 residues

were still used as crucial factors for analysis of the process

of penetration of the inhibitor into the M2 pore (Fig 4b)

Compared with amantadine, the free energy values and

binding sites were very different:

1 Most of the points in the free energy plots for riman-tadine were substantially lower than for amanriman-tadine, and the plots of the angles for rimantadine also fluc-tuated less than for amantadine (Fig 3) This means rimantadine bound more strongly than amantadine, in good agreement with the experimental observation that rimantadine is more efficient than amantadine for treat-ment of influenza A infection (Stephenson and Nichol-son 2001; Jefferson et al 2004)

2 The binding sites of rimantadine were also shrunk and were more specific than for amantadine in the SG and GH regions, except for the area in front of Val27 (Fig 3) After addition of a methyl group, the hydro-phobicity of the functional group including the primary amine group, had substantially increased Rimantadine bound closer to Val27 (a hydrophobic side chain) or

at the Ser22–Val27 (SV) region than amantadine The binding site at SG and GH was narrower and focused

on Ser31 and Gly34 (a polar and hydrophobic side chain) for the same purpose Therefore, the difference between the energy of the bound state for SG or GH was larger for rimantadine than for amantadine In other words, rimantadine bound strongly at S31

resi-Root Branch

Val27

Ser31

Gly34

His37

Fig 5 Illustration of the walking process of amantadine in the M2

pore

due, and, as a result, the probability of transition from

SG to the GH bound state was limited

Effects of mutations on penetration: loss of the

inhibitor trap Val27, the unstable binding place

Asn31, and the third gate Ala34 blocked the channel

to amantadine and rimantadine

The major importance of the Val27, Ser31, and Gly34

residues and their mutants V27A, S31N, and G34A was

strongly confirmed by the mechanism of resistance to

amantadine (Wang et al 1995; Hay et al 1985) In this part

of our work the mechanisms of drug resistance of mutants

of the M2 channel were determined from the energy profile

or from the process of penetration of the inhibitors

On the basis of the energy profile, we found that mutated residues led to different effects of amantadine and rimanta-dine on the M2 channel (Fig 6) It was found that:

1 V27A: the barrier at residue 27 had vanished;

2 S31N: the binding free energies in the SG region were wider and deeper; and

3 G34A: when mutation occurred at the Gly34 resi-due (low barrier) to become Ala34 (steep barrier), the inhibitors needed more energy to cross the barrier, because the steep barrier of the Ala34 residue was higher than that of Val27 residue and even higher than that of the His37 residue

In more detail, the results showed that:

-6

-4

-6

-4

2

-6 -4 2 4 6 0 13 19 15

Z (Å) -6

-4

B

-6

-8

-3

2

3

WT S31N

-6

-8

-3

2

3

V27A

-6 -4 2 4 6 0 13 19 15

Z (Å) -6

-8

-3

2

3

G34A

A

21 24

21 24

Gly34

Val27 Ser31

Gly34

Val27 Ser31

Fig 6 Energy profiles for amantadine (a) and rimantadine (b) in the

WT and three M2 channel variants: S31N, V27A, and G34A

-120 0

120

S31N WT

Z (Å) -120

0 120

A

18 21

120 0 -120

120 0 -120

18 21

Z (Å)

B

Fig 7 Angle profiles of amantadine (a) and rimantadine (b) in the

WT and S31N mutant

1 The unequal energies at the two sides of the

second-ary gate, Val27 (Fig 3), created a gate trap for the

inhibitors It was, therefore, easier for amantadine or

rimantadine to penetrate the pore than to move in the

opposite direction The loss of the secondary gate was

caused by V27A involved vanishing of the trap, and

the inhibitors could not stably bind inside the V27A

mutant pore Furthermore, the absence of the gate

Val27 increased water flux in the pore (Yi et al 2008)

and indirectly reduced inhibition by the Adamantane

cage

2 The deeper energy around the 31st mutant position

proved that both amantadine and rimantadine were

sufficiently bound But, as is apparent from Fig 7

the large width led to unstable binding of amantadine

and rimantadine (0–10 Å) Minor differences between

the profiles for the mutant S31N and wild type were

revealed in our virtual experiments

3 The mutant G34A contained in a third gate, which

pre-vented deeper penetration of the inhibitors

Each mutation has different effect on the inhibitor, but

the mechanism of drug-resistance of the M2 channel was

not apparent from our pathway docking results One

expla-nation could be that the virus replaced the Val27 gate,

which could longer prevent passage of the inhibitors

get-ting through, by a new higher-energy gate at Gly34 and

mutated residues in front of the third gate (S31N is

typi-cal mutant), which have strong affinity for amantadine and

rimantadine

Conclusions

To depict penetration of the M2 channel by amantadine and

rimantadine, the inhibitors were docked 500 times into the

wild type and its mutants (V27A, S31N, and G34A) by use

of pathway docking At every docking step, the grid box

was moved 0.05 Å, in the direction from the N-terminus to

the C-terminus, and the box size was adjusted to ensure the

inhibitors were inside the channel The energy and angle

profiles led to several significant findings

1 There was not only one binding site for amantadine

and rimantadine in the M2 channel but two positions,

at Ser31 (SG region) and Gly34 (GH region) The

inhibitors bound at each position with different

ener-gies or probabilities, and the binding energy in the SG

region was higher than that in the other The binding

energy difference was not large (0.1 kCal/mol) for

amantadine, which led to ease of transition between the

two positions However, it was different for

rimanta-dine with the added methyl group The binding affin-ity along the channel for rimantadine was higher and more concentrated than for amantadine, at Ser31 and Gly34; this is explained by the lower hydrophilicity of the functional groups

2 The hydrophobic cage and the primary amine group of amantadine and rimantadine may act as hooks by use

of which the ligands step inside the channel at Val27, Ser31, and Gly34 This explains the many empiri-cal binding positions and mutations spanning the M2 channel

3 The mechanism of drug resistance of M2 is probably

a result of three features of the mutants First, residues which interact strongly with the primary amine group

of the inhibitors (Ser31 in our work) are replaced by residues with lower affinities Second, water flux is increased by loss of the secondary gate Val27 in the V27A mutation Finally, a replacement gate G34A is created to prevent deeper penetration of the inhibitors and to replace Val27 in control of water flux inside the channel These processes cause unstable binding of inhibitors, thereby reducing inhibition by the hydro-phobic cage of the drugs and preserving the normal activity of the M2 channel

Acknowledgments The work was funded by the Vietnam National

Foundation for Science and Technology Development (NAFOSTED) under grant number 106.01-2012.66 Computing resources and sup-port provided by the Institute for Computational Science and Tech-nology, Ho Chi Minh City, are gratefully acknowledged We would like to thank Professors Thanh Truong and Mai Suan Ly for valuable advice.

References

Acharya R, Carnevale V, Fiorin G, Levine BG, Polishchuk AL, Balan-nik V, Samish I, Lamb RA, Pinto LH, DeGrado WF, Klein ML (2010) Structure and mechanism of proton transport through the transmembrane tetrameric M2 protein bundle of the influenza A virus Proc Natl Acad Sci USA 107:15075–15080

Cady SD, Hong M (2008) Amantadine-induced conformational and dynamical changes of the influenza M2 transmembrane proton channel Proc Natl Acad Sci USA 105:1483–1488

Cady SD, Mishanina TV, Hong M (2009) Structure of amantadine-bound M2 transmembrane peptide of influenza A in lipid bilay-ers from magic-angle-spinning solid-state NMR: the role of Ser31 in amantadine binding J Mol Biol 385:1127–1141 Cady SD, Schmidt-Rohr K, Wang J, Soto CS, DeGrado WF, Hong M (2010) Structure of the amantadine binding site of influenza M2 proton channels in lipid bilayers Nature 463:689–692

Gandhi CS, Shuck K, Lear JD, Dieckmann GR, DeGrado WF, Lamb

RA, Pinto LH (1999) Cu(II) inhibition of the proton transloca-tion machinery of the influenza A virus M2 protein J Biol Chem 274:5474–5482

Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Klene M, Li X, Knox

JE, Hratchian HP, Cross JB, Bakken V, Adamo C, Jaramillo

J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi

R, Pomelli C, Ochterski JM, Ayala PY, Morokuma K, Voth

GA, Salvador P, Dannenberg JJ, Zakrzewski VG, Dapprich S,

Daniels AD, Strain MC, Farkas O, Malick DK, Rabuck AD,

Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul AG,

Clifford S, Cioslowski J, Stefanov BB, Liu G, Liashenko A,

Piskorz P, Komaromi I, Martin RL, Fox DJ, Keith T, Al-Laham

MA, Peng CY, Nanayakkara A, Challacombe M, Gill PMW,

Johnson B, Chen W, Wong MW, Gonzalez C, Pople JA (2004)

Gaussian 03, revision C.02 Gaussian Inc, Wallingford

Hay AJ, Wolstenholme AJ, Skehel JJ, Smith MH (1985) The

molec-ular basis of the specific anti-influenza action of amantadine

EMBO J 4:3021–3024

Holsinger LJ, Lamb RA (1991) Influenza virus M2 integral

mem-brane protein is a homotetramer stabilized by formation of

disulfide bonds J Virology 183:32–43

Hu J, Fu R, Nishimura K, Zhang L, Zhou H-X, Busath DD,

Vijay-vergiya V, Cross TA (2006) Histidines, heart of the hydrogen

ion channel from influenza A virus: toward an understanding

of conductance and proton selectivity Proc Natl Acad Sci USA

103:6865–6870

Hu J, Asbury T, Achutha S, Li C, Bertram R, Quine JR, Fu R, Cross

TA (2007) Backbone structure of the amantadine-blocked

Trans-membrane domain M2 proton channel from influenza A virus

Biophys J 92:4335–4343

Humphrey W, Dalke A, Schulten K (1996) VMD-visual molecular

dynamics Mol Graph 14:33–38

Jefferson T, Deeks JJ, Demicheli V, Rivetti D, Rudin M (2004)

Aman-tadine and rimanAman-tadine for preventing and treating influenza A in

adults Cochrane Database Syst Rev 3:CD001169

Jing X, Ma C, Ohigashi Y, Oliveira FA, Jardetzky TS, Pinto LH, Lamb

RA (2008) Functional studies indicate amantadine binds to the

pore of the influenza A virus M2 proton-selective ion channel

Proc Natl Acad Sci USA 105:10967–10972

Kovacs FA, Denny JK, Song Z, Quine JR, Cross TA (2000) Helix tilt

of the M2 transmembrane peptide from influenza A virus: an

intrinsic property J Mol Biol 295:117–125

Kukol A, Adams PD, Rice LM, Brunger AT, Arkin TI (1999)

Experi-mentally based orientational refinement of membrane protein

models: a structure for the influenza A M2 H + channel J Mol

Biol 286:951–962

Lamb RA, Zebedee SL, Richardson CD (1985) Influenza virus

M2 protein is an integral membrane protein expressed on the

infected-cell surface Cell 40:627–633

Le L, Leluk J (2011) Study on phylogenetic relationship, variability,

and correlated mutations in M2 proteins of influenza virus A

PLoS One 6:e22970 doi: 10.1371/journal.pone.0022970

Lin TI, Schroeder C (2001) Definitive assignment of proton

selectiv-ity and attoampere unitary current to the M2 ion channel protein

of influenza virus J Virology 75:3647–3656

Nguyen H, Le L (2015) Steered molecular dynamics approach for

promising drugs for influenza A virus targeting M2 channel

pro-tein European Biophys J doi: 10.1007/s00249-015-1047-4

Nishimura K, Kim S, Zhang L, Cross TA (2002) The closed state of

a H+ channel helical bundle combining precise orientational

and distance restraints from solid state NMR Biochemistry

41:13170–13177

Ohigashi Y, Ma C, Jing X, Balannick V, Pinto LH, Lamb RA (2009)

An amantadine-sensitive chimeric BM2 ion channel of influenza

B virus has implications for the mechanism of drug inhibition

Proc Natl Acad Sci USA 106:18775–18779

Okada A, Miura T, Takeuchi H (2001) Protonation of histidine and histidine-tryptophan interaction in the activation of the M2 ion channel from influenza a virus Biochemistry 40:6053–6060 Pielak RM, Chou JJ (2010) Flu channel drug resistance: a tale of two sites Protein Cell 1:246–258

Pinto LH, Holsinger LJ, Lamb RA (1992) Influenza virus M2 protein has ion channel activity Cell 69:517–528

Sansom MS, Kerr ID (1993) Influenza virus M2 protein: a molecular modeling study of the ion channel Protein Eng 6:65–74 Schnell JR, Chou JJ (2008) Structure and mechanism of the M2 pro-ton channel of influenza A virus Nature 451:591–595

Sharma M, Yi M, Dong H, Qin H, Peterson E, Busath DD, Zhou H-X, Cross TA (2010) Insight into the mechanism of the influ-enza A proton channel from a structure in a lipid bilayer Science 330:509–512

Stephenson I, Nicholson KG (2001) Influenza: vaccination and treat-ment Eur Respir J 17:1282–1293

Stouffer AL, Acharya R, Salom D, Levine AS, Costanzo LD, Soto

CS, Tereshko V, Nanda V, Stayrook S, DeGrado WF (2008) Structural basis for the function and inhibition of an influenza virus proton channel Nature 451:596–599

Sugrue RJ, Hay AJ (1991) Structural characteristics of the M2 protein

of influenza A viruses: evidence that it forms a tetrameric chan-nel J Virology 180:617–624

Takeda M, Pekosz A, Shuck K, Pinto LH, Lamb RA (2002) Influenza

a virus M2 ion channel activity is essential for efficient replica-tion in tissue culture J Virology 76:1391–1399

Tang Y, Zaitseva F, Lamb RA, Pinto LH (2002) The gate of the influ-enza virus M2 proton channel is formed by a single tryptophan residue J Biol Chem 277:39880–39886

Tian C, Gao PF, Pinto LH, Lamb RA, Cross TA (2003) Initial struc-tural and dynamic characterization of the M2 protein trans-membrane and amphipathic helices in lipid bilayers Protein Sci 12:2597–2605

Tran L, Le L (2014) Recent progress and challenges in the computer-aided design of inhibitors for influenza A M2 channel proteins Med Chem Res doi: 10.1007/s00044-014-0964-6

Tran L, Choi SB, Al-Naiiar BO, Yusuf M, Wahab HA, Le L (2011) Discovery of potential M2 channel inhibitors based on the aman-tadine scaffold via virtual screening and pharmacophore mod-eling Molecules 16:10227–10255

Tran N, Tran L, Le L (2013) Strategy in structure-based drugs design for influenza A virus targeting M2 channel proteins Med Chem Res 12:6078–6088

Trott O, Olson AJ (2010) AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient opti-mization, and multithreading J Comput Chem 31:455–461 Venkataraman P, Lamb RA, Pinto LH (2005) Chemical rescue of his-tidine selectivity filter mutants of the M2 ion channel of influ-enza A virus J Biol Chem 280:21463–21472

Wang C, Takeuchi K, Pinto LH, Lamb RA (1993) Ion channel activity

of influenza A virus M2 protein: characterization of the amanta-dine block J Virology 67:5585–5594

Wang C, Lamb RA, Pinto LH (1995) Activation of the M2 ion chan-nel of influenza virus: a role for the transmembrane domain histi-dine residue Biophys J 69:1363–1371

Yi M, Cross TA, Zhou H-X (2008) A secondary gate as a mechanism for inhibition of the M2 proton channel by amantadine J Phys Chem B 112:7977–7979