- Email: [email protected]
An in-depth analysis of the biological functional studies based on the NMR M2 channel structure of influenza A virus
Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
An in-depth analysis of the biological functional studies based on the NMR M2 channel structure of influenza A virus Ri-Bo Huang a, Qi-Shi Du b,c,d,*, Cheng-Hua Wang c, Kuo-Chen Chou d a
Guangxi Academy of Sciences, 98 Daling Road, Nanning, Guangxi 530004, China College of Life Science and Technique, Guangxi University, 100 University Road, Nanning, Guangxi 530004, China c College of Chemistry and Life Science, Tianjin Normal University, Tianjin, 300074, China d Gordon Life Science Institute, San Diego, CA 92130, USA b
a r t i c l e
i n f o
Article history: Received 11 October 2008 Available online 6 November 2008
Keywords: Membrane protein channel Virus Drug-resistance Rimantadine Molecular wedge Allosteric inhibition mechanism Long-range interaction
a b s t r a c t The long-sought three-dimensional structure of the M2 proton channel of influenza A virus was successfully determined recently by the high-resolution NMR [J.R. Schnell, J.J. Chou, Structure and mechanism of the M2 proton channel of influenza A virus, Nature 451 (2008) 591–595]. Such a milestone work has provided a solid structural basis for studying drug-resistance problems. However, the action mechanism revealed from the NMR structure is completely different from the traditional view and hence prone to be misinterpreted as ‘‘conflicting” with some previous biological functional studies. To clarify this kind of confusion, an in-depth analysis was performed for these functional studies, particularly for the mutations D44N, D44A and N44D on position 44, and the mutations on positions 27–38. The analyzed results have provided not only compelling evidences to further validate the NMR structure but also very useful clues for dealing with the drug-resistance problems and developing new effective drugs against H5N1 avian influenza virus, an impending threat to human beings. Ó 2008 Elsevier Inc. All rights reserved.
Influenza A viruses have the ability to undergo changes by the mechanisms of antigenic drift and shift, resulting in new virus strains, which may be extremely toxic and drug resistant [1,2]. Given the fact that influenza shifts may occur every 20–30 years, the danger of future influenza A pandemics highlights the need to develop more-effective drugs. The threat of an impending influenza pandemic, possibly through mutations of the present deadly avian strain H5N1, has triggered a global effort to develop more-effective antivirus drugs. However, during the past several decades the great efforts in developing anti-influenza virus drugs have almost been futile due to the rapid mutations of the influenza virus, which causes the persistent resistance to the existing drugs. It is known now that the replication of influenza A viruses needs a pH-gated channel, the so-called M2 proton channel [3,4]. At low pH the channel is opened allowing protons transporting through the channel; at high pH the channel is closed and the proton conductivity is stopped [4–6]. Since the M2 proton channel was discovered, it has been the main target for finding drugs against influenza A virus. The adamantane-based drugs, amantadine and rimantadine [7], that target the M2 channel, have been used as first-choice antiviral drugs
* Corresponding author. Address: College of Life Science and Technique, Guangxi University, 100 University Road, Nanning, Guangxi 530004, China. Fax: +86 771 323 8107. E-mail address: [email protected] (Q.-S. Du) 0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2008.10.148
against community outbreaks of influenza A viruses for many years, but rates of viruses resistant to the adamantane-based drugs have been increasing globally. It was reported that the resistance to these drugs in humans, birds and pigs has reached more than 90% [1]. Recently, the long-sought 3D (dimensional) structure of the M2 protein channel has been successfully determined by the high-resolution NMR spectroscopy [8]. It has provided the detailed structural information and mechanism implication that may lead to the solution of drug-resistance problem of influenza A virus and help design more-effective medications [7]. Interestingly, the drug-binding site revealed from the NMR structure [8] is completely different from the traditional view. According to the traditional model, the inhibitor amantadine is plugged into the pore of the channel by binding at the site composed of Val27, Ala30, Ser31, and Gly34 [4]; binding of amantadine physically occludes the pore of M2 channel, preventing proton relay through water molecules. In contrast to that, it has been unambiguously observed from the high-resolution NMR structure [8] that the inhibitor rimantadine binds to an external, lipid-facing pocket near the Trp41 gate and forms 1–2 hydrogen bonds with Asp44. The drug thus bound functions as a molecular wedge that stabilizes the closed conformation of the channel gate, thereby raising the energy barrier for the closed channel to open in acidic condition [8]. Accordingly, the traditional model is actually the pore-blocking model while the molecular wedge model inferred
1244
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
from the NMR structure [8] is an allosteric one. The simple poreblocking model was speculated based on some functional mutation studies that worked for some of membrane channels [9] but is conflicting with the direct NMR observation of the M2 channel [8]. The binding mechanism is the key for the structure-based rational drug design. However, the novel molecular wedge allosteric mechanism revealed from the NMR structure may be difficult to understand due to its conceptually complete difference with the traditional model. Particularly, it is prone to be misinterpreted by those who are insisting on the traditional point of view (see, e.g. [10]). The present study was initiated in an attempt to revisit the previous functional studies and analyze their inferences based on the high-resolution NMR structure [8] in hope to clarify the confusions in this area.
A closed-up view of the interaction between residue Trp41 and Asp44 in the two adjacent chains of the M2 channel is given in Fig. 2, from which we can see that the Asp44 carboxyl forms an inter-chain hydrogen bond (2.09 Å) with the indole amine of Trp41 of the adjacent helix. The inter-chain hydrogen bond will tighten the proton gate to its closed state and maintains the four helices of the channel in a compact 4-helix bundle. But in the lower pH environment, the pH sensor His37 and the indole amine of Trp41 are protonated, so as to weaken the hydrogen bond between Trp41 and Asp44, making it easily broken by the repulsive interaction between the positively charged His37 residues. That is why the proton gate in a lower pH environment can be opened (switched on). During the open–close gating process, the tetrad Asp44 plays the role as a lock for the M2 channel. Drug-inhibition mechanism
Methods The Auto Dock 3.0 program [11] with the MMFF94 force field and force field charges [12] was used to perform the molecular docking of a rimantadine inhibitor to the tetrameric M2 channel. The atomic coordinates of the M2 channel (pdb code: 2RLF) determined by NMR [8] were taken from RCSB protein data bank at www.rcsb.org/pdb. The docking was conducted with the search range in a pocket consisting of polar residues Trp41, Asp44, Arg45 and hydrophobic residues Leu40, Ile42, and Leu43, which are around the ligand-binding site [8]. During the molecular docking operation, the ligand rimantadine was limited in a cubic box with side length of 20 Å. For the results thus generated, we recorded 25 most favorable positions, from which the one with the lowest total interaction energy was used as the complex structure of the M2 channel with rimantadine for further studies. The pKa value of Asp44 and Asn44 in the M2 channel with and without rimantadine ligand was computed using an improved PROPKA program [13,14]. The fundamental equation of the PROPKA program can be formulated as
To further understand the drug-binding site and its inhibition mechanism, let us analyze the complex structure (Fig. 2) obtained by the molecular docking operation as described in the Method section. It can be seen from the figure that the amino group of rimantadine forms two hydrogen bonds (2.75 and 1.81 Å) with the carboxyl of Asp44. Li et al. [13,14] have proved that the hydrogen bonding interaction is the prime determinant factor to the pKa value of Asp (carboxyl residue) in protein, and that its pKa value will shift downward by 0.60 by forming one hydrogen bond with it. Accordingly, two hydrogen bonds thus formed with rimantadine will lower its pKa by 1.20 (cf. Eq. (1)). As a consequence, the pKa value of Asp44 will be lowered down from the original model value of 3.80–2.60. The downward shift of the pKa value of the Asp44 will further strengthen its hydrogen bonding with Trp41 of the adjacent helix, thus tightening the lock of the channel gate in the acidic (low pH) condition. That is why after binding with amantadine/ rimantadine, the M2 proton gate channel is no longer opened even in the low pH environment. Mutations on position 44
pK a ¼ pK model þ DpK HB þ DpK charge þ DpK solv
ð1Þ
where pKmodel is a model value of an amino acid, DpKHB, DpKcharge, and DpKsolv are the corrections to the pKa value in a protein from the interactions of hydrogen bonds, atomic partial charges, and solvation effect, respectively. Among them the hydrogen bond interaction makes the main contribution [13,14].
Results and discussions Gating mechanism It is widely accepted that in the M2 channel the His37 tetrad serves as the pH sensor and the Trp41 tetrad as the proton gate. Being an ionic channel, the M2 channel must have a clear gating mechanism, i.e., it can be opened at lower pH value allowing the proton flow through the channel while closed at higher pH value blocking the proton flow. With the high-resolution experimental M2 channel structure available [8], we are able to get a clear picture about the proton-gating mechanism as illustrated below. Shown in Fig. 1 are the positions of several key residues in the M2 channel: the pH sensor His37, the proton gate Trp41, and the channel lock Asp44. These positions were determined with pH 7.5 [8] and hence correspond to a closed conformation. For such a closed M2 channel, the pore diameter of the pH sensor is 1.7 Å and that of the proton gate 1.4 Å [8]. Both the two distances are within the van der Waals interaction range and therefore the His37 tetrad and Trp41 tetrad can effectively block the proton flow (H2O.H+).
As shown above, the residue Asp44 plays an important role during the channel gating process. However, it was reported recently by Jing et al. [10] that the D44A mutated M2 channel could still be inhibited by the drug amantadine. In the same paper, these authors cited the work by Betakova et al. [6] that the M2 channel possesses a natural mutation D44N. Based on these, Jing et al. [10] claimed that the external drug-binding site as observed in the NMR structure [8] was not functionally important. Below, let us see what should be concluded according to the aforementioned new channel gating and drug-inhibition mechanism. Shown in Fig. 3 are three amino acid sequences of M2 channel: the first one has the same sequence as that of the NMR structure (PDB code 2RLF) [8], the second one corresponds to the Rostock strain (R-M2), and the third to the Weybridge strain (W-M2). The underline indicates the transmembrane part of the M2 channel. R-M2 and W-M2 are two wild type M2 proteins and their difference is in the positions 27, 38, and 44. In the R-M2 channel the position 44 is occupied by Asn44, instead of Asp44. However, it is incorrect to use the existence of the natural mutation D44N in R-M2 to invalidate the Asp44 being the drug-binding site, as claimed by Jing et al. [10]. Conversely, it should be used to further strengthen the allosteric binding mechanism revealed by the NMR structure [8], as elucidated below. In the natural mutation D44N of the M2 channel for the Rostock strain (R-M2) the Asn44 possesses the similar structure of Asp44 and hence can play the same roles as Asp44 does; this can be clearly seen in Fig. 3B and C. The residue Asn44 forms a hydrogen bond with Trp41 and two hydrogen bonds with the rimantadine
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
1245
Fig. 1. (A) The tetrameric M2 channel in complex with rimantadine determined by NMR [8]. In the closed state, the four tightly packed transmembrane helices form a lefthanded twisting bundle [19] and define a narrow channel. The ligand rimantadines bind outside the channel in the pocket composed of the residues Asp44, Trp41, and Arg45, which form the polar patch, as well as the hydrophobic wall composed of Leu40 and Leu43 of one TM helix and Ile42 of the adjacent TM helix. (B) The pore diameters at the positions of the key residues in the M2 channel: the tetrad His37 serves as the pH sensor, the Trp41 as the proton gate, and the Asp44 as the channel lock that switches the channel on or off. In the closed conformation the pore diameter of pH sensor is 1.7 Å and that of proton gate is 1.4 Å; both are in the van der Waals interaction range and hence able to block the proton conductance.
Fig. 2. A close-up view from the bottom of the channel: the interaction between rimantadine and the M2 proton channel. The rimantadine binds to Asp44 of Chain 1 through two hydrogen bonds (2.75 and 1.81 Å). In the closed conformation of the M2 channel the residue Asp44 of Chain 1 forms a hydrogen bond (2.09 Å) with the Trp41 of Chain 2, tightening the proton gate Trp41 to its closed state. The hydrogen bond between Trp41 and Asp44 holds the two adjacent helices, making the tetrameric channel in a compact conformation.
(Fig. 3C), just exactly like the Asp44 does (Fig. 3B). However, Asn44 has higher pKa value than that of Asp44 in the M2 channel, because it has an acyl group (–CONH2) instead of carboxyl group (–COOH). According to our calculation using the improved PROPKA program and Eq. (1) [13,14], the pKa value of Asn44 in the ligand-free conformation of the M2 channel is 7.27 and that in the rimantadine-binding conformation is 6.07. Both are much higher than the corresponding pKa values, 3.80 and 2.60, of Asp44 in the M2 channel. That is why the R-M2 channel can keep open in higher pH environment and is more difficult to be locked by the inhibitor amantadine or rimantadine. That is also why the R-M2 channel possesses a sevenfold greater proton conductance and drug-resis-
tance. This is completely consistent with the report by Betakova et al. [6] as saying: ‘‘substitution of asparagines for aspartic acid 44 in the M2 PR8-M2 (D44N) caused an increase in activity (resistant to amantadine) to a level comparable to that of the wt Rostock protein, confirming the more general influence of this particular substitution at residue 44 on M2 activity”. On the other hand, the substitution of Asn with Asp (N44D) at potion 44 of the RM2 channel, however, caused a significant decrease in activity (approximately to the levels found for W-M2), indicating that the change in this residue alone could account for the difference in the pH-modulating activities of the two wt channel proteins [6]. Accordingly, our interpretation based on the NMR structure [8] is fully consistent with the results as observed in [6]. Although Jing et al. [10] cited the paper of Betakova et al. [6] to oppose the NMR structure and its action rimantadine, actually the conclusion drawn from their own D44A mutation study is completely conflicting with the results from the mutation studies D44N and N44D conducted by Betakova et al. [6], as can be illustrated as follows. In their paper [10], Jing et al. claimed that after the mutation D44A, the M2 channel still worked normally and the drug amantadine was still active, and hence the Asp44 is not functionally important. However, it is clearly stated by Betakova et al. [6] that the position 44 possesses ‘‘the more general influence” and the residue on the position 44 (Asp44 or Asn44) ‘‘could account for the difference in the pH-modulating activities”. Therefore Jing et al. [10] misinterpreted the report of [6]. Besides, the reliability of their results [10] is highly questionable because they used an unnatural mutation, D44A, to conduct the mutational experiments that would completely change the polar property of the residue, significantly impairing the channel activity. Mutations on positions from 27 to 38 Before the high-resolution NMR structure is available, the binding site of amantadine/rimantadine with the M2 channel was speculated according to the biological functional studies [15]. This can be seen from the statement by Hay et al. [15] as saying that ‘‘influenza A viruses that mutate and become drug resistant contain mutations in the A/M2 TM domain and naturally arising mutations mostly occur at TM domain residues Val27, Ala30, Ser31, and
1246
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
Fig. 3. (A) The amino acid sequences of the M2 channel: NMR structure 2RLF, Rostock strain R-M2, and Weybridge strain W-M2. The underline indicates the transmembrane part of the M2 channel. Difference between R-M2 and W-M2 is at the positions 27, 38, and 44. (B) In the NMR structure (2RLF), there are two hydrogen bonds between Asp44 and rimantadine and one hydrogen bond between Asp44 and Trp41 of the adjacent helix. (C) After D44N mutation for the M2 channel followed by energy optimization, the similar hydrogen bonding pattern still exists and hence Asn44 can play the same roles as Asp44 does.
Gly34”. Therefore some authors concluded that amantadine was positioned around Ser31, in line with the central location of this residue in the constellation of amantadine-resistant mutations (on Val27, Ala30, Ser31, and Gly34) [15,16]. However, using such a simplistic logic to identify the binding site according to the information obtained solely from amino acid mutations is not always correct because no ‘‘long-range” interaction effects [17] whatsoever have been taken into account. As is well known, the relationship between the function of a protein and its constituent amino acids is extremely complicated. It is impossible to really find their correlation without a reliable 3D structure. The high-resolution NMR structure of the M2 channel can provide a solid basis in this regard. As shown in Fig. 2, the M2 channel is locked by the hydrogen bond between Asp44 of one helix and Trp41of the adjacent helix. Such a lock is triggered by binding the amantadine/rimantadine ligand to Asp44 through two hydrogen bonds outside the pore. The strengths of these three hydrogen bonds all depend on the orientations of the two residues. Therefore, their orientations may affect the lock of the M2 channel. Since the M2 channel is composed of four transmembrane helices, the orientation of Asp44 in one helix and that of Trp41 in the adjacent helix may be affected by the residue mutation on any of the positions from 27 to 38, and hence affect the biological function of the M2 channel as well. But it should be pointed out that this kind of functional change is caused by the impact on the primary binding site via the long-range interaction.
helices, hence affecting the strength of hydrogen bonds between Asp44 and Trp41 and between Asp44 and amantadine/rimantadine, eventually resulting in drug-resistance problem to adamantane-based drugs. Confused by the long-range effects, it was prone to make mistakes in identifying the real drug-binding site of the M2 channel and provide misleading information for the real target in drug design. With the high-resolution NMR structure of the M2 channel available [8], it is now clear that the real binding site for amantadine/rimantadine is at the site involving Trp41 and Asp44.
The drug-resistance of M2 channel
Conclusion
The aforementioned long-range interaction effects can be also used to elucidate why the natural and artificial drug-resistant mutations in the M2 channel occurred in a large range from positions 27–38 [15], and why the drug-resistant influenza A virus in humans, birds and pigs has reached more than 90% during last three decades [1]. According to the long-range interaction [17], all mutations before position 41, especially on positions 27–38, could affect the orientations of Trp41 and Asp44 in two adjacent
The 3D M2 channel structure determined recently by the highresolution NMR technique [8] has provided a solid basis for understanding the proton-gating and drug-inhibition mechanism from a conception quite different from the traditional view. However, the new conception is not easy to be accepted or prone to be misinterpreted. The analyses and discussions presented in this article can help clarify the confusions in these regards. Particularly, it is crystal clear according to the above analysis that the functional studies
Coordinate number ratio It was reported that one amantadine molecule could block the M2 channel [18], which led to a conclusion that the coordinate number ratio (Hill coefficient) between the M2 receptor and amantadine ligand was 1:1. According to the NMR structure for the M2 channel in complex with rimantadine, the ligand binds at four equivalent sites near the gate Trp41 on the lipid-facing side of the channel and stabilizes the closed conformation of the pore [8]. However, it does not mean that the coordinate number ratio between the receptor and the ligand must be 1:4. As we can see from the above analysis and Fig. 2, merely one rimantadine molecule can also block an M2 channel effectively. Therefore, the coordinate number ratio of the M2 channel to amantadine/rimantadine could be anywhere between 1:1 and 1:4.
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
conducted by Betakova et al. [6] are completely consistent with the unique role of position 44 and the importance of Asp44 (or Asn44) in the M2 channel, as revealed by the NMR structure [8]. Moreover, it has also become clear now that the mutations on positions from 27 to 38 may affect the orientations of Asp44 and Trp41, and hence lead to drug-resistance as well. However, this belongs to a kind of indirect or ‘‘long-range interaction” effect. It is Asp44 and Trp41 that form the primary binding site of rimantadine as observed from the NMR structure. The current analyses may stimulate and encourage new strategies for finding effective drugs against influenza A virus. Acknowledgment This work is financially supported by the National High-tech Research and Development Program (‘863’) of China under the project 2006AA020103. References [1] V.M. Deyde, X. Xu, R.A. Bright, M. Shaw, C.B. Smith, Y. Zhang, Y. Shu, L.V. Gubareva, N.J. Cox, A.I. Klimov, Surveillance of resistance to adamantanes among influenza A(H3N2) and A(H1N1) viruses isolated worldwide, J. Infect. Dis. 196 (2007) 249–257. [2] B.J. Smith, J.L. McKimm-Breshkin, M. McDonald, R.T. Fernley, J.N. Varghese, P.M. Colman, Structural studies of the resistance of influenza virus neuramindase to inhibitors, J. Med. Chem. 45 (2002) 2207–2212. [3] K. Martin, A. Helenius, Nuclear transport of influenza virus ribonucleoproteins: the viral matrix protein (M1) promotes export and inhibits import, Cell 67 (1991) 117–130. [4] R.J. Sugrue, A.J. Hay, Structural characteristics of the M2 protein of influenza A viruses: evidence that it forms a tetrameric channel, Virology 180 (1991) 617–624.
1247
[5] R.A. Kumpf, D.A. Dougherty, A mechanism for ion selectivity in potassium channels: computational studies of cation-pi interactions, Science 261 (1993) 1708–1710. [6] T. Betakova, F. Ciampor, A.J. Hay, Influence of residue 44 on the activity of the M2 proton channel of influenza A virus, J. Gen. Virol. 86 (2005) 181–184. [7] S. Borman, Flu virus proton channel analyzed: structures of key surface protein suggest different drug mechanisms, Chem. Eng. News 86 (2008) 53–54. [8] J.R. Schnell, J.J. Chou, Structure and mechanism of the M2 proton channel of influenza A virus, Nature 451 (2008) 591–595. [9] L.J. Holsinger, D. Nichani, L.H. Pinto, R.A. Lamb, Influenza A virus M2 ion channel prote: a structure-function analysis, J. Virol. 68 (1994) 1551–1563. [10] X. Jing, C. Ma, Y. Ohigashi, F.A. Oliveira, T.S. Jardetzky, L.H. Pinto, R.A. Lamb, Functional studies indicate amantadine binds to the pore of the influenza A virus M2 proton-selective ion channel, Proc. Natl. Acad. Sci. USA 105 (2008) 10967–10972. [11] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson, Automated docking using a Lamarckian Genetic Algorithm and empirical binding free energy function, J. Comput. Chem. 19 (1998) 1639–1662. [12] W.D. Cornell, P. Cieplak, C.I. Bayly, I.R. Gould, K.M. Merz Jr., D.M. Ferguson, D.C. Spellmeyer, T. Fox, J.W. Caldwell, P.A. Kollman, A second generation force field for the simulation of proteins and nucleic acids, J. Am. Chem. Soc. 177 (1995) 5179–5197. [13] H. Li, A.D. Robertson, J.H. Jensen, The determinants of carboxyl pKa values in turkey ovomucoid third domain, Proteins 55 (2004) 689–704. [14] H. Li, A.D. Robertson, J.H. Jensen, Very fast empirical prediction and rationalization of protein pKa values, Proteins 61 (2005) 704–721. [15] A.J. Hay, A.J. Wolstenholme, J.J. Skehel, M.H. Smith, The molecular basis of the specific anti-influenza action of amantadine, EMBO J. 4 (1985) 3021–3024. [16] M. Yi, T.A. Cross, H.X. Zhou, A secondary gate as a mechanism for inhibition of the M2 proton channel by amantadine, J. Phys. Chem. B 112 (2008) 7977– 7979. [17] K.C. Chou, Review: prediction of tight turns and their types in proteins, Anal. Biochem. 286 (2000) 1–16. [18] C. Wang, K. Takeuchi, L.H. Pinto, R.A. Lamb, Ion channel activity of influenza A virus M2 protein: characterization of the amantadine block, J. Virol. 67 (1993) 5585–5594. [19] K.C. Chou, G.M. Maggiora, G. Nemethy, H.A. Scheraga, Energetics of the structure of the four-alpha-helix bundle in proteins, Proc. Natl. Acad. Sci. USA 85 (1988) 4295–4299.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
An in-depth analysis of the biological functional studies based on the NMR M2 channel structure of influenza A virus Ri-Bo Huang a, Qi-Shi Du b,c,d,*, Cheng-Hua Wang c, Kuo-Chen Chou d a
Guangxi Academy of Sciences, 98 Daling Road, Nanning, Guangxi 530004, China College of Life Science and Technique, Guangxi University, 100 University Road, Nanning, Guangxi 530004, China c College of Chemistry and Life Science, Tianjin Normal University, Tianjin, 300074, China d Gordon Life Science Institute, San Diego, CA 92130, USA b
a r t i c l e
i n f o
Article history: Received 11 October 2008 Available online 6 November 2008
Keywords: Membrane protein channel Virus Drug-resistance Rimantadine Molecular wedge Allosteric inhibition mechanism Long-range interaction
a b s t r a c t The long-sought three-dimensional structure of the M2 proton channel of influenza A virus was successfully determined recently by the high-resolution NMR [J.R. Schnell, J.J. Chou, Structure and mechanism of the M2 proton channel of influenza A virus, Nature 451 (2008) 591–595]. Such a milestone work has provided a solid structural basis for studying drug-resistance problems. However, the action mechanism revealed from the NMR structure is completely different from the traditional view and hence prone to be misinterpreted as ‘‘conflicting” with some previous biological functional studies. To clarify this kind of confusion, an in-depth analysis was performed for these functional studies, particularly for the mutations D44N, D44A and N44D on position 44, and the mutations on positions 27–38. The analyzed results have provided not only compelling evidences to further validate the NMR structure but also very useful clues for dealing with the drug-resistance problems and developing new effective drugs against H5N1 avian influenza virus, an impending threat to human beings. Ó 2008 Elsevier Inc. All rights reserved.
Influenza A viruses have the ability to undergo changes by the mechanisms of antigenic drift and shift, resulting in new virus strains, which may be extremely toxic and drug resistant [1,2]. Given the fact that influenza shifts may occur every 20–30 years, the danger of future influenza A pandemics highlights the need to develop more-effective drugs. The threat of an impending influenza pandemic, possibly through mutations of the present deadly avian strain H5N1, has triggered a global effort to develop more-effective antivirus drugs. However, during the past several decades the great efforts in developing anti-influenza virus drugs have almost been futile due to the rapid mutations of the influenza virus, which causes the persistent resistance to the existing drugs. It is known now that the replication of influenza A viruses needs a pH-gated channel, the so-called M2 proton channel [3,4]. At low pH the channel is opened allowing protons transporting through the channel; at high pH the channel is closed and the proton conductivity is stopped [4–6]. Since the M2 proton channel was discovered, it has been the main target for finding drugs against influenza A virus. The adamantane-based drugs, amantadine and rimantadine [7], that target the M2 channel, have been used as first-choice antiviral drugs
* Corresponding author. Address: College of Life Science and Technique, Guangxi University, 100 University Road, Nanning, Guangxi 530004, China. Fax: +86 771 323 8107. E-mail address: [email protected] (Q.-S. Du) 0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2008.10.148
against community outbreaks of influenza A viruses for many years, but rates of viruses resistant to the adamantane-based drugs have been increasing globally. It was reported that the resistance to these drugs in humans, birds and pigs has reached more than 90% [1]. Recently, the long-sought 3D (dimensional) structure of the M2 protein channel has been successfully determined by the high-resolution NMR spectroscopy [8]. It has provided the detailed structural information and mechanism implication that may lead to the solution of drug-resistance problem of influenza A virus and help design more-effective medications [7]. Interestingly, the drug-binding site revealed from the NMR structure [8] is completely different from the traditional view. According to the traditional model, the inhibitor amantadine is plugged into the pore of the channel by binding at the site composed of Val27, Ala30, Ser31, and Gly34 [4]; binding of amantadine physically occludes the pore of M2 channel, preventing proton relay through water molecules. In contrast to that, it has been unambiguously observed from the high-resolution NMR structure [8] that the inhibitor rimantadine binds to an external, lipid-facing pocket near the Trp41 gate and forms 1–2 hydrogen bonds with Asp44. The drug thus bound functions as a molecular wedge that stabilizes the closed conformation of the channel gate, thereby raising the energy barrier for the closed channel to open in acidic condition [8]. Accordingly, the traditional model is actually the pore-blocking model while the molecular wedge model inferred
1244
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
from the NMR structure [8] is an allosteric one. The simple poreblocking model was speculated based on some functional mutation studies that worked for some of membrane channels [9] but is conflicting with the direct NMR observation of the M2 channel [8]. The binding mechanism is the key for the structure-based rational drug design. However, the novel molecular wedge allosteric mechanism revealed from the NMR structure may be difficult to understand due to its conceptually complete difference with the traditional model. Particularly, it is prone to be misinterpreted by those who are insisting on the traditional point of view (see, e.g. [10]). The present study was initiated in an attempt to revisit the previous functional studies and analyze their inferences based on the high-resolution NMR structure [8] in hope to clarify the confusions in this area.
A closed-up view of the interaction between residue Trp41 and Asp44 in the two adjacent chains of the M2 channel is given in Fig. 2, from which we can see that the Asp44 carboxyl forms an inter-chain hydrogen bond (2.09 Å) with the indole amine of Trp41 of the adjacent helix. The inter-chain hydrogen bond will tighten the proton gate to its closed state and maintains the four helices of the channel in a compact 4-helix bundle. But in the lower pH environment, the pH sensor His37 and the indole amine of Trp41 are protonated, so as to weaken the hydrogen bond between Trp41 and Asp44, making it easily broken by the repulsive interaction between the positively charged His37 residues. That is why the proton gate in a lower pH environment can be opened (switched on). During the open–close gating process, the tetrad Asp44 plays the role as a lock for the M2 channel. Drug-inhibition mechanism
Methods The Auto Dock 3.0 program [11] with the MMFF94 force field and force field charges [12] was used to perform the molecular docking of a rimantadine inhibitor to the tetrameric M2 channel. The atomic coordinates of the M2 channel (pdb code: 2RLF) determined by NMR [8] were taken from RCSB protein data bank at www.rcsb.org/pdb. The docking was conducted with the search range in a pocket consisting of polar residues Trp41, Asp44, Arg45 and hydrophobic residues Leu40, Ile42, and Leu43, which are around the ligand-binding site [8]. During the molecular docking operation, the ligand rimantadine was limited in a cubic box with side length of 20 Å. For the results thus generated, we recorded 25 most favorable positions, from which the one with the lowest total interaction energy was used as the complex structure of the M2 channel with rimantadine for further studies. The pKa value of Asp44 and Asn44 in the M2 channel with and without rimantadine ligand was computed using an improved PROPKA program [13,14]. The fundamental equation of the PROPKA program can be formulated as
To further understand the drug-binding site and its inhibition mechanism, let us analyze the complex structure (Fig. 2) obtained by the molecular docking operation as described in the Method section. It can be seen from the figure that the amino group of rimantadine forms two hydrogen bonds (2.75 and 1.81 Å) with the carboxyl of Asp44. Li et al. [13,14] have proved that the hydrogen bonding interaction is the prime determinant factor to the pKa value of Asp (carboxyl residue) in protein, and that its pKa value will shift downward by 0.60 by forming one hydrogen bond with it. Accordingly, two hydrogen bonds thus formed with rimantadine will lower its pKa by 1.20 (cf. Eq. (1)). As a consequence, the pKa value of Asp44 will be lowered down from the original model value of 3.80–2.60. The downward shift of the pKa value of the Asp44 will further strengthen its hydrogen bonding with Trp41 of the adjacent helix, thus tightening the lock of the channel gate in the acidic (low pH) condition. That is why after binding with amantadine/ rimantadine, the M2 proton gate channel is no longer opened even in the low pH environment. Mutations on position 44
pK a ¼ pK model þ DpK HB þ DpK charge þ DpK solv
ð1Þ
where pKmodel is a model value of an amino acid, DpKHB, DpKcharge, and DpKsolv are the corrections to the pKa value in a protein from the interactions of hydrogen bonds, atomic partial charges, and solvation effect, respectively. Among them the hydrogen bond interaction makes the main contribution [13,14].
Results and discussions Gating mechanism It is widely accepted that in the M2 channel the His37 tetrad serves as the pH sensor and the Trp41 tetrad as the proton gate. Being an ionic channel, the M2 channel must have a clear gating mechanism, i.e., it can be opened at lower pH value allowing the proton flow through the channel while closed at higher pH value blocking the proton flow. With the high-resolution experimental M2 channel structure available [8], we are able to get a clear picture about the proton-gating mechanism as illustrated below. Shown in Fig. 1 are the positions of several key residues in the M2 channel: the pH sensor His37, the proton gate Trp41, and the channel lock Asp44. These positions were determined with pH 7.5 [8] and hence correspond to a closed conformation. For such a closed M2 channel, the pore diameter of the pH sensor is 1.7 Å and that of the proton gate 1.4 Å [8]. Both the two distances are within the van der Waals interaction range and therefore the His37 tetrad and Trp41 tetrad can effectively block the proton flow (H2O.H+).
As shown above, the residue Asp44 plays an important role during the channel gating process. However, it was reported recently by Jing et al. [10] that the D44A mutated M2 channel could still be inhibited by the drug amantadine. In the same paper, these authors cited the work by Betakova et al. [6] that the M2 channel possesses a natural mutation D44N. Based on these, Jing et al. [10] claimed that the external drug-binding site as observed in the NMR structure [8] was not functionally important. Below, let us see what should be concluded according to the aforementioned new channel gating and drug-inhibition mechanism. Shown in Fig. 3 are three amino acid sequences of M2 channel: the first one has the same sequence as that of the NMR structure (PDB code 2RLF) [8], the second one corresponds to the Rostock strain (R-M2), and the third to the Weybridge strain (W-M2). The underline indicates the transmembrane part of the M2 channel. R-M2 and W-M2 are two wild type M2 proteins and their difference is in the positions 27, 38, and 44. In the R-M2 channel the position 44 is occupied by Asn44, instead of Asp44. However, it is incorrect to use the existence of the natural mutation D44N in R-M2 to invalidate the Asp44 being the drug-binding site, as claimed by Jing et al. [10]. Conversely, it should be used to further strengthen the allosteric binding mechanism revealed by the NMR structure [8], as elucidated below. In the natural mutation D44N of the M2 channel for the Rostock strain (R-M2) the Asn44 possesses the similar structure of Asp44 and hence can play the same roles as Asp44 does; this can be clearly seen in Fig. 3B and C. The residue Asn44 forms a hydrogen bond with Trp41 and two hydrogen bonds with the rimantadine
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
1245
Fig. 1. (A) The tetrameric M2 channel in complex with rimantadine determined by NMR [8]. In the closed state, the four tightly packed transmembrane helices form a lefthanded twisting bundle [19] and define a narrow channel. The ligand rimantadines bind outside the channel in the pocket composed of the residues Asp44, Trp41, and Arg45, which form the polar patch, as well as the hydrophobic wall composed of Leu40 and Leu43 of one TM helix and Ile42 of the adjacent TM helix. (B) The pore diameters at the positions of the key residues in the M2 channel: the tetrad His37 serves as the pH sensor, the Trp41 as the proton gate, and the Asp44 as the channel lock that switches the channel on or off. In the closed conformation the pore diameter of pH sensor is 1.7 Å and that of proton gate is 1.4 Å; both are in the van der Waals interaction range and hence able to block the proton conductance.
Fig. 2. A close-up view from the bottom of the channel: the interaction between rimantadine and the M2 proton channel. The rimantadine binds to Asp44 of Chain 1 through two hydrogen bonds (2.75 and 1.81 Å). In the closed conformation of the M2 channel the residue Asp44 of Chain 1 forms a hydrogen bond (2.09 Å) with the Trp41 of Chain 2, tightening the proton gate Trp41 to its closed state. The hydrogen bond between Trp41 and Asp44 holds the two adjacent helices, making the tetrameric channel in a compact conformation.
(Fig. 3C), just exactly like the Asp44 does (Fig. 3B). However, Asn44 has higher pKa value than that of Asp44 in the M2 channel, because it has an acyl group (–CONH2) instead of carboxyl group (–COOH). According to our calculation using the improved PROPKA program and Eq. (1) [13,14], the pKa value of Asn44 in the ligand-free conformation of the M2 channel is 7.27 and that in the rimantadine-binding conformation is 6.07. Both are much higher than the corresponding pKa values, 3.80 and 2.60, of Asp44 in the M2 channel. That is why the R-M2 channel can keep open in higher pH environment and is more difficult to be locked by the inhibitor amantadine or rimantadine. That is also why the R-M2 channel possesses a sevenfold greater proton conductance and drug-resis-
tance. This is completely consistent with the report by Betakova et al. [6] as saying: ‘‘substitution of asparagines for aspartic acid 44 in the M2 PR8-M2 (D44N) caused an increase in activity (resistant to amantadine) to a level comparable to that of the wt Rostock protein, confirming the more general influence of this particular substitution at residue 44 on M2 activity”. On the other hand, the substitution of Asn with Asp (N44D) at potion 44 of the RM2 channel, however, caused a significant decrease in activity (approximately to the levels found for W-M2), indicating that the change in this residue alone could account for the difference in the pH-modulating activities of the two wt channel proteins [6]. Accordingly, our interpretation based on the NMR structure [8] is fully consistent with the results as observed in [6]. Although Jing et al. [10] cited the paper of Betakova et al. [6] to oppose the NMR structure and its action rimantadine, actually the conclusion drawn from their own D44A mutation study is completely conflicting with the results from the mutation studies D44N and N44D conducted by Betakova et al. [6], as can be illustrated as follows. In their paper [10], Jing et al. claimed that after the mutation D44A, the M2 channel still worked normally and the drug amantadine was still active, and hence the Asp44 is not functionally important. However, it is clearly stated by Betakova et al. [6] that the position 44 possesses ‘‘the more general influence” and the residue on the position 44 (Asp44 or Asn44) ‘‘could account for the difference in the pH-modulating activities”. Therefore Jing et al. [10] misinterpreted the report of [6]. Besides, the reliability of their results [10] is highly questionable because they used an unnatural mutation, D44A, to conduct the mutational experiments that would completely change the polar property of the residue, significantly impairing the channel activity. Mutations on positions from 27 to 38 Before the high-resolution NMR structure is available, the binding site of amantadine/rimantadine with the M2 channel was speculated according to the biological functional studies [15]. This can be seen from the statement by Hay et al. [15] as saying that ‘‘influenza A viruses that mutate and become drug resistant contain mutations in the A/M2 TM domain and naturally arising mutations mostly occur at TM domain residues Val27, Ala30, Ser31, and
1246
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
Fig. 3. (A) The amino acid sequences of the M2 channel: NMR structure 2RLF, Rostock strain R-M2, and Weybridge strain W-M2. The underline indicates the transmembrane part of the M2 channel. Difference between R-M2 and W-M2 is at the positions 27, 38, and 44. (B) In the NMR structure (2RLF), there are two hydrogen bonds between Asp44 and rimantadine and one hydrogen bond between Asp44 and Trp41 of the adjacent helix. (C) After D44N mutation for the M2 channel followed by energy optimization, the similar hydrogen bonding pattern still exists and hence Asn44 can play the same roles as Asp44 does.
Gly34”. Therefore some authors concluded that amantadine was positioned around Ser31, in line with the central location of this residue in the constellation of amantadine-resistant mutations (on Val27, Ala30, Ser31, and Gly34) [15,16]. However, using such a simplistic logic to identify the binding site according to the information obtained solely from amino acid mutations is not always correct because no ‘‘long-range” interaction effects [17] whatsoever have been taken into account. As is well known, the relationship between the function of a protein and its constituent amino acids is extremely complicated. It is impossible to really find their correlation without a reliable 3D structure. The high-resolution NMR structure of the M2 channel can provide a solid basis in this regard. As shown in Fig. 2, the M2 channel is locked by the hydrogen bond between Asp44 of one helix and Trp41of the adjacent helix. Such a lock is triggered by binding the amantadine/rimantadine ligand to Asp44 through two hydrogen bonds outside the pore. The strengths of these three hydrogen bonds all depend on the orientations of the two residues. Therefore, their orientations may affect the lock of the M2 channel. Since the M2 channel is composed of four transmembrane helices, the orientation of Asp44 in one helix and that of Trp41 in the adjacent helix may be affected by the residue mutation on any of the positions from 27 to 38, and hence affect the biological function of the M2 channel as well. But it should be pointed out that this kind of functional change is caused by the impact on the primary binding site via the long-range interaction.
helices, hence affecting the strength of hydrogen bonds between Asp44 and Trp41 and between Asp44 and amantadine/rimantadine, eventually resulting in drug-resistance problem to adamantane-based drugs. Confused by the long-range effects, it was prone to make mistakes in identifying the real drug-binding site of the M2 channel and provide misleading information for the real target in drug design. With the high-resolution NMR structure of the M2 channel available [8], it is now clear that the real binding site for amantadine/rimantadine is at the site involving Trp41 and Asp44.
The drug-resistance of M2 channel
Conclusion
The aforementioned long-range interaction effects can be also used to elucidate why the natural and artificial drug-resistant mutations in the M2 channel occurred in a large range from positions 27–38 [15], and why the drug-resistant influenza A virus in humans, birds and pigs has reached more than 90% during last three decades [1]. According to the long-range interaction [17], all mutations before position 41, especially on positions 27–38, could affect the orientations of Trp41 and Asp44 in two adjacent
The 3D M2 channel structure determined recently by the highresolution NMR technique [8] has provided a solid basis for understanding the proton-gating and drug-inhibition mechanism from a conception quite different from the traditional view. However, the new conception is not easy to be accepted or prone to be misinterpreted. The analyses and discussions presented in this article can help clarify the confusions in these regards. Particularly, it is crystal clear according to the above analysis that the functional studies
Coordinate number ratio It was reported that one amantadine molecule could block the M2 channel [18], which led to a conclusion that the coordinate number ratio (Hill coefficient) between the M2 receptor and amantadine ligand was 1:1. According to the NMR structure for the M2 channel in complex with rimantadine, the ligand binds at four equivalent sites near the gate Trp41 on the lipid-facing side of the channel and stabilizes the closed conformation of the pore [8]. However, it does not mean that the coordinate number ratio between the receptor and the ligand must be 1:4. As we can see from the above analysis and Fig. 2, merely one rimantadine molecule can also block an M2 channel effectively. Therefore, the coordinate number ratio of the M2 channel to amantadine/rimantadine could be anywhere between 1:1 and 1:4.
R.-B. Huang et al. / Biochemical and Biophysical Research Communications 377 (2008) 1243–1247
conducted by Betakova et al. [6] are completely consistent with the unique role of position 44 and the importance of Asp44 (or Asn44) in the M2 channel, as revealed by the NMR structure [8]. Moreover, it has also become clear now that the mutations on positions from 27 to 38 may affect the orientations of Asp44 and Trp41, and hence lead to drug-resistance as well. However, this belongs to a kind of indirect or ‘‘long-range interaction” effect. It is Asp44 and Trp41 that form the primary binding site of rimantadine as observed from the NMR structure. The current analyses may stimulate and encourage new strategies for finding effective drugs against influenza A virus. Acknowledgment This work is financially supported by the National High-tech Research and Development Program (‘863’) of China under the project 2006AA020103. References [1] V.M. Deyde, X. Xu, R.A. Bright, M. Shaw, C.B. Smith, Y. Zhang, Y. Shu, L.V. Gubareva, N.J. Cox, A.I. Klimov, Surveillance of resistance to adamantanes among influenza A(H3N2) and A(H1N1) viruses isolated worldwide, J. Infect. Dis. 196 (2007) 249–257. [2] B.J. Smith, J.L. McKimm-Breshkin, M. McDonald, R.T. Fernley, J.N. Varghese, P.M. Colman, Structural studies of the resistance of influenza virus neuramindase to inhibitors, J. Med. Chem. 45 (2002) 2207–2212. [3] K. Martin, A. Helenius, Nuclear transport of influenza virus ribonucleoproteins: the viral matrix protein (M1) promotes export and inhibits import, Cell 67 (1991) 117–130. [4] R.J. Sugrue, A.J. Hay, Structural characteristics of the M2 protein of influenza A viruses: evidence that it forms a tetrameric channel, Virology 180 (1991) 617–624.
1247
[5] R.A. Kumpf, D.A. Dougherty, A mechanism for ion selectivity in potassium channels: computational studies of cation-pi interactions, Science 261 (1993) 1708–1710. [6] T. Betakova, F. Ciampor, A.J. Hay, Influence of residue 44 on the activity of the M2 proton channel of influenza A virus, J. Gen. Virol. 86 (2005) 181–184. [7] S. Borman, Flu virus proton channel analyzed: structures of key surface protein suggest different drug mechanisms, Chem. Eng. News 86 (2008) 53–54. [8] J.R. Schnell, J.J. Chou, Structure and mechanism of the M2 proton channel of influenza A virus, Nature 451 (2008) 591–595. [9] L.J. Holsinger, D. Nichani, L.H. Pinto, R.A. Lamb, Influenza A virus M2 ion channel prote: a structure-function analysis, J. Virol. 68 (1994) 1551–1563. [10] X. Jing, C. Ma, Y. Ohigashi, F.A. Oliveira, T.S. Jardetzky, L.H. Pinto, R.A. Lamb, Functional studies indicate amantadine binds to the pore of the influenza A virus M2 proton-selective ion channel, Proc. Natl. Acad. Sci. USA 105 (2008) 10967–10972. [11] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson, Automated docking using a Lamarckian Genetic Algorithm and empirical binding free energy function, J. Comput. Chem. 19 (1998) 1639–1662. [12] W.D. Cornell, P. Cieplak, C.I. Bayly, I.R. Gould, K.M. Merz Jr., D.M. Ferguson, D.C. Spellmeyer, T. Fox, J.W. Caldwell, P.A. Kollman, A second generation force field for the simulation of proteins and nucleic acids, J. Am. Chem. Soc. 177 (1995) 5179–5197. [13] H. Li, A.D. Robertson, J.H. Jensen, The determinants of carboxyl pKa values in turkey ovomucoid third domain, Proteins 55 (2004) 689–704. [14] H. Li, A.D. Robertson, J.H. Jensen, Very fast empirical prediction and rationalization of protein pKa values, Proteins 61 (2005) 704–721. [15] A.J. Hay, A.J. Wolstenholme, J.J. Skehel, M.H. Smith, The molecular basis of the specific anti-influenza action of amantadine, EMBO J. 4 (1985) 3021–3024. [16] M. Yi, T.A. Cross, H.X. Zhou, A secondary gate as a mechanism for inhibition of the M2 proton channel by amantadine, J. Phys. Chem. B 112 (2008) 7977– 7979. [17] K.C. Chou, Review: prediction of tight turns and their types in proteins, Anal. Biochem. 286 (2000) 1–16. [18] C. Wang, K. Takeuchi, L.H. Pinto, R.A. Lamb, Ion channel activity of influenza A virus M2 protein: characterization of the amantadine block, J. Virol. 67 (1993) 5585–5594. [19] K.C. Chou, G.M. Maggiora, G. Nemethy, H.A. Scheraga, Energetics of the structure of the four-alpha-helix bundle in proteins, Proc. Natl. Acad. Sci. USA 85 (1988) 4295–4299.