Molecular basis of R294K mutation effects of H7N9 neuraminidases with drugs and cyclic peptides: an in silico and experimental study

Abstract

In China, a recent outbreak of the new R294K H7N9 influenza has infected 134 people and killed 45 people since January 2014. Prof. Gao et al. reported that an R294K neuraminidase (Shanghai N9: R294K mutation; Anhui N9: no R294K mutation) results in multi-drug resistance with extreme oseltamivir resistance (over 100 000-fold). Herein, we report the findings of molecular simulations and computational alanine-scanning mutagenesis for cyclic peptide I with the neuraminidases of Shanghai N9 and Anhui N9 A. Our results suggest that cyclic peptide I can inhibit the Shanghai N9 and Anhui N9 influenza A virus neuraminidases. The peptide can provide efficient binding affinities to Shanghai N9 either by the five residues (Arg119, Ile277, Lys294, Arg372 and Tyr406) mutations or by the 13 residues (Val117, Gln137, Thr149, Hie151, Asp152, Ser154, Ile224, Asn296, Asn347, Asn348, Ile429, Pro433 and Lys434) analyzed by binding mode analysis. However, the four residues (Ile150, Arg153, Gln155 and Arg157) can obviously affect the binding interactions among the cyclic peptide.

---

1. Introduction

The influenza A virus contains two major surface glycoproteins, neuraminidase (NA) and hemagglutinin (HA). The balance of HA and NA activities has been considered to be critical for influenza virus infectivity and transmissibility.¹ H7N9 is a strain of influenza A virus and normally circulates amongst avian populations with some variants known to occasionally infect humans. In China, a recent outbreak of the new R294K H7N9 virus has infected 134 people and killed 45 people since January 2014. Prof. Gao et al. reported that a R294K neuraminidase (Shanghai N9: R294K mutation; Anhui N9: no R294K mutation) results in multi-drug resistance with extreme oseltamivir resistance (over 100 000-fold).¹ The transfer function (ΔG_bind = −RT ln(IC₅₀)) is used to transfer the IC₅₀ values to the experimental binding free energy (ΔG_bind) values. For oseltamivir (G39), the ΔG_bind of the wild type neuraminidase (Anhui N9) is equal to −12.91 kcal mol⁻¹ (IC₅₀ = 0.79 nM), and the ΔG_bind of the R294K mutation neuraminidase (Shanghai N9) is equal to −5.20 kcal mol⁻¹ (IC₅₀ = 214 770 nM). For zanamivir (ZMR), the ΔG_bind of the wild type neuraminidase (Anhui N9) is equal to −13.31 kcal mol⁻¹ (IC₅₀ = 0.41 nM), and the ΔG_bind of the R294K mutation neuraminidase (Shanghai N9) is equal to −10.10 kcal mol⁻¹ (IC₅₀ = 75.7 nM). For peramivir (BCZ), the ΔG_bind of the wild type neuraminidase (Anhui N9) is equal to −13.33 kcal mol⁻¹ (IC₅₀ = 0.40 nM), and the ΔG_bind of the R294K mutation neuraminidase (Shanghai N9) is equal to −12.91 kcal mol⁻¹ (IC₅₀ = 0.79 nM). The calculated binding affinities (ΔG_bind) of the BCZ, G39 and ZMR inhibitors are very close to the experimental data (Tables 3 and 4 and the calculation of binding affinities (ΔG_bind) of the BCZ, G39 and ZMR inhibitors are listed in the ESI†). Therefore, we used the experimental data of the BCZ, G39 and ZMR inhibitors to compare with the N9/cyclic peptides complex systems. Although zanamivir and peramivir can still inhibit Shanghai N9, the antiviral activity is lost in oseltamivir. Thus new antivirus drugs must therefore be developed.

Neuraminidases cleave the terminal linkage of the sialic acid receptor, which results in the release of the progeny virions from the infected host cells. In addition, neuraminidase plays an important role in facilitating the early process of the infection of lung epithelial cells by the influenza virus.² Because of its essential role in influenza virus replication and its highly conserved active sites, neuraminidase has been an attractive target for the development of novel anti-influenza drugs.^3–10 Neuraminidase has been grouped into nine different subtypes (N1–N9) based on antigenicity.¹¹ Additionally, neuraminidase is further classified into two groups: group 1 (N1, N4, N5 and N8) and group 2 (N2, N3, N6, N7 and N9), based on the primary sequence. For the R294K neuraminidase of the 2013 H7N9 influenza A virus (Shanghai N9), it has also been reported that five residues (Arg119, Ile277, Lys294, Arg372 and Tyr406) are the key residues for drug-neuraminidase binding.¹

Peptides can bind with exquisite specificity to their in vivo targets, resulting in exceptionally high potencies of action and relatively few off-target side effects. This high degree of selectivity in their interactions is the product of millions of years of evolutionary selection for complementary shapes and sizes from among a huge array of structural and functional diversity. Peptides can be fine-tuned to interact specifically with biological targets, such as potent endogenous hormones, growth factors, neurotransmitters, immunologic agents and signaling molecules. Although peptides have many valuable applications in medicine, so far peptide synthesis has still been severely limited by high clearance, poor membrane permeability, low systemic stability and high manufacture costs. Recently, more than 100 peptide drugs have reached the market, such as oxytocin (8 residues), calcitonin (32 residues, hypercalcemia, osteoporosis), teriparatide (34 residues, parathyroid hormone analog, osteoporosis) and enfuvirtide (36 residues, enfuvirtide, antiretroviral).^12–15 Cyclic peptides are polypeptide chains wherein the amino termini and carboxyl termini, amino termini and side chain, carboxyl termini and side chain, or side chain and side chain are linked with a covalent bond that generates the ring. A number of cyclic peptides have been discovered in nature and many cyclic peptides have been synthesized in the laboratory. Cyclic peptides have several applications in medicine and biology. Recently, a cyclic peptide (Table 1; USA patent: US20130261048) has been reported to efficiently inhibit the neuraminidases of influenza A or B viruses. Thus the objective of our research is to discover novel drugs for Shanghai N9 by using cyclic peptides as neuraminidase inhibitors. Alanine-scanning mutagenesis is a useful technique in the determination of the catalytic or functional sites of protein residues.¹⁶ Since alanine amino residues do not alter the main-chain conformation or impose extreme electrostatic or steric effects, alanine is the substitution residue of choice. Sometimes bulky amino acids such as phenylalanine or tyrosine are used in cases where conservation of the size of mutated residues is needed. This technique can also be used to determine whether the side chain of a specific residue plays a significant role in bioactivity. Computational alanine-scanning mutagenesis methods, developed from bioinformatics and molecular mechanics, can predict important residues of protein–small compound interactions. As a computational method, the computational alanine-scanning mutagenesis method¹⁷ can predict the important residues in protein–small compound interactions.¹⁸ Here the SIE (solvated interaction energy)¹⁷ alanine-scanning mutagenesis method was performed to validate the drug-resistance issue of the Shanghai N9 R294K H7N9 influenza A virus neuraminidase. In the present study, the Shanghai N9 and Anhui N9 with the cyclic peptide I were studied with molecular dynamics (MD) simulations and binding free energy calculations. The Shanghai N9 with the cyclic peptide was analyzed by computational alanine-scanning mutagenesis calculations. Our simulations aim to gain further insight into the binding interactions between Shanghai N9 and the cyclic peptide.

Table 1 Cyclic peptide

Number of cyclic peptide	Amino sequence in one-letter or drug name
I	CGQRETPEGAEAKPWYC

---

2. Results and discussion

2.1 Binding modes of the Shanghai N9 influenza A virus neuraminidase with the cyclic peptide

In comparison, the cyclic peptide I has different binding modes with the Shanghai N9 influenza A virus neuraminidase and the analysis of our simulations is shown in Tables 2 and 3 and Fig. S1.† The overall results of our simulations and the X-ray crystal information^1,19 suggest that Arg153, Asn347, Arg119, Arg372, Hie151, Lys294, Asp152, Asn348 and Asn347 can form hydrogen bonds with the R294K H7N9 neuraminidase. Asn200, Asn348, Ile150, Asp152, Arg119, Lys434, Hie151, Pro433, Ile429, and Ile224 can form non-bonding interactions with the R294K H7N9 neuraminidase. Comparing our results (most frequent residues in Tables 2 and 3 and Fig. S1†) and a recent report,¹ the 22 residues (Val117, Arg119, Gln137, Thr149, Ile150, Hie151, Asp152, Arg153, Ser154, Gln155, Arg157, Ile224, Ile277, Lys294, Asn296, Asn347, Asn348, Arg372, Tyr406, Ile429, Pro433 and Lys434) are subjected to be key residues for Shanghai N9. For the R294K neuraminidase of the 2013 H7N9 influenza A virus (Shanghai N9), it has also been reported that five residues (Arg119, Ile277, Lys294, Arg372 and Tyr406) are the binding key residues for drug-neuraminidase binding.¹ Tables 2 and 3 show that the cyclic peptide can interact with these binding key residues for drug-neuraminidase binding.¹

Table 2 Analysis of the binding modes of cyclic peptide I with the Shanghai N9 neuraminidase

Number of cyclic peptide	Hydrogen bonding	Non-bonding
I	Gln155, Asn347, Asn348, Arg372, Arg153, Ile150, Arg157, Gln137	Ile224, Asn296, Lys294, Tyr406, Pro433, Hie151, Lys434, Ile429, Arg119, Thr149, Ser154, Val117, Asp152

---

Table 3 The detail binding information of cyclic peptide I with the Shanghai N9 neuraminidase

Number of residue (cyclic peptide I)	Hydrogen bonding	Non-bonding
a Two hydrogen bonds.
Cys1	Null	Ans296, Ile224
Gly2	Null	Null
Gln3	Null	Null
Arg4	Null	Asp152
Glu5	Gln155^a	Ile224
Thr6	Null	Hie151, Asp152
Pro7	Glu137, Arg157	Hie151, Asp152
Glu8	Arg157,^a Ile150	Hie151, Pro433, Thr149, Ser154
Gly9	Null	Pro433, Arg119
Ala10	Null	Pro433
Glu11	Arg153	Hie151, Lys294, Asn296
Ala12	Null	Tyr406
Lys13	Null	Asn348
Pro14	Null	Null
Trp15	Asn347	Null
Tyr16	Null	Hie151, Lys294, Asn296
Cys17	Null	Hie151

---

2.2 Binding free energy of Shanghai N9 and Anhui N9 influenza A virus neuraminidase with cyclic peptide I

The binding free energy of each drug was obtained from a 900 ns MD simulation and the SIE method, with both processes using the same parameters. All the results are shown in Tables 4 and 5 and Fig. 1. For Shanghai N9, the binding free energy of cyclic peptide I is −8.26 ± 0.29 kcal mol⁻¹. For Anhui N9, the binding free energy of cyclic peptide I is −8.24 ± 0.53 kcal mol⁻¹. In comparison with our simulation and the experimental binding free energies of the three drugs (G39, ZMR and BCZ), cyclic peptide I may provide efficient binding affinities with the Anhui N9 and Shanghai N9 influenza A virus neuraminidases. The SIE calculations are shown in Tables 4 and 5.

Table 4 Binding free energies of the cyclic peptide/BZC/G39/ZMR with the Shanghai N9 influenza A virus neuraminidase

Number of cyclic peptide	Inter vdW	Inter coulomb	Reaction field	Cavity	Constant	Predicted ΔGbinding (kcal mol^−1)	Experimental ΔGbinding (kcal mol^−1)
a Our experimental result.
I	−90.05 ± 1.03	−90.94 ± 1.86	156.34 ± 1.33	−26.68 ± 0.36	−2.89	−8.26 ± 0.33	−8.56^a
BCZ	−36.70 ± 2.31	−135.21 ± 4.20	90.28 ± 3.07	−8.16 ± 0.17	−2.89	−11.35 ± 0.23	−12.911
G39	−21.03 ± 1.41	−97.16 ± 3.81	98.59 ± 3.02	−6.79 ± 0.07	−2.89	−5.65 ± 0.28	−5.201
ZMR	−30.15 ± 1.07	−110.31 ± 3.02	83.54 ± 2.47	−6.53 ± 0.07	−2.89	−9.74 ± 0.16	−10.101

---

Table 5 Binding free energies of the cyclic peptide/BZC/G39/ZMR with the Anhui N9 influenza A virus neuraminidase

Number of cyclic peptide	Inter vdW	Inter coulomb	Reaction field	Cavity	Constant	Predicted ΔGbinding (kcal mol^−1)	Experimental ΔGbinding (kcal mol^−1)
a Our experimental result.
I	−85.49 ± 1.17	−86.18 ± 1.57	149.74 ± 1.08	−29.19 ± 0.38	−2.89	−8.24 ± 0.38	−8.75^a
BCZ	−33.68 ± 1.83	−141.75 ± 3.81	86.56 ± 4.12	−7.87 ± 0.06	−2.89	−13.02 ± 0.37	−13.331
G39	−34.97 ± 2.74	−135.35 ± 4.57	83.82 ± 2.71	−7.76 ± 0.17	−2.89	−12.75 ± 0.31	−12.911
ZMR	−30.23 ± 2.13	−135.81 ± 2.96	81.81 ± 3.18	−7.85 ± 0.07	−2.89	−12.53 ± 0.63	−13.311

---

Fig. 1 The binding free energies of the cyclic peptide, oseltamivir (G39), peramivir (BCZ), oseltamivir (G39), and zanamivir (ZMR) to the Shanghai N9 and Anhui N9 influenza A virus neuraminidases.

2.3 Computational alanine-scanning mutagenesis calculations of the binding modes of cyclic peptide I with the Shanghai N9 influenza A virus neuraminidase

The 22 residues (Val117, Arg119, Gln137, Thr149, Ile150, Hie151, Asp152, Arg153, Ser154, Gln155, Arg157, Ile224, Ile277, Lys294, Asn296, Asn347, Asn348, Arg372, Tyr406, Ile429, Pro433 and Lys434), which were analyzed by binding mode analysis, were mutated to alanine for computational alanine-scanning mutagenesis calculations. The binding free energy of each drug was obtained from a 1000 ns MD simulation and the SIE method, with both processes using the same parameters. All the results are shown in Table S1† and Fig. 2. The computational alanine-scanning mutagenesis calculations of the 4 residues (Ile150, Arg153, Gln155 and Arg157) can obviously affect the binding interactions of the cyclic peptide. The ΔG_diff of Ile150, Arg153, Gln155 and Arg157 are −1.03, −1.11, −1.01 and −1.85 kcal mol⁻¹, respectively. The mutation of Ile150 caused a net reduction in hydrogen bonding and an inter coulomb/vdW interactions loss (coulomb interaction: −90.94 ± 1.86 to −87.17 ± 2.75; van der Waals (vdW) interaction: −90.05 ± 1.86 to −82.64 ± 1.48 kcal mol⁻¹). The mutation of Arg153 caused a net reduction in hydrogen bonding and an inter coulomb/vdW interactions loss (coulomb interaction: −90.94 ± 1.86 to −79.13 ± 2.71; vdW interaction: −90.05 ± 1.86 to −81.65 ± 1.47 kcal mol⁻¹). The mutation of Gln155 caused a net reduction in hydrogen bonding and an inter coulomb/vdW interactions loss (coulomb interaction: −90.94 ± 1.86 to −84.14 ± 2.72; vdW interaction: −90.05 ± 1.86 to −84.32 ± 1.47 kcal mol⁻¹). The mutation of Arg157 caused three net reductions in hydrogen bonding and an inter coulomb/vdW interactions loss (coulomb interaction: −90.94 ± 1.86 to −71.54 ± 2.51; vdW interaction: −90.05 ± 1.86 to −87.32 ± 1.11 kcal mol⁻¹). In comparison with Table S1† and Fig. 2, the differences in the binding free energies (ΔG_diff) of the residue mutations of Ile150, Arg153, Gln155 and Arg157 could affect the binding interactions with cyclic peptide I.

Fig. 2 Differences in the predicted binding free energies of the virtual alanine mutations of cyclic peptide I to the Shanghai N9 influenza A virus neuraminidase (Table S1†).

2.4 Neuraminidase enzyme assay

For Shanghai N9 and Anhui N9, the IC₅₀ values of the cyclic peptide I are 910 and 670 nM, respectively. The transfer function (ΔG_bind = −RT ln(IC₅₀)) is used to transfer the IC₅₀ values to the experimental binding free energy (ΔG_bind) values. For Shanghai N9/cyclic peptide I, the ΔG_bind is equal to −8.56 kcal mol⁻¹. For Anhui N9/cyclic peptide I, the ΔG_bind is equal to −8.75 kcal mol⁻¹. Although the IC₅₀ values of the cyclic peptide are greater than the IC₅₀ values of the ZMR and BCZ drugs, the IC₅₀ values of the cyclic peptide are less than the IC₅₀ values of G39 (Shanghai N9). We suggest that cyclic peptides can be potential drugs for Shanghai N9 and Anhui N9 influenza A virus disease treatment.

3. Materials and methods

3.1 3D structures design of cyclic peptide I and neuraminidases

The cyclic peptide structures were generated by a Discovery Studio visualizer. Then the cyclic peptide was subjected to optimize the 3D structures with VEGA ZZ software.²⁰ Initial cyclic peptides/neuraminidase (PDB ID: 4MXS for Shanghai N9; PDB ID: 4MWQ for Anhui N9) complex structures were generated by Autodock Vina software. Autodock Vina is a fast and accurate way to dock drugs into fixed protein binding sites, utilizing NNscore 2.0 (ref. 21) and several types of genetic algorithms. Four thousand conformations were obtained from docking for the cyclic peptide and then scored by the NNscore function. The conformations of the best NNscores were then selected for subsequent MD simulations and solvated interaction energy (binding free energy) calculations. Fig. 3 shows the overview of the Shanghai N9/cyclic peptide I complex structure.

Fig. 3 An overview of the Shanghai N9/cyclic peptide I complex structure. Shanghai N9 is shown in the cartoon model (colored in orange-yellow) and cyclic peptide I is shown in the ball-stick model (colored in red).

3.2 Computational models of the cyclic peptide I and neuraminidases

Our models were then calculated with the AMBER 14 (pmemd.cuda)^21–23 package using the AMBER FF12SB all-hydrogen amino acid parameter.²⁴ Each complex was solvated in TIP3P²⁵ water using a truncated hexahedron periodic box extending at least 10 Å from the complex. Nearly 10 000 water molecules were added to solvate the complex, with a resulting box size of nearly 80.21 × 82.71 × 90.12 ų. All MD simulations were performed in the canonical ensemble with a simulation temperature of 310 K, unless stated otherwise, by using a Verlet integrator with an integration time step of 0.002 ps and SHAKE constraints²⁶ of all covalent bonds involving hydrogen atoms. In the electrostatic interactions, atom-based truncation was performed using the PME²⁷ method, and the switch van der Waals function was used with a 2.00 nm cutoff for atom-pair lists. The complex structure was minimized for 100 000 conjugate gradient steps, and was then subjected to a 1000 ns isothermal, constant volume MD simulation. Moreover, the final structures from these simulations were used in the solvated interaction energy (SIE) calculations, and were used to analyze the binding modes of the cyclic peptides with Shanghai N9. The important residues of our results (binding mode analysis from Shanghai N9) and the 22 residues (Val117, Arg119, Gln137, Thr149, Ile150, Hie151, Asp152, Arg153, Ser154, Gln155, Arg157, Ile224, Ile277, Lys294, Asn296, Asn347, Asn348, Arg372, Tyr406, Ile429, Pro433 and Lys434) were subjected to computational alanine-scanning mutagenesis. Fig. 4 and 5 of the complex structures (both Anhui and Shanghai N9/cyclic peptide I/G39/ZMR/BCZ) were equilibrated at 100 ns. Each of the complex structures were sampled at 900 ns (99 000 snapshots) for binding free energy (SIE) calculations. The snapshots of Shanghai N9/cyclic peptide I were subjected to computational alanine-scanning mutagenesis calculations. Our models with the three drugs were then simulated with an AMBER package using the AMBER FF12SB all-hydrogen amino acid and general amber force field (GAFF) parameters. The geometries of the three compounds were fully optimized and their electrostatic potentials were obtained using single-point calculations, both at the Hartree–Fock level with the 6-31G(d,p) basis set using a GAUSSIAN 09 program. Subsequently, their partial charges were obtained with the restrained electrostatic potential (RESP) procedure using Antechamber. The initial structures are Shanghai N9 (PDB ID: 4MWW for G39; 4MWX for ZMR; 4MW0 for BCZ) and Anhui N9 complex structures (PDB ID: 4MWQ for G39; 4MWR for ZMR; 4MWV for BCZ). Then these complex structures were inserted into the TIP3P water box. All the MD simulations were performed in a canonical ensemble with a simulation temperature of 310 K, unless stated otherwise, by using the Verlet integrator with an integration time step of 0.002 ps and SHAKE constraints of all covalent bonds involving hydrogen atoms. In the electrostatic interactions, atom-based truncation was performed using the PME method, and the switch van der Waals function was used with a 2.00 nm cutoff for atom-pair lists. The complex structure was minimized for 100 000 conjugate gradient steps, and was then subjected to a 1000 ns isothermal, constant volume MD simulation. The final structures from these simulations were used in the solvated interaction energy (SIE) calculations.

Fig. 4 The potential energies of the complex structures (both Anhui and Shanghai N9/cyclic peptide I/G39/ZMR/BCZ) were equilibrated during the 1000 ns simulations.

Fig. 5 The complex structures (both Anhui and Shanghai N9/cyclic peptide I/G39/ZMR/BCZ) were equilibrated at 100 ns.

3.3 Solvated interaction energy (SIE) method and computational alanine-scanning mutagenesis

Our binding free energy calculations were performed by the SIE method. The SIE method is a kind of linear interaction energy (LIE) approach.^28–30 The SIE¹⁷ function to estimate the protein–ligand free energy is written as:

ΔG_bind(ρ,D_in,α,γ,C) = α × [E_C(D_in) + ΔG^R_bind(ρ,D_in) + E_vdW + γΔMSA(ρ)] + C (1)

where E_C and E_vdW are the intermolecular coulomb and van der Waals interaction energies in the bound state, respectively. These values were calculated using the AMBER molecular mechanics force field (FF99SB) with an optimized dielectric constant. ΔG^R_bind is the change in the reaction field energy between the bound and free states and is calculated by solving the Poisson equation with a boundary element method program, BRI BEM,³¹ and using a molecular surface generated with a variable-radius solvent probe.³² The ΔMSA term is the change in the molecular surface area upon binding. The following parameters are calibrated by fitting to the absolute binding free energies for a set of 99 protein–ligand complexes: AMBER van der Waals radii linear scaling coefficient (ρ), the solute interior dielectric constant (D_in), the molecular surface area coefficient (γ), the global proportionality coefficient related to the loss of configurational entropy upon binding (α), and a constant (C).²⁸ The optimized values of these parameters are α = 0.1048, D_in = 2.25, ρ = 1.1, γ = 0.0129 kcal (mol 0.1 nm²)⁻¹, and C = −2.89 kcal mol⁻¹. The 14 key residues from the binding modes analyzed were mutated to alanine for computational alanine-scanning mutagenesis calculations. These mutated complex structures were then sampled with 1000 ns MD simulations. Moreover, the final structures from these simulations were used in the SIE calculations. The SIE and computational alanine-scanning mutagenesis calculations were carried out with a Sietraj program.¹⁷

3.4 Neuraminidase enzyme assay protocol³³

A standard fluorometric enzyme assay³⁴ was adapted to measure the neuraminidase activity. Producer cell lysates were transfected with representative neuraminidase constructs and added to a neuraminidase fluorogenic substrate, 2′-4-methylumbelliferyl-N-acetylneuraminic acid (4-MUNANA; Sigma), to a final concentration of 100 μmol L⁻¹. The reactions were carried out in 50 μL of 33 mmol L⁻¹ MES (pH 6.5) containing 4 mmol L⁻¹ CaCl₂ in 96-well black Optiplates (BD Biosciences). A titration of the cyclic peptides were added to the reaction mixtures and then incubated in a 37 °C water bath for 1 hour. The reactions were terminated by adding 150 μL stop solution containing 0.5 mol L⁻¹ NaOH (pH 10.7) and 25% ethanol. The fluorescence of the released 4-methylumbelliferone was measured using a spectrophotometer (PerkinElmer). The excitation wavelength was set at 355 nm, and the emission wavelength was set at 460 nm. Samples were done in triplicate and the experiments repeated at least three times. The Shanghai N9 and Anhui N9 were obtained from Sino Biological Inc. The IC₅₀ value was calculated by Sigmaplot software.

4. Conclusions

Cyclic peptides are an unusual class of compounds, which were first discovered from microorganisms, due to their broad biological activities, such as antimicrobial, antiviral, immunosuppressive, and antitumor activities. Cyclic peptides are undergoing very active investigations as potential new sources of drugs and antibiotics. They are much more resistant to proteases than a linear peptide chain. This resistance to proteolysis means that they tend to survive the human digestive process. They can also bind proteins in the cell where traditional drugs cannot.³⁵ The N- to C-terminal cyclic peptides can provide the important principles and strategies that clearly resonate within the world of bioactive peptides and peptide toxins. The ability to transform peptide biologics into stable and orally active constituents represents a major pharmacological achievement. Cyclic peptide drug leads have gained the attention of the pharmaceutical industry. In this article, we used the Autodock Vina program, a Tip3 water solvent model, MD simulations techniques, a computational alanine-scanning mutagenesis method and the SIE method to predict the binding affinities in which peramivir (BCZ), oseltamivir (G39), zanamivir (ZMR) and cyclic peptide I interact with the Shanghai N9 and Anhui N9 influenza A virus neuraminidases. The calculation binding affinities (ΔG_bind) of the cyclic peptide I, BCZ, G39 and ZMR inhibitors are very close to the experimental data. From our SIE binding free energies simulation results, cyclic peptides can be potential drugs for Shanghai N9 and Anhui N9 influenza A virus disease treatment. For validating the drug-resistance issue, we performed the computational alanine-scanning mutagenesis simulation to study the binding affinities of the cyclic peptide I to Shanghai N9 neuraminidases. Cyclic peptide I can provide efficient binding affinities to Shanghai N9 either by the five residues (Arg119, Ile277, Lys294, Arg372 and Tyr406) mutations¹ or by the 13 residues (Val117, Gln137, Thr149, Hie151, Asp152, Ser154, Ile224, Asn296, Asn347, Asn348, Ile429, Pro433 and Lys434) analyzed by binding mode analysis. But the four residues (Ile150, Arg153, Gln155 and Arg157) can obviously affect the binding interactions among the cyclic peptide. When comparing with the three drugs (BCZ, G39 and ZMR), the IC₅₀ values of the cyclic peptide are greater than the IC₅₀ values of the ZMR and BCZ drugs. The IC₅₀ values of the cyclic peptide are less than the IC₅₀ values of G39 (Shanghai N9). Therefore, our approach theoretically suggests that cyclic peptide I has the potential to inhibit the Shanghai N9 and Anhui N9 influenza A virus neuraminidases.

Acknowledgements

The authors would like to thank the Kaohsiung Medical University of the Taiwan and the Ministry of Science and Technology of Taiwan, for supporting this research (Contract no. MOST 103-2113-M-037-007 and KMU-TP103C0).

Notes and references

1. Y. Wu, Y. Bi, C. J. Vavricka, X. Sun, Y. Zhang, F. Gao, M. Zhao, H. Xiao, C. Qin, J. He, W. Liu, J. Yan, J. Qi and G. F. Gao, Cell Res., 2013, 23, 1347–1355.
2. M.-B. L. Jennifer, Antiviral Res., 2000, 47, 1–17.
3. S.-Q. Wang, Q.-S. Du and K.-C. Chou, Biochem. Biophys. Res. Commun., 2007, 354, 634–640.
4. Y.-T. Wang, C.-h. Chan, Z.-Y. Su and C.-L. Chen, Biophys. Chem., 2010, 147, 74–80.
5. H. Liu, X. Yao, C. Wang and J. Han, Mol. Pharm., 2010, 7, 894–904.
6. R. E. Amaro, R. V. Swift, L. Votapka, W. W. Li, R. C. Walker and R. M. Bush, Nat. Commun., 2011, 2, 388.
7. D. Pan, H. Sun, C. Bai, Y. Shen, N. Jin, H. Liu and X. Yao, J. Mol. Model., 2011, 17, 2465–2473.
8. S. Chavan, S. Bhayye and M. Sobhia, Mol. Diversity, 2011, 15, 979–987.
9. B. K. Mai and M. S. Li, Biochem. Biophys. Res. Commun., 2011, 410, 688–691.
10. A. Sharma, A. Tendulkar and P. Wangikar, Med. Chem. Res., 2011, 20, 1445–1449.
11. P. M. Colman, Protein Sci., 1994, 3, 1687–1696.
12. E. K. Spanakis and D. E. Sellmeyer, Osteoporosis Int., 2014, 25, 1411–1414.
13. M. L. Greenberg and N. Cammack, J. Antimicrob. Chemother., 2004, 54, 333–340.
14. H.-J. Lee, A. H. Macbeth, J. H. Pagani and W. Scott Young 3rd, Prog. Neurobiol., 2009, 88, 127–151.
15. K. G. Saag, E. Shane, S. Boonen, F. Marín, D. W. Donley, K. A. Taylor, G. P. Dalsky and R. Marcus, N. Engl. J. Med., 2007, 357, 2028–2039.
16. F. Lefèvre, M.-H. Rémy and J.-M. Masson, Nucleic Acids Res., 1997, 25, 447–448.
17. Q. Cui, T. Sulea, J. D. Schrag, C. Munger, M.-N. Hung, M. Naïm, M. Cygler and E. O. Purisima, J. Mol. Biol., 2008, 379, 787–802.
18. D. J. Diller, C. Humblet, X. Zhang and L. M. Westerhoff, Proteins: Struct., Funct., Bioinf., 2010, 78, 2329–2337.
19. H. Yang, P. J. Carney, J. C. Chang, J. M. Villanueva and J. Stevens, J. Virol., 2013, 87, 12433–12446.
20. A. Pedretti, L. Villa and G. Vistoli, J. Comput.-Aided Mol. Des., 2004, 18, 167–173.
21. R. Salomon-Ferrer, D. A. Case and R. C. Walker, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2013, 3, 198–210.
22. D. A. Case, J. T. Berryman, R. M. Betz, D. S. Cerutti, T. E. Cheatham III, T. A. Darden, R. E. Duke, T. J. Giese, H. Gohlke, A. W. Goetz, N. Homeyer, S. Izadi, P. Janowski, J. Kaus, A. Kovalenko, T. S. Lee, S. LeGrand, P. Li, T. Luchko, R. Luo, B. Madej, K. M. Merz, G. Monard, P. Needham, H. Nguyen, H. T. Nguyen, I. Omelyan, A. Onufriev, D. R. Roe, A. Roitberg, R. Salomon-Ferrer, C. L. Simmerling, W. Smith, J. Swails, R. C. Walker, J. Wang, R. M. Wolf, X. Wu, D. M. York and P. A. Kollman, AMBER 2015, University of California, San Francisco, 2015.
23. R. Salomon-Ferrer, A. W. Götz, D. Poole, S. Le Grand and R. C. Walker, J. Chem. Theory Comput., 2013, 9, 3878–3888.
24. T. Steinbrecher, J. Latzer and D. A. Case, J. Chem. Theory Comput., 2012, 8, 4405–4412.
25. W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey and M. L. Klein, J. Chem. Phys., 1983, 79, 926–935.
26. J.-P. Ryckaert, G. Ciccotti and H. J. C. Berendsen, J. Comput. Phys., 1977, 23, 327–341.
27. T. Darden, D. York and L. Pedersen, J. Chem. Phys., 1993, 98, 10089–10092.
28. A. Perdih, U. Bren and T. Solmajer, J. Mol. Model., 2009, 15, 983–996.
29. U. Bren, V. Martínek and J. Florián, J. Phys. Chem. B, 2006, 110, 10557–10566.
30. J. Åqvist, C. Medina and J.-E. Samuelsson, Protein Eng., 1994, 7, 385–391.
31. E. O. Purisima, J. Comput. Chem., 1998, 19, 1494–1504.
32. S. Bhat and E. O. Purisima, Proteins: Struct., Funct., Bioinf., 2006, 62, 244–261.
33. M. Potier, L. Mameli, M. Bélisle, L. Dallaire and S. B. Melançon, Anal. Biochem., 1979, 94, 287–296.
34. J. Tisoncik-Go, K. Cordero and L. Rong, Virol. Sin., 2013, 28, 81–91.
35. P. Thapa, M. Espiritu, C. Cabalteja and J.-P. Bingham, Int. J. Pept. Res. Ther., 2014, 20, 545–551.

---

Footnote

† Electronic supplementary information (ESI) available: Table S1: detail information of free energy of computational alanine-scanning mutagenesis calculations of the cyclic peptide I with Shanghai N9 influenza A virus neuraminidase. See DOI: 10.1039/c5ra10068b

---