Biomimetic design: a programmed tetradecapeptide folds and auto-dimerizes as a stereochemically articulated receptor protein

Abstract

In order to mimic biogeny, protein-quaternary structure may be evolved to the desired structure specificity—as a fold, assembly, and sequence—in a hierarchy of independent design steps. In demonstration of the algorithm, octapeptides were accomplished to obtain the desired structural specificity as folds, assemblies, and sequences over a hierarchy of design steps; the approach was shown to extend the structure by allowing amino acid stereochemistry to be applied as a design variable. Moreover, it was proven to simplify design by allowing the programming to involve a hierarchy of independent steps. Illustrating the abovementioned approach of biomimetic design with involvement of amino acid stereochemistry as the design variable, we now report an example of functional design, in which a fourteen-residue sequence was accomplished to the desired binding specificity as a receptor protein, first in the fold stereochemically, then in the assembly spatially, and finally in the sequence chemically.

---

Introduction

Biology rests on the versatility of proteins to structurally exist as catalysts, regulators, and building blocks, which are critical for viability of the living cell. The basis may lie in the ability of the structures to adaptationally evolve in a hierarchy of independent selections, first in conformation as folds, then in spatial structure as assemblies, and finally in chemical structure as sequences. Having the higher level in structure adapt or evolve over the lower one in an independent step, the hierarchy may be crucial to biogeny as it would allow the useable structures to recur during protein evolution, while it would challenge design in the vast diversity of combinatorials possible. Design too may be in folds, assemblies, and sequences implemented in a hierarchy of structure selections. That a level of the natural hierarchy indeed may be programmed over the lower one in an independent step was considered for having the folds and assemblies of interest inverse programmed as the desired polypeptide sequences chemically.¹ Given successes of the algorithm,^2,3 we considered its extension for having not only chemical sequences but also conformational folds and assemblies of the folds adapted or programmed to the desired structure specificity computationally.⁴ To test the algorithm, chirality in α-carbon was considered, in addition to conformation in main chain and chemical structure in sequence, as a variable for having the scope of sequence-level design extended beyond what is possible when α-helix and β-sheet are the building blocks used, being stereochemically frozen as poly-L sequences. L- and D-α-amino acids were introduced as a design alphabet,⁵ which was proven to offer, fundamentally, the possibility of shape-specific design, i.e., bracelet-,⁶ boat-,⁷ canoe-,⁸ and cup-shaped⁹ folds were achieved, which illustrate diversity of the shapes possible. Stereochemistry may extend but will also complicate design in the combinatorials possible. To tackle the problem, design was approached as a hierarchy of steps, programming first folds—in isolation of spatial and chemical variables—stereochemically, then assemblies—in isolation of conformational and chemical variables—spatially, and finally sequences—in isolation of conformational and spatial variables—chemically. With this algorithm, a shigatoxin-like C5 homopentamer could be accomplished as a protein, which is remarkably small in its stereochemically articulated eight-residue folds compared with the sixty-eight-residue folds of the natural protein, which is poly-L in structure.⁴ To further test the design algorithm and stereochemistry as the variable, we now report an example of functional design. A fourteen-residue sequence is accomplished to the desired binding specificity as a receptor protein by its stepwise programming first as the desired fold stereochemically, then as the desired assembly spatially, and finally as the desired sequence chemically.

Bis-nitrophenylcarbonate (bNPC) and bis-nitrophenylphosphate (bNPP) were harnessed as the model ligands. Bilateral symmetry of the structures was exploited for approaching the receptor structure biomimetically, targeting it as an assembly C2 in symmetry, such as in the HIV protease active site.¹⁰ A homooligomer of specific stoichiometry and point-group symmetry—cyclic, dihedral, or cubic—may be designed on the basis of the packing specificity of the monomer as the building-block fold. Packing specificity of the fold will depend on its shape, which in the protein monomers is stereospecific for their topology as poly-L folds.⁵ With α-carbon chirality diversified, an assembly of interest may be adapted to the desired topological specificity in its building-block fold stereochemically. A desired assembly was thus approached by computational selection of stereochemistry and conformation in main chain, shape in fold of main chain, and chemical structure in side chains as the hierarchy of design variables.⁴ Stereochemically diverse homopolypeptides modeled computationally as random folds were thus first selected, with SYMMDOCK,¹¹ as the desired assemblies, and then inverse programmed, with IDeAS,²¹ as the desired amino acid sequences. Implemented over the random folds derived from the 256 stereoisomers possible for blocked octaleucine (Ac-Leu₈-NHMe), the method gave the sequences that spontaneously folded and assembled to the desired specificity as homopentamers.⁴ With shape in fold directing its self assembly, the pentamer of interest was easy to achieve as the monomer fold of interest. Similar design of monomer for its self assembly and ligand-binding ability as a receptor of interest is a more difficult challenge for computational design. For tackling the challenge, we approached design of the desired receptor structure biomimetically. A β-hairpin ordered as a homodimer was adapted for the desired binding specificity as a receptor protein first in the fold stereochemically and then in the sequence chemically.

Assembly of the desired structure can be specified by having the protomer fold interact over either its side chains or main chain. The reported C5 structure was specified with the interactions involving mainly its side chains.⁴ The homodimer in this study was specified with the interactions involving primarily its main chain. A type II′ β-hairpin, centered on its conformational nucleator in the ^DPro₇-Gly₈ structure, was targeted for self assembly over the interactions of main chain.^12,13 The dimer mutated in the sequence positions 4 and 11 from L to D structure, transforms, on well documented reasoning,⁵ to a boat-like homodimer. As illustrated in Fig. 1, the homodimer had a cleft for possible design as the desired receptor structure chemically. Only a limited number of interactions are likely to hold the monomers in the desired intermolecular assembly; thus, there is interest to determine if the dimer will assemble and bind the desired ligand successfully.

Fig. 1 Stereochemically bent β-hairpin as a C₂-symmetric homodimer (panel A) complexed with bis-nitrophenyl-phosphate (bNPP) as the binding ligand (panel B).

The dimer was targeted for chemical-level design as the desired receptor protein. Every sequence position, except ^DPro₇-Gly₈, was tested for a side chain to define the fold, assembly, and complex as a minimum-energy polypeptide sequence. Side chain polarity and aromaticity were applied as a heuristic to maximize hydrophobic gain and π–π interaction in the designed sequences as the desired proteins. Side chain coordinates were drawn from a database of natural rotamers.^14,15 The structures were applied directly in the sequence positions of L stereochemical structure and after a suitable symmetry transform, as the enantiomers in the positions of D stereochemical structure. Specific positions in the dimer loaded with the active site ligand were tested for the sterically compatible rotamers, which were evaluated as minimum energy pairs over the interacting positions in the bound complex using the dead-end-elimination calculation.¹⁶ Minimum energy sequences were computed over the selected rotamers with Monte-Carlo algorithm.¹⁷ Orientation-dependent hydrogen bond energies,¹⁸ solvent-accessibility based hydrophobic gain,¹⁹ and steric and dispersion-interaction calculations involving Lennard-Jones potential²⁰ were the energy functions applied. The search methods and energy functions were implemented with the software IDeAS.²¹ P1, P2, and P3 are the polypeptide sequences chosen for verifying the design. The interactions likely to define the dimer and its bound complex as the minimum-energy sequences are summarized in Table 1. In P1, Ile₂ substitutes Tyr₂ of P2 and P3 for hydrophobic effect or π–π interaction in ligand binding. In addition, Trp_6/9 and His_9/14 in all the sequences and Tyr₂ in P2 and P3 are provided for holding the ligand in π–π interaction.

Table 1 Specific interactions in complexes of peptide homodimers with ligands

Interactions	H-bond	Salt bridge	π–π/CH–π/hydrophobic
Intra-molecular	^2INH–^13VCO (1), ^13VNH–^2ICO (1), ^2YNH–^13LCO (2), ^13LNH–^2YCO (2), ^4ENH–^11KCO (1, 2, 3), ^11KNH–^4ECO (1, 2, 3), ^6LNH–^9WCO (1), ^9WNH–^6LCO (1), ^6WNH–^9ICO (2), ^9INH–^6WCO (2), ^6WNH–^9HCO (3), ^9HNH–^6WCO (3), ^9WNH–^7PCO (1), ^9INH–^7PCO (2), ^9HNH–^7PCO (3)	^DE4–^DK11 (1, 2, 3)	Y2–W6 (2, 3)
			W6–H9 (3)
		S1–H14 (1)	L6–W9 (1)
		D1–H14 (2)	W6–I9 (2)
		H9–S13 (3)	Y2–L13 (2)
			I2–V13, I2–L6 (1)
			I9–L13 (2)
Inter-molecular	^12DNH–^14HCO (1)	D12–H14 (1)	W9–W9 (1)
	^14HNH–^12DCO (1)	E12–K14 (3)	W6–W6 (2, 3)
	^12SNH–^14HCO (2)		H9–H9 (3)
	^14HNH–^12SCO (2)
	^12ENH–^14KCO (3)		V13–V13 (1)
	^14KNH–^12ECO (3)		L6–L6 (1)
			I9–I9, L13–L13 (2)
Peptide–ligand	HNH^δ (3)–CO L		W9 (1)–L
			Y2/W6 (2)–L
			Y2/W6/H9 (3)–L
			I2/L6/V13 (1)–L
			I9/L13 (2)–L

---

P1: Ac-S₁-I₂-N₃-^DE₄-S₅-L₆-^DP₇-G₈-W₉-T₁₀-^DK₁₁-D₁₂-V₁₃-H₁₄-NH₂

P2: Ac-D₁-Y₂-N₃-^DE₄-S₅-W₆-^DP₇-G₈-I₉-T₁₀-^DK₁₁-S₁₂-L₁₃-H₁₄-NH₂

P3: Ac-N₁-Y₂-S₃-^DE₄-D₅-W₆-^DP₇-G₈-H₉-T₁₀-^DK₁₁-E₁₂-S₁₃-K₁₄-NH₂

Results and discussion

MS and NMR

P1, P2, and P3 were made by manual solid-phase synthesis²⁶ and were determined to be >87% pure by HPLC (Fig. S5 ESI†). Appearance of requisite peaks in the ESI-MS spectrum validated the structures (Fig. S6 ESI†). ¹H NMR spectra in 90% D₂O–H₂O mixture reflect requisite resonances for the aromatic-H, peptide-NH, and other specific protons expected (Fig. S7 ESI†). The spectra were assessed for effect of diluting the solutions from 5.0 to 0.5 mM concentration. Appreciable changes occurred in selective peptide-NHs, which argue for possible involvement of the NHs in assembly on the basis of inter-molecular hydrogen bonding. Selective aromatic-Hs in P2 and P3, but not in P1, were also affected by dilution, arguing for possible involvement of the aromatic structures with folding and/or self-assembly of the sequences (Fig. S8 ESI†). Considering other evidence that will be discussed, the assembly implied in the mM concentration regime with NMR may reflect a higher order aggregation of the sequence as a homodimer structure. This is no surprise considering the notoriety of the β-hairpin fold to be susceptible to aggregation given that several peptides of the main chain have unpaired hydrogen bonds.

Circular dichroism and fluorescence

The sequences were assessed for assembly in the 20–200 μM concentration range with CD and fluorescence spectroscopy. As noted in Fig. 2 panel A, all sequences were similar in CD at 20 μM concentration but were remarkably structure specific in the high concentration CDs. A coupled minimum and maximum of ellipticity appeared with concentration in P2 and P3 and possibly was a coupled exciton of the aromatic side chains in the sequences.^27,28 The exciton was noted to be clearly absent in P1, which implies that Tyr₂ was involved in producing the excitons in P2 and P3; presumably, the interacting side chains are brought close by concentration dependent folding and/or self-assembly of the sequences. The observed CD change in P1 (Fig. S9 ESI†) may reflect the ordering of the main chain in the sequence following its concentration promoted self-assembly. The high-concentration CD in this case has the deceptive appearance of an α-helix CD but has the minima (200 and 215 nm) at the wavelength lower than the requirement at ∼208 and ∼222 nm. Moreover, with its multiple D residues, including a Pro, P1 cannot adopt the α-helix fold due to an unfavorable steric effect of the D residues. The lowest concentration CDs being similar suggests that the different sequences may be unordered as monomers that fold with self-assembly cooperatively. Conformed to the possibility, the molar ellipticities were plotted against concentration at 225 nm for all the sequences (Fig. S10 ESI†). The protomer folds may be comparable in intermolecular binding strength, conformed to the largely similar structures of the sequences. Fluorescence anisotropy was sigmoidal with concentration and has the midpoints coincide for all the sequences, as noted in Fig. 2, panel B. The concentration dependent change of anisotropy conforms to an increase of mass, which directly evidences assembly of the structures. Thus, combined evidence of CD and fluorescence indicated that the designed sequences were unordered monomers that folded with self-assembly upon increasing the polypeptide concentration.

Fig. 2 Coupled exciton in CD spectrum for P3 (panel A) and the associated changes of ellipticity at 225 nm in P1, P2, and P3 (inset panel A) and concomitant increase of fluorescence anisotropy (panel B) are sigmoidal functions of concentration.

Dynamic light scattering

The sequences were assessed with DLS at 40 and 100 μM concentrations. At 40 μM concentration, a single peak of hydrodynamic radius ∼4 nm appears (Fig. 3). The size corresponds with the expectation from either the unordered monomer or the ordered dimer; the equilibrium could not be diagnosed on this basis. The observation, however, did rule out any higher order aggregates present at the given concentration. At 100 μM concentration, particles of hydrodynamic radius ∼15 and ∼100 nm appeared; higher-order aggregates of the dimer may have been involved due to the peptides of the main chain that are unpaired in hydrogen bonds. Upon evidence of NMR and DLS, the sequences appeared to aggregate at >100 μM concentration. Upon combining evidence of DLS, CD, and fluorescence, the sequences could be unordered monomers at <20 μM concentration. At intermediate concentration, the CD, fluorescence, and DLS evidences argue that unfolded monomer equilibrates with the folded dimer and at higher concentrations, they fold to form oligomers.

Fig. 3 Results of DLS showing hydrodynamic radius of P3 at two concentrations of the peptide.

Isothermal titration calorimetry and AutoDock

Ligand binding was evaluated with fluorescence quenching, Isothermal Titration Calorimetry (ITC), and AutoDock.²⁹ Titration with bNPC and bNPP promoted quenching of fluorescence of the sequences, as is noted in Fig. S11 ESI.† The quenching was stronger at 120 μM concentration of the sequences than at 20 μM concentration. Presumably, the polypeptide receptor orders with concentration due to its self-assembly as a homodimer structure. The quenching data from 120 μM concentration experiment gave non-linear Stern Volmer plots (Fig. 4 panel A; S12 panel A ESI†). Static and dynamic modes of quenching may be involved. Fits to a modified Stern–Volmer equation given under the Materials and methods section are linear as is noted in Fig. 4 panel B, and S12 panel B ESI.† According to the derived interaction parameters given in Table 2, bNPC and bNPP are comparable in binding strength with the receptor structure.

Fig. 4 Plots of fluorescence quenching data of P3 with specific binding ligands according to the Stern–Volmer equation and a modified form of the equation.

Table 2 Binding parameters assessed with quenching of fluorescence and calculated with AutoDock

Peptide	Ligand	Ka (×10^3 M^−1)	ΔG^0 fluorescence (kJ mol^−1)	ΔG^0 AutoDock (kJ mol^−1)
P1	bNPC	2.4	−19.29	−18.4 ± 0.33
	bNPP	7.4	−22.08	−21.2 ± 0.56
P2	bNPC	9.5	−22.68	−18.1 ± 0.28
	bNPP	8.7	−22.48	−22.5 ± 0.13
P3	bNPC	1.2	−17.51	−20.4 ± 0.33
	bNPP	4.6	−20.89	−22.7 ± 0.07

---

A model for ligated complex of the receptor structure recovered as a central member of the largest cluster prepared with MD was evaluated with AutoDock.²⁹ The calculated energies of ligand binding, given in Table 2, agree with the estimates from fluorescence quenching experiment. It is possible that the modeled homodimer is the receptor protein involved in binding the desired ligand. Interaction between P3 and bNPP was assessed with ITC. The results are given in Fig. 5. The interaction parameters derived by assuming 1 : 1 stoichiometry between the dimer structure and the ligand are given as an inset in Fig. 5. The observed binding strength compares well with the estimates from fluorescence quenching experiment and AutoDock calculation. The homodimer in P3 may be the receptor structure binding the ligand. According to ITC, ligand binding is endothermic, as it is positive in enthalpy. This may manifest a desolvation penalty in the phosphate group of bNPP. However, ligand binding is weak and may be driven by entropy of hydrophobic gain. The comparable interaction strength of diverse sequences with geometrically distinct carbonate and phosphate ligands supports the possibility that being driven by the hydrophobic effect, ligand binding is low in structural specificity.

Fig. 5 Heat changes upon addition of 10 μL aliquots of bNPP (12.5 mM) into a solution of P3 in 250 μM concentration (upper panel) and the integrated heat changes (lower panel). Fit for a single-site binding model is shown as a solid line. The calculated thermodynamic parameters are shown in the inset.

Molecular dynamics

The designs were verified with molecular dynamics using the GROMACS package^22,23 with a gromos-96 force field that was modified for treatment of D residues. Simulations initiated in a desired fold, assembly, and ligated complex were run for 50 ns each in a periodic box of SPC water.²⁴ Harvested at 10 ps intervals, the sequences were clustered to a 0.15 nm RMSD cutoff over Cα of polypeptide structure using the Duara et al. algorithm.²⁵ Thermodynamic stability of a fold, assembly, and complex was judged on the basis of the number of clusters to which the structure drifts and the size of the minimum-energy cluster taking it to represent an approximate statistical weight of the desired structure in the quasi Boltzmann-like ensemble modeled. The results in Fig. S1–S4 ESI† justify the designs, but also imply that the structures may be rather weak in ordering as the desired folds, assemblies, and ligated complexes.

Conclusion

The adaptational versatility of proteins, in their hierarchic construction plan, presumably characterizes their design challenge a well; a vast diversity of combinatorials is possible and the search for the desired structural hierarchy is a challenge. Protein design challenge thus became ​ greatly simplified when the chemical level of the natural hierarchy was optimized in an independent step while having the conformational and spatial levels kept frozen, as the fold and assembly of interest.^1–3 The possibility of having different levels of the natural hierarchy optimized in independent steps prompted us to explore the biomimetic algorithm to have proteins designed in steps and to have the structures extended with inclusion of stereochemistry as an added variable of sequence. Exploring the design algorithm and amino acid stereochemistry as the added variable of sequence two successful designs were accomplished: a small C5 structure was accomplished in mimicry of the much larger natural protein shiogatoxin⁴ and a small receptor structure of C2 symmetry was now accomplished in this study in mimicry of the much larger natural protein HIV protease.

Experimental section

Fmoc-protected amino acids, reagents for solid phase synthesis, Rink amide AM resin, solvents, bis-nitrophenylcarbonate, and bis-nitrophenylphosphate were purchased from Sigma-Aldrich or Novabiochem-Merck.

Peptide synthesis

Synthesis was performed manually on a Rink amide AM resin using standard Fmoc chemistry and HOBt/DIC as the coupling reagents. Each coupling, monitored with Kaiser and chloranil tests, typically required about 6 hours. Deprotections were carried out with 30% (v/v) piperidine–DMF. N-terminus was acetylated (–NHCOCH₃) with Ac₂O : DIPEA : DMF in a 1 : 2 : 20 ratio. The cleavage of the final polypeptide and deprotection of the side chains were achieved with reagent K (82.5% TFA/5% dry-phenol/5% thioanisole/2.5% ethandithiol/5% water). The product precipitated with anhydrous diethyl ether was lyophilized from 1 : 4H₂O : ^tBuOH solution as a white powder. The final purification of the peptide was accomplished with HPLC over RP-18 (10 μm, 10 mm × 250 mm; Merck) eluted with ACN\H₂O\0.1%TFA\0–100% gradient.

ESI-MS

Mass spectra were obtained on a QTOF-ESI Mass Spectrometer. Positive ions were detected in linear/reflectron mode.

Nuclear magnetic resonance

NMR experiments were performed on a Bruker 800 MHz instrument equipped with a cryoprobe at 298 K in 9 : 1H₂O and D₂O with 0.1 mM 2,2-dimethyl-2-silapentane-5-sulfonate sodium salt (DSS) as an internal reference at 5 mM concentration. ¹H NMR data were processed with the TOPSPIN software of Bruker.

Circular dichroism (CD)

CD measurements were performed on a JASCO J-815 CD spectropolarimeter calibrated with d₁₀-camphoursulphonic acid. Data were collected at 298 K in a 0.2 cm path-length quartz cell with a 2 nm bandwidth in the 195–240 nm range. Scanning was carried out at 100 nm min⁻¹ with a 1.0 s time constant in 1 nm steps; five scans were averaged after baseline correction for the solvent. Peptides were prepared in the 20–200 μM range by optical measurements in an aqueous solution of water–methanol mixture at pH 7.0. The observations in millidegrees were converted to the mean residue ellipticity using a reported relation.³⁰

Fluorescence

Fluorescence measurements were taken on a Varian Cary Eclipse fluorimeter. Data were collected at 298 K in a 1 cm cell with a 280 nm excitation wavelength. The excitation and emission slits were 5 nm and scanning speed was 120 nm min⁻¹. The peptide solutions were prepared in the 10–100 μM concentration range by optical measurements of the peptide and of the standard NATA (SIGMA chemicals). Binding experiments were carried out at 20 and 120 μM concentration of the peptide and 100 μM concentrations of bNPC and bNPP ligands. The binding of the ligands with the peptide dimers was accompanied by quenching of Trp fluorescence. The fluorescence quenching data were analyzed with the Stern–Volmer equation as follows:

F₀/F = 1 + K_SV[Q]; (1)

where F₀ and F are the fluorescence intensities observed in absence and presence of quencher, respectively, [Q] is quencher concentration and K_SV is the Stern–Volmer quenching constant. In the event of nonlinear Stern Volmer plots, both static and dynamic quenching could be involved according to the following relation:

F₀/F = (1 + K_SV[Q])(1 + K_a[Q]); (2)

where the modified ‘Modified Stern–Volmer’ equation is as follows:

F₀/F₀ − F = (1/f_a) + (1/f_aK_a[Q]); (3)

where f_a is the fraction of fluorophore accessible to the quencher and K_a is the association constant of the quencher to the fluorophore. F₀/F₀ − F, against the reciprocal of ligand concentration [Q], is a linear equation that gives the association constant K_a and binding free energy calculated as follows:

ΔG = −RT ln K_a (4)

Fluorescence anisotropy

Steady state fluorescence anisotropy measurements were recorded with a Varian Cary Eclipse fluorimeter equipped with an emission-excitation polarizer. All the experiments are performed at 25 °C.

Dynamic light scattering

The size-distributions of the peptide samples were assessed by DLS using a Nano-ZS (Nanoseries) Malvern Instrument with a 35 mW diode laser source. The measurements were carried out at 25 °C and involved a 90° scattering angle. The samples were prepared in water using a 0.22 m filter with four hours of sonication. Processing of the fluctuating signal with a digital autocorrelator gave the particle diffusion coefficient and particle size with the Stokes–Einstein equation. The resulting correlation function was processed to a size-distribution for the particles using Malvern Software.

Isothermal titration calorimetry (ITC)

The experiments were performed on a VP-ITC micro calorimeter (Microcal, Inc.) at 298 K. The sample cell contained peptide in 250 μM concentration, as determined by an optical measurement, and the reference cell contained water. The 12.5 mM bis-nitrophenyl-phosphate (ligand) solution loaded into a 250 mL syringe was titrated into peptide solution in 10 μL aliquots in 25 steps at 4 min intervals. The change in enthalpy (DH) due to dilution was determined by titrating ligand into solvent as well as solvent into peptide solution. These backgrounds were subtracted from DH obtained for the corresponding ligand-peptide binding experiments prior to curve fitting. The background-subtracted data were fitted to a model describing a single binding site using the MicroCal software. The binding enthalpy (DH), entropy (DS), and association (K_a) constant were thus calculated.

Computational methods

Inverse algorithm for sequence design. The inverse-design package IDeAS implements a combination of a deterministic pruning algorithm dead end elimination and a stochastic search method Monte Carlo. Energy functions for a sequence solution involved an accessibility-based solvation energy, H-bond energy, coulomb energy, and entropy in the form of probability of side chain usage. The software uses a natural database of rotamers and their suitable symmetry transforms as the rotamers of the D structure. IDeAS has been validated in other examples of protein design.^31,32

Molecular dynamics. MD was performed using gromos-96-43a1 force field in the GROMACS package. Simulations were under NVT with a periodic boundary condition in a 4.8 nm cubic box for the monomer and a 5.5 nm box for the dimer. The SPC water model was used. Bond lengths were constrained with SHAKE.³³ A 2 femtosecond time step was used. Non-bonded interactions were cut-off at 1.4 nm using the shift function. The potential energy of the system, i.e., peptide and water, was minimized using the steepest descent algorithm with a 100 kJ mol⁻¹ nm⁻¹ tolerance limit with a sufficient number of steps to achieve convergence. Position restrained molecular dynamics was performed for 200 ps at 298 K to equilibrate the water molecules. Initial velocities were drawn from a Maxwellian distribution. The peptide and solvent were separately coupled to a Berendsen temperature bath using a 0.1 ps time constant. Conformational clustering was performed to a RMSD cut-off ≤0.15 nm for the Ca atoms.

Molecular docking. A flexible docking algorithm was implemented with AutoDock 4.0. Central member of the top cluster populating the MD trajectory over aromatic residues comprising the biding site was used as the receptor structure for calculating the binding energy of the receptor and its ligand. Using a genetic algorithm for ligand docking and an RMSD tolerance of 2 Å, structurally distinct conformational clusters of the ligand were ranked in increasing energy.

Acknowledgements

We acknowledge the DST (SR/S1/OC-74/2008) Government of India and the Department of Chemistry, IIT Bombay for the financial support and supercomputing facility “Corona”.

References

1. D. N. Bolon and S. L. Mayo, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 14274.
2. J. B. Siegel, A. Zanghellini, H. M. Lovick, G. Kiss, A. R. Lambert, J. L. St. Clair, J. L. Gallaher, D. Hilvert, M. H. Gleb, B. L. Stoddard, K. N. Houk, F. E. Michael and D. Baker, Science, 2010, 329, 309.
3. H. K. Privett, G. Kiss, T. M. Lee, R. Blomburg, R. A. Chica, L. M. Thomas, D. Hilvert, K. N. Hour and S. L. Mayo, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 3790.
4. P. Ghosh, A. U. Mushtaq and S. Durani, RSC Adv., 2012, 2, 3242.
5. S. Durani, Acc. Chem. Res., 2008, 41, 1301–1308.
6. S. Rana, B. Kundu and S. Durani, Chem. Commun., 2004, 21, 2462.
7. S. Rana, B. Kundu and S. Durani, Chem. Commun., 2005, 2, 207.
8. S. Rana, B. Kundu and S. Durani, Bioorg. Med. Chem., 2007, 15, 3874.
9. S. Rana, B. Kundu and S. Durani, Biopolymers, 2007, 87, 231.
10. J. Erickson, D. J. Neidhart, J. VanDrie, D. Kempf, X. C. Wang and D. W. Norbeck, Science, 1990, 249, 527.
11. D. S. Duhovny, Y. Inbar, R. Nussinov and H. J. Wolfson, Nucleic Acids Res., 2005, 33, 363.
12. S. M. Noble, V. E. Carnahan, L. B. Moore, T. Luntz, H. Wang, O. R. Ittoop, J. B. Stimmel, P. R. Davis-Searles, R. E. Watkins, G. B. Wisely, E. LeCluyse, A. Tripathy, D. P. McDonnell and M. R. Redinbo, Biochemistry, 2006, 45, 8579.
13. P. B. Stranges, M. Machius, M. J. Miley, A. Tripathy and B. Kuhlman, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 20562.
14. R. L. Dunbrack, Curr. Opin. Struct. Biol., 2002, 12, 431.
15. S. C. Lovell, J. M. Word, J. S. Richardson and D. C. Richardson, Proteins, 2000, 40, 389.
16. R. F. Goldstein, Biophys. J., 1994, 66, 1335.
17. N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller and E. J. Teller, Chem. Phys., 1953, 21, 1087.
18. D. B. Gordon, S. A. Marshall and S. L. Mayo, Curr. Opin. Struct. Biol., 1999, 9, 509.
19. B. I. Dahiyat and S. L. Mayo, Protein Sci., 1996, 5, 895.
20. D. N. Bolon, J. S. Marcus, S. A. Ross and S. L. Mayo, J. Mol. Biol., 2003, 329, 611.
21. R. Ranbhor, A. Tendulkar, A. Kumar, V. Ramakrishnan, K. Patel, K. R. Srivastava and S. Durani, unpublished.
22. E. Lindahl, B. Hess and D. van der Spoel, J. Mol. Model., 2001, 7, 306.
23. W. F. van Gunsteren, S. R. Billeter, A. A. Eising, P. H. Hunenberger, P. Kruger, A. E. Mark, W. R. P. Scott and I. G. Tironi, Biomolecular Simulation: The GROMOS96 Manual and User Guide, Hochschulverlag AG an der ETH Zurich, Zurich, Switzerland, 1996.
24. H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, J. Hermans and B. Pullman, Interaction models for water in relation to protein hydration, in Intermolecular Forces, Reidel Publishing Company, The Netherlands, 1981, p. 331.
25. X. Duara, W. F. van Gunsteren and A. E. Mark, Proteins: Struct., Funct., Genet., 1999, 34, 269.
26. W. C. Chan and P. D. White, Fmoc Solid Phase Peptide Synthesis: A Practical Approach, IRL Press Oxford, UK, 1989.
27. M. L. Waters, Curr. Opin. Chem. Biol., 2002, 6, 736.
28. C. A. Hunter, K. R. Lawson, J. Perkins and C. J. Urch, J. Chem. Soc., 2001, 2, 651.
29. G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner, R. K. Belew, D. S. Goodsell and A. J. Olson, J. Comput. Chem., 2009, 30, 2785.
30. G. D. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum Press, New York, 1996.
31. K. Patel, B. Goyal, A. Kumar, N. Kishore and S. Durani, J. Phys. Chem. B, 2010, 114, 16887.
32. K. Patel, K. R. Srivastava and S. Durani, Bioorg. Med. Chem., 2010, 18, 8270.
33. J. P. Ryckaert, G. Ciccoti and H. J. C. Berendsen, J. Comput. Phys., 1977, 23, 327.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra17350g

---