Modeling Peptide Structures Based On NMR Data
M. Madan Babu, Center for Biotechnology, Anna University
Introduction:
In the present study molecular model building for a synthetic, designed decapeptide is considered, using distance information from the nuclear Overhauser effect (nOe). Nuclear Magnetic Resonance provides a powerful method for determination of polypeptide structures. Interproton distances, typically in the order of 4.5 A are accurately measured using the nuclear Overhauser effect (nOe) data. The intensity of the nOe between a pair of protons can be related to the distance of separation between the proton pair.
Noe measurement thus, generate information about spatial proximity of protons in a molecule. These spatial information can be used to generate models of peptides giving rise to structures comparable with defined 3-Dimensional structure of polypeptides in solution. It is generally observed that such models are very close to the structure determined by X-Ray crystallography in the crystalline state
The sequence is : BOC - M - L - F -V - DP - A - L - V - V - F - OMe. The peptide is designed to adopt a b- Hairpin conformation nucleated by a DP-A turn segment.
Derivation of distance and angle restraints :
Assignment of distance limits for pairs of protons in a molecule is based on observation of nuclear Overhauser effect between the protons. Individual protons in the 500 MHz 1H NMR spectra of peptide 1 were assigned using a combination of two 2-dimensional experiments:
1. TOCSY - Total Correlation Spectroscopy
2. ROESY - Rotating frame Overhauser effect spectroscopy
A list of chemical shift assignments for the protons in the peptide 1 is given in the Table 1. NOe cross peaks in the ROESY spectra (appendix 1.1) were classified as strong , medium and weak by visual inspection of the contour levels. The following distance limits were assigned to these nOe intensities:
1. Strong ( 1.8 to 2.5 Å )
2. Medium ( 2.5 to 3.5 Å )
3. Weak intensity (3.5 to 4.5 Å )
Vicinal spin-spin coupling constants involving NH and Ca H protons provide a measure of dihedral angle f . In this study, 3Ja HNH coupling constants greater than 8.5 Hz were used as additional constraints for the structural model building. Table 2 lists the complete set of nOe and dihedral restraints used in the calculations of the structure. In these calculations, intra residue nOes were not applied. In a separate investigation hydrogen bonding restraints connecting the donors and the acceptors were also included as inputs for the calculations.These Hydrogen bonding interaction were experimentally obtained from the variable temperature experiments which was done from 280K to 315K in steps of 5K. It was noted that the amide protons of the residues Leu2, Val4, Leu7, and Val9, suffered considerably less chemical shift with temperature, thus can be inferred that there exists a Hydrogen bonding pattern in these residues.
Structure Calculation:
16 distance restraints and 6 angle restraints were used for a molecular dynamics structure calculation as schematized below :
Start with an extended structure
|
Minimize using restraints
|
Start Molecular dynamics with the minimizedstructure as the starting structure.
|
20 picoseconds of Equilibration
|
80 picoseconds of dynamics simulation.
|
Collect trajectories every picosecond for further analysis.
The structure calculations in the present study was started with an initial extended conformation of the peptide with the f, y values of 180o, 180o, except at the DPro5, where the f was kept at 45o . During the minimization and dynamics , the forcefield used was AMBER ( available in the Insight II software, Biosym Inc ® .), Omega values for residues Met1 to Val4 and Ala6 to Phe10 was forced to 180o (trans peptide bonds ) using a force constant of 5.0 kcal/mol/rad2 . The omega of DPro was not forced to 180o for we wanted the observed nOes to determine the trans/cis geometry of the peptide bond linking DPro5 and Val4. It should be emphasized here that peptides containing proline residues have propensity to adopt cis rotamer at the Xxx - Pro peptide bond.
The starting extended structure was subjected to 2500 iterations of Steepest descent minimization, followed by 2500 iterations of Conjugate Gradient minimization with the set of distance and dihedral restraints listed in table 2. The convergence of minimization was followed until the RMS Derivative was less than 0.01 KCal/Mol. In another study the same minimization protocol was carried out with the distance restraints, Hydrogen bonding restraints and angle restraints, available in table 2.1, 2.2 and 2.3
The energy minimized structure was then taken as the starting structure for the molecular dynamics simulation. A total simulation time of 100 picoseconds was carried out with the following protocol:
1. The system was allowed to equilibrate at 300K for 20 picoseconds
2. The dynamics simulation was continued for 80 picoseconds after the equilibration period, keeping the temperature constant at 300K.
Molecular dynamics trajectories were stored at every picosecond after the equilibriation period. The distance and dihedral force constants were kept at 25.00 KCal/Mol/Å and 50.0 kcal/mol/rad2 all through the simulation. The resulting molecular conformations at different time points during the simulation were evaluated for their total energy. 15 lowest energy conformations were selected for further analysis of experimental restraints violations. 13 structures out of this set did not violate the distance restraints more than 0.2 Å ( a program was written during the course of this work to check distance violation , available in the appendix 1.1), and hence were selected for final structural analysis.
Results and Discussion :
Structure calculations for the peptide using observed nOe restraints and dihedral restraints have resulted in a closely related family of beta hairpin conformations . Fig 1.1 illustrates the superimposition of 13 best structures obtained using the procedure described above. The RMS deviation between the backbone atoms of the structures with respect to an average structure ( Fig 1.2 , calculated using MOLMOL ) is 0.40 Å and 0.96 Å between all the heavy atoms. All these superimposable conformations have the f ,y values ( calculated using PSW , a program written by Prof. C. Ramakrishnan), which lie within the allowed region of the Ramachandran map ( Fig 1.3 ) demonstrating the stereochemical quality of these calculated conformations. b -Hairpin conformations in peptides are defined by two extended anti-parallel strands linked by a short loop ( 2-4 residues ). In the family of structures observed for this peptide, residues 1 to 4 and residues 7 to 10 have f ,y values populating the extended region of the ramachandran map. Table 3 lists the dihedral angles at all residues in the mean structure. Clearly, the turn segment in this b hairpin conformation consists of DPro - Ala residues. As expected f Dpro is limited to 60o +/- 20o. The Ramachandran angles at the turn residues in are ( 59.-112 ) and ( -69, 32 ) which are diagnostic of a type II ' b turns.
Hydrogen bonding analysis in all the calculated structures were evaluated for hydrogen bonding interactions using HBPLUS. The donor and acceptor atoms identified are given in table 2.3. A beta hairpin of 10 residues with a two residue turn is expected to have 4 intramolecular Hydrogen bonds between residues 2-9 and 4-7. In all this calculated structures, these expected Hydrogen bonds were observed. It must be noted here that the restraint set did not include Hydrogen bonding interactions. As demonstrated here, the structures determined using nOe derived distance restraints are sufficiently well defined to have proper hydrogen bonding interactions When this structure calculations were repeated with a modified restraint inputs, which includes Hydrogen Bonding interactions ( appendix 1.1 ), similar b hairpin conformations were observed for the peptide. Fig 1.4 illustrates the superposition of two average conformations
This structure calculations procedure with modified restraint input which includes Hydrogen bonding interactions was repeated to obtain conformations.
Super imposition of 13 best structures after analysis
Appendix
1.0 Program for analysis
of trajectories
2.0 Program for
generating restraint file for Insight
3.0 NMR Spectra