Dmitri Klimov, Professor
School of Systems Biology

           Curriculum Vitae | Contact

Research Interests
BINF690 Numerical Methods in Bioinformatics
BINF740 Introduction to Biophysics
BINF739/BIOL691 Molecular Modeling for Biologists
BINF741 Introduction to Computer Simulations of Biomolecules



Animations of protein folding/unfolding and aggregation of Alzheimer's Abeta peptides

Visualization of force-induced unfolding for lattice protein models

Trajectories for force-induced unfolding Monte Carlo simulations are obtained at constant force pulling the protein ends apart. The simulation temperature Ts (<TF) is selected according to the condition that at zero force the native state be stable. The selected force value Fs is large enough to make the stretched state thermodynamically dominant, when force is "on". The simulations start with the native conformation and performed until the rod-like conformations are achieved. In the absence of force this 36-mer sequence folds thermodynamically and kinetically by two-state mechanism. Monte Carlo simulations show that when force is applied unfolding begins with fast extension of sequence termini, while the hydrophobic core retains its structure. The major event in unfolding process is the cooperative unraveling of the sequence core, which proceeds via transient distinct conformational steps.

  1. unfolding trajectory 002
  2. unfolding trajectory 003

Visualization of molecular dynamics simulations of stretching of two-state beta-barrel

Molecular dynamics simulations of tension-induced unfolding of 46-mer beta-barrel is performed at constant force Fs and the temperature Ts, at which the folded state is stable in the absence of force. The initial conformations are equilibrated at Ts and Fs=0. The constant pulling force is selected according to the condition that completely extended state attains the lower energy than the native state when force is applied.

These simulations illustrate the close link between the sequence topology and the nature of unfolding dynamics. Both terminal strands are rich in hydrophobic residues and buried in the native structure, and display strong hydrophobic interactions between each other and the rest of the structure. As a consequence, the unfolding takes place very cooperatively without any discernible intermediates. Therefore, the unfolding of terminal strands is the rate-limiting step.

To view the animations for few generic unfolding trajectories follow these links:

  1. unfolding trajectory 001
  2. unfolding trajectory 003

Visualization of beta-hairpin folding

Our off-lattice model for 16-mer C-terminal hairpin from GB1 protein reproduces at quantitative level most experimenatal results. The model includes side chains, hydrogen bonding, and realistic amino acid interactions. The calculated native conformation devaites from the one deposited in PDB by only 2.5A. The folding kinetics has been studied using Langevin dynamics at T=0.82TF, at which the thermal average of the fraction of native contacts is 0.7. The folding, which is two-state under these conditions, involves three structural stages: (i) collapse and formation of hydrophobic cluster; (ii) zipping of hydrogen bonds starting with the turn and formation of native backbone; (iii) native side chain packing. The animations show two generic trajectories, which start with the random unfolding structures and last until the first passage to the native state.

  1. folding trajectory 435
  2. folding trajectory 654

Molecular dynamics simulations of the assembly of Abeta16-22 oligomers

In order to advance our understanding of the molecular basis of Alzheimer's disease we initiated molecular dynamics (MD) studies of aggregation of fragments of Abeta peptides. Our goal is (i) to obtain detailed kinetic picture of the assembly of Abeta peptides, (ii) to establish the set of interpeptide interactions, which drive the process of oligomerization, and (iii) to assess the role of sequence and external factors (such as chemical denaturants). For the first time the formation of Abeta oligomers was simulated starting with random initial conditions, without any "prebuilt" bias or structure. Antiparallel orientation of Abeta16-22 peptides in oligomers is preferred, because it maximizes favorable electrostatic and hydrophobic interpeptide interactions.

The fragment of MD trajectory visualizes the final stages of formation of an antiparallel registry of two Abeta16-22 peptides. The antiparallel docking is anchored by electrostatic contacts between oppositely charged terminal side chains (shown in blue and red) and is further stabilized by non-specific "glue" of hydrophobic interactions (hydrophobic residues are shown in green). The timescale, on which such ordered oligomers form, is approximately 10 nanoseconds. For clarity, the third peptide and water solvating the oligomer is not shown.

  1. MD trajectory of the assembly of Abeta16-22 oligomer