Simulations of Biotoxins and of PDZ Domains
X. Chen and Y. Deng
The toxin produced by the bacterium Clostridium botulinum is one of
the deadliest known [1]. Recently, a great deal of information about its
mode of action has become available through structural and other biophysical
studies [2]. The toxin itself (BoNT) is a protein of approximately 1300
residues, with two chains, a light chain (LC) of approximately 50 kDa and a
heavy chain (HC) of approximately 100 kDa. The two are linked by a disulfide
bond. The LC contains a Zn2+ ion, which acts within the cell to cleave a
protein necessary for neurotransmission leading to paralysis and death. This
process is believed to involve several steps: binding of the toxin to the
endosomal membrane, translocation through the lipid bilayer, and proteolysis
of specific neurotransmitters within the cell. Structural studies show that
the HC contains two domains responsible for binding and translocation, while
the LC contains the catalytic domain. Both the structure of the toxin
outside the cell (but not at endosomal pH) and the mechanism of action of
the zinc protease are reasonably well understood. But, the structures at low
pH and the mechanism of the translocation remain illusive.
The results of 0.2 micro-seconds simulated time for two different
temperatures and pH values are shown in Figure 1. We also investigated
the coupling between the histidine residues with the zinc ion as in Figure
2.
| Figure 1. Comparison
of the RMSD for the whole protein. |
Figure 2.
Structures of LC in LC only run at different pH and 37oC,
Red: pH 4.7; Blue: pH 7.0 |
It is well known that the pH dependence of these structures is not fully
described by the traditional method [3]. The protonation state is a function
of pH, but the change with pH depends on the protonation state, requiring a
self-consistent solution. A recent new approach combining molecular dynamics
and Monte Carlo simulations has been developed [4]. We have implemented this
method with the AMBER 8 [5] package to verify our results [6] .
We also simulated the disheveled PDZ domain with ligands [7]. The Disheveled
(Dvl) PDZ domain is believed to play an essential role in the canonical and
noncanonical Wnt signaling pathways, which are involved in embryo
development as well as in tumorigenesis. Also, it binds directly to Frizzled
(Fz) receptors. An organic molecule (NSC668036) from the National Cancer
Institute small-molecule library has been identified to be able to bind to
the Dvl PDZ domain [8]. We developed molecular dynamic simulations to
analyze the binding between them and analyzed their interactions in details.
Interesting phenomena were revealed by our simulations. As in Figure 3, we
found the convergence of the 30 structures of the NSC668036. The figure uses
colors to represent the relative RMSD of different structures after 10ns
simulation. Since the structures have zero RMSD when comparing with itself,
the diagonal line of the matrix is zero. An island with small numbers in the
matrix means that the relevant structures have similar conformation, and
thus, tend to converge to a similar structure.
|
  |
|
| Figure 3. Final state of PDZ domain after 10
ns simulation. |
Figure 4. RMSD of PDZ domain and ligand #10. |
Interesting phenomena were revealed by our simulations. As in Figure 4, we
found the convergence of the 30 structures of the NSC668036. The figure uses
colors to represent the relative RMSD of different structures after 10ns
simulation. Since the structures have zero RMSD when comparing with itself,
the diagonal line of the matrix is zero. An island with small numbers in the
matrix means that the relevant structures have similar conformation, and
thus, tend to converge to a similar structure.
References
-
[1] Singh, B.R. Nature Structural Biology 7: 617-619 (2000).
-
[2] Chen, X. and Deng, Y. J. Molecular Modeling. Conditionally accepted,
2006.
-
[3] Smith, L.J. et al. Proteins 36: 77 (1999).
-
[4] Mongan, J. et al. J. Comp. Chem. 25: 2038 (2004).
-
[5] Pearlman, D.A. et al. Comp. Phys. Comm. 91: 1-41 (1995).
-
[6] Chen, Y., Chen, X., and Deng, Y. J. Comp. Phys. To be submitted,
2006.
-
[7] Chen, X. and Deng, Y. J. Molecular Modeling. To be submitted, 2006.
-
[8] Shan, J. et al. Biochemistry 44(47): 15495-15503 (2005).
Last Modified: January 31, 2008
Please forward all questions about this site to:
Claire Lamberti
|
