pubmed.ncbi.nlm.nih.gov

An evolutionary strategy for all-atom folding of the 60-amino-acid bacterial ribosomal protein l20 - PubMed

️Sun Jan 01 2006

An evolutionary strategy for all-atom folding of the 60-amino-acid bacterial ribosomal protein l20

A Schug et al. Biophys J. 2006.

Abstract

We have investigated an evolutionary algorithm for de novo all-atom folding of the bacterial ribosomal protein L20. We report results of two simulations that converge to near-native conformations of this 60-amino-acid, four-helix protein. We observe a steady increase of "native content" in both simulated ensembles and a large number of near-native conformations in their final populations. We argue that these structures represent a significant fraction of the low-energy metastable conformations, which characterize the folding funnel of this protein. These data validate our all-atom free-energy force field PFF01 for tertiary structure prediction of a previously inaccessible structural family of proteins. We also compare folding simulations of the evolutionary algorithm with the basin-hopping technique for the Trp-cage protein. We find that the evolutionary algorithm generates a dynamic memory in the simulated population, which leads to faster overall convergence.

PubMed Disclaimer

Figures

**FIGURE 1**
Overlay of the folded and the native conformation of the bacterial ribosomal protein L20 in simulations A and B (*upper* and *lower*, respectively) with the corresponding C_β-C_β matrices. The upper triangle of the C_β-C_β matrix shows absolute, the lower relative deviations between the folded and the experimental structure, respectively. Each square encodes the deviation between the C_β-C_β distance of two amino acids in the experimental structure to the C_β-C_β distance of the same amino acids in the folded structure. Black (*gray*) squares, deviation of <1.50 Å (2.25 Å); white squares, large deviations.

**FIGURE 2**
(*Top*) Average and minimal energy (*upper part*) and average RMSB deviations (*lower part*) as a function of iteration number for simulations A and B with N = 50. (*Bottom*) Native score in the phases with N = 266 and N = 50 of both simulations versus the number of function evaluations (*solid line*, simulation A; *dotted line*, simulation B).

**FIGURE 3**
Energy (*upper*) and RMSB deviation (*lower*) of the best decoy in the final population of simulation B as a function of iteration number, indicating a continuous convergence of the simulation toward the native conformation.

**FIGURE 4**
Color-coded distance matrix of the final conformations of simulations A (*top*) and B (*bottom*). In each panel, the upper right (lower left) triangle encodes the backbone (full) RMSB deviation between the members of the population. The top row and leftmost column in each figure show the native conformation. Blue/green (1- to 4-Å range), similar structures; red (deviations of 8–10 Å), large deviations. The conformations are sorted by energy, starting with the best from the top.

**FIGURE 5**
Overlay of the native and the energetically best decoys in the simulation B native family: B1, B4, B6, B8, B10, and nonnative family: B2, B3, B5, B7, B9. The first substantially different decoy (B40) is shown in the bottom row.

**FIGURE 6**
Chart of the decision-making process, when a newly generated conformation (with energy E_new) is presented to the master process. The worst matching conformation in the active population is replaced by the new conformation, if the latter differs from all present conformations and is lower in energy than the lowest conformation. If there are similar conformations, the closest (by RMSB) is replaced, if its energy is higher.

**FIGURE 7**
Minimal (*red*) and average energies (*black*) in kcal/mol for the basin-hopping simulations (*dashed lines*) and the evolutionary algorithm (*solid lines*) as a function of the numerical effort per population member (in thousands of function evaluations).

Cited by

Simplified protein models: predicting folding pathways and structure using amino acid sequences.
Adhikari AN, Freed KF, Sosnick TR. Adhikari AN, et al. Phys Rev Lett. 2013 Jul 12;111(2):028103. doi: 10.1103/PhysRevLett.111.028103. Epub 2013 Jul 11. Phys Rev Lett. 2013. PMID: 23889448 Free PMC article.
Mutations as trapdoors to two competing native conformations of the Rop-dimer.
Schug A, Whitford PC, Levy Y, Onuchic JN. Schug A, et al. Proc Natl Acad Sci U S A. 2007 Nov 6;104(45):17674-9. doi: 10.1073/pnas.0706077104. Epub 2007 Oct 29. Proc Natl Acad Sci U S A. 2007. PMID: 17968016 Free PMC article.
A free-energy approach for all-atom protein simulation.
Verma A, Wenzel W. Verma A, et al. Biophys J. 2009 May 6;96(9):3483-94. doi: 10.1016/j.bpj.2008.12.3921. Biophys J. 2009. PMID: 19413955 Free PMC article.
Protein structure prediction by all-atom free-energy refinement.
Verma A, Wenzel W. Verma A, et al. BMC Struct Biol. 2007 Mar 19;7:12. doi: 10.1186/1472-6807-7-12. BMC Struct Biol. 2007. PMID: 17371594 Free PMC article.

References

1. Baker, D., and A. Sali. 2001. Protein structure prediction and structural genomics. Science. 294:93–96. - PubMed
1. Schonbrunn, J., W. J. Wedemeyer, and D. Baker. 2002. Protein structure prediction in 2002. Curr. Opin. Struct. Biol. 12:348–352. - PubMed
1. Liwo, A., P. Arlukowicz, C. Czaplewski, S. Oldizeij, J. Pillardy, and H. Scheraga. 2002. A method for optimising potential energy functions by a hierarchichal design of the potential energy landscape. Proc. Natl. Acad. Sci. U.S.A. 99:1937–1942. - PMC - PubMed
1. Moult, J., K. Fidelis, A. Zemia, and T. Hubbard. 2001. Critical assessment of methods of protein structure (CASP): round IV. Proteins. 45:2–7. - PubMed
1. Snow, C. D., H. Nguyen, V. S. Pande, and M. Gruebele. 2002. Absolute comparison of simulated and experimental protein folding dynamics. Nature. 420:102–106. - PubMed

An evolutionary strategy for all-atom folding of the 60-amino-acid bacterial ribosomal protein l20 - PubMed