Limits...
An estimate of the numbers and density of low-energy structures (or decoys) in the conformational landscape of proteins.

Vadivel K, Namasivayam G - PLoS ONE (2009)

Bottom Line: We show that the number of native-like structures for a polypeptide is, in general, an exponential function of its sequence length.The density of these structures in conformational space remains more or less constant and all the increase appears to come from an expansion in the volume of the space.These results are consistent with earlier reports that were based on other models and techniques.

View Article: PubMed Central - PubMed

Affiliation: Centre of Advanced Study in Crystallography & Biophysics, University of Madras, Tamilnadu, India.

ABSTRACT

Background: The conformational energy landscape of a protein, as calculated by known potential energy functions, has several minima, and one of these corresponds to its native structure. It is however difficult to comprehensively estimate the actual numbers of low energy structures (or decoys), the relationships between them, and how the numbers scale with the size of the protein.

Methodology: We have developed an algorithm to rapidly and efficiently identify the low energy conformers of oligo peptides by using mutually orthogonal Latin squares to sample the potential energy hyper surface. Using this algorithm, and the ECEPP/3 potential function, we have made an exhaustive enumeration of the low-energy structures of peptides of different lengths, and have extrapolated these results to larger polypeptides.

Conclusions and significance: We show that the number of native-like structures for a polypeptide is, in general, an exponential function of its sequence length. The density of these structures in conformational space remains more or less constant and all the increase appears to come from an expansion in the volume of the space. These results are consistent with earlier reports that were based on other models and techniques.

Show MeSH
Comparison with experimental structures.The best matched, i.e. lowest rmsd structure (black) superposed on the respective crystal structure (white) for each peptide. Only the backbone atoms are shown.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2663821&req=5

pone-0005148-g003: Comparison with experimental structures.The best matched, i.e. lowest rmsd structure (black) superposed on the respective crystal structure (white) for each peptide. Only the backbone atoms are shown.

Mentions: Finally, Table 1 (and Figure 3) gives a comparison of the generated structures with the respective experimental structure. For each peptide, some of the former are accurate replicas of the latter, and have very low values of the rmsd in atomic positions on superposition. This is true for the loop sequences as well, though the force field did not use any information about the flanking sequences, or about the interactions the residues in the loop make with the rest of the protein or with the atoms of the solvent. In addition, as we have discussed elsewhere [42], the MOLS search also identifies other low energy-structures observed by other techniques, both experimental and theoretical. For example, in the case of the neuropeptide Met-enkephalin the sample of 1500 structures contained the global energy minimum as revealed by other calculations [45]–[47], besides the structures seen in experiments [48]. Again these facts support our contention that, despite the relatively small number of structures generated, MOLS sampling covers conformational space thoroughly. In general, the energy values of the structures closest to the experimental results are not the lowest of all the generated structures. However they are within about 25 kcal/mol of the latter, with no short contacts or other unphysical interactions.


An estimate of the numbers and density of low-energy structures (or decoys) in the conformational landscape of proteins.

Vadivel K, Namasivayam G - PLoS ONE (2009)

Comparison with experimental structures.The best matched, i.e. lowest rmsd structure (black) superposed on the respective crystal structure (white) for each peptide. Only the backbone atoms are shown.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2663821&req=5

pone-0005148-g003: Comparison with experimental structures.The best matched, i.e. lowest rmsd structure (black) superposed on the respective crystal structure (white) for each peptide. Only the backbone atoms are shown.
Mentions: Finally, Table 1 (and Figure 3) gives a comparison of the generated structures with the respective experimental structure. For each peptide, some of the former are accurate replicas of the latter, and have very low values of the rmsd in atomic positions on superposition. This is true for the loop sequences as well, though the force field did not use any information about the flanking sequences, or about the interactions the residues in the loop make with the rest of the protein or with the atoms of the solvent. In addition, as we have discussed elsewhere [42], the MOLS search also identifies other low energy-structures observed by other techniques, both experimental and theoretical. For example, in the case of the neuropeptide Met-enkephalin the sample of 1500 structures contained the global energy minimum as revealed by other calculations [45]–[47], besides the structures seen in experiments [48]. Again these facts support our contention that, despite the relatively small number of structures generated, MOLS sampling covers conformational space thoroughly. In general, the energy values of the structures closest to the experimental results are not the lowest of all the generated structures. However they are within about 25 kcal/mol of the latter, with no short contacts or other unphysical interactions.

Bottom Line: We show that the number of native-like structures for a polypeptide is, in general, an exponential function of its sequence length.The density of these structures in conformational space remains more or less constant and all the increase appears to come from an expansion in the volume of the space.These results are consistent with earlier reports that were based on other models and techniques.

View Article: PubMed Central - PubMed

Affiliation: Centre of Advanced Study in Crystallography & Biophysics, University of Madras, Tamilnadu, India.

ABSTRACT

Background: The conformational energy landscape of a protein, as calculated by known potential energy functions, has several minima, and one of these corresponds to its native structure. It is however difficult to comprehensively estimate the actual numbers of low energy structures (or decoys), the relationships between them, and how the numbers scale with the size of the protein.

Methodology: We have developed an algorithm to rapidly and efficiently identify the low energy conformers of oligo peptides by using mutually orthogonal Latin squares to sample the potential energy hyper surface. Using this algorithm, and the ECEPP/3 potential function, we have made an exhaustive enumeration of the low-energy structures of peptides of different lengths, and have extrapolated these results to larger polypeptides.

Conclusions and significance: We show that the number of native-like structures for a polypeptide is, in general, an exponential function of its sequence length. The density of these structures in conformational space remains more or less constant and all the increase appears to come from an expansion in the volume of the space. These results are consistent with earlier reports that were based on other models and techniques.

Show MeSH