Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Feb 1:16:33.
doi: 10.1186/s12859-015-0471-x.

PCalign: a method to quantify physicochemical similarity of protein-protein interfaces

Shanshan Cheng  1 , Yang Zhang  2   3 , Charles L Brooks 3rd  4   5   6
Affiliations

PCalign: a method to quantify physicochemical similarity of protein-protein interfaces

Shanshan Cheng et al. BMC Bioinformatics. .

Abstract

Background: Structural comparison of protein-protein interfaces provides valuable insights into the functional relationship between proteins, which may not solely arise from shared evolutionary origin. A few methods that exist for such comparative studies have focused on structural models determined at atomic resolution, and may miss out interesting patterns present in large macromolecular complexes that are typically solved by low-resolution techniques.

Results: We developed a coarse-grained method, PCalign, to quantitatively evaluate physicochemical similarities between a given pair of protein-protein interfaces. This method uses an order-independent algorithm, geometric hashing, to superimpose the backbone atoms of a given pair of interfaces, and provides a normalized scoring function, PC-score, to account for the extent of overlap in terms of both geometric and chemical characteristics. We demonstrate that PCalign outperforms existing methods, and additionally facilitates comparative studies across models of different resolutions, which are not accommodated by existing methods. Furthermore, we illustrate potential application of our method to recognize interesting biological relationships masked by apparent lack of structural similarity.

Conclusions: PCalign is a useful method in recognizing shared chemical and spatial patterns among protein-protein interfaces. It outperforms existing methods for high-quality data, and additionally facilitates comparison across structural models with different levels of details with proven robustness against noise.

PubMed Disclaimer

Figures

Figure 1
Figure 1
The ROC curves for predicting highly related interfaces using three methods, PCalign, Ialign, and I2I-SiteEngine. As shown by the red and magenta curves, our method PCalign gives an AUC value of 0.970, and for the backbone set 0.955. In comparison, Ialign gives an AUC of 0.980. I2I-SiteEngine performs slightly worse, with those predicted by match score, total score and t-score having AUC values of 0.831, 0.884 and 0.909 respectively.
Figure 2
Figure 2
Recognition of interface similarity across unrelated interfaces by PCalign and Ialign. The comparison is based on two geometric criteria; fraction of aligned residues (coverage) and RMSD of aligned residues. (A) All-against-all pairwise comparison, with PCalign (Ialign) aligning on average 53.1 ± 13.3% (51.4 ± 13.8%) of residues with RMSD of 3.725 ± 0.371 Å (3.810 ± 0.473 Å). (B) All-against-all pairwise comparison, with PCalign (Ialign) aligning on average 54.8 ± 13.2% (51.4 ± 13.8%) of residues with RMSD of 3.686 ± 0.378 Å (3.810 ± 0.473 Å), where the chemical term in PCalign is turned off to capture geometric similarity only. (C) Closest unrelated interface in the set of 124 dimers, with PCalign (Ialign) aligning on average 68.4 ± 14.5% (68.3 ± 15.6%) of residues with RMSD of 3.483 ± 0.366 Å (3.563 ± 0.502 Å). (D) Closest unrelated interface in the set of 124 dimers, with PCalign (Ialign) aligning on average 70.1 ± 15.4% (68.3 ± 15.6%) of residues with RMSD of 3.466 ± 0.371 Å (3.563 ± 0.502 Å), considering the geometric part of the scoring function in PCalign only. In all scenarios, PCalign does slightly better than Ialign in recognizing geometric similarities across unrelated interfaces, and using a scoring function that considers both chemical and geometric properties in PC-score performs less well compared to using one that considers purely geometric properties in PC-score, due to the fact that this analysis uses purely geometric criteria.
Figure 3
Figure 3
Further performance comparison between PCalign and Ialign. When a method finds higher coverage with lower RMSD for a particular pair of interfaces compared, it is considered better for that case; with lower coverage and higher RMSD, it is considered worse. All cases of PCalign (plotted in green) outperforming Ialign (plotted in orange) and Ialign outperforming PCalign in the 185136 pairs compared are shown in the scatter plots and summarized in the bar plots. (A) In 50838 cases, PCalign is better than Ialign, when the chemical term in PC-score is turned on. (B) Ialign outperforms PCalign in 34087 cases in comparison. (C) PCalign has an odds ratio of 1.5 in finding a better structural alignment than Ialign. (D) When only geometrical property is considered in the scoring function, PCalign outperforms Ialign in 57639 cases. (E) Ialign is better than PCalign in 27790 cases. (F) PCalign outperforms Ialign with an odds ratio of 2.1 if only the physical environment of protein-protein interfaces is considered.
Figure 4
Figure 4
Three examples of viral mimicry resulting from convergent evolution. The first example is that of the M3 protein mimicking CCL2 in complexing with another CCL2 monomer (ABC), the second being the V protein competing with DDB2 in binding with DDB1 (DEF), and the third case being the G protein targeting the ephrin B2 ligand in similar ways with its native ephrin type-B receptor 4 (GHI). They are shown with the two complexes superimposed (ADG), with a focused view of the matched interfacial residues (BEH), and with just one binding site on the viral protein and that on the host protein it mimics (CFI). In all illustrations the viral protein is colored in blue, and the host protein it displaces is colored in cyan. The human target protein is colored red when bound with the viral protein, and orange when complexed with its cognate binding partner. The small spheres represent the Cα positions of all the interfacial residues present in the original complex, while the large spheres represent those which are structurally equivalent in the virus-host protein complex and in the endogenous complex. Figures are generated by the VMD software [35].
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Nooren IM, Thornton JM. Diversity of protein-protein interactions. EMBO J. 2003;22(14):3486–3492. doi: 10.1093/emboj/cdg359. - DOI - PMC - PubMed
    1. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, et al. The Protein Data Bank. Nucleic Acids Res. 2000;28(1):235–242. doi: 10.1093/nar/28.1.235. - DOI - PMC - PubMed
    1. Rossmann MG, Morais MC, Leiman PG, Zhang W. Combining X-ray crystallography and electron microscopy. Structure. 2005;13(3):355–362. doi: 10.1016/j.str.2005.01.005. - DOI - PMC - PubMed
    1. Bahadur RP, Chakrabarti P, Rodier F, Janin J. A dissection of specific and non-specific protein-protein interfaces. J Mol Biol. 2004;336(4):943–955. doi: 10.1016/j.jmb.2003.12.073. - DOI - PubMed
    1. Jones S, Thornton JM. Principles of protein-protein interactions. Proc Natl Acad Sci U S A. 1996;93(1):13–20. doi: 10.1073/pnas.93.1.13. - DOI - PMC - PubMed

Publication types

LinkOut - more resources