Frequency of errors in MCCE1.0 calculations Calculated vs experimental pKas


The data presented here represents the prediction of the ionization states of more than 35000 residues. There is a significant level of uncertainty in the pKas in this implementation of MCCE (MCCE1). Continuum electrostatics calculations tend to overestimate the interactions between surface charged groups.(1) Benchmark pKa studies on small proteins using no conformational sampling find better match to experimental data with dielectric constants of 10 to 20 reducing the interactions.(1) In contrast, MCCE calculations find good matches to benchmarks with dielectric constants of 4 to 8. The version used, here (MCCE1) has been benchmarked in calculations of known amino acid pKas (2) and Ems of hemes in small cytochromes.(3) pKa calculations of 166 ionizable residues in 12 small soluable proteins such as lysozyme and barnase showed root mean square deviation between the calculated and measured of 0.83 pH units with >90% having errors of < 1 pH unit using e=4.(2) However, this was carried out using a SOFT function that decreased very strong interactions by about 50%. This is needed because MCCE1 does not provide enough well solvated conformers for surface conformers. However, as all conformers are present in all calculations of pair-wise interactions adding more surface conformers increases the region with a low dielectric constant artificially increasing the interactions. This is not a problem for the calculations of dsolv here since these were determined in a calculation with only one conformer per residue (see (2) for a more complete description).

While ionization states for both surface and buried residues are presented here, the focus is on the ionization state of the buried charged groups. Analyses using MCCE and other continuum electrostatics methods of buried active sites in photosynthetic reaction centers(4-6), bacteriorhodopsin(7, 8), halorhodopsin,(9) and fumarate reductase(10) provide the best benchmarks for the analysis of buried charges. These sites are far from the surface so boundary errors are negligible. Calculations with large dielectric constants or with SOFT do not find good agreement with the data. The value of e=4, used here provides the best results. Thus, earlier calculations with MCCE1 suggest that the ionization states and even pKas of buried residues are likely to be well predicted.(6, 8) The predicted pKas without SOFT for surface residues, especially those in ion pairs are likely to overestimate the stability of the ionized form, often pushing acid pKas to be less than zero and bases to move to be over 14.(2) However, the ionization states at pH 7 are likely to be well estimated. A new version of the program (MCCE2) has recently become available which has much better rotamer sampling and significantly corrects the problems that arise from errors in the boundary. This provides a consistently good match to buried and surface residues using the same assumptions and parameters. Calculation of pKas for the set of proteins examined here is in progress.


1. Garcia-Moreno, E. B., and Fitch, C. A. (2004) Structural interpretation of pH and salt-dependent processes in proteins with computational methods, Methods Enzymology 380, 20-51.
2. Georgescu, R. E., Alexov, E. G., and Gunner, M. R. (2002) Combining conformational flexibility and continuum electrostatics for calculating pKa's in proteins, Biophys J. 83, 1731-1748.
3. Mao, J., Hauser, K., and Gunner, M. R. (2003) How cytochromes with different folds control heme redox potentials, Biochemistry 42, 9829-40.
4. Alexov, E. G., and Gunner, M. R. (1999) Calculated protein and proton motions coupled to electron transfer: electron transfer from QA- to QB in bacterial photosynthetic reaction centers, Biochemistry 38, 8253-70.
5. Rabenstein, B., Ullmann, G. M., and Knapp, E. W. (2000) Electron transfer between the quinones in the photosynthetic reaction center and its coupling to conformational changes, Biochemistry 39, 10487-96.
6. Zhu, Z., and Gunner, M. R. (2005) Energetics of quinone-dependent electron and proton transfers in Rhodobacter sphaeroides photosynthetic reaction centers, Biochemistry 44, 82-96.
7. Spassov, V. Z., Luecke, H., Gerwert, K., and Bashford, D. (2001) pKa alculations suggest storage of an excess proton in a hydrogen- bonded water network in bacteriorhodopsin, J Mol Biol 312, 203-19.
8. Song, Y., Mao, J., and Gunner, M. R. (2003) Calculation of proton transfers in Bacteriorhodopsin bR and M intermediates, Biochemistry 42, 9875-88.
9. Song, Y., and Gunner, M. R. (2005) Halorhodopsin and bacteriorhodopsin: comparing the mechanism of a Cl and a proton pump, manuscript in preperation.
10. Haas, A. H., and Lancaster, C. R. (2004) Calculated coupling of transmembrane electron and proton transfer in dihemic quinol:fumarate reductase, Biophys. J. 87, 4298-4315.