• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of narLink to Publisher's site
Nucleic Acids Res. Jul 1, 2005; 33(Web Server issue): W372–W375.
Published online Jun 27, 2005. doi:  10.1093/nar/gki365
PMCID: PMC1160126

PCE: web tools to compute protein continuum electrostatics

Abstract

PCE (protein continuum electrostatics) is an online service for protein electrostatic computations presently based on the MEAD (macroscopic electrostatics with atomic detail) package initially developed by D. Bashford [(2004) Front Biosci., 9, 1082–1099]. This computer method uses a macroscopic electrostatic model for the calculation of protein electrostatic properties, such as pKa values of titratable groups and electrostatic potentials. The MEAD package generates electrostatic energies via finite difference solution to the Poisson–Boltzmann equation. Users submit a PDB file and PCE returns potentials and pKa values as well as color (static or animated) figures displaying electrostatic potentials mapped on the molecular surface. This service is intended to facilitate electrostatics analyses of proteins and thereby broaden the accessibility to continuum electrostatics to the biological community. PCE can be accessed at http://bioserv.rpbs.jussieu.fr/PCE.

INTRODUCTION

Electrostatic interactions play a central role in protein structure and function (13). However, in spite of their importance, it is still difficult to treat electrostatic properties theoretically. One reason is that Coulomb's law is valid for charges immersed in an infinite medium of a uniform dielectric (4). Yet, many reactions take place at the interface between regions having different dielectric values. Analytical solution of the Poisson–Boltzmann equation can resolve this problem only in several particular cases. Thus, the numerical solution of the Poisson–Boltzmann equation, like the Finite Difference Poisson–Boltzmann method (FDPB) (5), appears to be a good model for describing macromolecules in water solutions (6,7). Some of these methods are implemented at: http://agave.wustl.edu/pdb2pqr/ (8), http://agave.wustl.edu/apbs/ (9) and http://honiglab.cpmc.columbia.edu/ (10).

The basic electrostatic property that can be calculated via PCE (protein continuum electrostatics) is the electrostatic potential, yet continuum electrostatic models have also proven to be useful for the calculation of pKa in proteins. We have implemented on the PCE server (http://bioserv.rpbs.jussieu.fr/PCE) the MEAD (macroscopic electrostatics with atomic detail) package based on the FDPB method for the computation of electrostatic potential values and pH titration behavior of titratable groups in proteins (11). In this approach, the protein is modeled as a low dielectric material (dielectric usually between 2.0 and 4.0) with embedded partial charges surrounded by a high dielectric medium (the solvent is regarded as a continuum with a dielectric value around 80.0), the atomic details of the protein structure are used to define the boundary and charge placement and the FDPB method is used to solve the resulting partial differential equations for the electrostatic potential.

METHODS AND IMPLEMENTATION

The PCE web service is driven by a modular, Python-written collection of routines, which provides non-interactive, high-throughput usage of the package MEAD. MEAD is based on the FDPB method to compute electrostatic potential values and pH titration behavior of titratable groups in proteins (11).

In the MEAD package, the electrostatic interactions are calculated with a continuum dielectric model where the protein and the water phase are represented as homogenous materials with a low (by default εin = 4) and high (by default εout = 80) dielectric constant ε, respectively. The dielectric boundary between protein and solvent is the molecular surface, which depends on the atomic radii and the radius of a solvent-sized spherical probe. The Poisson–Boltzmann equation is solved by means of a finite difference algorithm (5,1113). The user can use the default values for ε or, for example, use εin = εout (in this case the Poisson–Boltzmann equation will be resolved in a homogenous dielectric medium) or evaluate the effect of changing εin from 2 to 10 or 20 to simulate protein flexibility (14).

Electrostatic potentials calculations

In our implementation, the electrostatic potentials are computed by the program ‘potential’ of MEAD starting from the coordinate file. Our implementation adds hydrogen atoms (15) and assigns the Parse parameters (radii and charges) (16). In addition, it is possible to upload PQR files instead of PDB files, such format parameterized files can be obtained, e.g. at http://agave.wustl.edu/pdb2pqr/ (8) (at this site, users can also perform hydrogen-bond network optimization). The finite difference computation uses a cubic box with 1 Å lattice spacing with number of grid points depending on the protein size and centered on the protein.

Calculations of pKas of titratable groups

The overall methodology for calculating pKa, defined as the pH for which the average protonation of the group is 1/2, has been described elsewhere (12). This approach implies the calculation of the intrinsic pKa values of all titratable groups (defined as the pKa of one titratable group if all other titratable groups in the protein were held in their neutral form) as well as the pair-wise electrostatic interactions between the titratable groups:

  1. The intrinsic pKa is calculated by pKint = pKmod − (ΔΔGBorn + ΔΔGback)/2.3RT, where the pKmod is the pKa of model compound of amino acid in water determined experimentally. In the MEAD implementation, the pKmod values are as follows: Asp, 4.0; Glu, 4.4; His, 6.6 for the Nε2 or 7.0 for Nδ1; Tyr, 9.6; Lys, 10.4; Arg, 12.0; C-terminus, 4.0; and N-terminus, 7.5. The correction ΔΔGBorn takes into account the Born self-energy owing to the transfer of the amino acid group from water (high dielectric medium) to the protein (low dielectric medium). The correction ΔΔGback is the background interaction term owing to the interaction of the titratable group with non-titratable (permanent) charges in the protein. In our implementation, all permanent partial charges on all non-titratable groups are taken into consideration.
  2. The pair-wise charge interactions between the titratable groups can be represented as a matrix, Wij, where i and j label the interacting groups.
  3. As a result, the pKa of one titratable group i is represented by pKai=pKint,i+12qiqjWij, where qi and qj are the pH-dependent charges of the titratable groups i and j, respectively. All electrostatic energies (Wij, ΔΔGBorn and ΔΔGback) can be calculated by solving the Poisson–Boltzmann equation for the electrostatic potential generated by the protonated and deprotonated forms of each titratable group placed in protein and in water. For these computations, our implementation applies the MEAD program ‘multiflex’ to solve the electrostatic potentials by the FDPB method using two successive lattices with increasing grid resolutions, 1 and 0.25 Å, respectively.
  4. As seen in (ii), the charges qi and qj of the titratable groups at any given pH have to be additionally computed. Proteins can contain several tens to hundreds of titratable groups, thus leading to the problem of multiple site titration for the computation of pKa values. The exact calculation of a multiple site titration curve, given the intrinsic pKas and the pair-wise charge interactions, could be carried out at any given pH by a Boltzmann-weighted average over all of the possible protonation states of the protein (12). However, such calculations grow exponentially with the number of titratable groups (2N combinations for N titratable groups). In the MEAD package, this problem is overcome by introducing the Reduced Site approximation method (17). This approach assumes that a titratable group can be considered as non-titratable at a pH far away from its pKmod. In this case, this group can be assumed to be non-titratable for this pH, and the number of possible combinations to compute is reduced to 2N−1. Thus, our service proposes at present pKa calculations for <50 titratable groups. The computed titration curves, in our implementation, give a pH value where the protonation state of the titratable groups is 0.5, which is assigned to be the pK1/2 and can be interpreted as the pKa of this group.

IMPLEMENTATION

Potential and pKa calculations are presented as two distinct services. Both of them start from a coordinate file in the PDB Format (18). Users can either specify a PDB identifier or upload their own file (with or without hydrogens). The protein internal dielectric, solvent dielectric as well as ionic strength values can be adjusted interactively.

For the electrostatic potential calculations, it is in addition possible to specify coordinates at which the value of the potentials will be returned. Finally, the electrostatic potential values can then be used to color-code the molecular surface of the selected protein via DINO (Visualizing Structural Biology, http://www.dino3d.org). Some options allow to adjust the color gradients for positive and negative potential values or to specify a site in the protein that is of importance such that this residue becomes positioned at the center of the picture. Returned images are either static (four orthogonal views are proposed) or animated (rotation around the y-axis) in the GIF.

For the pKa calculations, it is presently only possible to activate Cys and Tyr as titratable, Asp, Glu, Lys, Arg and His being always considered as titratable. This is likely to evolve in future versions of the interface but was presently retained so as to lower the risk of getting improper results.

RESULTS

Electrostatic potentials

The users can download as results, color figures (in different formats, different orientations and resolutions together with animated GIF files) of 3D electrostatic potentials distribution on the protein surface. The users can select the energy range for the color-coding as well as decide on which region to focus the figure. In addition, the PQR file can be obtained. If the users specify a set of x, y, z coordinates, the server can return electrostatic potential values at these points.

As an example, if the users select lysozyme (PDB code 7lyz) as input file, they can obtain figures as follows (Figure 1). Such computations can be used to compare electrostatic potentials for wild-type proteins and mutants, in such cases the users need to upload their mutant PDB files.

Figure 1
Electrostatic potential distributions on the lysozyme surface from −3.0 kcal/mol/e (red) to +3.0 kcal/mol/e (blue).

pKas of titratable groups

The PCE server returns pKint, pK1/2, titration curves for all titratable groups and pair-wise charge interactions between titratable groups as well as the isoelectric point (pI). We have tested the implementation on several proteins taken from the PDB (18) and report the results for lysozyme (PDB entry 7lyz) and ribonuclease A (PDB entry 3rn3). Our computed pKa values (εin = 4) are compared with pKa values previously calculated or experimentally determined (Tables 1 and and2)2) (12,14). Small deviations between our data and the previously calculated pK1/2 are owing to the differences in some parameters in the finite difference calculations. Overall, a good correspondence between our pK1/2 and the experimental pKa values can be seen, except for few titratable groups. This seems to be due, in part, to the fact that proteins should have considerable mobility for the solvent-exposed side chains (19). These fluctuations can be simulated by using a higher εin as proposed in (14) or by computing pKa values on conformations selected from molecular dynamic trajectories (20). Other methods have been developed to account for protein flexibility during pKa calculations (2123).

Table 1
pKa calculation for lysozyme
Table 2
pKa calculation for ribonuclease

CONCLUSION AND FUTURE DIRECTIONS

PCE is presently based on MEAD and allows users to compute protein electrostatic potentials via a FDPB solver starting with a PDB file as query input. The electrostatic potential values can then be used to color-code the molecular surface of the molecules presently generated within DINO. Static images in different formats and with different resolutions, centered on specific residues, as well as animated GIF files can be downloaded. The PARSE parameters are used for the computations, but we plan to implement other forcefield parameters in a near future. Currently, non-protein or nucleic acid atoms are excluded from the calculations, but work is in progress to circumvent this limitation. Additional optimizations are in progress as we plan to define automatically the charge assignments depending on the calculated pKa values and then compute electrostatic potentials taking into account such information.

Users can also compute pKint, pK1/2, titration curves for all titratable groups and pair-wise charge interactions between titratable groups as well as isoelectric point of the protein. At present, no more than 50 titratable sites can be considered. However, we are working on the implementation of a Monte Carlo method to allow pKa computations for larger proteins.

Acknowledgments

This work was supported by the Inserm Institute via a ‘poste-vert’ to M.M. and the Avenir grant to B.V. RPBS was supported by ‘Action Conjointe CNRS-INSERM, Bioinformatique 2002’. Funding to pay the Open Access publication charges for this article was provided by INSERM.

Conflict of interest statement. None declared.

REFERENCES

1. Honig B., Nicholls A. Classical electrostatics in biology and chemistry. Science. 1995;268:1144–1149. [PubMed]
2. Warshel A., Aqvist J. Electrostatic energy and macromolecular function. Annu. Rev. Biophys. Biophys. Chem. 1991;20:267–298. [PubMed]
3. Bashford D. Macroscopic electrostatic models for protonation states in proteins. Front Biosci. 2004;9:1082–1099. [PubMed]
4. Sharp K.A., Honig B. Electrostatic interactions in macromolecules: theory and applications. Annu. Rev. Biophys. Biophys. Chem. 1990;19:301–332. [PubMed]
5. Warwicker J., Watson H.C. Calculation of the electric potential in the active site cleft due to alpha-helix dipoles. J. Mol. Biol. 1982;157:671–679. [PubMed]
6. Koumanov A., Ruterjans H., Karshikoff A. Continuum electrostatic analysis of irregular ionization and proton allocation in proteins. Proteins. 2002;46:85–96. [PubMed]
7. Rocchia W., Sridharan S., Nicholls A., Alexov E., Chiabrera A., Honig B. Rapid grid-based construction of the molecular surface and the use of induced surface charge to calculate reaction field energies: applications to the molecular systems and geometric objects. J. Comput. Chem. 2002;23:128–137. [PubMed]
8. Dolinsky T.J., Nielsen J.E., McCammon J.A., Baker N.A. PDB2PQR: an automated pipeline for the setup of Poisson–Boltzmann electrostatics calculations. Nucleic Acids Res. 2004;32:W665–W667. [PMC free article] [PubMed]
9. Baker N.A., Sept D., Joseph S., Holst M.J., McCammon J.A. Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl Acad. Sci. USA. 2001;98:10037–10041. [PMC free article] [PubMed]
10. Nayal M., Hitz B.C., Honig B. GRASS: a server for the graphical representation and analysis of structures. Protein Sci. 1999;8:676–679. [PMC free article] [PubMed]
11. Bashford D., Gerwert K. Electrostatic calculations of the pKa values of ionizable groups in bacteriorhodopsin. J. Mol. Biol. 1992;224:473–486. [PubMed]
12. Bashford D., Karplus M. pKa's of ionizable groups in proteins: atomic detail from a continuum electrostatic model. Biochemistry. 1990;29:10219–10225. [PubMed]
13. Gilson M.K., Honig B. Calculation of the total electrostatic energy of a macromolecular system: solvation energies, binding energies, and conformational analysis. Proteins. 1988;4:7–18. [PubMed]
14. Antosiewicz J., McCammon J.A., Gilson M.K. The determinants of pKas in proteins. Biochemistry. 1996;35:7819–7833. [PubMed]
15. Gautier R., Camproux A.C., Tuffery P. SCit: web tools for protein side chain conformation analysis. Nucleic Acids Res. 2004;32:W508–W511. [PMC free article] [PubMed]
16. Sitkoff D., Sharp K.A., Honig B. Accurate calculation of hydration free energies using macroscopic solvent models. J. Phys. Chem. 1994;98:1978–1988.
17. Bashford D., Karplus M. Multiple-site titration curves of proteins. J. Phys. Chem. 1991;95:9556–9561.
18. Berman H.M., Westbrook J., Feng Z., Gilliland G., Bhat T.N., Weissig H., Shindyalov I.N., Bourne P.E. The protein data bank. Nucleic Acids Res. 2000;28:235–242. [PMC free article] [PubMed]
19. Smith L.J., Sutcliffe M.J., Redfield C., Dobson C.M. Analysis of phi and chi 1 torsion angles for hen lysozyme in solution from 1H NMR spin–spin coupling constants. Biochemistry. 1991;30:986–996. [PubMed]
20. Miteva M.A., Brugge J.M., Rosing J., Nicolaes G.A., Villoutreix B.O. Theoretical and experimental study of the D2194G mutation in the C2 domain of coagulation factor V. Biophys. J. 2004;86:488–498. [PMC free article] [PubMed]
21. You T.J., Bashford D. Conformation and hydrogen ion titration of proteins: a continuum electrostatic model with conformational flexibility. Biophys. J. 1995;69:1721–1733. [PMC free article] [PubMed]
22. Georgescu R.E., Alexov E., Gunner M.R. Combining conformational flexibility and continuum electrostatics for calculating pK(a)s in proteins. Biophys. J. 2002;83:1731–1748. [PMC free article] [PubMed]
23. Alexov E. Role of the protein side-chain fluctuations on the strength of pair-wise electrostatic interactions: comparing experimental with computed pK(a)s. Proteins. 2003;50:94–103. [PubMed]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • PubMed
    PubMed
    PubMed citations for these articles
  • Substance
    Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...