BioComputing

   Bluues - Electrostatic properties of proteins based on generalized Born radii


Quick Help and References

Description
The Poisson-Boltzmann (PB) equation and its linear approximation have been widely used to describe biomolecular electrostatics [3]. Generalized Born (GB) models offer a convenient computational approximation for the more fundamental approach based on the Poisson-Boltzmann equation, and allows estimation of pairwise contributions to electrostatic effects in the molecular context.

This server concerns the implementation of the most common features of the electrostatic properties of proteins. The server first computes generalized Born radii, via a surface integral and then it uses generalized Born radii (using a finite radius test particle) to perform electrostic analyses. The output of the server can be summarized as follows:
  • Generalized Born radius of each atom
  • Electrostatic solvation free energy
  • pH-dependent properties
  • pKa of all ionizable groups
  • Electrostatic potential at the surface of the molecule


The results obtained are comparable to those obtained using state-of-the-art Poisson-Boltzmann solvers but this method but this method may be more efficient because it depends on the accuracy required and on the size of the system. Moreover, the algorithm produces Born radii which allow calculation of pairwise interactions, an information not directly available in the Poisson-Boltzmann based methods. The server is designed for easy usage with output in the form of easy comprehensible graphs, sorted tables and molecular viewers in JMol.

A more detailed description of the output will be provided later.
Input
E-Mail address
This is optional but if your do not intend in waiting for the job to finish or you want to store your process we highly recommend using an email.

Job tag for structure
An optional title for your submission. This will appear in the header of the output. We suggest you select one, especially if sending multiple queries, as they may be completed in a different order.

Structure

This is where you must provide information about your structure. There are 3 possibilities for providing the structure:
  • Entering the 4 letter PDB ID. For example, entering simply 9ANT will execute bluues on chains A, B, C, D, E, F for this PDB file. In addition, one can provide the chain ID in order to single out the calculation for a particular chain. PDB information will run the script PDB2PQR.py with the CHARMM forcefield [4]. For NMR PDB files the first model is only considered.
  • Upload a file in PDB format. The file must conform to the PDB specification. PDB information will run the script PDB2PQR.py with the CHARMM forcefield [4]. For NMR PDB files the first model is only considered.
  • Upload a file in PQR format. This file is a user defined file (generated from a PDB file) which was converted to the PQR file format [4].

Type of computation
Generalized Born (GB) radii and Potential at atomic surface
Both GB radii and Potential at atomic surface are executed by default since they are the most basic type of computation and they are efficient.
The computation of GB radii is performed by surface integrals. GB models provide efficient and accurate calculations for the electrostatic effects of the solvent [5-7]. The radii allow the calculation of interaction energy between any two charges and from a purely geometric point of view they give a measure of the atom depth in the structure.
The Potential at atomic surface is computed as the energy of interaction all atoms of the molecule with a unit test charge with a generalized Born radius equal to half the solvent probe radius, (i.e. 0.7Å by default). The potential is greatly influenced by partial or net charges on the molecule, but also by molecular shape. For more detailed descriptions see the bluues methods paper in the reference section.

pka shifts
This option will generate pH-dependent properties such as total charge, pH-dependent free energy of folding, and pKa of ionizable groups as well as titration curves for each titratable residue. Here we follow the approach of [8] with the advantage that the Green's functions are obtained directly within the GB approach whereas they require multiple calculations in the PB approach.In bluues methods paper(see reference section) the results were compared with the commonly used Propka2.0 [10] showing comparable (or better on the chosen dataset) performances.

Surface area
The solvent accessible surface area (i.e. the surface accessible by the center of a solvent probe sphere) is computed based on a grid mapping of all atoms. This is provided for completeness and/or further use. GB radii are computed based on this surface.

Parameters
  1. Solvent probe radius (Å): the radius used to define the solvent accessible surface. Default: 1.4 (Å) for water.
  2. Minimum atomic radius (Å): radii smaller than this value are reset to this value. E.g. for some forcefield hydrogens may receive very small van der Waals radii and this value resets them to the chosen value. Default: 1.0 Å
  3. Temperature (Kelvin): Default: 298.15K (i.e. 25 C)
  4. Salt radius (Å): the average radius of the ionic species. Default: 2.0 Å.
  5. Inner dielectric constant: the dielectric constant of the protein [9]. Default: 4.
  6. Ionic strength: The solution ionic strength. Default 0.150 M for physiological conditions.
Computation time
Using the surface area and pKa options:
  • Large protein: 523 amino acids, 10212 atoms takes approximately 28 mins.
  • Medium sized protein: 156 amino acids, 2486 atoms takes approximately 3 mins.
  • Small protein: 26 amino acids, 374 atoms takes approximately 1-2 mins.
  • NOTE: Currently we are in the process of upgrading our infrastructure which should improve computational time dramatically.
    Output
    Protein information
    Number of residues in protein : ℕ Number of atoms in protein : ℕ Number of chains in parent protein : ℕ Number of models : ℕ
    Available files
    Default options:
    1. Radii for each atom (*.gbr). A text file containing the columns:
      N  ATOM  RES RESN CH.     GBR4      GBR5      GBR6      GBR7      GBR8      GBR9
      
      where N: atom number, ATOM: atom identifier, RES: residue adentifier, RESN: residue number, CH.: chain identifier, GBRN: where N: 4,..,9 are the GB radii computed using the corresponding models. GBR6 is used for all subsequent calculations.
    2. Potential for each atom (*.srfatpot). A PDB file which can be loaded into most molecular viewers. The temperature column (B-factor) was replaced by the calculated surface potential (kJ/(mol q) for the protein.

    If Surface Area clicked:

    1. Residues and atoms accessible to solvent (*.area). A text file containing the area in A squared for the system, the single chains, the single residues and the single atoms.
    2. List of surface points (*.srf). A PDB file viewable in most molecular viewers (under the dot representation) where each surface point is treated as a PDB atom.
    3. List of surface points, surface normals and corresponding atom (*.at_r_n). A text file containing the number of the exposed atom, the surface point coordinates and the vector normal to the surface.

    If pKa clicked:

    1. Titration info for residues (*.titration). A text file with the following columns:
      ATOM  RES RESN CH.     pH  ionization
      
      where ATOM: the atom identifier, RES: the residue identifier, RESN: the residue number, pH: the simulated pH and ionization: the simulated charge on the atom.
    2. All titratable sites with contributions and other information (*.pka). A text file with columns:
      ATOM  RES RESN CH.     pKa    pKa_0 dpKa^self dpKa^bg  dpKa^ii    GBR6  Solv. exp.
      
      where ATOM: the atom identifier, RES: the residue identifier, RESN: the residue number in the PDB file, CH.: the chain identifier, pKa: the calculated pKa of the titratable atom,
      pKa_0: the model compound pKa, dpKa^self: the shift in pKa due to desolvation, dpKa^bg: the shift in pKa due to the interaction with other charges in the molecule with all titratable sites in their neutral state, dpKa^ii: the shift in pKa due to the interaction between titratable sites, GBR6: the GB radius according to the GBR6 model (a measure of the depth of the titratable site in the molecular structure), Solv. exp.: the solvent accessible surface area of the titratable site.
      Note that titration data are post-processed for computing the pKa, therefore titration data do not match necessarily the outputted pKa.
    3. pH-dependent total charge and the contribution to the free energy of folding versus the pH (*..ddg). A text file with the following columns:
      pH    DDG (kJ/mol) charge: the calculated charge at this pH
      
      where pH: is the simulated pH, DDG: pH-dependent folding free energy and charge: the calculated charge at this pH
    4. Folding energy vs. pH (*.ddg.ph_ddg.pdf) A graph of the information in (3) above.
    5. Charge vs. pH (*.ddg.ph_charge.pdf) A graph of the information in (3) above.
    Visualisation
    The molecular view and its surface potential is displayed using the JMOL visualisation tool.
    surface example
    Figure 1. The JMOL interface for the calculated surface potential. Quick buttons are available for coloring, movements and molecular representation. However, more detailed options for the JMOL viewer can be accessed by right clicking the protein.

    Folding free energy and charge. To complement the surface potential. Two images/graphs showing the folding free energy and charge as they vary against pH are shown beneath the JMOL visualisation ( these can also be viewed in higher resolution PDF's).
    DDG example
    charge example
    Figure 2. Quick overview of the folding free energy (top) and charge (bottom) as they vary against the simulate pH.

    Sorted tables. In order to find the highest and lowest values within our calculations tables sorted by important characteristics are produced by the server:
    1. Atoms sorted by Generalized Born Radii. This allows a quick analysis of the most buried and exposed atoms in the protein.
    2. Atoms sorted by surface potential. This allows a quick analysis of the most potential, positive or negative, atoms in the protein.
    3. GBR table
      Figure 3. Example sorted table for Generalized Born Radii. From this example it is clear that the most buried atom is 825 HE1 belonging to residue PHE 49 of chain A.
    4. Titratable residues sorted by pKa. pKa values are indicators of protein stability and activity of the protein and therefore may give evidence of protein function.
    5. GBR table
      Figure 4. Example sorted table for surface potential. From this example it is clear that the most positive potential is atom 516 HH11 belonging to residue ARG 31 of chain A.

      Links to the titration curve and data are also present in this table.
      GBR table
      Figure 5. Example sorted table by pKa. From this example it is clear that the highest pKa value is for residue ARG 52 in chain A, in particular atom CZ.






      Basic titration curve
      Acidic titration curve
      Figure 6. Top: the titration curve for the basic amino acid Lysine residue 13 (atom NZ). Bottom: the titration curve for the acidic amino acid Aspartate residue 25 (atom CG).
      How to visualise Bluues PDB files in PyMol
      In order to make the server more useful for experimentalists, the output should be easy to use in PyMol. The user can download the PDB file, potential for each atom, from the available files and examine the surface potential in PyMol by using the following command:
      PyMOL> spectrum b, blue_white_red, minimum=-5, maximum=5
      PyMOL> show surface
      
      where -5 and 5 are the minimum and maximum potentials and can be modified at will.
      In addition, if surface area option is used the PDF file, list of surface points, can also be used within PyMol in a similar manner.






      Examples
      Below is the link to sample output of the Bluues server. In addition, biological inferences are presented.

      Example 1   -    Complex antennapedia homeodomain with DNA (PDB ID: 9ANT).
      Chain A is protein while chain C and D are the DNA. The surface potential on the complexing part of the protein is strongly positive. First input 9ant into the PDB ID textfield and a into the chain ID textfield. For this example we examine the surface potential displayed in JMOL:

      1. The default display for JMOL is a wire and stick representation of the protein, red indicates positive potential while blue indicates negative potential.
      2. Click the Surface potential button to the display the calculated potential. Note this may take a little time.
      3. The atoms and bonds are now hidden by the surface. Click the Translucent button to display surface, atoms and bonds are now all colored by surface potential.
      4. The residues contacting in some form with DNA have been previously reported in the literature [11] (see figure 3) for 9ant. The residues include Arg 31, Leu 26, Arg 53, Tyr 25, Met 54, Gln 50, Ile 47, Arg 43, Asn 51, Tyr 8, Gln 6, Gln 44, Trp 48, Thr 13.
      5. Residue Arg 5 makes contact with the minor groove in the DNA.
      6. Ile 47 contacts the methyl of thymine 8 and the C8 of adenine 9 in the DNA.
      7. Asn 51 makes a pair of hydrogen bonds to adenine 9 in the DNA.
      8. The Diagram below reveals these important residues are involved in the positive face of the protein:
      9. surface diagrams
        Figure 7. (a) The initial surface displayed from the server with the positive potential displayed to the front. (b) The surface potential made transparent with the atoms and bonds colored also by potential. (c) Arg 5 is colored green, Asn 51 is colored black and Ile 47 is colored yellow. The other residues mentioned above are colored purple.
        surface diagrams
        Figure 8. On the contrary when rotated 180 degrees the protein contains a negative face. None of the highlighted residues are involved on this side of the protein.
      For protein-DNA complexes the polyelectrolytic aspects of DNA make electrostatic interactions very strong.


      Example 2   -    Tanford transition of beta-lactoglobulin (1BEB, full complex: leave chain blank). pH influences the structure of this protein through electrostatics. In particular, changes in the protonation state of titratable groups influence processes such as ligand acceptance or release, the % of disorder or flexibility and protein-protein association [12].

      1. First let's analyse the surface potential without the use of previous literature. We can find the top top potential atoms from the "Sorted tables". A possible procedure for finding the most potential in the surface might be to define a cut-off at (here we use +7 kJ/(mol.q)) and record residues which are not repeated. From the Surface potential table the following residues contained atoms with +7 kcal/mol.q or greater and were not repeated: 5, 6, 141, 100, 138, 69, 8, 101, 40, 60, 83, 77, 87, 150, 14, 79, 34, 91
      2. Reload the surface potential. Click translucent in order to see the ball and stick behind the surface potential.
      3. Open a console. Type select 5, 6, 141, 100, 138, 69, 8, 101, 40, 60, 83, 77, 87, 150, 14, 79, 34, 91. Then type color black.
      4. From the structure we can see numerous areas of interest. Let's keep it simple and examine the front of the protein.
      5. Right click the structure for more detailed JMOL options. Click view --> front.
      6. We can examine where these residues fall on the helices, loops and strands. Open 2 new tabs in your browser. copy the working bluues.html link into both tabs. In one tab simply orientate the structure to the front again to get an opaque surface (it is easier to see the high potnetials when it is not translucent). In the 2nd tab right click the structure for detailed JMOL options. Click style --> scheme --> cartoon.
      7. In the cartoon tab, select the residues again in the console and color them black.
      8. Figure 9 shows the JMOL structures you should have available in your 3 tabs.
      surface diagrams
      Figure 9. The beta-lactoglobulin (PDB code: 1beb) molecule orientated towards the front. left: The opaque surface potential. Middle: the translucent surface potential with residues containing maximum potential atoms colored black. Right: the cartoon with top residues marked in black.
    References

    If you use the server in work leading to publications:

      Please cite:


      Walsh I, Minervini G, Corazza A, Esposito G, Tosatto S.C.E. and Fogolari F.
      Bluues Server: electrostatic properties of wild-type and mutated protein structures.
      Bioinformatics. 28(16):2189-90.. (2012)

      Bluues Method:


      Fogolari F, Corazza A, Yarra V, Jalaru A, Viglino P and Esposito G.
      Bluues: a program for the analysis of the electrostatic properties of proteins based on generalized Born radii
      BMC Bioinformatics in press (2012)

      Other references:

    1. Fogolari F, Brigo A, Molinari H.
      The Poisson-Boltzmann equation for biomolecular electrostatics: a tool for structural biology.
      J Mol Recognit. 2002 15:377-92.

    2. Dolinsky TJ, Nielsen JE, McCammon JA and Baker NA.
      PDB2PQR: an automated pipeline for the setup of Poisson–Boltzmann electrostatics calculations
      Nucl. Acids Res. (2004) 32 (suppl 2):W665-W667.

    3. Mongan J, Svrcek-Seiler WA and Onufriev A.
      Analysis of integral expressions for effective Born radii.
      J Chem Phys. 2007 Nov 14;127(18):185101.

    4. Tjong H, Zhou HX
      GBr6: a parametrization free, accurate, analytical generalized Born method.
      J. Phys. Chem. 2007, 111:3055-3061.

    5. Grycuk T
      Deficiency of the Coulomb-field approximation in the generalized Born model: An improved formula for Born radii evaluation.
      J. Chem. Phys. 2003, 119:4817-4826.

    6. Antosiewicz J, McCammon JA, Gilson MK
      Prediction of pH-dependent properties of proteins.
      J. Mol. Biol. 1994, 238:415-436.

    7. Schutz CN, Warshel A.
      What are the dielectric "constants" of proteins and how to validate electrostatic models?
      Proteins. 2001 Sep 1;44(4):400-17.

    8. Li H, Robertson AD and Jensen JH.
      Very fast empirical prediction and rationalization of protein pKa values.
      Proteins 2005, 61:704–721.

    9. Fraenkel E, Pabo CO.
      Comparison of X-ray and NMR structures for the Antennapedia homeodomain-DNA complex.
      Nat Struct Biol. 1998 Aug;5(8):692-7.

    10. Fogolari F, Ragona L, Licciardi S, Romagnoli S, Michelutti R, Ugolini R and Molinari H.
      Electrostatic Properties of Bovine Beta-Lactoglobulin.
      PROTEINS: Structure, Function, and Genetics 39:317–330 (2000)