Description

Molsurfer is based on JSMol and replaces the former java based software. It uses the representation of the macromolecular interface generated by adsi (the interface module of ads, analytically defined molecular surfaces ) and JSmol, a JavaScript implementation of Jmol. (Java plugins have recently become very restrictive, due to many security concerns.)

It can be used to study protein-protein and protein-DNA/RNA interfaces. The 2D projections of the computed interface aid visualization of complicated interfacial geometries in 3D. Molecular properties, including hydrophobicity and electrostatic potential, can be projected onto the interface. MolSurfer can thereby aid exploration of molecular complementarity, identification of binding "hot spots" and prediction of the effects of mutations. MolSurfer can also facilitate the location of cavities at macromolecular interfaces.

Links to MolSurfer

You can generate links to molsurfer, allowing the analysis of interfaces in RCSB entries. For this the URL parameters pdbidinput, chainid1 and chainid2 needs to be specified in the following manner:

http://molsurfer.h-its.org/cgi-bin/molsurfer_master.cgi?pdbidinput=1hbb&chainid1=A&chainid2=B,C,D

Just replace the bold parts of the URL above with the information about your PDB entry.

Problems with PDB files

In some cases molsurfer will not work on some uploaded PDB files. If the files have missing columns or coordinates with large numbers. In these cases you can use this script to move the protein to the center of the coordinate system, resulting in smaller numbers for the coordinates. The script can be called with the pdb file to process as argument. It will write a file with '-moved-origin' appended. Use this to upload to molsurfer. The script requires python3, biopython and numpy installed. It was tested on linux only.

Details of Interface Mapping and Displayed Properties

• A quasi rectangular mesh with 1 Å spacing was generated on the analytically defined interface between these 2 macromolecules.
• Points for which the sum of the distances to the closest atoms of macromolecule I and II exceeds 6 Å were deleted.
• Heteroatoms that lie within 3 Å of any interface point were added.
• The properties of each macromolecule were projected onto every point of the interface; these are the properties assigned to the closest atom to the point.
• Residue hydrophobicities were assigned according to the residue name and following the parameters in Eisenberg D., Weiss R.M., Terwilliger T.C. and Wilcox W. (1982) Farad. Symp. Chem. Soc., 17, 109-120, namely:

  ALA   0.25       GLN  -0.69       LEU   0.53       SER  -0.26
  ARG  -1.80       GLU  -0.62       LYS  -1.10       THR  -0.18
  ASN  -0.64       GLY   0.16       MET   0.26       TRP   0.37
  ASP  -0.72       HIS  -0.40       PHE   0.61       TYP   0.02
  CYS   0.04       ILE   0.73       PRO  -0.07       VAL   0.54
  none of the above                 0.00
• Atomic hydrophobicities were assigned according to the atom name and follow  Eisenberg D., Wesson M., Yamashita M. (1989) Chem. Scrip., 29A, 217-221, namely:

  'NZ  LYS'  -38       'OE1 GLU'  -37         'C'    18
  'NH1 ARG'  -38       'OE2 GLU'  -37         'S'     5
  'NH2 ARG'  -38       'OD1 ASP'  -37         'O'    -9
                       'OD2 ASP'  -37         'N'    -9
  none of the above    0
• Atomic radii were also assigned from the previous reference, namely:

  'C'    1.9 A
  'S'    1.8 A
  'O'    1.4 A
  'N'    1.7 A
  none of the above    1.9 A
• Electrostatic potential of each (isolated) macromolecule was computed with the APBS program and the potential values were interpolated at each interface point.
• Additional plots are shown for the similarities of electrostatic potential, atomic and residue similarities as well as B-Factor similarity.

  The similarities are calculated according to Hodgkin and Richards (Journal of Quantum Chemistry, 1987). Regions where both sides of the interface are very similar have values of approx. 1. Anticorrelated areas have values of -1. Uncorrelated areas near 0.

• Additionally, the changes in binding free energy upon alanine mutation were assigned to each side-chain atom of the corresponding residue.  This data can be obtained from the Alanine Scanning Energetics database for example.

More details can be found in Analytically Defined (molecular) Surfaces website. The standalone version of MolSurfer can be downloaded from the previous MolSurfer website.

References

• R.R. Gabdoulline, R.C. Wade & D. Walther (2003) MolSurfer: a macromolecular interface navigator, Nucleic Acids Research, 31, 3349-3351. abstract and full text
• R.R. Gabdoulline, R.C. Wade & D. Walther. (1999) MolSurfer: two dimensional maps for navigating three-dimensional structures of proteins, Trends Biochem. Sci., 24, 285-287 . (excerpt)

Application Examples

MolSurfer applications citing the first reference above.

• Hough MA, Hall JF, Kanbi LD, Hasnain SS. (2001) Structure of the M148Q mutant of rusticyanin at 1.5 angstrom: a model for the copper site of stellacyanin, Acta Crystallogr D, 57, 355-360.
• Luedemann SK, Gabdoulline RR, Lounnas V, Wade RC, Substrate access to cytochrome P450cam investigated by molecular dynamics simulations: An interactive look at the underlying mechanisms, Internet. J. Chem. (2001) 4, 6. (abstract)
• Campbell JD, Biggin PC, Baaden M, Sansom MSP (2003) Extending the structure of an ABC transporter to atomic resolution: Modeling and simulation studies of MsbA, Biochemistry, 42 (13), 3666-3673.
• Flaus A, Rencurel C, Ferreira H, Wiechens N, Owen-Hughes T (2004) Sin mutations alter inherent nucleosome mobility, EMBO Journal, 23, 343-353.
• Wang T, Tomic S, Gabdoulline RR, Wade RC (2004) How optimal are the binding energetics of barnase and barstar? Biophys J. 87, 1618-30.
• Huelsmeyer M, Chames P, Hillig RC, Stanfield RL, Held G, Coulie PG, Alings C, Wille G, Saenger W, Uchanska-Ziegler B, Hoogenboom HR, Ziegler A, (2005) A Major Histocompatibility Complex Peptide-restricted Antibody and T Cell Receptor Molecules Recognize Their Target by Distinct Binding Modes (A Major Histocompatibility Complex Peptide-restricted Antibody and T Cell Receptor Molecules Recognize Their Target by Distinct Binding Modes), J. Biol. Chem. 280, 2972-2980.
• K�vesi I, Schay G, Yonetani T, Laberge M, Fidy J, (2005) High pressure reveals that the stability of interdimeric contacts in the R- and T-state of HbA is influenced by allosteric effectors: Insights from computational simulations, Biochim Biophys Acta, 1764, 516-21.

Authors

• The list of authors of the original version is given here
• A complete reimplementation in python, JSMol and HTML-Bootstrap was performed from Jui-Hung Yuan in 2018.
• The functionality to calculate electrostatic properties on downloaded PDB entries, an overhaul of the demonstration cases and other features was added by Dorothee Groß in 2022.
• General maintainance is done by Stefan Richter (mcmsoft(at)h-its.org)

Issues

• Uploading PDB files not conforming to the standards might cause problems. See the section "Problems with PDB files" on the main documentation page on how to resolve this.
• In case you use an ad blocker with your browser, the JSMol visualization of the protein might not be loaded (Error message "Network error"). In this case, disable your ad blocker or put the molsurfer on the exception list.
• Uploading two files, both without chain id causes an error
• Using two unrelated proteins causes an error

Changes

• Aug/2023 added a downloadable script to fix problems in PDB files with large numbers in coordinate files. See the section "Problems with PDB files" on the main documentation page on how to resolve this.
• July/2022 changed forcefield to AMBER and PDB2PQR to version 3.5.2 to allow also processing of nucleotides. Removed all HET Atoms in the pdb files before pdb2pqr processing. Use ionic strength of the original version (0.15 M instead of 0.05 M, also protein permittivity from 2 to 4 as in the original publication). Display all models in case of NMR structures (only the first one will be used for the calculation)
• 2022 added support for generating electrostatic properties when only PDB ids are submitted (Runs PDB2PQR in the background). Links to log files and data files provided.
• 2021 added similarity maps for hydrophobic and electrostatic properties. This is a reintroduction of a feature from the original implementation
• 2018 reimplementation with JSMol. No Java required any longer.
• Adding the following features:
  • Also added the feature to display B-Factor and electrostatic potential information in one run.
  • Support to start using PDB identifier
  • Dynamically load description and chain information for PDB entries. Chain information for files
  • Initial release using Java applet and webmol.