Cameron Mura: DOWSER page (mirror at UNC)

Overview of DOWSER and available DNA-related files

...

A current limitation of the DOWSER program is its inability to handle PDB input files containing nucleic acid. Therefore, a pair of DOWSER-specific data files has been constructed in order to enable DOWSER to properly treat input files containing DNA (RNA another day!). These two files provide: (i) the atomic dictionary and residue definitions for DNA (thereby extending DOWSER's built-in atomdict.db); and (ii) the DNA-specific Lennard-Jones nonbonded interaction parameters and atomic radii that are used for molecular surface calculations (thereby extending DOWSER's built-in atomparms.db). For the sake of consistency with the current DOWSER parameter set for proteins, values for all nonbonded parameters for DNA atoms (i.e., atomic partial charges and LJ constants) were drawn from the Gromos FF, as implemented in the latest GROMACS (v3.2.1) MD simulation package (e.g., see the ffgmx* files for more specific information).

There are two approaches to utilizing the DNA-specific files provided here: (i) one may either download the dictionary and parameter files (below) and then use DOWSER with the "-atomtypes" and "-atomparms" command-line switches; or (ii) one may directly modify these two data files in the DOWSER source code by applying a patch file (below).

For route (i), save the following two files in the "DATA/" subdir of your local DOWSER installation (i.e., value of env variable $DOWSER), and then run DOWSER as shown below:

mydna_dict.db -- an atomic dictionary file for DNA
mydna_parms.db -- a nonbonded parameters file for DNA
Sample execution:
$] $DOWSER/bin/dowser some_file_with_DNA.pdb -atomtypes $DOWSER/DATA/mydna_dict.db -atomparms $DOWSER/DATA/mydna_parms.db

For route (ii), simply go to the root of your DOWSER source tree and apply the following patch file (which was produced in unified diff format). This will allow DOWSER to be run on DNA-containing PDB files without having to specify the "-atomtypes" and "-atomparms" options.

mydna_DOWSER.patch -- patch file for DOWSER source
Download the patch to the proper subdir and apply it:
$] cd $DOWSER/..
$] wget mydna_DOWSER.patch
$] patch -p0 < mydna_DOWSER.patch

For both routes (i) and (ii), DNA "ATOM" lines in the input PDB file must adhere to the following two conventions:

Deoxynucleotide residue names should be specified as DX (right-justified in the 3-character wide residue name column), so change GUA to DG, ADE to DA, CYT to DC, THY to DT. Furthermore, 5' dephosphorylated residues should be specified as DX5 (so DG5, DA5, DT5, DC5 for 5' termini).
For the DNA glycosidic atom names, all apostrophes ("'") should be changed to asterisks ("*").

Sample images from the application of DOWSER to DNA-containing input structures
(click on these thumbnail images for larger-sized pop-up images)

Sample output produced by using DOWSER with the above nucleic acid-specific dictionary and parameter data files. The input in this illustrative example was a simple 20-nt DNA structure which was artifically distorted (by kinking the deoxyribophosphate backbone) so as to introduce some space between the strands (primarily in the left two-thirds of the helix, as illustrated in the left-most panel), thereby leading to successful placement of a few waters by DOWSER:

The image directly to the left was produced by application of the "dowserx" command, thus resulting in the identification of many more interstitial waters.

The illustrations below are from a more realistic sample application of DOWSER with the above DNA-specific data files, using a relatively large protein-DNA complex as the target structure for solvation. Note DOWSER's careful placement of a bridging water between a protein side chain (Lys) and DNA base (Ade) in the zoomed-in view (Lys...Ade in rightmost panel).