Protocol for re-refining a structure from PDB deposited structure factors using X-PLOR.


The overall flow chart is below:

Create structure file and formated coordinate file
       |
Designate test set of relections
       |
Rigid body refinement 
       |
Determine weighting for further refinement
       |                <------------------------------------------
Relax the molecule      <-----------------------------             |
       |                                              |            |
Simulated annealing refinement    Stage 2: HOH        |  Stage 3:  |
       |                                   Ligand     |   Increase |
Positional refinement                      Disulfides |   Density  |
       |                                              |   Weight   |
B-factor refinement     >-----------------------------             |
                        >------------------------------------------

The individual steps are below:

For 7rsa the sf file (r7rsasf.ent.Z) had the following format:

*CONTNT   H,K,L,FOBS,A,B                                                        
*FORMAT   (3I4,3F10.2)                                                          
*COORDS   7RSA                                                                  
 -23   0  14     11.76     -9.97      0.00                                      
 -23   0  16      7.13      5.15      0.00                                      
 -23   0  20     18.33    -22.35      0.00                                      
 -23   0  21      9.55     15.79      0.00                                      
 -23   1   2     10.53    -10.96      9.08                                      

We want only the first 4 columns so use these two awk scripts on the sf file:

{ blank = substr($0,1,1);
if (blank != "*") { printf("%4d%4d%4d%9.2f\n",$1,$2,$3,$4)}
}

and

{ printf(" INDE %5d%5d%5d FOBS=   %7.3f  SIGMA=     1.000\n",$1,$2,$3,$4)} 

which creates a sigma column and gives the following format:

 INDE   -23    0   14 FOBS=    11.760  SIGMA=     1.000
 INDE   -23    0   16 FOBS=     7.130  SIGMA=     1.000
 INDE   -23    0   20 FOBS=    18.330  SIGMA=     1.000
 INDE   -23    0   21 FOBS=     9.550  SIGMA=     1.000
 INDE   -23    1    2 FOBS=    10.530  SIGMA=     1.000

Run xplor with setup_free_r.inp to designate the test set reflections. The output will be called *.cv for "cross validation".

Extract only the ATOM records from the PDB file for initial refinement. Make sure the number of structures in the extracted ATOM records is the same as the number of molecules in the unit cell. (Moleman2 gives you the number of molecules in the unit cell).

Run xplor with generate.inp to create the *.psf file and the initial *.pdb file. You need the segment name information (from PDB file) for the generate.inp:

name="7RSA"

Run xplor rigid.inp to best position the whole molecule. You will need the following information from the sf file:

 evaluate ($SG="P2(1)" ) 
 evaluate ($a=30.18)
 evaluate ($b=38.40)
 evaluate ($c=53.32)
 evaluate ($alpha=90)
 evaluate ($beta=105.85)
 evaluate ($gamma=90)
 evaluate ($low_res=9.91
 evaluate ($high_res=1.26)

On the new coordinates (rigid.pdb) run xplor with check.inp (edit above parameters and filenames) to get a weiughting value for density fitting:

 suggested weight wa = 60398.2

Then run xplor with prepstage.inp (edit above parameters and filenames) to bring the geometry of the molecule into agreement with the parameter file you are using so the simulated annealing won't blow up. The new coordinates will be called prepstage.pdb.

Then run xplor with slowcool.inp (edit above parameters and filenames) to do the first atomic refinement.

Look at the R and Free-Rfor each trial in slowcool.out (search for "Accum" in the file) and choose the model that is best. Then use the appropriate coordinate file (slowcool_1.pdb or slowcool_2.pdb) as input for the next step. [In this case the R=0.3044, and free_R=0.3530 at this stage]

Do positional refinement. Run xplor with positional.inp (edited with above parameters and file names) and use the output from this run (positional.pdb) as the input for the next step.[In this case the R=0.3044, and free_R=0.3530 at this stage]

Do B-factor refinement. Run xplor with brefinement.inp (edited with above parameters and file names).[In this case the R=0.2593, and free_R=0.2923 at this stage]

Incorporate the heteroatoms and disulfide bonds and go back to prepstage for another cycle of slowcool, positional and b-factor refinement as described below.

Stage 2: Heteroatoms and Disulfides

The disulfides, water and tertiary butyl alcohol will be added using the generatewater.inp file.

The 7rsa structure contains tertiary butyl alcohol. You need a PDB file, a topology file and a parameter file for this molecule. These are available for many heteroatom structures from the Uppsala Software factory in many cases. Go to http://xplor.csb.yale.edu/hetero/ and look for your compoud. Then download the PDB, topology, and parameter files.

Edit the generatewater.inp file to read in all the topology files needed:

    @TOPPAR:tophcsdx.pro        {* for protein *}
    @TOPPAR:toph19.sol          {* for water *}
    @tbu.top                    {* for tert butyl alc *}

Also edit generatewater.inp to read in the necessary parameter files:

   @TOPPAR:protein_rep.param
   @tbu.par

Edit generatewater.inp to contain a disulfide patch for each disulfide:

  patch disu
    reference=1=( resid 26 )
    reference=2=( resid 84 )
  end

Edit generatewater.inp to read in the water and tertbutyl alc as segments:

 segment                                         {*Generate water *}
   name="7RSA"                              {*Should match columns 73-76 in*}
   chain
     coordinates @wat.pdb
   end
 end
 coordinates @wat.pdb

 segment                                         {*Generate tert butyl alc.*}
   name="7RSA"                              {*Should match columns 73-76 in*}
   chain
     coordinates @tbu.pdb
   end
 end
 coordinates @tbu.pdb

Note: Your water coordinate file (wat.pdb in this case) should have the following atom and residue names:

HETATM 2034  OH2 TIP3  201      22.921   5.252  18.826  1.00 10.71      7RSA2143
HETATM 2035  OH2 TIP3  202      28.804  13.214  24.058  1.00 12.94      7RSA2144
HETATM 2036  OH2 TIP3  203      30.134  18.460   7.098  1.00 12.84      7RSA2145

Run xplor with generatewater.inp to make new pdb and psf files now containing water and heteroatoms and disulfides. These will be the input for prepstage.

Then run xplor with prepstage.inp (edit above parameters and new filenames) to create a new coordinate file for the next step (prepstage.pdb). [In this case the R=0.217, and free_R=0.244 at this stage]

Then run xplor with slowcool.inp (edit above parameters and new filenames). [In this case the R=0.1997, and free_R=0.2626 at this stage]

Do positional refinement again. Run xplor with positional.inp (edited with above parameters and new filenames) and use the output from this run (positional.pdb) as the input for the next step.[In this case the R=0.1997, and free_R=0.2625 at this stage]

Do B-factor refinement again. Run xplor with brefinement.inp (edited with above parameters and new file names).[In this case the R=0.1879, and free_R=0.2536 at this stage]

The bond rmsd from ideality was 0.005 at this stage. So to improve the fit to density another cycle with increased weight (wa) was performed.

Stage 2: Increase Weighting of Density

Then run xplor with slowcool.inp using brefinement.pdb as input and wa=100000. [In this case the R=0.1856, and the free_R=0.2556 at this stage]

Run xplor with brefinement.inp using slowcool_1.pdb as input with wa=100000. [In this case the R=0.1838, and the free_r=0.2534 at this stage]