mutating DNA in DNA protein complex

[kailsen]

I am a graduate student with biochemistry background. Currently i am working on DNA protein interaction and recently came to know about 3DNA through hundreds of citations that your software has received.
I tried to use the software to understand the effect of DNA mutation on binding of the protein. The Transcription factor in question has the pdb code 1GJI. The DNA sequence in this complex is GGGTTTAAA... and i want to mutate this sequence to GGATAAAAA... Preferably,I want to make the changes in the DNA protein complex itself keeping the backbone same. I hope it will be possible using your software.
Is it advisable to use Rebuild program to get the desired changes incorporated?

thanking you in advance for your help

dnaserver

[Liron Bar]

I am doing sort of DNA mutagenesis with 3DNA. Basically what I do is
generating base pair parameters file from PDB file with DNA, than
changing DNA sequence there and the rebuilding DNA in PDB format with
rebuild program. However, rebuild program from 3DNA builds its own phosphate backbone which, albeit similar, is different than the original one.

Could someone tell whether it is possible to run rebuilding with 3DNA in a mode which preserves the original backbone conformation? (one
can write a script to copy phosphate backbone from the original PDB file
but maybe it is unnecessary).

Janek

[xiangjun]

Nice to see two questions in a row asking for the possibility of performing base mutations while keeping the backbone unchanged. Given this is such a common and clearly defined function, it is hard to imagine there is no such a handy standalone utility program from so many other resources to get the job done. Am I missing something here?

Even though no such a standalone program exists (yet) in 3DNA, this job is certainly doable with currently available tools in 3DNA, and in a conceptually very simple way. Indeed it is a case in my mind for quite some time. I will provide a new utility program, temporally named "mutate_bases" (anyone has suggestion for a better name?), which is very generic and easy to use in future release of 3DNA. For the time being, here is a working example to illustrate how this could be done in 3DNA. I am hoping either or both of you would be able to automate this job by writing a script and contribute it back to the community. Shorting of or in addition to that, please do share with the community on how the following recipe goes with your projects.

I am using pd0001.pdb,  the X-ray structure of the nucleosome core particle at 2.8 A solved by Luger. This structure is distributed with 3DNA, so you can easily access it.  I would like to mutate the 5th base-pair from A-T to G-C while keeping the DNA backbone conformation and protein untouched. This method is generic by allowing for multiple mutations and/or to a non Watson-Crick base-pair as well.

Code: [Select]

find_pair pd0001.pdb pd0001.inp
analyze pd0001.inp
frame_mol -1 ref_frames.dat pd0001.pdb pd0001_bp1.pdb

cp bp_step.par bp_step_mut.par
cp pd0001_bp1.pdb pd0001_mutok.pdb

# manually mutate the 5th bp from A-T to G-C in file 'bp_step_mut.par'; see attachment link below
rebuild -atomic bp_step_mut.par pd0001_mut.pdb
cat pd0001_mut.pdb pd0001_bp1.pdb > pd0001_cmb.pdb
rasmol pd0001_cmb.pdb

find_pair pd0001_mut.pdb pd0001_mut.inp

# for the 5th base-pair, compare 'pd0001.inp'
#  810 1093  0 #    5 | ....>I:...5_:[..A]A-----T[..T]:.288_:J<....  0.57  0.07 16.19  8.75 -3.48
# with 'pd0001_mut.inp'
#    5  288  0 #    5 | ....>A:...5_:[..G]G-----C[..C]:.288_:B<....  0.57  0.07 16.18  8.85 -3.48

# now manually replace coordinates of the BASE (including C1') in residue ....>I:...5_:[..A]A
# of file 'pd0001_mutok.pdb' with those from ....>A:...5_:[..G]G of file 'pd0001_mut.pdb'
# and similarly T[..T]:.288_:J< by C[..C]:.288_:B<

HTH,

Xiang-Jun

[Liron Bar]

Shorting of or in addition to that, please do share with the community on how the following recipe goes with your projects.


Xiang-Jun

Hi,

It worked for me, however two modifications to the procedure had to be added:
1. Preparation of input DNA file for 3DNA is often needed (changing HETATM to ATOM and new bases name to old ones i.e. DG to G, DA to A etc., changing non-standard numbering of residues)
2. Correcting residue and chain identifiers in the final mutated structure and correcting atom numbering. Because I am preparing this procedure for sort of bioinformatic course I wanted not to use any scripts but rather already available tools. I successfuly used VMD for correcting atom numbering.

Thanks a lot
Janek Kosinski

In case anbybody is interested, here is the exact procedure:

Prepare 2oaa.pdb file for mutagenesis with 3DNA program.
1.   Open 2OAA.pdb in Swiss PDB Viewer
2.   Select chain C (the first DNA chain) and renumber residues starting from 1 (Menu Edit -> Rename Current Layer -> Renumber Selected groups from : 1
3.   Select chain D (the second DNA chain) and renumber residues starting from 1 (Menu Edit -> Rename Current Layer -> Renumber Selected groups from : 1
4.   Select chain A, C and D and Save selected residues in new directory DNA_mutagenesis/ under the name 2oaaA.pdb
5.   Close Swiss PDB Viewer
6.   Open 2oaaA.pdb in WordPad
7.   Press Ctrl+H to open Find & Replace dialog. In “Find what” type HETATM and in “Replace with” type ATOM<space><space> (ATOM and two spaces, these spaces are important)
8.   Press Ctrl+H to open Find & Replace dialog. In “Find what” type  <space><space> D and in “Replace with” type <space><space><space>
9.   Save the file.
Make mutagenesis with 3DNA program.
1.   Open Windows Command Prompt. Go to DNA_mutagenesis/ directory.
2.   Execute:
find_pair 2oaaA.pdb 2oaaA.inp
3.   Run analyze
analyze 2oaaA.inp
4.   Run frame_mol to attach local helical frames and set the orientation of a structure.
frame_mol -1 ref_frames.dat 2oaaA.pdb 2oaaA_bp1.pdb
5.   Copy the files:
cp bp_step.par bp_step_mut.par
cp 2oaaA_bp1.pdb 2oaaA_mutok.pdb
6.   Manually mutate bp_step_mut.par
7.   Open bp_step_mut.par in WordPad. Change 6th base pair (T-A) to C-G
8.   Rebuild PDB file
rebuild -atomic bp_step_mut.par 2oaaA_mut.pdb
9.   Open 2oaaA_mut.pdb and 2oaaA_mutok.pdb in WordPad. Manually replace coordinates of the base (including C1`) in residue 6 of chain C of 2oaaA_mutok.pdb with those from residue 6 of chain A of 2oaaA_mut.pdb and in residue 6 of chain D of  2oaaA_mutok.pdb with those from  residue 6 of chain B of 2oaaA_mut.pdb. After this, in 2oaaA_mutok.pdb, adjust the names of residues and chains in mutated residues and residue numbers. Do not care at wrong atom numbering for mutated bases at the moment.
10.   To correct atom numbering open 2oaaA_mutok.pdb in VMD. After opening save the coordinates under new name 2oaaA_mutatedDNA.pdb.

[xiangjun]

Hi Janek,

Thanks for reporting back and giving a detailed account on how you get it done. You have provided the community a very useful working example from a user's prospective. This type of communication is what I value the most. Along the same line, I am hoping 'dnaserver' could follow your lead in getting back to 3DNA forum community.

To make it more accurate, your 4th step: running "frame_mol" with option "-1" is to re-orient the structure with reference to the first base-pair. This step is necessary because 3DNA rebuilt structure (from "rebuild") is always set w.r.t. to the first base-pair, and it has nothing to do with attaching local helical frames.

4. Run frame_mol to attach local helical frames and set the orientation of a structure.
frame_mol -1 ref_frames.dat 2oaaA.pdb 2oaaA_bp1.pdb

Properly aligning the original (crystal) structure and 3DNA rebuilt one is the essence of this method, and it is exactly the point of the following steps in my original post trying to illustrate.

Code: [Select]

rebuild -atomic bp_step_mut.par pd0001_mut.pdb
cat pd0001_mut.pdb pd0001_bp1.pdb > pd0001_cmb.pdb
rasmol pd0001_cmb.pdb
It would be helpful if you could revise your procedures slightly, and re-post   it, with necessary attachments, in the Users' Contributions section. Putting your recipe there will make it standing out so others can easily find it.

Once again, thanks for contributing back to the 3DNA forum!

Xiang-Jun

[Liron Bar]

Hi Janek,

Thanks for reporting back and giving a detailed account on how you get it done. You have provided the community a very useful working example from a user's prospective. This type of communication is what I value the most. Along the same line, I am hoping 'dnaserver' could follow your lead in getting back to 3DNA forum community.

No problem, however few issues need to be clarified before publishing the final version of the methodology:

To make it more accurate, your 4th step: running "frame_mol" with option "-1" is to re-orient the structure with reference to the first base-pair. This step is necessary because 3DNA rebuilt structure (from "rebuild") is always set w.r.t. to the first base-pair, and it has nothing to do with attaching local helical frames.

4. Run frame_mol to attach local helical frames and set the orientation of a structure.
frame_mol -1 ref_frames.dat 2oaaA.pdb 2oaaA_bp1.pdb

Ok, I have copied the description of the frame_mol from the manual without checking what -1 option does. Thanks, I will update the tutorial

Properly aligning the original (crystal) structure and 3DNA rebuilt one is the essence of this method, and it is exactly the point of the following steps in my original post trying to illustrate.

Code: [Select]

rebuild -atomic bp_step_mut.par pd0001_mut.pdb
cat pd0001_mut.pdb pd0001_bp1.pdb > pd0001_cmb.pdb
rasmol pd0001_cmb.pdbIt would be helpful if you could revise your procedures slightly, and re-post   it, with necessary attachments, in the Users' Contributions section. Putting your recipe there will make it standing out so others can easily find it.

Once again, thanks for contributing back to the 3DNA forum!

Xiang-Jun

I cannot uderstand these steps:

Code: [Select]

cat pd0001_mut.pdb pd0001_bp1.pdb > pd0001_cmb.pdb
rasmol pd0001_cmb.pdb
You generate pd0001_cmb.pdb file which you do not use later for the creation of the final structure. What do you do with rasmol apart from visualizing the structure? Why the procedure worked for me without including these two steps?

Thanks for your help, we will make sth suitable for the Users' Contributions section.

Cheers,
Janek

[xiangjun]

Dear Janek,

Indeed, the following two steps have no bearings on the result. I put them here simply for illustration purpose: i.e., to see with RasMol how the base atoms rebuilt from 3DNA overlap with their original crystal structure counterparts when properly aligned. As shown in the 3DNA paper (2003 NAR), usually the RMSD would be less than 0.05 angstroms.

Code: [Select]

cat pd0001_mut.pdb pd0001_bp1.pdb > pd0001_cmb.pdb
rasmol pd0001_cmb.pdbAnother minor issue relates to the C1' atoms: it really does not matter whether to include them when you perform the mutations. I included them in my previous post simply because they was generated by default with 3DNA.

Cheers, and thanks for your efforts in polishing your recipe!

Xiang-Jun
[hr:3ic50y69][/hr:3ic50y69][red:3ic50y69]Note[/red:3ic50y69] (added June 5, 2011):
[pre:3ic50y69]Please see the thread "change one base pair in a double-strand DNA structure file"
where the Perl script [mono:3ic50y69][red:3ic50y69]mutate_bp[/red:3ic50y69][/mono:3ic50y69] and the ANSI C program [mono:3ic50y69][red:3ic50y69]mutate_bases[/red:3ic50y69][/mono:3ic50y69] are introduced.[/pre:3ic50y69]