O3'-P distance too long after reconstruction

[Damien]

Dear Xiangjun,

I analyzed the DNA molecule contained in PDB 2AOR from Web 3DNA. I then downloaded the parameter file. And I performed a reconstruction from the parameter file.

The resulting structure build.pdb contains many O3'-P bonds having a too long length, up to 4 A.

May you tell me what I made wrong ?

Many thanks for your help

Damien

[xiangjun]

Hi Damien,

The fact that "O3'-P distance too long after reconstruction" is well expected; 3DNA built structures are accurate for the bases, with only an approximate sugar-phosphate backbone. See FAQ "How do I build nucleic acid structures with sugar-phosphate backbone?"

For cases with longer than the default 4.5 A O3'(i)--P(i+1) distance, the 3DNA rebuild program outputs info message as below (using your example, and with a B-DNA backbone conformation):

Code: [Select]

O3' (#317) and P (#331) on chain A have distance 5.3 over 4.5: no linkage assigned
O3' (#768) and P (#782) on chain B have distance 5.2 over 4.5: no linkage assigned
This approximate structure may serve as a starting point for energy minimization.

Xiang-Jun
 

[Damien]

Many many thanks Xiang-Jun for this prompt reply! I appreciate!

OK, I understand the approximation for the sugar-phosphate backbone.

I have two questions:

- The default O3'-P distance is 4.5A. From the ADIT server of the PDB, the expected value is 1.6A. What is in fact the range of the possible values?

- Could you indicate some tools alllowing to do energy minimization?

Thanks again,

Damien

[xiangjun]

- The default O3'-P distance is 4.5A. From the ADIT server of the PDB, the expected value is 1.6A. What is in fact the range of the possible values?

The O3'-P covalent bond distance is ~1.6 A, as you noticed on the ADIT server of the PDB. The 3DNA default of 4.5 A is purely an empirical value used to decide if to include the bond in a corresponding CONECT record in the rebuilt PDB file; it has no 'chemical' meaning. Have a look of build.pdb file using a text editor, you will see my point.

- Could you indicate some tools alllowing to do energy minimization?

AMBER should help. Some other MD packages or Phenix should also do the trick. Yet, I still failed to find a command-line tool that can 'regularize' the backbone to a reasonable geometry while keeping the base atoms fixed.

HTH,

Xiang-Jun

[Damien]

Many thanks Xiang-Jun for your help! I appreciate!

Damien

[xiangjun]

Update on 2013-09-06: please see my post titled "Restraint optimization of DNA backbone geometry using PHENIX" for a PHENIX-based method to solve the long O3'-P distance problem.

Xiang-Jun