Phosphodiester bond length

[sharanr2]

Hi,

I was wondering if there was a way to see phosphodiester bond lengths in the backbone of an RNA structure? The closest I have found is the Dp field in the dssr-torsion.txt file, however, if the distance appears to be too long, it is not calculated, but a "---" is there.

Please let me know if this is possible or if you need more information/examples. Thank you.

[xiangjun]

Hi,

Thanks for using DSSR and posting your question on the Forum.

I was wondering if there was a way to see phosphodiester bond lengths in the backbone of an RNA structure?

Could you please be specific as to what you want to achieve? As always, concrete examples help.

As mentioned in the referred to "dssr-torsions.txt" output file, Dp is the perpendicular distance of the 3' P atom to the glycosidic bond. Please see Chen et al. (2010): "MolProbity: all-atom structure validation for macromolecular crystallography." Acta Crystallogr D Biol Crystallogr, 66(1):12-21 for the original reference. Clearly, Dp cannot be defined is either 3' P or the corresponding C1'--RN9/YN1 glycosidic bond is missing, and that's what the "---" is for.

Xiang-Jun

[sharanr2]

Thank you for your quick response.

I am hoping to use DSSR to characterize "good" structures from "bad" ones for use in a scoring function for structure prediction. For example (structure attached), if there is a large gap in a backbone, I would like to be able to quantify how large that gap is (these gaps longer than actually possible). The two nucleotides are in sequential order (base 5 and base 6) in the same strand (A). The distance is around 8 Å. I have tried adding an explicit connect statement between the P and the O3' of two nucleotides but that did not seem to change anything.

[xiangjun]

Thanks for your followup and providing a sample structure. I will look into it and get back to you later.

Just a quick note: DSSR does not check the PDB CONECT record, so adding such info won't help.

Xiang-Jun

[xiangjun]

Hi,

I've looked into the pred.pdb structure you provided, and noticed unusual ordering of nucleotides (nts). As seen in the attached images, the nts on chain A (or chain B) are not continuous, as one would expect in a normal PDB structure. This out-of-order issue causes problem: for example, A.C5 and A.C6 on chain A are far off, while A.C5 and B.U6 are close. Since A.C5 and A.C6 are clearly not connected, DSSR does not calculate Dp, as designed.

I do not know which software was used to predict the structure. However, it is desirable that the RNA structure be properly formatted. Here the nts need to be properly ordered in the PDB file.

HTH,

Xiang-Jun

[sharanr2]

Thank you for your response.

Is there any parameter that would measure the length of that gaps in the backbone? In the output file, DSSR recognizes there is a break in the chain in the summary of structural features section.

In the actual structure, those two bases are bonded by phosphodiester linkages, making a kissing loop between the A and B chains.

[xiangjun]

The issue lies in the PDB coordinate file (pred.pdb). As is clear from the Jmol image of your structure, A.C5 and A.C6 (circled) are far off to be connected. The whole structure is broken into 3 segments. A quick look of pred.pdb in PyMOL leads to the same result. The DSSR output is consistent with the visual observations of Jmol and PyMOL. I do not think DSSR can do more before the obvious problem is fixed in your PDB data file.

Xiang-Jun