Unrecognized interactions

[Phosphoserine]

Dear Dr. Xiang-Jun,

I've recently compared the results of protein--DNA interaction analysis provided by 3DNA-SNAP to some (more primitive) analysis I did using my scripts. I've noticed that in some of the structures I was interested in the SNAP tool did not recognize some amino acid--nucleotide interactions, despite them containing inter-atomic distances shorter than the cut-off (4.5 A). See, for example, the structure 1TTU. Amino acids LYS 233, ARG 399, and VAL 403 are not mentioned in the .out file, despite each of them being involved in interactions with DNA residues in the 13-bp helix (there could also be others, I had used a shorter interaction cut-off in my analysis).

Is there any reason for omitting these interactions, or is it a bug ? Thought you might be interested in it.

Best regards,
David J.

[xiangjun]

Hi David,

Please try SNAP with the "--type=either" option. By default, only protein interactions with BASE atoms are reported. Using 'either' would also include backbone interactions. I'll document this option in the coming release of SNAP.

Please have a try and report back if that helps.

Xiang-Jun

[Phosphoserine]

Hi,

thank you for the answer. Yes, adding the option --type=either helped recognize the interactions between the mentioned residues and DNA. If I understand correctly, the section "List of X nucleotide/amino-acid interactions" while using --type=either should include all interactions in the respective protein--DNA complex PDB structure which fall into the chosen inter-atomic distance cut-off ? And choosing --type=base or --type=backbone should report only the interactions of amino acid residues which interact with the DNA bases or DNA backbone, respectively, but not with the other one ?

It seems to me that only this one and the following (base-pair/amino-acid interactions) sections seem to be affected by the --type option. What confused me is that the following two sections which specifically deal with the backbone interactions (pair-wise phosphate-group/amino-acid H-bonding interactions and pair-wise sugar/amino-acid H-bonding interactions) are present in the .out file even when --type=base is selected, and, in fact, seem to be unaffected by the --type option. Similarly, the last three sections (base/amino-acid H-bonding interactions, base/amino-acid pseudo pairs, base/amino-acid pseudo stacks) are unaffected when one chooses --type=backbone.

This is just my opinion, but maybe some confusion could be prevented if the two sections dealing with the DNA backbone H-bonds and the last three sections describing interactions with DNA base atoms were either omitted when one chooses --type=base or --type=backbone, respectively, or if it were at least documented that they are agnostic to the --type setting. Alternatively, it would be helpful if it were documented that only the nucleotide/amino-acid interactions and base-pair/amino-acid interactions are affected by the --type option. From my (limited) experience, people are often interested in obtaining the complete list of interactions present in some protein--DNA complex, and it would be helpful to have it explicitly stated that the List of X nucleotide/amino-acid interactions section provides this (when --type=either is chosen).

Best regards,
David J.

[xiangjun]

Hi David,

Thank you so much for your feedback! As you know, the SNAP program is still in beta, so I'm completely open to constructive suggestions like yours.

If I understand correctly, the section "List of X nucleotide/amino-acid interactions" while using --type=either should include all interactions in the respective protein--DNA complex PDB structure which fall into the chosen inter-atomic distance cut-off ? And choosing --type=base or --type=backbone should report only the interactions of amino acid residues which interact with the DNA bases or DNA backbone, respectively, but not with the other one ?

• --type=base -- must have contacts with base atoms, may contact DNA/RNA backbone
• --type=backbone -- must have contacts with backbone atoms, may contact base
• --type=either -- must have contacts with base atoms or DNA/RNA backbone, may contact both
• --type=both -- must have contacts with base atoms and DNA/RNA backbone

Some of the interactions you're interested may be considered vdw contacts or hydrophobic interactions, which are currently not listed under the various H-bonding sections.

I understand the crucial importance of hydrophobic interactions in protein folding, and protein-DNA/RNA recognition. However, I've not been able to find a pragmatic algorithm, other than simple (carbon/carbon) distances, to identify them given a PDB structure. I've been reading some papers on the topic recently. Do you have any suggestion on must-read papers?

I'll update SNAP to a new release by next week. In the meantime, please feel free to communicate back your ideas on how SNAP could be made most useful from a user's perspective.

Best regards,

Xiang-Jun

[Phosphoserine]

Hi Xiang-Jun,

I'm afraid that I don't know of any papers specifically on this topic. I didn't really mean to pick out hydrophobic interactions specifically in the example, I was mostly interested in learning why I can't see them in the most general "List of X nucleotide/amino-acid interactions" section.

But you are right, if hydrogen bonding and stacking interactions are described in detail, it sounds like a good idea to also dedicate a section to contacts between hydrophobic carbon atoms. It would also be very useful to have a section on water-mediated interactions, which are also often crucial for the complex stability, but I understand this would probably involve a lot of additional work. But I'm afraid I don't have any advice on, for example, what inter-atomic distance cut-offs or angles between various vectors should be considered if one were to detect these (hydrophobic or water-mediated) interactions.

Best regards,
David