Get major/minor grooves from structure with broken linkages

[ilibarra]

Dear Dr. Lu,

I have a PDB entry from which I'd like to get groove parameters. However, it seemst that because the DNA chains are not completely connected during the find_pair routine, the analyze routine cannot calculate the parameters. I have run this PDB entry as a whole, and also extracting the DNA chains I'm interested (please find them both attached in this post).

# command
find_pair 4xrs_dna.pdb

    2         # duplex
    8         # number of base-pairs
    1     1    # explicit bp numbering/hetero atoms
    1    30   1 #    1 + ....>L:...2_:[.DT]T-**+-A[.DA]:..34_:M<....   7.99   1.08  26.53  10.95  11.47
    4    28   0 #    2 | ....>L:...5_:[.DA]A-**--T[.DT]:..32_:M<....   5.43   0.84  27.23  10.31   8.48
    5    26   9 #    3 x ....>L:...6_:[.DC]C-**--T[.DT]:..30_:M<....   2.42   0.92  25.10  10.44   4.51
    8    24   0 #    4 | ....>L:...9_:[.DG]G-----C[.DC]:..28_:M<....   1.94   1.69  62.83  10.24   5.47
    9    23   9 #    5 x ....>L:..10_:[.DA]A-----T[.DT]:..27_:M<....   0.79   0.44  52.85   8.67  -0.69
   12    20   1 #    6 + ....>L:..13_:[.DA]A-**+-T[.DT]:..24_:M<....   1.12   0.86  61.41   8.40   4.90
   14    18   0 #    7 | ....>L:..15_:[.DT]T-**+-A[.DA]:..22_:M<....   3.48   0.53  63.53   9.27   7.71
   15    16   0 #    8 | ....>L:..16_:[.DG]G-**+-T[.DT]:..17_:L<....   7.51   1.27  52.73   6.84  11.70
##### Base-pair criteria used:     4.00     0.00    15.00     2.50    65.00     4.50     7.80 [ O N]
##### 6 non-Watson-Crick base-pairs, and 5 helices (2 isolated bps)
##### Helix #1 (1): 1
##### Helix #2 (2): 2 - 3  ***broken O3' to P[i+1] linkage***
##### Helix #3 (2): 4 - 5
##### Helix #4 (1): 6
##### Helix #5 (2): 7 - 8  ***broken O3' to P[i+1] linkage***

Is the main reason for this structure not being completed the broken linkages? is it possible to fix the linkages with 3DNA or increase the distance threshold to tolerate them, in order to recover the al the base pairs and then calculate the grooves?

At the moment, the groove parameters are empty.

(output from analyze 4xrs_dna.inp. Only eight base pairs from the DNA-duplex are recovered and shown).

Minor and major groove widths: direct P-P distances and refined P-P distances
   which take into account the directions of the sugar-phosphate backbones

   (Subtract 5.8 Angstrom from the values to take account of the vdw radii
    of the phosphate groups, and for comparison with FreeHelix and Curves.)

Ref: M. A. El Hassan and C. R. Calladine (1998). ``Two Distinct Modes of
     Protein-induced Bending in DNA.'' J. Mol. Biol., v282, pp331-343.

                  Minor Groove        Major Groove
                 P-P     Refined     P-P     Refined
   1 TA/TA       ---       ---       ---       ---
   2 AC/TT       ---       ---       ---       ---
   3 CG/CT       ---       ---       ---       ---
   4 GA/TC       ---       ---       ---       ---
   5 AA/TT       ---       ---       ---       ---
   6 AT/AT       ---       ---       ---       ---
   7 TG/TA       ---       ---       ---       ---

Thank you for any feedback or suggestion,

Ignacio

P.S. This is a similar post that reports a similar warning.
http://forum.x3dna.org/general-discussions/this-structure-has-broken-o3'-to-p(i1)-linkages/

[ilibarra]

I have been able to extract the groove values by running Curves+. I guess that's a solution for now .

Thanks,
Ignacio

[xiangjun]

Hi Ignacio,

I'm glad to know that you've been able to "extract the groove values by running Curves+". As noted in my blogpost "Curves+ vs 3DNA", Curves+ has unique features and is more complete in its calculation of groove parameters (depths and widths). It's a good idea to always give it a try even if 3DNA also works.

As far as 3DNA is concerned, the DNA structure in PDB entry 4xrs is highly deformed. The helical fragment automatically identified by 'find_pair' in such a case is not as desired. With missing phosphate groups between neighboring pairs, the  'analyze' program can no longer calculate groove widths. In this case, you could manually edit the output from 'find_pair'  before feeding into  'analyze' to get the desired results. Imagine the scenario where you do not have access to 'find_pair' and have to prepare an input to 'analyze' manually, as you did for Curves+.

Best regards,

Xiang-Jun