So for example there is a paper by Ramakrishnan and Qasba which presents a model in a 3d stereo picture it has UDP-Gal, Glc, Mn++, a piece of the lactalbumin protein (?) and a piece of the lactose synthase (Beta-1,4-galactosyltransferase 1, B4GALT1) showing how the reactants are being held in the reaction pocket:
I would like to make a 3D MOLFILE from this, or have one. Just of this reaction center, not the entire proteins. I figure there might be a 3D folding structure or domain database somewhere where I could get data that I could convert into the MOLFILE coordinates.
I looked it up the UNIPROT KB for B4GALT1 and did find some 3D models that show some things:
Here UDP Galactose and the Manganese ion. But I don's see the Glucosamine in any of those 3D models nor do they detail the orientation of the side chains of the amino acids in the reaction center, let alone the lactalbumin association.
Is there any way I could get to 3D models that are shown in this paper (figure above?) do the authors not usually submit these things to some public databases? I think the answer is yes. Because as I moved from that UNIPROT image to the source https://www.rcsb.org/structure/2AGD lo and behold they reference my authors, and there is a data file https://files.rcsb.org/view/2AGD.pdb, which then might be the datafile they submitted. There one can find 3D coordinates down to the atom level!
EDIT: as commented by @andselisk there was a problem with my using the wrong link. The paper actually says what the accession number of their submissions are. The right one is 1NQI.
So, for example from their picture I want to find W314 that was labeled in their picture as a tryptophan at position 314, and indeed there it is (between 313 and 315:
ATOM 2476 N GLY B 313 -1.909 -10.437 27.249 1.00 19.91 N
ATOM 2477 CA GLY B 313 -0.797 -10.788 26.392 1.00 19.21 C
ATOM 2478 C GLY B 313 -0.248 -9.505 25.787 1.00 17.82 C
ATOM 2479 O GLY B 313 -0.786 -8.425 26.012 1.00 16.12 O
ATOM 2480 N TRP B 314 0.786 -9.635 24.973 1.00 18.89 N
ATOM 2481 CA TRP B 314 1.444 -8.478 24.397 1.00 18.21 C
ATOM 2482 C TRP B 314 0.786 -7.805 23.194 1.00 18.60 C
ATOM 2483 O TRP B 314 0.168 -8.449 22.356 1.00 16.45 O
ATOM 2484 CB TRP B 314 2.884 -8.868 24.045 1.00 18.06 C
ATOM 2485 CG TRP B 314 3.720 -7.734 23.557 1.00 18.61 C
ATOM 2486 CD1 TRP B 314 4.518 -6.898 24.318 1.00 17.43 C
ATOM 2487 CD2 TRP B 314 3.810 -7.268 22.217 1.00 16.35 C
ATOM 2488 NE1 TRP B 314 5.088 -5.938 23.520 1.00 16.69 N
ATOM 2489 CE2 TRP B 314 4.676 -6.139 22.225 1.00 17.88 C
ATOM 2490 CE3 TRP B 314 3.246 -7.686 21.000 1.00 18.70 C
ATOM 2491 CZ2 TRP B 314 4.990 -5.426 21.059 1.00 16.06 C
ATOM 2492 CZ3 TRP B 314 3.560 -6.963 19.823 1.00 14.57 C
ATOM 2493 CH2 TRP B 314 4.425 -5.852 19.871 1.00 15.37 C
ATOM 2494 N GLY B 315 0.920 -6.481 23.150 1.00 19.02 N
ATOM 2495 CA GLY B 315 0.459 -5.701 22.011 1.00 17.76 C
ATOM 2496 C GLY B 315 -0.937 -5.121 21.947 1.00 17.26 C
ATOM 2497 O GLY B 315 -1.885 -5.672 22.513 1.00 16.32 O
ATOM 2498 N GLY B 316 -1.039 -3.964 21.282 1.00 14.94 N
ATOM 2499 CA GLY B 316 -2.332 -3.328 21.074 1.00 12.98 C
ATOM 2500 C GLY B 316 -3.062 -2.543 22.141 1.00 14.49 C
ATOM 2501 O GLY B 316 -3.869 -1.674 21.793 1.00 16.13 O
There are the atom positions that look like 3D coordinates, and I should be able to turn them into a MOLFILE.
And I suppose there are no bonds in this PDB file, or are there? So if I wanted to get bonds for the MOLFILE, I would just have to match the atoms with MOLFILES or SMILES for the amino acids and other compounds I guess that might work. The atoms seem to use labels designating them to the AA, like in the 213 glycine example:
ATOM 2476 N GLY B 313 -1.909 -10.437 27.249 1.00 19.91 N
ATOM 2477 CA GLY B 313 -0.797 -10.788 26.392 1.00 19.21 C
ATOM 2478 C GLY B 313 -0.248 -9.505 25.787 1.00 17.82 C
ATOM 2479 O GLY B 313 -0.786 -8.425 26.012 1.00 16.12 O
I assume that "CA" means alpha-C so the other "C" is the carboxy-C, and "N" the amino-N. For the tryptophan then I assume "N" and "C" being the amino-N and carboxy-C respectively, "CA", "CB", ... being alpha-C, beta-C, gamma-C, "CE" epsilon-, "CZ" zeta-, "CH" eta-, and the numbers count the splits of the paths, NE1 being the ring-N, therefore CD1 means that side on the NE1, while the CD2 branches to the other side. CE2 would continue from there into the other ring, while CE3 is a new branch from CD2 completing the ring with the N, and so on, probably:
ATOM 2480 N TRP B 314 0.786 -9.635 24.973 1.00 18.89 N
ATOM 2481 CA TRP B 314 1.444 -8.478 24.397 1.00 18.21 C
ATOM 2482 C TRP B 314 0.786 -7.805 23.194 1.00 18.60 C
ATOM 2483 O TRP B 314 0.168 -8.449 22.356 1.00 16.45 O
ATOM 2484 CB TRP B 314 2.884 -8.868 24.045 1.00 18.06 C
ATOM 2485 CG TRP B 314 3.720 -7.734 23.557 1.00 18.61 C
ATOM 2486 CD1 TRP B 314 4.518 -6.898 24.318 1.00 17.43 C
ATOM 2487 CD2 TRP B 314 3.810 -7.268 22.217 1.00 16.35 C
ATOM 2488 NE1 TRP B 314 5.088 -5.938 23.520 1.00 16.69 N
ATOM 2489 CE2 TRP B 314 4.676 -6.139 22.225 1.00 17.88 C
ATOM 2490 CE3 TRP B 314 3.246 -7.686 21.000 1.00 18.70 C
ATOM 2491 CZ2 TRP B 314 4.990 -5.426 21.059 1.00 16.06 C
ATOM 2492 CZ3 TRP B 314 3.560 -6.963 19.823 1.00 14.57 C
ATOM 2493 CH2 TRP B 314 4.425 -5.852 19.871 1.00 15.37 C
I can see it's doable, but lots of manual work would be involved for me to match this. Ideally someone would have SMILES or MOLFILES with these standard (?) atom labels annotated? Or any other source of bond data?
Has anyone done this before, to construct a MOLFILE from (parts of!) such a PDB structure? Perhaps is there some software that does this? Like the authors here bothered to draw out specific atom structures and then connecting them with these blue ribbons. How did they draw this?
EPILOGUE: the accepted answer really hits the nail on the head. It's perfect. For other side questions, reading my article carefully especially the end, answers most. The only problem is, the figure I initially showed is what the authors had not submitted to the PDB and it really is the most interesting moment in the functioning of the lactose synthase. It is a model, predicted, not measured. I tried to reach out to the authors, but it's a long time ago 2001, still, nowhere have I found an actual model of this.
