Overlapping things

Totally overlapping residues or ligands is an error that we observe with some frequency for flexible ligands such as sugars. If a residue or ligand has two alternate positions or rotamers, then both situations can be present in the refinement and in the PDB file, but their occupancies should add up to maximally 1.0 per spatial location.

It would also help if alternates would have a name that allows us to recognize them as alternates.

Example of two sugars being one

There are many examples of alternate sugar conformations that are not recorded as such, but as fully independent molecules. One example is 5ABP in which the two sugars are given as independent entities, but both with occupancy 0.5 for all atoms. 

REMARK   1  AUTH   N.K.VYAS,M.N.VYAS,F.A.QUIOCHO                        5ABP  14
REMARK   1  TITL   COMPARISON OF THE PERIPLASMIC RECEPTORS FOR          5ABP  15
REMARK   1  TITL 2 L-ARABINOSE-, D-GLUCOSE/D-GALACTOSE-, AND            5ABP  16
REMARK   1  TITL 3 D-RIBOSE.  STRUCTURAL AND FUNCTIONAL SIMILARITY      5ABP  17
REMARK   1  REF    J.BIOL.CHEM.                  V. 266  5226 1991      5ABP  18
....
REMARK   1  AUTH   F.A.QUIOCHO,N.K.VYAS                                 5ABP  21
REMARK   1  TITL   NOVEL STEREOSPECIFICITY OF THE L-ARABINOSE-BINDING   5ABP  22
REMARK   1  TITL 2 PROTEIN                                              5ABP  23
REMARK   1  REF    NATURE                        V. 310   381 1984      5ABP  24
REMARK   1  REFN   ASTM NATUAS  UK ISSN 0028-0836                  006  5ABP  25

The two sugars are given as independent entities, but both with occupancy 0.5 for all atoms.

NOTE

Figure 5. 5abp with two sugar molecules, one in red and one in yellow.

Figure 6. Same situation. Click on the picture for a blow-up of the sugars in a stick representation.

The REMARK cards in the PDB file 5abp suggest that there was no X-ray density to justify the presence of 2 forms of the sugar. 

REMARK   5 SINCE IT HAS BEEN SHOWN THAT ABP CAN BIND EITHER ALPHA OR    5ABP  42
REMARK   5 BETA ANOMERS OF D-GALACTOSE WITH ALMOST EQUAL AFFINITY,      5ABP  43
REMARK   5 BOTH ARE PROVIDED IN THE SAME SITE.                          5ABP  44

But final judgement on the reality of this situation will have to wait for re-refinement of the coordinates, which will be difficult because the experimentally determined reflections are unfortunately not available. The administrative encoding of the two sugars, however, will confuse many molecular visualisation, calculation, docking, etc. software packages.

Another example of overlapping sugars

JRNL        AUTH   V.KOENIG,L.VERTESY,T.R.SCHNEIDER
JRNL        TITL   CRYSTAL STRUCTURE OF THE ALPHA-AMYLASE INHIBITOR
JRNL        TITL 2 TENDAMISTAT AT 0.93 A
JRNL        REF    ACTA CRYSTALLOGR.,SECT.D      V.  59  1737 2003
JRNL        REF    J.BIOL.CHEM.                  V. 267 16467 1992

This file holds a series of GLA and GLB's (α-D-galactose and β-D-galactose). Two of these (GLA 901 and GLB 902; both in the B chain) overlap nearly completely. The same is true for the C and D chain bound galactoses, but the A chain binds (via its calcium) only to a GLB.

The structure of 1oko. Three of the sugars are overlapping GLA and GLB while one is GLB only. 

Example of the GLA and GLB atoms attached to the B chain.
HETATM 3716  C1  GLA B 901      -6.156  12.350  48.967  0.40 18.49           C
HETATM 3717  C2  GLA B 901      -6.897  13.450  48.246  0.40 16.03           C
HETATM 3718  C3  GLA B 901      -8.263  12.925  47.818  0.40 13.70           C
HETATM 3719  C4  GLA B 901      -8.113  11.640  47.021  0.40 13.11           C
HETATM 3720  C5  GLA B 901      -7.204  10.645  47.744  0.40 15.59           C
HETATM 3721  C6  GLA B 901      -6.946   9.389  46.892  0.40 16.01           C
HETATM 3722  O1  GLA B 901      -6.883  11.930  50.142  0.40 21.61           O
HETATM 3723  O2  GLA B 901      -7.088  14.564  49.092  0.40 17.42           O
HETATM 3724  O3  GLA B 901      -8.948  13.923  47.106  0.40 11.50           O
HETATM 3725  O4  GLA B 901      -7.593  11.985  45.748  0.40 12.20           O
HETATM 3726  O5  GLA B 901      -5.982  11.236  48.098  0.40 16.20           O
HETATM 3727  O6  GLA B 901      -6.244   8.391  47.598  0.40 18.17           O
HETATM 3728  C1  GLB B 902      -6.156  12.350  48.967  0.60 18.49           C
HETATM 3729  C2  GLB B 902      -6.897  13.450  48.246  0.60 16.03           C
HETATM 3730  C3  GLB B 902      -8.263  12.925  47.818  0.60 13.70           C
HETATM 3731  C4  GLB B 902      -8.113  11.640  47.021  0.60 13.11           C
HETATM 3732  C5  GLB B 902      -7.204  10.645  47.744  0.60 15.59           C
HETATM 3733  C6  GLB B 902      -6.946   9.389  46.892  0.60 16.01           C
HETATM 3734  O1  GLB B 902      -4.894  12.828  49.370  0.60 22.62           O
HETATM 3735  O2  GLB B 902      -7.088  14.564  49.092  0.60 17.42           O
HETATM 3736  O3  GLB B 902      -8.948  13.923  47.106  0.60 11.50           O
HETATM 3737  O4  GLB B 902      -7.593  11.985  45.748  0.60 12.20           O
HETATM 3738  O5  GLB B 902      -5.982  11.236  48.098  0.60 16.20           O
HETATM 3739  O6  GLB B 902      -6.244   8.391  47.598  0.60 18.17           O

Overlapping GLA and GLB. It doesn't look like it, but there are indeed two sugars. You see only one as the two sugars have the same coordinates.

Overlapping GLA and GLB.

Overlapping GLA and GLB.

The fourth sugar is just GLB.

The only error made is that there are more variables than necessary in the refinament, but why do it this way? It will only confuse the users...

In these cases of overlapping sugars with (or without...) alternate chiralities did the depositors choose two different residues that were placed on top of each other, rather than using one molecule with alternate atoms. Their reasoning might have been that the α and β forms are nor alternates from each other because they have different names, but it would be much easier for all software that works with PDB files to have them as alternates of each other.