Hi all,
Just a general question for the membrane protein crystallographers out there – is there a consensus on what to do with unidentified lipid/detergent density (in the event one cannot see a headgroup or other identifying feature)? This is something I have run into a few times with membrane protein structures solved from crystals grown by vapor diffusion, but it is rarely too much of a concern there, as only a few detergent/lipid molecules are normally identifiable.
In the case of structures from crystals grown in the lipidic cubic phase, however, this is something that really needs consideration, as often one can observe many alkyl tails, but headgroups are only rarely ordered. See the attached snapshot from Coot for an example where many such tails are observed packing in a roughly hexagonal array (the map here was calculated from a refinement with the “lipids” omitted, and slabbed parallel to the membrane plane).
I can think of three ways to address this, as follows:
1. Leave the densities unmodeled, as we cannot positively identify the components that they represent.
2. Model them as arbitrary alkyl fragments of monoolein, which is likely the major component of the “membrane” within the crystal.
3. Model them as generic alkane chains of different lengths.
I think we can exclude the first option right away. For the example from which the attached snapshot is derived, the R/Rfree with the lipids modeled as alkanes of various lengths is 0.204/0.247; Running the same refinement without the lipids leads to an R/Rfree of 0.243/0.265, so their inclusion obviously helps matters and I would therefore rather not leave this density unmodeled.
Both of the other ways are unsatisfactory for different reasons. On the one hand, I am reluctant to deposit a structure with alkanes in the coordinates, because obviously they are not there in the crystal – the protein was crystallized in a monoolein matrix, not petroleum. On the other hand, I am also reluctant to assign what are clearly alkyl chains as some arbitrary fragment of monoolein when firstly I am not sure that they are all monoolein (some will be detergent and some will be copurifiying lipids from the expression organism), and secondly it is just plain tedious to assign dozens of fragmentary lipid molecules as presumed fragments of monoolein.
I wonder if a solution might be to create new residues containing alkyl chains of various lengths, named something like U01, U02, U(n), where N is the length of the alkyl chain that fits the density. Sort of similar to the way one might leave UNK residues in a peptide of unidentified sequence. Or maybe there is a better way of doing this? Would appreciate any suggestions.
I would like to add a follow-up question regarding the implications of the presence of a quasi-ordered lipidic cubic phase for bulk solvent correctons, but that can wait for another time as this is already rather overlong.
Best,
Oliver.
|