If a protein involved in adhesion (cell adhesion molecule) is to be crystallized, does it 'aggregate' by binding to the region containing active binding residues?
No, as you have crystal contacts and materials in the experiments that cause precipitation (as PEGs or high salts) you may get other interaction modes which are not the lower energy.
The best is ti check it out by doing an energy minimization after having determined the structure.
balaji
Another important point is the dynamical view of binding: even if one crystallizes the right binding mode we need to bear in mind that it is just a picture of a event that can only be described in a movie. We are missing the dynamics of the bound and pre-bound system at realistic pressure, temperature and solvent composition. Binding dynamics influence dissociation constants, residence times etc...
Proteins (enzymes) are not always in their lowest energy states. You can have active site residues adopting unusual, high energy ("disallowed") conformations in order to reduce the reaction activation energy to go from substrate to product. Alleviation effect, the strain is taken by (several of) the active site residues. This is observed in crystallised enzymes and does not involve crystal packing effects. So the answer is no. Further the active site is usually away from the surface, well "hidden" from the outside (a cavity, a cleft, sometimes a channel leading to the active site)
This question becomes IMPORTANT only when considering how you intend to solve the structure.
Proteins always crystallize in minumum energy conformations. However, it is difficult to say whether the conformation observed in the crystal structure is the absolute minimum in the energy landscape. Most likely, it is not. As mentioned by Frederic, active site residues may adopt higher energy conformations which may push the overall energy of the structure higher, into one of the numerous local minima that the landscape may provide.
Dear Ravi,
in a crystal of a single CAM, you can observe of course only the self-interaction (homophilic). When you crystallize from a solution, usually only species that are allowed (local minima) and populated (global minima) in solution crystallize, so the observed interaction should be existent also in solution. A crystal will show you a global minimum structure, otherwise the crystal would not be stable. But one protein have several interactions in all directions in a polymeric crystal, so there is not a single interaction in its binding region, but several minima to partners in all space directions.
Due to the high concentration and unusual conditions (PEG,Salt, pH) you also populate weak interactions in the crystal and one way to distinguish these crystal contacts in the polymeric crystal from the contacts that are also present in physiological solution at a given concentration is their free energy of binding. A minimum under crystallisation condition may not be a minimum at physiological conditions!!!
The PISA server tries to estimate which contacts are "real" and which contacts are only present due to the crystal lattice. Remember that you need to expand the crystall symmetry eg. by Pymol to see all homophilic contacts, since the asymmetric unit rarely contain the biological unit.
And remember, a protein is never allone in a crystal, also when the asymmetric unit is a monomer...
Good luck,,
Marco
There are things like Oswald ripening, where big crystals grow at the expense of small ones. Or a thermodynamically more stable polymorph growing at the expense of a fast-growing one. So, _for those specific conditions _, the system goes to the minimum energy state, providing you wait long enough (indefinitely if there is a kinetic trap, e.g. diamond versus graphite). However, the specific conditions are different from the low energy states in solution, or interactions with a 2D surface. Or put in another way, the enzyme-substrate system differs from the enzyme-substrate-precipitant system (which is why one of them gives you crystals).
As for the crystal contacts coinciding with the adhesion contacts: the only way to make sure is to solve the structure and compare with fingerprinting or mutagenesis data. I think it should't, but one would need much more info to speculate, e.g. how does secretion work for your system etc. The ease with which you can over-express the protein could give hints.
In my experience, a protein has been crystallized into several different space groups and the packing is totally different. So the energy minimum we assumed to get is just the one picked by the crystallization condition.
If I have understood correctly work in the Alber group has suggested that many crystal structures solved at low (cryo-cooled) temperatures bias the structure to a low energy state. Room temperature crystallography of the same proteins can reveal alternate side chain conformations. See http://www.ncbi.nlm.nih.gov/pubmed/21918110 and http://www.ncbi.nlm.nih.gov/pubmed/20499387
On adhesion proteins I work previously on the CD2-CD48/CD58 adhesion system where CD2 crystallised as a homodimer that mimics the CD2/CD4(5)8 heterodimer, and we demonstrated by NMR that CD2 in solution formed weakly associated dimers. If interested see e.g. http://www.ncbi.nlm.nih.gov/pubmed/10526406 http://www.ncbi.nlm.nih.gov/pubmed/8634239 and other papers from e.g. EY Jones, SJ Davis on this system.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Computer Graphics and Animation
- 3D