Exercise 4. As proteins are made, they fold in specific three-dimensional structures that determine their functionality in the cell milieu. This folding process is assisted by other proteins called ‘chaperones’. Algorithms to predict the final 3D structure starting from the linear sequence of aminoacids are a current subject of study. Central to this is the fact that a number or proteins have been crystallized and their molecular three-dimensional structure established by X-ray diffraction, Nuclear Magnetic Resonance (NMR) or other methods. These better known proteins can act as templates on which to model unknown proteins of similar primary structure (upwards of 20% identity), which may be more difficult to crystallise in standard conditions. Here you will make use of a very fine piece of software to visualize the normal three-dimensional structure of rhodopsin, and try to correlate the occurrence of mutations of known clinical effect to the possible mechanisms whereby they disturb function. The purpose of this lab is to give you a feel for the type of question that can be asked in the process of linking genotype to phenotype.

Available structural data and their nature. Eukaryotic membrane proteins such as rhodopsin are difficult to crystallise, which in turn makes the obtaining of structural data, such as X ray diffraction patterns, difficult. In fact, no experimental model based on X-ray or NMR data existed for any vertebrate rhodopsin until August 2000 (Palczewski et al, Science 289;739-745). Rhodopsin models based on electron diffraction did exist (Schertler GFX, 1998), and although their level of resolution was insufficient to yield atomic-level structure, (about 6A vs <2A obtained with X-ray diffraction), they have been very useful in providing the main guidelines on which the more detailed data were be integrated.

The Protein Data Bank (PDB) includes two entries (1BOJ.pdb and 1BOK.pdb) for a theoretical model of the bundle of seven transmembrane helices (TMH) of the bovine rhodopsin (Pogozheva et al, 1997), in  association with trans- and cis-retinal respectively. The model draws data and seeks to reconcile results from  many other studies, such as primary sequence, electron diffraction profiles, limited proteolysis or mutagenesis. The stability of the molecule is optimised by a choice of side chain rotameres and by maximising the number of hydrogen bonds formed between the polar side chains within and between the helices, considering H-bonds as distance constraints. The model ignores, but is compatible with, the additional presence of eleven disulfide bonds between cystein residues that the model leaves in close proximity. Other rhodopsin-related structures found on PDB refer to proteins that interact with rhodopsin, and  to X-ray or NMR data on bacteriorhodopsin, crystallised in toto or in part. This is a considerably different protein that occurs as a crystalline lattice in the cell membrane of Halobacterium salinarium, easier to study with crystallographic methods.

            Verify that you have access to these tools, and then you can attempt the following exercises:

• Open 1BOJ.pdv from SPDBV 3.51 and get used to the main options in the menus. Go first to the rightmost 'Window' option at the top and open the control panel. On the 'sidechain' section of these window, deselect all the side chains. You will see a more manageable representation of the bundle of seven alpha helices characteristic of many membrane proteins. From left to right, you will find buttons to displace, enlarge and/or reduce, and rotate the molecule so as to provide a variety of points of view. You can measure distances and angles. You can identify each atom in each residue. Observe the position and structure of retinal in the retinal pocket.
• Open 1BOK.pdv and observe the difference. In both cases, tabulate the extent of the alpha helices.
• In both cases, tabulate the aminoacids that are within 3, 4 or 5 A from cis- or trans-retinal. To do this, open the select menu and choose 'Neighbours of selected aa', having selected retinal beforehand. Identify the aminoacid covanlenty bound to retinal.
• Using the 'Mutate' button, observe whether any of the mutations recorded in the Iowa database may act by altering the electronic or charge environment in the retinal pocket.
• Observe the delicate network of hydrogen bonds within the alpha helices. To do this, go to the Tools menu and and select 'Compute hydrogen bonds'. Tabulate all the hydrogen bonds in one of the helices, including the backbone structure and the side chains. Find if there are any inter-helical hydrogen bonds. Find if any of these is altered by pathogeneic mutations.
• Examine the alpha helices in the helical wheel, and make observations on the distribution of aminoacids in the polar projection.
• PDB entries 1KPN, 1KPW and 1KPX are theoretical models for the human cone photoreceptor pigments for blue, green and red, respectively. Notice that these models contain the intra and extracellular loops. Any other differences, among them and with rhodopsin? (NB Try the Sequence Alignment button at the beginning of the 'Protein Seq' page.
• Finally, write a 2 page long illustrated WORD document summarising your results and conclusions.