Mini assignment
This protein is formally a 1,2-dihydrovomilenine 19,20-reductase. Retrieve this protein sequence from NCBI (genbank WKU61918.1). When modelling a protein, best practice is to model the protein in its native state, with the correct oligomeric state and any structurally important co-factors. While you can often figure this out from sequence and structural clues alone, the easiest way to get a quick read is to look at close homologs with experimentally determined structures.
Use blast to search the pdb for homologs. Visit the PDB webpage of its closest structural homolog, and infer what the most likely oligomeric state of this protein is. Note that for this protein family, zinc is an important co-factor; if you scroll down the pdb page for any of these structures, you will find zinc listed under small molecules. Use AlphaFold 3 to predict the oligomeric structure of this protein. Be sure to include the zinc ion in your prediction, by adding 2 Zn2+ (one zinc per chain) under the ion tab. Make a figure showing, in panel A, the structure of all 5 predicted models superimposed (pymol align is fine for this purpose), as cartoon and sticks, and zinc as spheres, coloured by pLDDT.
Show the PAE graph for this oligomeric model as panel B. In the caption, assess the quality of the predicted structure (referring to both pLDDT and PAE, and how closely different features superimpose). Note any regions where you think the predicted structure is unreliable. Note – for all figures, use a white background, maximal quality, ray trace, fancy helices, and in general make sure that the figure looks polished and professional; include a suitable caption that explains to the reader exactly what they are looking at, as well as answers any questions are superimposing, where used. Pay attention to colour, framing, and use consistent viewpoints unless there is a concrete reason not to.
Using Chai-1, model the same oligomeric complex, but in this prediction, also model the substrate of the enzyme (1,2-dihydrovomilenine; PubCHEM has an entry) and the reducing agent, NADPH (note – I am telling you what the ligands are here. If you are working on a novel protein, you will need to look at several structures, and read the relevant papers, to understand how the protein family works). In my hands, Chai-1 crashed when asked to also model the zinc ion too, so you can omit this if necessary. Make a second figure showing a single chain from the top Alphafold and Chai-1 structures superimposed. Colour the Alphafold model white, and the Chai-1 model cartoon by pLDDT. Include the ligands as sticks. Note whether there are any substantive differences in the structures, or confidence scores. These figures should be from a consistent perspective.
For a second panel, zoom into the active site, showing predictions of the ternary complex superimposed on one another, focussing on the substrate and co-factor. Are the substrate and co-factor predicted to bind in the same way in each active site? What does this suggest to you? Using the results of your earlier BLAST search of the pdb, visit each page of each until you find one with substrate bound, and one where NADP(H) is bound (there should be one in the first handful of hits. Superimpose these structures on your Chai-1 structural prediction.
Locate the catalytic site in the structural homolog, using the site of substrate binding to guide you. Does the orientation of NADPH follow the experimental structure (co-factor binding is generally conserved)? Does Chai-1 place 1,2-dihydrovomilenine in the catalytic site? Compare your predictions; you may have a subset of the predictions which are in a more plausible binding mode than others. In which case, focus on the best available, not the one with the highest overall score (as this score is dominated by the protein). In any of the predictions, does the NADPH nicotinamide get placed correctly relative to the atoms that will receive the hydride ion (figure 1 of the paper shows the groups being reduced; the hydride ion is transferred from NC4 of the nicotinamide ring; this needs to be placed in van der Waals contact with the atom receiving the hydrogen).
Make a pair of additional panels showing a detailed cartoon and sticks representation showing the active site of your protein and the pdb comparator (zoom into this region – you need not show all of the NADPH). Show the cartoon, and then the substrate(s) and all residues in contact with the substrate(s) as sticks. Do not show backbone atoms for the sticks unless they contribute hydrogen bonds. Do not hide cartoons where you show sticks. Show any hydrogen bonds as dashed lines. Show a transparent all-white surface, made with hydrogen atoms included, that shows the shape of the pocket. You can use whatever colours you see fit for the protein, though O/N/S should be red/blue/yellow-orange, and the ligand and NADPH should be in a contrasting colour. One panel should show 1,2-dihydrovomilenine 19,20-reductase, the second should show the homolog. Both should be from the same perspective, with the same framing to allow comparison.
Does this pose represent a productive ternary complex, or would you need to find some alternative way to model this complex to fully visualize it?
Last Completed Projects
| topic title | academic level | Writer | delivered |
|---|