Team:MichiganState/Model
Modeling
Modeling and Analysis of the Formate Hydrogenlyase Activator Protein (FhlA) Structure
The primary amino acid sequence of the formate hydrogenlyase transcriptional activator protein (FhlA) was submitted to the Phyre2 web portal for protein modeling, prediction and analysis.4 This program compiled a list of existing protein structures in the Protein Data Bank3 which were partially homologous to FhlA. Results—individually referred to as hits—allowed 98% of the FhlA residues to be modeled above a 90% confidence threshold. Secondary structure alignment predictions for each hit were also provided by the Phyre2 analysis. The top hit (PDB: 5M7N)2 aligned to 38% of the modeled secondary structure, near the C-terminus of FhlA, with one hundred percent confidence. Nine other structures covered similar regions of the c-terminal half of FhlA at similarly high confidence.
Hits aligning with the FhlA N-terminus covered smaller regions of the protein, and those with total confidence were less numerous than those covering the c-terminus-half of FhlA. Of the hits corresponding to the FhlA N-terminus of the FhlA sequence, one crystal structure (PDB: 4BWI)1 had three formate molecules bound to it. Using PyMOL, the FhlA model was aligned with this structure. Figure 1 shows this overlay, with the FhlA model in green and the 4BWI structure in purple. Analysis of this overlay was completed to determine if—and, moreover, how—the formate binding affinity of the FhlA protein could be increased.
Three formate molecules can be seen bound to solvent-exposed areas on this protein structure (Figure 1, purple spheres). The overlay of the two structures was used as a guide to identifying similar regions on the FhlA model where formate could potentially bind. From this initial overlay, it appeared that only 4BWI residues 1-189 mapped well to the FhlA model. The rest of the 4BWI chain, residues 190-421, did not map continuously with the FhlA, although shared structural commonalities with the central domain of the FhlA model.
One formate is bound to the 4BWI chain along the continuous alpha helix. Notably, this binding takes place almost precisely where the 4BWI overlay ceases to align with the model. Although there is significantly more curvature to the corresponding helix in the FhlA model, if it were not for this curve, it is possible that the structures could align better than the initial overlay depicts (Figure 1). To examine whether residues 190 through 421 of 4BWI did align with the unmapped, central domain of the FhlA model, a more elaborate overlay was necessary. This was done by performing a second overlay of only 4BWI residues 190-421 to the FhlA model (Figure 2, pink). In Figure 2, residues 190-421 of the initial 4BWI overlay were hidden for clarity. This final overlay was used for further analysis to predict where formate may bind to the modeled FhlA, as well as how its affinity to bind may be increased through site-direct mutagenesis.
Based on the known formate binding position in 4BWI and the fact that the FhlA structure was predicted without considering any potential ligand interactions, we predict that the conformation of the FhlA protein may differ significantly from the model when formate is present. An intriguing hypothesis is that formate binding may induce a substantial conformational change in FhlA, perhaps related to the formate-induced activation of FhlA protein activities. If this hypothesis is correct, the FhlA residues involved in formate binding may be particularly important in driving a conformational change involved in protein activation. The corresponding residue for formate binding in 4BWI is R193, and we show an alignment of this region's primary amino acid sequence in Figure 3. This is not surprising because, as formate is a negative compound, it is likely attracted to, and stabilized by, the positive arginine side chain.
The alignment in Figure 3 shows that the FhlA residue most likely to correspond to the formate binding site in 4BWI is an asparagine in position 174, which is adjacent to another asparagine at position 173 (Figure 3; top). Asparagine has an electronegative oxygen atom on the end of its side chain, which is positioned toward the formate molecule bound to the 4BWI overlay. The two adjacent asparagine residues in the FhlA protein may inhibit formate from interacting with that region of the protein, as they are more electronegative than arginine, for example.
To test this prediction, we suggest one experimental approach could involve site-directed mutagenesis of FhlA N174 to mutate it into a positively charged arginine. Experiments should be performed to determine the level of transcriptional activation that is inducible by a range of formate concentrations in both this mutant FhlA and in the unmodified FhlA activator. Another site-direct mutation is recommended to mutate the asparagine at position 173 to a smaller, neutral glycine residue.
These mutations, in combination, are predicted to have the greatest effect on increasing the formate binding affinity of the FhlA. To more quantitatively determine the appropriateness of the residue selection for mutation, distances between the mutated residues and the formate molecule were compared to the distances between the corresponding 4BWI residues and said formate molecule. As presented in Figure 4, measurements were made between formate and 4BWIR193. Also in this figure, measurements between formate and N174R of the FhlA model are shown.
The distance between an oxygen atom of formate and the nearest atom of R193 in 4BWI was found to be about 6.8 angstroms. Measurements were made between this same formate molecule and the nearest atoms of the mutated N174R of the FhlA model. This residue was seen to be substantially closer to the formate molecule than that measured in 4BWI. A potential implication of this increase in proximity is that the residue’s affinity for formate. Though beyond the scope of this study, more direct analysis of formate binding affinity to FhlA and potential point mutants could be conducted using experiments such as isothermal titration calorimetry (ITC).
References
1 Anders, K., Angerer, V., Widany, G.D., Mroginski, M.A., von Stetten, D., Essen, L.-O., Structure of the phytochrome Cph2 from Synechocystis sp. PCC6803, The Journal of Biological Chemistry, Volume 288, 30 October 2013, Pages 35714-35725, http://dx.doi.org/10.2210/pdb4bwi/pdb
2 Cornaciu, I., Fernandez, I., Hoffmann, G., Carrica, M.C., Goldbaum, F.A., Marquez, J.A., Crystal structure of NtrX from Brucella abortus in complex with ATP processed with the CrystalDirect automated mounting and cryo-cooling technology, Journal of Molecular Biology, Volume 429, Issue 8, 21 April 2017, Pages 1192-1212, https://doi.org/10.1016/j.jmb.2016.12.022
3 Helen M. Berman, John Westbrook, Zukang Feng, Gary Gilliland, T. N. Bhat, Helge Weissig, Ilya N. Shindyalov, Philip E. Bourne, The Protein Data Bank, Nucleic Acids Research, Volume 28, Issue 1, 1 January 2000, Pages 235–242, https://doi.org/10.1093/nar/28.1.235
4 Kelley LA et al. The Phyre2 web portal for protein modeling, prediction and analysis, Nature Protocols, 10, 845-858 (2015).