January 2002 Research Report

Douglas Freymann (Northwestern) finally ceded the coordinates of one his stuctures of Ffh from Thermus aquaticus bound GMPPNP (1JPJ.pdb) to us on December 11, 2001. This has been employed in the modeling.

To try to determine some residues that may be important for GTP-Mg²⁺ binding, I determined the residues within 5 of the GDP in Thermus aquaticus Ffh (1NG1.pdb) and those within 5 of the GMPPNP, also in T. aquaticus Ffh (1JPJ.pdb). I also compiled a table from this and the literature of possible contacts (to be used in modeling) and any support of these hypothesis. After making many models in Swiss-Model and checking the given alignments, these are the final models.

1. The initial SRP model (contained in /home/elkins/srp/FFH/SwissModel) is the E. coli Ffh complete sequence modeled by Swiss-Model (http://www.expasy.ch/swissmod/SWISS-MODEL.html) against 5 templates of the same organism: each of the three complete whole protein chains of apo-Ffh from T. aquaticus (2FFH.pdb), the apo NG fragment of T. aquaticus Ffh (3NG1.pdb), and the GDP-Mg²⁺ bound Ffh from T. aquaticus. Though the alignment for the NG portion is improved in the final model of this portion of the protein, the 2FFH.pdb is the only template of the entire protein, including the M domain. For modeling the complex of the SRP and receptor, this piece should be merged with the new model as it should be at the interface of these two proteins.

2. The final SRP model (contained in /home/elkins/srp/FFH/Jan15FFHModel-REV) is the E. coli Ffh NG fragment (residues 5-298)(M was cleaved) produced by Swiss-Model in the Optimize mode. Four templates were given to Swiss-Model in the First Approach Mode. These include 1JPJ.pdb (T. aquaticus Ffh NG fragment bound GMPPNP), 1FTS.pdb (E. coli apo-FtsY NG fragment), 1NG1.pdb (T. aquaticus Ffh NG fragment bound GDP-Mg²⁺), and 1J8M.pdb (A. ambivalens apo-Ffh NG fragment). The output model and alignment (contained in /home/elkins/srp/FFH/Jan15FFHModel) was adjusted using the SwissPDBViewer program. The changes included shifting the insertion of the templates to residue HIS127 in the model and deleting the insertion in the templates at residue Lys130 so that beta-sheet was not disrupted. The project was saved in SwissPDBViewer as a project .pdb file and re-submitted to Swiss-Model in the Optimise Project Mode to obtain the final model (kffh.pdb). The Arg193 in the model corresponding to Arg191 in T. aquaticus is flipped out.

3. I executed the scripts to calculate pka values for the model using UHBD and the OPLS parameter set. Polar hydrogens were added using WhatIF for pH 7.0.Flipping observed in Asn79, Gln90, His195 and His127. Upon discussion with Rebecca, it was decided that some of my parameters to determine the pkas could be improved, next time I will use: T=308K, Protein dielectric both 5 15, Ionic Strength=150 mM, and the His residues will be included properly. Since including the GTP-Mg2+ is likely to alter these calculations, it was determined that I should dock the GTP-Mg2+ prior to re-calculation. I have been doing this.

4. I superimposed all of the template structures (1FTS, 1J8M, 1NG1, and 1JPJ) onto the model, kffh.pdb, using Razif's program sup2pdbs. This superimposes all identical Calpha atoms in the structures. Using these superimpositions, the GNPPMP from JPJ and the GDP-Mg2+ from 1NG1 could be visualized in the model. The model looks reasonably good and the residues hypothesized to be involved in GTP-Mg2+ binding are close and reasonable. The RMSD values from superimposition onto kffh are: 1NG1, 141 atoms superimposed, RMSD= 2.932; 1JPJ, 141 atoms superimposed, RMSD=3.037; 1J8M, 104 atoms superimposed, RMSD=0.959; 1FTS, 94 atoms superimposed; RMSD=2.938. These were combined into a single file allsup.pdb with the chains names as follows: 1JPJ- A, 1NG1-B, 1FTS-C, 1J8M-F, and KFFH with none.

5. However, this left me with a model with only a GTP analogue or GDP bound- not GTP-Mg2+. I downloaded all molecules bound GTP or GTP-Mg2+ from the Protein Databank to see both which conformations the GTP could exist in and to have a reasonable GTP structure to dock initially. These come from GTP binding mutant proteins. There were 8 structures (1A8R, 1A9C, 1CKM, 1CKN, 1FRW, 1HWX, 1QRA, and 521P). I have superimposed these all in InsightII and have a file with all included in the transformed position, however InsightII crashes every time it tries to GET this molecule (due to chain naming). There are 42 GTP structures. I now want to superimpose these structures to the GMPPNP and the GDP from the templates, however I have not determined a good way of doing this yet.

6. Once I have modeled in the GTP-Mg2+, the model will need some energy minimization so that we can see which contacts are made.

7. I have superimposed using 2 programs, PROSUP and one by Kay Diederichs, the model E. coli SRP kffh and the receptor 1fts on the nitrogenase iron protein (1N2C, segment A of A-H). This protein has low sequence homology, but high fold homology, with the SRP and SRP receptor. This was used to determine a SRP:SR model by Irmi. I have not yet fitted the 2 superimposed segments together into one pdb file. Using the PROSUP program, in the kffh model, 83 equivalent residues were superimposed (onto 1N2C-A) with an RMSD=2.98 and in the fts receptor, 85 equivalent residues were superimposed with an RMSD=2.89. This program gives multiple alignments that can be compared, however.

Other useful information (to be added to web page):

Switch I and II regions: respond to nucleotide binding.

Switch I region includes: (in Ras) recognizes the RBD (Ras-binding domain) of the effector. Low affinity contacts of this region enhance the exchange of the GEF (guanine exchange factor) domain.

Switch II region includes: makes contacts with the kinase domain for the activation of the effector. In Ras, consists of Ala59, Gly60, Gly75, and Gln61. The region is responsible for the early and late conformational switches due to nucleotide binding.

Conference: I attended the ESS conference held at the University of Heidelberg on January 25-27, 2002 entitled Flexibility and Function of Proteins.