February 2002 Research Report
The E. coli SRP model KFFH.pdband x-ray crystal structure of the E. coli SRP receptor 1FTS.pdb (Sinning et al.) were both superimposed on the nitrogenase iron protein structure 1N2C.pdb as previously done by Sinning et al. The nitrogenase iron protein file consisted of eight segments, only 4 of which (EF and GH) were paired homodimers. Superimposition of the EF and GH homdimers gave an RMSD of only 0.042313, indicating that these structures are equivalent. Superimposition of KFFH on the E protein yielded a RMSD of 0.843; the same superimposition on the F protein yielded an RMSD of 0.841. Since these segments of the homodimer also are nearly equivalent, KFFH was arbitrarily superimposed on the E protein and 1FTS on the F protein. KFFH and 1FTS were superimposed on the EF homodimer using the program sup2pdbs (R. Gabdoulline).Some of the residues found at the interface in the complex were 138, 139, 140, 165, 195, 196, and 248 in KFFH and 333, 335, 339, 387, 388, and 422 in 1FTS. As a result of the superimpositions, it was noted that Arg333 from 1FTS could form an H-bond with the AlF from the ADP-AlF originally contained in the 1N2C structure without minimization. Also in 1FTS, Val304 could interact with the AlF and with the beta phosphate of the ADP, Gly305 could interact with the beta phosphate of the ADP, and Glu475 could interact with the adenine ring. From KFFH, Asp250 could H-bond to the adenine ring of the ADP, Thr114 could interact with the ADP alpha phosphate, Thr113 and Gly 109 could interact with the ADP beta phosphate, and Ala110 could interact with the ADP. In the model of the SRP:SR produced by superimposition onto 1N2C, it is apparent that the GTP binding regions of both proteins lie directly at the interface. Also superimposed into the potential GTP binding sites of both the SRP and the SR were the GDP-Mg2+ from 1NG1, GMPPNP from 1JPJ, and 42 other GTP molecules extracted from pdbs in the protein databank. These GTPs give a wide range of potential conformations that the GTP may occupy.
The templates used to create KFFH (1JPJ, 1NG1, 1FTS, 1J8M) were superimposed on KFFH in order to compare the placement of potentially important structural residues in the proteins. Arg138(1J8M) is in a similar location to Arg140(KFFH), but the orientation of the sidechain nitrogens differ. Arg138(1JPJ) and Arg138(1NG1) differ inlocation to the other arginines but at more similar themseleves. They are further from the GTP and at the terminal N are about 90 degrees to each other, but generally follow the same path. They also face into the solvent and are about 180 degrees turned from KFFH. Arg333(1FTS) faces the GTP and mimics the Arg140(KFFH) position, but is turned inward and located slightly differently. Arg333 of 1FTS should move (flip) to accommodate the GTP. The Arg140(KFFH) faces the GTP and the complex interface. Surprisingly, in the 2 structures with the GDP-Mg2+ (1NG1) and GMPPNP (1JPJ), Arg138 faces away from the phosphate. This could be because the GDP-Mg2+ was there (short a phosphate), yielding a different positioning. The arginine from 1JPJ does reach in this direction, but due to its place in the helix, it is too short.
The ADP/GTP in the receptor is parallel to that in the SRP. This not likely correct but comes from the 1N2C positioning where the ADPs are also parallel and at the interface of the 1N2C monomers. There are also 2 Arg residues in the vicinity of the ADP in 1N2C, but neither are positioned close enough to h-bond the ADP phosphate groups.
In KFFH, there is a His at position 127 (insertion) that forces the Arg140 further down the alpha-helix in relation to the other structures which all lack this residue.
Arg191(1J8M) faces directly to the solvent but is right 90 degrees from Arg386(1FTS). Arg193(KFFH) parallels 1J8M. Arg191(1JPJ) and Arg191(1NG1) are turned inward in comparison and are buried. Arg386(1FTS) is also facing into the solvent. Arg193(KFFH) could possibly contact the GTP is flipped.
In 1NG1, the Mg waters are coordinated by protein sidechain, as opposed to mainchain, atoms: Asp187:OD1, Gln144:NE2, and Thr112:OG1. The Arg191:CZ is quite close and could potentially flip inwards.
For comparison, in the1WQ1 Ras-Ras-GAP-GDP-Mg2+ complex, Asn114 and Lys117 were found to coordinate to the guanine ring from above and Ala146 and Lys147 were found to coordinate to the guanine ring from below. Asp30 coordinates the O2 of the GDP, Asp119 coordinates GTP nitrogens N1 and N2, and Ser17 coordinates the alpha phosphate of the GDP. Gly115 coordinates the O4 of the GDP. Also in the area are Ala18 and Lys16.
For further calculations on GTP-Mg2+ bound forms of KFFH and 1FTS, three representative GTPs were chosen in which the phosphates lie in alternative positions, extended and bent. The chosen GTPs are denoted 312, 334, and 335 for KFFH and 509, 531, and 532 from 1FTS. These correspond, respectively, to GTP 13, 35, and 36 from the original 42 collected GTP conformations (from pdbs: 1A8R, 1A9C, 1CKM, 1CKN, 1FRW, 1HWX, 1QRA, and 521P). The Mg2+ was originally coordinated by 4 extra water molecules. These have been omitted from the calculations.
In order to calculate pkas and electrostatics using UHBD, hydrogens were added to the GTPs using InsightII for pH 7.0. Whatif was used to add polar hydrogens to the proteins KFFH and 1FTS. The waters were deleted from the hydrated Mg2+ ion so that just the Mg2+ is now present. A new qtable.dat was created to be used with the SRP and SR bound GTP-Mg2+. CHARMM19 parameters are used in lieu of OPLS parameters for the protein portion. Parameters for the GTP were created from CHARMM27 parameters for ATP and GDP. The total charge for the GTP molecule is -4.0. The Mg is assigned a charge of 2+ and radius for a 6-coordinated Mg. The input file for the pka calculations was set with the following values: T=308K, Protein dielectric: 5 (low pka homolow pka), 15 (mid pka), 78.5 (homohigh pka), and Ionic Strength=150 mM. The His and Cys residues were tallied and also included in the input file. The homohigh and homolow calculations were performed at Ionic Strength=0. pka values were computed for KFFH and 1FTS-bound GTP-Mg2+. UHBD electrostatics were computed for both GTP-Mg2+ bound KFFH and 1FTS using a 1 Angstrom spacing and a 1503 grid with both a protein dielectric of 2 and solvent dielectric of 78.5 and with both protein and solvent dielectric equal to 78.5. pka calculations were made for low (protein dielectric= 5.0, solvent dielectric= 78.5, ionic strength=150 mM), mid (protein dielectric=15.0, solvent dielectric=78.5, ionic strength=150 mM), high (protein dielectric=78.5, solvent dielectric=78.5, ionic strength=150 mM), homolow (protein dielectric=5.0, solvent dielectric=5.0, ionic strength=0 mM) and homohigh (protein dielectric=78.5, solvent dielectric=78.5, ionic strength=0 mM). These pka calculations were originally done for KFFH and 1FTS with GTP-Mg2+ with a solvated Mg2+ ion and are being redone. The results of the UHBD computations are as follows:
UHBD (E=electrostatic energy)
kffhgtp312: E (protein dielectric= 2, solvent dielectric=78.5): 0.171449E+05
E (protein and solvent dielectric=78.5): 0.484500E+03 kcal/mol
total charge on list of atoms: -8.110
kffhCOMgtp312.ins (-3.301424 to 0.9145226)
kffhCOMgtp312u.ins (-1.674428 to 0.3070242)
kffhgtp334: E (protein dielectric= 2, solvent dielectric=78.5): 0.170849E+05 kcal/mol
E (protein and solvent dielectric=78.5): 0.484504E+03 kcal/mol
total charge on list of atoms: -8.110
kffhCOMgtp334.ins (-3.340601 to 0.9195938)
kffhCOMgtp334u.ins (-1.529451 to 0.3066110)
kffhgtp335: E (protein dielectric= 2, solvent dielectric=78.5): 0.166485E+05 kcal/mol
E (protein and solvent dielectric=78.5): 0.470891E+03 kcal/mol
total charge on list of atoms: -8.110
kffhCOMgtp335.ins (-3.363526 to 0.9124098)
kffhCOMgtp335u.ins (-1.588834 to 0.3066230)
1ftsgtp509: E (protein dielectric= 2, solvent dielectric=78.5): 0.169931E+05 kcal/mol
E (protein and solvent dielectric=78.5): 0.480187E+03 kcal/mol
total charge on list of atoms: -9.240
1ftsCOMgtp509.ins (-2.880227 to 1.040530)
1ftsCOMgtp509u.ins (-1.814922 to 0.4106095)
1ftsgtp531: E (protein dielectric= 2, solvent dielectric=78.5): 0.174642E+05 kcal/mol
E (protein and solvent dielectric=78.5): 0.493380E+03 kcal/mol
total charge on list of atoms: -9.240
1ftsCOMgtp531.ins (-2.543298 to 1.040091);
1ftsCOMgtp331u.ins (-1.797642 to 0.4100578)
1ftsgtp532: E (protein dielectric= 2, solvent dielectric=78.5): 0.169652E+05 kcal/mol
E (protein and solvent dielectric=78.5): 0.478373E+03 kcal/mol
total charge on list of atoms: -9.240
1ftsCOMgtp532.ins (-2.006805 to 1.040118)
1ftsCOMgtp332u.ins (-1.394890 to 0.4091571)
Preliminary results with SDA for KFFH bound GTP312 (P1) and 1FTS bound GTP509 (P2) indicate that there seem to be 4 binding modes. Both proteins are highly negative in these calculations though. None correspond to the model produced by superimposition with 1N2C. The most highly probable from this analysis is one in which the GTP domains face the same direction but in which the 1FTS N-domain contacts the KFFH G-domain at a 45 degree angle of 1FTS off of KFFH. In this model, Glu198, Lys205, Asp202, Glu237, Lys233, and Asn230 from KFFH and Glu243, Arg246 and Asp225 from 1FTS seem to play a role in the interface.
Preliminary results with SDA for apo-KFFH (P1) and apo-1FTS (P2) indicate that the proteins are repelling each other because they are so highly negatively-charged.