Quantitative Assessment of Protein Structural Models by Comparison of H/D Exchange MS Data with Exchange Behavior Accurately Predicted by DXCOREX
Journal of the American Society for Mass Spectrometry
Tong Liua, Dennis Pantazatosb, Sheng Lia, Yoshitomo Hamuroc, Vincent J. Hilserd and Virgil L. Woods Jra*
*Virgil L. Woods Jr. M.D.
Univ. of Calif. San Diego
9500 Gilman Drive, mc 0656
La Jolla CA 92093
E-mail:
Supplementary Material
Accuracy of COREX-calculated exchange rates is improved by use of structures with explicitly-placed hydrogen: H-COREX
H-COREX, in which exchange competency is based on the accessible surface area (ASA) of explicitly placed amide hydrogen, was applied to five of the proteins used in the early validation studies of the original COREX algorithm, which had relied on calculation of amide nitrogen ASA to determine exchange competency[1]. The theoretical amide-specific exchange protection factors determined for hen egg-white lysozyme (HEL) by both the COREX algorithm and H-COREX are shown in Figure S1 along with the comparable NMR experimental data. While the results of the conventional COREX and H-COREX computations largely agree with each other and the NMR data, twelve of the 129 COREX-calculated amide-specific protection factors in HEL that disagreed with the NMR data were correctly calculated by H-COREX (black circles). Four additional proteins that had been previously employed for COREX validation [1] were similarly investigated using the H-COREX algorithm. Analysis of equine lysozyme (129 residues) demonstrated that H-COREX correctly calculated protection factors for the 10 residues with COREX calculations that were in disagreement with NMR determinations (Supplemental Material Figure S2). While COREX analysis of bovine pancreatic trypsin inhibitor (BPTI), turkey ovomucoid third domain (OMKTY) and staphylococcal nuclease (SN) produced protection factors that were in close agreement with NMR determinations, the few inaccuracies (3/56 residues for BPTI, 2/56 residues for OMKTY, and 1/141 residues for SN were correctly determined by H-COREX analysis[1].
Proteins Studied
For some proteins, the primary amino acid numbering system used in the published DXMS studies was different from the numbering system used in the corresponding PDB-published structures. In this paper, we adjusted the PDB numbering system to match that of the published data. Unless otherwise specified, deuteration levels are deuterons incorporated into probe peptides.
IB
The experimental deuteron incorporation of 18 peptide fragments of IB bound to NF-B following 2 minutes of on-exchange were available from Table 1, Pg 18954 of the referenced publication[2] and in the publication, the exchange results of 10 fragments were decorated onto the 3-D structure of IB (Fig. 5, Pg 18954). The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure of IB - NF-B complex (PDB ID: 1NFI:A, B, E)[3]. The N-terminal residues 66-69 of the protein were ignored as they were missing in the crystal structure. Thus, residues 70-282 of IB and all NF-B residues shown in the structure were subject to H-COREX calculation. Experimental data associated with the N-terminal IB peptide 66-80 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench is instituted were corrected by an average loss factor in the published study.
D/D RIIPKA
The experimental deuteron incorporation of 11 peptide fragments following 4 on-exchange time points were available from Fig. S2 in the Supplementary Material of the referenced publication[4]. The experimental data sets were further deconvoluted and results were represented in 2D ribbon maps of Fig. 5A(-), Pg 2988[4]. The D/D domain is known to dimerize at high affinity under the conditions employed in this study. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the D/D domain dimer(PDB ID: 1L6E:A, B)[5] and data deconvolution was performed using the method described in the Supplementary Material of the reference[4]. The C-terminal residues 47-48 at both chains were ignored as they were missing in the structure. Thus, residues 1-46 in the structure were subject to H-COREX calculation. The experimental data associated with N-terminal peptide fragment 37-48 and 42-48 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
Catalytic subunit of PKA E230Q mutation
The experimental deuteron incorporation of 76 peptide fragments following 6 on-exchange time points were available from Supplementary Fig. 2 of the referenced publication[6]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure(PDB ID: 1SYK)[7]. All residues (residues 1-350) in the structure were subject to H-COREX calculation. Subsequently, correlation analysis was performed between the experimental data and the theoretical data. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
-actinin CH2 domain
The experimental deuteron incorporation of 16 peptide fragments following 15 minutes of on-exchange were available from Table 1, Pg 2600 of the referenced publication[4]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure (PDB ID: 2EYI)[8]. All residues (residues 138-253) in the structure were subject to H-COREX calculation. Subsequently, correlation analysis was performed between the experimental data and the theoretical data. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
High-affinity Human Growth Hormone Variant (hGHv)
For unbound hGHv, the experimental deuteron incorporation of 17 peptide fragments following 4 on-exchange time points were available from Supplementary Figure 3 of the referenced publication[9]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure (PDB ID: 1HUW)[10]. Since residues 130-154 were missing in the published structure, only residues 1-129 and 155-191 in the structure were subject to H-COREX calculation. Experimental data associated with the fragments 127-146 and 149-157, as well as fragment 160-163 which is adjacent to the missing residues, were excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
For receptor bound hGHv, the experimental deuteration levels of 17 peptide fragments following 4 on-exchange time points were available from Supplementary Figure 3 of the referenced publication[9]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published complex structure (PDB ID: 1KF9: A, B, C)[11]. Residues 129-155 and the C-terminal residue 191 of hGHv were ignored as they were missing in the crystal structure. Thus, residues 1-128 and 156-190 of hGHv and all residues of receptor shown in the crystal structure were subject to H-COREX calculation, but only peptides in the hGHv protein with available corresponding DXMS data available were subject to correlation analysis. Experimental data associated with fragments 127-146, 149-157 and 185-191 as well as fragment 160-163 adjacent to the missing residues were excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
Mitogen-activated protein kinase p38
The experimental deuteron incorporation of 39 peptide fragments following 6 on-exchange time points were reported in the referenced publication[12-13],. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure (PDB ID: 1P38.). The N-terminal residues 1-14 and the C-terminal residues 366-371 were ignored as they were missing in the crystal structure. Residues 187-191 were also excluded in the calculation because they were deleted from the protein construct used for crystallization. Thus, residues 15-186 and 192-365 were subject to H-COREX calculation. Experimental data associated with peptide fragments 2-9, 180-193 and 357 – 368 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequentcorrelation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
AlkB
The experimental deuteron incorporation of 17 peptide fragments following 4 on-exchange time points were reported in the referenced publication[14] . The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure(PDB ID: 2FD8)[14]. The N-terminal residues 1-14 and the C-terminal residues 214-224 were ignored as they were missing in the crystal structure. Thus, residues 15-213 were subject to H-COREX calculation. Experimental data associated with peptide fragments 1-4, 5-19 and 203-224 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
A4V-Superoxide dismutase
The experimental deuteration levels of 11 peptide fragments following 4 on-exchange time points were reported in the referenced publication58, 59). The theoretical deuteration levels for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure(PDB ID: 1UXM)[15]. All residues in chain A (residues 1-153) in the structure were subject to H-COREX calculation. Subsequently, correlation analysis was performed between the experimental data and the theoretical data. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
WT- superoxide dismutase
The experimental deuteron incorporation of 11 peptide fragments following 4 on-exchange time points were reported in the referenced publication58, 59). The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure(PDB ID: 1PU0)[16]. All residues in chain A structure (residues 1-153), were subject to H-COREX calculation. Subsequently, correlation analysis was performed between the experimental data and the theoretical data. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
Free- PPAR LBD
The experimental deuteron incorporation of 23 peptide fragments following 6 on-exchange time points were reported in the referenced publication[17-18]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure (PDB ID: 1PRG: A, B)[18]. The N-terminal residues 1- 24 and the C-terminal residues 295 were ignored as they were missing in the crystal structure. Thus, residues 25-294 in both chain A and B were subject to H-COREX calculation. Experimental data associated with peptide fragments 2-29 and 288-295 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
Bound PPAR LBD
The experimental deuteron incorporation of 23 PPAR LBD peptides in the presence of rosiglitazone following 6 on-exchange time points were reported in the reverenced publication[17-18]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published complex structure (PDB ID: 2PRG: A, B, C)[18]. The N-terminal residues 1- 24 were ignored as they were missing in the crystal structure. Thus, residues 25-295 of PPAR LBD and all receptor residues shown in the crystal structure were subject to H-COREX calculation. Experimental data associated with peptide fragment 2-29 PPAR LBD had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
Cerezyme
The experimental deuteron incorporation of 31 peptide fragments following 4 on-exchange time points were reported in the referenced publication[19]. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the published structure (PDB ID: 2NT1)[19]. All residues in chain A (residues 1-497) in the structure were subject to H-COREX calculation. Subsequently, correlation analysis was performed between the experimental data and the theoretical data. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
GVIA-2 iPLA2 catalytic domain
The experimental deuteron incorporation of 22 peptide fragments following 7 on-exchange time points were available from Fig. 2, Pg23655 of the referenced publication[20] and additional data provided by the author of the paper (Dr. Yuan-Hao Hsu). In the publication, the exchange results were decorated onto a 3-D model of the GVIA-2 iPLA2 catalytic domain in Fig. 2, Pg23655. The theoretical deuteron incorporation for the same peptides and on-exchange times were calculated by DXCOREX analysis of the proposed model, with coordinates provided by the author (Dr. Yuan-Hao Hsu )[3]. The N-terminal residues 471-474, residues 606-621 and C-terminal residues 805-806 were ignored as they were missing in the proposed model. Thus, residues 475-604 and 623-804 were subject to H-COREX calculation. Experimental data associated with peptide fragments 473-480, 599-615, 614-629 and 804-806 had no corresponding determination in the DXCOREX calculation, and was excluded from the subsequent correlation analysis. Deuterium losses occurring after exchange quench were corrected by fragment specific loss factors measured during the published experiments.
References for Supporting Information
1. Hilser, V. J.Freire, E.: Structure-Based Calculation of the Equilibrium Folding Pathway of Proteins. Correlation with Hydrogen Exchange Protection Factors. J Mol Biol. 262, 756-772 (1996)
2. Truhlar, S. M., Torpey, J. W.Komives, E. A.: Regions of Ikappabalpha That Are Critical for Its Inhibition of Nf-Kappab.DNA Interaction Fold Upon Binding to Nf-Kappab. Proc Natl Acad Sci U S A. 103, 18951-18956 (2006)
3. Jacobs, M. D.Harrison, S. C.: Structure of an Ikappabalpha/Nf-Kappab Complex. Cell. 95, 749-758 (1998)
4. Full, S. J., Deinzer, M. L., Ho, P. S.Greenwood, J. A.: Phosphoinositide Binding Regulates Alpha-Actinin Ch2 Domain Structure: Analysis by Hydrogen/Deuterium Exchange Mass Spectrometry. Protein Sci. 16, 2597-2604 (2007)
5. Morikis, D., Roy, M., Newlon, M. G., Scott, J. D.Jennings, P. A.: Electrostatic Properties of the Structure of the Docking and Dimerization Domain of Protein Kinase a Iialpha. Eur J Biochem. 269, 2040-2051 (2002)
6. Yang, J., Garrod, S. M., Deal, M. S., Anand, G. S., Woods, V. L., Jr.Taylor, S.: Allosteric Network of Camp-Dependent Protein Kinase Revealed by Mutation of Tyr204 in the P+1 Loop. J Mol Biol. 346, 191-201 (2005)
7. Wu, J., Yang, J., Kannan, N., Madhusudan, Xuong, N. H., Ten Eyck, L. F.Taylor, S. S.: Crystal Structure of the E230q Mutant of Camp-Dependent Protein Kinase Reveals an Unexpected Apoenzyme Conformation and an Extended N-Terminal a Helix. Protein Sci. 14, 2871-2879 (2005)
8. Borrego-Diaz, E., Kerff, F., Lee, S. H., Ferron, F., Li, Y.Dominguez, R.: Crystal Structure of the Actin-Binding Domain of Alpha-Actinin 1: Evaluating Two Competing Actin-Binding Models. J Struct Biol. 155, 230-238 (2006)
9. Horn, J. R., Kraybill, B., Petro, E. J., Coales, S. J., Morrow, J. A., Hamuro, Y.Kossiakoff, A. A.: The Role of Protein Dynamics in Increasing Binding Affinity for an Engineered Protein-Protein Interaction Established by H/D Exchange Mass Spectrometry. Biochemistry. 45, 8488-8498 (2006)
10. Ultsch, M. H., Somers, W., Kossiakoff, A. A.de Vos, A. M.: The Crystal Structure of Affinity-Matured Human Growth Hormone at 2 a Resolution. J Mol Biol. 236, 286-299 (1994)
11. Schiffer, C., Ultsch, M., Walsh, S., Somers, W., de Vos, A. M.Kossiakoff, A.: Structure of a Phage Display-Derived Variant of Human Growth Hormone Complexed to Two Copies of the Extracellular Domain of Its Receptor: Evidence for Strong Structural Coupling between Receptor Binding Sites. J Mol Biol. 316, 277-289 (2002)
12. Wang, Z., Harkins, P. C., Ulevitch, R. J., Han, J., Cobb, M. H.Goldsmith, E. J.: The Structure of Mitogen-Activated Protein Kinase P38 at 2.1-a Resolution. Proc Natl Acad Sci U S A. 94, 2327-2332 (1997)
13. Hamuro, Y., Coales, S. J., Morrow, J., Griffin, P. R., Southern, M. R.C., W. P.: Application of Hydrogen/Deuterium-Exchange to P38 Mitogen-Activated Protein Kinase. Amer Biotechnol Lab 25, 28-30. (2007)
14. Yu, B., Edstrom, W. C., Benach, J., Hamuro, Y., Weber, P. C., Gibney, B. R.Hunt, J. F.: Crystal Structures of Catalytic Complexes of the Oxidative DNA/Rna Repair Enzyme Alkb. Nature. 439, 879-884 (2006)
15. Molnar, K. S., Karabacak, N. M., Johnson, J. L., Wang, Q., Tiwari, A., Hayward, L. J., Coales, S. J., Hamuro, Y.Agar, J. N.: A Common Property of Amyotrophic Lateral Sclerosis-Associated Variants: Destabilization of the Copper/Zinc Superoxide Dismutase Electrostatic Loop. J Biol Chem. 284, 30965-30973 (2009)
16. DiDonato, M., Craig, L., Huff, M. E., Thayer, M. M., Cardoso, R. M., Kassmann, C. J., Lo, T. P., Bruns, C. K., Powers, E. T., Kelly, J. W., Getzoff, E. D.Tainer, J. A.: Als Mutants of Human Superoxide Dismutase Form Fibrous Aggregates Via Framework Destabilization. J Mol Biol. 332, 601-615 (2003)
17. Nolte, R. T., Wisely, G. B., Westin, S., Cobb, J. E., Lambert, M. H., Kurokawa, R., Rosenfeld, M. G., Willson, T. M., Glass, C. K.Milburn, M. V.: Ligand Binding and Co-Activator Assembly of the Peroxisome Proliferator-Activated Receptor-Gamma. Nature. 395, 137-143 (1998)
18. Hamuro, Y., Coales, S. J., Morrow, J. A., Molnar, K. S., Tuske, S. J., Southern, M. R.Griffin, P. R.: Hydrogen/Deuterium-Exchange (H/D-Ex) of Ppargamma Lbd in the Presence of Various Modulators. Protein Sci. 15, 1883-1892 (2006)
19. Kornhaber, G. J., Tropak, M. B., Maegawa, G. H., Tuske, S. J., Coales, S. J., Mahuran, D. J.Hamuro, Y.: Isofagomine Induced Stabilization of Glucocerebrosidase. Chembiochem. 9, 2643-2649 (2008)
20. Hsu, Y. H., Burke, J. E., Li, S., Woods, V. L., Jr.Dennis, E. A.: Localizing the Membrane Binding Region of Group Via Ca2+-Independent Phospholipase A2 Using Peptide Amide Hydrogen/Deuterium Exchange Mass Spectrometry. J Biol Chem. 284, 23652-23661 (2009)
Figure Legends for Supporting Information
Figure S1. Natural logarithm of the COREX (blue) vs H-COREX (red)-calculated amide hydrogen exchange rate protection factors (LnPFcal) for hen egg-white lysozyme (HEL) are shown in the upper panel while NMR-determined exchange protection factors (LnPFexp) are shown in the lower panel (taken from Figure 3b reference [1]). The secondary structure of the protein is indicated at the top of panel. Proline has no exchangeable amide hydrogen, so lnPFs of residues flanking prolines are connected by dotted lines. Black circles indicate residues with protection factors correctly predicted by H-COREX vs COREX.
Figure S2. Equine Lysozyme: The natural logarithm of the COREX (blue) vsH-COREX (red)-calculated amide hydrogen exchange rate protection factors (LnPFcal) for equine lysozyme are shown in the upper panel while NMR-determined exchange protection factors (LnPFexp) are shown in the lower panel (taken from Figure 4b, reference [1]. The secondary structure of the protein is indicated at the top of panel. Proline has no exchangeable amide hydrogen, so lnPFs of residues flanking prolines are connected by dotted lines. Black circles indicate residues with protection factors correctly predicted by H-COREX vs COREX.