Solution structure of human IL-13 and implication for receptor binding.

Interleukin-13 has been implicated as a key factor in asthma, allergy, atopy and inflammatory response, establishing the protein as a valuable therapeutic target. The high-resolution solution structure of human IL-13 has been determined by multidimensional NMR. The resulting structure is consistent with previous short-chain left-handed four-helix bundles, where a significant similarity in the folding topology between IL-13 and IL-4 was observed. IL-13 shares a significant overlap in biological function with IL-4, a result of the common alpha chain component (IL-4Ralpha) in their respective receptors. Based on the available structural and mutational data, an IL-13/IL-4Ralpha model and a sequential mechanism for forming the signaling heterodimer is proposed for IL-13.

The discovery of anthranilic acid-based MMP inhibitors. Part 1: SAR of the 3-position.

A novel series of anthranilic acid-based inhibitors of MMP-1, MMP-9, and MMP-13 was prepared and evaluated both in vitro and in vivo. The most potent compound, 6e, has in vivo activity in a rat sponge-wrapped cartilage model.

MS/NMR: a structure-based approach for discovering protein ligands and for drug design by coupling size exclusion chromatography, mass spectrometry, and nuclear magnetic resonance spectroscopy.

A protocol is described for rapidly screening small organic molecules for their ability to bind a target protein while obtaining structure-related information as part of a structure-based drug discovery and design program. The methodology takes advantage of and combines the inherent strengths of size exclusion gel chromatography, mass spectrometry, and NMR to identify bound complexes in a relatively universal high-throughput screening approach. Size exclusion gel chromatography in the spin column format provides the high-speed separation of a protein-ligand complex from free ligands. The spin column eluent is then analyzed under denaturing conditions by electrospray ionization mass spectrometry (MS) for the presence of small molecular weight compounds formerly bound to the protein. Hits identified by MS are then individually assayed by chemical shift perturbations in a 2D 1H-15N HSQC NMR spectrum to verify specific interactions of the compound with the protein and identification of the binding site on the protein. The utility of the MS/NMR assay is demonstrated with the use of the catalytic fragment of human fibroblast collagenase (MMP-1) as a target protein and the screening of a library consisting of approximately 32 000 compounds for the identification of molecules that exhibit specific binding to the RGS4 protein.

Structural relatedness of distinct determinants recognized by monoclonal antibody TP25.99 on beta 2-microglobulin-associated and beta 2-microglobulin-free HLA class I heavy chains.

The association of HLA class I heavy chains with beta2-microglobulin (beta2m) changes their antigenic profile. As a result, Abs react with either beta2m-free or beta2m-associated HLA class I heavy chains. An exception to this rule is the mAb TP25.99, which reacts with both beta2m-associated and beta2m-free HLA class I heavy chains. The reactivity with beta2m-associated HLA class I heavy chains is mediated by a conformational determinant expressed on all HLA-A, -B, and -C Ags. This determinant has been mapped to amino acid residues 194-198 in the alpha3 domain. The reactivity with beta2m-free HLA class I heavy chains is mediated by a linear determinant expressed on all HLA-B Ags except the HLA-B73 allospecificity and on <50% of HLA-A allospecificities. The latter determinant has been mapped to amino acid residues 239-242, 245, and 246 in the alpha3 domain. The conformational and the linear determinants share several structural features, but have no homology in their amino acid sequence. mAb TP25.99 represents the first example of a mAb recognizing two distinct and spatially distant determinants on a protein. The structural homology of a linear and a conformational determinant on an antigenic entity provides a molecular mechanism for the sharing of specificity by B and TCRs.

High-resolution solution structure of the catalytic fragment of human collagenase-3 (MMP-13) complexed with a hydroxamic acid inhibitor.

The high-resolution solution structure of the catalytic fragment of human collagenase-3 (MMP-13) complexed with a sulfonamide derivative of a hydroxamic acid compound (WAY-151693) has been determined by multidimensional heteronuclear NMR. A total of 30 structures were calculated for residues 7-164 by means of hybrid distance geometry-simulated annealing using a total of 3280 experimental NMR restraints. The atomic rms distribution about the mean coordinate positions for the 30 structures is 0.43(+/-0.05) A for the backbone atoms, 0.80(+/-0.09) A for all atoms, and 0.47(+/-0.04) A for all atoms excluding disordered side-chains. The overall structure of MMP-13 is composed of a beta-sheet consisting of five beta-strands in a mixed parallel and anti-parallel arrangement and three alpha-helices where its overall fold is consistent with previously solved MMP structures. A comparison of the NMR structure of MMP-13 with the published 1.6 A resolution X-ray structure indicates that the major differences between the structures is associated with loop dynamics and crystal-packing interactions. The side-chains of some active-site residues for the NMR and X-ray structures of MMP-13 adopt distinct conformations. This is attributed to the presence of unique inhibitors in the two structures that encounter distinct interactions with MMP-13. The major structural difference observed between the MMP-13 and MMP-1 NMR structures is the relative size and shape of the S1' pocket where this pocket is significantly longer for MMP-13, nearly reaching the surface of the protein. Additionally, MMP-1 and MMP-13 exhibit different dynamic properties for the active-site loop and the structural Zn-binding region. The inhibitor WAY-151693 is well defined in the MMP-13 active-site based on a total of 52 distance restraints. The binding motif of WAY-151693 in the MMP-13 complex is consistent with our previously reported MMP-1:CGS-27023A NMR structure and is similar to the MMP-13: RS-130830 X-ray structure.

Solution structure of ZipA, a crucial component of Escherichia coli cell division.

ZipA, an essential component of cell division in Escherichia coli, interacts with the FtsZ protein at the midcell in one of the initial steps of septum formation. The high-resolution solution structure of the 144-residue C-terminal domain of E. coli ZipA (ZipA(185)(-)(328)) has been determined by multidimensional heteronuclear NMR. A total of 30 structures were calculated by means of hybrid distance geometry-simulated annealing using a total of 2758 experimental NMR restraints. The atomic root means square distribution about the mean coordinate positions for residues 6-142 for the 30 structures is 0.37 +/- 0.04 A for the backbone atoms, 0. 78 +/- 0.05 A for all atoms, and 0.45 +/- 0.04 A for all atoms excluding disordered side chains. The NMR solution structure of ZipA(185)(-)(328) is composed of three alpha-helices and a beta-sheet consisting of six antiparallel beta-strands where the alpha-helices and the beta-sheet form surfaces directly opposite each other. A C-terminal peptide from FtsZ has been shown to bind ZipA(185)(-)(328) in a hydrophobic channel formed by the beta-sheet providing insight into the ZipA-FtsZ interaction. An unexpected similarity between the ZipA(185)(-)(328) fold and the split beta-alpha-beta fold observed in many RNA binding proteins may further our understanding of the critical ZipA-FtsZ interaction.

NMR structure of free RGS4 reveals an induced conformational change upon binding Galpha.

Heterotrimeric guanine nucleotide-binding proteins (G-proteins) are transducers in many cellular transmembrane signaling systems where regulators of G-protein signaling (RGS) act as attenuators of the G-protein signal cascade by binding to the Galpha subunit of G-proteins (G(i)(alpha)(1)) and increasing the rate of GTP hydrolysis. The high-resolution solution structure of free RGS4 has been determined using two-dimensional and three-dimensional heteronuclear NMR spectroscopy. A total of 30 structures were calculated by means of hybrid distance geometry-simulated annealing using a total of 2871 experimental NMR restraints. The atomic rms distribution about the mean coordinate positions for residues 5-134 for the 30 structures is 0.47 +/- 0.05 A for the backbone atoms, 0. 86 +/- 0.05 A for all atoms, and 0.56 +/- 0.04 A for all atoms excluding disordered side chains. The NMR solution structure of free RGS4 suggests a significant conformational change upon binding G(i)(alpha)(1) as evident by the backbone atomic rms difference of 1. 94 A between the free and bound forms of RGS4. The underlying cause of this structural change is a perturbation in the secondary structure elements in the vicinity of the G(i)(alpha)(1) binding site. A kink in the helix between residues K116-Y119 is more pronounced in the RGS4-G(i)(alpha)(1) X-ray structure relative to the free RGS4 NMR structure, resulting in a reorganization of the packing of the N-terminal and C-terminal helices. The presence of the helical disruption in the RGS4-G(i)(alpha)(1) X-ray structure allows for the formation of a hydrogen-bonding network within the binding pocket for G(i)(alpha)(1) on RGS4, where RGS4 residues D117, S118, and R121 interact with residue T182 from G(i)(alpha)(1). The binding pocket for G(i)(alpha)(1) on RGS4 is larger and more accessible in the free RGS4 NMR structure and does not present the preformed binding site observed in the RGS4-G(i)(alpha)(1) X-ray structure. This observation implies that the successful complex formation between RGS4 and G(i)(alpha)(1) is dependent on both the formation of the bound RGS4 conformation and the proper orientation of T182 from G(i)(alpha)(1). The observed changes for the free RGS4 NMR structure suggest a mechanism for its selectivity for the Galpha-GTP-Mg(2+) complex and a means to facilitate the GTPase cycle.

Analysis by NMR spectroscopy of the structural homology between the linear and the cyclic peptide recognized by anti-human leukocyte antigen class I monoclonal antibody TP25.99*.

The anti-human leukocyte antigen (HLA) class I monoclonal antibody (mAb) TP25.99 has a unique specificity since it recognizes both a conformational and a linear determinant expressed on the beta(2)-mu-associated and beta(2)-mu-free HLA class I heavy chains, respectively. Previously, we reported the identification of a cyclic and a linear peptide that inhibits mAb TP25.99 binding to the beta(2)-mu-associated and beta(2)-mu-free HLA class I heavy chains (S. A. Desai, X. Wang, E. J. Noronha, Q. Zhou, V. Rebmann, H. Grosse-Wilde, F. J. Moy, R. Powers, and S. Ferrone, submitted for publication). The linear X(19) and cyclic LX-8 peptides contain sequence homologous to residues 239-242, 245, and 246 and to residues 194-198, respectively, of HLA class I heavy chain alpha(3) domain. Analysis by two-dimensional transfer nuclear Overhauser effect spectroscopy of the induced solution structures of the linear X(19) and cyclic LX-8 peptides in the presence of mAb TP25.99 showed that the two peptides adopt a similar structural motif despite the lack of sequence homology. The backbone fold is suggestive of a short helical segment followed by a tight turn, reminiscent of the determinant loop region (residues 194-198) on beta(2)-mu-associated HLA class I heavy chains. The structural similarity between the linear X(19) and cyclic LX-8 peptides and the lack of sequence homology suggests that mAb TP25.99 predominantly recognizes a structural motif instead of a consensus sequence.

NMR solution structure of the catalytic fragment of human fibroblast collagenase complexed with a sulfonamide derivative of a hydroxamic acid compound.

The solution structure of the catalytic fragment of human fibroblast collagenase (MMP-1) complexed with a sulfonamide derivative of a hydroxamic acid compound (CGS-27023A) has been determined using two-dimensional and three-dimensional heteronuclear NMR spectroscopy. The solution structure of the complex was calculated by means of hybrid distance geometry-simulated annealing using a combination of experimental NMR restraints obtained from the previous refinement of the inhibitor-free MMP-1 (1) and recent restraints for the MMP-1:CGS-27023A complex. The hydroxamic acid moiety of CGS-27023A was found to chelate to the "right" of the catalytic zinc where the p-methoxyphenyl sits in the S1' active-site pocket, the isopropyl group is in contact with H83 and N80, and the pyridine ring is solvent exposed. The sulfonyl oxygens are in hydrogen-bonding distance to the backbone NHs of L81 and A82. This is similar to the conformation determined by NMR of the inhibitor bound to stromelysin (2, 3). A total of 48 distance restraints were observed between MMP-1 and CGS-27023A from 3D 13C-edited/12C-filtered NOESY and 3D 15N-edited NOESY experiments. An additional 18 intramolecular restraints were observed for CGS-27023A from a 2D 12C-filtered NOESY experiment. A minimal set of NMR experiments in combination with the free MMP-1 assignments were used to assign the MMP-1 (1)H, 13C, and 15N resonances in the MMP-1:CGS-27023A complex. The assignments of CGS-27023A in the complex were obtained from 2D 12C-filtered NOESY and 2D 12C-filtered TOCSY experiments.

Homology model for oncostatin M based on NMR structural data.

Oncostatin M (OM) is a member of the cytokine family which regulates the proliferation and differentiation of a variety of cell types and includes interleukin-6 (IL-6), leukemia inhibitory factor (LIF), and granulocyte-colony stimulating factor (G-CSF). This family of proteins adopts a four-helix bundle fold with up-up-down-down topology and contains intramolecular disulfide bonds. Since an X-ray or NMR structure for OM is not currently available, a homology model for OM was determined from the X-ray structures of human growth hormone (hGH), LIF, and G-CSF where the alignment was based on secondary structure instead of sequence. The OM secondary structure was determined from NMR structural data, and the secondary structures for hGH, LIF, and G-CSF were obtained from the reported X-ray structures. The resulting homology model was refined using sequential NOE distance 13C restraints, chemical shift information, and a conformational database.

High-resolution solution structure of the inhibitor-free catalytic fragment of human fibroblast collagenase determined by multidimensional NMR.

The high-resolution solution structure of the inhibitor-free catalytic fragment of human fibroblast collagenase (MMP-1), a protein of 18.7 kDa, which is a member of the matrix metalloproteinase family, has been determined using three-dimensional heteronuclear NMR spectroscopy. A total of 30 structures were calculated by means of hybrid distance geometry-simulated annealing using a total of 3333 experimental NMR restraints, consisting of 2409 approximate interproton distance restraints, 84 distance restraints for 42 backbone hydrogen bonds, 426 torsion angle restraints, 125 3JNH alpha restraints, 153 C alpha restraints, and 136 C beta restraints. The atomic rms distribution about the mean coordinate positions for the 30 structures for residues 7-137 and 145-163 is 0.42 +/- 0.04 A for the backbone atoms, 0.80 +/- 0.04 A for all atoms, and 0.50 +/- 0.03 A for all atoms excluding disordered side chains. The overall structure of MMP-1 is composed of a beta-sheet consisting of five beta-strands in a mixed parallel and anti-parallel arrangement and three alpha-helices. A best-fit superposition of the NMR structure of inhibitor-free MMP-1 with the 1.56 A resolution X-ray structure by Spurlino et al. [Spurlino, J. C., Smallwood, A. M., Carlton, D. D., Banks, T. M., Vavra, K. J., Johnson, J. S., Cook, E. R., Falvo, J., and Wahl, R. C., et al. (1994) Proteins: Struct., Funct., Genet. 19, 98-109] complexed with a hydroxamate inhibitor yields a backbone atomic rms difference of 1.22 A. The majority of differences between the NMR and X-ray structure occur in the vicinity of the active site for MMP-1. This includes an increase in mobility for residues 138-144 and a displacement for the Ca(2+)-loop (residues 74-80). Distinct differences were observed for side-chain torsion angles, in particular, the chi 1 for N80 is -60 degrees in the NMR structure compared to 180 degrees in the X-ray. This results in the side chain of N80 occupying and partially blocking access to the active site of MMP-1.

Assignments, secondary structure and dynamics of the inhibitor-free catalytic fragment of human fibroblast collagenase.

Fibroblast collagenase (MMP-1), a 169-residue protein with a molecular mass of 18.7 kDa, is a matrix metalloproteinase which has been associated with pathologies such as arthritis and cancer. The assignments of the 1H, 15N, 13CO and 13C resonances, determination of the secondary structure and analysis of 15N relaxation data of the inhibitor-free catalytic fragment of recombinant human fibroblast collagenase (MMP-1) are presented. It is shown that MMP-1 is composed of a beta-sheet consisting of five beta-strands in a mixed parallel and antiparallel arrangement (residues 13-19, 48-53, 59-65, 82-85 and 94-99) and three alpha-helices (residues 27-43, 112-124 and 150-160). This is nearly identical to the secondary structure determined from the refined X-ray crystal structures of inhibited MMP-1. The major difference observed between the NMR solution structure of inhibitor-free MMP-1 and the X-ray structures of inhibited MMP-1 is the dynamics of the active site. The 2D 15N-1H HSQC spectra, the lack of information in the 15N-edited NOESY spectra, and the generalized order parameters (S2) determined from 15N T1, T2 and NOE data suggest a slow conformational exchange for residues comprising the active site (helix B, zinc ligated histidines and the nearby loop region) and a high mobility for residues Pro138-Gly144 in the vicinity of the active site for inhibitor-free collagenase. In contrast to the X-ray structures, only the slow conformational exchange is lost in the presence of an inhibitor.

Properly oriented heparin-decasaccharide-induced dimers are the biologically active form of basic fibroblast growth factor.

Interaction of basic fibroblast growth factor (FGF-2) with heparin or heparan sulfate proteoglycans (HSPGs) is required for receptor activation and initiation of biological responses. To gain insight into the mechanism of activation of the FGF receptor by FGF-2 and heparin, we have used NMR, dynamic light scattering, and HSPG-deficient cells and cell-free systems. The first 28 N-terminal residues in FGF-2 were found to be highly mobile and flexible, consistent with the disorder found in both the NMR and X-ray structures. The structure of an FGF-2-heparin-decasaccharide complex that binds to and activates the FGF receptor was compared to a heparin-tetrasaccharide-induced complex that does not promote an interaction with the receptor. The major change observed upon the addition of the tetrasaccharide to FGF-2 was an increase in the correlation time consistent with the formation of an FGF-2 dimer. The NMR line widths of FGF-2 in the presence of the decasaccharide are severely broadened relative to the tetrasaccharide, consistent with dynamic light scattering results which indicate FGF-2 is a tetramer. The interaction of these heparin species with FGF-2 does not induce a significant conformational change in the overall structure of FGF-2, but small chemical shift changes are observed in both heparin and receptor binding sites. A trans-oriented symmetric dimer of FGF-2 is formed in the presence of the tetrasaccharide whereas two cis-oriented dimers in a symmetric tetramer are formed in the presence of the decasaccharide. This suggests that the cis-oriented FGF-2 dimer is the minimal biologically active structural unit of FGF-2. These data allow us to propose a novel mechanism to explain the functional interaction of FGF-2 with heparin and its transmembrane receptor.

High-resolution solution structure of basic fibroblast growth factor determined by multidimensional heteronuclear magnetic resonance spectroscopy.

The high-resolution solution structure of recombinant human basic fibroblast growth factor (FGF-2), a protein of 17.2 kDa that exhibits a variety of functions related to cell growth and differentiation, has been determined using three-dimensional heteronuclear NMR spectroscopy. A total of 30 structures were calculated by means of hybrid distance geometry--simulated annealing using a total of 2865 experimental NMR restraints, consisting of 2486 approximate inteproton distance restraints, 50 distance restraints for 25 backbone hydrogen bonds, and 329 torsion angle restraints. The atomic rms distribution about the mean coordinate positions for the 30 structures for residues 29-152 is 0.43 +/- 0.03 A for the backbone atoms, 0.83 +/- 0.05 A for all atoms, and 0.51 +/- 0.04 A for all atoms excluding disordered side chains. The overall structure of FGF-2 consists of 11 extended antiparallel beta-strands arranged in three groups of three or four strands connected by tight turns and loop regions creating a pseudo-3-fold symmetry. Two strands from each group come together to form a beta-sheet barrel of six antiparallel beta-strands. A helix-like structure was observed for residues 131-136, which is part of the heparin binding site (residues 128-138). The discovery of the helix-like region in the primary heparin binding site instead of the beta-strand conformation described in the X-ray structures may have important implications in understanding the nature of heparin--FGF-2 interactions. A total of seven tightly bound water molecules were found in the FGF-2 structure, two of which are located in the heparin binding site. The first 28 N-terminal residues appear to be disordered, which is consistent with previous X-ray structures. A best fit superposition of the NMR structure of FGF-2 with the 1.9 A resolution X-ray structure by Zhu et al. (1991) yields a backbone atomic rms difference of 0.94 A, indicative of a close similarity between the NMR and X-ray structures.

Resonance assignments for Oncostatin M, a 24-kDa alpha-helical protein.

Oncostatin M (OM) is a cytokine that shares a structural and functional relationship with interleukin-6, leukemia inhibitory factor, and granulocyte-colony stimulating factor, which regulate the proliferation and differentiation of a variety of cell types. A mutant version of human OM in which two N-linked glycosylation sites and an unpaired cysteine have been mutated to alanine (N76A/C81A/N193A) has been expressed and shown to be active. The triple mutant has been doubly isotope-labeled with 13C and 15N in order to utilize heteronuclear multidimensional NMR techniques for structure determination. Approximately 90% of the backbone resonances were assigned from a combination of triple-resonance data (HNCA, HNCO, CBCACONH, HBHACONH, HNHA and HCACO), intraresidue and sequential NOEs (3D 15N-NOESY-HMQC and 13C-HSQC-NOESY) and side-chain information obtained from the CCONH and HCCONH experiments. Preliminary analysis of the NOE pattern in the 15N-NOESY-HMQC spectrum and the 13C alpha secondary chemical shifts predicts a secondary structure for OM consisting of four alpha-helices with three intervening helical regions, consistent with the four-helix-bundle motif found for this cytokine family. As a 203-residue protein with a molecular weight of 24 kDa, Oncostatin M is the largest alpha-helical protein yet assigned.

1H, 15N, 13C and 13CO assignments and secondary structure determination of basic fibroblast growth factor using 3D heteronuclear NMR spectroscopy.

The assignments of the 1H, 15N, 13CO and 13C resonances of recombinant human basic fibroblast growth factor (FGF-2), a protein comprising of 154 residues and with a molecular mass of 17.2 kDa, is presented based on a series of three-dimensional triple-resonance heteronuclear NMR experiments. These studies employ uniformly labeled 15N- and 15N-/13C-labeled FGF-2 with an isotope incorporation > 95% for the protein expressed in E. coli. The sequence-specific backbone assignments were based primarily on the interresidue correlation of C alpha, C beta and H alpha to the backbone amide 1H and 15N of the next residue in the CBCA(CO)NH and HBHA(CO)NH experiments and the intraresidue correlation of C alpha, C beta and H alpha to the backbone amide 1H and 15N in the CBCANH and HNHA experiments. In addition, C alpha and C beta chemical shift assignments were used to determine amino acid types. Sequential assignments were verified from carbonyl correlations observed in the HNCO and HCACO experiments and C alpha correlations from the HNCA experiment. Aliphatic side-chain spin systems were assigned primarily from H(CCO)NH and C(CO)NH experiments that correlate all the aliphatic 1H and 13C resonances of a given residue with the amide resonance of the next residue. Additional side-chain assignments were made from HCCH-COSY and HCCH-TOCSY experiments. The secondary structure of FGF-2 is based on NOE data involving the NH, H alpha and H beta protons as well as 3JHNH alpha coupling constants, amide exchange and 13C alpha and 13C beta secondary chemical shifts. It is shown that FGF-2 consists of 11 well-defined antiparallel beta-sheets (residues 30-34, 39-44, 48-53, 62-67, 71-76, 81-85, 91-94, 103-108, 113-118, 123-125 and 148-152) and a helix-like structure (residues 131-136), which are connected primarily by tight turns. This structure differs from the refined X-ray crystal structures of FGF-2, where residues 131-136 were defined as beta-strand XI. The discovery of the helix-like region in the primary heparin-binding site (residues 128-138) instead of the beta-strand conformation described in the X-ray structures may have important implications in understanding the nature of heparin-FGF-2 interactions. In addition, two distinct conformations exist in solution for the N-terminal residues 9-28. This is consistent with the X-ray structures of FGF-2, where the first 17-19 residues were ill defined.

Signal transduction in chemotaxis. A propagating conformation change upon phosphorylation of CheY.

The CheY protein from Escherichia coli and Salmonella typhimurium are among the best characterized proteins of the receiver domain family of two component signal transduction systems in bacteria. Phosphorylation of CheY plays a central role in bacterial chemotaxis. However, it is not entirely clear how its state of phosphorylation contributes to its function. Genetic evidence suggests that CheY changes its conformation upon phosphorylation. We present evidence for this conformation change by comparing the NMR 15N-1H correlation spectra of CheY.Mg2+ complex and phospho-CheY in the presence of magnesium. Large changes in chemical shift are used as indicators of chemical changes and probable structural changes in the protein backbone. Our observations suggest that significant structural changes occur in CheY upon phosphorylation and that these changes are distinct from the changes produced by magnesium ion binding. In addition to residues Asn-59 and Gly-65 that are immediately adjacent to the site of phosphorylation at Asp-57, a large number of other residues show significant chemical shift changes as a result of phosphorylation. These include Met-17, Val-21, Asn-23, Gly-39, Met-60, Met-63, Asp-64, Leu-66, Glu-67, Leu-68, Leu-69, Met-85, Val-86, Thr-87, Ala-88, Asn-94, Val-107, Lys-109, Thr-112, Ala-113, Ala-114, and Asn-121. These results appear inconsistent with the recent suggestion that phosphorylation produces the same structural changes as magnesium binding (Bellsolell, L., Prieto, J., Serrano, L., and Coll, M. (1994) J. Mol. Biol. 238, 489-495). We find that some regions change overlap with a genetically defined motor binding face. We therefore propose that the conformation switch modulates the interaction of CheY with its target, the flagellar motor. Other regions also change, possibly reflecting the many different functions of CheY homologues.