Structure of alpha-lytic protease complexed with its pro region

While the majority of proteins fold rapidly and spontaneously to their native states, the extracellular bacterial protease alpha-lytic protease (alphaLP) has a t(1/2) for folding of approximately 2,000 years, corresponding to a folding barrier of 30 kcal mol(-1). AlphaLP is synthesized as a pro-enzyme where its pro region (Pro) acts as a foldase to stabilize the transition state for the folding reaction. Pro also functions as a potent folding catalyst when supplied as a separate polypeptide chain, accelerating the rate of alphaLP folding by a factor of 3 x 10(9). In the absence of Pro, alphaLP folds only partially to a stable molten globule-like intermediate state. Addition of Pro to this intermediate leads to rapid formation of native alphaLP. Here we report the crystal structures of Pro and of the non-covalent inhibitory complex between Pro and native alphaLP. The C-shaped Pro surrounds the C-terminal beta-barrel domain of the folded protease, forming a large complementary interface. Regions of extensive hydration in the interface explain how Pro binds tightly to the native state, yet even more tightly to the folding transition state. Based on structural and functional data we propose that a specific structural element in alphaLP is largely responsible for the folding barrier and suggest how Pro can overcome this barrier.     

X-ray crystal structure of the yeast Kar3 motor domain complexed with Mg.ADP to 2.3 A resolution

The kinesin family of motor proteins, which contain a conserved motor domain of approximately 350 amino acids, generate movement against microtubules. Over 90 members of this family have been identified, including motors that move toward the minus or plus end of microtubules. The Kar3 protein from Saccharomyces cerevisiae is a minus end-directed kinesin family member that is involved in both nuclear fusion, or karyogamy, and mitosis. The Kar3 protein is 729 residues in length with the motor domain located in the C-terminal 347 residues. Recently, the three-dimensional structures of two kinesin family members have been reported. These structures include the motor domains of the plus end-directed kinesin heavy chain [Kull, F. J., et al. (1996) Nature 380, 550-555] and the minus end-directed Ncd [Sablin, E. P., et al. (1996) Nature 380, 555-559]. We now report the structure of the Kar3 protein complexed with Mg.ADP obtained from crystallographic data to 2.3 A. The structure is similar to those of the earlier kinesin family members, but shows differences as well, most notably in the length of helix alpha 4, a helix which is believed to be involved in conformational changes during the hydrolysis cycle.     

The yeast mic2 mutant is defective in the formation of mannosyl-diinositolphosphorylceramide

Three transmembrane glutamic acid residues play essential roles in the metal-tetracycline/H+ antiporter Tet(K) of Staphylococcus aureus [Fujihira et al., FEBS Lett. 391 (1996) 243-246]. In the putative hydrophilic loop region of the Tet(K) and Tet(L) proteins, six acidic residues are conserved. Asp74, Asp200, Asp318 and Glu381 are located on the putative cytoplasmic side, and Asp39 and Glu345 on the putative periplasmic side. These residues were replaced by a neutral amino acid residue or a charge-conserved one. In contrast to the transmembrane glutamic acid residues, the replacement of the two glutamic acid residues (Glu345 and Glu381) did not affect the tetracycline resistance level. Out of the other four aspartic acid residues, the only essential residue is Asp318, any replacement of which resulted in complete loss of the tetracycline resistance and transport activity. Asp318 is located in cytoplasmic loop 10-11 in the putative 14-transmembrane-segment topology of Tet(K). In the case of the tetracycline exporters of Gram-negative bacteria, the only essential acidic residue in the cytoplasmic loop region is located in loop 2-3 [Yamaguchi et al., Biochemistry 31 (1992) 8344-8348]. It may be a general role for tetracycline efflux proteins that three transmembrane and one cytoplasmic acidic residues are mandatory for the tetracycline transport function.     

Solution structure of human neuropeptide Y

The three-dimensional structure of synthetic human neuropeptide Y in aqueous solution at pH 3.2 and 37 degrees C was determined from two-dimensional 1H NMR data recorded at 600 MHz. A restraint set consisting of 440 interproton distance restraints inferred from NOEs and 11 backbone and 4 side-chain dihedral angle restraints derived from spin-spin coupling constants was used as input for distance geometry calculations on DIANA and simulated annealing and restrained energy minimization in X-PLOR. The final set of 26 structures is well defined in the region of residues 11-36, with a mean pairwise rmsd of 0.51 A for the backbone heavy atoms (N, C alpha and C) and 1.34 A for all heavy atoms. Residues 13-36 form an amphipathic alpha-helix. The N-terminal 10 residues are poorly defined relative to the helical region, although some elements of local structure are apparent. At least one of the three prolines in the N-terminal region co-exists in both cis and trans conformations. An additional set of 24 distances was interpreted as intermolecular distances within a dimer. A combination of distance geometry and restrained simulated annealing yielded a model of the dimer having antiparallel packing of two helical units, whose hydrophobic faces form a well-defined core. Sedimentation equilibrium experiments confirm the observation that neuropeptide Y associates to form dimers and higher aggregates under the conditions of the NMR experiments. Our results therefore support the structural features reported for porcine neuropeptide Y [Cowley, D.J. et al. (1992) Eur. J. Biochem., 205, 1099-1106] rather than the 'aPP' fold described previously for human neuropeptide Y [Darbon, H. et al. (1992) Eur. J. Biochem., 209, 765-771].     

The immediate 5'-flanking region of the rat phenylalanine hydroxylase-encoding gene

We have characterized the immediate (465 bp) 5'-flanking region of the rat phenylalanine hydroxylase (PAH)-encoding gene. This sequence shows considerable similarity to the 5'-flanking region of the human PAH gene [Konecki et al., Biochemistry 31 (1992) 8363-8368]. Both sequences lack obvious TATA elements; however, putative regulatory sites, including a potential cyclic AMP-response element and glucocorticoid response elements, are present.     

Three-dimensional structure of meso-diaminopimelic acid dehydrogenase from Corynebacterium glutamicum

Diaminopimelate dehydrogenase catalyzes the NADPH-dependent reduction of ammonia and L-2-amino-6-ketopimelate to form meso-diaminopimelate, the direct precursor of L-lysine in the bacterial lysine biosynthetic pathway. Since mammals lack this metabolic pathway inhibitors of enzymes in this pathway may be useful as antibiotics or herbicides. Diaminopimelate dehydrogenase catalyzes the only oxidative deamination of an amino acid of D configuration and must additionally distinguish between two chiral amino acid centers on the same symmetric substrate. The Corynebacterium glutamicum enzyme has been cloned, expressed in Escherichia coli, and purified to homogeneity using standard biochemical procedures [Reddy, S. G., Scapin, G., & Blanchard, J. S. (1996) Proteins: Structure, Funct. Genet. 25, 514-516]. The three-dimensional structure of the binary complex of diaminopimelate dehydrogenase with NADP+ has been solved using multiple isomorphous replacement procedures and noncrystallographic symmetry averaging. The resulting model has been refined against 2.2 A diffraction data to a conventional crystallographic R-factor of 17.0%. Diaminopimelate dehydrogenase is a homodimer of structurally not identical subunits. Each subunit is composed of three domains. The N-terminal domain contains a modified dinucleotide binding domain, or Rossman fold (six central beta-strands in a 213456 topology surrounded by five alpha-helices). The second domain contains two alpha-helices and three beta-strands. This domain is referred to as the dimerization domain, since it is involved in forming the monomer--monomer interface of the dimer. The third or C-terminal domain is composed of six beta-strands and five alpha-helices. The relative position of the N- and C-terminal domain in the two monomers is different, defining an open and a closed conformation that may represent the enzyme's binding and active state, respectively. In both monomers the nucleotide is bound in an extended conformation across the C-terminal portion of the beta-sheet of the Rossman fold, with its C4 facing the C-terminal domain. In the closed conformer two molecules of acetate have been refined in this region, and we postulate that they define the DAP binding site. The structure of diaminopimelate dehydrogenase shows interesting similarities to the structure of glutamate dehydrogenase [Baker, P. J., Britton, K. L., Rice, D. W., Rob, A., & Stillmann, T.J. (1992a) J. Mol. Biol. 228, 662-671] and leucine dehydrogenase [Baker, P.J., Turnbull, A.P., Sedelnikova, S.E., Stillman, T. J., & Rice, D. W. (1995) Structure 3, 693-705] and also resembles the structure of dihydrodipicolinate reductase [Scapin, G., Blanchard, J. S., & Sacchettini, J. C. (1995) Biochemistry 34, 3502-3512], the enzyme immediately preceding it in the diaminopimelic acid/lysine biosynthetic pathway.     

A transient precursor of the HIV-1 protease. Isolation, characterization, and kinetics of maturation

Recently, the mechanism of autoprocessing of the protease (PR) of the human immunodeficiency virus type 1 from the model polyprotein, MBP-DeltaTF-PR-DeltaPol, which contains the protease linked to short native flanking sequences (DeltaTF and DeltaPol) fused to the maltose binding protein (MBP) of Escherichia coli, was reported (Louis, J. M., Nashed, N. T., Parris, K. D., Kimmel, A. R., and Jerina, D. M. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 7970-7974). According to this mechanism, intramolecular cleavage of the N-terminal strands of the dimeric MBP-DeltaTF-PR-DeltaPol protein leads to the formation of the PR-DeltaPol intermediate, which is subsequently converted to the mature protease by cleavage of the C-terminal strands. We now report the purification and characterization of the PR-DeltaPol intermediate and the kinetics of its processing to the mature protease. Unlike the MBP-DeltaTF-PR-DeltaPol precursor, PR-DeltaPol has proteolytic activity similar to that of the mature enzyme at pH 5.0. The pH rate profile for kcat/Km is similar to that of the mature protease above pH 4.0. Although the PR-DeltaPol is more sensitive than the mature protease toward denaturing reagents, both the enzymatic activity and the intrinsic fluorescence of PR-DeltaPol are linearly dependent on the protein concentration, indicating that the protein is largely in its dimeric form above 10 nM. In contrast to the first-order kinetics observed for the proteolytic reaction at the N terminus of the protease, the proteolytic reaction at the C terminus of the protease is second order in protein concentration. These results are discussed in terms of a mechanism in which the C-terminally located DeltaPol peptide chains are cleaved intermolecularly to release the mature protease.     

Dictyostelium myosin 25-50K loop substitutions specifically affect ADP release rates

While most of the sequence of myosin's motor domain is highly conserved among various organisms and tissue types, the junctions between the 25 and 50 kDa domains and the 50 and 20 kDa domains are strikingly divergent. The 50-20K loop is positioned to interact with actin, while the 25-50K loop is situated nearer the ATP binding site [Rayment, I., et al. (1993) Science 261, 50-58]. Chimeric studies of the 50-20K loop [Uyeda, T. Q.-P., et al. (1994) Nature 368, 567-569; Rovner, A. S., et al. (1995) J. Biol. Chem. 270 (51), 30260-30263] have shown that this loop affects actin activation of ATPase activity. Given the function of myosin as a molecular motor, it was proposed that the 25-50K loop might specifically alter ADP release [Spudich, J. A. (1994) Nature 374, 515-518]. Here we study the role of this loop by engineering chimeras containing the Dictyostelium myosin heavy chain with loops from two enzymatically diverse myosins, rabbit skeletal and Acanthamoeba. The chimeric myosins complement the myosin null phenotype in vivo, bind nucleotide normally, interact normally with actin, and display wild-type levels of actin-activated ATPase activity. However, the rate of ADP release from the myosins, normally the slowest step involved in motility, was changed in a manner that reflects the activity of the donor myosin. In summary, studies of Dictyostelium myosin heavy chain chimeras have shown that the 50-20K sequence specifically affects the actin-activated ATPase activity [Uyeda, T. Q.-P., et al. (1994)] while the 25-50K sequence helps determine the rate of ADP release.     

Structural effects of the binding of GTP to the wild-type and oncogenic forms of the ras-gene-encoded p21 proteins

Molecular dynamics calculations have been performed to determine the average structures of ras-gene-encoded p21 proteins bound to GTP, i.e., the normal (wild-type) protein and two oncogenic forms of this protein, the Val 12- and Leu 61-p21 proteins. We find that the average structures for all of these proteins exhibit low coordinate fluctuations (which are highest for the normal protein), indicating convergence to specific structures. From previous dynamics calculations of the average structures of these proteins bound to GDP, major regional differences were found among these proteins [Monaco et al. (1995), J. Protein Chem., in press]. We now find that the average structures of the oncogenic proteins are more similar to one another when the proteins are bound to GTP than when they are bound to GDP [Monaco et al. (1995), J. Protein Chem., in press]. However, they still differ in structure at specific amino acid residues rather than in whole regions, in contradistinction to the results found for the p21-GDP complexes. Two exceptions are the regions 25-32, in an alpha-helical region, and 97-110. The two oncogenic (Val 12- and Leu 61-) proteins have similar structures which differ significantly in the region of residues 97-110. This region has recently been identified as being critical in the interaction of p21 with kinase target proteins. The differences in structure between the oncogenic proteins suggest the existence of more than one oncogenic form of the p21 protein that can activate different signaling pathways.     

Identification of the catalytic triad residues of porcine liver acylamino acid-releasing enzyme

Acylamino acid-releasing enzyme (AARE) [EC 3.4.19.1] is a tetrameric serine protease, which belongs to the oligopeptidase family and specifically removes acetyl amino acids from N-terminally acetylated peptides. By using diisopropyl fluorophosphate, we previously identified one of the residues comprising the catalytic triad of this enzyme as Ser587 [Miyagi, M. et al. (1995) J. Biochem. 118, 771-779]. To elucidate the other two residues forming the catalytic triad of this new serine-type protease, wild-type and four mutant AAREs, in which each candidate residue of the catalytic triad deduced from sequence alignment with other oligopeptidases was substituted by site-directed mutagenesis, were expressed in Escherichia coli as fusion proteins with short peptide chains at both N- and C-termini of a subunit of porcine liver enzyme. All of the recombinant AAREs were estimated to have similar conformational and quaternary structures to the native porcine liver enzyme from their CD spectra and behavior on gel-filtration, but the mutants in which Ala587, Asn675, or Tyr707 was substituted for Ser587, Asp675, or His707, respectively, did not show detectable hydrolytic activity toward acetyl-L-methionyl L-alanine. These facts suggest that Ser587, Asp675, and His707 are essential residues for the AARE activity and comprise the catalytic triad of the enzyme in this order. Thus, AARE has been shown to have a protease-like domain in its C-terminal region, as do other proteins classified as members of the oligopeptidase family.     

Construction, expression, and characterization of a novel fully activated recombinant single-chain hepatitis C virus protease

Efficient proteolytic processing of essential junctions of the hepatitis C virus (HCV) polyprotein requires a heterodimeric complex of the NS3 bifunctional protease/helicase and the NS4A accessory protein. A single-chain recombinant form of the protease has been constructed in which NS4A residues 21-32 (GSVVIVGRIILS) were fused in frame to the amino terminus of the NS3 protease domain (residues 3-181) through a tetrapeptide linker. The single-chain recombinant protease has been overexpressed as a soluble protein in E. coli and purified to homogeneity by a combination of metal chelate and size-exclusion chromatography. The single-chain recombinant protease domain shows full proteolytic activity cleaving the NS5A-5B synthetic peptide substrate, DTEDVVCCSMSYTWTGK with a Km and k(cat) of 20.0 +/- 2.0 microM and 9.6 +/- 2.0 min(-1), respectively; parameters identical to those of the authentic NS3(1-631)/NS4A(1-54) protein complex generated in eukaryotic cells (Sali DL et al., 1998, Biochemistry 37:3392-3401).     

Modeling the carbohydrate-binding specificity of pig edema toxin

The wild-type binding pentamer of Shiga-like toxin IIe (SLT-IIe) binds both the globotriaosylceramide (Gb3) and globotetraosylceramide (Gb4) cell surface glycolipids, whereas the double mutant GT3 (Q65E/K67Q) exhibits a marked preference for Gb3 [Tyrrell, G. J., et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 524-528]. We modeled three unique sites (sites 1-3) for binding of the carbohydrate moiety of Gb3 to GT3 and SLT-IIe, on the basis of the three sites observed for the SLT-I pentamer [Ling, H., et al. (1998) Biochemistry 37, 1777-1788]. Examination of the three sites in light of various mutation and binding data strongly suggested that one of the binding sites plays a role in the change of specificity observed for the GT3 mutant. We applied several modeling techniques, and developed a model for binding of the carbohydrate moiety of Gb4 to this site of the SLT-IIe binding pentamer. This model is consistent with a wide variety of mutation and binding data and clearly shows the importance of the terminal GalNAc residue of Gb4, as well as that of the two mutated residues of GT3, to the intermolecular interaction.     

Human immunodeficiency virus protease ligand specificity conferred by residues outside of the active site cavity

To gain greater understanding of the structural basis of human immunodeficiency virus (HIV) protease ligand specificity, we have crystallized and determined the structures of the HIV-1 protease (Val32Ile, Ile47Val, Val82Ile) triple mutant and simian immunodeficiency virus (SIV) protease in complex with SB203386, a tripeptide analogue inhibitor containing a C-terminal imidazole substituent as an amide bond isostere. SB203386 is a potent inhibitor of HIV-1 protease (Ki = 18 nM) but shows decreased inhibition of the HIV-1 protease (Val32Ile, Ile47Val, Val82Ile) triple mutant (Ki = 112 nM) and SIV protease (Ki = 960 nM). Although SB203386 binds in the active site cavity of the triple mutant in a similar fashion to its binding to the wild-type HIV-1 protease [Abdel-Meguid et al. (1994) Biochemistry 33, 11671], it binds to SIV protease in an unexpected mode showing two inhibitor molecules each binding to half of the active site. Comparison of these two structures and that of the wild-type HIV-1 protease bound to SB203386 reveals that HIV protease ligand specificity is imparted by residues outside of the catalytic pocket, which causes subtle changes in its shape. Furthermore, this work illustrates the importance of structural studies in order to understand the structure-activity relationship (SAR) between related enzymes.     

Potent peptide inhibitors of human hepatitis C virus NS3 protease are obtained by optimizing the cleavage products

In the absence of a broadly effective cure for hepatitis caused by hepatitis C virus (HCV), much effort is currently devoted to the search for inhibitors of the virally encoded protease NS3. This chymotrypsin-like serine protease is required for the maturation of the viral polyprotein, cleaving it at the NS3-NS4A, NS4A-NS4B, NS4B-NS5A, and NS5A-NS5B sites. In the course of our studies on the substrate specificity of NS3, we found that the products of cleavage corresponding to the P6-P1 region of the substrates act as competitive inhibitors of the enzyme, with IC50s ranging from 360 to 1 microM. A detailed study of product inhibition by the natural NS3 substrates is described in the preceding paper [Steinkühler, C., et al. (1997) Biochemistry 37, 8899-8905]. Here we report the results of a study of the structure-activity relationship of the NS3 product inhibitors, which suggest that the mode of binding of the P region-derived products is similar to the ground-state binding of the corresponding substrates, with additional binding energy provided by the C-terminal carboxylate. Optimal binding requires a dual anchor: an "acid anchor" at the N terminus and a "P1 anchor" at the C-terminal part of the molecule. We have then optimized the sequence of the product inhibitors by using single mutations and combinatorial peptide libraries based on the most potent natural product, Ac-Asp-Glu-Met-Glu-Glu-Cys-OH (Ki = 0.6 microM), derived from cleavage at the NS4A-NS4B junction. By sequentially optimizing positions P2, P4, P3, and P5, we obtained several nanomolar inhibitors of the enzyme. These compounds are useful both as a starting point for the development of peptidomimetic drugs and as structural probes for investigating the substrate binding site of NS3 by modeling, NMR, and crystallography.     

Essential lysine residues in the N-terminal and the C-terminal domain of human adenylate kinase interact with adenine nucleotides as found by site-directed random mutagenesis

To elucidate the minimum requirement of amino acid residues for the active center in human adenylate kinase (hAK1), we carried out random site-directed mutagenesis of key lysine residues (K9, K21, K27, K31, K63, K131, and K194), which were conserved in mammalian AK1 species, with the pMEX8-hAK1 plasmid [Ayabe, T., et al. (1996) Biochem. Mol. Biol. Int. 38, 373-381]. Twenty different mutants were obtained and analyzed by steady-state kinetics, and all mutants showed activity loss by Km and/or k(cat) effects on MgATP2-, AMP2-, or both. The results have led to the following conclusions. (1) Lys9 would appear to interact with both MgATP2- and AMP2- but to a larger extent than with AMP2-. (2) Lys21 is likely to play a role in substrate binding of both MgATP2- and AMP2- but more strongly affects MgATP2-. (3) Lys27 and Lys131 would appear to play a functional role in catalysis by interacting strongly with MgATP2-. (4) Lys31 would appear to interact with MgATP2- and AMP2- at the MgATP2- site. (5) Lys63 would be more likely to interact with MgATP2- than with AMP2-. (6) Lys194 in the flanking C-terminal domain would appear to interact not only with MgATP2- but also with AMP2- at the MgATP2- site by stabilizing substrate binding. The loss of the positively charged epsilon-amino group of lysine affects both the affinity for the substrate and the catalytic efficiency. Hence, hydrophilic lysine residues in hAK1 would appear to be essential for substrate-enzyme interaction with the coordination of some arginine residues, reported previously [Kim, H. J., et al. (1990) Biochemistry 29, 1107-1111].     

The prosequence of thermolysin acts as an intramolecular chaperone when expressed in trans with the mature sequence in Escherichia coli

The zinc metalloendopeptidase, thermolysin (EC 3.4.24.27) produced by Bacillus thermoproteolyticus serves as a model of important physiological enzymes such as neprilysin, angiotensin converting enzyme and endothelin converting enzyme. Thermolysin is synthesised as a pre-proenzyme, with an N-terminal prosequence of 204 residues and a mature sequence of 316 residues. The prosequence facilitates the folding of the denatured mature sequence in vitro and the cleavage of the peptide bond linking the pro and mature sequences occurs by an autocatalytic, intramolecular process. With the aim to study the role of the prosequence in vivo and to produce active mutants for structural studies, the mature sequence of thermolysin has now been expressed in Escherichia coli, either alone or with the prosequence as an independent polypeptide, i.e. in trans form. In addition, the mature sequence of an inactive mutant in which Glu143 involved in the catalytic process was replaced by Ala has also been expressed in trans with the prosequence. The results show that the pro-sequence is required to obtain active thermolysin and that a covalent link with the mature sequence is not necessary for the correct folding of the protease in vivo. Moreover, when expressed in E. coli (in trans with the prosequence), the yield of correctly folded E143A mutant was similar to that of the wild-type protease, whereas no mature enzyme was detected when it was expressed as a pre-proenzyme in Bacillus subtilis. These results demonstrate that the thermolysin prosequence acts as an intramolecular chaperone in vivo and open the way to structural studies of catalytic site mutants produced in large quantities in E. coli.     

RS-17053 (N-[2-(2-cyclopropylmethoxyphenoxy)ethyl]-5-chloro-alpha, alpha-dimethyl-1H-indole-3-ethanamine hydrochloride), a selective alpha 1A-adrenoceptor antagonist, displays low affinity for functional alpha 1-adrenoceptors in human prostate: implications for adrenoceptor classification

Hereditary developmental abnormalities of the upper or lower limbs in humans are easily recognizable phenotypes that can be used in the mapping and cloning of genes involved in normal human development. We studied a large Indian pedigree (UR002) with an autosomal dominant triphalangeal thumb (TPT) and polysyndactyly (PSD). The abnormalities were present only in the upper limbs, and the phenotype was fully penetrant. The expression of the phenotype was variable and ranged from unilateral TPT to bilateral TPT, preaxial du-, tri-, or quadruplication of the thumb, or syndactyly of multiple thumbs. There were 112 affected individuals in the pedigree. Previous linkage analyses on apparently similar phenotypes have identified a locus at 7q36 [Heutink et al., 1994, Nature Genet 6:287-291; Tsukurov et al., 1994]. To map the gene responsible for the TPT-PSD in family UR002, we performed linkage analysis in DNA from 47 affected and 7 normal individuals. Marker D7S550, located at 7q36, yielded a maximum LOD score of 11.31 at theta = 0.00. Additional markers in the region also showed no recombination. These data indicate that the gene responsible for the hand abnormality in pedigree UR002 maps to the same region as that in previous pedigrees with similar phenotype. Further analyses of recombinants among all the linked families by using new polymorphic markers will narrow the critical genomic region and facilitate positional cloning of the elusive gene.     

Residue lysine-34 in GroES modulates allosteric transitions in GroEL

The conserved residue Lys-34 in GroES was replaced by alanine and glutamic acid using site-directed mutagenesis. This residue is near the carboxy terminus of the mobile loop in GroES (residues 17-32) which becomes immobilized upon formation of the GroEL/GroES complex [Landry et al. (1993) Nature 364, 255-258]. Both charge neutralization (Lys-34-->Ala) and charge reversal (Lys-34-->Glu) at this position have little effect on the binding constant of GroES to GroEL, but they increase the enhancement by GroES of cooperativity in ATP hydrolysis by GroEL. This is reflected by a change in the Hill coefficient (at 10 mM K+) from 4.10 (+/- 0.22) in the presence of wild-type GroES to 5.17 (+/- 0.24) and 4.46 (+/- 0.14) in the presence of the GroES mutants Lys-34-->Ala and Lys-34-->Glu, respectively. The results are interpreted using the Monod-Wyman-Changeux (MWC) model for cooperativity [Monod et al. (1965) J. Mol. Biol. 12, 88-118]. They suggest that Lys-34 in GroES modulates the allosteric transition in GroEL by stabilizing a relaxed (R)-like state.