Substrate-assisted catalysis: molecular basis and biological significance.

Substrate-assisted catalysis (SAC) is the process by which a functional group in a substrate contributes to catalysis by an enzyme. SAC has been demonstrated for representatives of three major enzyme classes: serine proteases, GTPases, and type II restriction endonucleases, as well as lysozyme and hexose-1-phosphate uridylyltransferase. Moreover, structure-based predictions of SAC have been made for many additional enzymes. Examples of SAC include both naturally occurring enzymes such as type II restriction endonucleases as well as engineered enzymes including serine proteases. In the latter case, a functional group from a substrate can substitute for a catalytic residue replaced by site-directed mutagenesis. From a protein engineering perspective, SAC provides a strategy for drastically changing enzyme substrate specificity or even the reaction catalyzed. From a biological viewpoint, SAC contributes significantly to the activity of some enzymes and may represent a functional intermediate in the evolution of catalysis. This review focuses on advances in engineering enzyme specificity and activity by SAC, together with the biological significance of this phenomenon.

Elastase substrate specificity tailored through substrate-assisted catalysis and phage display.

The catalytic histidine of human neutrophil elastase was replaced with alanine (H57A) to determine if a substrate histidine could substitute for the missing catalytic group-'substrate-assisted catalysis'. H57A and wild-type elastase were recovered directly from Pichia pastoris following expression from a synthetic gene lacking the elastase pro sequence, thereby obviating the need for zymogen activation. Potential histidine-containing substrates for H57A elastase were identified from a phage library of randomized sequences. One such sequence, REHVVY, was cleaved by H57A elastase with a catalytic efficiency, k(cat)/K(M), of 2800 s(-1) M(-1), that is within 160-fold of wild-type elastase. In contrast, wild-type but not H57A elastase cleaved the related non-histidine containing sequence, REAVVY. Ten different histidine-containing linkers were cleaved by H57A elastase. In addition to the requirement for a P2 histidine, significant preferences were observed at other subsites including valine or threonine at P1, and methionine or arginine at P4. A designed sequence, MEHVVY, containing the preferred residues identified at each subsite proved to be a more favorable substrate than any of the phage-derived sequences. Extension of substrate-assisted catalysis to elastase suggests that this engineering strategy may be widely applicable to other serine proteases thereby creating a family of highly specific histidine-dependant proteases.

Antibody engineering.

Among the most important advances in antibody engineering of this past year is the advent of new tools to study the relationship between protein (including antibody) structure and function. Very rapid large-scale mutational analysis of antibodies is now possible by using in vitro transcription and translation. Ribosome display is a rapidly evolving technology for modifying antibody function that offers several potential advantages over phage display.

Contribution of domain interface residues to the stability of antibody CH3 domain homodimers.

Dimers of CH3 domains from human IgG1 were used to study the effect of mutations constructed at a domain-domain interface upon domain dissociation and unfolding, "complex stability". Alanine replacement mutants were constructed on one side of the interface for each of the sixteen interdomain contact residues by using a single-chain CH3 dimer in which the carboxyl terminus of one domain was joined to the amino terminus of the second domain via a (G4S)4 linker. Single-chain variants were expressed in Escherichia coli grown in a fermentor and recovered in yields of 6-90 mg L-1 by immobilized metal affinity chromatography. Guanidine hydrochloride-induced denaturation was used to follow domain dissociation and unfolding. Surprisingly, the linker did not perturb the complex stability for either the wild type or two destabilizing mutants. The CH3 domain dissociation and unfolding energetics are dominated by six contact residues where corresponding alanine mutations each destabilize the complex by >2.0 kcal mol-1. Five of these residues (T366, L368, F405, Y407, and K409) form a patch at the center of the interface and are located on the two internal antiparallel beta-strands. These energetically key residues are surrounded by 10 residues on the two external beta-strands whose contribution to complex stability is small (three have a Delta DeltaG of 1.1-1.3 kcal mol-1) or very small (seven have a Delta DeltaG of

A mutational analysis of binding interactions in an antigen-antibody protein-protein complex.

Alanine scanning mutagenesis, double mutant cycles, and X-ray crystallography were used to characterize the interface between the anti-hen egg white lysozyme (HEL) antibody D1.3 and HEL. Twelve out of the 13 nonglycine contact residues on HEL, as determined by the high-resolution crystal structure of the D1.3-HEL complex, were individually truncated to alanine. Only four positions showed a DeltaDeltaG (DeltaGmutant - DeltaGwild-type) of greater than 1.0 kcal/mol, with HEL residue Gln121 proving the most critical for binding (DeltaDeltaG = 2.9 kcal/mol). These residues form a contiguous patch at the periphery of the epitope recognized by D1.3. To understand how potentially disruptive mutations in the antigen are accommodated in the D1.3-HEL interface, we determined the crystal structure to 1.5 A resolution of the complex between D1.3 and HEL mutant Asp18 --> Ala. This mutation results in a DeltaDeltaG of only 0.3 kcal/mol, despite the loss of a hydrogen bond and seven van der Waals contacts to the Asp18 side chain. The crystal structure reveals that three additional water molecules are stably incorporated in the antigen-antibody interface at the site of the mutation. These waters help fill the cavity created by the mutation and form part of a rearranged solvent network linking the two proteins. To further dissect the energetics of specific interactions in the D1.3-HEL interface, double mutant cycles were carried out to measure the coupling of 14 amino acid pairs, 10 of which are in direct contact in the crystal structure. The highest coupling energies, 2.7 and 2.0 kcal/mol, were measured between HEL residue Gln121 and D1.3 residues VLTrp92 and VLTyr32, respectively. The interaction between Gln121 and VLTrp92 consists of three van der Waals contacts, while the interaction of Gln121 with VLTyr32 is mediated by a hydrogen bond. Surprisingly, however, most cycles between interface residues in direct contact in the crystal structure showed no significant coupling. In particular, a number of hydrogen-bonded residue pairs were found to make no net contribution to complex stabilization. We attribute these results to accessibility of the mutation sites to water, such that the mutated residues exchange their interaction with each other to interact with water. This implies that the strength of the protein-protein hydrogen bonds in these particular cases is comparable to that of the protein-water hydrogen bonds they replace. Thus, the simple fact that two residues are in direct contact in a protein-protein interface cannot be taken as evidence that there necessarily exists a productive interaction between them. Rather, the majority of such contacts may be energetically neutral, as in the D1.3-HEL complex.

Characterization of anti-anti-idiotypic antibodies that bind antigen and an anti-idiotype.

Two mouse monoclonal anti-anti-idiotopic antibodies (anti-anti-Id, Ab3), AF14 and AF52, were prepared by immunizing BALB/c mice with rabbit polyclonal anti-idiotypic antibodies (anti-Id, Ab2) raised against antibody D1.3 (Ab1) specific for the antigen hen egg lysozyme. AF14 and AF52 react with an "internal image" monoclonal mouse anti-Id antibody E5.2 (Ab2), previously raised against D1.3, with affinity constants (1.0 x 10(9) M-1 and 2.4 x 10(7) M-1, respectively) usually observed in secondary responses against protein antigens. They also react with the antigen but with lower affinity (1.8 x 10(6) M-1 and 3.8 x 10(6) M-1). This pattern of affinities for the anti-Id and for the antigen also was displayed by the sera of the immunized mice. The amino acid sequences of AF14 and AF52 are very close to that of D1.3. In particular, the amino acid side chains that contribute to contacts with both antigen and anti-Id are largely conserved in AF14 and AF52 compared with D1.3. Therapeutic immunizations against different pathogenic antigens using anti-Id antibodies have been proposed. Our experiments show that a response to an anti-Id immunogen elicits anti-anti-Id antibodies that are optimized for binding the anti-Id antibodies rather than the antigen.

Analysis of binding interactions in an idiotope-antiidiotope protein-protein complex by double mutant cycles.

The idiotope-antiidiotope complex between the anti-hen egg white lysozyme antibody D1.3 and the anti-D1.3 antibody E5.2 provides a useful model for studying protein-protein interactions. A high-resolution crystal structure of the complex is available [Fields, B. A., Goldbaum, F. A., Ysern, X., Poljak, R.J., & Mariuzza, R. A. (1995) Nature 374, 739-742], and both components are easily produced and manipulated in Escherichia coli. We previously analyzed the relative contributions of individual residues of D1.3 to complex stabilization by site-directed mutagenesis [Dall'Acqua, W., Goldman, E. R., Eisenstein, E., & Mariuzza, R. A. (1996) Biochemistry 35, 9667-9676]. In the current work, we introduced single alanine substitutions in 9 out of 21 positions in the combining site of E5.2 involved in contacts with D1.3 and found that 8 of them play a significant role in ligand binding (delta Gmutant-delta Gwild type > 1.5 kcal/mol). Furthermore, energetically important E5.2 and D1.3 residues tend to be juxtaposed in the crystal structure of the complex. In order to further dissect the energetics of specific interactions in the D1.3-E5.2 interface, double mutant cycles were carried out to measure the coupling of 13 amino acid pairs, 9 of which are in direct contact in the crystal structure. The highest coupling energy (4.3 kcal/mol) was measured for a charged-neutral pair which forms a buried hydrogen bond, while side chains which interact through solvated hydrogen bonds have lower coupling energies (1.3-1.7 kcal/mol), irrespective of whether they involve charged-neutral or neutral-neutral pairs. Interaction energies of similar magnitude (1.3-1.6 kcal/mol) were measured for residues forming only van der Waals contacts. Cycles between distant residues not involved in direct contacts in the crystal structure also showed significant coupling (0.5-1.0 kcal/mol). These weak long-range interactions could be due to rearrangements in solvent or protein structure or to secondary interactions involving other residues.

Hydrogen bonding and solvent structure in an antigen-antibody interface. Crystal structures and thermodynamic characterization of three Fv mutants complexed with lysozyme.

Using site-directed mutagenesis, X-ray crystallography, and titration calorimetry, we have examined the structural and thermodynamic consequences of removing specific hydrogen bonds in an antigen-antibody interface. Crystal structures of three antibody FvD1.3 mutants, VLTyr50Ser (VLY50S), VHTyr32Ala (VHY32A), and VHTyr101Phe (VHY101F), bound to hen egg white lysozyme (HEL) have been determined at resolutions ranging from 1.85 to 2.10 A. In the wild-type (WT) FvD1.3-HEL complex, the hydroxyl groups of VLTyr50, VHTyr32, and VHTyr101 each form at least one hydrogen bond with the lysozyme antigen. Thermodynamic parameters for antibody-antigen association have been measured using isothermal titration calorimetry, giving equilibrium binding constants Kb (M-1) of 2.6 x 10(7) (VLY50S), 7.0 x 10(7) (VHY32A), and 4.0 x 10(6) (VHY101F). For the WT complex, Kb is 2.7 x 10(8) M-1; thus, the affinities of the mutant Fv fragments for HEL are 10-, 4-, and 70-fold lower than that of the original antibody, respectively. In all three cases entropy compensation results in an affinity loss that would otherwise be larger. Comparison of the three mutant crystal structures with the WT structure demonstrates that the removal of direct antigen-antibody hydrogen bonds results in minimal shifts in the positions of the remaining protein atoms. These observations show that this complex is considerably tolerant, both structurally and thermodynamically, to the truncation of antibody side chains that form hydrogen bonds with the antigen. Alterations in interface solvent structure for two of the mutant complexes (VLY50S and VHY32A) appear to compensate for the unfavorable enthalpy changes when protein-protein interactions are removed. These changes in solvent structure, along with the increased mobility of side chains near the mutation site, probably contribute to the observed entropy compensation. For the VHY101F complex, the nature of the large entropy compensation is not evident from a structural comparison of the WT and mutant complexes. Differences in the local structure and dynamics of the uncomplexed Fv molecules may account for the entropic discrepancy in this case.

Crystal structure of an Fv-Fv idiotope-anti-idiotope complex at 1.9 A resolution.

Anti-idiotopic antibodies react with unique antigenic features, usually associated with the combining sites, of other antibodies. They may thus mimic specific antigens that react with the same antibodies. The structural basis of this mimicry is analyzed here in detail for an anti-idiotopic antibody that mimics the antigen, hen egg-white lysozyme. The crystal structure of an anti-hen-egg-white lysozyme antibody (D1.3) complexed with an anti-idiotopic antibody (E5.2) has been determined at a nominal resolution of 1.9 A. E5.2 contacts substantially the same residues of D1.3 as lysozyme, thus mimicking its binding to D1.3. The mimicry embodies conservation of hydrogen bonding: six of the 14 protein-protein hydrogen bonds bridging D1.3-E5.2 are structurally equivalent to hydrogen bonds bridging D1.3-lysozyme. The mimicry includes a similar number of van der Waals interactions. The mimicry of E5.2 for lysozyme, however, does not extend to the topology of the non-polar surfaces of E5.2 and lysozyme, which are in contact with D1.3 as revealed by a quantitative analysis of the contacting surface similarities between E5.2 and lysozyme. The structure discussed herein shows that an anti-idiotopic antibody can provide an approximate topological and binding-group mimicry of an external antigen, especially in the case of the hydrophilic surfaces, even though there is no sequence homology between the anti-idiotope and the antigen.

A mutational analysis of the binding of two different proteins to the same antibody.

The crystal structures of the complexes between the anti-hen egg white lysozyme (HEL) antibody D1.3 and HEL and between D1.3 and the anti-D1.3 antibody E5.2 have shown that D1.3 contacts these two proteins through essentially the same set of combining site residues [Fields, B. A., Goldbaum, F. A., Ysern, X., Poljak, R. J., & Mariuzza, R. A. (1995) Nature 374, 739-742]. To probe the relative contribution of individual residues to complex stabilization, single alanine substitutions were introduced in the combining site of D1.3, and their effects on affinity for HEL and for E5.2 were measured using surface plasmon resonance detection, fluorescence quench titration, or sedimentation equilibrium. The energetics of the binding to HEL are dominated by only 3 of the 13 contact residues tested (delta Gmutant-delta Gwild type > 2.5 kcal/mol): VLW92, VHD100, and VHY101. These form a patch at the center of the interface and are surrounded by residues whose apparent contributions are much less pronounced ( < 1.5 kcal/mol). This contrasts with the interaction of D1.3 with E5.2 in which most the contact residues (11 of 15) were found to play a significant role in ligand binding ( > 1.5 kcal/mol). Furthermore, even though D1.3 contacts HEL and E5.2 in very similar ways, the functionally important residues of D1.3 are different for the two interactions, with only substitutions at D1.3 positions VH100 and VH101 greatly affecting binding to both ligands. Thus, the same protein may recognize different ligands in ways that are structurally similar yet energetically distinct.

Crystal structure of the complex of the variable domain of antibody D1.3 and turkey egg white lysozyme: a novel conformational change in antibody CDR-L3 selects for antigen.

The crystal structure of the Fv fragment of the murine monoclonal anti-lysozyme antibody D1.3, complexed with turkey egg-white lysozyme (TEL), is presented. D1.3 (IgG1, kappa) is a secondary response antibody specific for hen egg-white lysozyme (HEL). TEL and HEL are homologous and differ in amino acid sequence in the antibody-antigen interface only at position 121. The side-chain of HEL residue Gln121 makes a pair of hydrogen bonds to main-chain atoms of the antibody light chain. In the D1.3-TEL structure, TEL residue His121 makes only one hydrogen bond with the light chain as a result of 129 degree and 145 degree change in peptide torsion angles for residues Trp92 and Ser93. Probably as a consequence of this conformational change, the D1.3-TEL association occurs at a much slower rate than the D1.3-HEL association. The D1.3-TEL complex is destabilized with respect to the D1.3-HEL interaction by the loss of two hydrogen bonds, exclusively due to the substitution of histidine for glutamine. While antibodies of secondary responses are indeed highly specific for antigen, this work demonstrates that by undergoing subtle conformational change antibodies can still recognize mutated protein antigens, albeit at a cost to affinity.

Protein motion and lock and key complementarity in antigen-antibody reactions.

Antibodies possess a highly complementary combining site structure to that of their specific antigens. In many instances their reactions are driven by enthalpic factors including, at least in the case of the reaction of monoclonal antibody D1.3 with lysozyme, enthalpy of solvation. They require minor structural rearrangements, and their equilibrium association constants are relatively high (10(7)-10(11) M-1). By contrast, in an idiotope--anti-idiotope (antibody-antibody) reaction, which is entropically driven, the binding equilibrium constant is only 1.5 x 10(5) M-1 at 20 degrees C. This low value results from a slow association rate (10(3) M-1 s-1) due to a selection of conformational states that allow one of the interacting molecular surfaces (the idiotope on antibody D1.3) to become complementary to that of the anti-idiotopic antibody. Thus, antibody D1.3 reacts with two different macromolecules: with its specific antigen, hen egg lysozyme, and with a specific anti-idiotopic antibody. Complementarity with lysozyme is closer to a "lock and key" model and results in high affinity (2-4 x 10(8) M-1). That with the anti-idiotopic antibody involves conformational changes at its combining site and it results in a lower association constant (1.5 x 10(5) M-1). Thus, an "induced fit" mechanism may lead to a broadening of the binding specificity but with a resulting decrease in the intrinsic binding affinity which may weaken the physiological function of antibodies.

Solvent rearrangement in an antigen-antibody interface introduced by site-directed mutagenesis of the antibody combining site.

The three-dimensional structure of a site-directed mutant of the bacterially expressed Fv fragment from monoclonal antibody D1.3, complexed to the specific antigen lysozyme has been determined to a nominal resolution of 1.8 A using X-ray diffraction data. The replacement of VL Trp92 by Asp allows two water molecules to occupy space taken by Trp92 in the wild-type complex, in agreement with a previous observation that water molecules play an important role in stabilizing this antigen-antibody complex. The equilibrium constant for the binding of the mutant Fv to the antigen decreases by three orders of magnitude (from 2.3 x 10(8) M-1 to 2.6 x 10(5) M-1). Titration calorimetry shows that this results from a smaller negative binding enthalpy (delta delta H = -16 kJ mol-1 at 24 degrees C), whereas the value of the binding entropy is not affected. Since in the complex between the mutated Fv and antigen the buried area has decreased relative to that of the wild-type Fv by about 150 A2, the contribution of the buried unit area to the decrease in free energy (delta Gzero) is approximately 117 J mol-1 (28 cal mol-1) per A2. The loss of interatomic contacts in replacing Trp by Asp permits an approximate calculation for the contribution of van der Waals interactions made by Trp92 in this complex, which gives an average of 2.1 kJ mol-1 (0.5 kcal mol-1) for contacts between carbon atoms.

Bound water molecules and conformational stabilization help mediate an antigen-antibody association.

We report the three-dimensional structures, at 1.8-A resolution, of the Fv fragment of the anti-hen egg white lysozyme antibody D1.3 in its free and antigen-bound forms. These structures reveal a role for solvent molecules in stabilizing the complex and provide a molecular basis for understanding the thermodynamic forces which drive the association reaction. Four water molecules are buried and others form a hydrogen-bonded network around the interface, bridging antigen and antibody. Comparison of the structures of free and bound Fv fragment of D1.3 reveals that several of the ordered water molecules in the free antibody combining site are retained and that additional water molecules link antigen and antibody upon complex formation. This solvation of the complex should weaken the hydrophobic effect, and the resulting large number of solvent-mediated hydrogen bonds, in conjunction with direct protein-protein interactions, should generate a significant enthalpic component. Furthermore, a stabilization of the relative mobilities of the antibody heavy- and light-chain variable domains and of that of the third complementarity-determining loop of the heavy chain seen in the complex should generate a negative entropic contribution opposing the enthalpic and the hydrophobic (solvent entropy) effects. This structural analysis is consistent with measurements of enthalpy and entropy changes by titration calorimetry, which show that enthalpy drives the antigen-antibody reaction. Thus, the main forces stabilizing the complex arise from antigen-antibody hydrogen bonding, van der Waals interactions, enthalpy of hydration, and conformational stabilization rather than solvent entropy (hydrophobic) effects.