Synthesis of bi-substrate state mimics of dihydropteroate synthase as potential inhibitors and molecular probes

The increasing emergence of resistant bacteria drives us to design and develop new antimicrobial agents. Pursuant to that goal, a new targeting approach of the dihydropteroate synthase enzyme, which serves as the site of action for the sulfonamide class of antimicrobial agents, is being explored. Using structural information, a new class of transition state mimics has been designed and synthesized that have the capacity to bind to the pterin, phosphate and para-amino binding sites. The design, synthesis and evaluation of these compounds as inhibitors of Bacillusanthracis dihydropteroate synthase is described herein. Outcomes from this work have identified the first trivalent inhibitors of dihydropteroate synthase whose activity displayed slow binding inhibition. The most active compounds in this series contained an oxidized pterin ring. The binding of these inhibitors was modeled into the dihydropteroate synthase active site and demonstrated a good correlation with the observed bioassay data, as well as provided important insight for the future design of higher affinity transition state mimics.     

Structural Advantage of Sugar Beet α-Glucosidase to Stabilize the Michaelis Complex with Long-chain Substrate

The α-glucosidase from sugar beet (SBG) is an exo-type glycosidase. The enzyme has a pocket-shaped active site, but efficiently hydrolyzes longer maltooligosaccharides and soluble starch due to lower K_m and higher k_cat/K_m for such substrates. To obtain structural insights into the mechanism governing its unique substrate specificity, a series of acarviosyl-maltooligosaccharides was employed for steady-state kinetic and structural analyses. The acarviosyl-maltooligosaccharides have a longer maltooligosaccharide moiety compared with the maltose moiety of acarbose, which is known to be the transition state analog of α-glycosidases. The clear correlation obtained between log K_i of the acarviosyl-maltooligosaccharides and log(K_m/k_cat) for hydrolysis of maltooligosaccharides suggests that the acarviosyl-maltooligosaccharides are transition state mimics. The crystal structure of the enzyme bound with acarviosyl-maltohexaose reveals that substrate binding at a distance from the active site is maintained largely by van der Waals interactions, with the four glucose residues at the reducing terminus of acarviosyl-maltohexaose retaining a left-handed single-helical conformation, as also observed in cycloamyloses and single helical V-amyloses. The kinetic behavior and structural features suggest that the subsite structure suitable for the stable conformation of amylose lowers the K_m for long-chain substrates, which in turn is responsible for higher specificity of the longer substrates.     

Detailed structural analysis of glycosidase/inhibitor interactions: complexes of Cex from Cellulomonas fimi with xylobiose-derived aza-sugars

Detailed insights into the mode of binding of a series of tight-binding aza-sugar glycosidase inhibitors of two fundamentally different classes are described through X-ray crystallographic studies of complexes with the retaining family 10 xylanase Cex from Cellulomonas fimi. Complexes with xylobiose-derived aza-sugar inhibitors of the substituted "amidine" class (xylobio-imidazole, K(i) = 150 nM; xylobio-lactam oxime, K(i) = 370 nM) reveal lateral interaction of the "glycosidic" nitrogen with the acid/base catalyst (Glu127) and hydrogen bonding of the sugar 2-hydroxyl with the catalytic nucleophile (Glu233), as expected. Tight binding of xylobio-isofagomine (K(i) = 130 nM) appears to be a consequence of strong interactions of the ring nitrogen with the catalytic nucleophile while, surprisingly, no direct protein contacts are made with the ring nitrogen of the xylobio-deoxynojirimycin analogue (K(i) = 5800 nM). Instead the nitrogen interacts with two ordered water molecules, thereby accounting for its relatively weaker binding, though it still binds some 1200-fold more tightly than does xylobiose, presumably as a consequence of electrostatic interactions at the active site. Dramatically weaker binding of these same inhibitors to the family 11 xylanase Bcx from Bacillus circulans (K(i) from 0.5 to 1.5 mM) is rationalized for the substituted amidines on the basis that this enzyme utilizes a syn protonation trajectory and likely hydrolyzes via a (2,5)B boat transition state. Weaker binding of the deoxynojirimycin and isofagomine analogues likely reflects the energetic penalty for distortion of these analogues to a (2,5)B conformation, possibly coupled with destabilizing interactions with Tyr69, a conserved, catalytically essential active site residue.     

Beta-galactosidase (Escherichia coli) has a second catalytically important Mg2+ site

It is shown here that Escherichia coli beta-galactosidase has a second Mg2+ binding site that is important for activity. Binding of Mg2+ to the second site caused the k(cat) (with oNPG as the substrate) to increase about 100 s(-1); the Km was not affected. The Kd for binding the second Mg2+ is about 10(-4)M. Since the concentration of free Mg2+ in E. coli is about 1-2 mM, the second site is physiologically significant. Non-polar substitutions (Ala or Leu) for Glu-797, a residue in an active site loop, eliminated the k(cat) increase. This indicates that the second Mg2+ site is near to Glu-797. The Ki values of transition state analogs were decreased by small but statistically significant amounts when the second Mg2+ site was occupied and Arrhenius plots showed that less entropic activation energy is required when the second site is occupied. These inhibitor and temperature results suggest that binding of the second Mg2+ helps to order the active site for stabilization of the transition state.     

Solid phase synthesis and binding affinity of peptidyl transferase transition state mimics containing 2'-OH at P-site position A76

All living cells are dependent on ribosomes to catalyze the peptidyl transfer reaction, by which amino acids are assembled into proteins. The previously studied peptidyl transferase transition state analog CC-dA-phosphate-puromycin (CCdApPmn) has important differences from the transition state, yet current models of the ribosomal active site have been heavily influenced by the properties of this molecule. One significant difference is the substitution of deoxyadenosine for riboadenosine at A76, which mimics the 3′ end of a P-site tRNA. We have developed a solid phase synthetic approach to produce inhibitors that more closely match the transition state, including the critical P-site 2′-OH. Inclusion of the 2′-OH or an even bulkier OCH₃ group causes significant changes in binding affinity. We also investigated the effects of changing the A-site amino acid side chain from phenylalanine to alanine. These results indicate that the absence of the 2′-OH is likely to play a significant role in the binding and conformation of CCdApPmn in the ribosomal active site by eliminating steric clash between the 2′-OH and the tetrahedral phosphate oxygen. The conformation of the actual transition state must allow for the presence of the 2′-OH, and transition state mimics that include this critical hydroxyl group must bind in a different conformation from that seen in prior analog structures. These new inhibitors will provide valuable insights into the geometry and mechanism of the ribosomal active site.     

Systematic optimization model and algorithm for binding sequence selection in computational enzyme design

A systematic optimization model for binding sequence selection in computational enzyme design was developed based on the transition state theory of enzyme catalysis and graph‐theoretical modeling. The saddle point on the free energy surface of the reaction system was represented by catalytic geometrical constraints, and the binding energy between the active site and transition state was minimized to reduce the activation energy barrier. The resulting hyperscale combinatorial optimization problem was tackled using a novel heuristic global optimization algorithm, which was inspired and tested by the protein core sequence selection problem. The sequence recapitulation tests on native active sites for two enzyme catalyzed hydrolytic reactions were applied to evaluate the predictive power of the design methodology. The results of the calculation show that most of the native binding sites can be successfully identified if the catalytic geometrical constraints and the structural motifs of the substrate are taken into account. Reliably predicting active site sequences may have significant implications for the creation of novel enzymes that are capable of catalyzing targeted chemical reactions.     

Structural, thermodynamic, and kinetic analyses of tetrahydrooxazine-derived inhibitors bound to beta-glucosidases

The understanding of transition state mimicry in glycoside hydrolysis is increasingly important both in the quest for novel specific therapeutic agents and for the deduction of enzyme function and mechanism. To aid comprehension, inhibitors can be characterized through kinetic, thermodynamic, and structural dissection to build an "inhibition profile." Here we dissect the binding of a tetrahydrooxazine inhibitor and its derivatives, which display Ki values around 500 nm. X-ray structures with both a beta-glucosidase, at 2 A resolution, and an endoglucanase at atomic (approximately 1 A) resolution reveal similar interactions between the tetrahydrooxazine inhibitor and both enzymes. Kinetic analyses reveal the pH dependence of kcat/Km and 1/Ki with both enzyme systems, and isothermal titration calorimetry unveils the enthalpic and entropic contributions to beta-glucosidase inhibition. The pH dependence of enzyme activity mirrored that of 1/Ki in both enzymes, unlike the cases of isofagomine and 1-deoxynojirimycin that have been characterized previously. Calorimetric dissection reveals a large favorable enthalpy that is partially offset by an unfavorable entropy upon binding. In terms of the similar profile for the pH dependence of 1/Ki and the pH dependence of kcat/Km, the significant enthalpy of binding when compared with other glycosidase inhibitors, and the tight binding at the optimal pH of the enzymes tested, tetrahydrooxazine and its derivatives are a significantly better class of glycosidase inhibitor than previously assumed.     

Inhibition of Rabbit Brain Prolyl Endopeptidase by N-Benzyloxycarbonyl-Prolyl-Prolinal, a Transition State Aldehyde Inhibitor

Prolyl endopeptidase cleaves peptide bonds on the carboxyl side of proline residues within a peptide chain. The enzyme readily degrades a number of neuropeptides including substance P, neurotensin, thyrotropin-releasing hormone, and luteinizing hormone-releasing hormone. The finding that the enzyme is inhibited by benzyloxycarbonyl-prolyl-proline, with a Ki of 50 microM, prompted the synthesis of benzyloxycarbonyl-prolyl-prolinal as a potential transition state analog inhibitor. Rabbit brain prolyl endopeptidase was purified to homogeneity for these studies. The aldehyde was found to be a remarkably potent inhibitor of prolyl endopeptidase with a Ki of 14 nM. This Ki is more than 3000 times lower than that of the corresponding acid or alcohol. By analogy with other transition state inhibitors, it can be assumed that binding of the prolinal residue to the S1 subsite and the formation of a hemiacetal with the active serine of the enzyme greatly contribute to the potency of inhibition. The specificity of the inhibitor is indicated by the finding that a variety of proteases were not affected at concentrations 150 times greater than the Ki for prolyl endopeptidase. The data indicate that benzyloxycarbonyl-prolyl-prolinal is a specific and potent inhibitor of prolyl endopeptidase and that consequently it should be of value in in vivo studies on the physiological role of the enzyme.     

How similar are enzyme active site geometries derived from quantum mechanical theozymes to crystal structures of enzyme‐inhibitor complexes? Implications for enzyme design

Quantum mechanical optimizations of theoretical enzymes (theozymes), which are predicted catalytic arrays of biological functionalities stabilizing a transition state, have been carried out for a set of nine diverse enzyme active sites. For each enzyme, the theozyme for the rate-determining transition state plus the catalytic groups modeled by side-chain mimics was optimized using B3LYP/6-31G(d) or, in one case, HF/3-21G(d) quantum mechanical calculations. To determine if the theozyme can reproduce the natural evolutionary catalytic geometry, the positions of optimized catalytic atoms, i.e., covalent, partial covalent, or stabilizing interactions with transition state atoms, are compared to the positions of the atoms in the X-ray crystal structure with a bound inhibitor. These structure comparisons are contrasted to computed substrate-active site structures surrounded by the same theozyme residues. The theozyme/transition structure is shown to predict geometries of active sites with an average RMSD of 0.64 A from the crystal structure, while the RMSD for the bound intermediate complexes are significantly higher at 1.42 A. The implications for computational enzyme design are discussed.     

Importance of hydrogen bonding for efficiency and specificity of the human mitochondrial DNA polymerase

To assess the contribution to discrimination afforded by base pair hydrogen bonding during DNA replication by the human mitochondrial DNA polymerase, we examined nucleoside mimics lacking hydrogen bond forming capability but retaining the overall steric shape of the natural nucleotide. We employed oligonucleotide templates containing either a deoxyadenosine shape mimic (dQ) or a deoxythymidine shape mimic (dF). Additionally, the nucleoside triphosphate analogs difluorotoluene deoxynucleoside triphosphate, 9-methyl-1-H-imidazo[(4,5)-b]pyridine deoxyribose triphosphate, and 4-methylbenzimidazole deoxyribose triphosphate (dZTP; another dATP shape mimic) were assayed. We used pre-steady state methods to determine the kinetic parameters governing nucleotide incorporation, k(pol) and K(d). In general, the loss of hydrogen bonding potential led to 2-3 kcal/mol reduction in ground state binding free energy, whereas effects on the maximum rate of polymerization were quite variable, ranging from negligible (dATP:dF) to nearly 4 kcal/mol (dZTP:dT). Although we observed only a 46-fold reduction in discrimination when dF was present in the template, there was a complete elimination of discrimination when dQ was present in the template. Our data with dF indicate that hydrogen bonding contributes 2.2 kcal/mol toward the efficiency of incorporation, whereas data with dQ (which may overestimate the effect due to poor steric mimicry) suggest a contribution of up to 6.8 kcal/mol. Taken together, the data suggest that sterics are necessary but not sufficient to achieve optimal efficiency and fidelity for DNA polymerase. Base pair hydrogen bonding contributes at least a third of the energy underlying nucleoside incorporation efficiency and specificity.     

Structural and kinetic analyses of arginine residues in the active site of the acetate kinase from Methanosarcina thermophila

Acetate kinase catalyzes transfer of the gamma-phosphate of ATP to acetate. The only crystal structure reported for acetate kinase is the homodimeric enzyme from Methanosarcina thermophila containing ADP and sulfate in the active site (Buss, K. A., Cooper, D. C., Ingram-Smith, C., Ferry, J. G., Sanders, D. A., and Hasson, M. S. (2001) J. Bacteriol. 193, 680-686). Here we report two new crystal structure of the M. thermophila enzyme in the presence of substrate and transition state analogs. The enzyme co-crystallized with the ATP analog adenosine 5'-[gamma-thio]triphosphate contained AMP adjacent to thiopyrophosphate in the active site cleft of monomer B. The enzyme co-crystallized with ADP, acetate, Al(3+), and F(-) contained a linear array of ADP-AlF(3)-acetate in the active site cleft of monomer B. Together, the structures clarify the substrate binding sites and support a direct in-line transfer mechanism in which AlF(3) mimics the meta-phosphate transition state. Monomers A of both structures contained ADP and sulfate, and the active site clefts were closed less than in monomers B, suggesting that domain movement contributes to catalysis. The finding that His(180) was in close proximity to AlF(3) is consistent with a role for stabilization of the meta-phosphate that is in agreement with a previous report indicating that this residue is essential for catalysis. Residue Arg(241) was also found adjacent to AlF(3), consistent with a role for stabilization of the transition state. Kinetic analyses of Arg(241) and Arg(91) replacement variants indicated that these residues are essential for catalysis and also indicated a role in binding acetate.     

Phenylalanine ammonia-lyase: Mirror-image packing of d- and l-phenylalanine and d- and l-transition state analogs into the active site

The binding of substrate and product analogs to phenylalanine ammonia-lyase (EC 4.3.1.5) from maize has been studied by a protection method. The ligand dissociation constants, K_L, were estimated from the variation with [L] of the pseudo-first-order rate constants for enzyme inactivation by nitromethane. The phenylalanine analogs d- and l-2-aminooxy-3-phenylpropionic acid showed K_L, values over 20,000-fold lower than the K_m for l-phenylalanine. From these and other K_L values it is deduced that when the enzyme binds l-phenylalanine the structural free energy stored in the protein is higher than when it binds the superinhibitors. Models for binding d- and l-phenylalanine and the superinhibitors are described. The enantiomeric pairs are considered to have similar K_L values because they pack into the active site in a mirror-image relationship. If the elimination reaction approximates to the least-motion course deduced on stereoelectronic grounds, the mirror-image packing of the superinhibitors into the active site mimics the conformation inferred for a transition state in the elimination. It appears, therefore, that structural changes take place in the enzyme as the transition state conformation is approached causing stored free energy to be released. This lowers the activation free energy for the elimination reaction and accounts for the strong binding by the above analogs.     

Dimensional mapping of the active site of cholesterol esterase with alkylboronic acid inhibitors

The cholesterol esterase-catalyzed hydrolysis of the water-soluble substrate p-nitrophenyl butyrate occurs via an acylenzyme mechanism, and is competitively inhibited by boronic acid transition state analog inhibitors. Accordingly, we undertook to dimensionally map the enzyme's active site via synthesis and characterization of a series of n-alkyl boronic acid inhibitors. The most potent of these is n-hexaneboronic acid, with a Ki = 13 +/- 1 microM, since inhibitor potency declines for both longer and shorter boronic acids. No inhibition is observed for methaneboronic acid and n-octaneboronic acid inhibits poorly, with a Ki of 7 mM. These results indicate that the ability of the enzyme to form tight complexes with boron-containing transition state analog inhibitors is sensitive to alkyl chain length. The trend in inhibitor potency is discussed in terms of substrate specificity of and transition state stabilization by cholesterol esterase, and has important implications for the design of optimal reversible inhibitors of the enzyme.     

Designing cyclic peptide inhibitor of dengue virus NS3-NS2B protease by using molecular docking approach

Peptides are preferred for designing inhibitors because of their high activity and specificity. Seven cyclopentapeptide inhibitors were designed in this study against dengue virus type 2 (DEN-2) NS3-NS2B protease: CKRRC, CGRRC, CRGRC, CRTRC, CTRRC, CKRKC and CRRKC. Docking analysis was performed to study the enzyme-inhibitor binding interactions. The free energy binding and estimated Ki values for all the inhibitors were found to be small (within micromolar range), indicating that the inhibitors bind considerably well to the binding site. The results showed that the cyclopentapeptide CKRKC was the best peptide inhibitor candidate with estimated free binding energy of -8.39 kcal/mol and Ki of 0.707 μM when compared to the standard inhibitor Bz-Nle-Lys-Arg-Arg-H that has been experimentally tested and shown to exhibit Ki value of 5.8 μM. Several modes of weak interactions were observed between the cyclopentapeptide CKRKC and the active site of DEN-2 NS3-NS2B protease. Thus, the cyclopentapeptide is proposed as a potential inhibitor to the NS3-NS2B protease activities of DEN-2. While these preliminary results are promising, further experimental investigation is necessary to validate the results.     

Transition state chirality and role of the vicinal hydroxyl in the ribosomal peptidyl transferase reaction

The ribosomal peptidyl transferase is a biologically essential catalyst responsible for protein synthesis. The reaction is expected to proceed through a transition state approaching tetrahedral geometry with a specific chirality. To establish that stereospecificity, we synthesized two diastereomers of a transition state inhibitor with mimics for each of the four ligands around the reactive chiral center. Preferential binding of the inhibitor that mimics a transition state with S chirality establishes the spatial position of the nascent peptide and the oxyanion and places the amine near the critical A76 2'-OH group on the P-site tRNA. Another inhibitor series with 2'-NH 2 and 2'-SH substitutions at the critical 2'-OH group was used to test the neutrality of the 2'-OH group as predicted if the hydroxyl functions as a proton shuttle in the transition state. The lack of significant pH-dependent binding by these inhibitors argues that the 2'-OH group remains neutral in the transition state. Both of these observations are consistent with a proton shuttle mechanism for the peptidyl transferase reaction.     

Batesian mimics influence mimicry ring evolution

Mathematical models of mimicry typically involve artificial prey species with fixed colorations or appearances; this enables a comparison of predation rates to demonstrate the level of protection a mimic might be afforded. Fruitful theoretical results have been produced using this method, but it is also useful to examine the possible evolutionary consequences of mimicry. To that end, we present individual-based evolutionary simulation models where prey colorations are free to evolve. We use the models to examine the effect of Batesian mimics on Müllerian mimics and mimicry rings. Results show that Batesian mimics can potentially incite Müllerian mimicry relationships and encourage mimicry ring convergence.     

Picomolar transition state analogue inhibitors of human 5'-methylthioadenosine phosphorylase and X-ray structure with MT-immucillin-A

Methythioadenosine phosphorylase (MTAP) functions solely in the polyamine pathway of mammals to remove the methylthioadenosine (MTA) product from both spermidine synthase (2.5.1.16) and spermine synthase (2.5.1.22). Inhibition of polyamine synthesis is a validated anticancer target. We designed and synthesized chemically stable analogues for the proposed transition state of human MTAP on the basis of the known ribooxacarbenium character at all reported N-ribosyltransferase transition states [Schramm, V. L. (2003) Acc. Chem. Res. 36, 588-596]. Methylthio-immucillin-A (MT-ImmA) is an iminoribitol tight-binding transition state analogue inhibitor with an equilibrium dissociation constant of 1.0 nM. The immucillins resemble the ribooxacarbenium ion transition states of N-ribosyltransferases and are tightly bound as the N4' cations. An ion pair formed between the iminoribitol cation and phosphate anion mimics the ribooxacarbenium cation-phosphate anion pair formed at the transition state and is confirmed in the crystal structure. The X-ray crystal structure of human MTAP with bound MT-Imm-A also reveals that the 5'-methylthio group lies in a flexible hydrophobic pocket. Substitution of the 5'-methylthio group with a 5'-phenylthio group gives an equilibrium binding constant of 1.0 nM. Methylthio-DADMe-immucillin-A is a pyrrolidine analogue of the transition state with a methylene bridge between the 9-deazaadenine group and the pyrrolidine ribooxacarbenium mimic. It is a slow-onset inhibitor with a dissociation constant of 86 pM. Improved binding energy with DADMe-immucillin-A suggests that the transition state is more closely matched by increasing the distance between leaving group and ribooxacarbenium mimics, consistent with a more dissociative transition state. Increasing the hydrophobic volume near the 5'-position at the catalytic site with 5'-phenylthio-DADMe-immucillin-A gave a dissociation constant of 172 pM, slightly weaker than the 5'-methylthio group. p-Cl-phenylthio-DADMe-immucillin-A binds with a dissociation constant of 10 pM (K(m)/K(i) value of 500000), the tightest binding inhibitor reported for MTAP. These slow-onset, tight-binding transition state analogue inhibitors are the most powerful reported for MTAP and have sufficient affinity to be useful in inhibiting the polyamine pathway.     

Inhibition of chymotrypsin by peptidyl trifluoromethyl ketones: determinants of slow-binding kinetics

A series of seven peptidyl trifluoromethyl ketone (TFK) inhibitors of chymotrypsin have been prepared which differ at the P1 and P2 subsites. Inhibition equilibria and kinetics of association and dissociation with chymotrypsin have been measured. The association rate of Ac-Phe-CF3 was measured at enzyme concentrations between 8 nM and 117 microM in order to examine the relation between the ketone/hydrate equilibrium of trifluoromethyl ketones and the "slow binding" by these inhibitors. The association rate decreases at high enzyme concentrations, indicating that TFK ketone is the reactive species and that conversion of TFK hydrate to ketone becomes rate limiting under these conditions. Inhibitors with hydrophobic side chains at P2 bind more tightly but more slowly to chymotrypsin, indicating that formation of van der Waals contacts between the P2 side chain and the His 57 and Ile 99 side chains of chymotrypsin is a relatively slow process. Inhibitor properties were compared to the Michaelis-Menten kinetic constants of a homologous series of peptide methyl ester and peptide amide substrates. Plots of log Ki vs log (kcat/Km) are linear with slopes of 0.65 +/- 0.2, indicating that these inhibitors are able to utilize 65% of the total binding energy between chymotrypsin and its hydrolytic transition state.     

Docking studies on glycoside hydrolase Family 47 endoplasmic reticulum alpha-(1-->2)-mannosidase I to elucidate the pathway to the substrate transition state

Alpha-(1-->2)-mannosidase I from the endoplasmic reticulum (ERManI), a Family 47 glycoside hydrolase, is a key enzyme in the N-glycan synthesis pathway. Catalytic-domain crystal structures of yeast and human ERMan1s have been determined, the former with a hydrolytic product and the latter without ligands, with the inhibitors 1-deoxymannojirimycin and kifunensine, and with a thiodisaccharide substrate analog. Both inhibitors were bound at the base of the funnel-shaped active site as the unusual 1C4 conformer, while the substrate analog glycon is a 3S1 conformer. In the current study, AutoDock was used to dock alpha-D-mannopyranosyl-(1-->2)-alpha-D-mannopyranose with its glycon in chair (1C4,4C1), half-chair (3H2,3H4,4H3), skew-boat (OS2,3S1,5S1), boat (2,5B,3,OB,B1,4,B2,5), and envelope (3E,4E,E3,E4) conformations into the yeast ERManI active site. Both docked energies and forces on docked ligand atoms were calculated to determine how the ligand distorts to the transition state. From these, we can conclude that (1) both 1C4 and OS2 can be the starting conformers; (2) the most likely binding pathway is 1C4-->3H2-->OS2-->3,OB-->3S1-->3E; (3) the transition state is likely to be close to a 3E conformation.     

Incorporating Motion into Investigations of mimicry

During the past thirty years, natural selection due to predation has been investigated with regard to prey motion in three areas that are relevant to the evolution of mimicry: (1) anti-apostatic selection, (2) locomotor mimicry, and (3) escape mimicry. Anti-apostatic selection, or selection against the odd individuals, arises when prey are at very high densities or when prey are Müllerian mimics. When prey are at high densities, motion of the prey increases selection against odd individuals. When the prey are Müllerian mimics, motion may also play an important role in strengthening selection against odd individuals. This may explain locomotor mimicry between Müllerian mimics. Locomotor mimicry arises when two distantly-related prey species appear alike in behaviour, and there is a corresponding suite of morphological, physiological, and biomechanical traits that the prey have in common. Locomotor mimicry has been demonstrated in Müllerian mimics. It is also predicted to occur in Batesian mimics but with important limitations due to selection by the predator for the prey to maintain the ability to escape if detected. Locomotor mimicry may also occur between palatable species that are alike as a result of unprofitable prey (or escape) mimicry. Escape mimicry arises when prey are difficult to capture. By frustration learning, the predator associates the colour of the prey with unprofitability. In all three instances, dis-similarity in colour or motion probably increases selection against the odd individual. In addition, the interaction of colour and motion gives rise to greater reliability of the signals to a specialist predator. However for a generalist predator, multiple component signals of the prey lead to errors in signal perception and greater risk of cheating. This revised version was published online in July 2006 with corrections to the Cover Date.