Reversibly bound and covalently attached ligands induce conformational changes in the omega loop, Cys69-Cys96, of mouse acetylcholinesterase.

We have used a combination of cysteine substitution mutagenesis and site-specific labeling to characterize the structural dynamics of mouse acetylcholinesterase (mAChE). Six cysteine-substituted sites of mAChE (Leu(76), Glu(81), Glu(84), Tyr(124), Ala(262), and His(287)) were labeled with the environmentally sensitive fluorophore, acrylodan, and the kinetics of substrate hydrolysis and inhibitor association were examined along with spectroscopic characteristics of the acrylodan-conjugated, cysteine-substituted enzymes. Residue 262, being well removed from the active center, appears unaffected by inhibitor binding. Following the binding of ligand, hypsochromic shifts in emission of acrylodan at residues 124 and 287, located near the perimeter of the gorge, reflect the exclusion of solvent and a hydrophobic environment created by the associated ligand. By contrast, the bathochromic shifts upon inhibitor binding seen for acrylodan conjugated to three omega loop (Omega loop) residues 76, 81, and 84 reveal that the acrylodan side chains at these positions are displaced from a hydrophobic environment and become exposed to solvent. The magnitude of fluorescence emission shift is largest at position 84 and smallest at position 76, indicating that a concerted movement of residues on the Omega loop accompanies gorge closure upon ligand binding. Acrylodan modification of substituted cysteine at position 84 reduces ligand binding and steady-state kinetic parameters between 1 and 2 orders of magnitude, but a similar substitution at position 81 only minimally alters the kinetics. Thus, combined kinetic and spectroscopic analyses provide strong evidence that conformational changes of the Omega loop accompany ligand binding.

nals by catalyzing hydrolysis of the neurotransmitter acetylcholine at a diffusion limited rate (2,3). The crystallographic structure of mouse AChE reveals a catalytic triad (Ser 203 , Glu 334 , and His 447 ) located at the bottom of a narrow active site gorge 20 Å in depth (4 -6). Because the cross-section of the physiological substrate acetylcholine is larger than the narrowest part of the gorge, the remarkably high turnover rate of AChE raises questions regarding substrate access to the catalytic site.
Molecular dynamic simulations suggest that rapid fluctuations of gorge width combined with diffusion facilitated by electrostatic forces could enhance substrate accessibility (7)(8)(9)(10). In addition, the high affinity and slowly dissociating complex of fasciculin and AChE retains slight residual catalytic activity (11,12), despite the occlusion of the active site gorge by fasciculin as shown in the crystal structures (5,13,14). Rapid fluctuations in residues lining the gorge walls may leave transient gaps at the fasciculin-AChE interface and may account for residual activity.
The large omega loop (⍀ loop), Cys 69 -Cys 96 , flanking the active site gorge in mouse AChE corresponds to the activation loop of Cys 60 -Cys 97 in Candida rugosa lipase, a related ␣/␤-fold hydrolase protein (15)(16)(17). Crystallographic studies of the lipase revealed that the activation loop occludes the active center in the absence of substrate but folds back in the presence of lipid substrate allowing its access. Although kinetic and structural studies of AChE have not revealed evidence for such large substrate, induced lid-like movements (18,19), high catalytic turnover rates for the cholinesterases might indicate that small amplitude motions along the ⍀ loop allow rapid access of incoming substrate and release of reaction product (19). To elucidate the nature of the ligand-dependent conformational changes of AChE, we have employed cysteine substitution mutagenesis and site-directed labeling with an environmentally sensitive fluorophore, acrylodan. The emission spectrum and quantum yield of the fluorophore are dependent on the effective dielectric constant and thus reflect the degree of solvent exposure and the local polarity experienced by the fluorophore (20). For example, when acrylodan is conjugated to a cysteine lining the gorge, upon fasciculin binding, it becomes sandwiched between the fasciculin loop and wall of the gorge, thereby becoming protected from solvent (20).
To examine further the role of the ⍀ loop in ligand binding, we have conjugated cysteines at various positions on the ⍀ loop and opposing gorge wall. Six single cysteine mutants were prepared for acrylodan conjugation (Fig. 1). Three were on the ⍀ loop as follows: L76C near the tip of the loop and E81C and E84C on their outer surface not lining the gorge. Two residues on the opposing face of the gorge H287C and Y124C were selected, along with a distal residue A262C whose temperature coefficient (B factor) would indicate flexible movement of an-other disulfide loop on which it resides (5,6). We examined the kinetics of substrate catalysis and inhibitor association with the modified enzymes, and we correlate these kinetic parameters with the spectroscopic changes in the conjugated acrylodan upon ligand association. Fluorescence measurements reveal changes in conformation reflected in the substituted side chains well removed from the active center gorge. The results suggest that ligand binding at the catalytic site allosterically alters the conformation of a specific segment of the ⍀ loop whereby gorge closure occurs and residue side chain positions distal to the binding site are affected.
Expression, Mutagenesis, and Purification of mAChE-Mouse AChE was produced by transfection of expression plasmid (pCDNA3, Invitrogen, San Diego, CA) containing an encoding cDNA where the AChE sequence was terminated at position 548. The plasmid was transfected into HEK293 cells. Cells were selected with G418 to obtain stable producing cell lines, and AChE was expressed as a secreted soluble enzyme in serum-free media (20). Mutant enzymes were generated by standard mutagenesis procedures, and cassettes containing the mutation were subcloned into pCDNA 3 (20). Nucleotide sequences of the cassettes were confirmed by double-stranded sequencing to ensure that spurious mutations were not introduced into the coding sequence. Affinity chromatography using (m-aminophenyl)trimethylammonium linked through a long chain to Sepharose CL-4B resin (Sigma) permitted one-step purification of AChE. From 4 to 6 liters of media, mutant and wild type enzyme were purified in quantities ranging between 5 and 25 mg, as described previously (22)(23)(24). Purity was ascertained by SDS-PAGE and by measurements of specific activity. Residue 262 is on a peripheral disulfide loop and in the crystal has a large thermal factor. A-D, Connolly surface representations of structure. A, unliganded AChE (6); B, TFK ϩ conjugated with AChE; note partial exposure of the white molecule, TFK ϩ , at the base of the gorge (33); C, fasciculin 2 bound AChE at the mouth of the gorge (5); D, fasciculin 2 complex with AChE, rotated 90 o . Acrylodan conjugated to E84C is shown in yellow; and acrylodan conjugated to E81C is shown in green. (Note in D the proximity between arginine 11 on Fas 2 and the acrylodan side chain at position 84).
Assay of Catalytic Activity-The spectrophotometric method of Ellman was used (25), and kinetic constants for acetylthiocholine hydrolysis were determined by fitting the observed rates as described (26). Titration of active sites with known concentrations of the irreversible phosphorylating agent, MEPQ, was accomplished by the method of Levy and Ashani (27).
Acrylodan Labeling-Mutant enzymes were pretreated with 0.25 mM dithiothreitol for 30 min at room temperature to ensure reduction of the introduced cysteine. Excess dithiothreitol was removed by use of a G-50 Sephadex spin column (Roche Molecular Biochemicals) equilibrated in 10 mM Tris, 100 mM NaCl, 40 mM MgCl 2 , pH 8.0. A volume of 1 l of acrylodan at 100 times the enzyme concentration was slowly mixed with the enzyme to achieve an ϳ5-fold molar excess of acrylodan to mutant enzyme. Labeling was allowed to proceed for at least 12 h at 4°C, and unreacted acrylodan was removed by size exclusion chromatography using Sephadex G-25 (Amersham Pharmacia Biotech) in 0.1 M sodium phosphate buffer, pH 7. Concentrations of acrylodan-labeled enzyme were determined from the maximal acrylodan absorbance found between 360 and 380 nm (⑀ ϳ16,400 M Ϫ1 cm Ϫ1 ). Stoichiometry of labeling of the various preparations, estimated from a comparison of enzyme concentration by protein (280 nm) to acrylodan (360 -380 nm) absorbance, ranged as follows: L76C, 0. Trifluoroacetophenone Inhibition-Picomolar amounts of enzyme in 0.01% bovine serum albumin in 0.1 M sodium phosphate buffer, pH 7.0, were reacted with TFK ϩ in the absence of substrate. Inhibition was monitored by measuring residual enzyme activity by removal of aliquots during the course of the reaction. Bimolecular rate constants of inhibition were determined by nonlinear fits of the data (28).
Spectrofluorometric Assays-Steady-state emission spectra were measured at room temperature using a Jobin Yvon/Spex FluoroMax II spectrofluorometer (Instrument S.A., Inc., Edison, NJ) with the excitation and emission bandwidths set at 5 nm. The excitation wavelength for acrylodan was set at 359 nm, and emission was monitored between 420 and 600 nm. Equilibrium dissociation constants, K d , for BW286c51 and edrophonium with the acrylodan-labeled enzyme were obtained by titration of a fixed quantity of labeled enzyme (54 -120 nM) with various concentrations of indicated inhibitors. K d values were obtained by monitoring the fractional decrease in the total area under the fluorescence emission curves from 420 to 600 nm for the acrylodan-labeled E84C or a limited segment of the emission between 450 and 485 nm for the acrylodan-labeled E81C. For ligands of high affinity such as BW286c51, where binding is nearly stoichiometric, data were fitted to Equation 1.
⌬F and ⌬F max are the change and maximum change in fluorescence, respectively; E t is the total enzyme concentration, and I t is the total inhibitor concentration. Association of TFK ϩ with acrylodan-labeled E81C and E84C was assessed from the kinetics of decrease in fluorescence at 470 and 477 nm respectively, following addition of a stoichiometric excess TFK ϩ at several concentrations. Data were fitted to a single exponential approach to equilibrium.
Association and dissociation rate constants of edrophonium and BW286c51 with E81C and E84C AChEs were determined from changes in the tryptophan fluorescence using a stopped-flow spectrophotometer as described previously (29). Time-dependent decreases in tryptophan fluorescence were observed upon excitation at 276 nm by means of a 305-nm emission cut-off filter.

Characterization of Substrate Hydrolysis and Fasciculin 2
Inhibition-The cysteine-substituted enzymes show kinetics of acetylthiocholine hydrolysis similar to wild type enzyme (Table  I and Scheme 1) suggesting that all mutant enzymes fold correctly despite the presence of the newly introduced cysteine. The K m value of E84C shows slightly less than a 4-fold increase, whereas the change in turnover rate, k cat , is minimal. Similar changes in kinetic constants were observed previously for E84Q mAChE (28). Since K m , in diffusion limited catalysis, depicts the initial encounter between substrate and enzyme, an increase in K m likely arises from the reduction of negative charge that electrostatically steers the cationic substrate into the active center gorge. Interestingly, a similar E81C mutation has little or no effect on substrate hydrolysis. Not all negatively charged residues around the active center appear to be involved equivalently in electrostatic steering.
Association and dissociation rates of fasciculin with A262C, H287C, and Y124C mutant enzymes were also found to be close to the rates with wild type enzyme (20). Fasciculin, at low concentrations, is also capable of associating with the mutant enzymes after acrylodan conjugation (Fig. 2). In addition, enzyme activity measurements of fasciculin-bound acrylodan conjugates show greater than 99% inhibition (data not shown).
Influence of Residue Modification on Inhibition by m-Trimethylammoniotrifluoromethylacetophenone-TFK ϩ binding to cysteine-substituted enzymes, both free and modified with acrylodan, was also examined (Table II). For E81C and E84C, the association rate constants (k on ) for TFK ϩ were obtained from measurements of enzyme activity. Although positions 81 and 84 are both spatially removed from TFK ϩ -binding site, k on SCHEME 1. In this scheme substrate can combine at two discrete sites to form two binary complexes, ES and SE (where S is substrate; E is enzyme; and P is product). Only ES results in substrate hydrolysis. For simplicity, S is assumed to combine equally well with E and ES. The efficiency of substrate hydrolysis of the ternary complex SES, as compared with ES, is reflected in the value of the parameter, b, the relative catalytic turnover of the ternary complex (26).
where [S] is substrate concentration, K SS is the substrate inhibition or activation constant, and b is the relative catalytic turnover of the ternary complex (12).
for E84C is slightly slower than that for wild type enzyme. By contrast, E81C shows no difference in the kinetic constants. Conjugation of acrylodan, a neutral naphthalene derivative, with E84C reduces k on of TFK ϩ 7-fold compared with unconjugated E84C, whereas conjugation of E81C with acrylodan only reduces k on of TFK ϩ slightly. For acrylodan-labeled mutants, k on was measured from the time-dependent decrease of fluorescence signal (Fig. 3).
Influence of Residue Modification on Inhibition by Noncovalent Active Site Inhibitors-A similar trend in inhibition kinetics was seen with noncovalent active site inhibitors such as edrophonium and BW286c51 (Table II). An increase over wild type K d of 2-fold occurs for edrophonium binding to E84C, and an 18-fold increase in K d is observed for BW286c51 binding. Similar increases in K d of edrophonium and BW286c51 were seen for E84Q human AChE (18). By comparison, E81C showed no alterations in ligand binding constants. For acrylodan-labeled mutants, K d was measured from the fluorescence signals of an equilibrium titration (Fig. 4). Acrylodan-labeled E84C shows K d increases of 10-fold for edrophonium and 3-fold for BW286c51 compared with unreacted E84C. For acrylodan-labeled E81C, only a slight increase in K d is seen for both ligands. The high concentration of acrylodan-labeled E81C required for equilibrium titrations precludes an accurate estimate of K d for high affinity ligands such as BW286c51.
Effect of Fasciculin on Acrylodan Fluorescence Emission-The peptide toxin, fasciculin, inhibits AChE by tightly capping the mouth of active center gorge ( Fig. 1) (11, 30 -32). Table III shows changes in emission maxima of acrylodan-labeled AChE mutants in the presence of fasciculin. There is no discernible change in fluorescence emission of acrylodan-conjugated A262C (20), consistent with the position 262 being distal to the fasciculin-binding site. The large hypsochromic shifts seen at both the 124 and 287 positions reflect solvent exclusion and an increase in hydrophobicity experienced by the fluorophores in the gorge upon fasciculin binding (20). For the ⍀ loop mutant, L76C, fasciculin binding produces a 40% increase in quantum yield but no change in emission maximum. Bathochromic shifts are found at both the 81 and 84 positions, with position 84 producing a shift of larger magnitude ( Fig. 2 and Table III).
Effect of Covalently Conjugated Active Site Inhibitors on Acrylodan Fluorescence Emission-Changes in emission maxima of acrylodan-labeled AChE mutants in the presence of conjugating trifluoroacetophenones are shown in Table IV. The trifluoroacetophenones inhibit the enzyme by conjugating to form a hemiketal at active site serine without dissociation of leaving group (33). Both the isosteric neutral and cationic trifluoroketones (TFK 0 and TFK ϩ ) produced no discernible changes in emission spectra of acrylodan conjugated at H287C and A262C, consistent with a fluorophore position distant from gorge base and hence not in direct contact with ligand. Remarkably, both TFK 0 and TFK ϩ produce a substantial bathochromic shift (at least 30 nm) with acrylodan-E84C. The trifluoroketones also produce spectral shift of intermediate value (20 nm (Table IV). Effect of Noncovalent Active Site Inhibitors on Acrylodan Fluorescence Emission-Noncovalent active site inhibitors, such as edrophonium, tacrine, and huperzine, associate primarily with the choline subsite at the base of active site gorge. Crystal structures of inhibitors bound to Torpedo californica AChE revealed that these ligands should have no direct contact with the conjugated fluorophore at all six cysteine-substituted sites (36,37). Upon edrophonium, tacrine, or huperzine association, alteration of acrylodan emission maxima is undetectable for positions 124, 287, and 262 (Table V). However, as seen for other ligands, acrylodan conjugated at E84C surprisingly shows a bathochromic shift of 33 nm (from 477 to 510 nm) upon inhibitor binding. A change of smaller magnitude is seen in the case of acrylodan-L76C (from 505 to 509 nm) and acrylodan-E81C (from 480 to 510 nm) with noncovalent active site inhibitors. Ligand binding results in a common emission maximum ( max ϳ510 nm) for acrylodan at the three ⍀ loop positions.
Effect of Bisquaternary Inhibitors on Acrylodan Emission Spectrum-Extended bisquaternary inhibitors, such as BW286c51 and decamethonium, belong to a class of inhibitors that interact with two binding sites of AChE simultaneously (32, 38 -39). The quaternary ammonium moiety on one end of the molecule associates with the Trp 86 residue that characterized the choline-binding site, whereas the other end resides near Trp 286 at the active site gorge rim. Table VI shows changes in emission maxima of acrylodan-labeled AChE mutants in the presence of bisquaternary inhibitors. No changes are observed at position 262. By contrast, both decamethonium and BW284c51 caused a pronounced hypsochromic shift and increase in quantum yields with acrylodan conjugated at Y124C and H287C. Addition of decamethonium produced a hypsochromic shift of 35 nm at position 124, and a modest 7 nm shift at position 287. BW284c51 has a similar effect; for the ⍀ loop mutants, L76C, E81C, and E84C, bathochromic shifts of similar magnitude to the monoquaternary ligands were observed (Tables V and VI).

Characteristics of Fluorescence from Acrylodan-conjugated
Cysteine Residues-Fluorescence emission of acrylodan is exquisitely sensitive to the dielectric constant of the solvent. In general, the fluorescence emission spectrum of acrylodan shifts toward the red (bathochromic) shift, and the quantum yield decreases as the polarity of solvent increases (20, 40 -42). This sensitivity to solvent polarity arises from the interaction of the excited state of acrylodan with its surrounding solvent. The excited state is more polar than the ground state and, as such, will interact with a polar solvent so as to align solvent dipoles. This alignment lowers the energy of the excited state and causes the red shift of the emission spectrum. Hence, an acrylodan-labeled enzyme with an emission maximum of 510 -525 nm likely reflects exposure of the side chain to solvent (20,42). On the other hand, acrylodan emission maxima in the range of 475-500 nm likely reflect solvent exclusion and a more hydrophobic environment surrounding the fluorophore. The time course of TFK ϩ reaction with acrylodan-E84C (Fig. 3) reveals a large spectral shift from 477 to 512 nm, indicating acrylodan conjugated at this position has moved to a more hydrophilic environment with TFK ϩ bound. The large spectral shift yields a clear isoemissive point, which arises when only two distinct emitting species are present, in this case the free enzyme and the TFK ϩ conjugate.
Influence of Residue Modification on Ligand Binding-The changes in emission spectra of acrylodan-labeled ⍀ loop residues 81 and 84 have been exploited to monitor ligand binding   (Table II). We observe that cysteine substitution and acrylodan conjugation at position 84 affect ligand binding kinetics but not at position 81. Cysteine substitution at position 84 has little influence on catalytic parameters derived from steady-state catalysis ( Table I). The K m of E84C increases less than 4-fold compared with the wild type enzyme. By contrast, a similar substitution at position 81 has no effect on ATCh steady-state catalysis. Precise quantitation of these catalytic parameters for the acrylodan-conjugated enzyme is complicated by incomplete modification by acrylodan. However, inhibitor association can be measured using the change in fluorescence signal (Table II).
Here we observe reductions in binding kinetics for several ligands (Table II) ranging between 1 or 2 orders of magnitude at position 84 but very little change at position 81. Although a portion of the reduction at position 84 is due to the cysteine substitution, acrylodan conjugation has a small, but significant (3-10-fold), influence on ligand binding. Even though both the 81 and 84 residues reside on the enzyme surface removed from the active center gorge, modification only at position 84 appreciably affects the energetics of ligand binding. The acrylodan moiety, whose dimension is slightly larger than the indole moiety of tryptophan, may impart steric restrictions to the region around the 84 site contributing to the energy cost in ligand binding. A small alteration in ligand binding energy (1.5-3.0 kcal/mol) is not unexpected if the conformation of ⍀ loop plays a role in ligand binding.
Velan et al. (18) have examined steady-state kinetics for a large number of ⍀ loop substitutions and truncations. Modification of Glu 84 and its neighboring residues were found to have limited effect on steady-state kinetics. Faerman et al. (19) inserted a cysteine at position 82 to pair with a second cysteine residing proximally in the body of the enzyme. Although it could not be firmly established that a disulfide bond formed, little change in kinetic parameters (K m and k cat ) was observed. Because of compensating contributions of the component primary constants, it is often difficult to correlate changes in steady-state kinetic parameters with structural perturbations. Our site-directed fluorophore labeling provides a physical assessment of the localized conformational change in the ⍀ loop. In cases where the fluorophore makes direct contact with the ligand, as for acrylodan-labeled Y124C and H287C with fasciculin, the energetic perturbations from substitution are larger, since complementarity of the binding site may be altered through the insertion of acrylodan side chain at the interface between the ligand and its binding site (20).
Acrylodan Modification at a Site Distal to the Active Center Core-We chose the A262C modification as a positional reference for a site distal to the active center. This residue is also located at the tip of a disulfide loop but is located ϳ30 Å away from the rim of the active center gorge. Crystallographic studies show this region to have a high temperature coefficient (B factor), indicative of substantial molecular motion of this surface residue. In fact, the position of this residue and its immediate neighbors is only secured in crystal forms where proximity of the symmetry-related AChE molecule limits its movement in the crystal structure (6).
Acrylodan substitutions at this position show a long wavelength emission ( max ϭ 517 nm) indicative of exposure to a hydrophilic environment (Table III). Moreover, none of the ligands studied, whether they are covalently attached to the active center (TFK or alkylphosphates), reversibly bound to the active center (edrophonium), span between the active center and peripheral site (decamethonium and BW286c51), or bind only to peripheral site (fasciculin), affect the spectroscopic properties of acrylodan conjugated at site 262 (Tables III-VI). This pattern indicates a lack of global conformational change affecting residue environments in a disulfide loops well removed from the active center (Fig. 1).
Residues Residing on the Active Center Gorge in Apposition with the ⍀ Loop-Residues 124 and 287 lie in close proximity to the ⍀ loop with H287C at the rim of the gorge and Y124C, residing just below the rim in the gorge interior (Fig. 1). The crystal structure of the complex shows fasciculin to "cap" these residues, and our previous studies show hypsochromic shifts of acrylodan upon fasciculin binding (20). None of the reversibly bound active center ligands (edrophonium, huperzine, and tacrine) induce a spectral shift at position 124 or 287. However, modest quenching is observed at position 124 upon binding of these active center ligands. The bisquaternary ligands, which should approach or come in close apposition with these residues, cause significant hypsochromic shifts. The large shift for decamethonium at position 124 may reflect the ability of the cluster of aromatic residues to collapse around the methylene chain of decamethonium enlodged within the active center gorge. Crystallographic studies show one quaternary ammonium of decamethonium to be consistently positioned in the vicinity of Trp 84 ; however, both the flexible side chain and the outermost quaternary group are found to assume multiple positions in the decamethonium-AChE complexes studied to date (6,29).
The distinct spectra observed for the two isosteric trifluoroacetophenone conjugates is surprising (Table IV). Covalent inhibition of cationic trifluoroacetophenone (TFK ϩ ) produces very little spectral shift of acrylodan at either position 124 or 287. This is consistent with the crystal structures where the trimethyl ammonio moiety of TFK ϩ forms a cation-interaction with Trp 86 , and the trifluoroacetophenone moiety forms a hemiketal bond with the active center serine 203 (33). However, the isosteric t-butyl congener (TFK 0 ) shifts the environ-  Table III. ment of residue 124 to that resembling a hydrophobic state. This difference suggests that the orientation of this hemiketal conjugate differs where the t-butyl group extends toward the gorge exit. TFK 0 inhibits the wild type enzyme 70-fold slower than TFK ϩ , presumably due to lack of cation-interaction and slightly different ligand orientation (21). Alkyl phosphorylation with small alkyl groups also has little influence on the environment at position 124 (Table IV).
⍀ Loop Substitutions-Our greatest surprise emerged from studies on the outer portion of the ⍀ loop, defined by residues between Cys 69 and Cys 96 , where we have examined three positions extending from the near tip of the loop (Leu 76 ) at the gorge rim descending toward the active center (Glu 81 and Glu 84 ). The residues modified are all on the outer surface and do not form the inner gorge wall. Since residues 81 and 84 carry acidic side chains, they might be expected to show solvent exposure in the native enzyme and not be involved in the internal stabilization of the loop, as is evident in the crystal structure of the mouse enzyme (5,6). In the absence of ligand, the spectra of the conjugated acrylodan moiety reveal different degrees of solvent exposure with the acrylodan at position 84 being the most protected in an hydrophobic environment, acrylodan at 81 being intermediate, and acrylodan at 76 being most exposed. Examination of crystal structures of mouse enzyme revealed a surface cavity near the side chain of the 84 site (5,6). The observed max likely reflects acrylodan buried in this surface cavity when conjugated to the 84 site (Fig. 1).
The presence of fasciculin causes a large bathochromic shift of acrylodan fluorescence at both the 81 and 84 positions, as well as increase in quantum yield of acrylodan at 76. The lack of a shift in emission seen for acrylodan at the 76 position may simply reflect a balance between a small environmental change at 76 upon ligand binding in general and partial solvent occlusion at this position by fasciculin. In the case of Glu 84 , the bathochromic shift likely reflects Arg 11 of fasciculin loop I coming in van der Waals contact with the 84 side chain and displacing acrylodan into a more polar environment. However, an explanation of the bathochromic shift at position 81 requires a more involved analysis. Although 81 is removed from the fasciculin-binding site, fasciculin has a sufficient molecular dimension to restrict the ⍀ loop so that the entire loop freezes or closes upon fasciculin binding. Thus, fasciculin binding may confer strain on the ␣-carbon backbone structure of the ⍀ loop such that the acrylodan side chain at positions 81 and 84 becomes exposed to the hydrophilic environment. The fact that substitutions at both positions yielded acrylodan spectra with equivalent emission maxima after ligand binding suggests a conformational involvement of the entire loop.
Similar to fasciculin, small ligands that bind to the active center produce a similar strain. All of the small ligands, whether reversibly bound or covalently attached, elicit marked changes in acrylodan emission with the largest spectral shift seen for E84C, an intermediate value seen for E81C, and only small change observed for L76C. In each case the conformational change induced by the ligand causes the acrylodan to move into a region of higher dielectric constant, presumably being more solvent-exposed. The pattern is remarkably consistent among the ligands, and only the small organophosphate when conjugated induces a shift of smaller magnitude. A likely explanation for the observed conformational changes is that ligand binding to the active center induces gorge closure, which is mediated throughout the ⍀ loop. The strain placed on the ␣-carbon backbone upon gorge closure causes the side chains to shift positions and become exposed to hydrophilic environment.
DeFarri et al. (43) have noted that the peripheral site inhibitor, thioflavin T, when bound to AChE, shows a large enhance-ment of fluorescence. Simultaneous binding of an active center ligand and thioflavin partially quenches the enhanced fluorescence of bound thioflavin. Radic and Taylor (29) have observed that bound active center ligands cause a partial quenching of the native tryptophan fluorescence in AChE. Since these ligands lack the spectral overlap for fluorescence resonance energy transfer, the bound ligand is likely to influence the connectivity between aromatic residues present in the gorge, thereby influencing fluorescence quantum yields. Taken together, these studies suggest that ligands induce conformational changes in AChE giving rise to a gorge conformation collapsed around the bound ligand. Our site-directed cysteine mutagenesis and fluorescence labeling studies allow one to delineate the involvement of particular residues on the ⍀ loop in this conformational change.
Cystallographic Structures and Solution Dynamics of the Acetylcholinesterase Complex-In the several crystal structures of AChE with conjugated or reversibly bound ligand that have been studied, little evidence for change in enzyme conformation has been detected with a difference of less than a root mean square of 1 Å 2 for the ␣-carbon backbone between the apoenzyme and the various complexes (4 -6, 13, 14, 33-35, 44). Changes in side chain orientation occur most notably in the phenyl ring at position 337 for certain reversible complexes (34) and phenylalanine 297, when bulky organophosphates are conjugated to the active site serine (44). However, based on the multiple positions of the outer trimethylammonio moiety in decamethonium for mouse (6) and Torpedo crystal structures (34), some flexibility may exist particularly within the gorge itself. Brownian dynamics often require reducing the radii of the attacking ligand or the residues lining the gorge in order to simulate the kinetics of diffusion-limited substrate access observed experimentally (45). Thus, all of crystal structures reported to date reveal a closed gorge with constrained dimensions. Our solution-based fluorescence studies provide the first physical evidence for localizing the ligand-induced conformational change to residues in the Cys 69 -Cys 96 ⍀ loop. These findings raise an interesting possibility that the unliganded enzyme exists in a rapidly converting conformational equilibrium between open and closed states, and both ligand binding and conditions of crystallization favor formation of a closed gorge state. In fact, analysis of the molecular dynamics of a solvated mouse AChE shows fluctuations yielding an average widening of the gorge over a 10-ns interval (46). Such opening and closing motions of the gorge may also be integral to the catalytic cycle of transacylation and deacylation during ester hydrolysis.