Thioflavin T Is a Fluorescent Probe of the Acetylcholinesterase Peripheral Site That Reveals Conformational Interactions between the Peripheral and Acylation Sites*

Three-dimensional structures of acetylcholinesterase (AChE) reveal a narrow and deep active site gorge with two sites of ligand binding, an acylation site at the base of the gorge, and a peripheral site near the gorge entrance. Recent studies have shown that the peripheral site contributes to catalytic efficiency by transiently binding substrates on their way to the acylation site, but the question of whether the peripheral site makes other contributions to the catalytic process remains open. A possible role for ligand binding to the peripheral site that has long been considered is the initiation of a conformational change that is transmitted allosterically to the acylation site to alter catalysis. However, evidence for conformational interactions between these sites has been difficult to obtain. Here we report that thioflavin T, a fluorophore widely used to detect amyloid structure in proteins, binds selectively to the AChE peripheral site with an equilibrium dissociation constant of 1.0 μm. The fluorescence of the bound thioflavin T is increased more than 1000-fold over that of unbound thioflavin T, the greatest enhancement of fluorescence for the binding of a fluorophore to AChE yet observed. Furthermore, when the acylation site ligands edrophonium or m-(N,N,N-trimethylammonio)trifluoroacetophenone form ternary complexes with AChE and thioflavin T, the fluorescence is quenched by factors of 2.7–4.2. The observation of this partial quenching of thioflavin T fluorescence is a major advance in the study of AChE for two reasons. First, it allows thioflavin T to be used as a reporter for ligand reactions at the acylation site. Second, it indicates that ligand binding to the acylation site initiates a change in the local AChE conformation at the peripheral site that quenches the fluorescence of bound thioflavin T. The data provide strong evidence in support of a conformational interaction between the two AChE sites.

Three-dimensional structures of acetylcholinesterase (AChE) reveal a narrow and deep active site gorge with two sites of ligand binding, an acylation site at the base of the gorge, and a peripheral site near the gorge entrance. Recent studies have shown that the peripheral site contributes to catalytic efficiency by transiently binding substrates on their way to the acylation site, but the question of whether the peripheral site makes other contributions to the catalytic process remains open. A possible role for ligand binding to the peripheral site that has long been considered is the initiation of a conformational change that is transmitted allosterically to the acylation site to alter catalysis. However, evidence for conformational interactions between these sites has been difficult to obtain. Here we report that thioflavin T, a fluorophore widely used to detect amyloid structure in proteins, binds selectively to the AChE peripheral site with an equilibrium dissociation constant of 1.0 M. The fluorescence of the bound thioflavin T is increased more than 1000-fold over that of unbound thioflavin T, the greatest enhancement of fluorescence for the binding of a fluorophore to AChE yet observed. Furthermore, when the acylation site ligands edrophonium or m-(N, N,Ntrimethylammonio)trifluoroacetophenone form ternary complexes with AChE and thioflavin T, the fluorescence is quenched by factors of 2.7-4.2. The observation of this partial quenching of thioflavin T fluorescence is a major advance in the study of AChE for two reasons. First, it allows thioflavin T to be used as a reporter for ligand reactions at the acylation site. Second, it indicates that ligand binding to the acylation site initiates a change in the local AChE conformation at the peripheral site that quenches the fluorescence of bound thioflavin T. The data provide strong evidence in support of a conformational interaction between the two AChE sites.
Acetylcholinesterase (AChE) 1 hydrolyzes the neurotransmitter acetylcholine at extremely high catalytic rates (1). Ligand binding studies (2) and x-ray crystallography (3) have revealed a narrow active site gorge some 20 Å deep with two separate ligand binding sites. During catalytic hydrolysis, the substrate acyl group is transferred briefly to residue Ser-200 2 in the acylation site at the bottom of the gorge. This site contains residues involved in a catalytic triad (His-440, Glu-327, Ser-200) and Trp-84, which binds to the trimethylammonium group of acetylcholine. The peripheral site near the mouth of the gorge includes, among others, residues Asp-72 and Trp-279. Recent investigations have shown that the peripheral site contributes to catalytic efficiency by transiently binding substrates on their way to the acylation site (4 -6).
The question of whether the peripheral site plays additional roles in the catalytic process has long been of interest. Changeux (7) was among the first to appreciate that AChE contained two distinct ligand binding sites and that there may be allosteric interactions between ligands bound at these sites involving conformational changes in the protein molecule. Ligands bound to the peripheral site can inhibit or accelerate reactions at the acylation site, and these effects have often been attributed to conformational interactions between the sites (8,9). However, direct evidence for such conformational interactions has been difficult to obtain. For example, three-dimensional structures of AChE complexes with two ligands specific for the acylation site, edrophonium and the cationic trifluoromethylketone TMTFA, showed no changes in the structure of the peripheral site (10,11). In this report we show that thioflavin T ( Fig. 1) binds specifically to the AChE peripheral site and is one of the most useful fluorescent probes of AChE yet discovered.
This fluorophore frequently is used to detect amyloid structure in proteins (12), but the three-dimensional structure of the AChE peripheral site shows no indication of the extensive ␤ sheet structure characteristic of amyloid. The intense fluorescence of thioflavin T bound to AChE is partially quenched by the binding of acylation site ligands in ternary complexes, and this quenching appears to result from a conformational interaction between the two sites. (13) was purified as outlined previously, and active site AChE concentrations were determined by assuming 450 units/nmol (6). 3 Thioflavin T chloride (Sigma) concentrations were assigned as 70% of the dry weight. Thioflavin T chloride recrystallized from water gave an extinction coefficient ⑀ 412 nm of 36,000 M Ϫ1 cm Ϫ1 . Concentrations of propidium iodide (Calbiochem) were determined with an extinction coefficient ⑀ 493 nm of 5900 M Ϫ1 cm Ϫ1 (2), and concentrations of edrophonium chloride (ethyl(3-hydroxyphenyl)dimethylammonium chloride; Sigma) were determined with ⑀ 271 nm of 3400 M Ϫ1 cm Ϫ1 from the dry weight (14). TMTFA (kindly provided by Dr. Daniel Quinn, University of Iowa) concentrations were calibrated by titration with AChE.

Materials-Recombinant human AChE
Steady State Inhibition of Enzyme-catalyzed Substrate Hydrolysis-Substrate hydrolysis rates were measured in buffer (20 mM sodium phosphate and 0.02% Triton X-100 at pH 7.0) at 25°C after addition of small aliquots of thioflavin T, acetylthiocholine, and 5,5Ј-dithiobis-(2nitrobenzoic acid) (to a final concentration of 0.33 mM) in a total volume of 1.0 ml. An Ellman assay (15) was used to measure formation of the thiolate dianion of 5,5Ј-dithiobis-(2-nitrobenzoic acid) at 412 nm (⌬⑀ 412 nm ϭ 14.15 mM Ϫ1 cm Ϫ1 ; (16)) for 1-3 min on a Varian Cary 3A spectrophotometer, and substrate concentrations were corrected for substrate depletion resulting from hydrolysis during this interval. It was assumed that acetylthiocholine concentrations were maintained at low enough values (Ͻ1.0 mM) to ignore substrate inhibition in the absence of inhibitors (4). In this case, a simplified scheme for inhibition by peripheral site ligands is given in Scheme 1.
In this scheme, inhibitor (I) can bind to each of the three enzyme species: E, ES, and EA. For example, ESI P represents a ternary complex with substrate (S) at the acylation site and I at the peripheral site (denoted by the subscript P). The acylation rate constant k 2 is altered by a factor a in this ternary complex. Reciprocal plots of Ϫ1 versus [S] Ϫ1 at all I concentrations here were linear, a result consistent with Scheme 1 (4,17), and slopes of these plots were calculated by weighted linear regression analyses that assumed that has a constant percent error. Plots of these slopes versus inhibitor concentration were fitted to Equation 1 by nonlinear regression analyses (Fig.P (BioSoft) version 6.0) with slope values weighted by the reciprocal of their variance (17).
K I is the equilibrium dissociation constant for I with E, and the experimental parameter ␣ is simply the ratio of the second order rate constant with saturating I to that in the absence of I (17). If thioflavin T (I1) can form a ternary complex with a second inhibitor (I2) and AChE, the residual concentration of free enzyme [E] in the presence of both inhibitors relative to the concentration of free enzyme ([E] [I2]ϭ0 ) when only I1 is present is given by Equation 2 (18).  (17), to justify the approximation that substrate hydrolysis was second order. Fluorescence Determinations of the Binding of Thioflavin T and Propidium to AChE-The fluorescence of either thioflavin T or propidium is enhanced when these ligands bind to the AChE peripheral site. Fluorescence was monitored on a Perkin-Elmer LS-50B luminescence spectrometer in 20 mM sodium phosphate buffer (pH 7.0) and 0.02% Triton X-100 thermostated at 23 Ϯ 1°C. Thioflavin T fluorescence was measured with excitation at 450 nm and emission from 470 to 500 nm with excitation and emission slits of 10 nm. Propidium fluorescence was monitored with excitation at 500 nm and emission from 590 to 630 nm with slits of 10 nm (2,18). Total areas under the fluorescence emission curves (F) were calculated (unless otherwise noted), and fluorescence contributions from scatter in the buffer and enzyme were subtracted.
In some experiments values of F were fitted by nonlinear regression analysis ( Fig.P) to Equation 3 (18).
In Equation 3, f L is the fluorescence intensity coefficient for free ligand, f EL is the fluorescence intensity coefficient for bound ligand, E tot is the total enzyme concentration, L tot is the total ligand concentration, K D is the equilibrium dissociation constant, and D ϭ E tot ϩ L tot ϩ K D . Data were fitted by nonlinear regression analysis (Fig.P) to Equation 3, either with L tot as the independent variable and E tot fixed or with the calculated E tot as the independent variable and L tot fixed, to give K D , f EL , and in some cases f L . In these analyses, values of F were weighted by assuming constant percent error.
In other experiments, Scheme 2 was employed to evaluate interactions in ternary complexes involving thioflavin T and a second ligand in fluorescence assays. In this scheme the affinity of thioflavin T (L) at the peripheral site (denoted by subscript P) of AChE (E) is characterized by the dissociation constant K L in the absence of an acylation site ligand (I) and by K L2 when I occupies the acylation site. When I was edrophonium, the binding of I to E and EL P was assumed to reach equilibrium instantaneously with dissociation constants K I and K I2 , respectively. Binding of I to E and EL p was much slower when I was TMTFA and occurred with association rate constants k I and k I2 and dissociation rate constants k ϪI and k ϪI2 , respectively. The program SCoP (Simulation Resources, Inc., Redlands, CA; version 3.51) was applied directly to the rate equations corresponding to Scheme 2 (17). This program solves differential equations with numerical solvers and allowed fitting of F at various inhibitor concentrations and/or times. The total concentrations E tot , L tot , and I tot , as well as f L , were inserted into the program as fixed input parameters. With edrophonium, the program simultaneously fit the fluorescence areas F to the edrophonium concentrations to determine five parameters: the three equilibrium constants (K L , K I , and K L2 ), f EL , and f EIL , the fluorescence intensity coefficient for the ternary complex. With TMTFA, K I also was fixed, and the program simulta-neously fit the time courses of fluorescence emission intensity at 480 nm (F 480 ) to the time after mixing to determine K L , K L2 , f EL , and f EIL plus two additional parameters, k I and k I2 . Fig. 2. Increasing concentrations of thioflavin T increase the slopes of the reciprocal plots in Fig. 2A, and a plot of these slopes versus thioflavin T concentration shows a slight curvature (Fig. 2B), consistent with a nonzero value of ␣ in Equation 1. In the classical equilibrium analysis of Scheme 1, ␣ is less than 1 when aK S /K S2 Ͻ 1. However, we have recently shown that a nonequilibrium analysis provides a more accurate interpretation, and in this analysis ␣ is less than 1 when k S2 Ͻ k S (17). The substrate association rate constant k S2 becomes smaller than k S when the bound peripheral site ligand slows the entry of substrate to the acylation site, an effect we have termed steric blockade. The value of ␣ of 0.05 obtained for thioflavin T here is similar to those found for the prototypic peripheral site inhibitors propidium and gallamine (␣ ϭ 0.02) (17), indicating that with bound thioflavin T k S2 /k S is less than 0.05. In addition to demonstrating a substantial steric blockade by thioflavin T, the data in Fig. 2 indicate a competitive inhibition constant K I for thioflavin T of 0.90 M.

Thioflavin T Inhibits Substrate Hydrolysis by AChE in a Manner Similar to Other Peripheral Site Inhibitors-A conventional steady state analysis of thioflavin T inhibition of AChE is shown in
Thioflavin T Binds to the Peripheral Site, and Its Affinity Is Not Altered by the Binding of Edrophonium to the Acylation Site-To demonstrate that thioflavin T binds to the peripheral site, we measured the ability of two ligands to competitively inhibit thioflavin T binding in a substrate hydrolysis assay (18). Propidium is specific for the AChE peripheral site (2), whereas edrophonium is specific for the AChE acylation site (10), and both inhibited substrate hydrolysis in the absence of thioflavin T, as shown by the open circles and dashed lines in Fig. 3. If either of these ligands was completely competitive with thioflavin T such that no ternary complex could form, the decrease in would correspond to the dotted lines in Fig. 3. This is essentially the case with propidium in Fig. 3A. The affinity of thioflavin T in the ternary complex decreased by a factor of 35 relative to its affinity in the free enzyme, a large change that was not significantly different from complete competition at the ligand concentrations employed. In contrast, the affinity of thioflavin T was no different in the ternary complex with edrophonium and in the free enzyme, and the fitted line with both inhibitors was superimposed on the dashed line for edrophonium alone in Fig. 3B. These data indicate that thioflavin T binds to the AChE peripheral site in a fashion that is essentially competitive with propidium.
The Fluorescence Enhancement of Thioflavin T Bound to AChE Is Much Larger than That of Propidium Bound to AChE-The fluorescence of thioflavin T was strikingly enhanced when it bound to AChE. An excitation peak at 448 nm and an emission peak at 486 nm became apparent, spectral characteristics similar to those of thioflavin T bound to A␤ amyloid (excitation maximum at 448 nm; emission maximum at 480 nm) (also see Ref. 12). To appreciate the fluorescence enhancement of thioflavin T when bound to the peripheral site of AChE, we compared it to the fluorescence enhancement observed with propidium bound to AChE. Prior to this report, propidium has been the fluorophore with the best reported fluorescence enhancement with AChE, and it is widely used as a reporter of ligand affinity at the AChE peripheral site (2,9,18). However, when AChE was titrated with varying concentrations of the two fluorophores (Fig. 4), the fluorescence enhancement f EL /f L for thioflavin T bound to AChE was greater than 1000, whereas that for propidium was about 7 (in agreement with a value of 10 obtained for propidium with Torpedo AChE (19)). Clearly, thioflavin T will be a more sensitive and versatile fluorescent reporter of ligand interactions with AChE than propidium has been. The titration in Fig. 4A also gave a K D for thioflavin T binding of 0.89 M, in good agreement with the K I obtained in Fig. 2. Repeating this titration and analysis . Lines correspond to normalized obtained by fitting the experimental data to Equation 2. In these analyses K 1 for thioflavin T was assumed to be 1.0 M (from Fig. 1). From the fitted curves without thioflavin T (dashed lines), K 2 for propidium was determined to be 1.87 Ϯ 0.09 M, and K 2 for edrophonium was determined to be 335 Ϯ 14 nM. Ratios of K 12 /K 1 from the fitted curves with 20 M thioflavin T (solid lines) were 35 Ϯ 9 with propidium and 1.01 Ϯ 0.05 with edrophonium. The dotted lines correspond to the curves calculated when no ternary complex forms (1/K 12 ϭ 0). with a higher fixed concentration of AChE (10 M) confirmed a 1:1 stoichiometry in the AChE-thioflavin T fluorescent complex (data not shown). Titrations also were conducted with a fixed concentration of thioflavin T (0.2 M) and varying concentrations of AChE (10 nM-20 M), and analysis with Equation 3 (18) again indicated a K D of 0.8 -0.9 M (data not shown). Propidium (100 M) was strongly competitive with thioflavin T in this titration protocol, because the apparent K D for thioflavin T increased by at least a factor of 50. Fig. 3B, a ternary complex of thioflavin T and edrophonium can form with AChE without a loss in affinity of either ligand for its AChE site. We therefore asked whether the binding of an acylation site ligand altered the fluorescence of bound thioflavin T. We first examined the effect of increasing edrophonium concentrations and observed a decrease in the fluorescence as edrophonium saturated the acylation site (Fig. 5). Fitting of the data to Scheme 2 gave values of K L for thioflavin T and K I for edrophonium that were in agreement with those from Figs. 2-4 and again indicated almost no change in the affinities of these ligands for AChE in the ternary complex (K L2 /K L ϭ 1.12 Ϯ 0.02, in agreement with K 12 /K 1 ϭ 1.01 Ϯ 0.05 from Fig. 3B). However, the binding of edrophonium decreased the fluorescence of thioflavin T in the ternary complex by a factor of 2.8. The observation of this partial quenching of thioflavin T fluorescence is a major advance in the study of AChE for two reasons. First, it allows thioflavin T to be used as a reporter for ligand reactions at the acylation site. Second, it indicates that ligand binding to the acylation site alters the environment or the configuration of thioflavin T bound to the peripheral site, a point that we return to below.

The Binding of Acylation Site Ligands to the AChE-Thioflavin T Complex Decreases the Fluorescence Enhancement of Bound Thioflavin T-According to
To illustrate the first point we examined TMTFA, which forms a hemiketal with Ser-200 in the acylation site that is an analog of the transition state formed by acetylcholine (11,20,21). The high affinity of TMTFA in this complex allows the time course of its binding to be monitored over a wide range of TMTFA concentrations. Fluorescence traces for nine combinations of TMTFA and thioflavin T concentrations were recorded, and three representative reactions are shown in Fig. 6. The nine reaction time courses were fitted simultaneously to Scheme 2 with the SCoP program. Over this range of association rates the TMTFA concentrations were too high to evaluate K I for TMTFA; so K I was fixed at 49 pM as determined previously (17), and six remaining parameters were determined. A wealth of information was obtained. First, a K L of 1.0 M for thioflavin T and an enhancement in fluorescence of bound thioflavin T (f EL /f L ) of 1600 were obtained, consistent with data in Figs. 4 and 5. Second, the affinities of thioflavin T and TMTFA in the ternary complex with AChE were at least as high as the affinities in their respective binary complexes (K L2 /K L ϭ K I2 /K I ϭ 0.56), in agreement with the previous observations for edrophonium in Figs. 3 and 5. Third, steric blockade of TMTFA binding by bound thioflavin T was evident (k I /k I2 ϭ 60), although the magnitude of the blockade was somewhat smaller than the nearly 400-fold decrease in association rate constant for TMTFA when propidium was bound to the peripheral site (17). Fourth, the binding of TMTFA decreased the fluorescence of thioflavin T in the ternary complex by a factor of 4.2. The extent of this quenching appears to be somewhat larger than that observed with edrophonium in Fig.  5, suggesting that the quenching of bound thioflavin T in ternary complexes with AChE may depend on the structure of the acylation site ligand. A quenching smaller than that obtained with either edrophonium or TMTFA was observed when Ser-200 was acylated with a methylethoxyphosphonyl group (data not shown), indicating that the acylation site ligand need not be cationic to quench the fluorescence of bound thioflavin T. DISCUSSION The discovery that the fluorescence of the AChE-thioflavin T complex is partially quenched when a ligand binds to the acylation site is of both practical and conceptual importance. In practical terms, this fluorescence change can be monitored to report on reactions at the acylation site. We illustrate this with TMTFA in Fig. 6, but similar protocols should allow the reactions of acylating agents like organophosphates to be measured over time courses as short as milliseconds. Conceptually, the partial quenching on the binding of acylation site ligands is the strongest evidence to date of conformational interaction between the peripheral and acylation sites of AChE. Although such conformational interaction has been widely invoked to account for a range of experimental data, close analysis indicates that previous evidence for such interactions is inconclusive.
A frequent assertion is that ligand binding to the peripheral site alters the conformation of the acylation site to reduce the efficiency of acylation or deacylation by substrates. This explanation has been offered to account for inhibition of substrate hydrolysis by peripheral site ligands (9) and, when the substrate itself can bind to the peripheral site, for substrate inhibition at high substrate concentrations (22). As noted by Taylor and Radic (23), ligand association with the peripheral site also may prevent access of a cationic substrate to the acylation site by physically obstructing substrate entry or by charge repulsion between ligand and substrate. Before an allosteric interaction can be invoked, contributions from these other factors must be eliminated. Over the past 3 years, we have demonstrated that bound peripheral site ligands do physically obstruct the entry of substrates as well as of other ligands into the acylation site. Bound propidium decreases the association and dissociation rate constants for the acylation site ligands huperzine A and TMTFA by factors of 10 -400. We have termed this effect steric blockade (17) and shown that similar decreases in the association rate constants for substrates and the dissociation rate constants for the alcohol products of substrate hydrolysis (e.g. thiocholine) are sufficient to account for the inhibition of substrate hydrolysis by peripheral site ligands (4,17) as well as for substrate inhibition (4). Steric blockade in our terminology is strictly a kinetic effect, and it can have no influence on substrate affinity or hydrolysis by slowly reacting substrates like organophosphates that equilibrate with AChE before acylation. However, 5-to 10-fold decreases in organophosphate affinity for propidium-bound AChE were observed with both a cationic and a neutral organophosphate, indicating an additional type of interaction (6,13). Molecular modeling calculations indicate that an organophosphate with a large leaving group would overlap with propidium if both ligands were placed in their expected sites in free AChE (13), and we have attributed these decreases in affinity to adjustments that accommodate both ligands in the ternary complex (6,13). Modeling calculations indicate that such steric overlap is not a factor in ternary complexes involving propidium-AChE and either huperzine A or TMTFA; yet ligand affinities in the ternary complexes also are 5-to 7-fold lower than in the corresponding binary complexes (17). We have suggested that the decreased affinities arise from electrostatic interactions between these cationic ligands (17). In summary, experimental data indicate that the binding of a peripheral site ligand to AChE can result in steric blockade, steric overlap, and/or electrostatic interaction that inhibits substrate hydrolysis or reduces ligand affinity at the acylation site. With small peripheral site ligands like propidium and gallamine, we see no evidence of an additional allosteric conformational effect that contributes to inhibition.
More compelling indications of conformational interaction between the peripheral and acylation sites of AChE have come from two other experimental approaches. First, the spectral properties of a pyrenebutyl methylphosphono group attached to Ser-200 in the acylation site are altered slightly by the binding of gallamine to the peripheral site (24). The authors attributed these alterations to an increase in polarizability of the pyrenebutyl environment, perhaps due to torsional movements of aromatic side chains in the vicinity of the pyrenebutyl moiety induced by the binding of the peripheral site ligand. Second, bound peripheral site ligands can accelerate the reaction of acylating agents with Ser-200 in the acylation site. Gallamine accelerated the acylation of AChE by dimethylcarbamyl fluoride (25), and peripheral site inhibitors including propidium and gallamine increased the first order phosphorylation rate constants for neutral organophosphates with AChE (6,13,26). This rate constant reflects the transfer of the organophosphate group to Ser-200 in the binary or ternary complexes (equivalent to k 2 and ak 2 , respectively, in Scheme 1). Before these effects can be attributed to a conformational change in the acylation site induced by ligand binding to the peripheral site, however, other explanations again must be eliminated. Foremost among alternative explanations is steric overlap as denoted above. One of the neutral organophosphates with accelerated first order phosphorylation when propidium is bound was also shown by molecular modeling to bind with AChE in a way that partially overlapped with the propidium binding site (13). We have argued that movement of the organophosphate to eliminate this overlap in the ternary complex could potentially induce strain in the bond to the organophosphate leaving group and thereby increase the phosphorylation rate constant (6). In like fashion, preliminary molecular modeling suggests that the pyrenebutyl methylphosphono group attached to Ser-200 is bulky enough to overlap partially with the peripheral site. Although the affinities of gallamine and propidium were similar for this modified AChE and for free AChE, a large thermodynamic interaction between the ligands is not necessary for there to be spectral changes due to the ligand proximity. Steric overlap thus remains a concern when conformational interactions between the peripheral and acylation sites are proposed. However, several pairs of peripheral site ligands and organophosphates (some with quite small leaving groups) showed the acceleration phenomenon (26). Because steric overlap is unlikely to generate the acceleration in all of these cases, it appears reasonable to infer a conformational change that promotes phosphorylation on ligand binding to the peripheral site.
Our data here are the first to indicate that ligand binding to the acylation site can alter the conformation of the peripheral site. The partial quenching of the fluorescence of bound thioflavin T when a ligand binds to the acylation site cannot result from fluorescence resonance energy transfer, because there is no spectral overlap between thioflavin T and either of the acylation site ligands edrophonium or TMTFA. However, the fluorescence of thioflavin T has been shown to depend on solvent viscosity (27), with a relationship previously described for a class of fluorescent dyes called molecular rotors (28,29). These dyes show increased fluorescence when introduced into high viscosity media, due to a decreased torsional relaxation. Binding to the AChE peripheral site could reduce the torsional mobility of thioflavin T and account for its enhanced fluorescence. It is less clear how the binding of an acylation site ligand could partially quench this fluorescence in the ternary complex, but the quenching would appear to require an increase in the torsional mobility of bound thioflavin T induced by a change in the local AChE conformation that is initiated in the acylation site. The exact location of thioflavin T when bound to the AChE peripheral site remains to be determined, but it is unlikely that simple proximity to the acylation site ligand could account for the quenching. Crystal structures of bound edrophonium and bound TMTFA show that both ligands bind near the base of the acylation site, well away from the region that defines the peripheral site (10,11). Furthermore, these ligands show no decrease in affinity in their ternary complexes with propidium-AChE, indicating no steric overlap. Ideally, one would like to support this proposed conformational interaction with additional data, but the conformational changes involved appear to be very small. Crystal structures of AChE and the TMTFA-AChE complex show a root mean square deviation of only 0.4 Å between equivalent C ␣ positions and no changes in the structure of the peripheral site (11). Electron paramagnetic resonance of AChE labeled with a spin-labeled organophosphate at Ser-200 gave no change in signal when propidium was bound to the peripheral site, although a small change was observed with monoclonal antibodies or the neurotoxin fasciculin, both of which bind to a larger surface area extending beyond the pe-ripheral site (30). Therefore, we conclude that the quenching of the fluorescence of AChE-bound thioflavin T by acylation site ligands is the only method to date that is sensitive enough to reveal a conformational interaction between the two sites. This interaction may provide a mechanism for bound peripheral site ligands to accelerate acylation site reactions.