Indoleamine Analogs as Probes of the Substrate Selectivity and Catalytic Mechanism of Serotonin N-Acetyltransferase*

Serotonin N-acetyltransferase (arylalkylamine N-ace-tyltransferase (AANAT)) catalyzes the reaction of serotonin (or tryptamine) with acetyl-CoA to formN-acetylserotonin (or N-acetyltryptamine) and is responsible for the melatonin circadian rhythm in vertebrates. This study evaluates a series of indoleamine analogs as alternate substrates of AANAT. 3-Indolepropylamine and 3-indolebutylamine were chemically synthesized and found to be processed by AANAT, although 20- and 60-fold less efficiently compared with the natural substrate serotonin, respectively. Racemic α-methyltryptamine andN ω-methyltryptamine were also shown to be substrates for AANAT, again with reducedk cat andk cat/K m compared with serotonin. The enzyme did exhibit ∼9:1 stereoselectivity for theR-enantiomer of α-methyltryptamine versus theS-enantiomer. By measuring the enzymatic ratesversus increasing buffer microviscosity, it was demonstrated that diffusional release of product is most likely the principal rate-determining step for the enzymatic transformation of tryptamine (which has similar k cat andk cat/K m compared with serotonin). Analysis of k cat andk cat/K m versuspH for the poor substrateN ω-methyltryptamine showed that an ionizable group on the enzyme with pK a ∼ 7, required to be in its deprotonated form, may be important in catalysis. The α-methyltryptamine analog α-trifluoromethyltryptamine was not processed by the enzyme, but served as a modest competitive inhibitor. Taken together with the pH-rate analysis, these results favor a model in which the serotonin substrate binds to the enzyme as the positively charged ammonium salt, and nucleophilicity of the amine is important in enzyme-catalyzed acetyl transfer.

Serotonin N-acetyltransferase (arylalkylamine N-acetyltransferase (AANAT) 1 ) is the enzyme responsible for the circadian rhythmic cycle of melatonin (5-methoxy-N-acetyltryptamine) (1). AANAT catalyzes the conversion of serotonin to N-acetylserotonin, the direct precursor to melatonin (Fig. 1). AANAT is expressed primarily in the pineal glands of vertebrates, where acetyltransferase activity rises in the night and falls in the daytime, regulating the production of melatonin. The recent cloning of AANAT (1) has facilitated advances in the understanding of its regulation (2) and enzyme mechanism (3). In principle, a detailed understanding of the substrate selectivity and catalytic mechanism of AANAT could lead to the development of specific AANAT inhibitors that might have therapeutic value in sleep and mood disorders.
AANAT is a 24-kDa protein that belongs to the motif A/B superfamily, which includes Ͼ150 proteins, many of which are believed to be acetyltransferase enzymes (1,3). This superfamily is defined by the presence of two weakly conserved tandem sequences of ϳ15 amino acids called motifs A and B. AANAT is itself highly conserved among vertebrates, displaying ϳ80% amino acid identity among sheep, chicken, rat, and human AANATs (1). Recently, recombinant sheep AANAT was overproduced and purified from Escherichia coli (3). Its kinetic mechanism was shown to be sequential (ternary complex) Ordered Bi Bi, with acetyl-CoA binding before serotonin (or the alternative substrate tryptamine) (3).
Previous studies of the substrate specificity of AANAT have revealed that the indole function is particularly important for efficient turnover (4). The design of specific AANAT inhibitors will be augmented by a thorough analysis of the tolerance of the AANAT active site for alternate substrates. The processing of alternative substrates can also yield insights into the catalytic and kinetic mechanisms of the enzyme. As part of a previous investigation of AANAT, a simple kinetic acetyltransferase assay was reported in which free CoASH was indirectly monitored by reaction with 5,5Ј-dithiobis(2-nitrobenzoic acid) (3). In principle, this assay allows a variety of serotonin analogs to be tested as AANAT substrates in a convenient way. In this study, we report the results of the effects of substrate alterations on AANAT processing.
Instruments-1 H, 13 C, and 19 F NMR experiments were performed on a Bruker 400-MHz spectrometer. Optical rotations were measured using a Jasco DIP-370 digital polarimeter.
3-Indolepropionoyl-O-benzylhydroxamate (5)-To a solution of 3-indolepropionic acid (3; 2.0 g, 10.6 mmol) in dimethylformamide (100 ml) was added Et 3 N (4.4 ml, 31.8 mmol) and O-benzylhydroxylamine hydrochloride (3.4 g, 21.1 mmol). 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (4.0 g, 21.1 mmol) was added to the solution at 0°C, and the resulting mixture was allowed to warm up to room temperature. After 48 h of stirring under an atmosphere of N 2 , undissolved material was removed from the reaction mixture by filtration, and the filtrate was concentrated in vacuo to remove dimethylform-amide. The resulting residue was partitioned between 0.5 N HCl (100 ml) and EtOAc (200 ml). The organic layer was washed with 100 ml of 1 N NaHCO 3 and brine; dried over MgSO 4 ; and concentrated in vacuo to give fairly pure product as a white solid (1.8 g, 58%), which was recrystallized from hot MeOH to give 1.2 g (first crop) of pure compound 5. TLC R F ϭ 0.24 (EtOAc/hexanes, 1:1). 1  , and the resulting mixture was refluxed under N 2 . After 3 days of refluxing, excess LiAlH 4 was quenched by slow addition of 0.5 N HCl (5 ml). The mixture was filtered, and the filtrate was diluted with EtOAc (50 ml) and 0.5 N HCl (25 ml). The aqueous layer was basified with 1 N NaOH to pH Ͼ10 and then extracted with EtOAc (2 ϫ 50 ml). The organic layers were combined, dried over MgSO 4 , and evaporated in vacuo to give 38 mg of compound 1 (yield based on recovered compound 5, 90%), which appeared as a single spot on TLC. R F ϭ 0.25 (CH 2 Cl 2 /MeOH/NH 4 OH, 100:10:2). For analytical purposes and for use in the enzyme assay, the material was further purified using preparative TLC to give compound 1 as a white solid, which was quantitatively converted to its hydrochloride salt and stored at Ϫ20°C until needed. 1  3-Indolebutylamine (2) (5)-Following the procedure described above for the preparation of compound 1, 3-indolebutyryl-O-benzylhydroxamate (6) (160 mg, 0.54 mmol) was reacted with LiAlH 4 (7.4 ml, 1.0 M in tetrahydrofuran) to give 7 mg of compound 2 (yield based on recovered compound 6, 80%). TLC R F ϭ 0.21 (CH 2 Cl 2 /MeOH/NH 4 OH, 100:10:2). 1  (Ϯ)-␣-Trifluoromethyltryptamine (9) (6)-Compound 9 was synthesized using the vinylnitroso cycloaddition strategy of Zimmer and Reissig (6) and gave 1 H NMR and mass spectra consistent with those reported (6). Measurement of the pK a for compound 9 was carried out by pH titration of the 1 H NMR spectrum of compound 9 with 3 mM compound 9 in 20% D 3 COD/D 2 O with 200 mM NaCl at 25°C. The pH was adjusted with NaOD and DCl, and the chemical shifts of the methinyl proton (␣ to the amine function and the trifluoromethyl group) were standardized versus trimethylsilylpropionate. The data were fit to a modified Hill equation (7) that fit smoothly to a one-proton transfer equilibrium with a calculated pK a of 5.4 Ϯ 0.1.
(Ϯ)-N-Acetyl-␣-methyltryptamine-A solution of ␣-methyltryptamine (53 mg, 0.3 mmol) in CH 2 Cl 2 (10 ml) was treated with acetic anhydride (30 l, 0.32 mmol) and Et 3 N (37 l, 0.5 mmol). The reaction flask was wrapped with aluminum foil, and the mixture was stirred at room temperature under N 2 . Reaction progress was monitored by TLC, and the reaction was essentially complete after 1 h. The reaction mixture was diluted with CH 2 Cl 2 to ϳ20 ml and successively washed with 0.5 N HCl (20 ml), 1 N NaHCO 3 (15 ml), H 2 O (15 ml), and brine. The organic layer was dried (MgSO 4 ) and then evaporated in vacuo to give the product as a clear oil (70 mg, 108% mass balance). 1 H and 13 C NMR analyses showed that excess mass was due to a minor contaminant of acetic anhydride. A sample for analysis was purified using preparative TLC. TLC R F ϭ 0.49 (CH 2 Cl 2 /MeOH, 10:1). 1  N-Acetyl-N -methyltryptamine-Following the procedure described above for the preparation of N-acetyl-␣-methyltryptamine, N -methyltryptamine (78 mg, 0.45 mmol) in CH 2 Cl 2 (10 ml) and Et 3 N (125 l, 0.9 mmol) was acetylated with acetic anhydride (85 l, 0.9 mmol) to give the N-acetyl-N -methyltryptamine product as a white solid (100 mg, 100%). cis-and trans-rotamers were observed in 1 H and 13 C NMR spectra about the N-CϭO bond at 298 K at a ratio of ϳ55:45. 1  Acetyltransferase Kinetic Assays of Amine Analogs-Preparation of purified recombinant sheep glutathione S-transferase-AANAT has been described, and it has shown to be nearly identical in its kinetic behavior to purified glutathione S-transferase-free AANAT (3). Assays of the amine analog substrates were performed essentially as previously reported using a spectrophotometric assay in which CoASH levels were monitored indirectly by reaction with 5,5Ј-dithiobis(2-nitrobenzoic acid) (3). To verify that N -methyltryptamine was forming the acetylated product, a large-scale reaction (20 mM, 3 ml) was performed, and N -acetyl-N -methyltryptamine was isolated and shown to have identical 1 H NMR and TLC properties compared with the authentic material. For K m measurements, at least five amine substrate concentrations ranging from Ͻ0.3 K m to Ͼ3 K m were employed in the presence of near-saturating concentrations of acetyl-CoA (Ͼ3 K m ). Rates were measured under initial conditions (Ͻ10% turnover of the limiting substrate). All assays were performed in duplicate, and rate values generally agreed within 10%. Kinetic constants were calculated by fitting the data to the Michaelis-Menten equation using a nonlinear curve fit (Kaleidograph), and the values are shown as means Ϯ S.E. in Table I.
␣-Trifluoromethyltryptamine (9) Inhibition Experiment-After showing that ␣-trifluoromethyltryptamine was not processed by AANAT (k cat /K m at least 1000-fold lower than for compound 7), it was tested as a competitive inhibitor using a Dixon analysis versus the substrate N -methyltryptamine. A range of five concentrations of compound 9 (up to 10 mM) was used in the inhibition experiment. The assay conditions were similar to those described previously (3) with minor modifications. MeOH (final concentration of 5%) was included in the assay to improve the solubility of compound 9. Controls demonstrated that this concentration of MeOH had no detectable effect on turnover rate. In addition, 0.1 M Mes (pH 6.1) with 0.1 M NaCl was used in place of the standard phosphate buffer (3). Two separate assays were performed at two fixed subsaturating concentrations (4 and 6 mM) of N -methyltryptamine (K m ϭ 4 mM). The concentration of acetyl-CoA (K m ϭ 0.4 mM) was kept constant and near-saturating (2 mM). Dixon plots were fit using linear regression. The K i was calculated assuming a competitive inhibition model (versus N -methyltryptamine) from the slope of the line using the steady-state kinetic equation (Equation 1).

FIG. 1. Serotonin N-acetyltransferase-catalyzed conversion of serotonin to N-acetylserotonin.
The calculated K i values from both experiments were in good agreement, and the values are shown as means Ϯ S.E. in Table I.
Stereochemical Studies-␣-Methyltryptamine (20 mM) and acetyl-CoA (20 mM) were reacted with AANAT (10 M) in a 3-ml reaction volume using previously described acetyltransferase conditions (pH 6.8 and sodium phosphate buffer) (3), and reaction progress was monitored with 5,5Ј-dithiobis(2-nitrobenzoic acid). After 20 min, when the reaction was shown to be ϳ50% complete, the reaction mixture was vortexed vigorously for 2 min. A white precipitate formed, which was removed by centrifugation (3000 ϫ g, swinging bucket). The resulting solution was diluted with CH 2 Cl 2 (10 ml) and H 2 O (10 ml). The aqueous layer (pH ϳ7) was separated and extracted with additional CH 2 Cl 2 (3 ϫ 10 ml). The combined organic layers were dried (MgSO 4 ) and concentrated in vacuo to give 6.8 mg (54%) of N-acetyl-␣-methyltryptamine. The neutral aqueous layer was basified to pH ϳ11 using 1 N NaOH and then extracted with CH 2 Cl 2 (3 ϫ 10 ml). The combined extracts were dried (MgSO 4 ) and concentrated in vacuo to give 5.5 mg (44%) of recovered ␣-methyltryptamine. The identity and purity of both materials (Ͼ95%) were verified by 1 H NMR. The concentrations of the solutions used for measuring optical rotations were verified independently by measuring UV absorption at ϭ 279 nm. The specific rotations are shown under "Results" and gave similar results on two separate occasions. To rule out the possibility that transfer of the acetyl group had occurred non-enzymatically at these concentrations of ␣-methyltryptamine and acetyl-CoA, the assay mixture, in the absence of AANAT, was stored at 30°C for 2 h. No acetyl transfer could be detected (Ͻ5%) under these conditions. Viscosity Studies-All assays were performed with the added viscogen sucrose to achieve the desired viscosity. Relative viscosities were previously measured with a Canon-Fenske viscometer at 30°C as described previously (8). Sucrose solutions used to achieve the viscosities in this work were 1.0 (0% (w/v) sucrose), 1.55 (16%), 1.99 (23%), 2.50 (30%), and 3.10 (35%). Measurements of k cat and K m were performed as described previously (3), except that Mops at pH 7.4 and 100 mM NaCl were employed. Under these conditions in the presence of saturating concentrations of N -methyltryptamine or tryptamine, the K m for acetyl-CoA was found to be 0.3 mM and did not significantly change with added sucrose. The K m for tryptamine at near-saturating concentrations of acetyl-CoA in the absence of sucrose was found to be 0.12 mM and was Ͻ0.1 mM with added sucrose. The K m for N -methyltryptamine at near-saturating concentrations of acetyl-CoA in the absence of sucrose was found to be 4 mM and fell with increasing sucrose concentrations as described under "Results." The k cat values under these assay conditions without added sucrose were 27 and 2.5 s Ϫ1 for tryptamine and N -methyltryptamine, respectively.
Rate Versus pH Profile-Using N -methyltryptamine as substrate, the measurements for k cat and K m were performed in the pH range 6 -8.6 following the general spectrophotometric assay procedures previously described (3). Accurate rate measurements were not possible above pH 8.6 because of high background aminolysis/hydrolysis rates of acetyl-CoA. All reactions were initiated with enzyme (which was maintained at pH 6.8, but did not significantly affect the final pH), and the activities proved to be linear for at least 3 min, suggesting that altered kinetics were not caused by protein instability under the assay conditions. Enzyme assays showed Michaelis-Menten kinetic behavior at each pH investigated. All assays were performed with near-saturating concentrations of acetyl-CoA (Ͼ3 K m , verified at each pH) throughout the pH range. The following buffers were used to obtain the pH values described in the experiments: pH 6.1-6.7 (Mes), pH 6.7-7.6 (Mops), and pH 7.7-8.6 (Epps). Fits for the k cat data were performed using Equation 2, where H is the proton concentration and K 1 is the dissociation constant for the ionizable group that facilitates the reaction in its deprotonated form. Fits for the k cat /K m data were performed using Equation 3, where H is the proton concentration, K 1 is the dissociation constant for the ionizable group that facilitates the reaction in its deprotonated form, and K 2 is the dissociation constant for the ionizable group that facilitates the reaction in its protonated form. Both equations were fit using the computer program Kinetasyst II (Intellikinetics), and the kinetic constants Ϯ S.E. are shown under "Results."

RESULTS
Homologated Tryptamines-The known substrates serotonin and tryptamine contain an ethyl side chain extending from the indole 3-position. The initial objective of these studies was to determine the tolerance of AANAT for substrate alkyl chain length variations. We thus set out to prepare the alkyl chainextended propylamine (1) and butylamine (2) analogs of tryptamine (5). Synthesis of these compounds was achieved in two steps, starting from the corresponding commercially available carboxylic acids 3 and 4 (Fig. 2). The carboxylic acids were converted to the hydroxamic acid derivatives 5 and 6 and then exhaustively reduced with LiAlH 4 to afford the desired amines 1 and 2.
Kinetic analysis of these compounds was carried out with recombinant sheep AANAT in the presence of saturating levels of acetyl-CoA and with an indirect spectrophotometric assay that monitors CoASH formation. Both of the chain-extended indoleamine substrates 1 and 2 were found to be AANAT substrates and displayed normal Michaelis-Menten kinetic behavior, although with somewhat reduced catalytic efficiency compared with serotonin (Table I). Propylamine analog 1 was 20fold reduced in catalytic efficiency, exhibiting both a higher K m and lower k cat compared with serotonin. Butylamine analog 2 was only 3-fold less active as a substrate compared with propylamine analog 1, and the difference was primarily in K m . These results show that AANAT has an optimal preference for indole substrates with 3-ethylamine side chains, corresponding to the natural substrate, but is reasonably tolerant of side chain extension by at least two more methylene groups from the indole ring.
Methyl-substituted Tryptamine Analogs-Another important issue related to AANAT selectivity is the ability to accommodate ␣-branching or nitrogen substitution (secondary amines). Tryptophan (both D-and L-isomers) was shown not to be processed by AANAT. The commercially available racemic ␣-methyltryptamine (7) and N -methyltryptamine (8) were next evaluated as AANAT substrates (Fig. 3). Both of these analogs were utilized as substrates by AANAT and displayed Michaelis-Menten behavior, although again with diminished catalytic efficiency compared with tryptamine (Table I). In addition to monitoring amine-induced CoASH generation in the presence of AANAT, direct characterization of the acetamide product derivatives of compounds 7 and 8 was accomplished using TLC and 1 H NMR, ruling out indoleamine-promoted acetyl-CoA hydrolysis. Racemic ␣-methyltryptamine (7)  for compound 8 with AANAT compared with the standard substrate tryptamine was ϳ300-fold reduced, it was still much greater (Ͼ300-fold) than the non-enzyme-catalyzed background acetylation rate. That ␣-methyltryptamine and Nmethyltryptamine were AANAT substrates allowed further investigation into active-site stereochemical accessibility and enzymatic rate-determining step(s).
Stereochemical Studies-Most enzyme reactions demonstrate stereoselectivity because of the well defined structure of enzyme active sites. Although racemic ␣-methyltryptamine (7) was shown to be a substrate for AANAT, the spectrophotometric assay described above was insufficient to reveal potential stereochemical selectivity. A larger scale reaction of AANAT with racemic ␣-methyltryptamine (7) was allowed to proceed to ϳ50% completion over 20 min, and the product and the unreacted starting material were isolated in pure form. The specific rotation ([␣] D 25 ) of the recovered ␣-methyltryptamine from the enzyme reaction was found to be ϩ26°Ϯ 3°(ϩ32°ϭ 100% S-enantiomer (9)), indicating that it was a 9:1 mixture of S/Risomers. Furthermore, the specific rotation of the enzymatic product N-acetyltryptamine (ϩ9°) confirmed it to be principally the R-enantiomer by demonstrating its similar magnitude to and opposite sign from (Ϫ8.4°) the chemical acetylation product of the enantiomerically enriched (S)-␣-methyltryptamine (i.e. the recovered starting material from the AANAT reaction). These results establish unequivocally that AANAT shows stereoselectivity for ␣-substituted tryptamine analogs and, in the case of ␣-methyl substitution, that the R-enantiomer is processed more efficiently than the S-enantiomer.
Viscosity Studies-Previous studies have shown that for many enzyme reactions, the rate-determining step can be diffusional release of product rather than the chemical transformation step (10). A particularly effective method for elucidating the nature of the rate-determining step(s) is to examine the effect of increasing viscosity on the kinetics of the reaction (10). By increasing the viscosity of the solution with a microviscogen such as sucrose, rates of diffusional steps are slowed, whereas unimolecular processes are theoretically unaffected. However, large quantities of compounds such as sucrose can cause conformational or other "artifactual" effects on enzyme reactions by interacting with the enzyme, substrate(s), or both. To control for these non-diffusional effects, it is helpful to have "poor substrates" for which the rate-determining step can be reason-ably assumed to be chemical rather than diffusional (11). The identification of N -methyltryptamine (8) as an AANAT substrate with a substantially reduced k cat (ϳ10-fold lower) compared with tryptamine appeared to fulfill the requirement of a poor substrate.
The viscosity of the reaction buffer was manipulated by the addition of sucrose, and the effects on k cat and k cat /K m with AANAT were examined with N -methyltryptamine in the presence of near-saturating acetyl-CoA. The slope of the plot of k cat -control/k cat -viscogen versus relative viscosity can give a quantitative estimate of the role of product diffusional release in the kinetic scheme. A slope of 1 indicates a maximal effect (product release is completely rate-determining), and a slope of 0 indicates no effect (product release is much faster than the chemical step). In the experiment, the slope was found to be ϩ0.16 Ϯ 0.03 with N -methyltryptamine as substrate and ϩ0.75 Ϯ 0.08 with tryptamine as substrate (Fig. 4).
The slope of a plot of (k cat /K m -control)/(k cat /K m -viscogen) can provide information about the "commitment to catalysis" of the substrate as well as information about artifactual effects. High commitment to catalysis indicates that the substrate off-rate is much slower than the (forward) chemical transformation step (10). In the extreme case, in a plot of (k cat /K m -control)/(k cat /K mviscogen) versus the relative viscosity ratio, the slope should approach 1. When the substrate off-rate is much faster than the chemical (forward) step, the slope should approach 0. In this case, the experimental slope with N -methyltryptamine as substrate was measured to be Ϫ0.14 Ϯ 0.03 (data not shown). With tryptamine as substrate, the K m values were too small to be accurately measured (Ͻ100 M) at high sucrose. The theoretical slope with the poor substrate N -methyltryptamine with a high K m (Ͼ1 mM) and a slow rate would be 0, and the small slope observed may indicate a small, but relatively insignificant non-diffusional effect of sucrose on the enzyme reaction. Overall, the viscosity effects suggest that for the Nmethyltryptamine reaction, the chemical step is largely ratedetermining, whereas for the tryptamine reaction, product diffusional release is the primary rate-determining step.
Rate Versus pH Profiles-Investigations of rate versus pH can sometimes provide information about the properties of ionizable groups that participate in enzymatic catalysis. For example, one way AANAT might accelerate the acetyltransferase reaction is to bind the neutral amine substrate in preference to the protonated ammonium form. This would effectively lead to the experimental observation that the pK a for the amine substrate would be depressed in the ternary complex.  Analyses of k cat versus pH generally reveal pK a values of ionizable groups in the bound enzyme-substrate complex, whereas plots of log(k cat /K m ) versus pH should reflect pK a values of ionizable groups present in the free enzyme and/or substrate (12). Measurements of rate versus pH were carried out with the poor substrate N -methyltryptamine to maximize the possibility of identifying groups that play a role in the chemical step. The results shown in Fig. 5 indicate that there is an ionizable group with pK a ϳ 7 (pK 1 ) present in both the free and bound complexes that is required to be in its deprotonated form. This group (with pK 1 ) facilitates acetyl transfer in its deprotonated form. This is very likely to be an enzymatic group since there are no groups on compound 8 in free solution with such a pK a (the secondary amine of compound 8 should have a pK a of ϳ10 in free solution). In addition, in the plot of log(k cat /K m ) versus pH, there was evidence of a second group (pK 2 ) with pK a ϭ 8.5. However, the value of pK 2 is near the upper limit of the pH range where it was technically feasible to perform kinetic studies, somewhat limiting its reliability.
␣-Trifluoromethyltryptamine-Information about electronic requirements of the tryptamine substrate is important for understanding catalysis and potentially of great use for inhibitor design. Lowering the amine pK a by electronegative atom substitution in the substrate would result in a larger concentration of the neutral amine at pH 7. Depending on the mechanism of AANAT, this could lead to a substrate with a reduced K m , or it could lead to an inhibitor since a neutral amine with a reduced pK a is generally less nucleophilic than a neutral amine with a higher pK a . Since the atomic radius of fluorine is only 0.2 Å greater than that of hydrogen, fluorine is close to isosteric with hydrogen, but can significantly change electronic properties because of its strong electronegativity. The use of fluorine substitution in substrate analogs has been a powerful way to probe enzyme mechanism and to generate inhibitors (13). ␣-Trifluoromethyltryptamine (9) was synthesized as shown in Fig. 6.
The synthesis of ␣-trifluoromethyltryptamine was achieved following the method of Zimmer and Reissig (6). Commercially available 3-bromo-1,1,1-trifluoroacetone was converted to the corresponding oxime 10 by reaction with hydroxylamine hydrochloride. The oxime was used to generate the vinylnitroso derivative, which was used in situ to react in a 4 ϩ 2 cycloaddition and to produce the 3-substituted indole 11. Reduction of this material with LiAlH 4 led to generation of racemic ␣-trifluoromethyltryptamine (9). The ␣-amine pK a for compound 9 was measured to be 5.4 Ϯ 0.1, ϳ4.4 units lower than that of a corresponding unsubstituted amine (e.g. serotonin with pK a ϭ 9.8 (14)).
In contrast to the isosteric ␣-methyltryptamine, there was no detectable AANAT-catalyzed acetyl transfer to ␣-trifluoromethyltryptamine (at least 1000-fold lower than the k cat /K m for ␣-methyltryptamine). It was thus tested as a potential competitive inhibitor and was found to be a modest inhibitor with K i ϭ 3 mM, a value similar to the K m for ␣-methyltryptamine (Table  I). This suggests that ␣-trifluoromethyltryptamine can bind about as well as ␣-methyltryptamine to AANAT, but may lack the nucleophilicity to undergo reaction.

DISCUSSION
Homologated Tryptamines-Analysis of the effects on catalytic processing of alkyl chain extension of the ethylamine function of tryptamine contributes to an understanding of the geometric constraints enforced by the AANAT active site. Although a 20-fold decreased catalytic efficiency with the propylamine versus the ethylamine function may seem considerable, it is much less than is observed in perturbing many other enzyme-substrate interactions by the addition of an extra methylene to a functional group undergoing reaction. For example, methylene insertion achieved by an aspartate-to-glutamate substitution in a protein-tyrosine kinase was shown to reduce catalytic efficiency by 10,000-fold (15). Moreover, further extension of the alkyl side chain of the propylamine analog by an extra methylene as present in the butylamine analog (2) resulted in a lowering of the k cat /K m by only 3-fold more. It can reasonably be speculated that relatively weak hydrophobic interactions are involved in holding the tryptamine substrate in place for nucleophilic attack.
The lack of strict side chain recognition allows a more flexible approach to inhibitor design. For example, substitution of the ␤-position of tryptamine with a halide would be very unstable because it would be in the "benzylic" position (16). The corresponding halide substitution of the ␤-position in compound 1 would not be complicated by the same problem. The scaffolds of compounds 1 and 2 thus show promise for the development of mechanistic probes and potent inhibitors of AANAT.
Methyltryptamine Analogs-Both ␣-branching and N-methyl substitution create steric bulk around the nucleophilic amine without significantly affecting basicity. The methyl substitutions present in compounds 3 and 4 reduce the catalytic efficiency of AANAT processing, but still allow for acetyltransferase reaction. The R-enantiomer of compound 3 is much better tolerated than the S-enantiomer for enzyme processing. cleophilicity in the enzyme reaction. Assume that the ␤ nuc for the chemical step is 0.8, similar to related chemical model reactions (21). The difference in pK a values of the amines of the trifluoromethyl-substituted compound 9 compared with ␣-methyltryptamine (7) is 4.4 units, so the predicted rate reduction due to decreased basicity of compound 9 would be 10 3.5 -fold. Disruption in binding and/or orientation due to fluorines could further reduce the rate. Even a 10 3.5 -fold reduction is below the background experimental detection level (10 3 -fold lower than for compound 7) and is consistent with the experimental inability to detect turnover with substrate 9, assuming a transition state with ␤ nuc ϭ 0.8. Arylamines such as aniline (pK a ϭ 4.7) have been reported not to be substrates for AANAT, whereas the related compound phenethylamine (pK a ϭ 9.8) is an effective substrate (1). Based on the arguments above, the decreased nucleophilicity of arylamines would be insufficient to permit efficient AANAT-catalyzed acetyl transfer.
The pH-rate analyses did reveal an ionizable group with pK a ϳ 7 (pK 1 ) for both the free and bound enzyme complexes. 2 Although the role and identity of this residue are not known at present, a reasonable speculation is that it might be an enzyme histidine imidazole. Such a pK a would be consistent with a previously proposed histidine in catalysis whose role was suggested based on mutagenesis studies (1). Further analysis will be necessary to clarify this connection.
Conclusions-Recombinant sheep AANAT catalyzes acetyl transfer to a diverse array of tryptamine analogs. This relatively liberal acceptance of amine substrates offers hope for the generation of potent AANAT inhibitors. Despite this flexibility, AANAT can show stereoselectivity for ␣-branched substrates. Such behavior is useful knowledge for inhibitor design. It may also have utility in the preparation of enantiomerically pure tryptamine analogs. With physiologic substrates, it is likely that the chemical step is fast and that product release is largely rate-determining. This kinetic feature needs to be taken into account in the interpretation of mutagenesis and other perturbations of catalytic mechanism and substrate affinity. It could also be important in studies of biological regulation of acetyl-transferase activity. Based on studies with the ␣-trifluoromethyltryptamine analog and pH-rate profiles, the favored model for the catalytic mechanism of AANAT involves binding of the positively charged serotonin ammonium species, which subsequently becomes deprotonated and converted to a strong nucleophile during catalysis.