Kinetic Analysis of the Catalytic Mechanism of Serotonin N- Acetyltransferase (EC 2.3.1.87)*

Serotonin N- acetyltransferase (arylalkylamine N-acetyltransferase, AANAT, EC 2.3.1.87) is the penultimate enzyme in melatonin biosynthesis. This enzyme is of special biological interest because large changes in its activity drive the large night/day rhythm in circulating melatonin in vertebrates. In this study the kinetic mechanism of AANAT action was studied using bacterially expressed glutathione S- transferase (GST)-AANAT fusion protein. The enzymologic behavior of GST-AANAT and cleaved AANAT was essentially identical. Two-sub-strate kinetic analysis generated an intersecting line pattern characteristic of a ternary complex mechanism. The dead end inhibitor analog desulfo-CoA was competitive versus acetyl-CoA and noncompetitive versus tryptamine. Tryptophol was not an alternative substrate but was a dead end competitive inhibitor versus tryptamine and an uncompetitive inhibitor versus acetyl-CoA, indicative of an ordered binding mechanism requiring binding of acetyl-CoA first. N- Acetyltryptamine, a reaction product, was a noncompetitive inhibitor versus tryptamine and uncompetitive with respect to acetyl-CoA. Taken together these results sup-port an ordered BiBi ternary complex (sequential) kinetic mechanism for AANAT and provide a framework for inhibitor design. n M final concentration) that had been prediluted (10–100-fold) in 50 m M sodium phosphate, 500 m M NaCl, 2 m M EDTA, and 0.05 mg/ml bovine serum albumin in the absence of reducing agent immediately before use and maintained on ice during the assay. The reactions were quenched for 0–3 min with 0.6 ml of a buffer containing guanidinium-HCl (3.2 M ), sodium phosphate (0.1 M ), pH 6.8. These mixtures were treated with 0.1 ml of DTNB (2 m M , 0.1 M sodium phosphate, pH 6.8, 10 m M EDTA), vortexed, and allowed to stand for 5 min before absorbance readings were performed at 412 nm (thiophenolate-quantified assuming e 5 13.7 3 10 3 M 2 1 cm 2 1 ) (19). Background absorbances (with all compo-nents added including enzyme) were measured and subtracted from the total absorbance. A background correction was made for each acetyl-CoA concentration because acetyl-CoA had a small contaminant of free thiol (1–2% presumably free CoASH). The rate of conversion of acetyl-CoA to CoASH in the absence of amine was negligible over the course of the assay. Activity was linear with time for at least 3 min at high (2 m M ) and low (0.1 m M ) acetyl-CoA. Velocity measurements were made under initial conditions where reaction of the limiting substrate did not exceed 10%. Radiochemical Assay— A modification of an established radiochemi- cal assay (12) was used in which the concentration of tryptamine (1 m M ) is incubated with [ 14 C]acetyl-CoA (1 m M ; specific activity 5 1.24 Ci/mol) and N- [ 14 C]acetyltryptamine was measured. apparent

Serotonin N-acetyltransferase (arylalkylamine N-acetyltransferase, AANAT, 1 EC 2.3.1.87) is the penultimate enzyme in the synthesis of the pineal hormone melatonin (5-methoxy-N-acetyltryptamine). This enzyme is of special interest because large increases in its activity are responsible for the large increase in circulating melatonin levels in vertebrates (1,2). The nocturnal increase in melatonin plays a critical role in circadian biology and in the precise timing of seasonal changes in physiology, including reproductive activity. It is also thought to play a role in human mood and behavior (2).
AANATs constitute a family of approximately 23-kDa proteins which are ϳ80% identical at the amino acid level (4 -9). Highest levels of AANAT protein and mRNA occur in the pineal gland; lower levels are found in the retina; and trace levels occur in the brain, pituitary, and other tissues (4,5). AANAT activity is linked closely to enzyme protein amounts, which reflects regulation at transcriptional and post-transcriptional levels (10).
AANATs are not homologous to any other protein. However, the AANAT family is a member of a superfamily of proteins defined by the presence in tandem of two weakly conserved, approximately 15-amino acid sequences (motifs A and B). There are Ͼ150 members of this superfamily. Approximately one-third are either known acetyltransferases or share a high degree of homology with known acetyltransferases (5,10). 2 This and the results of site-directed mutagenesis analysis suggest that these conserved motifs may participate directly in substrate binding or the catalytic process of acetyl transfer. The motif A/B superfamily does not include the well studied family of liver arylamine N-acetyltransferases (EC 2.3.1.5) (11).
The catalytic action of arylamine N-acetyltransferases has been determined to be a ping-pong reaction mechanism involving formation of a covalent acetyl-S-cysteine enzyme intermediate (14 -18). As of yet, however, the catalytic mechanism of AANAT has not been established. This reflects in part the limited availability and instability of the native enzyme. Understanding the catalytic mechanism of action of AANAT is of importance for two reasons. One is that it provides a model for other members of the motif A/B superfamily. The second is that this information may play a role in the generation of specific AANAT inhibitors that may be important as therapeutic agents and as tools to probe the physiological role of melatonin. The recent cloning of AANATs (10) has now made it possible to prepare large amounts of purified recombinant enzyme for kinetic analysis, and here we present the first results of such investigations.

Expression and Purification
The entire open reading frame of the DNA encoding sheep AANAT in the plasmid vector pET15b (5) was excised with XhoI and religated in-frame into pGEX-4T-1 and transformed into the Escherichia coli strain BL21(DE3)pLysS. A frozen stock of this strain (10 l) was used to inoculate 25 ml of Luria Broth containing ampicillin (100 g/ml) and chloramphenicol (34 g/ml) in a culture flask and grown overnight at 37°C in a floor shaker. The culture was used to inoculate 2 liters of Luria Broth divided into three Erlenmeyer flasks also containing ampicillin (100 g/ml) and chloramphenicol (34 g/ml) (1:100 v/v) and grown at 37°C in a floor shaker until the absorption at 595 nm was equal to 0.5-0.6. The flasks were cooled to room temperature by standing at 4°C for 20 min and then treated with isopropyl-1-thio-␤-Dgalactopyranoside (to a final concentration of 0.2 mM). The cultures were maintained at 24°C for an additional 5 h. The cells were pelleted by centrifugation (4°C, 5,000 ϫ g, 10 min) and the cell paste (5.2 g) snap frozen with liquid N 2 and stored at Ϫ80°C. The cell paste was resuspended in 30 ml of lysis buffer (1 ϫ phosphate-buffered saline, 10 mM dithiothreitol, 10% glycerol, 10 mM EDTA, pH-6.9), and the suspension was lysed by passage through a French pressure cell at 12,000 p.s.i. Insoluble protein and cell debris were removed by centrifugation (4°C, 27,000 ϫ g, 30 min followed by 4°C, 100,000 ϫ g, 120 min), and the supernatant (25 ml) was snap frozen with liquid N 2 and stored at Ϫ80°C. The thawed solution was then incubated with Nutator (Fisher Scientific) mixing with 2 ml of glutathione-agarose (100 mg of dried resin, swollen with 20 ml of H 2 O, then pre-equilibrated in lysis buffer by washing two times with 10 ml, pelleting in a 50-ml centrifuge tube at 2,000 ϫ g in a swinging bucket centrifuge for 5 min) at room temperature for 30 min. The mixture was centrifuged at 2,000 ϫ g for 5 min and the supernatant carefully pipetted away. The pelleted resin was resuspended and washed two times with lysis buffer ϩ 263 mM NaCl (10 ml each) at 4°C. Subsequently, the glutathione-agarose resin was incubated at room temperature for 30 min on a Nutator with 20 ml of lysis buffer ϩ 113 mM NaCl ϩ 50 mM glutathione (whose pH was adjusted to 7 with 8 N NaOH). The resin was pelleted, and the supernatant (20 ml) was recovered and dialyzed (2 ϫ 500 ml) at 4°C against storage buffer (4.3 mM sodium phosphate, 1.4 mM potassium phosphate, 337 mM NaCl, 2.7 mM KCl, 5 mM dithiothreitol, 1 mM EDTA, 10% glycerol, pH 6.9). The protein concentration postdialysis was 0.5 mg/ml (total yield 10 mg) as determined by Bradford assay referenced to bovine serum albumin standard. Purity was approximately 90% as determined by 10% SDSpolyacrylamide gel electrophoresis (Coomassie staining). The protein was stored (Ϫ80°C) at 0.5 mg/ml concentration or after Centricon (Amicon Inc., Beverly, MA) ultrafiltration, at 4.2 mg/ml, and maintained stable enzyme activity for at least 4 months.
The GST-AANAT fusion protein was cleaved with thrombin and then purified according to the manufacturer's instructions (Pharmacia) to produce GST-free AANAT and protein concentration determined by Bradford assay.

Enzyme Assays
AANAT activity was measured primarily using a newly developed spectrophotometric assay. An established radiochemical assay was also used with minor modifications (12).
DTNB Product Detection Assay-This assay is based on the detection of CoASH generated during acetyl transfer by reaction with the thiol reagent DTNB (19). This assay was typically performed using a buffer containing 0.05 M sodium phosphate, pH 6.8, 500 mM NaCl, 2 mM EDTA, 0.05 mg/ml bovine serum albumin, variable acetyl-CoA (0.1-3 mM), and variable tryptamine (0.05-1 mM) at 30°C in 0.3 ml in 1.5-ml microcentrifuge tubes. Reactions were initiated with enzyme (3 l, 5-30 nM final concentration) that had been prediluted (10 -100-fold) in 50 mM sodium phosphate, 500 mM NaCl, 2 mM EDTA, and 0.05 mg/ml bovine serum albumin in the absence of reducing agent immediately before use and maintained on ice during the assay. The reactions were quenched for 0 -3 min with 0.6 ml of a buffer containing guanidinium-HCl (3.2 M), sodium phosphate (0.1 M), pH 6.8. These mixtures were treated with 0.1 ml of DTNB (2 mM, 0.1 M sodium phosphate, pH 6.8, 10 mM EDTA), vortexed, and allowed to stand for 5 min before absorbance readings were performed at 412 nm (thiophenolate-quantified assuming ⑀ ϭ 13.7 ϫ 10 3 M Ϫ1 cm Ϫ1 ) (19). Background absorbances (with all components added including enzyme) were measured and subtracted from the total absorbance. A background correction was made for each acetyl-CoA concentration because acetyl-CoA had a small contaminant of free thiol (1-2% presumably free CoASH). The rate of conversion of acetyl-CoA to CoASH in the absence of amine was negligible over the course of the assay. Activity was linear with time for at least 3 min at high (2 mM) and low (0.1 mM) acetyl-CoA. Velocity measurements were made under initial conditions where reaction of the limiting substrate did not exceed 10%.
Radiochemical Assay-A modification of an established radiochemical assay (12) was used in which the concentration of tryptamine (1 mM) is incubated with [ 14 C]acetyl-CoA (1 mM; specific activity ϭ 1.24 Ci/mol) and N-[ 14 C]acetyltryptamine was measured.
Comparative Analysis of Assays-The apparent specific activities measured with both methods were essentially identical (Ͻ20% difference). All assays were performed at least twice with duplicate measurements typically within 10%. Absorbance drift was minimal with fresh solutions (presumably because of slow air oxidation) over the 20 -30 min necessary for assay completion.  Table I. Two-substrate Kinetic Measurements-Two-substrate kinetic analysis was performed with substrate concentrations given in Fig. 3, and the data were fitted to the sequential (ternary complex) mechanism equation (Equation 2) using the computer program KinetAsyst II TM (Intel-liKinetics, State College, PA) based on the algorithms of Cleland (20), using a nonlinear least squares approach. Kinetic constants Ϯ S.E. are shown in Table II. K ma ϭ K m of acetyl-CoA in this work, K ia ϭ dissociation constant for acetyl-CoA (dissociation constant to free enzyme where acetyl-CoA binds prior to tryptamine), K mb ϭ K m of tryptamine. Fitting to a ping-pong mechanism gave a significantly larger (5-fold) sum of squares of the residuals.
Kinetic Measurements with Inhibitors-Competitive inhibition kinetic analysis was done by fitting all of the data points to the linear competitive inhibition equation of KinetAsyst II TM based on the algorithms of Cleland (20), using a nonlinear least squares approach. The fixed substrate was assumed to be saturating. Kinetic constants Ϯ S.E. are shown in Table  III.
Noncompetitive inhibition kinetic analysis was done by fitting all of the data points to the linear noncompetitive inhibition equation of KinetAsyst II TM based on the algorithms of Cleland (20), using a nonlinear least squares approach. Kinetic constants Ϯ S.E. are shown in Table III. Uncompetitive inhibition kinetic analysis was done by fitting all of the data points to the linear uncompetitive inhibition equation of Kin-etAsyst II TM based on the algorithms of Cleland (20), using a nonlinear least squares approach. Kinetic constants Ϯ S.E. are shown in Table III. The abbreviations are K ii ϭ K i intercept and K is ϭ K i slope based on double-reciprocal plot analysis according to the nomenclature of Cleland (20). The data for individual experiments with each inhibitor versus a varied substrate were fit to all three inhibitor models. Choice of kinetic fit was based on a combination of visual inspection and comparison of S.E. values and residuals for all three inhibition types applied to the data sets (20). In the cases where uncompetitive inhibition was assigned, there were no significant improvements in the standard errors or sum of squares of the residuals (less than 2-fold) by including the extra inhibitory constant K is . In the cases where competitive models were assigned, there were no significant improvements in the S.E. or the sum of squares of the residuals (less than 2-fold) by including the extra inhibitory constant K ii . The lines drawn through the data points in the figures are derived from the fitted equations above.

RESULTS
Enzyme Production-Expression of the sheep GST-AANAT fusion plasmid in E. coli resulted in the production of active soluble GST-AANAT fusion protein (Ͼ5 mg/liter of culture). Purification using glutathione affinity chromatography afforded nearly homogeneously pure recombinant protein with the predicted molecular mass (approximately 50 kDa) as determined by SDS-polyacrylamide gel electrophoresis (see Fig. 2).
GST-free AANAT was obtained by thrombin cleavage of GST-AANAT (Fig. 2). The resulting product had kinetic characteristics essentially identical to those of GST-AANAT (Table  I). GST-AANAT was used for further kinetic analysis because it was found to be more stable and easier to work with. The k cat of 25 s Ϫ1 for recombinant AANAT is similar to the reported k cat (80 s Ϫ1 ) for a pure N-acetyl-CoA-dependent acetyltransferase (22). Furthermore, it was unlikely that the enzyme contained a large fraction of inactive material because preparations obtained using a variety of purification protocols had essentially identical turnover numbers.
Assay Development-Using the DTNB assay, GST-AANAT reactions display linear activity versus time for at least 3 min in the absence of reducing agents, and enzyme activity is linear with respect to enzyme concentration up to 500 nM. After background subtraction, interference from trace reducing agents from the enzyme preparations or the acetyl-CoA was shown to be inconsequential. There is insignificant CoASH formation in the absence of amine substrate in the enzyme range employed. As little as 1 nmol of product formation (3 M) is detected reliably in a 0.3-ml reaction.
K m values for tryptamine and acetyl-CoA obtained with the DTNB assay showed good agreement with published values obtained with native sheep pineal AANAT 3 (3); specific activity values were approximately 1,000-fold higher for recombinant protein (3). These values are also in a convenient range to perform kinetic mechanism studies as outlined below. The k cat(app) and K m(app) generated with serotonin were nearly indistinguishable from those obtained with tryptamine (see Table I).
Two-substrate Kinetics-K m values were obtained for tryptamine using a range of acetyl-CoA concentrations. A double-reciprocal analysis of these data formed an intersecting line pattern (Fig. 3) that is characteristic of a ternary complex (sequential) mechanism. In contrast, a ping-pong mechanism is typically characterized by a parallel line pattern (21).  Desulfo-CoA as a Dead End Analog Inhibitor-Dead end analog inhibitors are compounds that resemble one substrate or the other but are unable to serve as substrates because of structural differences. Desulfo-CoA lacks the terminal sulfur atom of CoASH (Fig. 4) and is a potential dead end inhibitor of AANAT. It was found to be a linear competitive inhibitor versus acetyl-CoA at saturating tryptamine concentration with a K is of 1 mM, only 2-fold larger than the dissociation constant of acetyl-CoA (0.51 mM) (Fig. 5). Desulfo-CoA was also tested as a GST-AANAT inhibitor at fixed acetyl-CoA concentration and vary-ing concentrations of tryptamine (Fig. 5). It was found to be noncompetitive versus tryptamine as opposed to uncompetitive, the latter being the expected pattern if tryptamine binds first in an ordered ternary complex mechanism (21).
Tryptophol as a Dead End Analog Inhibitor-The structural analog to tryptamine in which the primary amine function is replaced by a hydroxyl function (tryptophol) (see Fig. 4) could theoretically be an alternative AANAT substrate. To investigate this possibility, 1 mM tryptophol was incubated with 500 nM GST-AANAT in the presence of 1 mM acetyl-CoA for up to 3 min. Under these conditions no detectable transfer took place, indicating that tryptophol is acetylated at least 400-fold less efficiently by GST-AANAT compared with tryptamine.
Because tryptophol was not a substrate, it was evaluated as a potential GST-AANAT inhibitor. As expected, tryptophol was a linear competitive inhibitor of tryptamine (Table III). In studies with tryptophol as the inhibitor and acetyl-CoA as the varied substrate, at subsaturating concentration of tryptamine there was a clear uncompetitive pattern of inhibition as reflected in the parallel line pattern of Fig. 6. This pattern indicates that acetyl-CoA must bind earlier than tryptophol for tryptophol inhibition to take place.
N-Acetyltryptamine Is a Product Inhibitor-N-Acetyltryptamine is one of the two products of the N-acetyltransferase reaction (see Fig. 1), and its inhibitory behavior was evaluated with GST-AANAT because reaction products can be useful diagnostic tools for kinetic mechanism studies. N-Acetyltryptamine was found to be a clear noncompetitive inhibitor versus tryptamine as the varied substrate ( Fig. 7 and Table  III). N-Acetyltryptamine was found to be an uncompetitive inhibitor versus acetyl-CoA at fixed, subsaturating tryptamine concentration (tryptamine concentration Ϸ K m of tryptamine, Table III). DISCUSSION DTNB-based AANAT Assay-Previously published kinetic assays for AANAT have relied on radioactive incorporation of acetyl ( 3 H or 14 C) into the acetylated product (3,12,23) or high pressure liquid chromatographic analysis (24,25). These assays were developed for high sensitivity as required for detection of enzyme activity in small biological samples. However, they require complex extraction procedures and are not prac-  The best fit was to a sequential (ternary complex) mechanism. For details, see "Materials and Methods"; for kinetic constants related to this plot, see Table II. tical for routine analysis of large numbers of samples as required for detailed mechanistic studies or inhibitor screens. The less complicated DTNB assay described in this report meets this requirement.
The potential difficulty with this approach was that reducing agents such as dithiothreitol and ␤-mercaptoethanol react with DTNB (Ellman's reagent). Indeed, GST-AANAT is somewhat unstable in the absence of reducing agents, losing greater than 50% activity within 2 h at 4°C (data not shown). However, the enzyme was stable within the short incubation periods used in these studies (less than 3 min). It was shown unequivocally that CoASH generation was coupled tightly to N-acetyltryptamine formation, and the activity was linear with time and enzyme concentration. An attractive feature of the DTNB detection assay is that it allows any potential amine substrate to be tested easily, providing it does not react with DTNB. Of note, a continuous spectrophotometric assay with DTNB was not possible because DTNB inhibited GST-AANAT activity.
Kinetic Mechanism of GST-AANAT-A generally useful approach to kinetic analysis of two-substrate enzymes is one in which both substrates are varied within the same experiment. It is well accepted that a double-reciprocal plot that results in intersecting lines suggests a ternary complex mechanism and that a parallel line pattern is characteristic of a ping-pong mechanism (21). Previous experiments using this approach to analyze crude preparations on AANAT from rat and bird pineal glands suggest that different mechanisms were involved (23,26). This difference is inconsistent with the high homology among vertebrate AANATs, especially within the putative binding domains and the putative catalytic site (4,10). The reported differences in the apparent mechanism of catalysis may reflect contaminants in the partially purified preparations as the assays appeared to be performed under similar conditions of pH and ionic strength. The potential problem with contaminants is avoided by the use of purified expressed GST-AANAT as described in this report.
The most important advance in this study was the evidence that a ternary complex mechanism is involved, as indicated by the clear intersecting line pattern. It should be noted that although this analytical approach has correctly predicted the mechanistic behavior of the best characterized acetyltransferases (15,27,28) it is not impossible that a covalent enzyme intermediate occurs. However, if a covalent enzyme intermediate occurs, it must form after both substrates are bound and decompose before either product leaves.
The next issue addressed was the order of binding of substrates which precedes ternary complex formation. Three major schemes are possible: (i) ordered with acetyl-CoA binding first; (ii) ordered with tryptamine binding first; and (iii) random substrate binding. To discriminate among these possibil- Acetyl-CoA concentration was fixed at 0.2 mM. Fit is to a linear noncompetitive inhibitor model. For details of the assays and calculations, see "Materials and Methods"; for the kinetic constants, see Table III. Acetyl-CoA concentration was fixed at 1 mM. Fit is to a linear noncompetitive inhibitor model. For details of the assay and calculations, see "Materials and Methods"; for the kinetic constants, see Table III. ities dead end inhibitors were used (21).
Desulfo-CoA, a dead end analog of acetyl-CoA (see Fig. 4), was shown to be a linear competitive inhibitor with respect to acetyl-CoA and a noncompetitive inhibitor with respect to tryptamine. These results rule out an ordered mechanism where tryptamine binds before acetyl-CoA. In such a mechanism, desulfo-CoA would have been uncompetitive with respect to tryptamine. The desulfo-CoA experiments leave open the possibilities that there is random binding of acetyl-CoA and tryptamine or ordered binding of acetyl-CoA before tryptamine. Another interesting point is the apparent similarity in affinity which GST-AANAT displays toward acetyl-CoA (K d ϭ 0.51 mM) and desulfo-CoA (K i ϭ 1 mM). It suggests that the thioester function contributes little binding energy in the ground state complex.
Tryptophol (Fig. 4) was next evaluated as a GST-AANAT substrate/inhibitor. It was shown that replacement of the amino function of tryptamine with a hydroxy group prevents enzyme-catalyzed acetyl transfer. The lack of reactivity suggests that the nucleophilicity of the amine is critical for enzyme-catalyzed reaction because tryptophol can bind with reasonable affinity to AANAT as demonstrated by its inhibitory behavior (see below). Interestingly, the O-acetyltransferase carnitine acetyltransferase is able to process the amino-substrate at a k cat only 13-fold lower than the normal hydroxy substrate (29). This altered reactivity between the two enzymes suggests that there may be a mechanistic difference in the chemical steps catalyzed between these two classes of acetyltransferases (O and N). Tetrahydrodipicolinate N-succinyltransferase also shows no reactivity toward the corresponding oxygen analog (30).
As expected, tryptophol proved to be a linear competitive inhibitor of GST-AANAT versus the varied substrate tryptamine. It was a clear uncompetitive inhibitor versus acetyl-CoA. Fitting the data to a noncompetitive fit gave no significant lessening of the residuals and gave a K is that was more than 10 times higher than the K ii with a very large error (Ϯ100%). These results strongly suggest that acetyl-CoA must bind before tryptamine to the enzyme, i.e. that there is an ordered mechanism. Although this inhibitory pattern is compatible with a ping-pong kinetic mechanism, a ping-pong mechanism is ruled out by the intersecting line pattern in the two substrate kinetic analysis (Fig. 3). The ordered binding suggests either (i) a conformational change in the protein which causes the tryptamine binding pocket to become accessible only after acetyl-CoA binds or (ii) tryptamine undergoes an important, direct noncovalent binding interaction with acetyl-CoA in the enzyme active site. Differentiation between these possibilities awaits further structural studies.
CoASH was not evaluated as a reversible inhibitor in the spectrophotometric assay because it reacts with DTNB. It also forms (CoAS) 2 in the absence of reducing agents, and the use of reducing agents in the radiochemical assay would require extensive kinetic characterization of the enzyme in the presence of such reagents, which is beyond the scope of this investigation. The reaction product N-acetyltryptamine is a noncompetitive inhibitor versus tryptamine and an uncompetitive inhibitor versus acetyl-CoA. This strongly suggests that AANAT obeys an ordered BiBi ternary complex mechanism with Nacetyltryptamine being the first product released followed by CoASH. The noncompetitive inhibition pattern versus tryptamine likely is caused by the binding of N-acetyltryptamine to both the acetyl-CoA-bound GST-AANAT form as well as the CoASH-bound GST-AANAT form. The lack of a slope effect in the inhibition of N-acetyltryptamine versus acetyl-CoA (at subsaturating tryptamine) presumably stems from the fact that the chemical step is very weakly reversible since a thioester bond (⌬G hydrolysis of acetyl-CoA ϭ Ϫ7.5 kcal/mol) is exchanged for an amide bond (⌬G hydrolysis of propionamide ϭ Ϫ2.1 kcal/mol) (21,31).
It has not been established that any of the acetyltransferases that have been kinetically characterized are members of the motif A/B superfamily. Accordingly, demonstration that AANAT obeys a ternary complex, ordered BiBi mechanism creates a precedent for other acetyltransferases in the motif A/B superfamily, including a eukaryotic histone N-acetyltransferase believed to be important in the regulation of gene expression (32). Delineating the kinetic mechanism also develops a framework for an approach to inhibitor design. Therefore, tryptamine analog inhibitors are unlikely to be potent at blocking AANAT action at low levels of acetyl-CoA. It remains to be seen whether bisubstrate analog inhibitors can be effective.