Novel bisubstrate analog inhibitors of serotonin N-acetyltransferase: The importance of being neutral

Article abstract of DOI:10.1016/S0045-2068(03)00081-6

Linker modified novel bisubstrate analog inhibitors 4-7 for serotonin N-acetyltransferase (arylalkylamine N-acetyltransferase, AANAT) have been designed and synthesized. Examination of these inhibitors with AANAT in vitro suggested that: (i) linker hydrogen bonding makes only modest contributions to the affinity of bisubstrate analog inhibitors studied; (ii) greater than or equal to four methylene groups between the indole and the coenzyme A (CoASH) moieties are required for a bisubstrate analog inhibitor to achieve strong AANAT inhibition; (iii) the AANAT active site appears not to accommodate positively charged linkers as well as neutral ones; and (iv) substrate amine pK_a depression may constitute one strategy for AANAT substrate recognition and catalysis. The results reported here have enhanced our understanding of AANAT substrate recognition/catalysis, and are important for novel inhibitor design. Since AANAT belongs to the GCN5-related N-acetyltransferase (GNAT) superfamily, our experimental strategies should find applications for other acetyltransferases.

Full text of DOI:10.1016/S0045-2068(03)00081-6

BIOORGANIC
CHEMISTRY
Bioorganic Chemistry 31 (2003) 398–411
www.elsevier.com/locate/bioorg
Novel bisubstrate analog inhibitors of serotonin
N-acetyltransferase: the importance
of being neutral
Weiping Zheng and Philip A. Cole*
Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine,
7
25 North Wolfe Street, Baltimore, MD 21205, USA
Received 25 April 2003
Abstract
Linker modified novel bisubstrate analog inhibitors 4–7 for serotonin N-acetyltransferase
arylalkylamine N-acetyltransferase, AANAT) have been designed and synthesized. Exami-
(
nation of these inhibitors with AANAT in vitro suggested that: (i) linker hydrogen bonding
makes only modest contributions to the affinity of bisubstrate analog inhibitors studied;
(
(
ii) greater than or equal to four methylene groups between the indole and the coenzyme A
CoASH) moieties are required for a bisubstrate analog inhibitor to achieve strong AANAT
inhibition; (iii) the AANAT active site appears not to accommodate positively charged
linkers as well as neutral ones; and (iv) substrate amine pK depression may constitute one
a
strategy for AANAT substrate recognition and catalysis. The results reported here have
enhanced our understanding of AANAT substrate recognition/catalysis, and are important
for novel inhibitor design. Since AANAT belongs to the GCN5-related N-acetyltransferase
(GNAT) superfamily, our experimental strategies should find applications for other acetyl-
transferases.
Ó 2003 Elsevier Science (USA). All rights reserved.
*
Corresponding author. Fax: 1-410-614-7717.
E-mail address: pcole@jhmi.edu (P.A. Cole).
0
045-2068/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0045-2068(03)00081-6

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
Keywords: Serotonin N-acetyltransferase; Catalysis; Inhibitors; pKa
399
1
. Introduction
Serotonin N-acetyltransferase (arylalkylamine N-acetyltransferase, AANAT, EC
.3.1.87) catalyzes the acetyl transfer from acetyl-coenzyme A (acetyl-CoA) to sero-
2
tonin to form acetyl-serotonin (Fig. 1), which is the rate-limiting step in the biosyn-
thesis of the circadian hormone melatonin (5-methoxy-N-acetyltryptamine) from
serotonin [1]. Besides the natural substrate serotonin, AANAT also efficiently cata-
lyzes acetyl transfer to other arylalkylamines such as tryptamine and phenethyl-
amine.
The mechanism and inhibition of AANAT have drawn a lot of attention [2]. AA-
NAT inhibitors are potential molecular probes for the biological roles played by the
hormone melatonin, and these inhibitors may be potential therapeutics for sleep and
mood disorders. Moreover, studies on AANAT may expand our understanding of
the GCN5-related N-acetyltransferase (GNAT) superfamily which include some his-
tone N-acetyltransferases and aminoglycoside N-acetyltransferases in addition to
AANAT [3]. Studies on AANATÕs catalytic mechanism have revealed it obeys an or-
dered Bi Bi sequential (ternary complex) mechanism with acetyl-CoA binding pre-
ceding arylalkylamine substrate [4]. These kinetic studies have led to the design
and synthesis of AANAT bisubstrate analog inhibitors with two different linker
types, i.e., amide- and ketone-linked compounds represented by 1, 2, and 3 (Fig. 2)
[
5]. Bisubstrate analog inhibitors have been used along with AANAT to solve the
Fig. 1. AANAT-catalyzed acetyl transfer reaction.

400
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
Fig. 2. The structures of AANAT bisubstrate analog inhibitors.
X-ray structure of the AANAT–bisubstrate analog complex which has shed further
insight into molecular recognition and catalytic mechanism [6].
We were interested in further varying the linker structure in these bisubstrate an-
alog inhibitors to potentially enhance inhibition and gain new mechanistic insights.
Of note, the indole and coenzyme A (CoASH) moieties were kept intact in our design
due to their mimicry of the substrates tryptamine and acetyl-CoA, respectively. These
newly designed AANAT bisubstrate analog inhibitors (4–7, Fig. 2) allow us to ex-
plore the nature of hydrogen bonding and electrostatics in molecular recognition.
2
. Materials and methods
2
.1. General
All reagents for chemical synthesis were purchased from Aldrich. Reagents for en-
0
zyme inhibition assays were purchased from Sigma: CoASH, 5,5 -dithiobis(2-nitro-
0
benzoic acid) (DTNB), N-(2-hydroxyethyl)piperazine-N -(3-propanesulfonic acid)
EPPS) and assay-grade tryptamine hydrochloride; Pharmacia Biotech.: acetyl-
(
CoA; Fisher Scientific: sodium phosphate, guanidinium hydrochloride, and EDTA.
All commercially available reagents were used as purchased without further purifica-
1
tion. H NMR spectra were obtained with a Bruker AMX 300 spectrometer
(
300 MHz), and chemical shift values (d) were expressed as parts per million
(
ppm) relative to tetramethylsilane (TMS). Electrospray ionization mass (ESI-MS)
spectra were obtained with API 150EX (PE Biosystems). High resolution mass
HRMS) spectra were recorded on a MALDI DE-STR machine at the mass spec-
trometry facility of the University of California, Riverside. Silica gel (Merck, 63–
(
ꢀ
2
00 mesh, 60 A) was used for flash column chromatography. Analytical thin layer
chromatography (TLC) was performed on Sigma-Aldrich silica gel 60 F 254 TLC
plates (thickness: 200 lm). Preparative HPLC isolation of bisubstrate analog inhib-
itors was performed on a reverse-phase C-18 column (25 ꢀ 2.14 cm) (Microsorbtm-
1
00, Rainin). Analytical HPLC used to monitor the progress of the CoASH

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
401
conjugation reactions with alkylhalides was performed on a reverse-phase C-18 col-
umn (25 ꢀ 0.46 cm) (Microsorb-MVTM-100, Rainin), eluted with a gradient of
5
(
0 mM aqueous potassium phosphate buffer (pH 4.5) (mobile phase A) and MeOH
mobile phase B) (0–3 min, 0% B; 3–15 min, linear increase to 65% B; 15–20 min, lin-
ear increase to 100% B; and 20–25 min, linear decrease to 0% B) (1.0 mL/min). All
final bisubstrate analog inhibitors were >96% pure based on analytical HPLC pro-
files with two different solvent systems employing the same method. (i) 50 mM aque-
ous potassium phosphate buffer (pH 4.5) (mobile phase A) and MeOH (mobile phase
B); (ii) water (0.05% TFA) (mobile phase A) and acetonitrile (0.05% TFA) (mobile
phase B). The recombinant sheep glutathione S-transferase (GST)-serotonin N-ace-
tyltransferase (GST-AANAT) (ꢁ90% pure) was expressed in Escherichia coli and pu-
rified by the glutathione–agarose affinity column, as described previously, and it has
been shown that nearly identical acetyltransferase kinetic behavior was observed for
GST-AANAT and GST-free AANAT [4]. Therefore, GST-AANAT was used for all
the AANAT inhibition assays described in this report.
2
.2. Synthesis of 1-bromo-5-(3-indolyl)pentane (8)
To a stirred cooled (ice-water bath) solution of 1-hydroxy-5-(3-indolyl)pentane
12) [7] (145 mg, 0.71 mmol) and carbon tetrabromide (390 mg, 2.1 mmol) in 3 mL
(
of anhydrous Et O was added triphenyl-phosphine (309 mg, 1.19 mmol) in small
2
portions. The reaction mixture was stirred at 0 °C for 15 min before the supernatant
was collected and the oily residue was washed with Et O twice. The combined or-
ganic phases were concentrated in vacuo and the residue was subjected to silica
2
gel flash column chromatography (EtOAc:hexane ¼ 1:10; t.l.c. R
f
ꢁ0.25), affording
1
6
1
7
2
3 mg (33%) of 8 as a slightly yellow oil. H NMR (CDCl
3
) d (ppm) 7.91 (bs,
H), 7.60 (d, J ¼ 7:8 Hz, 1H), 7.36 (d, J ¼ 8:4 Hz, 1H), 7.19 (t, J ¼ 8:1 Hz, 1H),
.11 (t, J ¼ 7:4 Hz, 1H), 6.99 (bs, 1H), 3.41 (t, J ¼ 6:9 Hz, 2H), 2.78 (t, J ¼ 7:5 Hz,
H), 1.87–1.97 (m, 2H), 1.70–1.80 (m, 2H), 1.50–1.60 (m, 2H). ESI-MS: 266
þ
(
[M + H] ).
2
.3. Synthesis of 1-bromo-4-(3-indolyl)butane (9) [8]
(
a) To a stirred cooled (ice-water bath) solution of 3-indolebutyric acid (408 mg,
.0 mmol) in 3 mL of anhydrous THF was added dropwise BH
ꢂ THF (1.0 M in
. The reaction mixture was stirred at room tem-
perature for 3 h before water was added cautiously to destroy excess BH . THF was
2
3
THF) (6.0 mL, 6.0 mmol) under N
2
3
removed in vacuo and the product was extracted into EtOAc which was dried over
anhydrous Na SO , filtered, and concentrated in vacuo. The resulting residue was
2
4
subjected to silica gel flash column chromatography (EtOAc:hexane ¼ 1:2), giving
1
2
30 mg (61%) of the intermediate 1-hydroxy-4-(3-indolyl)butane. H NMR (CDCl
3
)
d (ppm) 7.93 (bs, 1H), 7.61 (d, J ¼ 7:8 Hz, 1H), 7.36 (d, J ¼ 8:1 Hz, 1H), 7.19 (t,
J ¼ 7:5 Hz, 1H), 7.11 (t, J ¼ 7:4 Hz, 1H), 6.99 (bs, 1H), 3.69 (t, J ¼ 6:6 Hz, 2H),
2
of compound 8 from 12, 48 mg (16%) of compound 9 was obtained as a brownish
.80 (t, J ¼ 7:2 Hz, 2H), 1.60–1.86 (m, 4H). (b) In the same manner as the synthesis

402
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
solid from the above-obtained 1-hydroxy-4-(3-indolyl)butane (230 mg, 1.22 mmol)
after silica gel flash column chromatography (EtOAc:hexane ¼ 1:12; t.l.c. R
f
ꢁ0.20
1
in EtOAc:hexane 1:10). H NMR (CDCl
1
(
3
) d (ppm) 7.91 (bs, 1H), 7.61 (d, J ¼ 7:5 Hz,
H), 7.36 (d, J ¼ 8:1 Hz, 1H), 7.20 (t, J ¼ 7:5 Hz, 1H), 7.13 (t, J ¼ 7:4 Hz, 1H), 6.99
bs, 1H), 3.45 (t, J ¼ 6:6 Hz, 2H), 2.81 (t, J ¼ 7:5 Hz, 2H), 1.81–2.06 (m, 4H).
2
.4. Synthesis of 1-bromo-3-(3-indolyl)propane (10) [8]
In the same manner as the preparation of compound 9, 62 mg (31%) of 10 was
obtained from 3-indolepropionic acid (387 mg, 2.05 mmol) after two steps via the in-
1
termediate 1-hydroxy-3-(3-indolyl)propane. t.l.c. R
NMR (CDCl
) d (ppm) 7.95 (bs, 1H), 7.62 (d, J ¼ 7:8 Hz, 1H), 7.37 (d, J ¼ 7:8 Hz,
H), 7.21 (t, J ¼ 7:5 Hz, 1H), 7.13 (t, J ¼ 7:4 Hz, 1H), 7.04 (bs, 1H), 3.45 (t,
J ¼ 6:6 Hz, 2H), 2.95 (t, J ¼ 7:2 Hz, 2H), 2.21–2.31 (m, 2H).
f
ꢁ0.10 in EtOAc:hexane 1:10; H
3
1
2
.5. Synthesis of N-chloropropyl-2-(3-indolyl)ethylamine (13)
To a stirred solution of tryptamine (352 mg, 2.2 mmol) in 4 mL of anhydrous tet-
rahydrofuran (THF) was added dropwise 1-bromo-3-chloropropane (99 lL,
.0 mmol) under N
at room temperature. The mixture was stirred at 32–34 °C
1
2
(2 h), 40–45 °C (6 h), and room temperature (overnight) before water was added to
dissolve the precipitate formed. THF was removed in vacuo and EtOAc was added.
The mixture was acidified with 6N aqueous HCl at 0 °C, and the water layer was
washed with Et O to remove any unreacted 1-bromo-3-chloropropane before it
was basified to c.a. pH 12 with a 6N aqueous NaOH solution at 0 °C. EtOAc was
added and the resulting solution was washed with water, dried over anhydrous
2
MgSO
to preparative TLC (CH
(
4
. EtOAc removal in vacuo afforded a brown oily residue which was subjected
2
Cl
33%) of the isolated product 13 as a viscous colorless oil. H NMR (CDCl
2
:MeOH:NH
4
OH ¼ 100:10:2; R
f
ꢁ0.6), affording 77 mg
1
3
) d
(
ppm) 8.26 (bs, 1H), 7.63 (d, J ¼ 7:8 Hz, 1H), 7.36 (d, J ¼ 8:1 Hz, 1H), 7.20 (t,
J ¼ 7:2 Hz, 1H), 7.12 (t, J ¼ 7:4 Hz, 1H), 7.03 (bs, 1H), 3.57 (t, J ¼ 6:6 Hz, 2H),
þ
2
.98 (bs, 4H), 2.79 (t, J ¼ 6:9 Hz, 2H), 1.88–1.97 (m, 2H). ESI-MS: 237 ([M + H] ).
t
2
.6. Synthesis of N- butyloxycarbonyl-N-chloropropyl-2-(3-indolyl)ethylamine (14)
To a stirred cooled (ice-water bath) suspension of 13 (53 mg, 0.224 mmol) in
.5 mL of THF was added dropwise a solution of di-tert-butyl dicarbonate (49 mg,
.225 mmol) in 0.5 mL of THF. The reaction mixture was heated at 50–62 °C for
.5 h before THF was removed in vacuo. The resulting oily residue was subjected
1
0
1
to preparative TLC (EtOAc:hexane ¼ 1:3; R ꢁ0.5), affording 24 mg (32%) of 14 as
f
1
a viscous colorless oil. H NMR (CDCl
1
3
) d (ppm) 8.04 (bs, 1H), 7.64 (d, J ¼ 7:5 Hz,
H), 7.36 (d, J ¼ 8:1 Hz, 1H), 7.20 (t, J ¼ 6:9 Hz, 1H), 7.12 (t, J ¼ 7:4 Hz, 1H), 7.01
(
ESI-MS: 337 ([M + H] ).
bs, 1H), 3.51 (bs, 4H), 3.33 (bs, 2H), 3.00 (bs, 2H), 1.98 (bs, 2H), 1.35–1.52 (m, 9H).
þ

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
403
2.7. Typical procedure for the conjugation of CoASH with alkyl halides (8–10, 14) to
provide analogs 4, 5, 6, and 15
To a stirred solution of an alkyl halide in freshly degassed MeOH was added
dropwise a solution of CoASH (0.5–1.0 equivalent) in a 1.0 M aqueous buffer solu-
tion of triethylammonium bicabonate (pH 8.1–8.5) at room temperature. The reac-
tion mixture was stirred at room temperature or at 50–60 °C (for the conjugation
between 7 and CoASH, where one equivalent of NaI was also present) for 15–24 h
before MeOH was removed in vacuo. EtOAc was added to extract away any unre-
acted alkyl halide. The aqueous phase was lyophilized overnight, and the residue was
subjected to preparative HPLC to isolate the desired products. The HPLC column
was eluted with a gradient of 50 mM aqueous potassium phosphate buffer (pH
4
.5) (mobile phase A) and MeOH (mobile phase B) (0–5 min, 0% B; 5–35 min, linear
increase to 65% B; 35–50 min, linear increase to 100% B; and 50–55 min, linear de-
crease to 0% B) (10 mL/min), and was monitored at 260 nm. The isolated salt-con-
taining product was desalted by the same preparative HPLC column eluted with
double deionized (d.d.) water for 30 min, followed by MeOH to elute the product.
The collected fractions were concentrated in vacuo and lyophilized to give the final
1
products as white puffy solids which were characterized with H NMR and HRMS,
and used for the AANAT inhibition assays.
2
.8. Analog 4
By using the above-described typical conjugation procedure, analog 4 was pre-
1
pared in 42% yield from compound 8. HPLC t
(
R
(19.2 min); H NMR (CD
ppm) 8.55 (s, 1H), 8.19 (s, 1H), 7.48 (d, J ¼ 7:8 Hz, 1H), 7.30 (d, J ¼ 8:1 Hz,
3
OD) d
1
H), 7.04 (t, J ¼ 6:9 Hz, 1H), 6.98 (s, 1H), 6.95 (t, J ¼ 6:9 Hz, 1H), 6.12 (d,
J ¼ 6:0 Hz, 1H), 4.86–4.98 (m, 2H), 4.49 (bs, 1H), 4.26 (t, J ¼ 5:1 Hz, 2H), 4.07
(
s, 1H), 4.00 (dd, J ¼ 9:6, 5.4 Hz, 1H), 3.56 (dd, J ¼ 9:9, 4.2 Hz, 1H), 3.46 (t,
J ¼ 6:6 Hz, 2H), 3.29 (t, J ¼ 6:6 Hz, 2H), 2.72 (t, J ¼ 7:2 Hz, 2H), 2.57 (t,
J ¼ 7:2 Hz, 2H), 2.51 (t, J ¼ 7:2 Hz, 2H), 2.41 (t, J ¼ 6:9 Hz, 2H), 1.65–1.75 (m,
2
H), 1.55–1.65 (m, 2H), 1.40–1.52 (m, 2H), 1.06 (s, 3H), 0.82 (s, 3H). HRMS Calcd.
þ
for C34
H
51KN
8
O
16
P
3
S ([M + K] ) 991.1994; Found: 991.2039.
2
.9. Analog 5
By using the above-described typical conjugation procedure, analog 5 was pre-
1
pared in 20% yield from compound 9. HPLC t (19.4 min); H NMR (CD OD) d
R
3
(
1
ppm) 8.55 (s, 1H), 8.19 (s, 1H), 7.49 (d, J ¼ 7:8 Hz, 1H), 7.30 (d, J ¼ 8:1 Hz,
H), 7.04 (t, J ¼ 6:9 Hz, 1H), 6.99 (s, 1H), 6.95 (t, J ¼ 6:9 Hz, 1H), 6.12 (d,
J ¼ 5:7 Hz, 1H), 4.88–4.96 (m, 2H), 4.48 (bs, 1H), 4.26 (bs, 2H), 4.07 (s, 1H), 4.00
dd, J ¼ 10:2, 4.8 Hz, 1H), 3.56 (dd, J ¼ 9:6, 3.9 Hz, 1H), 3.45 (t, J ¼ 6:9 Hz, 2H),
(
3
.28 (t, J ¼ 6:6 Hz, 2H), 2.74 (t, J ¼ 6:9 Hz, 2H), 2.55 (t, J ¼ 7:2 Hz, 4H), 2.40 (t,
J ¼ 6:9 Hz, 2H), 1.72–1.84 (m, 2H), 1.55-1.69 (m, 2H), 1.06 (s, 3H), 0.81 (s, 3H).
þ
HRMS Calcd. for C33
H
49KN
8
O
16
P
3
S ([M + K] ) 977.1837; Found: 977.1906.

404
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
2
.10. Analog 6
By using the above-described typical conjugation procedure, analog 6 was pre-
1
R 2
pared in 25% yield from compound 10. HPLC t (18.1 min); H NMR (D O) d
(
1
ppm) 8.55 (s, 1H), 8.20 (s, 1H), 7.65 (d, J ¼ 8:1 Hz, 1H), 7.49 (d, J ¼ 8:1 Hz,
H), 7.24 (t, J ¼ 7:2 Hz, 1H), 7.17 (s, 1H), 7.15 (t, J ¼ 7:2 Hz, 1H), 6.17 (d,
J ¼ 5:4 Hz, 1H), 4.65 (bs, 1H), 4.31 (bs, 2H), 4.08 (s, 1H), 3.90 (dd, J ¼ 10:2,
4
.2 Hz, 1H), 3.62 (dd, J ¼ 9:9, 4.5 Hz, 1H), 3.48 (t, J ¼ 6:3 Hz, 2H), 3.34 (t,
J ¼ 6:6 Hz, 2H), 2.84 (t, J ¼ 7:2 Hz, 2H), 2.67 (t, J ¼ 6:6 Hz, 2H), 2.60 (t,
J ¼ 7:2 Hz, 2H), 2.44 (t, J ¼ 6:6 Hz, 2H), 1.90–2.02 (m, 2H), 0.94 (s, 3H), 0.79 (s,
þ
3
H). HRMS Calcd. for C32
H
48
N
8
O
16
P
3
S ([M + H] ) 925.2122; Found: 925.2116.
2
.11. Analog 15
By using the above-described typical conjugation procedure, analog 15 was pre-
1
pared in 8% yield from compound 14. HPLC tR (21.1 min); H NMR (D O) d
2
(
1
(
ppm) 8.53 (s, 1H), 8.20 (s, 1H), 7.61 (d, J ¼ 8:1 Hz, 1H), 7.46 (d, J ¼ 7:8 Hz,
H), 7.21 (t, J ¼ 8:1 Hz, 1H), 7.06–7.15 (m, 2H), 6.15 (d, J ¼ 5:4 Hz, 1H), 4.62
bs, 1H), 4.28 (bs, 2H), 4.05 (s, 1H), 3.87 (dd, J ¼ 10:2, 4.2 Hz, 1H), 3.60 (dd,
J ¼ 9:9, 4.5 Hz, 1H), 3.47 (bs, 4H), 3.32 (bs, 2H), 3.02–3.28 (m, 2H), 2.95 (bs,
2
5
1
H), 2.60 (bs, 2H), 2.40–2.50 (m, 4H), 1.52–1.78 (m, 2H), 1.42 (bs, 4H), 1.20 (bs,
ꢃ
H), 0.92 (s, 3H), 0.78 (s, 3H). HRMS Calcd. for C H N O P S ([M ) H] )
39
59
9
18
3
066.2912; Found: 1066.2992.
2
.12. Analog 7
To a stirred cooled (ice-water bath) solution of 15 (5 mg, 4.7 lmol) in 439 lL of
d.d. water was added dropwise 430 lL of trifluoroacetic acid (TFA). The reaction
mixture was stirred at room temperature for 105 min before it was basified to pH
8–9 using a 1.0 M aqueous triethylammonium bicarbonate buffer. The whole mixture
was lyophilized overnight before it was subjected to preparative HPLC to isolate and
desalt the product 7. The HPLC methods used were the same as those for isolating/
1
desalting analog 15. 3.7 mg (81%) of analog 7 was obtained. HPLC t
NMR (D
J ¼ 8:1 Hz, 1H), 7.12 (s, 1H), 7.10 (t, J ¼ 8:1 Hz, 1H), 7.00 (t, J ¼ 7:5 Hz, 1H),
.99 (d, J ¼ 3:0 Hz, 1H), 4.54 (bs, 1H), 4.22 (bs, 2H), 3.96 (s, 1H), 3.79 (dd,
R
(15.0 min); H
2
O) d (ppm) 8.43 (s, 1H), 8.05 (s, 1H), 7.44 (d, J ¼ 7:8 Hz, 1H), 7.32 (d,
5
J ¼ 10:2, 4.5 Hz, 1H), 3.54 (dd, J ¼ 10:5, 4.2 Hz, 1H), 3.41 (t, J ¼ 6:3 Hz, 2H),
3
.20–3.28 (m, 4H), 3.00–3.08 (m, 4H), 2.50 (t, J ¼ 6:6 Hz, 4H), 2.39 (t, J ¼ 6:3 Hz,
2
C H N O P S ([M + H] ) 968.2544; Found: 968.2610.
H), 1.78–1.90 (m, 2H), 0.86 (s, 3H), 0.73 (s, 3H). HRMS Calcd. for
þ
34
53
9
16
3
2
.13. Assay of AANAT inhibition by bisubstrate analog inhibitors 4–7
AANAT activity was measured as described previously employing a spectro-
photometric assay in which the enzymatic product CoASH concentrations were

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
405
Table 1
In vitro inhibition of AANAT acetyltransferase activity by bisubstrate analog inhibitors
Inhibitor
1
2
3
4
5
6
7
a
b
b
c
K
i
(lM)
0.048
0.067
0.032
0.36
0.31
0.84
5.84
a
i
Apparent inhibition constants (K ) were derived from the Dixon plots, assuming all inhibitors are
competitive versus acetyl-CoA, as demonstrated for analog 1 [5a]. The standard error for these mea-
surements is estimated to be ꢄ20%.
b
See [9].
See [5b].
c
determined indirectly by monitoring its reaction product with DTNB at 412 nm [4].
Assay reactions (pH 6.8 or 8.0) containing fixed concentrations of substrates acetyl-
CoA (0.3 mM) and tryptamine (0.3 mM) (similar to their K
five concentrations of a bisubstrate analog inhibitor (varied around K
with AANAT (14–56 nM) and the reactions were allowed to proceed for 2 min at
0 °C before being quenched with a 3.2 M aqueous guanidinium hydrochloride solu-
m
values [4]), and at least
i
) were initiated
3
tion (pH 6.8). Rate measurements were made under initial conditions, i.e., limiting
substrate turnover is 610%. All measurements were made in duplicate and data
agreed within 20% in each duplicate run. A background (quenched before adding
AANAT) was run for each inhibitor concentration. Inhibition kinetics were analyzed
with Dixon plots that were shown to be linear in all cases. Apparent inhibition con-
i
stants (K ) (Table 1) were estimated from the Dixon plots, assuming all inhibitors are
competitive versus acetyl-CoA and non-competitive versus tryptamine, as demon-
strated previously for analog 1 [5a]. Although these K values are only approximate
i
for inhibitors with slow-binding behavior that has been described for analog 1, the
prototype of this class of bisubstrate analog inhibitors reported in this work and pre-
viously [5b,9], good agreement between the K
* obtained from more extensive time-course measurements has been demonstrated
previously for analog 1 [9].
i
obtained from this analysis and the
K
i
3
. Results and discussion
Analogs 4–6 were designed to investigate the dependence of the inhibitory po-
tency on the distance between the indole and the CoASH moieties in AANAT bi-
substrate analog inhibitors without the complications of hydrogen bonding groups
in the linker. Comparative studies of compound 1, 2, 3, and 4 allow an estimate of
the contribution of amide and ketone groups in the linker with respect to overall
binding affinity. ‘‘Reduction’’ of the amide linkage of 2 to a 2° amino group gave
analog 7. An intriguing feature of analog 7 is its chargeable linker, as compared to
the neutral ones in all other current AANAT bisubstrate analog inhibitors. Studies
on 7 permit an examination of the effects of linker positive charge on binding in-
teractions with AANAT. Furthermore, the ionization state of the linker amino
group of enzyme-bound 7 at a certain pH can potentially reflect that of the amino

406
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
group of tryptamine (or the natural substrate serotonin) in the AANAT active
site.
Analogs 4–6 were prepared from the coupling at room temperature between Co-
ASH and the corresponding 3-indolealkylbromides (8–10) in a mixture of 1.0 M trie-
thylammonium carbonate buffer (pH 8.1–8.5) and freshly degassed MeOH under N
Scheme 1). The reaction progress was monitored with analytical HPLC, and the
2
(
conjugates were isolated by preparative HPLC. 3-Indolealkylbromides 8–10 were
synthesized according to the procedures in Schemes 2 and 3. Intermediate 12 was
prepared from indole according to the literature scheme and conditions as described
[
7]. 3-Indolepentylbromide (8) was obtained by treating 12 with CBr
anhydrous Et O. Alkylbromide 9 and 10 were prepared from the corresponding
-indolealkanoic acid using the literature scheme [8] with several modifications
employed.
Scheme 4 depicts the synthesis of analog 7. Controlled alkylation of tryptamine
with 1-bromo-3-chloropropane in anhydrous THF gave in 33% of yield intermediate
4 3
and PPh in
2
3
þ
ꢃ
Scheme 1. (i) Et
3
NH ꢂ HCO (pH 8.1–8.5)/MeOH.
3
2 5 4 4 3 2
Scheme 2. (i) Mg, C H I, anisole; (ii) glutaric anhydride; (iii) LiAlH /THF; (iv) CBr /PPh /Et O.

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
407
Scheme 3. (i) BH
3
4 3 2
ꢂ THF/THF; (ii) CBr /PPh /Et O.
t
þ
ꢃ
Scheme 4. (i) 1-bromo-3-chloropropane/THF; (ii) ( Boc)
2
O/THF; (iii) Et
3
NH ꢂ HCO (pH 8.4)/MeOH;
3
2
(iv) TFA/H O (1:1).
t
1
3 whose secondary amino group was then protected as Boc carbamide (14). The
t
rather modest yield (32%) for this Boc protection step may stem from competing in-
tramolecular attack by the indole or amine. The coupling between CoASH and in-
termediate 14 was carried out in a mixture of 1.0 M triethylammonium carbonate
buffer (pH 8.4) and freshly degassed MeOH at 50–55 °C under N . The reaction pro-
2
gress was monitored with analytical HPLC, and the product 15 was isolated by pre-
parative HPLC. Treating precursor 15 with 50% aqueous trifluoroacetic acid (TFA)
solution gave analog 7 that was isolated by preparative HPLC.
The identities of the final products (4–7) were confirmed by NMR and high res-
olution MS (HRMS) analyses.

408
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
Analogs 4–7 were evaluated as AANAT inhibitors on the purified sheep enzyme
in vitro, as described previously [9]. Bioassay data were analyzed by Dixon plots,
with results being recorded in Table 1.
The similar potency of inhibitors 1, 2, and 3 observed in our previous studies [5]
suggested that the hydrogen bond between the amide nitrogen of 1 and the backbone
carbonyl oxygen of Met159 on AANAT, observed on the structure of AANAT com-
plexed with 1 [6b], is not energetically critical for inhibitor binding. However, it is
possible that the amide carbonyl group contributes to the inhibitor binding affinity
by hydrogen bonding to the backbone amide NH of L124 on AANAT, as observed
in the co-crystal structure of AANAT and 1 [6b]. It is apparent from Table 1 that
inhibitor 4 is 5- to 10-fold less potent as compared to inhibitors 1, 2, and 3. This sug-
gested that the hydrogen bond acceptor capability of the carbobyl group is likely to
be worth approximately 1.4 kcal/mol in enhancing affinity.
The essentially identical inhibitory potency (K ) for analogs 4 and 5 suggested that
i
the linkers of these compounds can each confer the necessary optimal spacing and
orientation of the indole and CoASH moieties for binding to AANAT. However,
the increased spacing in compound 4 versus 5 may induce offsetting effects such as
increased hydrophobic interactions with AANAT mediated by 4 but greater entropic
loss due to its longer and more flexible linker.
The 2.5-fold decreased inhibitory potency of 6 as compared to 4 or 5 suggest that
optimal orientation of the indole and CoASH moieties for enzyme binding is difficult
to reach by shortening the linker length further. This observation further implies
that, in contrast to arylethylamines [2a], arylmethylamines should serve as poor sub-
strates of AANAT. It is conceivable that, in order for acetylation to occur, these
poor substrates have to be pulled away from the optimal binding environment.
When we designed inhibitor 7, we considered the plausible mechanistic model that
arylalkylamine substrate binds to AANAT as the positively charged ammonium salt
and is subsequently deprotonated by a general base during catalysis [2,10]. Under
these circumstances, the bisubstrate analog inhibitor 7 might be expected to display
higher binding affinity as compared to the neutral amide-linked inhibitor 1 or 2
(
Fig. 2). However, 7 was observed to be a 122-fold weaker inhibitor than 1 and 2.
This observation suggested that the enzyme appears to have a preference for the
binding of linker uncharged bisubstrate analogs rather than positively charged ones.
Since ‘‘charge-neutral’’ hydrogen bonds tend to be more energetically favorable than
‘
‘neutral–neutral’’ [11], these results suggest that the contribution of linker NH hy-
drogen bonding to bisubstrate analog affinity are not dominant. Indeed, as men-
tioned above, prior results with the ketone analog 3 indicated that the amide
nitrogen in the linker was not critical for affinity [5b].
Further insight into the ionization state of the linker amine of 7 in the AANAT
active site and the importance of hydrogen bonding is provided by comparing the
behaviors of compounds 4 and 7. Analog 4 lacks the NH-group from the linker in
analog 7 but is otherwise a nice isostere. Analog 4 was shown to be a 16-fold stronger
inhibitor than analog 7. Since the affinity of 1, 2, and 3 for AANAT are very similar,
we reasoned that it is the presence or absence of the potential for positive charge on
the linker of 7 versus 4 that results in the observed difference in binding affinity [12].

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
409
It is not unreasonable to expect that the interactions of indole and CoASH moieties
with AANAT are conserved for inhibitors 1, 2, 3, 4, and 7. Indeed, essentially iden-
tical AANAT complexes with 1, 2, or 7 were observed in X-ray crystallography stud-
ies (S. K. Burley and PAC, unpublished results). This structural conservation should
also justify our assumption that all inhibitor 1-related bisubstrate analog inhibitors
were competitive versus substrate acetyl-CoA and non-competitive versus substrate
tryptamine, as demonstrated for 1 [5a], and expected for AANAT which obeys an
ordered Bi Bi kinetic mechanism. Thus, it is deduced that the positively charged form
of 7 binds AANAT 16-fold more weakly than the neutral form of 7. Based on Eq. (1)
which should hold at low pH (6.8) when the linker amine is largely protonated in free
solution:
½
amino-pKa ðactive siteÞꢅ ꢃ ½amino-pKa ðfree solutionÞꢅ
log½ðK -4Þ=ðK -7Þꢅ
we estimate that the pK
¼
i
i
ð1Þ
a
is ꢁ1.2 lower for the linker amino group of inhibitor 7 in
the AANAT active site versus in free solution.
Inhibitors 7 and 4 were further evaluated at a higher pH (8.0). A 60-fold decrease
in potency was observed for inhibitor 4 at pH 8.0 versus pH 6.8, however, the cor-
responding decrease was 30-fold for inhibitor 7. These results further support the
concept that linker amine pK
7
a
depression is associated with the binding of inhibitor
to the AANAT active site. The drastic decrease in affinity of 4 for AANAT likely
results from AANAT ionizable groups that affect binding at the indole and/or the
CoASH sites. Thus, changes in ionization from one or more groups of AANAT ap-
pear to weaken the interaction between AANAT and both bisubstrate analogs 4 and
7
as the pH is raised from 6.8 to 8.0. However, this weakening is modestly offset by
the presumed increased proportion (by ꢁ12%, assuming pK 8.8, see below) at pH
a
8
.0 versus 6.8 of the neutral linker form of inhibitor 7 which is assumed to have
1
a
6-fold higher affinity for AANAT. Since tryptamine has a pK of 10 for its primary
amino group in free solution [13], it is calculated that the AANAT active site likely
depresses the substrate amine pK
from 10 to ꢁ8.8 (see Eq. (1)). These findings sug-
gest that substrate amine pK depression and general base-mediated substrate am-
a
a
monium ion deprotonation may work synergistically during AANAT substrate
recognition and catalysis. AANAT may use general base(s) to directly or indirectly
(
via one or more water molecules) deprotonate the substrate ammonium ion [2b,6].
His120 and/or His122 may serve as the general base(s) [2b,6]. The maximum that a
general base might be expected to contribute to rate enhancement at pH 7 is 60-fold.
This is calculated based on the fact that about 1/60th of the tryptamine (assuming
a
pK 8.8 in the active site) is already the neutral amine at pH 7. Such a 60-fold rate
enhancement is in fair agreement with the effect (40-fold) observed with the double
His to Gln mutant [2b].
In summary, linker modified novel AANAT bisubstrate analog inhibitors 4–7
have been designed and synthesized. Evaluation of these inhibitors with AANAT
in vitro has suggested that: (i) linker hydrogen bonding makes only modest contri-
butions to the affinity of bisubstrate analog inhibitors studied; (ii) greater than or

410
W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
equal to four methylene groups between the indole and CoASH moieties are re-
quired for a bisubstrate analog inhibitor to achieve strong AANAT inhibition;
(
as well as neutral ones; and (iv) substrate amine pK depression may constitute one
iii) the AANAT active site appears not to accommodate positively charged linkers
a
strategy for AANAT substrate recognition and the catalysis. The results reported
here have enhanced our understanding of AANAT substrate recognition and catal-
ysis and will guide future inhibitor design. Since AANAT belongs to the GNAT
superfamily, our experimental strategies and conclusions may extend to other impor-
tant acetyltransferase enzymes.
Acknowledgments
Financial support from the Ellison Medical Foundation and National Institutes
of Health are highly appreciated. The NMR studies were performed in the Biochem-
istry NMR Facility at Johns Hopkins University, which was established by a grant
from the National Institutes of Health (GM 27512) and a Biomedical Shared Instru-
mentation Grant (1S10-RR06262-0).
References
[
1] (a) D.C. Klein, J.L. Weller, Science 169 (1970) 1093–1095;
b) S.L. Coon, P.H. Roseboom, R. Baler, J.L. Weller, M.A. Namboodiri, E.V. Koonin, D.C. Klein,
Science 270 (1995) 1681–1683;
(
(
(
(
c) J. Borjigin, M.M. Wang, S.H. Snyder, Nature 378 (1995) 783–785;
d) D.C. Klein, P.H. Roseboom, S.L. Coon, Trends Endocrinol. Metab. 7 (1996) 106–112;
e) D.C. Klein, S.L. Coon, P.H. Roseboom, J.L. Weller, M. Bernard, J.A. Gastel, M. Zatz, P.M.
Iuvone, I.R. Rodriguez, V. Begay, J. Falcon, G.M. Cahill, V.M. Cassone, R. Baler, Recent Prog.
Horm. Res. 52 (1997) 307–357.
2] (a) W. Zheng, P.A. Cole, Curr. Med. Chem. 9 (2002) 1187–1199;
[
[
(
(
b) K.A. Scheibner, J. De Angelis, S.K. Burley, P.A. Cole, J. Biol. Chem. 277 (2002) 18118–18126;
c) N. Beaurain, C. Mesangeau, P. Chavatte, G. Ferry, V. Audinot, J.A. Boutin, P. Delagrange,
C. Bennejean, S. Yous, J. Enzyme Inhib. Med. Chem. 17 (2002) 409–414.
3] (a) D.E. Sterner, S.L. Berger, Microbiol. Mol. Biol. Rev. 64 (2000) 435–459;
(b) F. Javier Teran, M. Alvarez, J.E. Suarez, M.C. Mendoza, J. Antimicrob. Chemotherapy 28 (1991)
333–346.
[
[
4] J. De Angelis, J. Gastel, D.C. Klein, P.A. Cole, J. Biol. Chem. 273 (1998) 3045–3050.
5] (a) E.M. Khalil, P.A. Cole, J. Am. Chem. Soc. 120 (1998) 6195–6196;
(b) C.M. Kim, P.A. Cole, J. Med. Chem. 44 (2001) 2479–2485.
[
6] (a) A.B. Hickman, D.C. Klein, F. Dyda, Mol. Cell 3 (1999) 23–32;
(
(
b) A.B. Hickman, M.A. Namboodiri, D.C. Klein, F. Dyda, Cell 97 (1999) 361–369;
c) E. Wolf, J. De Angelis, E.M. Khalil, P.A. Cole, S.K. Burley, J. Mol. Biol. 317 (2002) 215–224.
[
[
[
7] A.H. Jackson, B. Naidoo, Tetrahedron 25 (1969) 4843–4852.
8] E. Benghiat, P.A. Crooks, J. Med. Chem. 26 (1983) 1470–1477.
9] E.M. Khalil, J. De Angelis, M. Ishii, P.A. Cole, Proc. Natl. Acad. Sci. USA 96 (1999) 12418–12423.
[
[
10] E.M. Khalil, J. De Angelis, P.A. Cole, J. Biol. Chem. 273 (1998) 30321–30327.
11] A.R. Fersht, J.P. Shi, J. Knill-Jones, D.M. Lowe, A.J. Wilkinson, D.M. Blow, P. Brick, P. Carter,
M.M. Waye, G. Winter, Nature 314 (1985) 235–238.

W. Zheng, P.A. Cole / Bioorganic Chemistry 31 (2003) 398–411
411
[
12] Even though the linker of analog 7 is slightly longer than that of 4, it is reasonable to expect that this
linker length variation per se does not contribute to the observed binding affinity difference between 7
and 4 based on our previous finding that elongation of the linker in 1 by one methylene group (giving
analog 2) has minimal impact on binding affinity [9]
[
13] Merck and Co., Merck Index, nineth ed., Rahway, NJ (1976) 1095.