Stereoselective synthesis of a conformationally defined cyclohexyl carnitine analogue that binds CPT-1 with high affinity

Article abstract of DOI:10.1016/S0968-0896(99)00080-2

Carnitine (1, 3-hydroxy-4-trimethylammoniobutyrate) is important in mammalian tissue as a carrier of acyl groups. In order to explore the binding requirements of the carnitine acyltransferases for carnitine, we designed conformationally defined cy- clohexyl carnitine analogues. These diastereomers contain the required gauche conformation between the trimethylammonium and hydroxy groups but vary the conformation between the hydroxy and carboxylic acid groups. Here we describe the synthesis and biological activity of the all-trans diastereomer (2), which was prepared by the ring opening of trans-methyl 2,3-epoxycylohex- anecarboxylate with NaN₃ . Racemic 2 was a competitive inhibitor of neonatal rat cardiac myocyte CPT-1 (K(i) 0.5 mMfor racemic 2; K(m) 0.2 mM for L-carnitine) and a noncompetitive inhibitor of neonatal rat cardiac myocyte CPT-2 (K(i) 0.67 mM). These results suggest that 2 represents the bound conformation of carnitine for CPT-1. (C) 1999 Published by Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0968-0896(99)00080-2

Bioorganic & Medicinal Chemistry 7 (1999) 1505±1511
Stereoselective Synthesis of a Conformationally De®ned
Cyclohexyl Carnitine Analogue that Binds CPT-1 with High
Anity
Tracy L. Hutchison, a Ashraf Saeed, a Paul E. Wolkowicz, b Jeanie B. McMillin c
and Wayne J. Brouillette a,*
aDepartment of Chemistry, University of Alabama at Birmingham, Birmingham, AL 35294, USA
bDepartment of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA
cDepartment of Pathology and Laboratory Medicine, University of Texas Health Science Center, Houston, TX 77030, USA
Received 8 September 1998; accepted 14 December 1998
AbstractÐCarnitine (1, 3-hydroxy-4-trimethylammoniobutyrate) is important in mammalian tissue as a carrier of acyl groups. In
order to explore the binding requirements of the carnitine acyltransferases for carnitine, we designed conformationally de®ned cy-
clohexyl carnitine analogues. These diastereomers contain the required gauche conformation between the trimethylammonium and
hydroxy groups but vary the conformation between the hydroxy and carboxylic acid groups. Here we describe the synthesis and
biological activity of the all-trans diastereomer (2), which was prepared by the ring opening of trans-methyl 2,3-epoxycylohex-
anecarboxylate with NaN3. Racemic 2 was a competitive inhibitor of neonatal rat cardiac myocyte CPT-1 (Ki 0.5 mM for racemic 2; Km
0.2 mM for l-carnitine) and a noncompetitive inhibitor of neonatal rat cardiac myocyte CPT-2 (Ki 0.67 mM). These results suggest
that 2 represents the bound conformation of carnitine for CPT-1. # 1999 Published by Elsevier Science Ltd. All rights reserved.
Introduction
the mitochondria and peroxisomes. Carnitine octanoyl-
transferase (COT) is selective for transferring medium
chain acyl groups and is found in peroxisomes and to a
lesser extent in microsomes. It is believed that COT is
responsible for shuttling medium chain acylcarnitines
out of peroxisomes. Carnitine palmitoyltransferases 1
and 2 (CPT-1 and CPT-2) are responsible for the trans-
fer of long chain acyl groups and are essential for the
entry of fatty acids into mitochondria prior to b-oxida-
tion.7±10 Recently, CPT-1 has been found to exist in two
isoforms with di€erent physical and kinetic properties,
liver CPT-1 (L-CPT-1) and skeletal muscle CPT-1 (M-
CPT-1). Rat heart mitochondria contain both isoforms
of CPT-1, and the minor component is identical to L-
CPT-1.
Carnitine is widely distributed in nature, and research
over the last three decades has provided a better under-
standing of the biochemical roles for carnitine.1 These
include mitochondrial long-chain fatty acid oxidation,
bu€ering the mitochondrial acyl CoA/CoA couple, a
scavenger system for acyl groups, and branched-chain
amino acid metabolism. Most of carnitine's functions
involve the transfer of an acyl residue from an acyl-CoA
to the b-hydroxy group followed by translocation
from one cellular compartment to another.2±4 In this
process, R-carnitine serves as a substrate for a reversible
reaction catalyzed by the carnitine acyltransferases, a
class of enzymes which di€er in their chain length
speci®city (see Fig. 1). Carnitine acetyltransferase (CAT)
is important for the transfer of short chain acyl
groups, particularly acetate, and for maintaining the
CoASH/acetyl-SCoA ratio.5,6 CAT is present in both
In order to probe potential di€erences among the car-
nitine binding sites on the di€erent carnitine acyl-
transferases, we have designed rigid carnitine analogues
2±5, which utilize a cyclohexyl framework. Because the
bulky trimethylammonium group locks the cyclohexane
ring in a single chair conformation, these carnitine ana-
logues contain de®ned spatial relationships between the
quaternary ammonium, hydroxyl, and carboxylate
moieties. The syntheses and biological evaluations
Key words: Carnitine; carnitine acyltransferase; enzyme inhibition;
conformationally de®ned analogue; bound conformation.
* Corresponding author. Tel.: +1-205-934-4747; fax: +1-205-934-
2543.
0968-0896/99/$ - see front matter # 1999 Published by Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00080-2

1506
T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
Figure 1. Reaction of carnitine with an Acyl-CoA.
(CAT and CPT-2) for 3±5 have been previously descri-
bed.11 Here we report the stereoselective synthesis of 2
and compare the e€ects of 2±5 on neonatal rat cardiac
myocyte CPT-1 and CPT-2 and pigeon breast CAT.
the active conformation of carnitine for binding to the
carnitine acyltransferases. However, the conformation
of carnitine about C2±C3 involved in binding to the
carnitine acyltransferases is not yet known. To address
this question, our laboratory has designed a series of
rigid carnitine analogues (2±5 in Fig. 2) which utilize a
cyclohexyl ring to control the stereochemistry. These
carnitine analogues all maintain the gauche relationship
between the hydroxy and trimethylammonium groups
(C3±C4) while varying the stereochemistry between the
hydroxy and carboxylic acid groups (C2±C3). Com-
pounds 3±5 (Fig. 2) each mimic one of the three low
energy conformations of carnitine about C2±C3, and
their preparation and biological evaluations were pre-
viously described.11 Although 3 and 5 were selective
inhibitors of CAT over CPT-2 (CPT-1 was not eval-
uated), 3±5 were only low mM inhibitors of carnitine
acyltransferases. We therefore pursued compound 2.
Compound 2, like analogue 3, mimics conformation A
about C2±C3 but, unlike 3, contains all of the func-
tional groups in equatorial positions. The evaluation of
2 was desirable since, relative to 3, it contains the cy-
clohexyl ring residues in a di€erent region of space and
Results and Discussion
Chemistry
We have been interested in evaluating di€erences in
carnitine binding requirement among the di€erent car-
nitine acyltransferases.11±17 Gandour et al. previously
identi®ed the relative populations for the three possible
low energy conformations of carnitine about C2±C3 in
1
solution using H NMR techniques (1(A)±1(C) in Fig.
2).18 It was found that all three low-energy conforma-
tions about C2±C3 are signi®cantly populated, although
two of these were dominant (Fig. 2). About C3±C4,
carnitine was reported to exist almost exclusively in a
conformation, which places the hydroxy and bulky tri-
methylammonium group in a gauche relationship. In fact,
this conformation about C3±C4 was suggested to be
Figure 2. Solution conformations for carnitine (1(A)±(C)) and their relative populations (%) as compared to the conformationally de®ned cyclo-
hexyl carnitine analogues 2±5. Compounds 2±5 all maintain the gauche arrangement about C3±C4 while varying the conformation about C2±C3.
Note that compounds 2 and 3 both mimic conformation 1(A) of carnitine but contain functional groups in di€erent axial versus equatorial
arrangements.

T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
1507
may more readily present key functional groups to the
carnitine binding site. Such information may prove
useful for designing selective inhibitors of individual
carnitine acyltransferases as therapeutic agents.
separated by HPLC to give pure 11. Azide 11 was
reduced with H2 and Pd/C to yield the acetate salt 13.22
Reductive alkylation of 13 under Eschweiler±Clarke
conditions23 provided dimethylamine 14. Quaterniza-
tion with chloromethane followed by hydrolysis of the
ester provided the ®nal product 2.
Since the synthetic approaches previously published for
3±5 could not provide usable quantities of compound 2,
an alternative approach to this compound was developed.
Compound 2 was synthesized from 3-bromocyclo-
hexene (6) as shown in Scheme 1. The conversion of 6
to 7 has been reported by Davies and Whitman19 using
sodium cyanide in N-methyl-2-pyrrolidinone. We
employed a variation of this method and found that
bromide 6 could be cleanly converted to the nitrile 7 by
stirring with KCN and 18-crown-6 in CH2Cl2 for
several days. Compounds 8±10 were then prepared
according to literature procedures.19,20 Ester 8 was
obtained by treating the nitrile 7 with methanolic HCl,
and 8 underwent stereoselective epoxidation with per-
acetic acid. The desired trans-epoxide 9 was obtained in
a ratio of 3:2 over the cis-epoxide 10, which was deter-
mined by comparing the integrations of the C2 protons
The relative con®guration of 2 was determined from
1
coupling constants between H2, H3, and H4 using H
NMR. Compound 2 is locked in a single chair con-
formation since the bulky trimethylammonium group
can only occupy the equatorial position. This places H4
axial, which is used as the reference proton in deter-
mining the relative con®guration. The residual splittings
between H2, H3, and H4 were determined by single
frequency o€-resonance decoupling experiments (recor-
ded at 400 MHz).
The chemical shifts for H2 (2.52), H3 (3.38), and H4
(4.08) were initially assigned by comparison to diaster-
1
eomers 3±5. The H NMR spectrum indicated that H2,
H3, and H4 were all axial based upon the following
observations. The presence of a pseudo triplet (over-
lapping doublet of doublets) for H3 with large coupling
constants (10.0 and 10.1 Hz) indicated two axial±axial
interactions. These assignments were con®rmed by
decoupling experiments. Irradiation of H2 resulted in the
collapse of H3 into a doublet (residual J3,4=10.1 Hz). A
similar result was obtained by irradiating H4 causing
collapse of H3 into a doublet (residual J2,3=10.0 Hz).
This con®rmed that H3 has axial±axial couplings to
both H2 and H4. The irradiation of H3 collapsed both
H2 and H4 to doublets of doublets. Each doublet of
doublets contained large splittings (J2,7a=12.1 Hz,
J4,5a=12.0 Hz), as well as small splittings (J2,7e=3.7 Hz,
J4,5e=3.8 Hz). This shows that H2 and H4 both have
axial±axial coupling, as well as axial±equatorial cou-
pling, con®rming that H2, H3, and H4 are all axial.
1
by H NMR. These diastereomers were then separated
by ¯ash chromatography.
Attempts were made to improve the stereoselectivity of
epoxidation of compound 8. It was thought that a
bulkier ester might increase the selectivity of epoxida-
tion by providing steric hindrance to the `cis' face of the
ring. The methyl ester of 8 was changed to a t-butyl
ester and epoxidized with mCPBA and peracetic acid at
40, 0, and 25 ꢀC. In all cases, the selectivity of the
epoxidation was not improved.
Following a procedure reported by Swift and Swern for
the conversion of 1,2-epoxycyclohexane to trans-2-azi-
docyclohexanol,21 epoxide 9 underwent ring opening
with NaN3 to give a 5:4 mixture of 11 and 12, which was
Scheme 1.

1508
T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
Enzyme kinetics assays
Conclusions
Racemic 2 was evaluated as an inhibitor of puri®ed
pigeon breast carnitine acetyltransferase (CAT) in the
forward direction (acylcarnitine formation), as well as
mitochondrial carnitine palmitoyltransferases 1 and 2
(CPT-1 and CPT-2) in cultured neonatal rat cardiac
myocytes (in the direction of acylcarnitine formation)
(Figs. 3 and 4). The e€ects of 3±5 on CPT-2 and CAT
were previously reported, and for comparison purposes
in the present study these were also assayed with CPT-1
(Figs. 5 and 6). As reported earlier,11 racemic 3 and 5
are modest inhibitors of CAT but do not inhibit CPT-2.
Studies performed here with CPT-1 (Table 1) reveal that
5 is a weak competitive inhibitor, while compound 3 is
speci®c for the inhibition of CAT. In contrast to 3±5,
racemic 2 did not inhibit CAT. However, 2 was an
e€ective noncompetitive inhibitor of CPT-2 (Ki=
0.67 mM). The only competitive inhibition for 2 was
observed with CPT-1, and the Ki for racemic 2 (0.5 mM)
was similar to the Km observed for l-carnitine
(0.2 mM). This is the most potent inhibition observed
for 2±5 and suggests that the carnitine conformation in
2 may represent the bound conformation of carnitine
with CPT-1. These results also suggest that properly
designed carnitine analogues may be useful as selective
small molecule inhibitors of carnitine acyltransferases.
While all of the rigid cyclohexyl carnitine analogues
contain the same hydroxy to trimethylammonium (O to
N) distances, the hydroxy to carboxy distances vary
considerably. The results from evaluation of these com-
pounds as inhibitors of carnitine acyltransferases (Table
1) permit formulation of the following structural
hypotheses. (1) The poor binding of cyclohexyl carnitine
analogues to CAT and CPT-2 may result from inter-
ference by the cyclohexyl ring residues. (2) The biologi-
cally active conformation of carnitine for CPT-1 is
contained in both rigid analogues 2 and 3, but 3 con-
tains the cyclohexyl ring residues in a region of space
that interferes with binding. (3) The carnitine binding
site of CPT-1 is distinctly di€erent from that for CAT
and CPT-2.
Experimental
General
Melting points were determined on an Electrothermal
melting point apparatus and are uncorrected. The 1H
NMR spectra were recorded on a Bruker DRX-400
(400 MHz) or a Bruker ARX-300 (300 MHz) spectrometer.
The 13C NMR spectra were recorded on a Bruker
Figure 3. Lineweaver±Burke plot of compound 2 and CPT-1, where
S=l-carnitine (concentration=M).
Figure 5. Lineweaver±Burke plot of compound 4 and CPT-1, where
S=l-carnitine (concentration=M).
Figure 4. Lineweaver±Burke plot of compound 2 and CPT-2, where
S=l-carnitine (concentration=M).
Figure 6. Lineweaver±Burke plot of compound 5 and CPT-1, where
S=l-carnitine (concentration=M).

T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
1509
Table 1. Kinetic constants for conformationally de®ned carnitine analogues with selected carnitine acyltransferasesa
Compound
Ki, mM, CPT-1b (Type Inhib)
Ki, mM, CPT-2b (Type Inhib)
Ki, mM, CATc (Type Inhib)
(R,S)-2
(R,S)-3
(R,S)-4
(R,S)-5
0.5 (Competitive)
No E€ect
9.5 (Noncompetitive)
9.3 (Competitive)
0.67 (Noncompetitive)
No E€ectd
No E€ect
4.1 (Competitive)d
3.7 (Competitive)d
2.9 (Competitive)
No E€ectd
5.3 (Competitive)d
a
All assays were in the forward direction (acylcarnitine formation) and used varying concentration of l-carnitine (Km 0.2 mM for CPT-1 and
0.2 mM for CPT-2).
Assayed in cultured neonatal rat cardiac myocytes.
Puri®ed pigeon breast CAT was utilized.
Taken from ref 11.
b
c
d
DRX-400 (100 MHz) or a Bruker ARX-300 (75 MHz).
Chemical shifts (ppm) are referenced to internal TMS
1.9±2.0 (br s, 2H), 2.9±3.0 (m, 1H), 3.7 (s, 3H), 5.6±5.7
and 5.75±5.8 (m, 2H).
1
for H and 13C NMR spectra, when recorded in CDCl3
solvent, and samples recorded in D2O were referenced
to either H2O (1H NMR) or dioxane (13C NMR). IR
spectra were obtained on a Bruker Vector 22 FT-IR or
a Perkin±Elmer 1310 IR. Mass spectra were recorded by
electrospray on a Perkin±Elmer Sciex API-III spectro-
meter. Flash chromatographic separations were per-
formed on Baker silica gel, 40 mm, and TLC was
performed on Whatman brand silica gel plates (250 mm
layer, 5Â10 cm). HPLC separations utilized a Rainin
Dynamax silica column on a Rainin Rabbit HPLC sys-
tem. All liquid reagents were distilled prior to use.
Methanol was distilled over magnesium methoxide
and methylene chloride was distilled over phosphorus
pentoxide. Acetone was dried by distilling over potas-
sium permanganate and then over anhydrous calcium
sulfate. All other solvents were reagent grade, obtained
from Fisher Scienti®c or Aldrich Chemical Company,
and used without further puri®cation. Elemental ana-
lyses were performed at Atlantic Microlab of Atlanta,
Georgia.
Methyl trans-2,3-epoxycyclohexanecarboxylate (9) and
methyl cis-2,3-epoxycyclohexanecarboxylate (10). To a
solution of 8 (16.6 g, 118 mmol) in CHCl3 (250 mL) was
added 32% peracetic acid (90.0 mL, 28.8 g, 427 mmol)
and anhydrous sodium acetate (16.7 g, 203 mmol). The
reaction mixture was stirred at 25ꢀC for 16 h. The mix-
ture was extracted with CHCl3 (4Â50 mL), the com-
bined organic extracts were washed with 5% NaHCO3
(4Â100 mL), and the organic layer was dried (Na2SO4).
The solvent was removed in vacuo to give a mixture of
the cis and trans epoxyesters (17.1 g, 92.9%). Integra-
1
tion of the C2 proton resonances in the H NMR spec-
trum gave the ratio of trans (C2 resonance at 3.4 ppm)
to cis (C2 resonance at 3.45 ppm) as 3:2. Part of the
mixture (5.0 g) was chromatographed (33% ether/hex-
anes) on a ¯ash silica column (5Â25 cm) to ®rst provide
9 as an oil (2.3 g, Rf 0.47) followed by a mixture of 9 and
10 (0.70 g). Further elution gave pure 10 as an oil (1.5 g,
1
Rf 0.31). The H NMR spectra for 9 and 10 were iden-
tical to those reported by Davies and Whitman.20
Cyclohex-2-enecarbonitrile (7). Dry potassium cyanide
(34.5 g, 530 mmol) and 18-crown-6 (600 mg, 2.20 mmol)
were added to a solution of 3-bromocyclohexene (6,
30.0 g, 186 mmol) in dry CH2Cl2 (90 mL). The reaction
mixture was stirred at 25ꢀC for 4 days. The unreacted
KCN was ®ltered and washed with CH2Cl2. The solvent
was removed from the ®ltrate on a rotary evaporator to
give 7 (19.4 g, 97.2%) as an oil: bp 83±84ꢀC/20 mm Hg
(lit.13 bp 89ꢀC/23 mm Hg).
For 9: 1H NMR (CDCl3) d 1.3±1.5 (m, 3H), 1.7±1.9 (m,
2H), 2.0±2.1 (m, 1H), 2.9 (dd, 1H), 3.2 (m, 1H), 3.4 (d,
1H), 3.7 (s, 3H).
For 10: 1H NMR (CDCl3) d 1.20±1.94 (m, 6H), 2.81±2.89
(m, 1H), 3.19±3.25 (m, 1H), 3.45 (t, 1H), 3.78 (s, 3H).
Methyl 3-azido-2-hydroxycyclohexanecarboxylate (11)
and methyl 2-azido-3-hydroxycyclohexanecarboxylate
(12). Compound 9 (3.50 g, 22.7 mmol) was added to a
¯ame dried round bottom ¯ask containing dry metha-
nol (75 mL) under a nitrogen atmosphere. Dry NaN3
(2.58 g, 39.7 mmol) and NH4Cl (1.83 g, 34.2 mmol) were
added to the reaction mixture. This mixture was heated
to re¯ux for 14 h. The reaction mixture was cooled to
room temperature and the solvent was removed, leaving
an oil. The oil was taken up in water (75 mL) and
extracted with ethyl acetate (3Â50 mL). The combined
organic extracts were washed with saturated NaCl
(2Â50 mL) and dried (Na2SO4). Integration of the
methyl ester resonances in the 1H NMR spectrum
revealed a 1:1.1 ratio of 11:12. The mixture was ®ltered
through a bed of silica gel on a sintered glass funnel
and concentrated in vacuo to yield a mixture of 11 and
12 (3.65 g, 81.8%). This mixture was separated on a
Methyl cyclohex-2-enecarboxylate (8). Anhydrous
hydrogen chloride was bubbled into a re¯uxing mixture
of 7 (8.60 g, 80.3 mmol) in dry methanol (100 mL, 79.1 g,
2.47 moles) for 60 min. The reaction mixture was heated
at re¯ux for an additional 60 min and stirred at room
temperature for 22 h. The reaction mixture was poured
over ice (300 g) and stirred until the ice had melted. The
mixture was extracted with ether (4Â50 mL). The com-
bined ether layers were washed with 5% NaHCO3
(3Â50 mL) and dried (Na2SO4), and the solvent was
removed in vacuo to give 8 as an oil (Rf 0.45, 30% ethyl
acetate/hexanes). The 1H NMR spectrum was iden-
tical to that reported by Davies and Whitham.19 This
product was carried on without further puri®cation.
1H NMR (CDCl3) d 1.4±1.6 (m, 1H), 1.7±1.9 (m, 3H),

1510
T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
Rainin Dynamax silica HPLC column (21.4Â250 mm,
60 A, 20% ethyl acetate/hexane, 489 mg/injection as a
54% ethyl acetate stock solution) using a ¯ow rate of
5 mL/min to give 11 (1.47 g, 32.9%; tR=57.6 min). IR
1
a pear-shaped pressure bottle and cooled to 78ꢀC.
Chloromethane was bubbled through the solution for
10 min. The bottle was capped and warmed to 30ꢀC for
5 days. The solid was ®ltered and washed with anhy-
drous acetone (5.0 mL) to give 15 (183 mg, 58.7%): mp
217±218ꢀC (ethyl acetate/hexanes). IR (KBr) 3250±3100
(neat) 3700±3000 (OH), 2100 (N3), 1710 (CˆO) cm
;
1H NMR (CDCl3) d 1.27±1.45 (m, 3H), 1.80±1.85 (m,
1H), 1.95±2.10 (m, 2H), 2.36±2.42 (m, 1H), 3.15 (d, 1H),
3.25±3.32 (m, 1H), 3.65±3.76 (dd, 1H), 3.73 (s, 3H); 13C
NMR (CDCl3) d 174.24, 74.28, 65.13, 51.99, 49.58,
29.80, 27.64, 23.31; MS (ES) m/z 200 (M+H)+. Anal.
calcd for C8H13N3O3: C, 48.26; H, 6.60; N, 21.10.
Found: C, 48.01; H, 6.56; N, 20.86.
1
(OH), 1736 (CˆO) cm 1; H NMR (D2O) d 1.30±1.55
(m, 3H), 1.79±1.90 (m, 2H), 2.20±2.28 (m, 1H), 2.52±
2.62 (m, 1H, J2±7e=3.48 Hz, J2±7a=11.05 Hz), 3.11 (s,
9H), 3.34±3.45 (m, 1H, J4±5e=3.79 Hz, J4±5a=11.94Hz),
3.64 (s, 3H, J2±3=J3±4=10.18 Hz), 4.08 (dd, 1H); 13C
NMR (D2O with 0.01% dioxane) d 175.76, 75.94, 70.77,
53.22, 52.65, 52.29, 27.51, 25.26, 23.09; MS (ES) m/
z=216 (M+H)+. Anal. calcd for C9H22ClNO3: C,
52.56; H, 8.84; N, 5.57; Cl, 13.93. Found: C, 52.36; H,
8.66; N, 5.46; Cl, 13.74.
For 12 (1.38 g, 30.9%; tR=74.2 min): IR (neat) 3700±
;
1H NMR
1
3000 (OH), 2100 (N3), 1720 (CˆO) cm
(CDCl3) d 1.45±1.90 (m, 6H), 2.26 (s, 1H), 2.95±3.05 (m,
1H), 3.72 (s, 3H), 3.83 (t, 1H), 4.13±4.19 (br s, 1H); 13C
NMR (CDCl3) d 173.4, 68.35, 64.38, 51.80, 42.09, 29.31,
23.94, 19.01; MS (ES) m/z 200 (M+H)+. Anal. calcd
for C8H13N3O3: C, 48.26; H, 6.60; N, 21.10. Found: C,
48.05; H, 6.56; N, 20.85.
2-Hydroxy-3-(trimethylammonio)cyclohexanecarboxylic
acid, chloride (2). A solution of 15 (100 mg, 398 mmol)
in 17% HCl (4.0 mL) was heated to re¯ux for 1.5 h. The
reaction mixture was cooled to rt, concentrated to dry-
ness in vacuo, and dried overnight under high vacuum
to yield 2 as a white solid (90.9 mg, 96.0%): mp 207±
Methyl 3-ammonio-2-hydroxycyclohexanecarboxylate, ace-
tate (13). Compound 11 (1.00 g, 5.02 mmol), methanol
(25.0 mL), acetic acid (10.7 mL), and 10% Pd/C (0.41 g)
were combined in a hydrogenation pressure bottle. The
reaction mixture was shaken on a Parr shaker (50 psi
H2) for 1 h. The catalyst was ®ltered and washed with
hot deionized water. The solvent was removed from the
®ltrate and dried under vacuum to yield 13 (1.07 g,
91.5%) as a white solid: mp 156±158ꢀC (ethyl acetate);
IR (KBr) 3500±2250 (OH), 1724 (CˆO) cm 1; 1H NMR
(D2O) d 1.34±1.46 (m, 3H), 1.74±2.22 (m, 2H), 1.81 (s,
3H), 2.43±2.49 (m, 1H), 2.98±3.14 (m, 1H), 3.62±3.68 (dd,
1H), 3.65 (s, 1H); 13C NMR (D2O) d 183.75, 178.93,
74.31, 57.79, 55.44, 53.12, 31.22, 30.56, 25.80, 25.21; MS
(ES) m/z 174 (M+H)+. Anal. calcd for C10H19NO5: C,
51.52; H, 8.23; N, 6.01. Found: C, 51.49; H, 8.18; N, 5.89.
208ꢀC (methanol/ether); IR (KBr) 3650±3100 (OH),
1
1718 (CˆO) cm
;
1H NMR (D2O) d 1.41±1.54 (m,
3H), 1.89±1.95 (m, 2H), 2.17±2.29 (m, 1H), 2.53±2.60
(m, 1H, J2±7a=12.11 Hz, J2±7e=3.70 Hz), 3.17 (s, 9H),
3.40±3.46 (m, 1H, J4-5a=11.95 Hz, J4±5e=3.77 Hz), 4.11
(dd, 1H, J3±4=J2±3=10.11 Hz); 13C NMR (D2O with
0.01% dioxane) d 177.09, 75.83, 70.55, 53.10, 52.25,
27.61, 25.17, 23.09; MS (ES) m/z=202 (M+H). Anal.
calcd for C10H20ClNO3: C, 50.52; H, 8.50; Cl, 14.91; N,
5.89. Found: C, 50.36; H, 8.41; Cl, 14.75; N, 5.68.
Cultured neonatal rat cardiac myocyte CPT-1 and CPT-
2 assays. The details of these procedures have been pre-
viously published.17 Brie¯y, neonatal cardiac myocytes
were isolated and cultured in 35 mm, 12-well plates
(2Â105 cells/well) as described by McMillin et al.24 For
the CPT-1 assay, cell membranes were permeabilized
with 10 mM digitonin (10 min at 37ꢀC). After removal
of the permeabilization medium, 0.5 mL of CPT assay
medium bu€ered with 10 mM HEPES, pH 7.0, and
containing 1% BSA, was added to each well. Palmitoyl-
CoA was added to give a ®nal concentration of 30 mM.
The assay was initiated by adding increasing con-
centrations of [14C]carnitine (speci®c activity=2000±
3000 dpm/nmol) to a series of cell cultures (®nal con-
centrations varied from 0.08 to 0.40 mM). The rates of
[14C]palmitoylcarnitine formation were linear for
40 min. CPT-2 assays were carried out separately in a
similar manner, except the mitochondrial membranes of
the cultured cells were permeabilized with 0.16% Triton
X-100 to inactivate CPT-1 activity and to provide latent
CPT-2 activity. The permeabilized medium was
removed from each well and added to a solution
(0.5 mL) containing 10 mM HEPES (pH 7.0) and 1%
BSA. The ®nal concentration of palmitoyl-CoA was 75
mM. Increasing concentrations of l-[14C]carnitine (spe-
ci®c activity=2000±3000 dpm/nmol) were then added to
a series of cell cultures (0.08±0.40 mM ®nal concentra-
tion) to initiate the reaction. The initial rates, which
were linear, were used to determine CPT-2 activity. The
Methyl 3-(N,N-dimethylamino)-2-hydroxycylohexane-
carboxylate (14). Compound 13 (580 mg, 2.49 mmol)
was dissolved in deionized water (22.0 mL) and added to
a hydrogenation pressure bottle. Formaldehyde (37%,
1.70 mL, 629 mg, 22.7 mmol) and 10% Pd/C (300 mg)
were added to the reaction mixture, which was then
shaken on a Parr shaker (50 psi H2) for 18 h. The cata-
lyst was ®ltered and washed with hot deionized water.
The ®ltrate was allowed to cool to room temperature,
saturated with potassium carbonate, and extracted with
CHCl3 (3Â50 mL). The combined organic layers were
dried (Na2SO4) and concentrated in vacuo to give 14 as
a white solid (455 mg, 91.0%): mp 50±52ꢀC (ether/hex-
anes). IR (KBr) 3550±3100 (OH), 1736 (CˆO) cm 1; 1H
NMR (CDCl3) d 1.11±1.53 (m, 3H), 1.72±1.88 (m, 3H),
2.21±2.42 (m, 2H), 2.26 (s, 6H), 3.58 (dd, 1H), 3.73 (s,
3H); 13C NMR (CDCl3) d 176.86, 72.09, 70.17, 53.77,
52.25, 42.22, 30.06, 26.20, 22.00; MS (ES) m/z 202
(M+H)+. Anal. calcd for C10H19NO3: C, 59.68; H,
9.52; N, 6.96. Found: C, 59.51; H, 9.41; N, 7.04.
Methyl 2-hydroxy-3-(trimethylammonio)cyclohexane-
carboxylate, chloride (15). A solution of 14 (250 mg,
1.24 mmol) in anhydrous acetone (5.0 mL) was added to

T. L. Hutchison et al. / Bioorg. Med. Chem. 7 (1999) 1505±1511
1511
Ki values of 2 for CPT-1 or CPT-2 were determined by
adding inhibitor to a ®nal concentration of 1 mM in
each well. After gentle shaking at 37ꢀC for 20 min, the
reaction was quenched with butanol-saturated 0.73 M
HCl. The amount of radioactive product in the butanol
extract was determined by scintillation counting.
11. Brouillette, W. J.; Saeed, A.; Abuelyaman, A.; Hutchison,
T. L.; Wolkowicz, P. E.; McMillin, J. B. J. Org. Chem. 1994,
59, 4297.
12. Saeed, A.; McMillin, J. B.; Wolkowicz, P. E.; Brouillette,
W. J. J. Med. Chem. 1994, 37, 3247.
13. Brouillette, W. J.; Puckett, S. W.; Comber, R. N.; Hosmer,
C. A. Biochem. Pharmacol. 1993, 46, 1671.
14. Brouillette, W. J.; Abuelyaman, A. S.; Hosmer, C. A.; Com-
ber, R. N.; Saeed, A. Org. Prep. Proced. Int. 1993, 25, 243.
15. Comber, R. N.; Brouillette, W. J. J. Org. Chem. 1987, 52,
2311.
References
16. Hosmer, C. A.; Comber, R. N.; Brouillette, W. J. J. Org.
Chem. 1985, 50, 3627.
17. Saeed, A.; McMillin, J. B.; Wolkowicz, P. E.; Brouillette,
W. J. Arch. Biochem. Biophys. 1993, 305, 307.
18. Colucci, W. J.; Gandour, R. D.; Mooberry, E. A. J. Am.
Chem. Soc. 1986, 108, 7141.
1. Sewell, A. C.; Bohles, H. J. Eur. J. Pediatr. 1995, 154, 871.
2. Bieber, L. L. Ann. Rev. Biochem. 1988, 57, 261.
3. Bremer, J. Phys. Rev. 1983, 63, 1420.
4. Bieber, L. L.; Emaus, R.; Valkner, K.; Farrell, S. Fed. Proc.
1982, 41, 2858.
5. Hoppel, C. L. In The Enzymes of Biological Membranes;
Martonosi, A., Ed., Plenum: New York, 1976; Vol. 2, pp 119±
143.
6. Colucci, W. J.; Gandour, R. D. Bioorg. Chem. 1988, 16, 307.
7. Fritz, I. B. Adv. Lipid Research 1963, 1, 285.
8. McGarry, J. D.; Foster, D. W. Annu. Rev. Biochem. 1980,
49, 395.
9. Lilly, K.; Chung, C.; Kerner, J.; Van Renterghem, R.; Bie-
ber, L. L. Biochem. Pharmacol. 1992, 43, 361.
10. Chatterjee, B.; Song, C. S.; Kim, J. M.; Roy, A. K. Bio-
chemistry 1988, 27, 9000.
19. Davies, S. G., Whitham, G. H., J. Chem. Soc., Perkin
Trans. 1 1976, 21, 2279.
20. Davies, S. G.; Whitham, G. H. J. Chem. Soc., Perkin
Trans. 1 1977, 5, 572.
21. Swift, G.; Swern, D. J. Org. Chem. 1967, 32, 511.
22. Vince, R.; Daluge, S. J. Med. Chem. 1974, 17, 578.
23. Keller, F.; Engle, R. R.; Kolhs, M. W. J. Med. Chem.
1963, 6, 202.
24. McMillin, J. B., Hudson, E. K., Buja, L. M. In Methods in
Toxicology: Mitochondrial Dysfunction, Lash, L., Jones, D.,
Eds., Academic: San Diego, 1993, Vol. 2, pp 301±309.