Investigating the role of the geminal dimethyl groups of coenzyme A: Synthesis and studies of a didemethyl analogue

Article abstract of DOI:10.1016/S0968-0896(00)00189-9

An analogue 2 of coenzyme A (CoA) has been prepared in which the geminal methyl groups are replaced with hydrogens. An NMR titration study was conducted and shifts in frequency of protons in the pantetheine portion of the molecule upon titration of the adenine base were observed as has been previously reported with CoA. These studies indicate that the geminal dimethyl groups are not essential for adoption of a partially folded conformation in solution. Based on ¹H-¹H coupling constants, the distribution of conformations about the carbon-carbon bonds in the region of the methyl deletion were estimated. The results suggest that the conformer distribution is similar to that of CoA, but with small increases in population of the anti conformers. A simple model compound containing the didemethyl pantoamide moiety was prepared and subjected to similar conformational analysis. The coupling constants and predicted conformer distribution were almost identical to that of the CoA analogue, indicating that the conformer distribution is controlled by local interactions and not influenced by interactions between distant parts of the CoA molecule. The acetyl derivative of 2 was a fairly good substrate for the acetyl-CoA utilizing enzymes carnitine acetyltransferase, chloramphenicol acetyltransferase, and citrate synthase, with 1.3- to 10-fold increased K(m) values and 2.5- to 11-fold decreases in V(max). The combined results indicate that the geminal dimethyl groups of CoA have modest effects on function and minimal effects on conformation. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00189-9

Bioorganic & Medicinal Chemistry 8 (2000) 2451±2460
Investigating the Role of the Geminal Dimethyl Groups of
Coenzyme A: Synthesis and Studies of a Didemethyl Analogue
a
a
b
b
Kurt W. Vogel, Lucy M. Stark, Pranab K. Mishra, Wei Yang
b,
and Dale G. Drueckhammer *
aDepartment of Chemistry, Stanford University, Stanford, CA 94305, USA
bDepartment of Chemistry, State University at Stony Brook, Stony Brook, NY 11794, USA
Received 24 March 2000; accepted 9 June 2000
AbstractÐAn analogue 2 of coenzyme A (CoA) has been prepared in which the geminal methyl groups are replaced with hydro-
gens. An NMR titration study was conducted and shifts in frequency of protons in the pantetheine portion of the molecule upon
titration of the adenine base were observed as has been previously reported with CoA. These studies indicate that the geminal
1
1
dimethyl groups are not essential for adoption of a partially folded conformation in solution. Based on H± H coupling constants,
the distribution of conformations about the carbon±carbon bonds in the region of the methyl deletion were estimated. The results
suggest that the conformer distribution is similar to that of CoA, but with small increases in population of the anti conformers. A
simple model compound containing the didemethyl pantoamide moiety was prepared and subjected to similar conformational
analysis. The coupling constants and predicted conformer distribution were almost identical to that of the CoA analogue, indicating
that the conformer distribution is controlled by local interactions and not in¯uenced by interactions between distant parts of the
CoA molecule. The acetyl derivative of 2 was a fairly good substrate for the acetyl-CoA utilizing enzymes carnitine acetyltransfer-
ase, chloramphenicol acetyltransferase, and citrate synthase, with 1.3- to 10-fold increased Km values and 2.5- to 11-fold decreases
in Vmax. The combined results indicate that the geminal dimethyl groups of CoA have modest e€ects on function and minimal
e€ects on conformation. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
CoA transferase but greatly stabilizes the transition
2
,3
state leading to and from this intermediate. This
results in a large reduction in the energy of activation
for this reaction. NMR has been used to study the con-
Coenzyme A (CoA) 1 and its acyl derivatives play a
central role in various aspects of metabolism, being used
1
4�7
and in enzyme
by about 4% of all enzymes. CoA is a fairly complex
formation of CoA both in solution
complexes8 and the conformation in the region of the
geminal dimethyl groups has received substantial atten-
tion. The structures of several CoA-utilizing enzymes
complexed with CoA or a CoA derivative have been
solved by X-ray crystallography10 and NMR. These
studies have shown that CoA exists both in solution and
in enzyme complexes with primarily a gauche con-
,9
molecule, though only the thiol group is modi®ed in the
reactions involved in the biological function of CoA.
The remainder of the molecule provides a handle for
recognition and binding by CoA-utilizing enzymes but
more speci®c roles for individual functionality are
unclear. One especially interesting feature of the CoA
molecule is the geminal dimethyl groups at C2 . The
toxicity of a-ketobutyrate was suggested to be due to
the incorporation of a-ketobutyrate into the pantoic
acid domain of CoA to form a disfunctional mono-
demethyl analogue, based on the observation that the
11
00
00
00
formation about the C1 ±C2 bond (see Fig. 1 for num-
bering). The geminal dimethyl groups might be expected
to induce or at least reinforce this conformation due to
the geminal-dialkyl e€ect.1
2,13
However, an analogue of
2
toxicity could be overcome with pantetheine. Jencks
CoA having these methyl groups deleted has not been
available for direct study of the in¯uence of the methyl
groups on the conformation and enzymatic activity.
and co-workers demonstrated that the functionality of the
pantoate moiety of CoA, which includes these geminal
dimethyl groups, destabilizes the covalent CoA±enzyme
intermediate in the reaction of the enzyme 3-oxoacid-
We report here the synthesis of the analogue of CoA 2
in which the geminal methyl groups are replaced with
hydrogens. An NMR titration study indicated shifts in
frequency of protons in the pantetheine portion of the
*Corresponding author. Tel.: +1-631-632-7923; fax: +1-631-632-7960;
e-mail: dale.drueckhammer@sunysb.edu
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00189-9

2
452
K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
Figure 1. Structures of coenzyme A and a didemethyl analogue.
molecule upon titration of the adenine base, as has been
4,5
previously reported with CoA. These studies indicate
that the geminal dimethyl groups are not essential for
adoption of a partially folded conformation in solution.
1
1
H± H coupling constants were measured to analyze the
conformer distribution about the carbon±carbon bonds
in the region of the methyl deletion. The results suggest
that the conformer distribution is similar to that of
CoA, but with a small increase in population of the anti
conformers. Conformational analysis of a simple model
compound containing the didemethyl pantoamide moiety
gave almost identical results to the didemethyl CoA
analogue, indicating that the conformer distribution is
controlled by local interactions. The acetyl derivative of
Scheme 1.
ATP and the ®nal two enzymes of CoA biosynthesis as
shown in Scheme 2. Prior to preparative conversion of 8
to 2, kinetic analysis of 8 as a substrate for phospho-
pantetheine adenylyltransferase was performed in a
coupled enzyme assay with pyrophosphate-dependent
phosphofructokinase.21 The Vmax for racemic 8 was
almost 6-fold lower than that for the natural substrate
phosphopantetheine, and the Km value of 5.7 mM was
73-fold higher than that for the natural substrate. Kinetic
analysis of the product didemethyldephospho-CoA with
the ®nal enzyme of CoA biosynthesis, dephospho-CoA
kinase was not performed. The enzymes of Scheme 2
were isolated from Corynebacterium (formerly Brevi-
2
was a fairly good substrate for three acetyl-CoA uti-
lizing enzymes. The role of the geminal dimethyl groups
in CoA is discussed in light of these results.
Results
Synthesis of a didemethyl analogue of CoA
bacterium) ammoniagenes and immobilized in poly-
The ®rst
acrylamide gel as described previously.2
2,23
enzyme, phosphopantetheine adenylyltransferase cata-
The CoA analogue 2 having the geminal pair of methyl
groups replaced with hydrogens was prepared using the
last two enzymes of CoA biosynthesis to perform the ®nal
steps. The didemethyl analogue of phosphopantetheine 8,
which served as the substrate for these ®nal two enzymatic
steps was prepared nonenzymatically as shown in Scheme
lyzed the coupling of 8 with the a-phosphate of ATP to
give the 3 -dephospho analogue of 2. Inorganic pyro-
0
phosphatase was included to hydrolyze the pyrophos-
phate formed in this reaction to make the reaction irre-
versible. Dephospho-CoA kinase catalyzed the selective
1
4
0
1
. The disul®de of N-thioethyl-3-chloropropionamide 3
1
phosphorylation of the 3 -hydroxyl group by ATP to
give the CoA analogue 2. Phosphoenolpyruvate and
5
was reacted with sodium azide in DMF followed by
reduction of the azide with triphenylphosphine in aqu-
eous THF to form the amine 4. The reaction with
triphenylphosphine also reduced the disul®de to the
thiol, which was oxidized back to the disul®de upon
exposure to air. Reaction of 4 with the racemic t-butyl-
pyruvate kinase were included to regenerate ATP from
24
1
6
17
the ADP formed in this step. The product 2 was puri®ed
by ion-exchange chromatography on DEAE cellulose,
followed by reverse-phase HPLC.
18
dimethylsilyl-protected a-hydroxy-g-butyrolactone 5
gave the protected pantetheine analogue 6. Reaction of
with dibenzyl phosphorochloridate gave the dibenzyl
NMR analysis of 2
1
6
Initial comparison of the H NMR spectrum of 2 with
that of CoA revealed di€erences in the protons of the
modi®ed pantoate moiety. The singlets at 0.71 and
1
9
phosphate triester 7. Reaction with HF removed the
t-butyldimethylsilyl protecting group and was followed
by reaction with trimethylsilyl chloride and lithium bro-
mide to convert the phosphate benzyl esters to the tri-
2
0
methylsilyl esters. Removal of the trimethylsilyl groups
upon addition of water and reduction of the disul®de with
dithiothreitol formed the phosphopantetheine analogue 8.
The didemethyl analogue of phosphopantetheine 8 was
converted to the corresponding CoA analogue 2 using
Scheme 2.

K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
2453
0
0
0.84 ppm for the diastereotopic methyl groups at C2 of
CoA are replaced with complex multiplets of the C2
decoupled spectrum showed an apparent doublet of
triplets for each proton. Again, the large coupling con-
stant is attributed to geminal coupling of about 14.4 Hz.
The remaining triplet structure for each proton indicates
that each proton has approximately equal coupling
00
protons of 2 at 1.63 and 1.96 ppm (numbering is shown
in Fig. 1). Based on coupling constants and relative
chemical shift, the resonance at 1.63 ppm was assigned
0
0
00
to the pro-S proton (H2 ) and the resonance at 1.96 to
S
constants with both C1 protons. For the proton at
0
which for CoA gives a singlet at 3.94 ppm, appeared
0
00
the pro-R proton (H2 ) (see discussion section). H3 ,
R
1.63 ppm, these coupling constants are about 4.8 Hz
while for the proton at 1.96 ppm they are about 7.3 Hz.
The coupling constants for the didemethylpantoate
moiety of 2 are summarized in Table 2.
0
near 4.1 ppm and overlapped with the C-5 protons of
0
0
the ribose moiety. The protons of C-1 , which give dis-
tinct signals at 3.80 and 3.54 ppm for CoA, appeared as
an overlapping multiplet at 3.9 ppm for 2.
Synthesis and NMR analysis of a model for the modi®ed
pantoate moiety of 2
1
H NMR spectra of 2 were obtained in D O at 13 dif-
2
ferent pD values ranging from pD 2.35 to 13.43. Pro-
tons of the pantetheine moiety exhibited changes in
chemical shift as the pD crossed the pK of the adenine
N-Benzyl 2,4-dihydroxybutanamide 10 was prepared by
reaction of benzyl amine with racemic 2-hydroxy-g-
butyrolactone 9 (Scheme 3). Similarly to 2, the protons
a
0
00
1
base (pK =4.3), the 3 -phosphate (pK =6.3) and the
a
at C2 in the H NMR spectrum of 10 were well sepa-
rated (numbering analogous to CoA). As with 2, cou-
a
thiol group (pK =9.8), as was observed previously with
a
4
,5
00
determined by decoupling the protons at C1 while
00
CoA. Table 1 shows the di€erences in chemical shift
pling constants between H3 and the C2 protons were
0
0
00
00
00
00
coupling constants of the C2 protons with the protons
of the protons at C5 , C6 , C8 and C9 on going from
pH 3.62 to 5.40, along with the previously reported
changes in chemical shift of the equivalent protons of
CoA. As with CoA, the same protons exhibited smaller
changes in chemical shift (0.003±0.007 ppm) upon titra-
0
0
00
00
at C1 were determined by decoupling H3 . The cou-
pling constants for 10 are also given in Table 2 and are
very similar to the analogous coupling constants in 2.
0
00
tion of the 3 -phosphate group and the protons at C8
0
0
and C9 exhibited additional up®eld shifts upon ioniza-
tion of the thiol group. All of these results are very
similar to the NMR titration results previously reported
Fitting of coupling constants with rotamer populations
0
0
00
For analysis of the conformation about the C1 ±C2
5
00
00
with CoA.
and C2 ±C3 bonds of 2 and 10, coupling constants for
each proton pair in each low energy staggered con-
formation were predicted for 2,4-dihydroxybutanamide-
4-phosphate using Macromodel.25 The rotamer desig-
1
1
Analysis of the H± H coupling constants in the region
of the modi®ed pantoate moiety of 2 was performed.
Analysis was hampered by the fact that the resonances
6
,7
nations as used previously are shown in Figure 2. The
predicted coupling constants between H3 and the pro-
0
0
00
S and pro-R protons at C2 are given in Table 3. Like-
for the protons at C1 overlapped with each other and
0
0
00
wise, the predicted coupling constants between each of
proton H3 overlapped with protons of the ribose ring.
0
0
However, the protons at C2 were well separated from
each other and from other protons in the molecule. In
the fully coupled spectrum, the multiplicities of these
protons were too complex to analyze, thus decoupling
00
00
the C1 protons and each of the C2 protons are given
in Table 4. The populations of each low-energy rotamer
about the C1 ±C2 and C2 ±C3 bonds which best ®t
the predicted and measured coupling constants were
then calculated, with the results shown in Table 5.
Shown for comparison are the rotamer populations for
00
00
00
00
0
0
experiments were performed. When the C1 protons
0
0
26
were decoupled, the signals for the C2 protons were
each simpli®ed to a doublet of doublets. Both exhibited
a large coupling constant (14.4 Hz), which is attributed
to the geminal coupling. The smaller splittings corre-
CoA 1 as predicted by Wu et al.7
0
0
spond to coupling by H3 and give coupling constants
0
0
of about 2.7 Hz and 9.0 Hz for the C2 protons at 1.96
and 1.63, respectively.
Table 2. Measured coupling constants for 2 and 10
2
10
0
0
00
Analysis of the C2 protons when H3 was decoupled
was made more complex by the presence of three cou-
pling constants for each proton, due to splitting by both
00
00
00
00
H2
R
H2
S
H2
R
H2
S
0
0
0
0
0
0
H1
H1
H3
S
4.8
4.8
9.0
7.3
7.3
2.7
5.5
5.5
8.5
7.3
7.3
3.6
0
0
R
C1 protons in addition to the geminal splitting. The
Table 1. Changes in chemical shift of protons of the pantetheine
moiety of 1 and 2 upon titration of the adenine base
0
0
00
00
00
H9
H5
H6
H8
a
Ád 1
Ád 2
0.034
0.035
0.035
0.032
0.031
0.022
0.021
0.019
b
a
Data from ref 5.
Di€erence in chemical shift at pD 3.62 versus pD 5.40.
b
Scheme 3.

2
454
K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
2
2
as a substrate for CoA-utilizing enzymes
was converted to the acetyl derivative by reaction with
thiophenyl acetate. The resulting acetyl-CoA analogue
was tested as a substrate for the acetyl-CoA-utilizing
enzymes carnitine acetyltransferase, chloramphenicol
acetyltransferase, and citrate synthase. The resulting Km
and Vmax values are given in Table 6. With carnitine
acetyltransferase, the Vmax was 9% of that observed
with acetyl-CoA while the Km value was increased only
slightly. With chloramphenicol acetyltransferase, the
Vmax for the analogue was about 40% that with acetyl-
CoA while the Km was increased about 8-fold. With
citrate synthase, the Vmax was 11% of that with acetyl-
CoA while the K was about 10-fold higher.
m
Discussion
Figure 2. Newman projections of the low-energy rotamers about the
00
00
00
00
C2 ±C3 and C1 ±C2 bonds of coenzyme A.
It is unclear why CoA contains the gem-dimethyl func-
tionality in the pantoate moiety, but a possible issue is
biosynthetic accessibility. The biosynthetic pathway for
0
0
00 00
S R
and H2
27
Table 3. Predicted coupling constants for H3 with H2
pantoic acid 13 (Scheme 4) illustrates that the methyl
groups originate from a-keto isovaleric acid 11, which is
derived from the transamination of valine or in bacteria
by the pathways of branched-chain amino acid bio-
synthesis.28 Reaction of 11 with methylene tetra-
hydrofolate produces ketopantoate 12 which is followed
by reduction to form 13. If pyruvate were substituted
for 11 in this pathway and the subsequent steps of CoA
biosynthesis, the product would be the didemethyl analo-
gue 2 while substitution by a-ketobutyrate would result in
Rotamer
1
2
3
0
0
0
0
H2
H2
S
11.5
1.9
2.5
3.5
3.7
11.5
R
0
0
00
00
Table 4. Predicted coupling constants for H1
0
S
and H1
R
with H2
S
0
29
and H2
R
a product with a single methyl group at this position.
a-Ketobutyrate is an intermediate in the threonine to
isoleucine biosynthetic pathway, and a byproduct in the
homocysteine/serine to cysteine pathway while pyruvic
acid plays a role in many cellular processes, thus both
Rotamer
5
4
6
0
0
00
00
00
00
00
H1
H1
S
H1
R
H1
S
H1
R
H1
S
R
28
would be expected to be present in cells. Thus, the
0
0
0
0
H2
H2
S
2.7
1.9
1.1
11.8
3.7
11.8
12.0
3.1
11.5
0.7
2.3
3.4
inclusion of these methyl groups is not imposed by the
availability of biosynthetic building blocks. A possible
role of the methyl groups in prebiotic synthesis of CoA
has also been discussed.30
R
Table 5. Calculated conformer distributions for 1, 2, and 10 (%)
0
0
00
00
00
00
C1 ±C2 ±C3 ±C4
O±C1 ±C2 ±C3
1
2
3
4
5
6
1a
2
36
72
59
62
24
27
2
4
14
64
49
39
28
11
14
8
40
47
1
0
aData from ref 7.
Scheme 4.
Table 6. Comparison of acetyl-CoA and acetyl-2 as substrates for acetyl-CoA utilizing enzymesa
Carnitine
acetyltransferase
Chloramphenicol
acetyltransferase
Citrate
synthase
K
m
(mM)
rel. Vmax
K
m
(mM)
rel. Vmax
K
m
(mM)
rel. Vmax
Acetyl-CoA
Acetyl-2
60 (5)
77 (7)
1
0.09 (0.02)
32 (3)
250 (30)
1
0.41 (0.1)
7 (1)
68 (8)
1
0.11 (0.03)
aNumbers in parentheses are standard error.

K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
2455
A likely role of the geminal methyl groups is that they
would be expected to in¯uence the conformation of the
molecule. Geminal dialkyl groups have been shown to
have large e€ects on conformation. A classical illustra-
tion of this e€ect is in the observed rates of reaction of
disul®des.4,5 These studies showed an up®eld shift in
protons of the pantetheine moiety of CoA as the pH was
increased above the pK of the adenine amino group
a
0
(pK =4.3) and the 3 -phosphate (pK =6.3). These pH
a
a
e€ects were observed in CoA, dephospho-CoA as well
as thioesters and disul®des of CoA, but not in phos-
phopantetheine. These distant e€ects were attributed to
intramolecular interactions between the pantetheine
moiety and the adenine base in a folded conformation
of CoA in which the pantetheine tail is coiled around the
adenine base. The increased shielding at higher pH was
attributed to a decrease in the population of this folded
conformation. The intramolecular interaction respon-
sible for these e€ects has not been addressed. Interac-
tion between protonated adenine and the pyrophos-
phate group might be possible, though the chemical
shift e€ects are observed in parts of the pantetheine
moiety quite distant from the pyrophosphate group. Lee
and Sarma estimated that no more than 30% of CoA
exists in the folded conformation on a time-average
3,3-disubstituted glutarate monoesters. The 3,3-
dimethylglutarate monoester 14b cyclizes 19-fold faster
than the unsubstituted glutarate monoester 14a (Fig.
3
1
3
). The rate enhancement is attributed to an increase
in the concentration of the gauche conformation of the
C2±C3 and C3±C4 bonds required for cyclization. The
e€ect of the methyl groups is even more striking in the
equilibrium for lactone formation from pantoic acid 16b
and its demethylated equivalent 2,4-dihydroxybutyric
acid 16a. While the Keq for lactone formation from
pantoic acid 16b in aqueous solution is 270, the Keq for
2,4-dihydroxybutyric acid 16a is only 0.67 under the
same conditions (Fig. 3).
30
The CoA analogue lacking the geminal dimethyl groups
was prepared to study the e€ect of deletion of these
methyl groups on the conformation and function of
CoA. The ®nal two steps of the synthesis were performed
using the enzymes phosphopantetheine adenylyltrans-
ferase and dephosphocoenzyme A kinase, which cata-
lyze the ®nal two steps of CoA biosynthesis. These
enzymes have been used in the synthesis of other CoA
4
basis. The folded and `open' conformations were pos-
tulated to interconvert via rotation about the pyrophos-
phate moiety. A similar folded conformation was
proposed previously for benzoyl-CoA.33 Studies reported
here of the pH dependence on chemical shift of protons
of the modi®ed pantetheine moiety of 2 gave very simi-
lar results to those observed with CoA. While the pD
range studied pushes the limits of accurate pH meter
readings, the in¯ection points are all within the range of
reliable measurements. Of the data included in Table 1,
2
3,32
analogues in this lab.
The phosphopantetheine ana-
logue 8 used for these ®nal two steps was racemic.
Phosphopantetheine adenylyltransferase has been
shown to be highly selective for the natural R-enantio-
00
00
00
the di€erences in Ád for H5 , H6 , and H9 are certainly
within experimental error and the di€erence for H8 may
3
2
00
mer of a phosphopantetheine analogue, thus it is
expected that the `natural' R-enantiomer of 8 was con-
verted preferentially to the dephospho-CoA analogue
and subsequently to the CoA analogue 2. The actual Km
value for the R-enantiomer of 8 is expected to be
also be. This suggests that deletion of the methyl groups
of CoA has no signi®cant e€ect on the existence of the
proposed folded conformation of CoA in solution.
2
.8 mM or one-half the Km value measured for the
Coupling constants between the protons of the modi®ed
pantoate moiety of 2 were determined to analyze the
conformation of 2 and for comparison of its conforma-
tion with that of natural CoA. The coupling constants
racemic compound. Though the yield of the ®nal two
steps was low, apparently due to inecient conversion of
8 to 2, enough product was obtained for several studies
to be performed.
00 00
between H3 and the C2 protons were easily deter-
00
mined upon decoupling of the C1 protons. The cou-
pling constants were initially used to tentatively assign
Several NMR studies of the conformation of CoA have
been reported. Lee and Sarma and subsequently Kiere et
al. reported the e€ects of pH on the H NMR spectrum of
CoA, dephospho-CoA, and some CoA thioesters and
0
0
00
the H2 and H2 chemical shifts. The coupling con-
R
S
1
00
stant of 2.7 Hz between H3 and the proton at 1.96 ppm
indicates that these two protons are almost exclusively
gauche to each other. The coupling constant of 9.0 Hz
between H3 and the proton at 1.63 ppm indicates that
these protons are largely anti to each other. It has been
shown that CoA exists as a mixture of conformers 1 and
about the C2 ±C3 bond (Fig. 2), with conformation 3
estimated to represent only 2% of the dynamic popula-
tion. If 2 is assumed to also exist primarily as a mixture
of conformers 1 and 2, with little contribution of con-
former 3, the signal at 1.96 ppm must be assigned to the
00
00 00
2
7
00
pro-R proton. This proton is gauche to H3 in both of
the populated conformers while the pro-S proton is
00
gauche and anti to H3 in the two populated con-
formers and thus has the larger coupling constant. The
opposite assignment would require that conformers 2
and 3 be predominant, with little population of con-
former 1. Such a dramatic change in rotamer popula-
tions appears especially unlikely given the fairly good
Figure 3. E€ects of geminal dimethyl groups on cyclization reactions.

2
456
K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
activity of the acetyl derivative of 2 with the three
enzymes tested. Furthermore, loss of the geminal dialkyl
e€ect would appear to favor an increase in population
of rotamer 1 for 2 relative to CoA, thus a dramatic
decrease in the population of rotamer 1 appears very
improbable. This assignment is further supported by
conformation. While the coupling constants could per-
haps be rationalized based on a predominantly anti
00
00
conformation about the C2 ±C3 bond, the coupling
constants are clearly not consistent with a single con-
00
00
formation about the C1 ±C2 bond.
0
0
comparison of the chemical shifts of the C2 protons
with those of the corresponding methyl groups of CoA.
The greater estimated population of anti conformers (1
and 6) about the C2 ±C3 and C1 ±C2 bonds for 2
0
0
00
00
relative to CoA is consistent with the expectation that
00
1
The more up®eld methyl signal in the H NMR spec-
trum of CoA has been assigned to the pro-S methyl
00
the methyl groups at C2 of CoA would tend to stabilize
the gauche conformations about these bonds. However,
the di€erences in estimated conformer populations are
very modest and appear to be substantially smaller than
the e€ects observed in other compounds, such as the
classical glutarate monoesters. It is likely that other
factors are important such as the gauche e€ect, which
tends to favor the positioning of polar groups gauche to
a neighboring C±C bond and anti to a C±H bond as has
group while the more down®eld signal has been
7
assigned to the pro-R methyl group. It is thus con-
0
0
sistent that the pro-S proton at C-2 of 2 is more up®eld
than the pro-R proton.
The relative rotamer populations estimated based on
coupling constants (Table 5) suggest that for 2 as with
00
00
CoA, the population of rotamer 3 about the C2 ±C3
bond is very small. This can be seen more simply and
3
4
been extensively studied with ¯uorinated alkanes. This
e€ect in 2 and 10 could mirror the conformational e€ect
of the methyl groups of 1, thus helping to rationalize the
similarities in predicted rotamer populations. Dipole±
dipole interactions, especially involving the polar C±O
0
0
00
conclusively in that H2 and H3 must be almost
S
exclusively gauche to each other, as in rotamers 1 and 2
to give the small coupling constant of 2.7 Hz. Any sig-
ni®cant population of rotamer 3 would give rise to lar-
ger coupling. The data also suggests that for 2, there is a
greater population of the anti conformation 1 than is
observed with CoA. The coupling constant of 9.0 Hz
00
00
bonds at C1 and C3 may also play a role in the con-
former distribution and in the interplay between con-
00
00
00
00
formations about the C2 ±C3 and C1 ±C2 bonds, as
has been studied in some detail for 1,3-di¯uoro-
propane.35
0
0
00
measured between H2 and H3 is almost large enough
S
to be consistent with the anti rotamer 1 being the exclu-
sive conformation, though the signi®cant population of
rotamer 2 gives a better ®t with the coupling constants
predicted by Macromodel. The data suggests that the
The di€erences in Vmax and Km for acetyl-2 relative to
acetyl-CoA with three acetyl-CoA utilizing enzymes as
shown in Table 6, and for 8 relative to phospho-
pantetheine with phosphopantetheine adenylyltransfer-
ase may be attributed in part to the in¯uence of the
methyl groups on the relative population of the con-
formation of CoA recognized by the enzymes. Another
factor could be the favorable transfer of the geminal
dimethyl groups from aqueous solution to a more
hydrophobic environment of an enzyme active site. The
simple kinetic data does not permit clear distinction of
possible e€ects. However, conformational e€ects would
probably only in¯uence the Km by a€ecting the con-
centration of the active conformation. The only slight
increase in Km for acetyl-2 versus acetyl-CoA with car-
nitine acetyltransferase further suggests that the methyl
groups do not have a major e€ect on the conformation
of CoA, or at least on the population of the conformer
of acetyl-CoA bound by carnitine acetyltransferase.
Other factors including hydrophobic e€ects might
in¯uence both Km and Vmax by di€erential e€ects on
binding in the ground state and transition state com-
plexes. The hydrophobic e€ect of a methyl group was
predicted by Pauling to provide a maximum 0.9 kcal/
mol of binding energy based on the energy of transfer of
a methyl group from an aqueous to hydrophobic envir-
onment,36 though e€ects of up to 3.6 kcal/mol have
0
0
00
conformation of 10 about the C2 ±C3 bond is similar to
that of 2, though with a moderately greater population
of conformer 3 relative to 1 for 10.
0
0
00
Regarding the conformation about the C1 ±C2 bond,
for CoA the population of the anti conformation 6 has
been determined to be small while the population of the
gauche conformation 4 is estimated to be more than dou-
ble that of the other gauche conformation 5. Estimates for
2
6
indicate that the population of the anti conformation
is somewhat greater than that observed in CoA while
as with CoA, the gauche conformation 4 is more popu-
lated than the other gauche conformation 5. While the
ratios of conformers 4 to 5 are di€erent between 1 and
2, it is not clear if this di€erence is signi®cant given the
potential error. The estimated conformer distribution
for 10 is again very similar to that for 2.
The estimated rotamer populations of Table 5 are sub-
ject to errors in both the measured and predicted cou-
pling constants. While there is thus some uncertainty in
the accuracy of the estimated conformer distributions,
some de®nite conclusions can be made. First, the cou-
pling constants and the estimated rotamer populations
for 2 and the simple model compound 10 are very simi-
lar. This indicates that the conformation of the pantoate
moiety is controlled by short-range interactions, with
possible long-range interactions such as those predicted
between the adenine and pantetheine moieties not having
a major in¯uence on the local conformation of the
pantoate moiety. The second conclusion is that the
structure of 2, like CoA, is not dominated by a single
3
7
been observed in the aminoacyl-tRNA synthetases.
The largest e€ect of deletion of the methyl groups on
enzyme activity was observed in the utilization of 8 as a
substrate for phosphopantetheine adenylyltransferase in
the synthesis of 2. This might be expected based on the
proximity of the reaction to the site of modi®cation.

K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
2457
Further analysis of the recently reported crystal struc-
ture of the phosphopantetheine adenylyltransferase
CDCl and D O, respectively. Mass spectral analysis
3 2
was performed at the University of California, Riverside
Mass Spectrometry facility, Riverside, CA.
38
may provide further insight into the basis for the 6-fold
decrease in Vmax and 36-fold increase in Km (assuming
only the R-enantiomer of 8 is a substrate). For carnitine
acetyltransferase, there is no crystal structure or other
information regarding the conformation of CoA in the
enzyme complex with which to compare the kinetic
results. For both citrate synthase and chloramphenicol
acetyltransferase, the anti conformation 1 about the
1
7
N -Mercaptoethyl -3- aminopropionamide disul®de (4).
To a solution of 314 (10.1 g, 30.4 mmol) in dimethyl sulf-
oxide (120 mL) was added sodium azide (5.9 g, 91 mmol)
ꢀ
and the reaction mixture was stirred at 50 C overnight.
Water (100 mL) was added and the precipitate was col-
lected via ®ltration, rinsed with water, and dried in
0
0
00
1
C2 ±C3 bond and the gauche conformation 5 about
0
vacuo overnight to yield a white solid (5.3 g). H NMR
0
00
the C1 ±C2 bond is observed in the crystal structures
of the enzyme complexes with substrate and substrate
(400 MHz, CDCl ): d 6.55 (s, 1H), 3.63 (overlapping
3
triplets, 4H), 2.85 (t, 2H), 2.5 (t, 2H). The solid was
dissolved in THF (110 mL) and triphenylphosphine
(8.08 g, 30.8 mmol) and water (0.67 mL, 37.2 mmol)
were added. The reaction mixture was stirred under
nitrogen at room temperature overnight. THF was
evaporated in vacuo, and the residue was redissolved in
methylene chloride (300 mL). The organic layer was
extracted with water (3Â100 mL) and the aqueous layers
were combined and lyophilized to yield a white solid
(3.88 g, 13.2 mmol, 43% yield): 1H NMR (200 MHz,
6
analogues. Interestingly, NMR analysis of the sub-
strate complex of chloramphenicol acetyltransferase has
suggested that the gauche conformation 3 about the
0
0
00
8
C2 ±C3 bond may be important. The data of Table 5
predict a modest increase in the population of con-
former 1 and a similar decrease in the population of
conformer 5 in 2 relative to CoA. In simplest terms, this
would predict approximately equal populations of the
1,5 conformation for 2 versus CoA, though this does not
consider that the conformations about these adjacent C±
C bonds may in¯uence each other. It thus appears that
conformational e€ects resulting from deletion of the
methyl groups are not a major factor in the di€erences
in kinetic data for acetyl-CoA versus acetyl-2. Thus,
both conformational and enzyme kinetic studies suggest
that the geminal dimethyl groups do not play a major
conformational role in the function of CoA.
D O): d 3.51 (t, 2H), 2.93 (t, 2H), 2.85 (t, 2H), 2.42 (t,
2H).
2
3-O-tert-Butyldimethylsilyl-2,2-didemethylpantethine (6).
A solution of 4 (3.07 g, 10.44 mmol) and 518 (4.54 g,
21 mmol) in methanol (90 mL) was stirred at room
temperature overnight. The reaction mixture was con-
centrated in vacuo, and the remaining oil was puri®ed
on silica gel with 3:1 ethyl acetate/methanol. Fractions
containing 6 were combined and concentrated under
reduced pressure and the residue was dissolved in ethyl
acetate, ®ltered to assure removal of silica gel, and con-
Summary
1
The studies of the didemethyl analogue of CoA 2
reported here do not support a major conformational
role of the geminal dimethyl groups of CoA. The pH-
dependent chemical shifts of protons of the pantetheine
moiety support a folded conformation for 2, similar to
CoA. Conformational populations estimated based on
coupling constants suggest only modest increases in anti
relative to gauche conformations of the pantoate moiety
upon deletion of the geminal dimethyl groups. While
the methyl groups of CoA may have a small e€ect on
the conformation which may result in a similar e€ect on
its activity in enzymatic reactions, other e€ects such as
hydrophobic interactions are probably of greater
importance.
centrated to yield a yellow oil (6.12 g, 81% yield): H
NMR (200 MHz, CDCl ): d 4.25 (dd, 1H), 3.7 (t, 2H),
3
3.5 (m, 4H), 2.8 (t, 2H), 2.45 (m, 2H), 1.9 (m, 1H), 1.7
(m, 1H), 0.9 (s, 9H), 0.12 (s, 3H), 0.10 (s, 3H).
3-O-tert-Butyldimethylsilyl-2,2-didemethylpantethine-1-
O-dibenzylphosphate (7). To a solution of N-chloro-
succinimide (5.33 g, 39.9 mmol) in toluene (100 mL) was
added dibenzyl phosphite (8.7 mL, 39.5 mmol). The
solution was stirred for 2 h under nitrogen, during
which time the reaction ¯ask was cooled in an ice bath
as needed to prevent re¯uxing. The resulting dibenzyl
phosphorochloridate solution was used immediately in
the next reaction.
Compound 6 (3.1 g, 4.2 mmol) was dissolved in pyridine
(
25 mL) and the solvent was evaporated in vacuo to
Experimental
remove residual water. The residue was redissolved in
dry pyridine (50 mL) and cooled in an acetonitrile±dry
ice bath. The dibenzyl phosphorochloridate solution was
added to the reaction ¯ask over 10 min. The reaction
mixture was allowed to warm to room temperature and
was stirred under nitrogen overnight. Water (30 mL)
was added and the resulting mixture was concentrated
under reduced pressure to a brown oil. The residue was
redissolved in ethyl acetate (200 mL) and washed
sequentially with aq sulfuric acid (1 M, 2Â50 mL), aq
sodium bicarbonate (1 M, 2Â50 mL), and saturated aq
sodium sulfate (1Â50 mL). The organic layer was dried
General experimental
Enzymes and enzyme substrates and cofactors were
obtained from Sigma. Other chemicals were obtained
from Aldrich. Dichloromethane, toluene, and acetonitrile
were distilled from calcium hydride. Tetrahydrofuran
was distilled from sodium benzophenone radical. NMR
experiments were conducted on Varian Gemini 200,
Varian XL-400, and Varian Inova 500 and 600 MHz
1
spectrometers. In H NMR experiments, TMS (0 ppm)
and HOD (4.8 ppm) were used as internal reference in

2
458
K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
over MgSO and concentrated under reduced pressure
4
enzymes phosphopantetheine adenylyltransferase (3
units) and dephosphocoenzyme A kinase (3 units),
which were partially puri®ed from Corynebacterium
ammoniagenes and coimmobilized in polyacrylamide gel
as described previously.23 The resulting suspension was
slowly stirred at room temperature for 6 days. After 2
days, additional pyruvate kinase (50 units) and inor-
ganic pyrophosphatase (50 units) were added, and after
4 days additional ATP (100 mg, 170 mmol), phospho-
enolpyruvate (40 mg, 200 mmol), pyruvate kinase (50
units) and inorganic pyrophosphatase (50 units) were
added. The reaction was monitored by analytical
reverse phase HPLC (Rainin Microsorb C-18 column,
4.6 mmÂ25 cm, ¯ow rate 1 mL/min, 5 min at 5% metha-
nol in 50 mM phosphate bu€er, pH 4.5 followed by a
linear gradient to 60% methanol over 12 min and then
maintained at 60% methanol) and after 6 days no more
8 (retention time 12.1 min) was observed. The reaction
mixture was centrifuged to remove immobilized
enzymes, which were further washed with water
(3Â20 mL). The combined reaction solution and washes
were diluted to 80 mL with water, and the pH adjusted
to 2.5 with aqueous HCl. The solution was loaded onto
a DEAE cellulose column (2.5Â20 cm) and eluted with a
linear gradient of LiCl (0±0.2 M) in 3 mM HCl (800 mL
to yield a golden oil. The product was puri®ed on silica
gel with 3:1 ethyl acetate/methanol. Fractions contain-
ing 7 were combined and concentrated under reduced
pressure and the residue was dissolved in ethyl acetate,
®ltered to assure removal of silica gel, and concentrated
1
to yield 7 as a yellow oil (2.22 g, 42% yield): H NMR
(
200 MHz, CDCl ): d 7.35 (m, 10H), 5.0 (d, 4H), 4.2 (m,
3
3
2
H), 3.5 (m, 4H), 2.75 (t, 2H), 2.4 (t, 2H), 1.9±2.1 (m,
H), 0.9 (s, 9H), 0.1 (s, 6H).
2,2-Didemethylpantethine-1-O-phosphate (8-disul®de).
To a solution of 7 (1.38 g, 1.09 mmol) in acetonitrile
7 mL) was added aqueous hydro¯uoric acid (0.18 mL,
8%). The reaction mixture was stirred for 2.5 h at
(
4
room temp. Chloroform (100 mL) was added, and the
organic layer was washed with sat aq NaCl (2Â50 mL),
dried over MgSO , and concentrated in vacuo to yield a
4
1
yellow oil (0.80 g). H NMR (200 MHz, CDCl ): d 7.35
3
(
2
m, 10H), 5.0 (m, 4H), 4.2 (m, 3H), 3.5 (m, 4H), 2.75 (t,
H), 2.4 (t, 2H), 1.9±2.1 (m, 2H). To a solution of the
resulting oil (0.80 g, 0.79 mmol) in dry acetonitrile
6 mL) was added lithium bromide (0.29 g, 3.29 mmol).
(
Acetonitrile was evaporated in vacuo to remove residual
water and the residue was redissolved in dry acetonitrile
ꢀ
(
added and the reaction mixture was stirred under nitro-
6 mL). Chlorotrimethylsilane (2 mL, 15.7 mmol) was
total volume) at a ¯ow rate of 0.5 mL/min at 4 C. 2
eluted between 0.12 and 0.16 M LiCl and was detected
by reaction of aliquots of individual fractions with
DTNB at pH 8.0. The fractions containing 2 were
combined, lyophilized, and further puri®ed by reverse-
phase HPLC (5 min at 5% methanol in 10 mM phos-
phate bu€er, pH 4.5 followed by a linear gradient to
40% methanol over 30 min, 2 eluted between 16 and
ꢀ
gen at 50 C overnight. The mixture was concentrated
under reduced pressure, dissolved in water (5 mL), washed
with ether, and lyophilized. The resulting mixture con-
taining inorganic salts was puri®ed by preparative C-18
reverse-phase HPLC, monitoring at 270 nm, and using a
gradient of methanol in 0.2% aqueous tri¯uoroacetic
1
acid. Fractions containing 8-disul®de were lyophilized
1
20 min) to give 2 (12.1 mg, 6%). H NMR (400 MHz,
to obtain a white solid: H NMR (400 MHz, D O): d
D O): d 8.54 (s, 1H), 8.30 (s, 1H), 6.14 (d, 1H,
2
2
4
.25 (dd, 1H, J=5.8, 3.3 Hz), 3.96 (m, 2H), 3.49 (m, 4H),
J=5.6 Hz), 4.53 (s, 1H), 4.15±4.20 (m, 3H), 3.90±4.02
(m, 2H), 3.39 (t, 2H, J=6.2 Hz), 3.26 (t, 2H, J=6.4 Hz),
2.54 (t, 2H, J=6.6 Hz), 2.38 (t, 2H, J=6.6 Hz), 1.95±
2.05 (m, 1H), 1.60±1.72 (m, 1H). 31P NMR (161.9 MHz,
D O, external H PO reference): d � 10.72, � 10.33, 0.38.
2
.82 (t, 2H, J=6.3 Hz), 2.46 (t, 2H, J=6.6 Hz), 2.06±2.14
3
1
(
m, 1H), 1.75±1.84 (m, 1H); C NMR (100 MHz, D O):
2
d 175.36, 173.08, 67.06, 61.69, 37.12, 35.62, 34.73, 34.46,
+
�
3 calcd
C H N O P S m/z 657.1066, found 657.1097.
3.45; HRMS (FAB): [M� H ]
for
2
3
4
+
HRMS (FAB): [MH ] calcd for C H N O P S m/z
19 31 7 16 3
1
8
35
4
14
2
2
7
38.076, found 738.077.
Determination of kinetic parameters for 8 with
phosphopantetheine adenylyltransferase
1
pH-Dependent NMR shift studies of 2. The 400 MHz H
NMR spectrum of a solution of 2 (2.3 mg, 3 mmol) and
trimethylsilyl-1-propane sulfonic acid (0.22 mg, 1 mmol)
To a solution of pyrophosphate reagent (Sigma Chemi-
cal Company, 1.0 mL) containing added ATP (5 mM)
was added phosphopantetheine adenylytransferase (0.01
units) and 8-disul®de (1.0 to 10.3 mM, preincubated
with 1 M DTT at pH 7 for 1 hour to reduce the dis-
ul®de) and the decrease in absorbance at 340 nm was
observed over time.
in 700 mL D O was recorded with suppression of the
2
HOD signal. Proton shifts were referenced relative to
the methyl groups of trimethylsilyl-1-propane sulfonic
acid at 0 ppm. The pH was adjusted between spectra
with aqueous DCl or NaOD, and pD was taken as the
pH meter reading+0.4. NMR spectra were recorded
across the pD range of 2.35 to 13.43.
0
0
00
2
,2 -Didemethyl-coenzyme A (2). To a solution of 8-
disul®de (181 mg, 275 mmol) in HEPES bu€er (2 mL,
Synthesis of 10. To a solution of 9 (1.49 g, 11.2 mmol) in
THF (26 mL) was added benzyl amine (2.39 g,
22.3 mmol). The solution was stirred at 55 C for 4 h,
ꢀ
0
2
.1 M, pH 8.1) was added dithiothreitol (42.3 mg,
75 mmol). The solution was incubated for 1 hour, after
ꢀ
then cooled to 4 C. Aq HCl (10%, 10 mL) was added
which it was added to a solution of HEPES bu€er
10 mL, 0.1 M, pH 8.1) containing ATP (231 mg,
00 mmol), MgCl hydrate (81 mg, 400 mmol), and phos-
(
and the solution was extracted with ethyl acetate
(3Â20 mL). The combined organic layers were con-
centrated under reduced pressure and the remaining oil
was puri®ed on silica gel with ethyl acetate/hexane
4
2
phoenolpyruvate (83.2 mg, 400 mmol). The soluble
enzymes pyruvate kinase (50 units) and inorganic pyro-
phosphatase (50 units) were added along with the
(R =0.42 in 4:1 ethyl acetate/hexane) to yield 10 (1.93 g,
f

K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
2459
1
9
5
1
.23 mmol): H NMR (300 MHz, DMSO-d ): d 7.2 (m,
3. Fierke, C. A.; Jencks, W. P. J. Biol. Chem. 1986, 261, 7603.
Whitty, A.; Fierke, C. A.; Jencks, W. P. Biochemistry 1995, 34,
6
H), 5.5 (bs, 1H), 4.3 (d, 2H), 4.0 (m, 1H), 3.5 (m, 2H),
.9 (m, 1H), 1.6 (m, 1H).
11678.
4. Lee, C.-H.; Sarma, R. H. J. Am. Chem. Soc. 1975, 97, 1225.
5. Keire, D. A.; Robert, J. M.; Rabenstein, D. L. J. Org.
Synthesis of acetyl-2. To a solution of 2 (5 mg, 6.8 mmol)
in HEPES bu€er (1.0 mL, 0.025 M, pH 8.2) was added a
solution of S-phenylthioacetate (0.1 mL, 0.75 mmol) in
acetonitrile (0.9 mL). The resulting turbid solution was
stirred vigorously for 1 hour at room temperature. The
mixture was allowed to separate into two phases which
were separated and the aqueous layer was extracted
with ether (5Â4 mL) and adjusted to pH 4.5 with aq
HCl. The product was puri®ed by reverse-phase HPLC
Chem. 1992, 57, 4427.
. D'Ordine, R. L.; Paneth, P.; Anderson, V. E. Bioorg.
Chem. 1995, 23, 169.
6
7
1
. Wu, W.-J.; Tonge, P. J.; Raleigh, D. P. J. Am. Chem. Soc.
998, 120, 9988.
8. Barsukov, I. L.; Lian, L.-Y.; Ellis, J.; Sze, K.-H.; Shaw, W.
V.; Roberts, G. C. K. J. Mol. Biol. 1996, 262, 543.
9. Wu, W.-J.; Anderson, V. E.; Raleigh, D. P.; Tonge, P. J.
Biochemistry 1997, 36, 2211.
1
0. For reviews of X-ray structures of CoA-utilizing enzymes
see: Engel, C.; Wierenga, R. Curr. Opin. Struct. Biol. 1996, 6,
90. Modis, Y.; Wierenga, R. Structure 1998, 6, 1345. For more
(
5 min at 5% methanol in 10 mM phosphate bu€er, pH
4
3
.5 followed by a linear gradient to 45% methanol over
0 min, product eluted between 19 and 24 min) to give
7
recently reported structures see: Benning, M. M.; Wesenberg, G.;
Liu, R.; Taylor, K. L.; Dunaway-Mariano, D.; Holden, H. M. J.
Biol. Chem. 1998, 273, 33572. Wybenga-Groot, L. E.; Draker,
K.; Wright, G. D.; Berghuis, A. M. Structure 1999, 7, 497.
Hickman, A. B.; Namboodiri, M. A. A.; Klein, D. C.; Dyda, F.
Cell 1999, 97, 361. Trievel, R. C.; Rojas, J. R.; Sterner, D. E.;
Venkataramani, R. N.; Wang, L.; Zhou, J.; Allis, C. D.; Berger,
S. L.; Marmorstein, R. Proc. Natl. Acad. Sci. USA 1999, 96,
1
acetyl-2 (2.96 mg, 60%). H NMR (400 MHz, D O): d
2
8
(
2
.54 (s, 1H), 8.28 (s, 1H), 6.16 (d, 1H, J=6.0 Hz), 4.57
s, 1H), 4.15±4.27 (m, 3H), 3.95±4.05 (m, 2H), 3.40 (t,
H, J=6.6 Hz), 3.30 (t, 2H, J=5.8 Hz), 2.92 (t, 2H,
J=6.2 Hz), 2.37 (t, 2H, J=6.6 Hz), 2.30 (s, 1H), 2.00±
.10 (m, 2H), 1.62±1.72 (m, 1H).
2
8
931. Clements, A.; Rojas, J. R.; Trievel, R. C.; Wang, L.; Ber-
Enzyme assays. Assays of carnitine acetyltransferase
were conducted in potassium phosphate bu€er (25 mM,
pH 7.5), containing 4,4 -dithiopyridine (PDS, 0.2 mM),
l-carnitine (0.2 mM), carnitine acetyltransferase (0.03
units, from pigeon breast) and acetyl-CoA or acetyl-2
ger, S. L.; Marmorstein, R. EMBO J. 1999, 18, 3521.
11. Lin, Y.; Fletcher, C. M.; Zhou, J.; Allis, C. D.; Wagner,
G. Nature 1999, 400, 86.
0
1
2. Allinger, N. L.; Zalkow, V. J. Org. Chem. 1960, 25, 701.
Wheeler, O. H.; de Rodriguez, E. E. G. J. Org. Chem. 1964, 29,
227.
(
20±200 mM) in a total volume of 1 mL by monitoring
the increase in absorbance of the thiopyridine anion at
24 nm. Assays of chloramphenicol acetyltransferase
1
13. Kirby, A. J. Adv. Phys. Org. Chem. 1980, 17, 183.
14. Airapetyan, G. M.; Zherebchenko, P. G.; Bystrova, V.
M.; Benevolenskaya, Z. V.; Kuleshova, N. D.; Linkova, M.
G.; Kildisheva, O. V.; Knunyants, I. L. Bull. Acad. Sci. USSR,
Div. Chem. Sci. 1967, 317.
3
were conducted in Tris bu€er (10 mM, pH 7.8) contain-
ing chloramphenicol (50 mM), DTNB (1 mM), chlor-
amphenicol acetyltransferase (0.03 units, from
Escherichia coli) and acetyl-CoA (10 mM to 100 mM) or
acetyl-2 (50±500 mM) by monitoring the increase in
absorbance of the nitrothiobenzoate dianion at 412 nm.
Citrate synthase assays were conducted in Tris bu€er
15. Howson, W.; Kitteringham, J.; Mistry, J.; Mitchell, M. B.;
Novelli, R.; Slater, R. A.; Swayne, G. T. G. J. Med. Chem.
1
1
1
1
988, 31, 352.
6. Knouzi, M.; Vaultier, M.; Carrie, R. Bull. Soc. Chim. Fr.
985, 815.
7. Oiry, J.; Pue, J. Y.; Imbach, J. L.; Fatome, M.; Sentenac-
(0.1 M, pH 8), containing DTNB (0.1 mM), oxalo-ace-
tate (0.5 mM), citrate synthase (0.03 units, from porcine
Roumanou, H.; Lion, C. J. Med. Chem. 1986, 29, 2217.
18. Dauben, W. G.; Hendricks, R. T.; Pandy, B.; Wu, S. C.;
Zhang, X.; Luzzio, M. J. Tetrahedron Lett. 1995, 44, 519.
heart) and acetyl CoA (4±40 mM) or acetyl-2 (20±
200 mM) by monitoring the increase in absorbance at
1
9. Shemin, D., Ed. Biochemical Preparations; John Wiley &
Sons: New York, 1957; Vol. 5, p 1.
0. McKenna, C. E.; Higa, M. T.; Cheung, N. H.; McKenna,
M. Tetrahedron Lett. 1977, 155.
412 nm.
2
Acknowledgements
2
2
1. O'Brien, W. E. Anal. Biochem. 1976, 76, 423.
2. Martin, D. P.; Drueckhammer, D. G. Biochem. Biophys.
This work was supported by National Science Founda-
tion grant MCB9722936. The NMR facility at Stony
Brook is supported by National Science Foundation
grant CHE9413510. We thank Professor Dan Raleigh
for helpful discussions and Dr. Fang Liu for assistance
with NMR experiments.
Res. Commun. 1993, 192, 1155.
3. Martin, D. P.; Bibart, R. T.; Drueckhammer, D. G. J.
Am. Chem. Soc. 1994, 116, 4660.
24. Hirschbein, B. L.; Mazenod, G. M.; Whitesides, G. M. J.
Org. Chem. 1982, 47, 3765.
2
25. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Kis-
kamp, R.; Lipton, M.; Cau®eld, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J. Comput. Chem. 1990, 11, 440.
26. Pachler, K. G. R. Spectrochim. Acta 1964, 20, 581.
27. Teller, J. H.; Powers, S. G.; Snell, E. E. J. Biol. Chem. 1976,
251, 3780. Aberhart, D. J. J. Am. Chem. Soc. 1979, 101, 1354.
References and Notes
1. Lee, C.-H.; Chen, A. F. In The Pyridine Nucleotide Co-
enzymes; Everse, J., Anderson, B., You, K., Eds.; Academic:
New York, 1982; p 189.
28. Epelbaum, S.; LaRossa, R. A.; VanDyk, T. K.; Elkayam,
T.; Chipman, D. M.; Barak, Z. J. Bacteriol. 1998, 180, 4056.
29. Ketopantoate hydroxymethyltransferase, the enzyme that
catalyzes the addition of the hydroxymethyl group to 11, has been
shown to accept a-ketobutyrate but not pyruvate as a substrate;
Powers, S. G.; Snell, E. E. J. Biol. Chem. 1976, 251, 3786.
2
. Primerano, D.; Burns, R. O. J. Bacteriol. 1982, 150, 1202.
La Rossa, R. A.; VanDyk, T. K. J. Gen. Microbiol. 1989, 135,
209.
2

2
460
K. W. Vogel et al. / Bioorg. Med. Chem. 8 (2000) 2451±2460
34. Kunttu, H.; Raaska, T.; Rasanen, M. Theochem 1989,
3
3
5
1
3
0. Miller, S. L.; Schlesinger, G. J. Mol. Evol. 1993, 36, 308.
1. Bruice, T. C.; Pandit, K. P. J. Am. Chem. Soc. 1960, 82,
858. Lightstone, F. C.; Bruice, T. C. J. Am. Chem. Soc. 1997,
19, 9103.
184, 277.
35. Wu, D.; Tian, A.; Sun, H. J. Phys. Chem. 1998, 102, 9901.
36. Pauling, L. Festschr. Prof. Dr. Arthur Stoll Siebzigsten
Geburstag 1958, 597.
2. Bibart, R. T.; Vogel, K. W.; Drueckhammer, D. G. J.
Org. Chem. 1999, 64, 2903.
3. Mieyal, J. J.; Webster, L. T.; Siddiqui, U. A. J. Biol.
Chem. 1974, 249, 2633.
37. Fersht, A. R.; Shindler, J. S.; Tsui, W.-C. Biochemistry
1980, 19, 5520.
38. Izard, T.; Geerlof, A. EMBO J. 1999, 18, 2021.
3