Cyclohexyl-octahydro-pyrrolo[1,2-a]pyrazine-Based Inhibitors of Human N-Myristoyltransferase-1

Article abstract of DOI:10.1124/jpet.103.061572

N-Myristoyltransferase (NMT) is an emerging therapeutic target that catalyzes the attachment of myristate to the N terminus of an acceptor protein. We have developed a medium-throughput assay for screening potential small molecule inhibitors of human NMT-1 consisting of recombinant enzyme, biotinylated peptide substrate, and [³H]myristoyl-CoA. Approximately 16,000 diverse compounds have been evaluated, and significant inhibition of NMT was found with 0.8% of the compounds. From these hits, we have identified the cyclohexyl-octahydro-pyrrolo[1,2-a]pyrazine (COPP) chemotype as inhibitory toward human NMT-1. Thirty-two compounds containing this substructure inhibited NMT-1, with IC₅₀ values ranging from 6 μM to millimolar concentrations, and a quantitative structure-activity relationship equation (r² = 0.72) was derived for the series. The most potent inhibitor (24, containing 9-ethyl-9H-carbazole) demonstrated competitive inhibition for the peptide-binding site of NMT-1 and noncompetitive inhibition for the myristoyl-CoA site. Computational docking studies using the crystal structure of the highly homologous yeast NMT confirmed that 24 binds with excellent complementarity to the peptide-binding site of the enzyme. To evaluate the ability of 24 to inhibit NMT activity in intact cells, monkey CV-1 cells expressing an N-myristoylated green fluorescent protein (GFP) fusion protein were treated with a known NMT inhibitor or with 24. Each compound caused the redistribution of GFP from the plasma membrane to the cytosol. Furthermore, 24 inhibits cancer cell proliferation at doses similar to those that inhibit protein myristoylation. Overall, these studies establish an efficient assay for screening for inhibitors of human NMT and identify a novel family of inhibitors that compete at the peptide-binding site and have activity in intact cells.

Full text of DOI:10.1124/jpet.103.061572

0022-3565/04/3091-340–347$20.00
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics
JPET 309:340–347, 2004
Vol. 309, No. 1
61572/1135042
Printed in U.S.A.
Cyclohexyl-octahydro-pyrrolo[1,2-a]pyrazine-Based Inhibitors
of Human N-Myristoyltransferase-1
Kevin J. French,1 Yan Zhuang, Randy S. Schrecengost, Jean E. Copper, Zuping Xia,1 and
Charles D. Smith
Department of Pharmacology, Penn State College of Medicine, Hershey, Pennsylvania
Received October 13, 2003; accepted December 12, 2003
ABSTRACT
N-Myristoyltransferase (NMT) is an emerging therapeutic target site of NMT-1 and noncompetitive inhibition for the myristoyl-
that catalyzes the attachment of myristate to the N terminus of CoA site. Computational docking studies using the crystal
an acceptor protein. We have developed a medium-throughput structure of the highly homologous yeast NMT confirmed that
assay for screening potential small molecule inhibitors of hu- 24 binds with excellent complementarity to the peptide-binding
man NMT-1 consisting of recombinant enzyme, biotinylated site of the enzyme. To evaluate the ability of 24 to inhibit NMT
peptide substrate, and [3H]myristoyl-CoA. Approximately activity in intact cells, monkey CV-1 cells expressing an N-
16,000 diverse compounds have been evaluated, and signifi- myristoylated green fluorescent protein (GFP) fusion protein
cant inhibition of NMT was found with 0.8% of the compounds. were treated with a known NMT inhibitor or with 24. Each
From these hits, we have identified the cyclohexyl-octahydro- compound caused the redistribution of GFP from the plasma
pyrrolo[1,2-a]pyrazine (COPP) chemotype as inhibitory toward membrane to the cytosol. Furthermore, 24 inhibits cancer cell
human NMT-1. Thirty-two compounds containing this sub- proliferation at doses similar to those that inhibit protein myris-
structure inhibited NMT-1, with IC50 values ranging from 6 ␮M toylation. Overall, these studies establish an efficient assay for
to millimolar concentrations, and a quantitative structure-activ- screening for inhibitors of human NMT and identify a novel
ity relationship equation (r2 ϭ 0.72) was derived for the series. family of inhibitors that compete at the peptide-binding site and
The most potent inhibitor (24, containing 9-ethyl-9H-carbazole) have activity in intact cells.
demonstrated competitive inhibition for the peptide-binding
The term protein myristoylation refers to the covalent at- quires only myristoyl-CoA and a protein containing a suit-
tachment of myristic acid to specific proteins. This process
has been shown to target proteins to specific subcellular
membranes (Boutin, 1997), stabilize protein conformation
(Chow et al., 1987; Zheng et al., 1993; Resh, 1996), and
facilitate additional lipidations (Robbins et al., 1995; Yur-
chak and Sefton, 1995; Song et al., 1997; van’t Hof and Resh,
1999). Although the reaction mainly occurs as a cotransla-
tional event, it has been shown to occur after cleavage of
some substrates (Zha et al., 2000). N-Myristoyltransferase
(NMT; E.C. 2.3.1.97) catalyzes this addition onto the N-ter-
minal glycine residue of specific proteins. The reaction re-
able peptide sequence and occurs through an ordered Bi Bi
mechanism (Rudnick et al., 1991; Rocque et al., 1993; Bhat-
nagar et al., 1998). The list of myristoylated proteins is
diverse, including viral proteins, G␣ proteins, tyrosine ki-
nases of the Src family, and nitric oxide synthase. Gene
replacement studies indicate that NMT is essential for via-
bility of Cryptococcus neoformans (Lodge et al., 1994). In
addition, mutation or deletion of NMT results in recessive
lethality in Saccharomyces cerevisiae (Duronio et al., 1989,
1991) and Candida albicans (Weinberg et al., 1995). Thus,
NMT appears to be an essential enzyme in both prokaryotes
and higher organisms. In humans, NMT has more complex
expression patterns, including a shortened alternative splice
variant lacking the ribosomal targeting sequence, which oc-
curs in several tissue types (McIlhinney et al., 1998). Addi-
tionally, Giang and Cravatt (1998) have recently cloned and
characterized a second isoform (NMT-2) that demonstrates
This work was supported by National Institutes of Health Grant R24
CA788243 (to C.D.S.).
1 Current address: Apogee Biotechnology Corporation, P.O. Box 916, Her-
shey, PA 17033.
Article, publication date, and citation information can be found at
http://jpet.aspetjournals.org.
DOI: 10.1124/jpet.103.061572.
ABBREVIATIONS: NMT, N-myristoyltransferase; COPP, cyclohexyl-octahydro-pyrrolo[1,2-a]pyrazine; DMSO, dimethylsulfoxide; GST, glutathi-
one S-transferase; myr-phe, N-myristoyl-L-phenylalanine; N-myr-GFP, N-myristoylated green fluorescent protein; PBS, phosphate-buffered
saline; QSAR, quantitative structure-activity relationship; GFP, green fluorescent protein; MOE, molecular operating environment; MOPP,
methyl-octahydro-pyrrolo[1,2-a]pyrazine; 2-HM, 2-hydroxymyristic acid.
340

Inhibitors of Human NMT-1
341
similar activities to NMT-1. These findings reveal that NMT vector encoding only the GST portion was used. The recombinant
proteins were expressed and purified on immobilized GSH columns
according to the specifications of Amersham Biosciences Inc. The
column elutates containing GST or GST-hNMT-1 were then dialyzed
in PBS overnight to remove glutathione and stored at Ϫ80°C in 20%
glycerol until use.
is an essential enzyme whose expression and processing are
complex.
Several pathogenic states are linked to undesired NMT
activity. The Rous Sarcoma Virus transforms cells through
expression of the tyrosine kinase p60src. Kamps et al. (1986)
demonstrated that mutation of the N-terminal glycine of
p60src conserves its kinase activity but blocks its ability to
associate with membranes and to transform cells. The Hu-
man Immunodeficiency Virus-1 gag precursor protein under-
goes N-myristoylation, and the inhibitor N-myristoyl glycinal
diethylacetal results in accumulation of immature gag pre-
cursors and inhibition of viral replication (Furuishi et al.,
1997). NMT also appears to play a role in cancer. Increases in
NMT expression and activity have been observed in colorec-
tal adenocarcinomas (Raju et al., 1997), gallbladder carcino-
mas (Rajala et al., 2000), and the murine leukemia cell line
L1210 (Boutin et al., 1993). Furthermore, several protein
tyrosine kinases of the Src family are well established onco-
genes that require myristoylation for activity (Resh, 1994).
Importantly, a number of studies have shown that c-Src is a
critical regulator of human cancers, including breast carci-
nomas (Garcia et al., 2001). Additional studies have shown
that Src may be a therapeutic target for the treatment of
osteoporosis (Sawyer et al., 2001).
Because of the dependence of yeast on NMT activity for
growth, the consideration of NMT as a therapeutic target has
primarily focused on the development of NMT inhibitors to
be used as antifungal agents. The pioneering work of Gordon
and colleagues led to the design and synthesis of antifungal
peptidomimetic and dipeptide analog inhibitors for C. albi-
cans NMT that are selective versus human NMT while dem-
onstrating low nanomolar potencies toward the enzyme (De-
vadas et al., 1995, 1997; Nagarajan et al., 1997). Other
recently described inhibitors include amino acid derivatives
(Boutin et al., 1993; Furuishi et al., 1997; Tabuchi et al.,
1997; Cordo et al., 1999) and nonhydrolyzable myristoyl-CoA
analogs (Heuckeroth and Gordon, 1989; Paige et al., 1989;
Wagner and Retey, 1991). However, the development of these
compounds may be limited by their poor pharmacokinetic
properties. Thus, we have undertaken a project to design,
synthesize, and evaluate small molecule inhibitors of human
NMT. We report here the discovery of the novel chemotype
cyclohexyl-octahydro-pyrrolo[1,2-a]pyrazine (COPP) that in-
hibits human NMT-1 with micromolar potency.
Myristoyl-L-phenylalanine was synthesized and purified as fol-
lows. Myristic acid was coupled with an equimolar amount of N-
hydroxysuccinimide by dicyclohexylcarbodiimide in ethyl acetate at
room temperature for 20 h (yield, 85%). The product was then re-
acted with an equimolar amount of L-phenylalanine (Sigma-Aldrich;
enantiomeric excess 98%) in a 50:50 solution of tetrahydrofuran/0.1
M sodium bicarbonate at room temperature for 20 h. After acidifica-
tion with 1 N hydrochloric acid (final pH, ϳ2), the final product was
a white solid crystallized from chloroform and petroleum ether
(yield, 60%). The identity and purity of the product was confirmed by
1H- and 13C-NMR analyses. The peptide substrate, NH2-
GNAASARRK-biotin, was synthesized by the Macromolecular Core
Facility at the Penn State College of Medicine using standard solid-
phase synthetic techniques.
[3H]Myristoyl-CoA Synthesis. [3H]Myristoyl-CoA was synthe-
sized by adding 0.1 U of Pseudomonas sp. Acyl-coenzyme A synthase
(Sigma-Aldrich) to 200 ␮l of buffer (20 mM Tris-HCl, 1 mM dithio-
threitol, 10 mM MgCl2, and 0.1 mM EGTA, pH 8.0) containing 1.3
mM Coenzyme A (lithium salt), 5.4 mM ATP, and 0.8 ␮M [8,9-
3H]myristic acid (50 Ci/mmol; American Radiolabeled Chemicals, St.
Louis, MO). The mixture was incubated at 37°C with shaking for 1 h,
and unreacted [3H]myristic acid was removed by acidification and
extraction with hexane. The extent of conversion to [3H]myristoyl-
CoA was determined to be 95%, and the product was stored at Ϫ20°C
until its use.
Assay of NMT Activity. The NMT assays were performed in
streptavidin-coated 96-well plates and consisted of the following
reagents: 2 ␮g of purified GST-hNMT-1 fusion protein, 23 nM NH2-
GNAASARRK-biotin (stoichiometric with the amount of bound
streptavidin), 0.5 ␮l of 1 mg/ml test compound in DMSO, and NMT
assay buffer (30 mM Tris-HCl, pH 7.4, 0.5 mM EGTA, 0.45 mM
EDTA, 4.5 mM 2-mercaptoethanol, and 0.05% Tween 20). The reac-
tions were initiated with the addition of [3H]myristoyl-CoA (final
concentration, 60 nM) to a final volume of 100 ␮l and were incubated
at 25°C with shaking for 30 min. Reactions were terminated by
aspiration of the incubation mixture. The bound biotinylated peptide
was washed three times with 200 ␮l of 30 mM Tris-HCl containing
0.05% Tween 20 and eluted with 300 ␮l of 0.1 M glycine, pH 2.8.
After 1 h, the eluted peptide was transferred to scintillation vials,
and the amount of bound 3H was quantified to determine NMT
activity.
Competition Assays. The peptide substrate NH2-GNAASARRK-
biotin was bound to stoichiometric excesses of streptavidin-coated
polystyrene beads (Polysciences, Warrington, PA). In all competition
studies, reactions were stopped before 10% of the substrate under-
went myristoylation to ensure valid kinetic analyses. For peptide
competition assays, the substrate concentration was varied from 0 to
500 ␮M and incubated with 5 ␮g of purified GST-hNMT-1 fusion
Materials and Methods
Unless otherwise noted, all chemicals and reagents were pur-
chased from Sigma-Aldrich (St. Louis, MO). The library of com- protein, 160 nM [3H]myristoyl-CoA, DMSO or 6 ␮M 24, and NMT
pounds used for screening was purchased from the ChemBridge assay buffer in a total volume of 50 ␮l. Reactions were incubated at
Corporation (San Diego, CA) and was provided as individual com- 25°C with shaking for 15 min. The beads were washed twice with
pounds dissolved in DMSO (1 mg/ml). Specific cyclohexyl-octahydro- assay buffer by centrifugation, transferred to scintillation vials, and
pyrrolo[1,2-a]pyrazine compounds in the library were synthesized 3H was quantified to determine NMT activity. For myristoyl-CoA
previously as described (Likhosherstov et al., 1993). The identities of competition assays, the peptide concentration was 100 ␮M, whereas
all compounds were confirmed by NMR analyses, and high-perfor- myristoyl-CoA (mixed in a 1:25 ratio with nonradioactive myristoyl-
mance liquid chromatography-mass spectroscopic analyses indicated CoA) concentration varied from 0 to 20 ␮M. Kinetic constants were
that each was of 95% or greater purity (data not shown).
calculated by regression analyses using the Prism 3.0 software pack-
The cDNA for human NMT-1 was a kind gift from Dr. Jeffrey age (GraphPad Software, San Diego, CA).
Gordon (Washington University, St. Louis, MO). NMT-1 cDNA was
Computational Studies. The chemical characteristics of each
subcloned into the pGEX-5x-3 bacterial expression vector (Amer- compound in the COPP set were determined using the 2D QSAR
sham Biosciences Inc., Piscataway, NJ), resulting in the generation descriptors as calculated by the program QuaSAR-Model in the MOE
of a GST-hNMT-1 fusion protein. As a control, the unmodified pGEX
software package (version 2002.03; Chemical Computing Group Inc.,

342
French et al.
Montreal, Quebec, Canada) running on a Dell Dimension 4500S with
an Intel Pentium 4 2.0 GHz processor. These descriptors were re-
lated to the biological activity of the compounds using the IC50 data
for inhibition of NMT activity, along with the molecular structure of
each COPP analog. Detailed information about descriptors used in
the QSAR can be found at http://www.chemcomp.com/Journal_
of_CCG/Features/descrip.html.
The three-dimensional structure of the yeast NMT with bound
myristoyl-CoA and peptide substrate analogs [S-(2-oxo)pentadecyl-
CoA and SC-58272] was obtained from the Protein Data Bank using
the accession code 2nmt (Bhatnagar et al., 1998). Modeling and
energy analyses were performed using the program DOCK in the
MOE software package with the default parameter settings. Prior to
docking of 24, the peptide analog was computationally removed, and
hydrogen atoms were added to the protein. Binding site analyses
were conducted within a three-dimensional docking box encompass-
ing the entire enzyme, using simulated annealing (Monte Carlo
method) and tabu search protocols. Both methods were set to opti-
mize spatial contacts as well as electrostatic interactions as deter-
mined by the molecular mechanics force field (MMFF94). Acceptance
tests were conducted at the end of each docking cycle until an
optimal solution (lowest total energy) was attained.
Cellular NMT Activity Assays. Whole-cell assay systems for the
evaluation of inhibitors of protein farnesylation, palmitoylation, and
N-myristoylation have been recently developed in this laboratory
(J. E. Copper, D. W. Dexter, and C. D. Smith, manuscript submitted
for publication). In these assays, various lipidation signals have been
fused onto enhanced green fluorescent protein (GFP), and these
proteins are expressed by transient transfection of CV-1 cells. The
subcellular localization of the modified GFPs can then be determined
by fluorescence microscopy. The N-myr-GFP plasmid was created by
Fig. 1. NMT assay. Major components of the NMT activity assay are
indicated, and details are provided under Materials and Methods. NMT
represents recombinant GST-hNMT-1.
toyl-CoA, and test compound or DMSO. Conducting the assay
in streptavidin-coated 96-well microtiter plates resulted in
binding of the biotinylated peptide to the plates and therefore
allowed rapid separation of the [3H]myristic acid substrate
from the [3H]myristoylated peptide product.
The results of a typical NMT assay are depicted in Fig. 2.
As shown, there was a very low background level of nonspe-
cific binding of [3H]myristoyl-CoA to the plates (ϪNMT).
Also, GSH column-purified lysate from E. coli expressing a
control vector encoding GST alone showed no NMT activity.
In contrast, addition of recombinant GST-hNMT-1 resulted
in very large increases in bound radioactivity, thereby pro-
viding a sensitive method of detecting inhibitors. The intra-
assay coefficient of variation was approximately 3%, whereas
interassay variation was approximately 10%. The addition of
0.5% DMSO, the vehicle used in the chemical library, caused
only a small decrease in NMT activity. A positive control,
myristoyl-L-phenylalanine, was tested at the published IC50
concentration of 100 ␮M (Boutin et al., 1991) and did in fact
inhibit the recombinant human NMT by approximately 50%.
An example of a hit in the screening assay, 24, is also indi-
cated for comparison. Therefore, a sensitive and robust
fusing sequence corresponding to the
N terminus of Fyn
(MGCVQCKTK) to GFP using polymerase chain reaction and cloning
into the pZEOSV2- expression vector (Invitrogen, Carlsbad, CA).
LipofectAMINE 2000 transfection reagent (Invitrogen) was used to
introduce the GFP construct into CV-1 monkey kidney cells growing
on glass coverslips. The cells were incubated with 24 or the solvent
(DMSO) for 48 h. Cells were fixed with 4% paraformaldehyde in PBS,
mounted on slides, and epifluorescence microscopy was performed
using a Nikon ECLIPSE E600 microscope at 100ϫ magnification
with excitation and emission wavelengths set at 465 and 495 nm,
respectively.
Cytotoxicity Assays. The following human tumor cell lines were
obtained from the American Type Culture Collection (Manassas, VA)
and grown in Dulbecco’s modified Eagle’s medium containing 10%
fetal bovine serum and 50 ␮g/ml gentamycin following standard
techniques: DU145, HepG2, HT29, MCF-7, MDA-MB-231, Panc-1,
SKOV3, and A-498. To determine the effects of 24 on the prolifera-
tion of the cells, cells were plated into 96-well microtiter plates and
allowed to attach for 24 h. Varying concentrations of 24 were added
to individual wells, and the cells were incubated for an additional
72 h. At the end of this period, the number of viable cells was
determined using the sulforhodamine-binding assay (Skehan et al.,
1990). The percentage of cells killed is calculated as the percent of
decrease in sulforhodamine-binding compared with control cultures.
Results
Development of the NMT Screening Assay. Raju et al.
(1996) have shown that functional human NMT can be ex-
pressed in Escherichia coli, indicating that the enzyme does
not require post-translational modification for activity.
Therefore, we expressed human NMT-1 fused to GST (to
facilitate purification) in E. coli to develop a medium-
throughput screen for inhibitors of this enzyme (Fig. 1). NMT
activity was measured using purified GST-hNMT-1, biotin-
Fig. 2. An example of NMT assay data. The NMT assay was conducted in
the absence of any recombinant (ϪNMT), with purified recombinant GST
or with GST-hNMT-1 in the presence of DMSO, 6 ␮M 24, or 100 ␮M
myristoyl-L-phenylalanine. Values represent the mean Ϯ S.D. of NMT
ylated peptide with the sequence GNAASARRK, [3H]myris- activity in triplicate samples in a representative experiment.

Inhibitors of Human NMT-1
343
TABLE 1
method of screening for inhibitors of human NMT has been
established.
Compound number, R substituent, and IC50 value for inhibition of
recombinant hNMT-1 by COPPs
Identification of the COPP Chemotype of NMT In-
hibitors. To facilitate screening of the compound library
consisting of approximately 16,000 “drug-like” small mole-
cules, we developed a combinatorial method in which GST-
hNMT-1 was simultaneously incubated with eight com-
pounds. Each compound was tested at a final concentration
of 5 ␮g/ml, which corresponds to concentrations ranging from
10 to 20 ␮M. Samples that inhibited NMT activity by at least
60% were then deconvoluted, where each of the eight com-
pounds was analyzed individually. This level of inhibition
was observed in approximately 0.8% of the compounds
tested. Of the approximately 120 compounds identified as
inhibitors of NMT-1, a subset containing the common struc-
ture COPP was readily apparent. Substructure searching of
the library database indicated that 32 compounds with the
COPP chemotype and 20 containing a methyl-octahydro-pyr-
rolo[1,2-a]pyrazine (MOPP) moiety were available for analy-
ses. These compounds were then individually tested at vary-
ing concentrations to determine their potencies for inhibiting
purified hNMT-1. The IC50 values for inhibition of hNMT-1,
along with the structures of the compounds, are listed in
Table 1.
The majority of the COPP compounds had IC50 values near
or below 100 ␮M, making them adequate lead compounds for
further examination. A wide range of potencies was observed,
with 24 demonstrating the highest potency (6 ␮M) and six
compounds being extremely weak inhibitors of NMT. MOPP
compounds with similar structures but lacking the cyclo-
hexyl moiety were mostly inactive, indicating the importance
of the cyclohexyl component for inhibition of NMT (data not
shown).
Compound
ϪR
GST-hNMT-1
IC50 ␮M
No.
1
cyclohexyl
113 Ϯ 7
590 Ϯ 127
121 Ϯ 8
419 Ϯ 30
49 Ϯ 13
6400 Ϯ 310
44 Ϯ 6
Ͼ8000
Ͼ8000
284 Ϯ 46
240 Ϯ 23
48 Ϯ 5
2
3-pyridinyl
3
2-methylbenzyl
4-methoxybenzyl
4
5
4-methylsulfanylbenzyl
4-methoxycarbonylbenzyl
4-trifluoromethylbenzyl
6
7
8
2-fluorobenzyl
2-nitrobenzyl
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4-methoxy-3-methylbenzyl
3,4-dimethoxybenzyl
3,4-dichlorobenzyl
2,6-dichlorobenzyl
77 Ϯ 13
89 Ϯ 16
6000 Ϯ 250
66 Ϯ 13
50 Ϯ 2
3,4-difluorobenzyl
2,6-difluorobenzyl
4-methoxy-2,5-dimethylbenzyl
5-bromo-2,4-dimethoxybenzyl
2-bromo-4,5-dimethoxybenzyl
2-nitro-4,5-dimethoxybenzyl
2-naphthalenyl
4-methoxy-naphthalene-1-ylmethyl
2-methoxy-naphthalene-1-ylmethyl
2-nitro-benzo͓1,3͔dioxol-1-ylmethyl
9-ethyl-9H-carbazole-3-ylmethyl
10b,10c-dihydro-pyrene-1-ylmethyl
6-methyl-chromene-4-one-3-ylmethyl
4-bromo-thiophen-2-ylmethyl
5-(2-nitro-phenyl)-furan-2-ylmethyl
5-(3-nitro-phenyl)-furan-2-ylmethyl
5-(4-nitro-phenyl)-furan-2-ylmethyl
2-(2,6,6-trimethyl-cyclohex-1-enyl)-ethyl
3-(2-methoxy-phenyl)-allyl
78 Ϯ 1
60 Ϯ 6
42 Ϯ 5
35 Ϯ 5
28 Ϯ 8
63 Ϯ 15
6 Ϯ 1
QSAR Model for Inhibition of hNMT-1 by COPPs. The
number of COPP compounds evaluated was sufficient to per-
mit ligand-based interaction analyses using QSAR. The Qua-
SAR-Model program in the MOE software package was used
to calculate a large number of two-dimensional chemical
descriptors for each of the COPP compounds indicated in
Table 1 and five MOPP compounds with measurable IC50
values (40–7300 ␮M). These descriptors were then combined
with the biological data [-log(IC50) for inhibition of hNMT-1]
70 Ϯ 8
Ͼ8000
Ͼ8000
45 Ϯ 2
41 Ϯ 1
36 Ϯ 4
43 Ϯ 4
1600 Ϯ 200
and analyzed for predictive equations using partial least- together, these descriptors indicate that specific aspects of
squares analyses. Through this process, a relationship be- molecular size, shape, and polarity are critical determinants
tween the activity of the COPP compounds and their respec- for inhibition of NMT.
tive structural characteristics was calculated. The optimized
Mechanism of Inhibition of NMT by COPP Com-
QSAR equation consisting of 37 descriptors is summarized in pounds. The mechanism of inhibition by the most potent
Table 2. Statistical parameters for this equation include: root COPP compound, 24, was examined using peptide and myr-
mean square error, 0.50; correlation coefficient (r2), 0.72; istoyl-CoA substrate competition experiments with GST-
cross-validated root mean square error, 1.02; and cross-vali- hNMT-1. The results of these studies are summarized in
dated r2, 0.18. The relationship between the experimental Table 3. In the absence of inhibitor, the Km for myristoyl-CoA
and calculated IC50 values for inhibition of NMT is indicated compares very well with the literature value of 2.6 ␮M
in Fig. 3. All predicted IC50 values were within 1-log of the (Rocque et al., 1993). The affinity of the enzyme for the
experimental IC50 values with the exception of those for 14 biotinylated peptide is somewhat lower than the affinity of
and 27. The descriptors making the largest contributions to hNMT for the soluble peptide GNAASARR (Km, 6 ␮M;
the equation include: zagreb (an index describing the molec- Rocque et al., 1993). However, it is well established that the
ular connectivity and shape of the heavy atoms in the mole- affinity for the peptide is highly sequence-sensitive, so that
cule); weight (the molecular weight of the compound); Km values Ͼ100 ␮M are not unusual. For studies of the
vsa_hyd (a pharmacophore feature that approximates the ability of 24 to interact with the peptide-binding site of NMT,
sum of Van der Waal surface areas of hydrophobic atoms); the peptide concentration was varied from 0 to 500 ␮M,
and WeinerPath and WeinerPol (adjacency and distance ma- whereas the myristoyl-CoA concentration was kept constant
trix descriptors for path and polarity, respectively). Taken in the assays. The presence of 24 at its IC50 concentration

344
French et al.
TABLE 2
TABLE 3
QSAR parameters for inhibition of hNMT-1
Michaelis-Menten parameters for hNMT-1
Values represent the mean Ϯ S.D. for samples in one of three similar experiments.
Descriptor labels, correlation coefficients, and relative contributions of the 37 de-
scriptors used in the optimized QSAR model are given. An explanation of the
descriptor labels can be found at: http://www.chemcomp.com/Journal_of_CCG/Fea-
tures/descrip.html.
Peptide
Myristoyl-CoA
Vmax
Addition
Km
Vmax
Km
Relative
Importance
Descriptor Label
Coefficient
␮M
pmol/min
␮M
pmol/min
DMSO
181 Ϯ 1
252 Ϯ 2
0.084 Ϯ 0.012
0.076 Ϯ 0.006
3.0 Ϯ 1.8
2.6 Ϯ 2.0
0.057 Ϯ 0.005
0.034 Ϯ 0.001
VdistMa
WeinerPath
WeinerPol
A_count
B_count
chi0v
0.0161
Ϫ2.0716
Ϫ1.7817
0.3237
0.004
0.561
0.483
0.088
0.112
0.020
0.044
0.007
0.085
0.071
0.095
0.055
0.148
0.513
0.183
0.135
0.234
0.001
0.001
1.000
0.249
0.082
0.229
0.103
0.224
0.151
0.254
0.012
0.271
0.391
0.026
0.027
0.151
0.065
0.229
0.074
0.179
6 ␮M 24
0.4147
decreased the Km for the peptide relative to the solvent
control. However, inhibition was overcome by the addition of
excess peptide, as demonstrated by the similar Vmax values,
indicating that 24 acts as a competitive inhibitor with respect
to the peptide substrate. To determine the effects of 24 on the
myristoyl-CoA binding site, the concentration of myristoyl-
CoA was varied from 0 to 100 ␮M, whereas the peptide
concentration was held constant in the assays. As shown in
Table 3, the Km for myristoyl-CoA was unchanged by the
presence of 24. In contrast, the Vmax decreased in the pres-
ence of 24, indicating that it acts as a noncompetitive inhib-
itor with respect to the lipid substrate. Overall, these kinetic
analyses indicate that 24, and most likely the rest of the
COPP family, inhibit NMT activity by competing with the
substrate peptide for binding to the enzyme.
Docking of COPP-24 to NMT. The competition experi-
ments described above provided important structural data
that the COPP compounds interact with the peptide-binding
site of hNMT-1. This was further explored using molecular
modeling to develop a three-dimensional model of COPP
interaction with NMT. Although the three-dimensional
structure of human NMT remains to be determined, the
structure of NMT from S. cerevisiae has been solved at 2.9 Å
resolution (Bhatnagar et al., 1998). Sequence comparisons of
human and yeast NMT indicate a very high level of sequence
similarity. This homology is even more apparent in the active
site, with residues involved in peptide binding being very
highly conserved between the human and yeast isoforms. As
shown in Table 4, 25 of the 26 residues that are involved in
peptide binding are identical in the human isoform. Further-
more, 16 residues are absolutely conserved in all known
NMT isoforms. Additionally, biochemical comparisons of the
human and yeast enzymes have demonstrated that they
share a degree of similarity (Rocque et al., 1993). Thus, the
yeast NMT structure is likely to serve as an acceptable model
until human NMT structure data are available.
0.0739
chi0v_C
chi1v
Ϫ0.1627
Ϫ0.0267
Ϫ0.3120
0.2606
chi1v_C
chi0
chi0_C
Ϫ0.3518
0.2038
Ϫ0.5455
Ϫ1.8947
0.6767
Ϫ0.4979
0.8628
chi1
chi1_C
Weight
a_heavy
a_nC
b_heavy
VAdjEq
VAdjMa
Zagreb
Ϫ0.0001
0.0054
3.6917
PEOE_VSA_HYD
Q_VSA_HYD
Q_VSA_POS
Kier1
0.9180
0.3018
Ϫ0.8456
0.3802
KierA1
0.8288
Kier2
0.2431
Apol
0.9392
Mr
Ϫ0.0432
Ϫ0.9990
1.4436
a_hyd
vsa_hyd
SlogP
0.0965
SMR
Ϫ0.1007
0.5563
SMR_VSA4
SMR_VSA5
vdw_area
vdw_vol
LogP(o/w)
0.2381
Ϫ0.8456
0.2737
0.6595
As shown in Fig. 4A, the X-ray structure of the yeast NMT
with bound myristoyl-CoA and peptide substrate analogs
[S-(2-oxo)pentadecylCoA and SC-58272, respectively] was ob-
tained from the Brookline Protein Data Bank (accession code
2nmt). The binding sites for the two substrates are in close
proximity, and amino acid residues critical for providing the
“oxyanion hole” that is necessary for polarization of the myr-
istoyl-CoA amide have been identified (Bhatnagar et al.,
1998). To determine the binding site for 24, SC-58272 was
computationally removed, and hydrogen atoms were added to
protein. Default parameter settings were used, and interac-
tion energies for binding of 24 to a three-dimensional box
encompassing the entire protein were calculated using a
molecular mechanics force field (MMFF94) and simulated
annealing and tabu search protocols. Both methods were set
to optimize spatial contacts and electrostatic interactions. At
Fig. 3. Regression analysis of QSAR model for predicting COPP potency.
The experimental IC50 (Table 1) and the calculated IC50 for each of the 32
COPP and five MOPP compounds are shown. The best-fit linear regres-
sion line and 95% confidence intervals are shown.

Inhibitors of Human NMT-1
345
TABLE 4
with 24. This model suggests that the carbazole side chain is
the most potent group tested, since it can closely interact
with more NMT residues than smaller groups.
Contact residues for SC-58272 and COPP-24
The models of NMT with bound peptide analog SC-58272 or COPP-24 were used to
define amino acid residues in close proximity to these ligands. Residues within 3.0 Å
of SC-58272 are indicated by an X. Residues within 3.0 Å of the cyclohexyl or
carbazole groups, or 4.5 Å of the N2 or N5 groups of COPP-24 (as indicated below) are
designated as CH, CB, N2, or N5, respectively.
Analysis of the contact residues of NMT with the peptide
analog or 24 indicated that 16 of the 25 residues of NMT that
lie within 3 Å of the peptide analog are also shared with the
modeled binding site for 24 (Table 4). There is only one
residue (Phe-419) in close proximity to the binding site for 24
but not the peptide analog. It should be noted that the amino
acid residues are numbered according their position in the
full-length yeast NMT protein, following the example of
Bhatnagar (1998). Furthermore, of the 16 residues that are
in close proximity to 24, 13 are absolutely conserved in all
known NMT isoforms. These findings, in conjunction with
the kinetic competition studies, indicate that the COPP com-
pounds inhibit NMT activity by interacting with the peptide-
binding site.
Position
Residue
SC-58272
COPP-24
104**
105**
106**
107*
108*
111*
113**
115**
169**
205**
206*
207**
219**
221**
223*
234**
235*
334
414**
416**
417**
418**
419*
Val
Glu
Asp
Arg
Asp
Phe
Phe
Tyr
Asn
Thr
Ala
Gly
Tyr
His
Pro
Phe
Thr
Gly
Gly
Gly
Asp
Gly
Phe
Leu
Met
Leu
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CH/N2
N5
CB/N2/N5
N5
Cellular Effects of COPP-24. The assays described
above demonstrate the ability of COPP-containing com-
pounds to inhibit purified human NMT. To determine their
efficacy in intact cells, a novel assay was developed in which
the N terminus of GFP was modified to allow its myristoyl-
ation (N-myr-GFP) by endogenous NMT. This involves tran-
sient transfection of CV-1 cells with a plasmid encoding GFP
to which a myristoylation signal patterned after that of the
src family member Fyn (MGCVQCKTK) is fused. This results
in the myristoylation of the N-terminal glycine residue and
the palmitoylation of a nearby cysteine residue, and these
modifications of GFP target it to the plasma membrane (Fig.
5). Treatment of the N-myr-GFP-transfected cells with mil-
limolar concentrations of a known, very low-potency inhibitor
of NMT, 2-hydroxymyristic acid (2HM), resulted in redistri-
bution of GFP to the cytosol. Similar effects are seen when
the cells are treated with micromolar concentrations of 24,
i.e., displacement of N-myr-GFP to the cytosol. The N-termi-
nal sequence of the N-myr-GFP construct is different from
the peptide substrate used in the screening assays, demon-
strating that the COPP compounds can inhibit the myristoyl-
ation of a variety of proteins. This method provides evidence
that the inhibitors detected in the screen with recombinant
NMT can also block endogenous NMT activity in whole cells.
Since NMT is essential for the viability of yeast, and since
several proliferative signaling pathways in cancer cells in-
volve myristoylated proteins, the effects of 24 on the prolif-
eration of a panel of human tumor cell lines were determined.
As indicated in Table 5, 24 caused dose-dependent inhibition
of the proliferation of a variety of tumor cell lines. The IC50
values for 24 fell within a relatively narrow concentration
range (0.9–14.8 ␮M), suggesting that this class of compound
may be useful for the treatment of tumors that derive from a
variety of tissues. The compound also displayed sensitivity
toward paired tumor and nontumor cells, as higher potency
was observed in the human breast tumor cell line MCF-7
versus the nontumorgenic cell line MCF-10A. Of interest,
colon-derived HT29 cells demonstrated increased sensitivity
to 24, and previous work has indicated a close association of
activation of Src (which requires myristoylation for activity)
and colon cancer (Sawyer et al., 2001). Overall, the data
indicate that 24 and, likely, related COPPs have excellent
CB
CH
CH/N5
CH
CB
CB
CB
CB
CB
CB/N2
CH
CH
420**
454*
455*
X
X
X
* Residues conserved between the yeast and human isoforms.
** Residues absolutely conserved in all known NMTs.
the end of each refinement cycle, acceptance tests were run
until an optimal solution (lowest energy) was attained. After
simulated annealing was run to convergence, the lowest-
energy solution indicated that 24 binds to the enzyme in the
peptide-binding groove very close to the catalytic residues.
This is demonstrated in Fig. 4B, in which the protein has
been hidden, leaving the myristoyl-CoA analog (left side) and
an overlay of SC-58272 and 24 (right side). Compound 24
bound within the peptide-binding site in each iteration
throughout the simulation. As shown, the COPP group binds
near the site of the second and third residues of the peptide
with the pyrrolo[1,2-a]pyrazine system positioned above and
away from the peptide backbone, whereas the cyclohexyl
moiety lies in the plane of the peptide substrate. This orien-
tation appears to allow interaction of the cyclohexyl group
with hydrophobic residues of NMT, and loss of this interac-
tion probably explains the decreased activities of MOPP com-
pounds in which a methyl group is substituted for the cyclo-
hexyl moiety. Like the cyclohexyl group, the carbazole side
chain of 24 lies in plane with the substrate peptide, extending
through a large portion of the peptide-binding groove. This
orientation may be further stabilized by the presence of ar-
omatic residues of NMT, allowing in ␲-stacking interactions potential utility as anticancer agents.

346
French et al.
Fig. 4. Modeling of NMT interactions with 24.
Left panel, the crystal structure of yeast NMT
with bound myristoyl-CoA (left side) and SC-
58272 (right side, green) substrate analogs.
Right panel, protein-hidden model of binding of
the myristoyl-CoA analog (left side) and an over-
lay of SC-58272 (magenta) and 24 (yellow) in the
lowest total energy configuration.
TABLE 5
Cytotoxicity of 24 toward human tumor cell lines
Values represent the mean Ϯ S.E.M. of four experiments.
Cell Line
Tissue of Origin
IC50
␮M
DU145
Prostate
14.8 Ϯ 2.9
4.4 Ϯ 1.2
3.7 Ϯ 0.6
0.9 Ϯ 0.3
5.6 Ϯ 2.9
11.9 Ϯ 1.3
12.6 Ϯ 1.3
13.2 Ϯ 3.6
5.8 Ϯ 1.4
HepG2
Liver
HT29
MCF-7
MCF-10A B
MDA-MB-231
Panc-1
SKOV3
A-498
Colon
Breast
Breast nontransformed
Breast
Pancreas
Ovary
Kidney
IC50, concentration of 24 that kills 50% of the cells.
demonstrated potency toward hNMT-1, indicating that it is
an attractive scaffold for medicinal chemistry efforts. It is
important to note that changing the substituent connected to
the COPP nucleus markedly altered potency. For example,
substitution of the 2-benzyl position with an electron-donat-
ing group (3) produced substantially more potent compounds
than substitution with an electron-withdrawing group (8 and
9). Substitution of two fluorine atoms (15) for two chlorine
atoms (13) greatly reduced the potency of the benzyl-substi-
tuted COPPs. In general, compounds with multiple conju-
gated ring systems were more potent than substituted ben-
zyl, phenyl, or thiophenyl compounds. Overall, these results
show that the COPP chemotype inhibits human NMT-1 ac-
tivity at micromolar levels, and the structure of the side
chain can dramatically influence activity. Further medicinal
chemistry efforts will be required to achieve similar potencies
as previously described peptidomimetics (Devadas et al.,
1997) and nonhydrolyzable myristoyl-CoA analogs (Paige et
al., 1989). Structure-activity relationships obtained from the
screen revealed that most MOPP compounds showed no po-
tency. However, there were several that inhibited hNMT-1
activity (data not shown). This is most likely due to the
presence of potent side chains, which were not present in the
Fig. 5. Inhibition of endogenous NMT by 24. CV-1 monkey kidney cells
were transfected with N-myr-GFP as described under Materials and
Methods. The cells were treated with DMSO, 1 mM 2HM, or 4 ␮M 24 for
48 h before the subcellular distribution of GFP was assessed by fluores-
cence microscopy. Images represent typical results in multiple experi-
ments.
Discussion
Because of the increasing evidence that protein N-myris- initial COPP library. New COPP analogs will be synthesized
toylation plays important roles in several pathogenic states, that will incorporate these side chains and enrich the exist-
we sought to identify small-molecule inhibitors of human ing structure-activity relationship series.
NMT. Previous drug development endeavors were aimed to-
Kinetic studies revealed that the most potent COPP-based
ward designing inhibitors specific toward fungal NMT activ- inhibitor is competitive at the peptide-binding site of NMT.
ity (Masubuchi et al., 2001; Ebiike et al., 2002). We hypoth- Since COPP 24 inhibition was removed with the addition of
esize that human NMT is an attractive target for cancer and excess peptide substrate, it is unlikely that this chemotype is
viral therapeutics. To this end, we have developed an effi- a false positive from the screen and is acting in a direct
cient assay for screening using recombinant human enzyme fashion. It may be desirable to target inhibitors toward the
and a biotinylated substrate peptide and have identified a peptide-binding site for several reasons. Peptidomimetics
series of moderately potent inhibitors with activity both in tend to be more drug-like than analogs of the lipid substrate
vitro and with intact cells. The COPP chemotype consistently would be. However, peptidomimetics demonstrate poor intes-

Inhibitors of Human NMT-1
347
Likhosherstov AM, Peresada VP, and Skoldinov AP (1993) New route to the synthe-
sis of octahydropyrrolo[1, 2-a] pyrazines. Khim-Farm Zh 27:46–48.
Lodge JK, Jackson-Machelski E, Toffaletti DL, Perfect JR, and Gordon JI (1994)
Targeted gene replacement demonstrates that myristoyl-CoA: protein N- myris-
toyltransferase is essential for viability of Cryptococcus neoformans. Proc Natl
Acad Sci USA 91:12008–12012.
Masubuchi M, Kawasaki K, Ebiike H, Ikeda Y, Tsujii S, Sogabe S, Fujii T, Sakata K,
Shiratori Y, Aoki Y, et al. (2001) Design and synthesis of novel benzofurans as a
new class of antifungal agents targeting fungal N-myristoyltransferase. Part 1.
Bioorg Med Chem Lett 11:1833–1837.
McIlhinney RA, Young K, Egerton M, Camble R, White A, and Soloviev M (1998)
Characterization of human and rat brain myristoyl-CoA:protein N- myristoyl-
transferase: evidence for an alternative splice variant of the enzyme. Biochem J
333:491–495.
Nagarajan SR, Devadas B, Zupec ME, Freeman SK, Brown DL, Lu HF, Mehta PP,
Kishore NS, McWherter CA, Getman DP, et al. (1997) Conformationally con-
strained [p-(omega-aminoalkyl)phenacetyl]-L-seryl-L-lysyl dipeptide amides as
potent peptidomimetic inhibitors of Candida albicans and human myristoyl-CoA:
protein N-myristoyl transferase. J Med Chem 40:1422–1438.
Paige LA, Zheng GQ, DeFrees SA, Cassady JM, and Geahlen RL (1989) S-(2-
oxopentadecyl)-CoA, a nonhydrolyzable analogue of myristoyl-CoA, is a potent
inhibitor of myristoyl-CoA:protein N-myristoyltransferase. J Med Chem 32:1665–
1667.
Rajala RV, Radhi JM, Kakkar R, Datla RS, and Sharma RK (2000) Increased
expression of N-myristoyltransferase in gallbladder carcinomas. Cancer 88:1992–
1999.
Raju RV, Datla RS, and Sharma RK (1996) Expression of human N-myristoyltrans-
ferase in Escherichia coli. Comparison with N-myristoyltransferases expressed in
different tissues. Mol Cell Biochem 155:69–76.
Raju RV, Moyana TN, and Sharma RK (1997) N-Myristoyltransferase overexpres-
sion in human colorectal adenocarcinomas. Exp Cell Res 235:145–154.
Resh MD (1994) Myristylation and palmitylation of Src family members: the fats of
the matter. Cell 76:411–413.
tinal absorptive properties. Furthermore, lipids, such as the
nonhydrolyzable myristoyl-CoA analogs, are poor drugs,
demonstrating excessively rapid clearance, low volumes of
distribution, and systemic toxicity. Experience with the de-
velopment of inhibitors of farnesyltransferase has clearly
demonstrated the advantage of peptidomimetics and bisub-
strate analogs over farnesyl analogs. Finally, acyl-CoAs are
substrates for a number of enzymes, and therefore, achieving
selectivity for specific acyltransferases is likely to be difficult.
The studies described herein provide methods and lead com-
pounds for the development of inhibitors of human NMT. The
COPP compounds, as small molecules, may achieve improved
pharmacokinetic properties versus the peptidomimetic and
lipid analogs described above. Continuing studies will seek to
optimize these leads and to determine their in vivo therapeu-
tic potential.
References
Bhatnagar RS, Futterer K, Farazi TA, Korolev S, Murray CL, Jackson-Machelski E,
Gokel GW, Gordon JI, and Waksman G (1998) Structure of N-myristoyltransferase
with bound myristoylCoA and peptide substrate analogs. Nat Struct Biol 5:1091–
1097.
Boutin JA (1997) Myristoylation. Cell Signal 9:15–35.
Boutin JA, Clarenc JP, Ferry G, Ernould AP, Remond G, Vincent M, and Atassi G
(1991) N-myristoyl-transferase activity in cancer cells. Solubilization, specificity
and enzymatic inhibition of a N-myristoyl transferase from L1210 microsomes.
Eur J Biochem 201:257–263.
Resh MD (1996) Regulation of cellular signalling by fatty acid acylation and preny-
lation of signal transduction proteins. Cell Signal 8:403–412.
Robbins SM, Quintrell NA, and Bishop JM (1995) Myristoylation and differential
palmitoylation of the HCK protein-tyrosine kinases govern their attachment to
membranes and association with caveolae. Mol Cell Biol 15:3507–3515.
Rocque WJ, McWherter CA, Wood DC, and Gordon JI (1993) A comparative analysis
of the kinetic mechanism and peptide substrate specificity of human and Saccha-
romyces cerevisiae myristoyl-CoA:protein N-myristoyltransferase. J Biol Chem
268:9964–9971.
Rudnick DA, McWherter CA, Rocque WJ, Lennon PJ, Getman DP, and Gordon JI
(1991) Kinetic and structural evidence for a sequential ordered Bi Bi mechanism
of catalysis by Saccharomyces cerevisiae myristoyl-CoA:protein N-myristoyltrans-
ferase. J Biol Chem 266:9732–9739.
Sawyer T, Boyce B, Dalgarno D, and Iuliucci J (2001) Src inhibitors: genomics to
therapeutics. Expert Opin Investig Drugs 10:1327–1344.
Skehan P, Storeng R, Scudiero D, Monks A, McMahon J, Vistica D, Warren JT,
Bokesch H, Kenney S, and Boyd MR (1990) New colorimetric cytotoxicity assay for
anticancer-drug screening. J Natl Cancer Inst 82:1107–1112.
Boutin JA, Ferry G, Ernould AP, Maes P, Remond G, and Vincent M (1993) Myris-
toyl-CoA:protein N-myristoyltransferase activity in cancer cells. Purification and
characterization of a cytosolic isoform from the murine leukemia cell line L1210.
Eur J Biochem 214:853–867.
Chow M, Newman JF, Filman D, Hogle JM, Rowlands DJ, and Brown F (1987)
Myristylation of picornavirus capsid protein VP4 and its structural significance.
Nature (Lond) 327:482–486.
Cordo SM, Candurra NA, and Damonte EB (1999) Myristic acid analogs are inhib-
itors of Junin virus replication. Microbes Infect 1:609–614.
Devadas B, Freeman SK, Zupec ME, Lu HF, Nagarajan SR, Kishore NS, Lodge JK,
Kuneman DW, McWherter CA, Vinjamoori DV, et al. (1997) Design and synthesis
of novel imidazole-substituted dipeptide amides as potent and selective inhibitors
of Candida albicans myristoylCoA:protein N-myristoyltransferase and identifica-
tion of related tripeptide inhibitors with mechanism-based antifungal activity.
J Med Chem 40:2609–2625.
Song KS, Sargiacomo M, Galbiati F, Parenti M, and Lisanti MP (1997) Targeting of
a G alpha subunit (Gi1 alpha) and c-Src tyrosine kinase to caveolae membranes:
clarifying the role of N-myristoylation. Cell Mol Biol (Noisy-Le-Grand) 43:293–303.
Tabuchi M, Okamoto H, Furutani T, Azuma M, Ooshima H, Otake T, Kawahata T,
and Kato J (1997) Inhibition of octapeptide N-myristoylation by acyl amino acids
and acyl alkanolamines. J Enzyme Inhib 12:27–36.
van’t Hof W and Resh MD (1999) Dual fatty acylation of p59(Fyn) is required for
association with the T cell receptor zeta chain through phosphotyrosine-Src ho-
mology domain-2 interactions. J Cell Biol 145:377–389.
Wagner AP and Retey J (1991) Synthesis of myristoyl-carba(dethia)-coenzyme A and
S-(3-oxohexadecyl)-coenzyme A, two potent inhibitors of myristoyl-CoA:protein N-
myristoyltransferase. Eur J Biochem 195:699–705.
Weinberg RA, McWherter CA, Freeman SK, Wood DC, Gordon JI, and Lee SC (1995)
Genetic studies reveal that myristoylCoA:protein N-myristoyltransferase is an
essential enzyme in Candida albicans. Mol Microbiol 16:241–250.
Yurchak LK and Sefton BM (1995) Palmitoylation of either Cys-3 or Cys-5 is
required for the biological activity of the Lck tyrosine protein kinase. Mol Cell Biol
15:6914–6922.
Devadas B, Zupec ME, Freeman SK, Brown DL, Nagarajan S, Sikorski JA, Mc-
Wherter CA, Getman DP, and Gordon JI (1995) Design and syntheses of potent
and selective dipeptide inhibitors of Candida albicans myristoyl-CoA:protein N-
myristoyltransferase. J Med Chem 38:1837–1840.
Duronio RJ, Rudnick DA, Johnson RL, Johnson DR, and Gordon JI (1991) Myristic
acid auxotrophy caused by mutation of S. cerevisiae myristoyl-CoA:protein N-
myristoyltransferase. J Cell Biol 113:1313–1330.
Duronio RJ, Towler DA, Heuckeroth RO, and Gordon JI (1989) Disruption of the
yeast N-myristoyl transferase gene causes recessive lethality. Science (Wash DC)
243:796–800.
Ebiike H, Masubuchi M, Liu P, Kawasaki K, Morikami K, Sogabe S, Hayase M, Fujii
T, Sakata K, Shindoh H, et al. (2002) Design and synthesis of novel benzofurans as
a new class of antifungal agents targeting fungal N-myristoyltransferase. Part 2.
Bioorg Med Chem Lett 12:607–610.
Furuishi K, Matsuoka H, Takama M, Takahashi I, Misumi S, and Shoji S (1997)
Blockage of N-myristoylation of HIV-1 gag induces the production of impotent
progeny virus. Biochem Biophys Res Commun 237:504–511.
Garcia R, Bowman TL, Niu G, Yu H, Minton S, Muro-Cacho CA, Cox CE, Falcone R,
Fairclough R, Parsons S, et al. (2001) Constitutive activation of Stat3 by the Src
and JAK tyrosine kinases participates in growth regulation of human breast
carcinoma cells. Oncogene 20:2499–2513.
Zha J, Weiler S, Oh KJ, Wei MC, and Korsmeyer SJ (2000) Posttranslational
N-myristoylation of BID as a molecular switch for targeting mitochondria and
apoptosis. Science (Wash DC) 290:1761–1765.
Zheng J, Knighton DR, Xuong NH, Taylor SS, Sowadski JM, and Ten Eyck LF (1993)
Crystal structures of the myristylated catalytic subunit of cAMP-dependent pro-
tein kinase reveal open and closed conformations. Protein Sci 2:1559–1573.
Giang DK and Cravatt BF (1998) A second mammalian N-myristoyltransferase.
J Biol Chem 273:6595–6598.
Heuckeroth RO and Gordon JI (1989) Altered membrane association of p60v-src and
a murine 63-kDa N-myristoyl protein after incorporation of an oxygen-substituted
analog of myristic acid. Proc Natl Acad Sci USA 86:5262–5266.
Kamps MP, Buss JE, and Sefton BM (1986) Rous sarcoma virus transforming
protein lacking myristic acid phosphorylates known polypeptide substrates with-
out inducing transformation. Cell 45:105–112.
Address correspondence to: Dr. Charles D. Smith, Department of Pharma-
cology, H078, Penn State College of Medicine, Hershey, PA 17033. E-mail:
cdsmith@psu.edu