Module assembly for protein-surface recognition: Geranylgeranyltransferase I bivalent inhibitors for simultaneous targeting of interior and exterior protein surfaces

Article abstract of DOI:10.1002/chem.200701634

Synthetic chemical probes designed to simultaneously targeting multiple sites of protein surfaces are of interest owing to their potential application as site specific modulators of protein-protein interactions. A new approach toward bivalent inhibitors of mammalian type I geranylgeranyltransferase (GGTase I) based on module assembly for simultaneous recognition of both interior and exterior protein surfaces is reported. The inhibitors synthesized in this study consist of two modules linked by an alkyl spacer; one is the tetrapeptide CVIL module for binding to the interior protein surface (active pocket) and the other is a 3,4,5-alkoxy substituted benzoyl motif that contains three aminoalkyl groups designed to bind to the negatively charged protein exterior surface near the active site. The compounds were screened by two distinct enzyme inhibition assays based on fluorescence spectroscopy and incorporation of a [ ³H]-labeled prenyl group onto a protein substrate. The bivalent inhibitors block GGTase I enzymatic activity with K_i values in the submicromolar range and are approximately one order of magnitude and more than 150 times more effective than the tetrapeptide CVIL and the methyl benzoate derivatives, respectively. The bivalent compounds 6 and 8 were shown to be competitive inhibitors, suggesting that the CVIL module anchors the whole molecule to the GGTase I active site and delivers the other module to the targeting protein surface. Thus, our module-assembly approach resulted in simultaneous multiple-site recognition, and as a consequence, synergetic inhibition of GGTase I activity, thereby providing a new approach in designing protein-surface-directed inhibitors for targeting protein-protein interactions.

Full text of DOI:10.1002/chem.200701634

DOI: 10.1002/chem.200701634
Module Assembly for Protein-Surface Recognition:
Geranylgeranyltransferase I Bivalent Inhibitors for Simultaneous Targeting
of Interior and Exterior Protein Surfaces
Shinnosuke Machida,[a] Kakeru Usuba,[b] Michelle A. Blaskovich,[c] Akiko Yano,[b]
Kazuo Harada,[b] Saïd M. Sebti,[c] Nobuo Kato,[a] and Junko Ohkanda*[a]
1392
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1392 – 1401

FULL PAPER
Abstract: Synthetic chemical probes
designed to simultaneously targeting
multiple sites of protein surfaces are of
interest owing to their potential appli-
cation as site specific modulators of
protein–protein interactions. A new ap-
proach toward bivalent inhibitors of
mammalian type I geranylgeranyltrans-
ferase (GGTase I) based on module as-
sembly for simultaneous recognition of
both interior and exterior protein sur-
faces is reported. The inhibitors synthe-
sized in this study consist of two mod-
ules linked by an alkyl spacer; one is
the tetrapeptide CVIL module for
binding to the interior protein surface
(active pocket) and the other is a 3,4,5-
alkoxy substituted benzoyl motif that
contains three aminoalkyl groups de-
signed to bind to the negatively
charged protein exterior surface near
the active site. The compounds were
screened by two distinct enzyme inhibi-
tion assays based on fluorescence spec-
troscopy and incorporation of a [3H]-la-
beled prenyl group onto a protein sub-
strate. The bivalent inhibitors block
GGTase I enzymatic activity with Ki
values in the submicromolar range and
are approximately one order of magni-
tude and more than 150 times more ef-
fective than the tetrapeptide CVIL and
the methyl benzoate derivatives, re-
spectively. The bivalent compounds 6
and 8 were shown to be competitive in-
hibitors, suggesting that the CVIL
module anchors the whole molecule to
the GGTase I active site and delivers
the other module to the targeting pro-
tein surface. Thus, our module-assem-
bly approach resulted in simultaneous
multiple-site recognition, and as a con-
sequence, synergetic inhibition of
GGTase I activity, thereby providing a
new approach in designing protein-sur-
face-directed inhibitors for targeting
protein–protein interactions.
Keywords: enzymes · GGTase I ·
inhibitors
mimetics
·
proteins
·
proteo-
AHCTREUNG
Introduction
Hamilton and co-workers used an elegant approach to
create synthetic mimics of the antibody complementarity de-
termining region (CDR) with nonpeptidic scaffolds, such as
calixarenes[12,13] and porphyrins,[14–16] that accumulate multi-
ple functional groups for targeting protein surfaces. These
large synthetic antibody mimics were shown to require a
large molecular surface for strong binding to the target. This
approach demonstrated that the functionality of natural an-
tibodies can be possibly mimicked by rationally designed
synthetic molecules. However, in these examples, the molec-
ular weight of the synthetic scaffold was still large. Increas-
ing the binding affinity while decreasing the size of the mol-
ecule that is necessary for selective targeting is a challenging
problem. To overcome this problem, we propose a strategy
for protein-surface recognition that involves anchoring an
exterior surface-binding module to an interior surface-bind-
ing module. The resulting bivalent compounds are expected
to show increased affinity towards the target protein while
leading to a decrease in the size of the protein-surface rec-
ognition scaffold required owing to a synergetic effect. One
approach to test this hypothesis is the case of enzyme-inhibi-
tor design in which a substrate mimic with high binding af-
finity and selectivity to the enzyme active site is linked to an
exterior binding module that targets the appropriate sites on
the protein surface. The resulting bivalent compounds are
predicted to increase binding affinity (taking advantage of
DG), and therefore the size of exterior binding module
should decrease compared with that required for binding to
the target by itself.
Protein–protein interactions play a central role in numerous
biological processes, including programmed cell death, pro-
liferation and differentiation, and aging. In humans, roughly
40000 to 200000 protein interactions are estimated to
exist.[1] Low-molecular-weight compounds that modulate
such interactions in a selective manner are useful as probes
to elucidate biological functions as well as drug leads for
new therapeutic agents to target specific protein interfaces.
Much effort has recently focused on the rational design of
inhibitors based on proteomimetic strategies[2,3] to modulate
protein–protein interactions, and a number of studies have
been reported with success in this field.[4–6] For example, a
general solution called “protein grafting” has been devel-
oped for transferring the functional epitope into a rigid min-
iature protein scaffold.[7,8] Similar approaches in combina-
tion with phase-display technique have also recently been
developed with b-sheet templates[9,10] and b-hairpin cyclic
peptidomimetics.[11]
[a] S. Machida, Prof. N. Kato, Prof. J. Ohkanda
The Institute of Scientific and Industrial Research (ISIR)
Osaka University, 8–1 Mihogaoka, Ibaraki, Osaka 567–0047 (Japan)
Fax : (+81)6-6879-8474
E-mail: johkanda@sanken.osaka-u.ac.jp
[b] K. Usuba, A. Yano, Prof. K. Harada
Department of Life Sciences
Tokyo Gakugei University, Koganei, Tokyo 184–8501 (Japan)
Strategies for bivalent or bisubstrate receptors have been
utilized in recent chemical-biology studies, such as protein
dimerization,[17,18] lead identification,[19] protein modifica-
tions,[20] and enzyme inhibitors.[21–30] Srivastava, Mallik, and
co-workers have developed an efficient strategy to target
the surface-exposed histidine side chains by using an imino-
[c] M. A. Blaskovich, Prof. S. M. Sebti
Drug Discovery Program at H. Lee Moffitt Cancer Center and Re-
search Institute, Department of Molecular Medicine
University of South Florida, Tampa, FL 33612–9497 (USA)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2008, 14, 1392 – 1401
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1393

J. Ohkanda et al.
diacetate copper(II) complex.[24,31] In these examples, how-
ever, the motifs for the protein surface were designed for
pinpoint binding of one or two amino acid side chains in-
stead of targeting the large (ꢀ1600 Š2) surface areas that
are typically involved in protein–protein interactions.[2–4] For
modulating protein-surface functions, the anchoring strategy
of a relatively large exterior binding motif with a smaller in-
terior binding motif would provide a new approach for de-
signing site-specific protein-surface receptors. As far as we
are aware, there are no examples of nonpeptidic bivalent
enzyme inhibitors that target the large protein surface areas
that are involved in native protein–protein interactions.
Type I geranylgeranyltransferase (GGTase I)[32] is a heter-
odimeric zinc metalloenzyme and a member of the protein
prenyltransferase family. This family of enzymes catalyzes
the first step in a lipid post-translational modification that
affects nearly 0.5% of cellular proteins.[33] Mammalian
GGTase I is responsible for transferring a C20 geranylgeran-
yl group from geranylgeranyl pyrophosphate (GGPP) to a
cysteine residue of the carboxy-terminus CAAX tetrapep-
tide of a target substrate protein; where C is cysteine, AA is
an aliphatic dipeptide, and, in most cases, X is leucine or
phenylalanine.[34] During the last decade, protein prenylation
has become a major focus in anticancer research, and the
major effort in developing prenyltransferase inhibitors fo-
subunit protein surface near the entrance to the active
pocket.[56–58] A number of studies have demonstrated that an
unusual polylysine region near the carboxyl terminus of K-
Ras4B is a critical determinant for geranylgeranylation cata-
lyzed by GGTase I when FTase is inhibited.[59–62] Therefore,
it is most likely that when the enzyme binds to K-Ras4B,
the electrostatic attraction triggers the protein–protein inter-
action at the protein surface of GGTase I. In other words,
GGTase I can be regarded as an enzyme that utilizes two
distinct protein surfaces for substrate recognition in its
native task, one is the interior surface of the substrate-bind-
ing cavity and the other is the protein exterior surface near
the active site.
We therefore applied the concept of bivalent inhibitor
design for simultaneous targeting of both surfaces of GGTa-
se I (Figure 1). The interior surface-binding module
cused on
a related enzyme called farnesyltransferase
(FTase). The aim was the specific inhibition of malignant
transformation caused by mutated Ras and other farnesylat-
ed proteins.[35–37] The preclinical evaluation of FTase inhibi-
tors was successful, with little toxicity, and several com-
pounds are currently in clinical trials.[38]
Figure 1. A schematic representation of the inhibitor design for simulta-
neous targeting of interior and exterior protein surfaces. Bivalent hybrid
compound consist of an exterior surface-binding module (red) derived
from a methyl gallate, an alkyl spacer (green), and an interior surface-
binding module, Cys-Val-Ile-Leu-OH tetrapeptide.
Recently, GGTase I has been gaining much attention[39,40]
owing to its potential to be a new target, not only for anti-
cancer drugs,[41–45] but also for other diseases such as smooth
muscle hyperplasia[46] and hepatitis C.[47] Its substrate pro-
teins, such as RhoA,[48] RhoC,[49] Rac1,[50] Cdc42,[51] and
RalA,[52] have been found to be implicated in promoting tu-
morigenesis and cancer metastasis. Very recently, a study
using a GGTase I deficient mouse model has shown that
GGTase I deficiency apparently improves the percentage of
survival in mice with K-Ras-induced lung cancer, supporting
the potential of GGTase I to be a clinical target.[44] A further
significant observation that has been reported is that K-
Ras4B, the most frequently mutated form of Ras in human
tumors, becomes a substrate for GGTase I when FTase is
blocked, and retains full biological activity.[53–55] As at least
5–10 times more proteins are geranylgeranylated than are
farnesylated and the complex signaling pathways involving
geranylgeranylated proteins have not been fully character-
ized, the development of potent and selective GGTase I in-
hibitors would provide useful tools for protein prenylation
studies.
(Figure 1, blue) should bind into the active pocket in a se-
lective manner and should anchor the whole molecule near
the pocket, whereas the exterior surface-binding module
(Figure 1, red) requires relatively divergent structural fea-
tures and multiple positively charged groups for the electro-
static interaction with the GGTase I surface. To explore the
minimal and favorable distance in between the modules, we
used different length spacers (Figure 1, green).
Herein, we describe the rational design of GGTase I biva-
lent inhibitors for simultaneous targeting of interior and ex-
terior protein surfaces based on the module assembly strat-
egy described in Figure 1.
Results and Discussion
Inhibitor design and chemistry: According to the GGTase I
crystal structure, the substrate-binding pocket opens into the
extensive a–b subunit interface and extends into the hydro-
phobic funnel-shaped cavity of the b subunit with an ap-
proximate inner diameter of 15 Š and a depth of 14Š. [63]
Near the entrance of the binding pocket, there is a charac-
teristic cluster of acidic amino acid residues on the a-subunit
protein surface where Glu125a, Glu160a, Glu161a,
Glu187a, Asp191a, Glu229a, and Asp230a are located on
Mammalian GGTase I consists of a 48-kDa a subunit and
a 43-kDa b subunit, which possesses the hydrophobic active
pocket. The crystal structures of the ternary complex of
GGTase I bound to the peptide and a GGPP analogue have
revealed that there is a characteristic acidic region on the a-
1394
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1392 – 1401

Protein-Surface Recognition
FULL PAPER
Scheme 1. Modules and hybridized compounds synthesized in this study.
the edge of the pocket lip. The compounds prepared on the
basis of the design strategy are summarized in Scheme 1.
For the interior surface binding module, Cys-Val-Ile-Leu-
OH tetrapeptide (1), the CAAX sequence that is seen in
GGTase I substrate proteins Rap-2b, and several others,[64]
became our first choice because of its affinity to GGTase I
with selectivity (IC50/GGTase I=2.3 mm; IC50/FTase=
11 mm).[65] The size of the acidic surface of GGTase I is ap-
out the CVIL moiety was installed. The resulting model
showed that the trisubstituted benzoyl part provides an ap-
propriate shape to bind the targeted acidic region of GGTa-
se I (Figure 2).
2
proximately 90 Š and contains the seven acidic amino acid
residues as well as several hydrophobic or neutral residues,
such as Leu156a, Ala157a, Gln195a, and Asn194a. A mo-
lecular-modeling study showed that the methyl-3,4,5-trial-
koxy-substituted benzoate derived from b-homophenylala-
nine derivative (3) possibly provides a complementary size
(ꢀ84Š 2). In a further attempt to reduce the size of the exte-
rior surface-binding module, n-propyl amine substituted de-
rivative (4), in which three benzyl groups in 3 were re-
moved, was also designed. These compounds were readily
synthesized by alkylation of methyl gallate with the corre-
sponding N-protected amino alkyl bromides, followed by de-
protection (see the Supporting Information). These exterior
surface-binding modules were attached to the N terminal of
1 by spacers of different lengths to give the bivalent hybrid
compounds 5–9.
To check whether the bivalent molecular sizes and shapes
are relevant for the simultaneous two-site recognition, a su-
perimposed model structure of 5 and GGTase I was built
based on the previously reported X-ray crystal structure of
the ternary complex of GGTase I[58] bound to a GGPP ana-
logue and KKKSKTKCVIL. Within the complex, the por-
tion of KKKSKTK was removed and the structure of 5 with-
Figure 2. A superimposed model of 5 (yellow, CPK) and GGTase I crystal
structure (PDB: 1N4Q). The catalytic Zn ion (green, CPK), GGPP ana-
logue (stick); The bright area (right) and shadowed portions are the a
and b subunits, respectively. Negatively charged and positively charged
surfaces are shown in red and blue, respectively. DS Viewer 6.0.
Chem. Eur. J. 2008, 14, 1392 – 1401
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1395

J. Ohkanda et al.
Initial screening for inhibition activity by fluorescent
enzyme assay: To examine whether the module-assembly
approach results in bivalent inhibitors with better potency
towards GGTase I, the compounds were initially screened at
5 mm for their ability to inhibit GGTase I in a fluorescent
assay[66] by using geranylgeranyl pyrophosphate (GGPP;
5.0 mm) and the environmentally sensitive fluorogenic sub-
strate, N-dansyl-Gly-Cys-Val-Ile-Leu-OH (DansGCVIL;
1.0 mm, Km =(0.28ꢁ0.04) mm; dansyl= 5-(dimethylamino)-
naphthalene-1-sulfonyl; see Scheme S21 in the Supporting
Information) in Tris-HCl buffer solution (Tris=tris(hydroxy-
methyl)aminomethane). This substrate increases the fluores-
cent intensity upon geranylgeranylation at the cysteine thiol
group. Recombinant mammalian GGTase I was expressed
and purified by a method previously reported.[67] The fluo-
rescent increase was monitored for 5 min and the percent-
age of inhibition was calculated by comparison with the
standard slope, which was taken from the reaction in the ab-
sence of inhibitors (see the inset in Figure 3A). The percent-
age of inhibition of GGTase I at a compound concentration
of 5 mm is shown in Figure 3A. As expected, the various
modules of the bivalent compounds were weak GGTase I in-
hibitors on their own. For example, CVIL (1), which has
been shown to be a competitive inhibitor for GGTase I,[65]
inhibited 42% of the enzyme activity (see the inset of Fig-
ure 3A); and compounds 2–4 inhibited the enzyme by 2–
26%. However, the bivalent compound 6, in which the mod-
ules 1 and 3 were hybridized, and compound 8, in which
modules 1 and 4 were linked, were both found to be more
effective than either module alone and showed more than
87% inhibition. Concentration–response studies showed
that the compound 6 inhibition curve is sigmoidal with an
IC50 value of 1.0 mm and a Ki of (0.22ꢁ0.04) mm, both of
which were calculated by using the Cheng–Prusoff equation
(Figure 3B).[68] In contrast, the curves for the modules 1, 2,
and 3 were apparently shifted to the higher concentration
region. For example, the concentration required for 50% in-
hibition in the case of the tetrapeptide CVIL (1) was 6.9 mm,
which was approximately seven times higher than that of 6.
Concentration–response curves of the bivalent compounds
(5–7) containing spacers of different lengths are summarized
in Figure 4A. The results show that the spacer length affects
the inhibitory activity of the bivalent compounds. Com-
pound 7 (n=3), which consists of a shorter spacer for the
C2 unit than 6 (n=5), was slightly less active than 6 (Ki =
(0.48ꢁ0.11) mm). On the other hand, compound 5 (n=11),
with a spacer length twice as long as that of 6 was less
potent. This suggests that for compounds with the b-homo-
phenylalanine type module, the longer spacers diminish the
binding affinity to GGTase I owing to the entropical disad-
vantage.[69] However, as shown in Figure 4B, compound 8,
which has an equal length of spacer to that of 5 and a much
simpler exterior module, restores the activity with a Ki value
of (0.42ꢁ0.09) mm. Elimination of the three benzyl groups
from the exterior binding module in 5 should result in in-
creased flexibility as well as decreased hydrophobicity
around the three amino side chains of the benzoyl moiety in
Figure 3. Fluorescent GGTase I inhibition assays were carried out by
using GGPP (5 mM) and the fluorogenic substrate, DansGCVIL (1 mM)
in 50 mm Tris-HCl, pH 7.5 at 308C. A) The results of primary screening
of the % inhibition of GGTase I in the presence of 5 mm of module com-
pounds (1–4) and the bivalent compounds (6 and 8). Inset: the time
course fluorescence change at 520 nm (ex: 340 nm) in the absence of
(blue) or in the presence of 5 mm of inhibitors (1, blue; 6, red). The
slopes were used for the calculations of percentage of inhibition. B) Con-
centration–response curves of the modules (1, 2, and 3) and the corre-
sponding hybridized compound 6 plotted the % inhibition against the in-
hibitor concentrations. The IC50 values were then used to calculate the in-
hibitory constants, Ki, by the Cheng–Prusoff equation (see refer-
ence [68]). The standard deviation is given for n=3. dIf =change in fluo-
rescence intensity.
8 (calculated log P values: ꢂ0.9 for 3, 5.08 for 4).[70] There-
fore, a possible explanation for the significant difference in
the activity might be 1) with the long spacer (n=11), the
three 1-amino-1-benzyl-propyloxy groups in 5 do not take a
suitable conformation for binding to the targeted surface
owing to rigidity, whereas 1-amino-propyloxy groups are
more flexible, thus the spacer length (n=11) was able to de-
liver the moiety to the targeted site; or 2) the high hydro-
phobicity of 5 might result in random binding or possible ag-
gregation. This observation prompted us to synthesize com-
pound 9 in which the better aminopropyl module and the
1396
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1392 – 1401

Protein-Surface Recognition
FULL PAPER
more widely used method, which measures the incorpora-
tion of [3H]GGPP and [3H]FPP (farnesylpyrophosphate)
into the protein substrates, H-Ras-CVLL and H-Ras-CVLS,
respectively, by using previously described methods.[71] The
moderate inhibition activities against GGTase I of the tetra-
peptides 1 and 2 were confirmed as shown in Table 1. No in-
Table 1. IC50 values for inhibition of GGTase I and FTase in vitro by
modules (1–4) and bivalent inhibitors (5–8).
IC50 [mm][a]
Cmpd
GGTase I
FTase
1
2
3
4
5
6
7
8
4.8ꢁ1.7
1.4ꢁ0.5
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
0.98ꢁ0.45
0.64ꢁ0.06
0.66ꢁ0.05
0.60ꢁ0.03
[a] Where the standard deviation is given for n=3, or n=1.
hibition activity against FTase was detected with 1 or 2,
which is in agreement with the fact that CVIL is a selective
inhibitor for GGTase I over FTase.[65] Neither benzoate de-
rivative 3 nor 4 showed any activity. All the bivalent com-
pounds (5–8) were confirmed to be potent for GGTase I
with IC50 values in the submicromolar range (Table 1) and
most importantly, no inhibition was detected against FTase
(IC50: GGTase I/FTase>167 in the case of 8). The selectivity
observed here clearly indicates that the CVIL module in
each bivalent compound recognizes and binds to the GGTa-
se I active pocket but not to the FTase pocket. These results
demonstrate that specificity is not compromised by our
design strategy.
The spacer-length effect was also observed in this assay,
showing that compounds 6 and 7 (n=5 and 3, respectively)
were more potent (IC50 =0.64and 0.66 mm, respectively)
than compound 5 (n=11, IC50 =0.98 mm). It is noteworthy
that compound 8, a with smaller and less hydrophobic exte-
rior binding module than 6, was found to be equally effec-
tive with an IC50 value of 0.60 mm, which was an improve-
ment of eight times and greater than 167 times those of the
corresponding modules, 1 and 4, respectively. Again, these
results confirmed that the molecular size of the hybrid in-
hibitors can be optimized by appropriately combining the
modules.
Figure 4. A) Dose–response curves for A) the bivalent compounds (5, 7)
with different a different spacer length than 6 and B) the methyl gallate
derivative (4) compared with bivalent compound 8 and 9, which consist
of the methyl gallate as the exterior surface-binding module. The data set
for 6 (A) and 5 (B) are shown for comparison (dashed line). The stan-
dard deviation values are given for n=3.
optimal length of spacer (n=5) were combined. As shown
in Figure 4B, 9 indeed showed improved activity compared
with 8 with an IC50 value of 0.62 mm and a Ki of (0.14ꢁ
0.03) mm, which has a similar activity to 6 and with a reduced
molecular weight of 270. This suggests that the molecular
size can be tuned by choosing 1) the minimal critical func-
tional groups for the exterior binding module and 2) the
most appropriate spacer length for the combination of mod-
ules. Notably, compound 4 showed no apparent inhibitory
activity by itself (Figure 4B). The weak activity of modules
1 and 4, and the synergistic effect seen with the bivalent in-
hibitor 9, validates our module assembly approach that tar-
gets two critical sites on GGTase I.
Mode of inhibition: We next wanted to confirm that the in-
hibition mode of the bivalent inhibitor and the CVIL tetra-
peptide was identical. To this end, Lineweaver–Burk analy-
sis was carried out for peptide 1 and the bivalent com-
pounds 6 and 8. The data were analyzed by using equations
for competitive, noncompetitive, and uncompetitive inhibi-
tion models. The data set for 1 clearly fit the competitive
model (Figure 5A; Ki =(0.97ꢁ0.19) mm). As shown in Fig-
ure 5B for 6, these data also showed the best fit to the com-
petitive model with an intersection point on the y axis (Ki =
Selectivity in GGTase I inhibition: To confirm the observa-
tion seen in the fluorescent assay and to evaluate the selec-
tivity of our bivalent inhibitor for GGTase I over the related
enzyme, FTase, we further tested the compounds (1–8) by a
Chem. Eur. J. 2008, 14, 1392 – 1401
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1397

J. Ohkanda et al.
active-site-bound form is essential to dictate whether S-far-
nesylation occurs. Assuming that this is also the case in
GGTase I, which is likely owing to its functional similarity,
the binding of the exterior module to the protein surface
may trigger structural changes of the interior binding
module CVIL to avoid S-geranylgeranylation.
Conclusion
We have developed a new series of bivalent inhibitors for
GGTase I based on a module assembly strategy for simulta-
neous targeting of interior and exterior protein surfaces.
Coupling of CVIL tetrapeptide and a 3,4,5-trisubstituted
benzoate derivative provided compounds 6 and 8 (Ki =0.15
and 0.21 mm, respectively) and improved the inhibitory abili-
ty by approximately one order of magnitude and more than
150-fold compared with the corresponding modules. An ex-
planation for the moderate additivity effect observed with
the bivalent compounds compared with tetrapeptide 1 might
be due to the remarkably weak affinity of the exterior bind-
ing modules 3 and 4. All bivalent compounds were found to
be highly selective for GGTase I over FTase (>167 times).
The spacer length was found to affect the potency, and very
long spacers (n=11) diminished the activity as seen in the
case of 5, which is consistent with a previous report.[69]
Simply changing the exterior binder in 5 to one with less hy-
drophobic and more flexible side chains restored the activi-
ty, as seen in 8. Thus, bivalent inhibitors targeting the exteri-
or and the interior surfaces can be minimized by choosing
the most appropriate modules and the spacer lengths. Inter-
estingly, geranylgeranylation was not detected, at least in
the case of 6, suggesting that the exterior surface binding
module may alter the CVIL conformation in the active
pocket. Although it remains moderate, the synergetic inhibi-
tory activity observed in this study strongly suggests that the
exterior binding modules were specifically delivered to the
targeted protein surface. Thus, we believe that the module-
assembly strategy for designing bivalent protein inhibitors
described herein provides a general strategy for specific tar-
geting of the protein surface, which can be extended to
other enzyme families for increased selectivity or for modu-
lating protein–protein interactions.
Figure 5. Kinetic analysis of the inhibition of GGTase I by CVIL tetra-
peptide 1 and bivalent inhibitor 6. GGTase I was treated with varying
concentrations of A) the tetrapeptide 1 (0.1, 1, and 5 mm) and B) the bi-
valent compound 6 (0.5, 1, and 3 mm) with the substrate concentration in-
creasing from 0.1 to 1 mm. [GGPP]=5 mm, T=293 K.
(0.15ꢁ0.01) mm for 6 and (0.21ꢁ0.03) mm for 8; see Fig-
ure S2 in the Supporting Information for 8).
As we used the native substrate peptide sequence for the
interior surface binder, we wanted to clarify whether gera-
nylgeranylation occurred on the thiol group of the bivalent
compounds. The HPLC analysis showed that in the absence
of inhibitor, the conversion of DansGCVIL to
DansGC(GG)VIL was completed within 5 minutes (see Fig-
ure S3A in the Supporting Information), whereas under the
same conditions but in the presence of 20 mm 6, the geranyl-
geranylation of DansGCVIL was completely suppressed
(see Figure S3B in the Supporting Information). Moreover,
no change in the peak area of 6 nor new-product peak gen-
eration was detected for up to 50 minutes incubation with 6,
indicating that geranylgeranylation did not occur on the
thiol group of 6 (see Figure S3C in the Supporting Informa-
tion). This is consistent with previous work with FTase in
which some CAAX tetrapeptide mimetic inhibitors are not
farnesylated. The crystal structural study has revealed that
these FTase inhibitors bind in a distinct conformation from
the substrate,[72] suggesting that the conformation in the
We are currently extending this strategy to isozyme selec-
tive inhibitors and to dual inhibitors targeting characteristi-
cally distinct and identical protein-surface structures.
Experimental Section
Materials and instruments: Reagents and solvents were obtained from
commercial sources without further purification unless otherwise noted.
For the fluorescent enzyme assay, RF-5300PC equipped with a SA-100
temperature controller (Sansyo) was used. HPLC data were obtained
from PU-2086 and a UV-2075 detector (JASCO) with Inertsil (C18-re-
verse phase column) column (GL Science). 1H and 13C NMR spectra
were recorded on a JEOL JNM-LA 400 spectrometer. Low- and high-
resolution mass spectra were collected by a JEOL JMS-T100 LC mass
1398
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1392 – 1401

Protein-Surface Recognition
FULL PAPER
spectrometer. Elemental analyses were performed on 2400CHN (Perkin
Elmer).
(m, 4H, -CH2CH2CO2CH3 and -CONHCH2CH2), 1.76–1.91 (m, 6H, 3-, 4-,
and 5-OCH2CH2), 2.20 (t, J=7.1 Hz, 2H, -CH2CO2CH3), 2.67–2.73 (m,
6H, 3-, 4-, and 5-PhCH2), 3.18–3.23 (m, 2H, -CONHCH2), 3.84–3.99 (m,
9H, 3-, 4-, 5-OCH2CH2CH), 6.59 (d, J=8.8 Hz, 1H, 4-NHBoc), 6.81 (d,
J=8.5 Hz, 1H, 3- and 5-NHBoc), 7.09–7.26 (m, 17H, 3,4,5-Ph and benzo-
yl) and 8.34(m, 1H, benzoyl-CONH) and 11.9 ppm (br s, 1H, -CO 2H);
HRMS–FAB (m/z): [M++H] calcd for C58H81N4O12 [M++H]: 1025.5851;
found: 1025.5835.
Synthesis of 3,4,5-tris(3-amino-4-phenyl-1-butoxy)benzoic acid methyl
ester trifluoroacetate (3):
A solution of methyl gallate (101 mg,
0.549 mmol), [(1S)-3-bromo-1-(phenylmethyl)propyl]carbamic acid tert-
butyl ester (31; 700 mg, 2.15 mmol; see the Supporting Information), po-
tassium carbonate (444 mg, 3.21 mmol) in N,N-dimethylformamide
(DMF; 20 mL) was stirred at 408C for 24h. After removal of DMF, the
product was extracted with 10% CHCl3 in AcOEt (600 mL) and 10%
citric acid, and the organic layer was dried over Na2SO4. The crude prod-
uct was purified by SiO2 column chromatography (gradient from CHCl3
alone, to CHCl3:AcOEt=1:1, and then to CHCl3:MeOH=50:1) to give
PyBop (50 mg, 96 mmol) in CH2Cl2 (1 mL) was added to a solution of the
free acid 38 (72 mg, 70 mmol), H-CysACHTREU(NG Trt)-Val-Ile-Leu-OtBu (Trt=trityl;
66 mg, 89 mmol), HOBt (25 mg, 0.16 mmol), and DIEA (25 mL,
015 mmol) in DMF (6 mL) at 08C. After stirring at room temperature
for 17 h, concentration, extraction with CHCl3, and purification by size-
exclusion column chromatography (Sephadex LH-20, CHCl3/MeOH=
1:1) gave the fully protected product (41; see the Supporting Informa-
3,4,5-trisACHTREUNG(3-N-(tert-butoxy)amino-4-phenyl-1-butoxy)benzoic acid methyl
ester (32; see the Supporting Information) as a white solid (464 mg,
91%). m.p.: 167–1688C; 1H NMR (400 MHz, CDCl3): d=1.38 (m, 27H,
Boc”3 (Boc=tert-butyloxycarbonyl), 1.82–2.10 (m, 6H, 3-, 4-, and 5-
OCH2CH2), 2.77–2.87 (m, 6H, 3,4, 5-CH2Ph), 3.88 (s, 3H, -OCH3), 4.07–
4.12 (m, 9H, 3,4,5-OCH2CH2CH), 4.69 (br s, 2H, 3- and 5-NH), 5.28 (br
s, 1H, 4-NH) and 7.17–7.28 ppm (m, 17H, aryl H); HRMS–FAB (m/z):
[M++H] calcd for C53H72N3O11, 926.5167; found, 926.5152. Trifluoroacetic
acid (TFA; 500 mL) in CH2Cl2 (500 mL) was added to a solution of 32
(25 mg, 27.0 mmol) and the mixture was stirred at 08C for 1 h. After
evaporation, Et2O was added to the residue and the resulting white pre-
cipitate was collected by centrifugation to give compound 3 (30 mg,
100%). HPLC purity, 99%; 1H NMR (400 MHz, CDCl3): d=2.11–2.15
(m, 6H), 2.86 (m, 3H), 3.18 (m, 3H), 3.67 (m, 3H), 3.88 (s, 3H, OCH3),
3.95 (s, 1H), 4.10 (s, 1H), 4.20 (s, 2H), 4.34 (s, 2H), 7.15–7.34 (m, 17H),
8.07–8.15 ppm (br s, 9H, NH3); HRMS–FAB (m/z): [M++H] calcd for
C38H48N3O5, 626.3594; found, 626.3605.
tion) as
a
white solid (86 mg, 70%). m.p.: 179–1818C; 1H NMR
(400 MHz, CDCl3): d=0.68–0.86 (m, 18H, g-CH3 Val, g-CH3 Ile, d-CH3
Ile, and 2d-CH3 Leu), 0.97–1.04(m, 1H, g-CH Leu), 1.13–1.86 (m, 54H,
g-CH2 Ile and b-CH2 Leu, b-CH Ile, b-CH Val, 3-, 4-, and 5-OCH2CH2,
-CONHCH2ACTHER(UNG CH2)3, OtBu and 3-, 4-, 5-Boc), 2.08 (m, 2H, b-CH2, Cys),
2.33 (m, 2H, -CH2CONH), 2.67–2.73 (m, 6H, 3-, 4-, and 5-CH2Ph), 3.17
(m, 2H, -CONHCH2), 3.84–3.98 (m, 9H, 3-, 4-, and 5-OCH2CH2CH),
4.12–4.27 (m, 4H, a-CH Cys, Val, Ieu, and Leu), 6.55 (d, J=8.4Hz, 4-
NHBoc), 6.78 (d, J=8.8 Hz, 2H, 3- and 5-NHBoc), 7.09–7.33 (m, 32H,
aryl-H), 7.50 (d, J=8.9 Hz, 1H, NH), 7.88 (d, J=8.8 Hz, 1H, NH), 8.09
(d, J=8.3 Hz, 1H, NH) 8.16 (d, J=7.9 Hz, 1H, NH), 8.31 ppm (m, 1H,
CONH); elemental analysis calcd (%) C101H138N8O16S·1.0H2O: C 68.52,
H 7.97, N 6.33; found: C 68.57, H 7.87, N 6.22; HRMS–FAB (m/z): [M+
+H] calcd for C101H139N8O16S1, 1752.003; found, 1752.0000.
This compound was deprotected by treatment with 50% TFA in CH2Cl2
in the presence of 5% triethylsilane at 08C for 1 h. After concentration,
the residue was suspended in Et2O, sonicated, and centrifuged to collect
the resulting white solid, which was washed with Et2O several times, and
dried to give 6, a colorless white powder (21 mg, 100%). m.p.: 154–
1628C; 1H NMR (400 MHz, CDCl3): d=0.78–0.86 (m, 18H, 2g-CH3 Val,
g-, d-CH3 Ile, and 2d-CH3 Leu), 1.02–1.11 (m, 1H, g-CH Leu), 1.24–1.96
Synthesis of N-[6-(3,4,5-tris(3-amino-4-phenyl-1-butoxy)benzolyamino)-
hexylcarbonyl]-l-cisteinyl-l-valyl-l-isoleucyl-l-leucine
trifluoroacetate
(6): 1m KOH (1.3 mL, 1.3 mmol) was added to a solution of 32 (43 mg,
46.4 mmol) in CH2Cl2 (4mL) and MeOH (8 mL) and the mixture was re-
fluxed for 15.5 h. Evaporation, extraction with CHCl3 (60 mL”3) and
10% citric acid (25 mL), and concentration gave the corresponding car-
boxylic acid (33, see the Supporting Information) as a white solid (43 mg,
1
(m, 18H, 3,4,5-OCH2CH2, -CONHCH2ACHTRE(UNG CH2)3, b-CH Val, b-CH and g-
100%). m.p.: 141–1448C; H NMR (400 MHz, CDCl3): d=1.38 (m, 27H,
CH2 Ile, and b-CH2 Leu), 2.13–2.18 (m, 2H, -CH2CONH), 2.62–2.78 (m,
2H, b-CH2 Cys), 2.83–2.97 (m, 6H, 3,4,5-PhCH2), 3.90–3.93 (m, 2H, a-
CH”2), 4.09–4.20 (m, 7H, 3-, 4-, and 5-OCH2 and a-CH), 4.40–4.45 (m,
1H, a-CH), 7.19–7.34(m, 17H, 3-, 4-, and 5-Ph and benzoyl), 7.79–7.86
(m, 2H, -NHCH), 8.06–8.14(m, 2H, -N HCH), 8.36–8.39 ppm (m, 1H,
PhCONH); HRMS–FAB (m/z): [M++H] calcd for C69H104N8O10S1,
1153.6735; found, 1153.6702.
Boc”3), 1.82–2.05 (m, 6H, 3-, 4-, and 5-OCH2CH2), 2.80–2.87 (m, 6H, 3-,
4-, and 5-CH2Ph), 4.09–4.13 (m, 9H, 3-, 4-, and 5-OCH2CH2CH), 4.69
(br s, 2H, 3 and 5-NH), 5.28 (br, 1H, 4-NH) and 7.17–7.28 ppm (m, 17H,
aryl H); HRMS–FAB (m/z): [M++H] calcd for C52H7N3O11, 912.5010;
found, 912.4990. Benzotriazol-1-yloxytris(pyrrolidino)phosphonium hexa-
fluorophosphate (PyBop; 67 mg, 0.13 mmol) in CH2Cl2 (1 mL)was added
to a solution of 33 (90 mg, 98.7 mmol), 6-aminohexanoic acid methyl ester
hydrochloride (19 mg, 0.10 mmol), N-hydroxy-benzotriazole (HOBt;
31 mg, 0.20 mmol), and N,N-diisopropylethylamine (DIEA; 34 mL,
0.20 mmol) in DMF (5 mL) at 08C and the mixture was stirred at RT for
13 h. After concentration, the product was extracted with CHCl3
(200 mL), 10% citric acid, 5% NaHCO3, and brine. The crude product
was purified by SiO2 column chromatography (CHCl3/MeOH=20:1) to
afford the desired product (35; see the Supporting Information) as a
white solid (101 mg, 98%). m.p.: 190–1958C; 1H NMR (400 MHz,
[D6]DMSO): d=1.13–1.29 (m, 29H, 3-, 4-, and 5-Boc and
Protein expression and purification: The plasmids, GGTase I a subunit in
pAlter-Ex2 (pAlter-Ex2-GGTa) and b subunit in pET28a (pET28a-
GGTb) were kindly provided by Prof. Casey from Duke University. The
protein was expressed and purified as previously reported.[42,67] Briefly,
the plasmids were successively transformed into BL21 (DE3) E. coli cells
(Novagen), which grew until the optical density reached 0.6. GGTase I
expression was induced by the addition of isopropyl-b-d-thiogalactoside
(0.4m m) and ZnSO4 (0.5 mm), and the cells were harvested after 4h and
lysed. The soluble fraction of the cell lysate was bound to nickel agarose
resin (Qiagen) by incubation for 1 h at 48C in a buffer solution contain-
ing Tris-HCl (20 mm, pH 7.7), NaCl (300 mm), imidazole (5 mm), dithio-
threitol (DTT; 1 mm), and protease inhibitor mix. The proteins were
eluted with increasing concentrations of imidazole, and the appropriate
fractions were immediately dialyzed and concentrated. The purity of the
protein solution was checked by SDS-PAGE gels (>90%).
-CH2CH2CH2CO2CH3),
1.45–1.58
(m,
10H,
-CH2CH2CO2CH3,
-CONHCH2CH2 and 3-, 4-, 5-OCH2CH2), 1.77–1.90 (m, 6H, -OCH2CH2),
2.29 (t, J=7.3 Hz, 2H, -CH2CO2CH3), 2.66–2.78 (m, 6H, 3,4,5-CH2Ph),
3.18–3.23 (m, 2H, -CONHCH2), 3.56 (s, 3H, -CO2CH3), 3.83–3.99 (m,
9H, 3-, 4-, and 5-OCH2CH2CH), 6.55 (d, J=8.3 Hz 1H, 4-NHBoc), 5.30
(d, J=8.4Hz, 2H, 3 and 5-N HBoc), 7.08 (s, 2H, benzoyl) 7.16–7.25 (m,
15H, 3-, 4-, and 5-Ph) and 8.33 ppm (m, 1H, NH); LRMS (ESI) [M+
+Na] calcd for C65H82N4O12Na, 1061; found, 1061; elemental analysis
calcd (%) for C59H82N4O12: C 68.18, H 7.95, N 5.39; found: C 67.99, H
7.90, N 5.30.
General conditions for fluorescent enzyme assay: The inhibition activities
of the synthetic compounds against GGTase I were measured by a kinetic
assay by using the fluorogenic substrate, dansyl-Gly-Cys-Val-Ile-Leu
(DansGCVIL), which was prepared by solid-phase peptide synthesis (see
Figures S21 and S22 in the Supporting Information). The peptide buffer
solution (53 mm Tris-HCl, pH 7.5, 0.1 mm ethylenediaminetetraacetate
(EDTA), 0.020% n-dodecyl-b-d-maltoside, 5.0 mm DTT) was used for
preparation of DansGCVIL stock solution, and the assay buffer solution
(50 mm Tris-HCl, pH 7.50, 1.2 mm MgCl2, 1.2 mM ZnCl2, 0.023% n-dode-
Compound 35 (85 mg, 82 mmol) was hydrolyzed in a similar manner as
described above by using KOH to give the corresponding carboxylic acid
38 (see the Supporting Information) as a pale yellow amorphous solid
(87 mg, 100%). m.p.: 168–1728C; 1H NMR (400 MHz, [D6]DMSO): d=
1.14–1.30 (s, 29H, 3-, 4-, and 5-Boc and CH2CH2CH2CO2CH3), 1.46–1.55
Chem. Eur. J. 2008, 14, 1392 – 1401
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1399

J. Ohkanda et al.
cyl-b-d-maltoside, 5 mm DTT) was used for enzyme dilution and for run-
ning the kinetic assay. Commercially purchased GGPP ammonium salt
methanol solution from Sigma–Aldrich was diluted with 25 mm
NH4HCO3 to 500 mm and stored at ꢂ808C in aliquots and further diluted
to 110 mm before use (methanol content less than 5%). The peptide sub-
strate, DansGCVIL, was dissolved in the peptide buffer solution (approx-
imately 500 mm), and the concentration was determined from a standard
curve of A340 versus the concentration of dansylglycine in the same buffer
solution. Inhibitors were first dissolved in DMSO solution (1 or 30 mm)
and further diluted with peptide buffer solution to various concentrations
(1.1 mm to 2.2 mm), which were then used for the assay. The GGTase I so-
lution (125 mM) was freshly diluted with the assay buffer solution prior to
use (685 nm). The assays were performed at 308C by using a thermostat-
ed curvette holder. Solutions of DansGCVIL (10 mL), GGPP (10 mL),
and inhibitor (10 mL) were added to the assay buffer solution (180 mL) in
a 500-mL tube, which was incubated in a 308C water bath for 5 min. The
GGTase I solution (10 mL) was added to this solution; the mixture was
vortexed, quickly transferred to the cuvette, and the fluorescent intensity
change at 520 nm (ex: 340 nm) was monitored for 5 min. The final con-
centrations of each component were; [DansGCVIL]=1 mm, [GGPP]=
5 mm, [GGTase I]=31 nm, [Inhibitors]=0–100 mm (the content of DMSO
in the final solution was less than 2%). The experiments for each concen-
tration of inhibitor were repeated at least three times.
compound, 10 mg H-Ras or H-Ras-CVLL substrate, and 0.5 mCi/sample
of either [3H] FPP or [3H] GGPP. Samples were precipitated by using tri-
carboxylic acid (TCA) and then filtered onto glass-fiber filters; unbound
[3H] FPP or [3H] GGPP was washed through the filters. Samples were
counted in a scintillation counter and the activity was compared with ve-
hicle controls to obtain IC50 values.
Acknowledgements
Financial support from Hayashi Memorial Foundation and a Grant-in-
Aid from the Ministry of Education, Culture, Sports, Science and Tech-
nology is acknowledged. We are grateful to Professor Patrick Casey for
providing us the plasmids. J.O. sincerely thanks Professors Andrew Ham-
ilton and Indraneel Ghosh for helpful discussions.
[1] P. Bork, L. J. Jensen, C. V. Merg, A. K. Ramani, I. Lee, E. M. Mar-
cotte, Curr. Opin. Struct. Biol. 2004, 14, 292–299.
[2] B. P. Orner, J. T. Ernst, A. D. Hamilton, J. Am. Chem. Soc. 2001,
123, 5382–5383.
The raw data points were fit linearly and the resulting slopes were used
to calculate IC50 values, which were determined by plotting the data as
percent inhibition versus the log of the concentration of the inhibitors
used. The data were fit to the sigmoidal Equation (1) by using SigmaPlot
(version 10, Systat software, Inc.).
[3] S. Fletcher, A. D. Hamilton, Curr. Opin. Chem. Biol. 2005, 9, 632–
638.
[4] H. Yin, A. D. Hamilton, Angew. Chem. 2005, 117, 4200–4235;
Angew. Chem. Int. Ed. 2005, 44, 4130–4163.
[5] L. Pagliaro, J. Felding, K. Audouze, S. J. Nielsen, R. B. Terry, C.
Krog-Jensen, S. Bucher, Curr. Opin. Chem. Biol. 2004, 8, 442–449.
[6] M. R. Arkin, J. A. Wells, Nat. Rev. Drug Discovery 2004, 3, 301–
317.
[7] J. A. Kritzer, R. Zutshi, M. Cheah, F. A. Ran, R. Webman, T. M.
Wongjirad, A. Schepartz, ChemBioChem 2006, 7, 29–31.
[8] S. E. Rutledge, H. M. Volkman, A. Schepartz, J. Am. Chem. Soc.
2003, 125, 14336–14347.
ðbꢂaÞ
ð1Þ
y ¼ a þ
1 þ 10ðcꢂxÞd
In Equation (1), a and b are the minimal and maximal values of percent-
age of inhibition, x is the logarithm of the inhibitor concentration, c is
the IC50 value, and d is the Hill slope. The IC50 values were further con-
verted to the inhibition constant Ki values by using the Cheng–Prusoff
equation (Equation (2)).[68]
[9] S. Rajagopal, S. C. Meyer, A. Goldman, M. Zhou, I. Ghosh, J. Am.
Chem. Soc. 2006, 128, 14356–14363.
[10] T. J. Smith, C. I. Stains, S. C. Meyer, I. Ghosh, J. Am. Chem. Soc.
2006, 128, 14456–14457.
IC50
ꢀ
ꢁ
Ki ¼
[11] R. L. A. Dias, R. Fasan, K. Moehle, A. Renard, D. Obrecht, J. A.
Robinson, J. Am. Chem. Soc. 2006, 128, 2726–2732.
[12] Y. Hamuro, M. Calama, H. Park, A. D. Hamilton, Angew. Chem.
1997, 109, 2797–2800; Angew. Chem. Int. Ed. 1997, 36, 2680–2683.
[13] H. S. Park, Q. Lin, A. D. Hamilton, Proc. Natl. Acad. Sci. USA
2002, 99, 5105–5109.
[14] H. C. Zhou, L. Baldini, J. Hong, A. J. Wilson, A. D. Hamilton, J.
Am. Chem. Soc. 2006, 128, 2421–2425.
[15] K. Groves, A. J. Wilson, A. D. Hamilton, J. Am. Chem. Soc. 2004,
126, 12833–12842.
[16] R. K. Jain, A. D. Hamilton, Org. Lett. 2000, 2, 1721–1723.
[17] J. E. Gestwicki, G. R. Crabtree, I. A. Graef, Science 2004, 306, 865–
869.
[18] S. L. Hussey, S. S. Muddana, B. R. Peterson, J. Am. Chem. Soc. 2003,
125, 3692–3693.
ð2Þ
½Sꢃ
1 þ
Km
As shown in Equation (2), [S] is the substrate concentration and Km
((0.28ꢁ0.04) mm) is the Michaelis constant of the substrate for the
enzyme obtained from a Micahelis–Menten plot (see Figure S1 in the
Supporting Information).
The Lineweaver–Burk analysis was carried out by using different sub-
strate and inhibitor concentrations with a range of 0.1–1 mm and 0.1–
5.0 mm, respectively. In all the cases, the GGPP concentration was fixed
to 5 mm for the saturated condition. The Lineweaver–Burk plot of 1/v
versus 1/[S] was analyzed by competitive, noncompetitive, and uncompe-
titive models with SigmaPlot and the best fit was obtained by a competi-
tive model in which the initial velocity v is given by Equation (3).
[19] D. A. Erlanson, J. W. Lam, C. Wiesmann, T. N. Luong, R. L. Sim-
mons, W. L. Delano, I. C. Choong, M. T. Burdett, W. M. Flanagan,
D. Lee, E. M. Gordon, T. OꢁBrien, Nat. Biotechnol. 2003, 21, 308–
314.
Vmax
ꢀ
ꢁꢀ
ꢁ
n ¼
ð3Þ
Km
½Sꢃ
½Iꢃ
i
1 þ
1 þ K
[20] Y. Takaoka, H. Tsutsumi, N. Kasagi, E. Nakata, I. Hamachi, J. Am.
Chem. Soc. 2006, 128, 3273–3280.
[21] J. Dimaio, B. Gibbs, D. Munn, J. Lefebvre, F. Ni, Y. Konishi, J. Biol.
Chem. 1990, 265, 21698–21703.
[22] T. Steinmetzer, B. Y. Zhu, Y. Konishi, J. Med. Chem. 1999, 42, 3109–
3115.
[23] K. M. Jude, A. L. Banerjee, M. K. Haldar, S. Manokaran, B. Roy, S.
Mallik, D. K. Srivastava, D. W. Christianson, J. Am. Chem. Soc.
2006, 128, 3011–3018.
In Equation (3), v is arbitrary fluorescence units per second, Vmax is the
maximum velocity of the reaction, and [I] is the inhibitor concentration.
In vitro GGTase I and FTase activity assays with [3H]-labeled substrates:
In vitro inhibition assays of mammalian FTase and GGTase I were per-
formed by measuring the incorporation of [3H] FPP (GE Healthcare, Pis-
cataway, NJ) and [3H] GGPP (Perkin-Elmer Life Sciences, Boston, MA)
into wild-type H-Ras (FTase) and H-Ras-CVLL (GGTase I), respectively,
as previously described.[71] Briefly, approximately 20mg of 60000”g post-
microsomal supernatant from SF-9 cells expressing human GGTase I or
FTase was incubated in the presence of an increasing concentration of
[24] A. L. Banerjee, M. Swanson, B. C. Roy, X. Jia, M. K. Haldar, S.
Mallik, D. K. Srivastava, J. Am. Chem. Soc. 2004, 126, 10875–10883.
1400
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1392 – 1401

Protein-Surface Recognition
FULL PAPER
[25] K. Enander, G. T. Dolphin, L. Baltzer, J. Am. Chem. Soc. 2004, 126,
4464–4465.
[48] R. Khosravi-Far, P. A. Solski, G. J. Clark, M. S. Kinch, C. J. Der,
Mol. Cell. Biol. 1995, 15, 6443–6453.
[26] B. A. Grzybowski, A. V. Ishchenko, C. Y. Kim, G. Topalov, R. Chap-
man, D. W. Christanson, G. M. Whitesides, E. I. Shakhnovich, Proc.
Natl. Acad. Sci. USA 2002, 99, 1270–1273.
[27] P. A. Boriack, D. W. Christianson, J. Kingery-Wood, G. M. White-
sides, J. Med. Chem. 1995, 38, 2286–2291.
[49] E. A. Clark, T. R. Golub, E. S. Lander, R. O. Hynes, Nature 2000,
406, 532–535.
[50] P. L. Joyce, A. D. Cox, Cancer Res. 2003, 63, 7959–7967.
[51] R. G. Qiu, A. Abo, F. McCormick, M. Symons, Mol. Cell. Biol. 1997,
17, 3449–3458.
[28] N. Schaschke, D. Gabrijelcic-Geiger, A. Dominik, C. P. Sommerhoff,
ChemBioChem 2005, 6, 95–103.
[29] K. Shen, P. A. Cole, J. Am. Chem. Soc. 2003, 125, 16172–16173.
[30] A. Profit, A. T. R. Lee, D. S. Lawrence, J. Am. Chem. Soc. 1999,
121, 280–283.
[31] B. C. Roy, A. L. Banerjee, M. Swanson, X. G. Jia, M. K. Haldar, S.
Mallik, D. K. Srivastava, J. Am. Chem. Soc. 2004, 126, 13206–13207.
[32] M. C. Seabra, Y. Reiss, P. J. Casey, M. S. Brown, Cell 1991, 65, 429–
434.
[52] K. Lim, A. T. Baines, J. J. Fiordalisi, M. Shipitsin, L. A. Feig, A. D.
Cox, C. J. Der, C. M. Counter, Cancer Cell 2005, 7, 533–545.
[53] E. C. Lerner, Y. M. Qian, A. D. Hamilton, S. M. Sebti, J. Biol.
Chem. 1995, 270, 26770–26773.
[54] D. B. Whyte, P. Kirschmeier, T. N. Hockenberry, I. Nunez-Oliva, L.
James, J. J. Catino, W. R. Bishop, J. Pai, J. Biol. Chem. 1997, 272,
14459–14464.
[55] C. A. Rowell, J. J. Kowalczyk, M. D. Lewis, A. M. Garcia, J. Biol.
Chem. 1997, 272, 14093–14097.
[33] W. W. Epstein, D. Lever, L. M. Leining, E. Bruenger, H. C. Rilling,
Proc. Natl. Acad. Sci. USA 1991, 88, 9668–9670.
[56] K. L. Terry, P. J. Casey, L. S. Beese, Biochemistry 2006, 45, 9746–
9755.
[34] P. J. Casey, J. LipidRes. 1992, 33, 1731–1740.
[35] K. Zhu, A. D. Hamilton, S. M. Sebti, Curr. Opin. Invest. Drugs 2003,
4, 1428–1435.
[57] T. S. Reid, S. B. Long, L. S. Beese, Biochemistry 2004, 43, 9000–
9008.
[58] J. S. Taylor, T. S. Reid, K. L. Terry, P. J. Casey, L. S. Beese, EMBO J.
2003, 22, 5963–5974.
[59] G. L. James, J. L. Goldstein, M. S. Brown, J. Biol. Chem. 1995, 270,
6221–6226.
[60] F. L. Zhang, P. Kirschmeier, D. Carr, L. James, R. W. Bond, L.
Wang, R. Patton, W. Windsor, R. Syto, R. Zhang, W. R. Bishp, J.
Biol. Chem. 1997, 272, 10232–10239.
[61] S. B. Long, P. J. Casey, L. S. Beese, Structure 2000, 8, 209–222.
[62] I. Smalera, J. M. Williamson, W. Baginsky, B. Leiting, P. Mazur, Bio-
chim. Biophys. Acta 2000, 1480, 132–144.
[63] K. T. Lane, L. S. Beese, J. LipidRes. 2006, 47, 681–699.
[64] T. S. Reid, K. L. Terry, P. J. Casey, L. S. Beese, J. Mol. Biol. 2004,
343, 417–433.
[65] S. L. Moores, M. D. Schaber, S. D. Mosser, E. Rands, M. B. OꢁHara,
V. M. Garsky, M. S. Marshall, D. L. Pompliano, J. B. Gibbs, J. Biol.
Chem. 1991, 266, 14603–14610.
[66] W. G. Stirtan, C. D. Poulter, Biochemistry 1997, 36, 4552–4557.
[67] H. Fu, J. F. Moomaw, C. R. Moomaw, P. J. Casey, J. Biol. Chem.
1996, 271, 28541–28548.
[68] Y. Cheng, W. H. Prusoff, Biochem. Pharmacol. 1973, 22, 3099–3108.
[69] D. K. Srivastava, K. M. Jude, A. L. Banerjee, M. Haldar, S. Mano-
karan, J. Kooren, S. Mallik, D. W. Christianson, J. Am. Chem. Soc.
2007, 129, 5528–5537.
[70] P. Broto, G. Moreau, C. Vandycke, Eur. J. Med. Chem. 1984, 19, 71–
78.
[71] A. Vogt, Y. Qian, T. F. McGuire, A. D. Hamilton, S. M. Sebti, Onco-
gene 1996, 13, 1991–1999.
[72] S. B. Long, P. J. Hancock, A. M. Kral, H. W. Hellinga, L. S. Beese,
Proc. Natl. Acad. Sci. USA 2001, 98, 12948–12953.
[36] F. G. Njoroge, A. G. Taveras, J. Kelly, S. Remiszewski, A. K. Mal-
lams, R. Wolin, A. Afonso, A. B. Cooper, D. F. Rane, Y. T. Liu, J.
Wong, B. Vibulbhan, P. Pinto, J. Deskus, C. S. Alvarez, J. D. Rosario,
M. Connolly, J. Wang, J. Desai, R. R. Rossman, W. R. Bishop, R.
Patton, L. Wang, P. Kirschmeier, M. S. Bryant, A. A. Nomeir, C. C.
Lin, M. Liu A. T. McPhail, R. J. Doll, V. M. Girijavallabhan, A. K.
Ganguly, J. Med. Chem. 1998, 41, 4890–4902.
[37] L. Wang, G. T. Wang, X. Wang, Y. Tong, G. Sullivan, D. Park, N. M.
Leonard, Q. Li, J. Cohen, W. Gu, H. Zhang, J. L. Bauch, C. G.
Jakob, C. W. Hutchins, V. S. Stoll, K. Marsh, S. H. Rosenberg, H. L.
Sham, N. Lin, J. Med. Chem. 2004, 47, 612–626.
[38] A. D. Basso, P. Kirschmeier, W. R. Bishop, J. LipidRes. 2006, 47,
15–31.
[39] M. H. Gelb, L. Brunsveld, C. A. Hrycyna, S. Michaelis, F. Tamanoi,
W. C. V. Boorhis, H. Waldmann, Nat. Chem. Biol. 2006, 2, 518–528.
[40] M. R. Philips, A. D. Cox, J. Clin. Invest. 2007, 117, 1223–1225.
[41] J. Sun, J. Ohkanda, D. Coppola, H. Yin, M. Kothare, B. Busciglio,
A. D. Hamilton, S. M. Sebti, Cancer Res. 2003, 63, 8922–8929.
[42] Y. K. Peterson, P. Kelly, C. A. Weinbaum, P. J. Casey, J. Biol. Chem.
2006, 281, 12445–12450.
[43] H. Peng, D. Carrico, V. Thai, M. A. Blaskovich, C. Bucher, E. E. Pu-
sateri, S. M. Sebti, A. D. Hamilton, Org. Biomol. Chem. 2006, 4,
1768–1784.
[44] A. M. Sjogren, K. M. Andersson, M. Liu, B. A. Cutts, C. Karlsson,
A. M. Wahlstrom, M. Dalin, C. Weinbaum, P. J. Casey, A. Tarkow-
ski, B. Swolin, S. G. Young, M. O. Bergo, J. Clin. Invest. 2007, 117,
1294–1304.
[45] S. Castellano, H. D. G. Fiji, S. S. Kinderman, M. Watanabe, P. D.
Leon, F. Tamanoi, O. Kwon, J. Am. Chem. Soc. 2007, 129, 5843–
5845.
[46] B. S. Zuckerbraun, J. E. Barbato, A. Hamilton, S. Sebti, E. Tzeng, J.
Surg. Res. 2005, 124, 256–263.
[47] J. Ye, C. Wang, J. Rhea Sumpter, M. S. Brown, J. L. Goldstein, J. Mi-
chael Gale, Proc. Natl. Acad. Sci. USA 2003, 100, 15865–15870.
Received: October 15, 2007
Revised: November 28, 2007
Published online: January 17, 2008
Chem. Eur. J. 2008, 14, 1392 – 1401
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1401