Development of tripeptidyl farnesyltransferase inhibitors.

Article abstract of DOI:10.1016/S0960-894X(02)00227-5

The first example of tripeptide inhibitors of farnesyltransferase with sub-micromolar inhibition activity was developed based on the fact that CVFM is not a substrate for farnesyltransferase.

Full text of DOI:10.1016/S0960-894X(02)00227-5

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1599–1602
Development of Tripeptidyl Farnesyltransferase Inhibitors
Hee-Yoon Lee,a,* Jeong-Hun Sohna and Byoung-Mog Kwonb
aCenter for Molecular Design and Synthesis, Department of Chemistry and School of Molecular Science(BK21),
Korea Advanced Institute of Science and Technology, Daejon 305-701, Republic of Korea
bKorea Research Institute of Bioscience and Biotechnology, Daejon 305-333, Republic of Korea
Received 28 January 2002; accepted 6 April 2002
Abstract—The first example of tripeptide inhibitors of farnesyltransferase with sub-micromolar inhibition activity was developed
based on the fact that CVFM is not a substrate for farnesyltransferase. # 2002 Elsevier Science Ltd. All rights reserved.
Significant progress has been made in the development
of non-cytotoxic anti-tumor agents during the last three
decades.1 Among them, inhibitors of farnesyltransferase
have been developed extensively for cancer chemo-
therapy since blockage or attenuation of the Ras farne-
sylation pathway was shown to suppress Ras related
cancer cell growth.2 During the farnesylation step,
CaaX motif present at the C-terminal of Ras where C is
cysteine, a is an aliphatic amino acid and X is preferably
serine or methionine, is recognized along with farnesyl
pyrophosphate (FPP) by farnesyltransferase and then
the cysteine of CaaX is farnesylated by the enzyme. It
was also found out that the CaaX tetrapeptide sequence
renders enough specificity for recognition by farnesyl-
transferase and even is farnesylated at the cysteine site.
While various substrate-mimetic inhibitors of farnesyl-
transferase were developed based either on tetrapeptide,
FPP, or even on the farnesylated peptides,3 only the
peptide-based inhibitors were successfully developed as
potent inhibitors of farnesylation.
might be more space for hydrophobic interaction than a
space for the binding of phenyl group of a2 and room
for extending the hydrophobic side chain of a2 into the
FPP binding site.
We became interested in exploring a possibility of fur-
ther shortening the tetrapeptide CaaX to the tripeptide
Caa in an anticipation that tripeptide analogues that
consist of Ca1a2 with a2-side chain having a large
hydrophobic group without both terminal X residue
and the carboxylate groupwould find enough hydro-
phobic binding energy from both the a2 binding site and
FPP binding site (Fig. 1). We hoped that the large
hydrophobic group would occupy the binding site
where the phenyl group blocked the entry of FPP and
also picked up the extra binding energy by extending
itself toward the FPP binding site.
To establish the structure–activity relationship(SAR) of
tripeptides for the binding pocket, we prepared Ca1a2
with varying size of the side chain of a2. To simplify the
SAR, a1 residue was fixed as valine and all the variation
was placed on the a2 residue. Substituted tyrosine, and
meta-tyrosine with various hydrophobic groups as esters
attached to their hydroxyl groups were prepared to
explore the extent of the hydrophobic interaction with
During the search for farnesyltransferase inhibitors from
nonpeptidyl or natural sources,4 we became interested in
the observation that some of the tetrapeptidyl inhibitors
of farnesyltransferase such as CVFM, CVWM and
CVYM were not substrates for farnesyltransferase. The
overall structures5 of ternary complexes using a FPP
analogue and peptides containing CVIM from X-ray
crystallographic study strongly suggested that the
phenyl rings of the peptidyl inhibitors blocked entry of
FPP to the binding site and prevented farnesylation of
these peptide-inhibitors. Thus it appears that there
*Corresponding author. Tel.:+82-42-869-2835; fax:+82-42-869-8370;
e-mail: leehy@kaist.ac.kr
Figure 1. Designed tripeptide-farnesyltransferase inhibitors.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00227-5

1600
H.-Y. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1599–1602
the binding pocket and to explore a possibility of
extending the interaction into the FPP binding site.6
Serine analogues were also prepared to introduce flex-
ibility to the side chain that could allow the side chain to
find the FPP binding site without much conformational
restriction.
removal of benzyl group7,9 after removing the chiral
auxiliary. The chiral auxiliary was replaced with methyl
ester using MeMgBr in methanol solvent10 to produce
m-tyrosine methyl ester. Derivatives of Cys-Val-Ser-
OMe, were also prepared in the same manner as in the
preparation of the Cys-Val-Tyr-OMe series.
Derivatives of Cys-Val-Tyr-OMe were prepared by
EDC coupling of (l)-tyrosine methyl ester hydrochlo-
ride with N-Cbz-valine followed by deprotection of Cbz
groupand subsequent EDC coupling with N-Boc-S-Tr-
(l)-cysteine without protecting the phenolic OH group
of tyrosine (Scheme 1). The hydroxyl groupof tyrosine
was esterified with acid chloride or with carboxylic acid
using EDC and DMAP to produce 4. Cysteine of 4 was
deprotected using trifluoroacetic acid and triethylsilane
to yield Y.
The inhibitory activity of the compounds for farnesyl-
transferase was evaluated as in vitro IC50 values. In
vitro IC50 values were determined against partially pur-
ified bovine farnesyltransferase using SPA assay (scin-
tillation proximity assay, Amersham, Arlinton
heights).11 3H-FPP and a biotin-linked C-terminal
undecapeptide of K-Ras4B (KKKSKTKCVIM) were
used as the substrates. The radioactivity captured by the
SPA beads were counted by a Packard Topcount, and
data were stored and analyzed in an Oracle-based data-
base. CVFM, one of the most potent tetrapeptide inhi-
bitors and its methyl ester was used as the references
with IC50 values of 0.06 and 5.8 mM respectively in our
farnesyltransferase inhibition studies.
Cys-Val-meta-Tyr-OMe analogues were synthesized
from meta-(l)-tyrosine methyl ester in the same manner
as the syntheses of Y series starting from meta-tyrosine
methyl ester. meta-Tyrosine methyl ester was synthe-
sized from m-hydroxybenzaldehyde according to the
literature preparation (Scheme 2).7 After protection of
the phenolic hydroxyl group of m-hydroxybenzaldehyde
as the benzyl ether, Wittig reaction followed by selective
hydrogenation of the double bond with platinum on
charcoal and subsequent ester hydrolysis produced 5.
After attaching the benzyl oxazolidinone as the chiral
auxiliary via the pivaloyl mixed anhydride method,8
stereoselective introduction of a-amino groupwas
accomplished via a-azidation followed by catalytic
reduction of the azide that was accompanied by the
Since all the prepared compounds did not possess the
crucial terminal carboxyl groupfor the binding activity,
all the synthesized compounds were screened first at the
concentration of 1 mg/mL for the inhibitory activity
(Table 1). While all the compounds showed appreciable
inhibitory activity, CVS-OMe that lacks the extended
hydrophobic group at a2 did not show any inhibitory
activity. This result demonstrated the importance of the
hydrophobic binding group at a2 and it also showed
that as the hydrophobic group became larger, inhibitory
activity became higher. This result indicates that the
hydrophobic group at the a2 plays an important role in
binding to the Ca1a2X box of farnesyltransferase but the
hydrophobic binding pocket of farnesyltransferase has
the limited space in it as the hydrophobic group
attached to a2 became large the binding activity did not
improve after a certain point. To evaluate the SAR
more accurately, IC50 values of the compounds were
Table 1.
Compda
R
%
Inhibitionb
Compda
R
%
Inhibitionb
Y1
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Y9
Y10
Me
6.3
26.7
79.6
74
Ym1
Ym2
Ym3
Ym4
Ym5
S1
i-Pr
30
Scheme 1. (a) N-Cbz-valine, EDC, HOBT, DIPEA, DMF, 16h, 83%;
(b) H2, 10% Pd/C, MeOH, 30 min, 94%; (c) N-Boc-S-Tr-cysteine, EDC,
HOBT, THF, 80%; (d) RCO2H, EDC, DMAP, CH2Cl2 or RCOCl,
DIPEA, CH2Cl2, 75–88%; (e) Et3SiH, TFA, CH2Cl2, 58–86%.
i-Pr
61
Ph
69.3
17
i-Bu
Pentyl
Bn
42
Me
Ph
12
32
69.3
57
Heptyl
17
S2
p-Ph-Ome
p-Tolyl
p-Ph-Cl
10
S3
67
Ph
23.4
12.6
0
S4
75
S5
38.6
16.3
CVS-OMe
S6
Scheme 2. (a) NaH, BnBr, TBAI, THF, 85%; (b) Ph3PCHCO2Me
PhH, reflux, 89%; (c) H2, 10% Pt/C, AcOEt, 99%; (d) LiOH, THF/
H2O, 90%; (e) PivCl, DIPEA, THF, À78 ꢀC, N-(Li)-(R)-phenylox-
azolidinone, 87%; (f) KHMDS, THF, trisyl azide, À78 ꢀC, AcOH, rt,
51%; (g) MeMgBr, MeOH, 0 ꢀC, 81%; (h) H2, Pd/C, AcOEt, 86%.
aConcentration of sample; 1 mg/mL.
b
%
{control-blank1(no enzyme, no sample)}].
Of inhibition=100[1À{sample-blank2(sample, no enzyme)}/

H.-Y. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1599–1602
1601
measured and the result is summarized in the Table 2.
The fact that CVFM methyl ester (CVFM-OMe) is 100-
fold less potent than CVFM (IC50=60 nM) clearly
indicated that the presence of a free carboxylate at the
C-terminus is important for a strong inhibition of the
enzyme. The IC50 value of Y1 indicated that deletion
of the 4th amino acid decreased the activity by another
10-fold from CVFM-OMe.
binding site we were able to reduce the size of the pep-
tide inhibitors to tripeptide from tetrapeptide. Though
there have been reports of tripeptide analogues of far-
nesyltransferase inhibitors that either mimic tetrapep-
tides or tripeptidyl-FPP, this is the first example of the
tripeptide based analogues of farnesyltransferase inhi-
bitors with sub-micromolar inhibitory activity.12 While
our investigation has not been able to fully extend the
hydrophobic binding interaction to the FPP binding
site, the current result strongly suggests that one could
pick up more hydrophobic binding energy from CaaX
based nonpeptidyl inhibitors13 since the hydrophobic
groups of those inhibitors appear to reach only to the
part of the hydrophobic group binding pocket of a2.
Currently, we are pursuing to find effective bisubstrate
inhibitors of farnesyltransferase, since combination of
Y3, Ym2 and Ym3 into another hydrophobic group
could also lead us to design better inhibitors as our SAR
result in connection to the X-ray ternary structures of
farnesyltransferase with inhibitors suggests that the
binding site of side chain of a2 residue should exist in
close proximity to the FPP binding site. We should be
able to extend this hydrophobic binding interaction into
the FPP binding site with the properly located extension
of the side chain.
Substituting the methyl groupof Y1 with a larger group
improved the binding activity. However, when the sub-
stituent became longer than the iso-butyl group( Y5–
Y10), no more binding energy was gained. This result
indicates that the space in the hydrophobic binding
pocket for a2 might not be as large as we hoped for to
pick up enough binding energy to compensate for the
lost binding energy from the lack of the carboxyl group.
Interestingly, compounds with branched side chains
showed markedly improved activity, especially when the
branching is at the b-position of the ester group (Y2–
Y4). This result is quite interesting since coincidentally
the iso-butyl side chain of arteminolide4a played an
important role in inhibiting farnesyltransferase. When
the substitution was moved from para-position to meta-
position of the phenyl ring, the extended hydrophobic
groupdid not seem to fit into the binding pocket well as
only Ym3 showed a significantly improved binding
activity. When serine analogues that were prepared with
an anticipation that would be free from the conforma-
tional restriction imposed by the phenyl rings of tyr-
osines were tested, improved binding activity was
observed with substituted benzoate esters attached to
the serine (S1–S4). When R groupof S was smaller or
bigger than the phenyl group (S5, S6), only a small
binding energy gain was observed. All these results
strongly suggest that there is a binding site for an
extended hydrophobic group of a2 and supplementing
this binding energy to the tripeptide inhibitors is large
enough to compensate the loss of the 4th amino acid
unit from CVFM-OMe and even offset a part of the
binding energy lost from the lack of the terminal car-
boxylate group.
References and Notes
1. (a) Goldenberg, G. J.; Moore, M. J.; Wright, J. E.;
Ludlum, D. B.; Kelland, L. R.; Sweatman, T. W.; Israel, M.;
Giovanella, B. C.; Pommier, Y.; Bissery, M.-C.; Lavelle, F.;
Zunino, F.; Capranico, G.; Romerdahl, C. A.; Brana, M. F.;
Schor, N. F.; Stetler-Stevenson, W. G.; Hendrzak, J. A.;
Brunda, M. J.; Ingber, D. E.; Crooke, S. T.; Sausville, E. A.;
Longo, D. L.; Blattler, W. A.; Chari, R. V. J.; Lambert, J. M.;
Sebolt-Leopold, J. S.; Schrier, P. I.; Osanto, S. In Cancer
Therapeutics: Experimental and Clinical Agents; Teicher, B. A.,
Ed.; Humana: Totowa, 1997; Vol. 1, p3. (b) Traxler, P.; Bold,
G.; Buchdunger, E.; Caravatti, G.; Furet, P.; Manley, P.;
O’Reilly, T.; Wood, J.; Zimmermann, J. Med. Res. Rev. 2001,
21, 499.
2. Gibbs, R. A.; Zahn, T. J.; Sebolt-Leopold, J. S. Curr. Med.
Chem. 2001, 8, 1437.
3. (a) Patel, D. V.; Young, M. G.; Robinson, S. P.; Hunihan,
L.; Dean, B. J.; Gordon, E. M. J. Med. Chem. 1996, 39, 4197.
(b) Schlitzer, M.; Bohm, M.; Sattler, I.; Dahse, H.-M. Bioorg.
Med. Chem. 2000, 8, 1991 and references therein.
In summary, we were able to identify a hydrophobic
binding pocket associated with the binding site of a2
residue and FPP binding site, and by utilizing this
4. (a) Lee, S.-H.; Kim, M.-J.; Bok, S. H.; Lee, H.; Kwon, B.-
M.; Shin, J.; Seo, Y. J. Org. Chem. 1998, 63, 7111. (b) Lee, H.-
Y.; Sohn, J.-H.; Kim, H. Y. Tetrahedron Lett. 2001, 42, 1695.
5. (a) Long, S. B.; Casey, P. J.; Beese, L. S. Structure 2000, 8,
209. (b) Strickland, C. L.; Windsor, W. T.; Syto, R.; Wang, L.;
Bond, R.; Wu, Z.; Schwartz, J.; Le, H. V.; Beese, L. S.; Weber,
P. C. Biochemistry 1998, 37, 16601.
Table 2.
Compd
R
IC50 (mM)
Compd
R
IC50 (mM)
Y1
Y2
Me
48
12
Ym1
Ym2
i-Pr
10
i-Pr
8.4
6. Attachment of hydrophobic groups as allylic or benzylic
ether linkage that could provide more flexibility than ester
linkage turned out to be unstable under the screening condi-
tion and produced unreliable inhibitory activities.
7. Banteli, R.; Brun, I.; Hall, P.; Metternich, R. Tetrahedron
Lett. 1999, 40, 2109.
8. Evans, D. A.; Gage, J. R.; Leighton, J. L. J. Am. Chem.
Soc. 1992, 114, 9434.
9. Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J.
Am. Chem. Soc. 1990, 112, 4011.
Y3
Y4
Y5
Y7
Y8
Y9
Y10
0.5
0.82
7.5
22
Ym3
Ym4
S1
Ph
1.1
17
i-Bu
Pentyl
Heptyl
Ph
0.9
S2
p-Ph-OMe
p-Tolyl
1.8
8.1
20
S3
0.92
0.82
Ph
S4
p-Ph-Cl
43
CVFM-OMe
CVFM
5.8
0.06
10. Evans, D. A.; Britton, T. C.; Dorow, R. L.; Dellaria, J. F.,
Jr. Tetrahedron 1988, 44, 5525.

1602
H.-Y. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1599–1602
11. Reiss, Y.; Goldstein, J. L.; Seabra, M. C.; Casey, P. J.;
Brown, M. S. Cell 1990, 62, 81.
3011. (b) Williams, T. M.; Bergman, J. M.; Brashear, K.; Bre-
slin, M. J.; Dinsmore, C. J.; Hutchinson, J. H.; MacTough,
S. C.; Stump, C. A.; Wei, D. D.; Zartman, C. B.; Bogusky,
M. J.; Culberson, C.; Buser-Doepner, C.; Hamilton, K. A.;
Koblan, K. S.; Kohl, N.; Liu, D.; Lobell, R. B.; Mosser, S. D.;
O’Neill, T. J.; Rands, E.; Schaber, M. D.; Wilson, F.; Sen-
derak, M. D.; Motzel, S. L.; Gibbs, J. B.; Graham, S. L.;
Heimbrook, D. C.; Hartman, G. D.; Oliff, A. I.; Huff, J. R. J.
Med. Chem. 1999, 42, 3779.
12. (a) Patel, D. V.; Patel, M. M.; Robinson, S. S.; Gordon,
E. M. Bioorg. Med. Chem. Lett. 1994, 4, 1883. (b) Kothare,
M. A.; Ohkanda, J.; Lockman, J. W.; Qian, Y.; Blaskovich,
M. A.; Sebti, S. M.; Hamilton, A. D. Tetrahedron 2000, 56,
9833.
13. (a) Qian, Y.; Marugan, J. J.; Fossum, R. D.; Vogt, A.;
Sebti, S. M.; Hamilton, A. D. Bioorg. Med. Chem. 1999, 7,