Conformational restriction of flexible ligands guided by the transferred NOE experiment: Potent macrocyclic inhibitors of farnesyltransferase [25]

Full text of DOI:10.1021/ja003673q

J. Am. Chem. Soc. 2001, 123, 2107-2108
Conformational Restriction of Flexible Ligands
2107
Guided by the Transferred NOE Experiment: Potent
Macrocyclic Inhibitors of Farnesyltransferase
Christopher J. Dinsmore,*,† Michael J. Bogusky,†
J. Christopher Culberson,§ Jeffrey M. Bergman,†
Carl F. Homnick,† C. Blair Zartman,† Scott D. Mosser,‡
Michael D. Schaber,‡ Ronald G. Robinson,‡
Kenneth S. Koblan,‡ Hans E. Huber,‡ Samuel L. Graham,†
George D. Hartman,† Joel R. Huff,† and Theresa M. Williams†
Departments of Medicinal Chemistry
Molecular Systems, and Cancer Research
Merck Research Laboratories
West Point, PennsylVania 19486
ReceiVed October 16, 2000
Figure 1. Superposition of two representative lowest-energy FTase-bound
conformations of 2 consistent with trNOE-derived restraints.
The optimization of enzyme inhibitor potency and specificity
is an important goal of drug design since both properties contribute
to clinical efficacy and safety. Restricting an inhibitor’s confor-
mation to one recognized by the enzyme increases potency by
lowering the entropic barrier to complex formation, and could
potentially enhance specificity by limiting its interactions with
other macromolecules.1 In lieu of detailed structural characteriza-
tion of enzyme-inhibitor interactions, the transferred nuclear
Overhauser effect (trNOE) NMR method has proven valuable in
defining conformations of ligands weakly associated with mac-
romolecules.2 However, despite the considerable implications of
the trNOE technique for drug design, there are few instances of
the method playing an influential role in inhibitor optimization,3
and none is directed at designing specific conformational con-
straints. This report describes the design of a highly potent
macrocyclic enzyme inhibitor based on the trNOE structure of a
conformationally flexible analogue.
structure-activity relationships of the clinical candidate 1,6 we
found that the related FTI 2, with diminished inhibitory activity
(FTase IC50 475 nM vs 2 nM), was an appropriate ligand for the
trNOE experiment.
In the absence of added enzyme, NMR spectroscopic evaluation
of 2 reveals no defined solution conformation. NMR-derived
intramolecular distance constraint data was generated in the
presence of the putative FTase‚FPP complex. Ligand-competition
experiments with a potent peptidomimetic FTI served to disqualify
non-active-site bound contributions. The calculated lowest-energy
structures depict folded conformations with the cyanophenyl group
flanking the piperazinone ring (Figure 1).7a Stabilization of this
orientation by covalent linkage of the cyanophenyl and piperazi-
none N-aryl substituent in a macrocycle appeared to be an
attractive approach to optimize the properties of 1.
The synthesis of a macrocyclic version of 1 is described in
Scheme 1. The piperazinone 8, prepared by a Mitsunobu
cyclodehydration reaction,8 was reductively coupled with aldehyde
5 to give 9. Compound 9 was subjected to a tandem base-
promoted arylmethanesulfonate deprotection and SNAr cyclization9
to give the cyclophane 10 in good yield. Interestingly, 10 exhibits
planar chirality, and its enantiomeric conformers are readily
resolved by chiral HPLC, due to a sufficient activation energy
for their interconversion.10
The calculated lowest-energy structure of (+)-107c (Figure 2,
gray) bears close resemblance to available FTase-bound confor-
mations of 2 (Figure 1), especially with regard to the relative
positions of the piperazinone, imidazole, and cyanophenyl rings.11
Variable temperature 1H NMR studies of (+)-10 revealed
Farnesyltransferase (FTase) is an important posttranslational
processing enzyme that prenylates proteins using farnesylpyro-
phosphate (FPP) and enables the participation by some in signal
transduction during cell proliferation.4 Inhibitors of this enzyme
(FTIs) are promising antitumor agents, and several are currently
being evaluated in human clinical trials.5 In our investigations of
(4) (a) Kato, K.; Cox, A. D.; Hisaka, M. M.; Graham, S. M.; Buss, J. E.
Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6403-6407. (b) Rowinsky, E. K.;
Windle, J. L.; Von Hoff, D. D. J. Clin. Oncol. 1999, 17, 3631-3652.
(5) (a) Oliff, A. Biochim. Biophys. Acta 1999, 1423, C19-C30. (b) End,
D. W. InVest. New Drugs 1999, 17, 241-258. (c) Gibbs, J. B. J. Clin. InVest.
2000, 105, 9-13.
† Department of Medicinal Chemistry.
§ Department of Molecular Systems.
‡ Department of Cancer Research.
(1) (a) Freidinger, R. M.; Veber, D. F.; Perlow, D. S.; Brooks, J. R.;
Saperstein, R. Science 1980, 210, 656-658. (b) Liskamp, R. M. J. Recl. TraV.
Chim. Pays-Bas 1994, 113, 1-19.
(6) Williams, T. M.; et al., Merck & Co., Inc. unpublished data.
(7) (a) Distance restraints from trNOE data for 2 were used to generate
conformations using JG (Kearsley, S. Merck & Co., Inc., unpublished). These
were minimized within Macromodel (ref 7b) using the MMFF force field
with a 4r distance-dependent dielectric, and with trNOE distance constraints
applied. (b) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990,
11, 440-467. (c) Same as 7a, but without distance restraints. (d) Same as 7a,
(2) Recent reviews: (a) Roberts, G. C. K. Curr. Opin Biotechnol. 1999,
10, 42-47. (b) Moore, J. M. Biopolymers 1999, 51, 221-243. (c) Stockman,
B. J. Prog. Nucl. Magn. Reson. Spectrosc. 1998, 33, 109-151.
(3) (a) Gonnella N. C.; Bohacek, R.; Zhang, X.; Kolossvary, I.; Paris, C.
G.; Melton, R.; Winter, C.; Hu, S.-I.; Ganu, V. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 462-466. (b) Williams, T. M.; Bergman, J. M.; Brashear, K.; Breslin,
M. J.; Dinsmore, C. J.; Hutchinson, J. H.; MacTough, S. C.; Stump, C. S.;
Wei, D. D.; Zartman, C. B.; Bogusky, M. J.; Culberson, J. C.; Buser-Doepner,
C.; Davide, J.; Greenberg, I. B.; Hamilton, K. A.; Koblan, K. S.; Kohl, N. E.;
Liu, D.; Lobell, R. B.; Mosser, S. D.; O’Neill, T. J.; Rands, E.; Schaber, M.
D.; Wilson, F.; Senderak, E.; 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, 3779-3784.
1
but with solution H NMR NOE distance restraint data for (+)-10.
(8) Weissman, S. A.; Lewis, S.; Askin, D.; Volante, R. P.; Reider, P. J.
Tetrahedron Lett. 1998, 39, 7459-7462.
(9) Dinsmore, C. J.; Zartman, C. B. Tetrahedron Lett. 1999, 40, 3989-
3990.
(10) Kinetic data for the racemization of (-)-10 in DMSO: Ea ) 28.4
kcal/mol; t1/2 (100 °C) ) 60 min; t1/2 (120 °C) ) 8.5 min; ∆Hq ) 26.9 kcal/
mol; ∆Sq ) 12.4 eu.
10.1021/ja003673q CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/08/2001

2108 J. Am. Chem. Soc., Vol. 123, No. 9, 2001
Communications to the Editor
Scheme 1a
Table 1. FTase Inhibition Data and Effect of Constraining Bond
Rotations
effect of constraint
on binding energy
(kcal/mol)c
number
IC50
Ki
per
bond
available
compd (nM)a
(nM)b
∆∆G°
bond rotationsd
(+)-10
0.1 0.20 ( 0.06
-
-
9-6 ) 3e
7
7
1
2
0.9 ( 0.3
0.9 (1.8) 0.2 (0.4)
(6.6) (1.6)
11 5490
nd
a Concentration of compound required to reduce the FTase-catalyzed
incorporation of [3H]FPP into recombinant Ras-CVIM by 50%. b KI
vs K-4B-Ras. c Overall free energy difference from the equation ∆∆G°
) -RT[ln(Ki,cyclic/Ki,acyclic)] (ref 16). Values in parentheses are from
estimated Ki’s vs Ras CVIM (KM 18 nM), which are derived from the
IC50 values using the Cheng-Prusoff equation (ref 18). d Independently
rotatable dihedral angles, excluding bonds in aromatic rings, amides,
and C-CtN. The three N-CH2-CH2-N bonds are also excluded,
since these variables are dependent on the N-CH2-CO-N dihedrals.
e Six degrees of freedom are lost on macrocyclization (ref 17).
a Reagents and conditions: (a) Zn(CN)2, Pd(PPh3)4, DMF, 80 °C, 6
h, 96%. (b) NBS, AIBN, CCl4, reflux, 3 h, 43%. (c) 1-Trityl-4-(acetoxy-
methyl)imidazole, EtOAc, reflux, 20 h; MeOH, reflux, 30 min, 93%. (d)
LiOH, 2:1 THF-H2O, 0 °C, 2 h, 83%. (e) SO3‚Py, Et3N, DMSO, 45
min, 86%. (f) ClCH2COCl, aq Na2CO3, EtOAc, 0 °C, 2 h. (g) MsCl,
Et3N, DMF, 0 °C, 16 h, 79% for 2 steps. (h) H2N(CH2)2OH, i-PrOAc, rt,
18 h. (i) DBAD, n-Bu3P, EtOAc, 40 °C, 10 h; HCl, EtOH, rt to 0 °C,
78%, 2 steps. (j) Na(AcO)3BH, 5, 4 Å sieves, DCE, 0 °C to room tem-
perature, 14 h, 91%. (k) Cs2CO3, DMSO, 0.05 M, 80 °C, 3 h, 86%.
The FTase inhibitory activity for macrocycle (+)-10 was
assessed relative to open-chain analogues (Table 1).12 In principle,
compounds 1 and 1113 may adopt conformations similar to (+)-
10 without suffering significant intramolecular destabilizing
interactions.14 Because of close structural similarities among all
three compounds, differences in their enthalpic contributions to
binding should be small, assuming similar binding modes.15 The
enhanced relative activity for (+)-10, therefore, suggests that the
entropic disadvantage of complex formation is significantly
reduced by incorporation of the macrocyclic constraint. However,
the extent of this entropic penalty (estimated as ∆∆G°)16 is
dependent on the position of ring cleavage, even though the
number of available bond rotations in 1 and 11 is the same.17
The penalty for binding of 1 on a per-bond basis16 relative to
(+)-10 may be an underestimate, since it is possible that the
chlorine atom of 1 contributes to binding by occupying a position
which is unavailable to the corresponding naphthalene CH group
in (+)-10. Furthermore, the greater penalty for 11 may be an
overestimate because of a possible increase in the enthalpy of
binding due to an inability of the imidazole and cyanophenyl rings
to occupy the optimal relative positions14 or due to diminished
binding of an N-H versus N-alkyl imidazole. Alternatively, 11
may exhibit a different binding mode relative to (+)-10.15
The macrocyclic conformational constraint imparts a striking
change in the mechanism of inhibition of the FTase-related
enzyme geranylgeranyltransferase-I (GGTase-I). Despite very
similar structural elements, 1 is competitive with the geranylgera-
nylpyrophosphate substrate (IC50 98 nM),6 while (+)-10 is
competitive with the K-Ras-derived peptide substrate used in the
assay (IC50 300 nM).12 This difference in mechanism suggests
disparate binding modes in the GGTase-I active site for acyclic
1 and macrocyclic (+)-10.
Figure 2. Superposition of the calculated lowest-energy conformation
of (+)-10 (gray) and the average solution structure as determined by 1H
NMR in d4-methanol (yellow).
differential line broadening below 0 °C, indicating conformational
exchange, and NOE data obtained at room temperature are
consistent with a mixture of available species. As such, the
calculated solution structure represents an average of available
conformations (Figure 2, yellow),7d largely derived from mobility
of the piperazinone ring.
(11) The assignment of absolute configuration for (+)-10 is based on X-ray
crystallographic studies of the (+)-10‚FPP‚FTase complex (Beese, L. S.;
Taylor, J. Duke University Medical Center, unpublished data).
(12) For assay descriptions, see refs 3b and 6. IC50 (-)-10 ) 0.4 nM.
(13) Compound 11 was prepared from the naphthol corresponding to 8 by
arylation (Cs2CO3, 2-fluorobenzonitrile), deprotection (HCl), and reductive
alkylation with imidazole-4-carboxaldehyde (Na(AcO)3BH).
(14) The N-(1-naphthyl) analogue of 1 (IC50 2 nM), although equipotent
to 1, cannot bind with the same exact conformation as (+)-10 due to
overlapping hydrogen atoms. The same may be true for 11 vs (+)-10.
(15) Lineweaver-Burke analyses showed that (+)-10 and 1 are competitive
with K-Ras protein, and not competitive with FPP. The mechanism of binding
for 11 is unknown and is assumed to be the same as for 1 and (+)-10, based
on similar structure.
In summary, the trNOE structure of a conformationally flexible
member of a well-developed inhibitor structural series was used
to design a macrocyclic analogue which exhibited both enhanced
FTase and altered GGTase-I inhibition properties. This suggests
that trNOE technology can contribute significantly to inhibitor
optimization well into late stages of design.
Supporting Information Available: trNOE data for 2, overlay of 2
and (+)-10, procedures and characterization data for compounds 3-10,
and data for the interconversion of (+)-10 and (-)-10 (PDF). This material
is available free of charge via the Internet at http:/pubs.acs.org.
(16) For a good discussion of this analysis, see: Khan, A. R.; Parrish, J.
C.; Fraser, M. E.; Smith, W. W.; Bartlett, P. A.; James, M. N. G. Biochemistry
1998, 37, 16839-16845 and references therein.
(17) Go, N.; Scheraga, H. A. Macromolecules 1970, 3, 178-187.
(18) Cheng, Y.-C.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
JA003673Q