Synthesis and evaluation of chiral bicyclic proline FKBP12 ligands

Article abstract of DOI:10.1016/S0960-894X(03)00758-3

As part of our ongoing program to explore novel structural classes of FKBP12 ligands, we herein wish to report a new class of FKBP12 ligands containing chiral bicyclic proline analogues. Details of the synthetic routes, together with preliminary biologica

Full text of DOI:10.1016/S0960-894X(03)00758-3

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3867–3870
Synthesis and Evaluation of Chiral Bicyclic Proline
FKBP12 Ligands
David C. Limburg, Bert E. Thomas, IV, Jia-He Li, Mike Fuller, Dawn Spicer, Yi Chen,
Hongzhi Guo, Joseph P. Steiner, Gregory S. Hamilton and Yong-Qian Wu*
Guilford Pharmaceuticals Inc., Research Department, 6611 Tributary St., Baltimore, MD 21224, USA
Received 18 December 2002; accepted 12 June 2003
Abstract—As part of our ongoing program to explore novel structural classes of FKBP12 ligands, we herein wish to report a new
class of FKBP12 ligands containing chiral bicyclic proline analogues. Details of the synthetic routes, together with preliminary
biological activity, will be presented.
# 2003 Elsevier Ltd. All rights reserved.
The discovery that small molecule ligands for the pepti-
dyl-prolyl isomerase (PPIase) FKBP12 possess powerful
neuroprotective and neuroregenerative properties in
vitro and in vivo suggests that they may have thera-
peutic utility in treating neurodegenerative diseases such
as Parkinson’s Disease.1,2 The neurotrophic effects of
these compounds are independent of the immunosup-
pressive pathways by which drugs such as FK506 and
rapamycin operate.3,4 Immunosuppressant FK506 pos-
sesses two distinct binding domains: an immunophilin-
binding domain that binds to FKBP12 and an ‘effector
domain’ that mediates the interaction of the drug–
immunophilin complex with the secondary protein tar-
get (Fig. 1). Recently, we reported that small
molecules that mimic only the FKBP-binding portion of
FK506, such as GPI 1046 (Fig. 1), lack immunosup-
pressant activity but are extraordinarily potent neuro-
trophic agents in vitro and promote neuroregeneration
in vivo.5
Previous work by ourselves and other groups exploring
the SAR of small molecules that mimic only the FKBP-
binding domain portion of FK506 have focused largely
on esters, thioesters, amides and ketones of proline or
pipecolic acid.6À8 As indicated by the bound conforma-
tion of FK506, additional steric volume may be toler-
ated above the plane of the pipecolic or prolyl ring. The
observation that the adamantane ring system fits well
Figure 1. FK506 and small molecule mimetics of the FKBP12 binding domain.
*Corresponding author. Tel.:+1-410-631-6823; fax:+1-410-631-6848; e-mail: wuy@guilfordpharm.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00758-3

3868
D. C. Limburg et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3867–3870
into the proline-binding pocket of FKBP129 suggests
that bridged bicyclic structures can serve as effective
ligands. Workers at Agouron reported that bicyclic
structures such as 1 and 2 are effective FKBP inhibitors
with 9.2 and 0.28 mM inhibition constants, respectively
(Fig. 2).9,10 We nowreport the synthesis and pre-
liminary biological evaluation of bicyclic, bridged ana-
logues of our previously described prolyl FKBP ligands,
exemplified by 8a.
acid and BF3 etherate followed by cyclopentadiene at
À78 ꢀC, yielding the cyclo addition adduct 3 in a highly
exo and diastereoselective manner.12 The observed exo/
endo ratio of 94:6 was consistent with reports in litera-
ture.13,14 The major adduct 3(exo) was easily separated
from the minor adduct 3(endo) by a silica gel column as a
single diastereoisomer. Both structures were confirmed
using 1D/2D NMR and NOE spectral techniques.
.
Scheme 2 illustrates the synthesis of the desired (N-
glyoxyl)prolyl esters and thioesters from the bicyclic
intermediates. Hydrogenation of benzylamine 3(exo)
using Pearlman’s catalyst resulted in 4. Moderate pres-
sure (70 psi) was required to achieve both deprotection
of the benzyl amine and saturation of the double bond.
Acylation of 4 with methyl chlorooxoacetate provided
the oxamate 5 in good yield. Reaction of the oxamate
with the appropriate Grignard reagents at À78 ꢀC
resulted in a selective and high-yielding reaction at the
more electrophilic keto carbonyl of the keto amide
moiety, furnishing the glyoxylate 6. Basic hydrolysis of
the ethyl ester using 2 N LiOH gave the free acid 7. The
final ester and thioester compounds (exo-analogues 8a,
8c, 8e, 8g, 8i) were obtained by coupling the free acid 7
with various alcohols and thiols using the DCC ester-
ification method. Similarly, synthesis from 3(endo) led
to the final endo-compounds (8b, 8d, 8f, 8h, 8j). Table 1
presents the structures of the compounds prepared.
One of the most widely used synthetic tools in the for-
mation of asymmetric bicyclic ring structures has been
the Diels–Alder reaction due to its high stereospecificity
and regioselectivity.11 Recently, there have been numer-
ous reports in literature describing the synthesis of cyc-
lic a-amino acid derivatives, in particular, aza-Diels–
Alder reactions performed with chiral imines resulting
in optically active a-amino acids.12
Scheme 1 shows the synthesis used to prepare the key
intermediates 3. Ethyl glyoxylate and (R)-1-phenylethyl
amine were condensed with a 5 A molecular sieves in
methylene chloride to form (R)-1-phenylethyl imine in
situ. The chiral imine was treated with trifluoroacetic
Similar to the aforementioned synthesis of the (N-
glyoxyl)ester and thioester series, several (N-sulfonyl)-
ester and thioester analogues were prepared as seen in
Scheme 3. Reaction of aminoester 4 with varying sulfo-
nyl chlorides afforded sulfonamides 9. Basic hydrolysis
of the ester yielded the free acid 10, which was followed
by esterification, giving the final ester and thioester
Figure 2. FKBP12 ligands.
Scheme 1. Reagents and conditions: (a) 5 A sieves, CH2Cl2, 0 ꢀC; (b) TFA, BF3 etherate, cyclopentadiene.
Scheme 2. Reagents and conditions: (a) 20% Pd(OH)2, EtOAc, 70 psi; (b) chlorooxoacetate, TEA, CH2Cl2, 0 ꢀC; (c) 1.25 equiv 1,1-dimethylpropyl
magnesium chloride, THF, À78 ꢀC, 3 h; (d) 2 N LiOH, MeOH, 3 h; (e) HO-R1 or HS-R1, 1,3-dicyclohexylcarbodiimide, 4-dimethylaminopyridine,
CH2Cl2, overnight.

D. C. Limburg et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3867–3870
3869
Scheme 3. Reagents and conditions: (a) ClSO2R2, TEA, CH2Cl2, 0 ꢀC; (b) 2 N LiOH, MeOH, 3 h; (c) HO-R1 or HS-R1, 1,3-dicyclohexyl-
carbodiimide, 4-dimethylaminopyridine, CH2Cl2, overnight.
-1-(3,3-dimethyl-1,2-dioxopentyl)-2-piperdinecarboxylate
(SB3) bound to FKBP12 (PDB code: 1FKG) was used
for all experiments.16 SB3 was used as a template to
Table 1. (N-Glyoxyl)prolyl esters and thioesters
Compd
Adduct
X
n
R1
IC50 (nM)
build both enantiomers of each bicyclic compound into
the active site of FKBP12. Each compound was mini-
mized in the active site with FKBP12 held rigid.17
8a
8b
8c
8d
8e
8f
8g
8h
8i
exo
endo
exo
endo
exo
endo
exo
endo
exo
O
O
O
O
O
O
S
S
S
S
3
3
3
3
4
4
3
3
2
2
C6H5
C6H5
Pyridine
Pyridine
C6H5
C6H5
C6H5
C6H5
CH(C6H5)2
CH(C6H5)2
670
>25,000
15,000
>25,000
710
5100
2300
22,000
980
>25,000
All of the bicyclic compounds fit into the active site of
FKBP12. The pharmacophore binding elements in the
active site utilized by potent ligands are two hydro-
phobic pockets (one prolyl binding pocket and a prox-
imal spherical binding pocket) and two hydrogen bond
donors (Tyr82 side-chain OH and Ile56 backbone NH).
The exo-analogues (l-enantiomer mimics) are predicted
to bind similarly to SB3, filling both hydrophobic
pockets and binding the Ile56 backbone NH and the
Tyr82 OH. Figure 3 shows the predicted binding mode
of compound 8a complexed with FKBP12 as an exam-
ple. Two less productive binding modes are predicted
for the endo-analogues (d-enantiomer mimics). In both
binding modes, the hydrophobic pockets are filled,
though the orientation of the bicyclic group is different
in each binding mode. In the first binding mode an
optimal interaction with the Tyr82 OH is maintained
but at the expense of a weaker interaction with the Ile56
backbone NH [2.0 A (l) vs 2.5 A (d)]. In the second
binding mode an optimal interaction with the Ile56
backbone NH is retained, which results in the oblitera-
tion of the interaction with the Tyr82 OH. Since opti-
mal interactions with the four pharmacophore binding
elements can not be maintained in either of the pre-
dicted binding modes for the endo-analogues, it is pre-
dictable that they have lower IC50s than the
corresponding exo-analogues.
8j
endo
Table 2. N-Sulfonyl esters and thioesters
Compd
Adduct
X
n
R1
R2
IC50 (nM)
11a
11b
11c
11d
exo
exo
exo
exo
O
S
O
S
3
3
3
3
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
CH2C6H5
CH2C6H5
650
910
1300
720
compounds (11a–d). Table 2 presents the structures of
the compounds prepared.
Inhibition of the rotamase activity of FKBP12 by test
compounds was assayed as described by Kofron,15
using the peptide N-succinyl Ala-Leu-Pro-Phe-p-nitro-
anilide (Bachem) as the substrate. The IC50 values were
obtained from three independent estimations with
deviations generally lower than 15%, and used as mea-
sures of relative ligand binding affinities. The results of
FKBP12 inhibitory activity are included in Tables 1 and
2. From the structure–activity data, it appears that the
exo-analogues, which mimic the natural l-proline/
l-pipecolic acid, showsignificant activity with IC 50s
ranging from high nanomolar to lowmicromolar ( 8c is
an exception). The IC50s are reduced by approximately
one order of magnitude when compared to their pro-
plyl/pipecolyl ring counterparts.4 In contrast, the endo-
analogues lose significant inhibitory activity with most
of endo compounds failing to showactivity up to a 25
mM concentration. This is consistent with our previous
observations of l and d stereoisomers of prolyl
compounds.7 As shown in Table 2, N-sulfonyl esters
and thioesters are generally equipotent to their glyoxyl
analogues.
In order to evaluate the neuroprotective and neuro-
trophic properties of these constrained bicyclic com-
pounds, we have tested representative compounds to an
in vivo model of neurodegeneration. MPTP lesioning of
dopaminergic neurons in mice was used as an animal
model of Parkinson’s disease as described previously.5
Four-week-old male CD1 white mice were dosed ip with
30 mg/kg of MPTP for 5 days. In a ‘post-MPTP’ pro-
tocol, compounds 8a and 11a, or vehicle, were adminis-
tered 10 mg/kg po, beginning on the third day following
cessation of MPTP treatment, for 5 days. At 18 days
following MPTP treatment, the animals were sacrificed
and the striata were dissected and homogenized.
Immunostaining was performed on sagital and coronal
brain sections using anti-tyrosine hydoxylase Ig to
quantitate survival and recovery of dopaminergic neu-
rons. In animals treated with MPTP and vehicle, a sub-
stantial loss of 60–70% of functional dopaminergic
terminals was observed as compared to non-lesioned
Molecular modeling experiments were performed to
better understand the SAR of the bicyclic compounds.
The crystal structure of (1R)-1,3-diphenyl-1-propyl-(2S)

3870
D. C. Limburg et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3867–3870
G. S.; Dawson, T. M.; Hester, L.; Snyder, S. H. Nat. Med.
1997, 3, 421.
4. Hamilton, G. S.; Huang, W.; Connolly, M. A.; Ross, D. T.;
Guo, H.; Valentine, H. L.; Suzdak, P. D.; Steiner, J. P. Bioorg.
Med. Chem. Lett. 1997, 7, 1785.
5. Steiner, J. P.; Hamilton, G. S.; Ross, D. T.; Valentine,
H. L.; Guo, H.; Connolly, M. A.; Liang, S.; Ramsey, C.; Li,
J.-H.; Huang, W.; Howorth, P.; Soni, R.; Fuller, M.; Sauer,
H.; Nowotnick, A.; Suzdak, P. D. Proc. Natl. Acad. Sci.
U.S.A. 1997, 94, 2019.
6. Hamilton, G. S.; Steiner, J. P. Curr. Pharm. Des. 1997, 3,
405.
7. Hamilton, G. S.; Wu, Y. Q.; Limburg, D.; Wilkinson, D.;
Vaal, M.; Li, J.-H.; Thomas, C.; Huang, W.; Sauer, H.; Ross,
D. T.; Soni, R.; Chen, Y.; Guo, H.; Howorth, P.; Valentine,
H.; Liang, S.; Spicer, D.; Fuller, M.; Steiner, J. J. Med. Chem.
2002, 16, 3549.
8. Wu, Y. Q.; Wilkinson, D.; Limburg, D.; Li, J-H.; Sauer,
H.; Ross, D. T.; Liang, S.; Spicer, D.; Valentine, H.; Fuller,
M.; Guo, H.; Howorth, P.; Soni, R.; Chen, Y.; Steiner, J.;
Hamilton, G. S. J. Med. Chem. 2002, 16, 3550.
9. Babine, R. E.; Bleckman, T. M.; Kissinger, C. R.; Sho-
walter, R.; Pelletier, L. A.; Lewis, C.; Tucker, K.; Moomaw,
E.; Parge, H. E.; Villafranca, J. E. Bioorg. Med. Chem. Lett.
1995, 5, 1719.
Figure 3. The predicted binding mode of 8a complexed to FKBP12. 8a
and FKBP12 residues Ile56 and Tyr82 are shown in capped sticks. The
surface of FKBP12 is shown in yellow.
10. Babine, R. E.; Bleckman, T.; Littlefield, E. S.; Parge, H. E.;
Pelletier, L. A. K.; Lewis, C. T.; French, J. V.; Imbacuan, M.;
Katoh, S.; Tatlock, J. H.; Showalter, R. E.; Villafranca, J. E.
Bioorg. Med. Chem. Lett. 1996, 6, 385.
11. Stella, L.; Abraham, H. Tetrahedron Lett. 1990, 31, 2603.
12. Sodergren, M.; Andersson, P. Tetrahedron Lett. 1996, 37,
7577.
13. Waldmann, H.; Braun, M. Liebigs Ann. Chem. 1991, 1045.
Nakano, H.; Kumagai, N.; Kabuto, C.; Matsuzaki, H.;
Hongo, H. Tetrahedron: Asymmetry 1995, 6, 1233.
14. Abraham, H.; Stella, L. Tetrahedron 1992, 48, 9707. Bai-
ley, P.; Brown, G.; Korber, F.; Reed, A.; Wilson, R. Tetra-
hedron: Asymmetry 1991, 2, 1263.
animals. However, animals that received these com-
pounds subsequent to lesioning with MPTP were effec-
tive in restoring striatal innervention. Administered
orally, compounds 8a and 11a regenerated striated
dopaminergic terminals to about 48 and 25% of control
levels, respectively.
In conclusion, we have used the hetero Diels–Alder
reaction to synthesize a novel series of chiral bicyclic
proline ligands that are mimetics of the FKBP-binding
domain of FK506. exo Analogues of these compounds
were shown to retain good binding affinity toward
FKBP12 while some representative compounds have
shown efficacy in a mouse model of Parkinson’s disease,
suggesting their potential therapeutic utility in treating
degenerative disorders of the nervous system.
15. Kofron, J. L.; Kuzmic, P.; Kishore, V.; Colon-Bonilla, E.;
Rich, D. H. Biochemistry 1991, 30, 6127.
16. Holt, D. A.; Luengo, J. I.; Yamashita, D. S.; Oh, H. J.;
Konialian, A. L.; Yen, H. K.; Rozamus, L. W.; Brandt, M.;
Bossard, M. J.; Levy, M. A.; Eggleston, D. S.; Liang, J.;
Schultz, L. W.; Stout, T. J.; Clardy, J. J. Am. Chem. Soc. 1993,
115, 9925.
17. All of the molecular modeling was performed in the
SYBYL modeling environment.18 The Tripos forcefield19 with
Gasteiger–Huckel atomic charges20 were used in all calcula-
tions. A non-bonded cutoff of 25 A was utilized with a dis-
tance dependent dielectric.
18. SYBYL1 6.7.1; Tripos Inc.: 1699 South Hanley Rd., St.
Louis, MO 63144, USA.
References and Notes
1. Snyder, S. H.; Sabatini, D. E. M.; Lai, M. M.; Steiner, J. P.;
Hamilton, G. S.; Suzdak, P. D. Trends Pharm. Sci. 1998, 19,
21.
2. Hamilton, G. S.; Steiner, J. P. J. Med. Chem. 1998, 41,
5119.
3. Steiner, J. P.; Connolly, M. A.; Valentine, H. L.; Hamilton,
19. Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. J.
Comp. Chem. 1989, 10, 982.
20. Gasteiger, J.; Marsili, M. Tetrahedron 1980, 36, 3219.