Inhibitors of phenylethanolamine N-methyltransferase devoid of α2-adrenoceptor affinity

Article abstract of DOI:10.1016/j.bmcl.2005.08.033

A series of 3-trifluoromethyl-1,2,3,4-tetrahydroisoquinolines was synthesized and evaluated for their phenylethanolamine N-methyltransferase (PNMT) inhibitory potency and affinity for the α₂-adrenoceptor. Although their PNMT inhibitory potency decreased compared with corresponding 3-methyl-, 3-hydroxymethyl- or 3-unsubstituted-THIQs, some of them showed good selectivity due to their extremely low α₂-adrenoceptor affinity.

Full text of DOI:10.1016/j.bmcl.2005.08.033

Bioorganic & Medicinal Chemistry Letters 15 (2005) 5319–5323
Inhibitors of phenylethanolamine N-methyltransferase devoid
of a2-adrenoceptor affinityq
Gary L. Grunewald,* Jian Lu, Kevin R. Criscione and Cosmas O. Okoro
Department of Medicinal Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence, KS 66045, USA
Received 5 July 2005; revised 5 August 2005; accepted 9 August 2005
Available online 19 September 2005
Abstract—A series of 3-trifluoromethyl-1,2,3,4-tetrahydroisoquinolines was synthesized and evaluated for their phenylethanolamine
N-methyltransferase (PNMT) inhibitory potency and affinity for the a2-adrenoceptor. Although their PNMT inhibitory potency
decreased compared with corresponding 3-methyl-, 3-hydroxymethyl- or 3-unsubstituted-THIQs, some of them showed good selec-
tivity due to their extremely low a2-adrenoceptor affinity.
Ó 2005 Elsevier Ltd. All rights reserved.
Phenylethanolamine N-methyltransferase (PNMT; EC
2.1.1.28) catalyzes the last step of epinephrine biosyn-
thesis.2 Inhibitors of PNMT are potential pharmacolog-
ical tools for the study of the function of epinephrine in
the central nervous system (CNS).3–6 Previous studies
have found that 1,2,3,4-tetrahydroisoquinolines (THI-
Qs) are potent inhibitors of PNMT (Table 1).7–13 How-
ever, most of these inhibitors either display significant
affinity for the a2-adrenoceptor (e.g., 1,2),10,14 which
complicates the interpretation of their biological effects,
or those inhibitors that are selective for PNMT are too
polar to pass the blood–brain barrier (BBB) (e.g., 3,
4).8,12,15 In light of these limitations, there remains an
ongoing interest in the development of a PNMT inhibi-
tor that is both sufficiently lipophilic to penetrate the
BBB and is devoid of affinity for the a2-adrenoceptor.
whereas the endogenous ligands at the a2-adrenoceptor
would be largely protonated) or the steric hindrance of
the trifluoromethyl moiety.16 Since THIQs with lipophilic
7-substituents usually show good PNMT inhibitory
potency but also good a2-adrenoceptor affinity,10 com-
pounds 5b, 5c, and 5e were proposed to explore the possi-
bility of increasing selectivity while maintaining the
PNMT inhibitory potency of these THIQs. Second, the
trifluoromethyl moiety would also increase the lipophilic-
ity of these compounds thereby increasing the possibility
of these THIQs to cross the BBB. Compounds 5h and 5j
were proposed to take advantage of this feature as THIQs
with 7-aminosulfonyl substituents generally possess good
PNMT inhibitory potency and poor a2-adrenoceptor
affinity, but are usually too polar for BBB penetration.8,12
To further enhance their PNMT inhibitory potency, the
enantiomers of several of the most potent 3-trifluorom-
ethyl-THIQs in this study were also prepared and
evaluated.18
Previously, several 3-trifluoromethyl-THIQs (i.e., 5a, 5d,
5f, 5g, and 5i in Table 2) were studied.16 Although these
compounds displayed decreased inhibitory potency for
PNMT compared to similarly substituted 3-methyl- or
3-hydroxymethyl-THIQs,17 their potentially attractive
features suggested a solution to the common problems
of most PNMT inhibitors. First, this series of inhibitors
showed extremely low affinity for the a2-adrenoceptor
(Ki, 400–3900 lM), possibly due to the low pKa of the
3-trifluoromethyl-THIQ amine (e.g., for 5a, pKa = 4.86;
it would be largely unprotonated at physiological pH,
Compounds 5b and 5c were synthesized from amine 6
(Scheme 1). Lactam 8 had been synthesized previously,16
but the yield of the cyclization of 7 was improved from
53% (with polyphosphoric acid at 140 °C) to 86% (with
POCl3/SnCl4 at 110°C). Hydrogenation of 9 gave
aniline 10, which was converted to iodo-lactam 11
via a Sandmeyer iodination.15 Reduction of 11 with
BH3ÆTHF gave 5b. Treatment of 11 with FSO2CF2-
CO2CH3 and CuI in DMF gave the trifluoromethyl ana-
logue (12), which was reduced to 5c.
Keywords: Phenylethanolamine-N-methyltransferase; Inhibitor; a2-
Adrenoceptor; Selectivity; Blood–brain barrier.
q
.
THIQs 5e, 5h, and 5j were synthesized from aniline 10
(Scheme 2), which was converted to chloro-lactam 13,
See Ref. 1.
Corresponding author. E-mail: ggrunewald@ku.edu
*
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.08.033

5320
G. L. Grunewald et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5319–5323
Table 1. In vitro human PNMT (hPNMT) and a2-adrenoceptor affinities of some PNMT inhibitors
Compound
Ki (lM SEM)
Selectivity a2/hPNMT
ClogPb
a
hPNMT
a2
1c
2e
3f
0.0031 0.0006d
0.093 0.007
0.28 0.02c
0.021 0.005
0.22 0.04
100 10
7
2.4
2.90
2.72
360
27,000
À0.29
À0.93
4g
0.052 0.004h
1400 200
a In vitro activities for the inhibition of [3H]clonidine binding to the a2-adrenoceptor.
b Calculated logP values (ClogP function in SYBYL 6.9; Ref. 28).
c Ref. 7.
d Ref. 11.
e Ref. 10, previous data were for bovine PNMT.
f Ref. 8.
g Ref. 12.
h Ref. 13.
Table 2. In vitro activities of 3-trifluoromethyl-THIQs
Compound
R
Ki (lM SEM)
Selectivity a2/hPNMT
ClogPb
a
hPNMT
23
a2
5a( )c
H
I
2
400 40
17
2.87
4.00
5b( )
R-(+)
S-(À)
1.9 0.1
0.94 0.06
4.2 0.2
>1000
d
>530
d
5c( )
R-(+)
S-(À)
CF3
0.98 0.11
0.50 0.03
1.9 0.1
>1000
d
>1000
3.76
d
5d( )c
Br
Cl
3.2 0.3
>1000
>310
3.74
3.59
5e( )
R-(+)
S-(À)
0.99 0.08
0.46 0.02
1.9 0.2
>1000
d
>1000
d
5f( )c
R-(+)
S-(À)
NO2
2.3 0.1
1.2 0.1
4.4 0.3
1400 200
d
610
2.62
d
5g( )c
5h( )
5i( )c
5j( )
CN
21
60
41
2
6
2
2900 300
>1000
140
>17
95
2.31
2.19
1.23
1.04
SO2NHEt
SO2CH3
SO2NH2
3900 500
8.0 0.7
>1000
>125
a In vitro activities reported for the inhibition of binding of [3H]clonidine at the a2-adrenoceptor.
b Calculated logP values (ClogP function in SYBYL 6.9; Ref. 28).
c Ref. 16, previous data were for bovine PNMT.
d These values were not determined because of the extremely low a2-adrenoceptor affinity or the solubility of the corresponding racemates.
followed by reduction with BH3ÆTHF to afford 5e.
Treatment of lactam 8 with chlorosulfonic acid (neat)
gave 14, which was reacted with ethylamine or ammoni-
um hydroxide to give 15a or 15b, followed by reduction
with BH3ÆTHF to give 5h or 5j.15
The enantiomers of 5b, 5c, 5e, and 5f were separated by
chiral HPLC.19 The absolute configurations of 5c and 5e
were established by the X-ray structure analysis of (R)-
5cÆHCl and (R)-5eÆHCl.20 To establish the absolute con-
figurations of 5b and 5f, (R)-5b and (R)-5f were prepared

G. L. Grunewald et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5319–5323
5321
Figure 1. This figure shows the amino acids (carbon is white, nitrogen
is blue, oxygen is red, sulfur is yellow, and bromine and fluorine are
green) interacting with 7-iodo-THIQ (2; stick model, carbon is
magenta) within the active site of hPNMT. Asn39 and Val269 are
not shown. A Connolly (solvent accessible) surface shows lipophilic
areas in brown, hydrophilic areas in blue, and neutral areas in green.
The docking results of R-5b (stick model, carbon is white) and R-18a
(stick model, carbon is yellow) are also shown. Hydrogens are not
shown for clarity.
Scheme 1. Reagents: (a) ClCO2Me, pyridine, CHCl3, 92%; (b) POCl3,
SnCl4, 86%; (c) H2SO4, KNO3, 90%; (d) H2, PtO2, MeOH, 99%;
(e) HCl, NaNO2; (f) CuI, KI, 61% from 10; (g) BH3ÆTHF, THF, 78%;
(h) FSO2CF2CO2Me, CuI , DMF, 27%.
The crystal structure of hPNMT complexed with 2 and
S-adenosyl-L-homocysteine (AdoHcy) has recently been
published.26 To help explain the reduced hPNMT inhib-
itory potency of the 3-trifluoromethyl-THIQs, molecu-
lar docking calculations were conducted using
AutoDock 3.027 and Sybyl 6.9.28 Compound 5b was
docked into the hPNMT active site based on the
X-ray crystal structure of hPNMT-2-AdoHcy (Fig. 1).
The docking study indicated that the binding orienta-
tion of 5b is similar to that of 2 in the active site of
PNMT, except that the THIQ ring of 5b is shifted about
˚
1 A away from Glu219 and Tyr222 as compared with 2.
Previous studies have shown that substitution of a meth-
yl or hydroxymethyl group at the 3-position of THIQs
increased the potency of THIQs for PNMT.17 However,
these studies also indicated that the active site of PNMT
has a limited amount of steric bulk tolerance for 3-sub-
stituents of THIQs.17 The comparisons of 5d and 5f with
the corresponding 3-methyl, 3-ethyl, and 3-isopropyl
substituted THIQs25 are shown in Table 3. A values
(conformational energies of cyclohexanes, ÀDG0) of
these 3-substituents are also included as indicators of
group sizes.29 As the size of the 3-substituent increases
from methyl to isopropyl (A value increases from 1.70
to 2.15 kcal/mole), the hPNMT inhibitory potency of
these compounds decreases. Although a fluorine (van
Scheme 2. Reagents: (a) HCl, NaNO2; (b) CuCl, KCl, 76% from 10;
(c) HSO3Cl, 77%; (d) EtNH2ÆHCl, EtOAc/Na2CO3, 87% for 15a; (e)
NH4OH, CH3CN, 71% for 15b; (f) BH3ÆTHF, THF, 85% for 5e, 94%
for 5h, 63% for 5j.
from amine (R)-6 (97% ee), which was synthesized
according to literature procedures.21
Radiochemical assays described previously were used to
determine human PNMT (hPNMT) inhibition con-
stants11,22 and a2-adrenoceptor binding affinities.23,24
˚
In general, the 3-trifluoromethyl-THIQs (Table 2)
showed decreased hPNMT inhibitory potency as com-
pared to the similarly 7-substituted 3-methyl-,25 3-hy-
droxymethyl-12,13 or 3-unsubstituted-THIQs.11 For
example, compared with THIQs 2 (Ki = 0.093 lM)
and 3 (Ki = 0.28 lM), both 3-trifluoromethyl-THIQ
analogues show decreased potency at hPNMT (5b,
Ki = 1.9 lM; 5j, Ki = 8.0 lM). For the hPNMT inhibito-
ry potency of different 3-trifluoromethyl-THIQs, com-
pounds bearing a lipophilic 7-substituent (e.g., 5b, 5c,
and 5e) showed higher potency than compounds bearing
a hydrophilic 7-substituent (e.g., 5g, 5i, and 5j).
der Waals radius 1.47 A) is a close isosteric substitution
for a hydrogen (van der Waals radius 1.20 A),30 a triflu-
˚
oromethyl group is considerably larger than a methyl
group. Studies have shown that the trifluoromethyl moi-
ety is close to the size of an isopropyl group.31 Interest-
ingly, when comparing their affinities for hPNMT,
3-trifluoromethyl-THIQs (i.e., 5d, 5f) displayed Ki val-
ues close to those of corresponding 3-isopropyl-THIQs
(i.e., 18a, 18b). Furthermore, a docking study of 18a
(Fig. 1) also showed a similar shift of the molecule in
the active site of PNMT as observed in the study of
5b. In the crystal structure of the hPNMT–2–AdoHcy

5322
G. L. Grunewald et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5319–5323
Table 3. In vitro activities of ( )-3-alkyl-7-nitro and -7-bromo-THIQs
Compound
R3
R7
A value (R3)a
Ki (lM SEM)
a2 Ki (lM SEM)b
hPNMT
c
16a( )
CH3
C2H5
Br
Br
1.70
1.75
2.15
2.10
1.70
1.75
2.15
2.10
0.017 0.005
0.48 0.02
4.4 0.3
1.1 0.1
1.2 0.1
3.9 0.3
17a( )c
18a( )c
5d( )d
CH(CH3)2
CF3
CH3
Br
Br
3.2 0.3
>1000
31 1
16b( )c,e
17b( )c
18b( )c
5f( )d
NO2
NO2
NO2
NO2
0.072 0.005
0.49 0.03
4.6 0.3
C2H5
CH(CH3)2
CF3
28 0.3
36 0.3
1400 200
2.3 0.1
a A value: conformational energy (kcal/mol) (Ref. 29).
b In vitro activities reported for the inhibition of binding of [3H]clonidine at the a2-adrenoceptor.
c Ref. 25.
d Ref. 16, previous data were for bovine PNMT.
e Ref. 12.
complex, the aliphatic amine of 2 forms a hydrogen
bond to the carboxylate of Glu219.26 In addition, the
aromatic ring of 2 is sandwiched between Phe182 and
Asn39,26 suggesting the presence of an aromatic p–p
stacking interaction between the aromatic ring of 2
and Phe182. The shift of the THIQ ring of 5b as predict-
ed by docking studies (Fig. 1), possibly induced by the
unfavorable steric interaction of the 3-trifluoromethyl
moiety, could disrupt both the H-bond interaction with
Glu219 and the p–p stacking interaction with Phe182.
can greatly decrease the affinity of THIQ for the a2-
adrenoceptor. On the other hand, the steric hindrance
of the trifluoromethyl group is likely to be the major rea-
son for the reduced PNMT inhibitory potency of these
compounds. THIQs 5c and 5e are two highly selective
sub-micromolar inhibitors of PNMT with high lipophil-
icity (ClogP values greater than that of 1) that should
enable them to cross the BBB.
Acknowledgments
None of this series of 3-trifluoromethyl-THIQs dis-
played significant affinity toward the a2-adrenoceptor,
although the limited solubility of some compounds pre-
cluded determination of exact Ki values. Previous com-
parative molecular field analysis (CoMFA) studies
on a series of THIQs indicated an area of steric bulk
intolerance at the 3-position of THIQs for the a2-adre-
noceptor.32 Interestingly, when comparing the a2-adren-
oceptor affinities of THIQs in Table 3, no significant
changes are observed from the 3-methyl-THIQs (i.e.,
16a, 16b) to the 3-isopropyl-THIQs (i.e., 18a, 18b).
However, dramatic decreases in a2-adrenoceptor affini-
ties are noticed between 3-isopropyl-THIQs and 3-triflu-
oromethyl-THIQs (i.e., 5d, 5f) even though the
trifluoromethyl moiety is close in size to the isopropyl
group. This comparison reinforces the hypothesis that
the significant reduction of the a2-adrenoceptor affinities
of the 3-trifluoromethyl-THIQs is due mainly to the de-
crease in pKa of the THIQ amine rather than due to
unfavorable steric interactions.16
This research was supported by NIH Grant HL 34193.
We thank David VanderVelde and Sarah Neuenswander
of the University of Kansas Nuclear Magnetic Reso-
nance Laboratory for their assistance. We thank Doug-
las Powell of the X-ray crystallography laboratory
(National Science Foundation Grant CHE-0079282)
for the crystal structures of (R)-5cÆHCl and (R)-5eÆHCl.
We also thank Gerald Lushington of the University of
Kansas Molecular Graphics and Modeling Laboratory
and Todd Williams of the University of Kansas Mass
Spectrometry Laboratory for their assistance.
References and notes
1. The contents of this paper were taken in large part from the
Ph.D. dissertation, Univeristy of Kansas, 2005, Jian Lu.
2. Axelrod, J. J. Biol. Chem. 1962, 237, 1657.
3. Mefford, I. N. Prog. Neuropsychopharmacol. Biol. Psychi-
atry 1988, 12, 365.
4. Fuller, R. W. Annu. Rev. Pharmacol. Toxicol. 1982, 22, 31.
5. Saavedra, J. M.; Grobecker, H.; Axelrod, J. Science 1976,
191, 483.
6. Perry, B. D.; Stolk, J. M.; Vantini, G.; Guchhait, R. B.;
UꢀPrichard, D. C. Science 1983, 221, 1297.
7. Pendleton, R. G.; Kaiser, C.; Gessner, G. J. Pharmacol.
Exp. Ther. 1976, 197, 623.
Comparison of the hPNMT inhibitory activities of the
enantiomers of the most potent racemates in Table 2
(i.e., 5b, 5c, 5e, and 5f) shows that the R-enantiomers
are approximately 4-fold as potent as their correspond-
ing S-enantiomers. This result is consistent with the
observations of other 3-substituted THIQs.18,33
8. Pendleton, R. G.; Gessner, G.; Weiner, G.; Jenkins, B.;
Sawyer, J.; Bondinell, W.; Intoccia, A. J. Pharmacol.
Exp.Ther. 1979, 208, 24.
In conclusion, as a strong electron-withdrawing group
at the 3-position of THIQ, the trifluoromethyl moiety

G. L. Grunewald et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5319–5323
5323
9. Bondinell, W. E.; Chapin, F. W.; Girard, G. R.; Kaiser,
C.; Krog, A. J.; Pavloff, A. M.; Schwartz, M. S.; Silvestri,
J. S.; Vaidya, P. D.; Lam, B. L.; Wellman, G. R.;
Pendleton, R. G. J. Med. Chem. 1980, 23, 506.
10. Grunewald, G. L.; Dahanukar, V. H.; Jalluri, R. K.;
Criscione, K. R. J. Med. Chem. 1999, 42, 118.
11. Wu, Q.; Criscione, K. R.; Grunewald, G. L.; McLeish, M.
J. Bioorg. Med. Chem. Lett. 2004, 14, 4217.
12. Grunewald, G. L.; Dahanukar, V. H.; Teoh, B.; Criscione,
K. R. J. Med. Chem. 1999, 42, 1982.
13. Grunewald, G. L.; Romero, F. A.; Criscione, K. R.
J. Med. Chem. 2005, 48, 134.
obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK [fax: +44 01223
336033 or e-mail: data_request@ccdc.cam.ac.uk].
21. Soloshonok, V. A.; Ono, T. J. Org. Chem. 1997, 62, 3030.
22. Romero, F. A.; Vodonick, S. M.; Criscione, K. R.;
McLeish, M. J.; Grunewald, G. L. J. Med. Chem. 2004,
47, 4483.
23. UꢀPrichard, D. C.; Greenberg, D. A.; Snyder, S. H. Mol.
Pharmacol. 1977, 13, 454.
24. Grunewald, G. L.; Romero, F. A.; Seim, M. R.; Criscione,
K. R.; Deupree, J. D.; Spackman, C. C.; Bylund, D. B.
Bioorg. Med. Chem. Lett. 2005, 15, 1143.
14. Toomey, R. E.; Horng, J. S.; Hemrick-Luecke, S. K.;
Fuller, R. W. Life Sci. 1981, 29, 2467.
15. Grunewald, G. L.; Caldwell, T. M.; Li, Q.; Slavica, M.;
Criscione, K. R.; Borchardt, R. T.; Wang, W. J. Med.
Chem. 1999, 42, 3588.
16. Grunewald, G. L.; Caldwell, T. M.; Li, Q.; Criscione, K.
R. J. Med. Chem. 1999, 42, 3315.
17. Grunewald, G. L.; Sall, D. J.; Monn, J. A. J. Med. Chem.
1988, 31, 824.
25. Grunewald, G. L.; Romero, F. A.; Chieu, A. D.; Fincham,
K. J.; Bhat, S. R.; Criscione, K. R. Bioorg. Med. Chem.
2005, 13, 1261.
26. McMillan, F. M.; Archbold, J.; McLeish, M. J.; Caine, J.
M.; Criscione, K. R.; Grunewald, G. L.; Martin, J. L.
J. Med. Chem. 2004, 47, 37.
27. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.;
Hart, W. E.; Belew, R. K.; Olson, A. J. J. Comput. Chem.
1998, 19, 1639.
18. Grunewald, G. L.; Caldwell, T. M.; Li, Q.; Dahanukar, V.
H.; McNeil, B.; Criscione, K. R. J. Med. Chem. 1999, 42,
4351.
28. SYBYL, version 6.9; Tripos, 1699 South Hanley Rd., St.
Louis, MO, 63144, USA.
29. Hirsch, J. A. Top. Stereochem. 1967, 1, 199.
30. Bondi, A. J. Phys. Chem. 1964, 68, 441.
31. Bott, G.; Field, L. D.; Sternhell, S. J. Am. Chem. Soc.
1980, 102, 5618.
32. Grunewald, G. L.; Caldwell, T. M.; Dahanukar, V. H.;
Jalluri, R. K.; Criscione, K. R. Bioorg. Med. Chem. Lett.
1999, 9, 481.
33. Grunewald, G. L.; Romero, F. A.; Criscione, K. R.
J. Med. Chem. 2005, 48, 1806.
19. The enantiomers (>98% ee) of 5b, 5c, 5e, and 5f were
separated on a CHIRALCEL OJ semi-prep column
(Chiral Technologies, Inc. West Chester, PA, USA) with
hexanes/2-propanol (70/30) as the mobile phase.
20. Crystallographic data of (R)-5cÆHCl and (R)-5eÆHCl have
been deposited with the Cambridge Crystallographic Data
Centre as supplementary publication numbers CCDC
279450 and CCDC 279451. Copies of the data can be