Design, synthesis and pharmacological test of a quinoline based, nonpeptidic analogue of neurotensin(8-13)

Article abstract of DOI:10.1039/a700894e

The design, synthesis and pharmacological testing of a quinoline based, nonpeptidic analogue of neurotensin(8-13) using our reported multiple template approach to developing nonpeptidic mimetics of neuropeptides are reported. The newly synthesized quinoli

Full text of DOI:10.1039/a700894e

View Article Online / Journal Homepage / Table of Contents for this issue
Design, synthesis and pharmacological test of a quinoline based,
nonpeptidic analogue of neurotensin(8–13)
Feng Hong,a Yuan-Ping Pang,*,a Bernadette Cusack b and Elliott Richelsonb
a
Neurochemistry Research, Mayo Foundation for Medical Education and Research,
4500 San Pablo Road, Jacksonville, FL 32224, USA
b Neuropsychopharmacology Research, Mayo Foundation for Medical Education and Research,
4500 San Pablo Road, Jacksonville, FL 32224, USA
The design, synthesis and pharmacological testing of a quinoline based, nonpeptidic analogue of
neurotensin(8–13) using our reported multiple template approach to developing nonpeptidic mimetics of
neuropeptides are reported. The newly synthesized quinoline analogue is found to be less active in binding
to the neurotensin receptors than previously reported mimetics-1 and -2 which are partial nonpeptidic
analogues of neurotensin(8–13). H owever, the present study led to the discovery of a mistake in a
literature procedure for alkylation at position-3 of ethyl indole-2-carboxylate and a consequent mistake in
the synthesis of mimics-1 and -2. With the computational and experimental results of the quinoline
analogue, along with mimics-1 and -2 with the correct structures, which, accordingly, served as a blind test
for the multiple template approach, the concept of this approach has been experimentally validated.
NH
Introduction
We previously reported the development of two partial nonpep-
tidic neurotensin(8–13) mimetics (mimics-1 and -2, Fig. 1)
whose Arg8Arg9Pro10 portion is replaced with substituted
indole-2-carboxylates as partially flexible non-peptidic equiva-
lents by our multiple template approach.1–3 Neurotensin(8–
13) is a peptide fragment of neurotensin (pGlu1Leu2Tyr3-
Glu4Asn5Lys6Pro7Arg8Arg9Pro10Tyr11Ile12Leu13) and is more
biologically relevant than the parent peptide.4 Full nonpeptidic
neurotensin(8–13) mimetics can be used for potential treat-
ments of neuropsychiatric diseases such as schizophrenia and
Parkinson’s disease and for evaluating further the physiological
and putative pathological roles of the neurotensin receptors.
The multiple template approach is a method for converting a
peptide of interest into a nonpeptidic mimic, which can then be
used as a therapeutic drug. The principle of this approach is to
convert a vast number of conformers of a peptide to a small
number of partially flexible molecules that can individually
mimic a different portion of the conformers of the native pep-
tide and altogether mimic all the conformers available to the
native peptide. Each of these partially flexible molecules con-
sists of a template of different size responsible for mimicking a
certain portion of the conformers of the native peptide. Given
the conformational flexibilities of the designed molecules and
the conformational flexibility of the native peptide, one could
determine how many templates are required to be tested. By
testing all the partially flexible molecules, one should arrive at a
molecule that covers the receptor-bound conformation of the
native peptide and fits to the receptor to achieve the desired
biological functions.
HN
NH2
(CH2)6
O
(H2C)5
NH
H2N
NH TyrIleLeuOH
N
H
NH
O
Mimic-1
NH
HN
NH2
(CH2)6
NH TyrIleLeuOH
N
H
O
O
(H2C)5
NH
H2N
Mimic-2
NH
O(CH2)5NH2
O
H
N
N
N
H
H2N(CH2)5O
O
OH
Mimic-3
Fig. 1 Structures of mimics-1 and -2 reported previously and newly
designed mimic 3
To continue our efforts into developing full nonpeptidic
neurotensin(8–13) mimetics by the multiple template approach,
properties of mimic-2 will be retained or slightly altered when
the indole template is replaced by the quinoline template.1 This
is because the difference in the triangular size between the
indole and quinoline templates is about 1.2 Å on one side of the
triangles (see Fig. 2). According to the multiple template
approach, a difference in distance of 1.2 Å can be nullified
by the conformational flexibility of the designed molecule.1 The
structures of the templates were optimized by quantum
mechanics calculations employing the Gaussian 94 program
with the 6-31G* basis set at the Hartree–Fock level.5
we designed, synthesized and tested
analogue of mimic-2 (mimic-3, Fig. 1). We report herein the
results and implications of the studies of this new mimic.
a full nonpeptidic
Design
The notion of replacing the indole template used in mimics-1
and -2 with the quinoline template in the structure of mimic-3
was a further test of the multiple template approach. For this
purpose, we wanted to demonstrate that the pharmacological
J. Chem. Soc., Perkin Trans. 1, 1997
2083

View Article Online
(a)
b = 4.4
OH
b = 4.9
O(CH2)2CH3
a = 3.6
a = 3.6
N
O
N
N
3 equiv. KOH, DMSO,
OH 18-crown-6,
c = 7.2
O
N
c = 6.8
O(CH2)2CH3
N
1-iodopropane,
3 h, 90%
N
OH
O
CH3(CH2)2O
O
room temp.
b = 5.7
a = 3.7
1
b = 6.1 O
N
a = 3.7
Model-1
N
N
N
c = 7.2
(b)
OH
O
O
c = 6.2
OH–, 1-iodopropane
Fig. 2 Triangular sizes (Å) represented by the indole-2-carboxylic
amide (top) and quinoline-2-carboxylic amide (below) nuclei
OH
O(CH2)2CH3
N
N
Substitution of the last two residues of neurotensin(8–13) by
the 1-adamantyl group was based on our finding that the full
nonpeptidic analogue of mimic-2 in which the IleLeu was
replaced by the 1-adamantyl group demonstrated a two-fold
increase in affinity for human neurotensin receptors relative to
mimic-2.6 Use of amino groups in mimic-3 was based on
reported results that the analogues of mimics-1 and -2 in which
the two guanidino groups are replaced by amino groups dem-
onstrated a two- to three-fold decrease in affinity for the neuro-
tensin receptors.3 It was our intention to test the quinoline tem-
plate first and to transform the amino group into a guanidino
group later, if the quinoline template was experimentally found
to be appropriate. For synthetic efficiency, two ether linkages
were used in mimic-3 since 4,8-dihydroxyquinoline-2-carboxy-
lic acid was commercially available.
OH
O
CH3(CH2)2O
Model-2
O
1
Scheme 2 Model study of O-alkylation
O(CH2)5OTBDMS
OH
i, 69%
O(CH2)5OTBDMS
OH
N
N
TBDMSO(CH2)5O
O
OH
O
2
1
ii, 76%
O(CH2)5OH
O(CH2)5N3
iii, 81%
O(CH2)5OH
O(CH2)5N3
Synthesis
N
N
HO(CH2)5O
O
N3(CH2)5O
O
Mimic-3 was designed to make use of commercially available
materials as building blocks for its synthesis (Scheme 1).
Assembly of the synthons in Scheme 1 could be easily carried
4
3
iv, 93%
O(CH2)5N3
Mimic-3
OH
N
N3(CH2)5O
O
OH
5
H2N
+
CO2H
v, 90%
+
OH
H2N
N
O(CH2)5N3
OH
OH
O
O
1
H
N
vi, 95%
Scheme 1 Retrosynthetic analysis of mimic-3
Mimic-3
N
N
H
N3(CH2)5O
O
out by standard peptide chemistry. Introduction of the two
alkylamino chains of mimic-3 to 4,8-dihydroxyquinoline-2-
carboxylic acid 1 required, however, investigation to establish
reaction conditions that could exclusively yield O-alkylation of
compound 1 and to solve the problem that many O-alkylation
methods could not be employed because of the poor solubility
of compound 1 in most organic solvents.
According to our model study on alkylation with 1-iodo-
propane, we found that the reaction conditions in Scheme 2(a)
employing three equivalents of KOH and a catalytic amount of
18-crown-6 at room temperature for three h could yield
exclusive O-alkylation in high yield. The evidence for exclusive
O-alkylation of 4,8-dihydroxyquinoline-2-carboxylic acid in
our model study is: (i) that the observed chemical shifts of the
ten sp2 carbons ranged from 101 to 166 ppm, and (ii) that
the chemical shifts of the three methylene carbons attached to
the three heteroatoms of 4,8-dihydroxyquinoline-2-carboxylic
acid were 67.3, 70.1 and 70.4 ppm, respectively. One would
observe the chemical shift for one of the ten sp2 carbons in the
range of 180 to 200 ppm and the chemical shift for one of the
three methylene carbons at less than 50 ppm, if N-alkylation
occurs as shown in Scheme 2(b).
6
OBut
Scheme
3
Synthesis of mimic-3. Reagents and conditions: i,
I(CH2)5OTBDMS, KOH, a trace of 18-crown-6, DMSO, room temp.,
12 h; ii, TBAF, THF, room temp., 2 h; iii, HN3, PPh3, DEAD, CH2Cl2,
0 ЊC, 1.5 h; iv, 2  KOH, MeOH, THF, room temp., 1 h; v, O-But-Tyr-
NH-Ada, DCC, HOBt, DMF, 40 ЊC, 12 h; vi, H2, 10% Pd–C, EtOAc–
MeOH (2:1), conc. HCl, room temp. 12 h.
The synthesis of mimic-3 as outlined in Scheme 3 thus
started from O-alkylation of 4,8-dihydroxyquinoline-2-carb-
oxylic acid with 1-(tert-butyldimethylsilyloxy)-5-iodopentane.7
Similar reaction conditions using 1-(tert-butyldimethylsilyl-
oxy)-5-iodopentane were accordingly used to prepare inter-
mediate 2. The same evidence for the exclusive O-alkylation was
observed in the NMR spectra of intermediate 2. Intermediate 2
was then transformed to intermediate 5 according to the pub-
lished protocol.4 Additional support for the exclusive O-alkyl-
ation was also observed in (i) the carbon NMR spectrum of the
subsequent intermediate 5 where the chemical shifts of the two
methylene carbons directly attached to the two ethereal oxygen
atoms of the 4,8-dihydroxyquinoline moiety are almost identi-
2084
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
Table
1
Comparison of binding affinities for neurotensin,
O
(H2C)5
NH
neurotensin(8–13) and mimics-2 and -3 at human and mouse neuro-
tensin receptors
HN TyrIleLeuOH
H2N
N
O
Kd [n]
NH
(CH2)6
HN
NH2
Human neurotensin
receptor in CHO–K1
membranes
Mouse neurotensin
receptor in N1E-115
intact cells
NH
Compound
Mimic 1
Neurotensin(8–13)
Neurotensin
Mimic-2
0.14 ± 0.01 (4)
1.6 ± 0.06 (93)
14 500 ± 1 200 (4)
76 900 ± 6 600 (4)
0.61 ± 0.02 (3)11
8.2 ± 0.6 (15)
2 600 ± 300 (6)
nd
HN TyrIleLeuOH
O
Mimic-3
N
Kd = apparent equilibrium dissociation constant. Values are geometric
mean ± S.E.; each value is the mean of duplicate determinations made
in independent experiments; values in parentheses indicate n value;
nd = no data.
O
(CH2)6
(H2C)5
NH
HN
NH2
H2N
NH
Mimic 2
NH
cal at 70.0 ppm. Coupling of intermediate 5 to the intermediate
of O-But-Tyr-NH-Ada 10 in the presence of dicyclohexylcar-
bodiimide (DCC) and 1-hydroxybenzotriazole (HOBt) in
DMF afforded the precursor of mimic-3 (6) in good yield.
Intermediate 10 was synthesized from two commercially avail-
able starting materials in a high yield according to Scheme 4.
Fig. 3 The correct structures of Mimics 1–2
Discussion
The observed relatively low affinity of mimic-3 suggested that
the multiple template approach was incorrect. As explained
for the design of mimic-3, if the theory is correct, mimic-3
would show a similar affinity for the neurotensin receptors as
mimic-2.
O
H
N
+
Fmoc
OH
H2N
To confirm this implication, we re-examined all the experi-
mental results of mimics 1–3 and found that this implication
was not relevant. We first found a mistake in the literature pro-
cedure for alkylation at position-3 of ethyl indole-2-carboxylate
with ethyl 3-iodopropanoate using K2CO3 as base in CH3CN.12
According to such reagents and reaction conditions along with
the reported NMR data of ‘2-carboxyindole-3-propanoic
acid’,12 we think that the alkylation actually occurred at the
nitrogen atom, and not at position-3. This is evident from the
reported chemical shift (a triplet at 4.77 ppm) of the two
protons of the methylene group attached to the indole ring of
the ‘2-carboxyindole-3-propanoic acid’. The correct structure
prepared from indole-2-carboxylate by the reported procedure
should be2-carboxyindole-1-propanoicacid. Inevitably, employ-
ing this literature procedure for the synthesis of mimics-1 and
-2, alkylations of the indole intermediates with 1-(tert-butyl-
dimethylsilyloxy)-6-iodohexane actually occurred at position-
1, and not at position-3 as reported previously.1–3,13 Evidence for
the N-alkylation of mimics-1 and -2 is (i) disappearance of the
NH stretch peaks of the indole ring of mimics-1 and -2 in the
IR spectra, (ii) disappearance of the signals of protons attached
7
OBut
8
i, 100%
O
H
N
ii, 97%
O-But-Tyr-NH-Ada
Fmoc
N
H
10
OBut
9
Scheme 4 Synthesis of O-But-Tyr-NH-Ada (10). Reagents and condi-
tions: i, DCC, HOBt, DMF, room temp., 12 h; ii, piperidine, THF, room
temp., 2 h.
Precursor 6 was efficiently converted into mimic-3 in its tris-
hydrochloric acid salt form by catalytic hydrogenation in the
presence of drops of concentrated HCl, thus reducing the azido
groups into amine groups and deprotecting the tert-butyl group.
Purification by reversed-phase HPLC eluting with MeCN–TFA
yielded mimic-3 in its tris(trifluoroacetate) salt form.
1
to the nitrogen atoms of mimics-1 and -2 in the H NMR
spectra, and (iii) four triplets for four sets of protons of the
methylene groups attached to the heteroatoms observed in the
range of 4.97 to 3.57 ppm. One would observe a strong NH
stretch peak around 3335 cmϪ1 in the IR spectrum, a signal for
the nitrogen proton in the NMR spectrum, and only three trip-
lets in that range of chemical shift, if the alkylations had
occurred at the indole position-3. The correct structures of
mimics-1 and -2 are now shown in Fig. 3.
Given the correct structures of mimics-1 and -2, we recalcu-
lated the sizes of the triangles of the indole template (see Fig.
4). The sizes of the new triangles represented by the indole
template with correct linkages as depicted in Fig. 4 reveal
that differences in the triangular size between the indole and
quinoline templates are more than 1.2 Å1 on one side of the
triangles (see Figs. 2 and 4). This explains why mimic-3 is sig-
nificantly less active in binding than mimic-2. The sizes of the
new triangles of mimics-1 and -2 (Fig. 4) also provided an
important blind test of the multiple template approach. If the
sizes of the triangles in the correct structures of mimics-1 and
-2 are beyond the range of the triangles (i.e. a = 2.9 to 3.8 Å,
Pharmacology
Mimic-3 in its tris(trifluoroacetate) salt form was tested for its
ability to compete for [3H]neurotensin binding at the human
neurotensin receptors (NTR) that had been stably transfected
in CHO–KI cells. Cell culture and radioligand binding assays
were performed using previously described methods.8–10 The
binding affinities (Kd) of neurotensin, neurotensin(8–13) and
mimics-2 and -3 are listed in Table 1. We present values previ-
ously derived at the mouse neurotensin receptors in intact N1E-
115 cells for comparison. Neurotensin(8–13) was the most
potent at both the human and mouse NTR. Additionally, the
order of potency for the remaining compounds stayed the same
in both cell lines. An examination of the Kd values for mimics-3
and -2 reveals a more than five-fold decrease in binding affinity
of mimic-3 over mimic-2. Additionally, mimic-2 exhibited a
decreased affinity at the human neurotensin receptor in com-
parison to the mouse neurotensin receptors.
J. Chem. Soc., Perkin Trans. 1, 1997
2085

View Article Online
O
(5-tert-Butyldimethylsilyloxy)pentyl 4,8-bis[(5-tert-butyl-
c = 8.3
c = 8.2
dimethylsilyloxy)pentyloxy]quinoline-2-carboxylate 2
N
Dry DMSO (20 ml) was added to a mixture of 4,8-
dihydroxyquinoline-2-carboxylic acid (1025 mg, 5 mmol),
KOH powder (1.0 g, 15.2 mmol) and a catalytic amount of 18-
crown-6 under N2. After the mixture was stirred at room temp.
for 3 h, 1-iodo-5-(tert-butyldimethylsilyloxy)pentane (5.9 g, 18
mmol) was introduced to the mixture via a syringe. The reac-
tion was quenched with cold water after 12 h stirring at room
temp. The ethyl acetate extract was washed with water and then
with brine, dried (Na2SO4) and evaporated. Flash chroma-
tography on silica gel eluting with ethyl acetate–hexanes (1:9)
gave intermediate 2 (2.8 g, 69%) as a light yellow oil; νmax(neat)/
cmϪ1 1746, 1715, 1256 and 1100; δH (300 MHz; CDCl3) 0.03–
0.04 (18 H, m, SiCH3), 0.87–0.90 [27 H, m, C(CH3)3], 1.45–1.68
(12 H, m, 2-H2 ϩ 4-H2), 1.85–2.12 (6 H, m, 3-H2), 3.62–3.70 (6
H, m, 5-H2OSi), 4.19 (2 H, t, J 7.0, 1-H2OAr), 4.27 (2 H, t, J 6.3,
1-H2OAr), 4.42 (2 H, t, J 7.2, CO2CH2), 7.07 (1 H, d, J 7.2,
5-ArH), 7.47 (1 H, t, J 8.1, 6-ArH), 7.56 (1 H, s, 3-ArH) and
7.77 (1 H, d, J 7.5, 7-ArH); δC(75.46 MHz; CDCl3) 6.7, 18.0,
22.0, 22.2, 25.7, 28.1, 28.5, 32.2, 32.4, 62.5, 62.6, 62.7, 65.7,
68.5, 68.8, 100.8, 109.5, 112.9, 123.2, 127.5, 140.4, 147.8,
155.3, 162.2 and 165.7; m/z (ESI) 806 [M ϩ H]ϩ {Found
(HRMS): m/z 806.5242. Calc. [M ϩ H]ϩ for C43H80NO7Si3:
806.5243}.
N
N
N
a = 4.1
b = 4.3
a = 4.1
b = 4.9
O
c = 7.3 N
a = 2.5
b = 4.9
O
O
N
N
a = 2.5
N b = 4.3
c = 6.8
Fig. 4 Triangular sizes represented by the indole-2-carboxylic amide
nucleus. Distances (Å) were measured in the structures optimized by
quantum mechanics calculations. Hydrogen atoms are not shown for
clarity.
b = 4.2 to 6.9 Å and c = 3.4 to 10.3 Å) governed by the entire
conformers of neurotensin(8–13), as calculated from the pep-
tide,1 one could conclude that the multiple template approach is
wrong. In this case mimics-1 and -2 would mimic conformers
that are not available to neurotensin(8–13) and yet were
experimentally found to be moderately active. As apparent in
Fig. 4, we found that none of the triangular sizes is beyond the
entire conformational range of neurotensin(8–13). Therefore,
new calculations of the template size and the binding data of
mimics-1 to -3 support the theory of the multiple template
approach. With the data obtained on these mimics, we have
experimentally demonstrated that in the multiple template
approach some templates (i.e. the indole template) can lead to
active mimics and others (i.e. the quinoline template) cannot,
although finding the indole template for active mimics-1 and -2
turned out to be serendipitous.
5-Hydroxypentyl 4,8-bis(5-hydroxypentyloxy)quinoline-2-carb-
oxylate 3
To a solution of intermediate 2 (400 mg, 0.48 mmol) in 5 ml
of THF was added TBAF (1  in THF; 2.2 ml) at room
temp. under N2. The reaction was quenched with water after
2 h stirring at room temp. The ethyl acetate extract was
washed with water and then with brine, dried (Na2SO4) and
concentrated. Flash chromatography on silica gel eluting with
ethyl acetate–hexane (85:15) yielded intermediate 3 (170 mg,
76%) as a colourless oil: νmax(neat)/cmϪ1 3373br, 1720, 1269
and 1068; δH (300 MHz; CDCl3) 1.55–1.80 (12 H, m, 2-
H2 ϩ 4-H2), 1.85–2.14 (6 H, m, 3-H2), 2.55 (3 H, s, OH),
3.58–3.75 (6 H, m, 5-H2O), 4.19 (2 H, t, J 6.1, 1-H2OAr),
4.25 (2 H, t, J 7.5, 1-H2OAr), 4.43 (2 H, t, J 6.3, CO2CH2),
7.06 (1 H, d, J 7.4, 5-ArH), 7.47 (1 H, t, J 8.0, 6-ArH), 7.55
(1 H, s, 3-ArH) and 7.76 (1 H, d, J 8.4, 7-ArH); δC(75.46
MHz; CDCl3) 22.3, 22.4, 28.1, 28.4, 28.6, 32.2, 62.2, 62.4,
66.0, 68.7, 69.0, 101.1, 109.7, 113.1, 123.4, 127.8, 140.4, 147.9,
155.3, 162.5 and 165.9; m/z (ESI) 464 [M ϩ H]ϩ {Found
(HRMS): m/z 464.2649. Calc. [M ϩ H]ϩ for C25H38NO7:
464.2648}.
Conclusion
We have found that the literature procedure for alkylation at the
3-position of ethyl indole-2-carboxylate employing K2CO3 as
base in CH3CN is in error, and that the related alkylated struc-
tures of mimics-1 and -2 have been mistakenly reported. The
newly synthesized nonpeptidic mimic-3, which was designed on
the basis of the incorrect structure of mimic-2, was therefore
less active in binding than mimic-2. With the results of mimic-3
and mimics-1 and -2 with the correct structures, which accord-
ingly served as a blind test for the used approach, we experi-
mentally validated the concept of the multiple template
approach. Further studies are required to demonstrate the prac-
ticality of the multiple template approach and will be reported
in due course.
Experimental
Tetrahydrofuran and diethyl ether were distilled from sodium
benzophenone ketyl prior to use. DMSO was dried with CaH2.
Methylene chloride was distilled from P2O5 prior to use. Sol-
vents used for chromatography were purchased in 5 gallon
drums, redistilled from an all-glass apparatus and stored in
glass bottles. Hexanes refers to light petroleum with bp 64–
69 ЊC. Silica gel 60 (Merck, 230–400 mesh ASTM for flash
chromatography) was used for column chromatography. TLC
was performed on Merck silica gel 60F-254 (0.25 mm, pre-
coated on glass). Other reagents were used as supplied by the
Aldrich Chemical Co. and Lancaster Synthesis Inc. NMR spec-
tra were taken on a Bruker AC-300 instrument. Chemical shifts
are reported in δ units with reference to Me4Si (δ = 0.00 ppm)
5-Azidopentyl 4,8-bis(5-azidopentyloxy)quinoline-2-carboxylate
4
To a solution of intermediate 3 (800 mg, 1.72 mmol) and
triphenylphosphine (1.6 g, 6.1 mmol) in CH2Cl2 (20 ml) was
added HN3 (1.6  in CH2Cl2; 5 ml) at 0 ЊC under argon.
Diethylazodicarboxylate (DEAD) (1 ml, 6.4 mmol) was then
introduced dropwise to the mixture with stirring at the same
temperature. The reaction was quenched with H2O after 1.5 h
stirring at 0 ЊC. The CH2Cl2 extract was washed with brine,
dried (Na2SO4) and concentrated. Flash chromatography on
silica gel eluting with ethyl acetate–hexanes (3:7) gave inter-
mediate 4 (754 mg, 81%) as a light yellow oil; νmax(neat)/cmϪ1
2099, 1800, 1723, 1265 and 1078; δH (300 MHz; CDCl3) 1.55–
1.78 (12 H, m, 2-H2 ϩ 4-H2), 1.85–2.15 (6 H, m, 3-H2), 3.25–
3.35 (6 H, m, 5-H2N3), 4.20 (2 H, t, J 8.4, 1-H2OAr), 4.27 (2 H,
t, J 6.3, 1-H2OAr), 4.43 (2 H, t, J 6.9, CO2CH2), 7.07 (1 H, d, J
7.4, 5-ArH), 7.48 (1 H, t, J 8.1, 6-ArH), 7.56 (1 H, s, 3-ArH)
and 7.77 (1 H, d, J 8.7, 7-ArH); δC(75.46 MHz; CDCl3) 23.2,
23.3, 23.4, 28.1, 28.4, 28.5, 28.53, 28.6, 51.2, 65.6, 68.4, 68.7,
101.0, 109.7, 113.2, 123.4, 127.8, 140.5, 147.9, 155.3, 162.3 and
165.9; m/z (ESI) 539 [M ϩ H]ϩ {Found (HRMS): m/z 539.2846.
Calc. [M ϩ H]ϩ for C25H35N10O4: 539.2843}.
1
for H spectra or CDCl3 (δ = 77.00 ppm) for 13C spectra as
internal standards. J Values are recorded in Hz. Ar, φ and Fn
denote quinoline, phenyl and fluorene rings, respectively. Mass
spectra were obtained on a FINNIGAN MAT-900 instrument.
High resolution mass data were collected by employing EI at 70
eV with PFK reference or by using ESI with a reference
material of PEG 400. Melting points were determined in
open capillary tubes on a Gallenkamp capillary melting point
apparatus and are uncorrected.
2086
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
4,8-Bis(5-azidopentyloxy)quinoline-2-carboxylic acid 5
1-H2OAr), 4.71 (1 H, q, J 7.0, COCH), 6.64 (2 H, d, J 8.1,
To a solution of intermediate 4 (700 mg, 1.3 mmol) in 6 ml of
MeOH and 4 ml of THF was added aqueous KOH (2 ; 6 ml)
at room temp. The mixture was stirred at room temp. for 1 h
and then acidified to pH 3 with 20% aqueous HCl. The ethyl
acetate extract was washed with H2O and then with brine, dried
(Na2SO4) and concentrated to give intermediate 5 (517 mg,
93%) as a yellow oil without further purification; νmax(neat)/
cmϪ1 3533, 2097, 1765, 1728, 1267 and 1076; δH (300 MHz;
CDCl3) 1.60–1.82 (8 H, m, 2-H2 ϩ 4-H2), 1.98–2.12 (4 H, m,
3-H2), 3.36 (4 H, t, J 6.2, 5-H2N3), 4.25 (2 H, t, J 6.4, 1-H2OAr),
4.39 (2 H, t, J 6.9, 1-H2OAr), 7.20 (1 H, d, J 7.2, 5-ArH), 7.60
(1 H, t, J 7.8, 6-ArH), 7.74 (1 H, s, 3-ArH), 7.84 (1 H, d, J 8.1,
7-ArH) and 9.56 (1 H, s, CO2H); δC(75.46 MHz; CDCl3) 23.1,
23.2, 28.1, 28.4, 28.5, 51.1, 69.0, 69.8, 99.8, 111.3, 113.7, 122.9,
128.5, 134.9, 147.9, 152.3, 162.7 and 165.5; m/z (ESI) 428
[M ϩ H]ϩ {Found (HRMS): m/z 428.2050. Calc. [M ϩ H]ϩ for
C20H26N7O4: 428.2046}.
3,5-φH), 7.04 (2 H, d, J 8.2, 2,6-φH), 7.30 (1 H, d, J 7.8,
ArH), 7.49–7.58 (2 H, m, ArH), 7.65–7.90 (8 H, m, NϩH3 ϩ
CONHAda ϩ ArH), 8.59 (1 H, d, J 8.4, ArCONH) and 10.07
[1 H, s, 1-ArϩH or PhOH (not likely but not impossible)];
δC(75.46 MHz; [2H6]DMSO) 22.4, 22.5, 26.7, 26.8, 27.8, 28.2,
28.8, 36.0, 38.4, 40.9, 51.0, 54.1, 68.4, 98.5, 111.1, 113.0, 114.9,
122.6, 127.0, 127.7, 130.3, 139.1, 149.2, 154.6, 155.9, 162.2,
163.0 and 169.6; m/z (ESI) 672 [M ϩ H]ϩ {Found (HRMS): m/z
672.4145. Calc. [M ϩ H]ϩ for C39H54N5O5: 672.4125}.
(2S)-N-Adamantyl-3-[(4-tert-butoxy)phenyl]-2-(9-fluorenyl-
methoxycarbonylamino)propanamide 9
A solution of N-Fmoc-tert-butyl--tyrosine (1.4 g, 3.05 mmol),
1-adamantylamine (468 mg, 3.1 mmol), DCC (693 mg, 3.4
mmol) and HOBt (455 mg, 3.4 mmol) in dry DMF (10 ml) was
stirred at room temp. under N2 for 12 h. The mixture was
poured into saturated aqueous NH4Cl and extracted with ethyl
acetate. The extract was washed with H2O and then with brine,
dried (Na2SO4) and concentrated. Flash chromatography on
silica gel eluting with EtOAc–hexanes (3:7) afforded inter-
mediate 9 (1.8 g, 100%) as a white solid; mp 92.1–93.6 ЊC;
νmax(CHCl3)/cmϪ1 3310br, 1694, 1665, 1536, 1257 and 1043;
δH (300 MHz; CDCl3) 1.33 [9 H, s, C(CH3)3], 1.57 [6 H, s,
CH2(Ada)], 1.80 [6 H, s, CH2(Ada)], 2.01 [3 H, s, CH(Ada)],
2.78–2.88 (1 H, m, φCH2R), 3.08–3.15 (1 H, m, φCH2R), 4.10–
4.24 (1 H, m, 9-FnH), 4.22 (1 H, t, J 6.9, COCH), 4.32–4.45 (2
H, m, CO2CH2), 4.93 (1 H, s, CONHAda), 5.54 (1 H, d, J 6.9,
FmocNH), 6.94 (2 H, d, J 6.0, 3,5-φH), 7.13 (2 H, d, J 6.0,
2,6-φH), 7.32 (2 H, t, J 6.5, FnH), 7.41 (2 H, t, J 7.3, FnH), 7.59
(2 H, d, J 7.4, FnH) and 7.76 (2 H, d, J 7.5, FnH); δC(75.46
MHz; CDCl3) 28.7, 29.2, 36.1, 38.9, 41.3, 47.1, 52.0, 56.9, 66.9,
78.3, 119.9, 124.2, 125.0, 127.0, 127.6, 129.9, 131.6, 141.2,
143.7, 154.2, 155.7 and 169.3; m/z (ESI) 593 [M ϩ H]ϩ {Found
(HRMS): m/z 593.3364. Calc. [M ϩ H]ϩ for C38H45N2O4:
593.3379}.
N-[(2S)-1-Adamantylamino-3-(4-tert-butoxyphenyl)-1-oxo-
propan-2-yl]-4,8-bis(5-azidopentyloxy)quinoline-2-carboxamide
6
To a solution of intermediate 5 (200 mg, 0.47 mmol) in
anhydrous DMF (10 ml) were added (2S)-N-adamantyl-2-
amino-3-[(4-tert-butoxy)]propanamide 10 (210 mg, 0.52 mmol),
DCC (116 mg, 0.56) and HOBt (80 mg, 0.56). The mixture was
quenched with cold H2O after 12 h stirring at 40 ЊC. The ethyl
acetate extract was washed with H2O and then with brine, dried
(Na2SO4) and concentrated. Flash chromatography on silica gel
eluting with ethyl acetate–hexanes (4:6) gave intermediate 6
(329 mg, 90%) as a light yellow oil; νmax(neat)/cmϪ1 3337, 2098,
1660, 1504 and 1080; δH (300 MHz; CDCl3) 1.31 [9 H, s,
C(CH3)3], 1.55–2.15 (27 H, m, AdaH ϩ 2-H2 ϩ 3-H2 ϩ 4-H2),
3.05 (1 H, dd, J 9.1 and 13.3, φCH2R), 3.23 (1 H, dd, J 5.4, and
13.4, φCH2R), 3.34 (2 H, t, J 6.6, 5-H2N3), 3.42 (2 H, t, J 6.4,
5-H2N3), 4.22 (2 H, t, J 6.2, 1-H2OAr), 4.29 (2 H, t, J 6.3,
1-H2OAr), 4.56–4.65 (1 H, m, COCH), 5.21 (1 H, s, CON-
HAda), 6.94 (2 H, d, J 8.3, 3,5-φH), 7.08 (1 H, d, J 8.6, 5-ArH),
7.25 (2 H, d, J 8.3, 2,6-φH), 7.46 (1 H, t, J 8.0, 6-ArH), 7.65 (1
H, s, 3-ArH), 7.77 (1 H, d, J 8.3, 7-ArH) and 8.97 (1 H, d, J 7.9,
ArCONH); δC(75.46 MHz; CDCl3) 23.3, 23.5, 28.3, 28.5, 28.7,
28.8, 29.2, 36.1, 38.4, 41.3, 51.1, 51.3, 51.8, 55.9, 68.4, 68.7,
78.1, 98.6, 110.0, 113.4, 123.3, 124.2, 127.1, 129.9, 132.1, 139.7,
149.2, 154.1, 154.9, 162.6, 164.5 and 168.8; m/z (ESI) 781
[M ϩ H]ϩ {Found (HRMS): m/z 780.4583. Calc. [M ϩ H]ϩ for
C43H58N9O5: 780.4561}.
(2S)-N-Adamantyl-2-amino-3-[(4-tert-butoxy)phenyl]propan-
amide 10
To a solution of intermediate 9 (1.8 g, 3.04 mmol) in THF (15
ml) was added piperidine (1 ml). The solvent was evaporated
after 2 h stirring at room temp. Flash chromatography on silica
gel eluting first with ethyl acetate–hexanes (1:1) and then with
ethanol–ethyl acetate (15:85) gave intermediate 10 (1.1 g, 97%)
as a white solid; mp 102.5–103.6 ЊC; νmax(neat)/cmϪ1 3334, 1657,
1516, 1166 and 905; δH (300 MHz; CDCl3) 1.34 [9 H, s,
C(CH3)3], 1.5 (2 H, s, NH2), 1.64–1.70 [6 H, m, CH2(Ada)],
1.94–1.99 [6 H, m, CH2(Ada)], 2.04–2.08 [3 H, m, CH(Ada)],
2.71 (1 H, dd, J 8.5 and 13.7, φCH2R), 3.09 (1 H, dd, J 4.6 and
13.7, φCH2R), 3.43 (1 H, dd, J 4.6 and 8.4, COCH), 6.80 (1 H,
s, CONHAda), 6.93 (2 H, d, J 8.4, 3,5-φH) and 7.10 (2 H, d, J
8.3, 2,6-φH); δC(75.46 MHz; CDCl3) 28.8, 29.4, 36.3, 40.4, 41.4,
51.0, 56.8, 78.3, 124.2, 129.7, 132.7, 154.0 and 173.1; m/z (ESI)
371 [M ϩ H]ϩ {Found (HRMS): m/z 371.2703. Calc. [M ϩ H]ϩ
for C23H35N2O2: 371.2699}.
N-[(2S)-1-Adamantylamino-3-(4-hydroxyphenyl)-1-oxopropan-
2-yl-4,8-bis(5-aminopentyloxy)quinoline-2-carboxamide
tris(trifluoroacetate) salt (mimic-3 in its salt form)
To a solution of intermediate 6 (100 mg, 0.13 mmol) in ethyl
acetate (10 ml) and methanol (5 ml) was added 10% Pd–C (20
mg) followed by six drops of concentrated HCl. The mixture
was filtered through Celite after 12 h stirring under a H2 atmos-
phere at room temp. The residue concentrated from the filtrate
was dissolved in 0.1% TFA (2 ml) and 80% acetonitrile in 0.1%
TFA (1 ml), and purified by reversed-phase HPLC on a Vydak
C8 (15–20 µm particle size, 250 mm length × 22 mm id) column
using gradient elution; starting with 90% of buffer A (0.1%
TFA in H2O) and then linearly increasing the concentration of
buffer B (prepared from 20% of buffer A in MeCN) at t = 1 min
to 90% over 30 min, at a flow rate of 8 cm3 minϪ1 with the
detector wavelength set at 220 nm, gave mimic-3 as its tris(tri-
fluoroacetate) salt (>95% pure). The retention time of mimic-3
was 25.94 min. After lyophilization, a yellow powder was iso-
lated; mp 164.4–166.5 ЊC; νmax(neat)/cmϪ1 3422br, 1655, 1508
and 1078; δH (300 MHz; [2H6]DMSO) 1.55–1.75 [14 H, m,
CH2(Ada) ϩ 2-H2 ϩ 4-H2], 1.86–2.0 [10 H, m, CH2(Ada) ϩ
3-H2], 2.02 [3 H, s, CH(Ada)], 2.80–3.01 (6 H, m, 5-H2Nϩ ϩ
φCH2R), 4.21 (2 H, t, J 5.8, 1-H2OAr), 4.32 (2 H, t, J 5.7,
Propyl 4,8-bis(propyloxy)quinoline-2-carboxylate (model-1)
Dry DMSO (20 ml) was added to a mixture of 4,8-
dihydroxyquinoline-2-carboxylic acid 1 (248 mg, 1.2 mmol),
KOH powder (270 mg, 3.96 mmol) and a catalytic amount of
18-crown-6 under N2. The mixture was stirred at room temp. for
1 h and then 1-iodopropane (734 mg, 4.32 mmol) was added.
The reaction was quenched with cold water after 2 h stirring at
room temp. The ethyl acetate extract was washed with water
and then with brine, dried (Na2SO4) and evaporated to dryness.
Flash chromatography on silica gel eluting with ethyl acetate–
hexanes (3:7) gave the product (330 mg, 83%) as a light yellow
solid; mp 65.8–66.4 ЊC; νmax(neat)/cmϪ1 1742, 1713, 1265, 1119
and 1072; δH (300 MHz; CDCl3) 1.05–1.16 (9 H, m, 3-H3), 1.86–
2.13 (6 H, m, 2-H2), 4.15 (2 H, t, J 7.1, OCH2), 4.23 (2 H, t, J
J. Chem. Soc., Perkin Trans. 1, 1997
2087

View Article Online
4 K. S. Kanba, S. Kanba, A. Nelson, H. Okazaki and E. Richelson,
J. Neurochem., 1988, 50, 131.
5 Gaussian94, Gaussian, Inc., Carnegie Office Park, Building Six,
Pittsburgh, PA 15106.
6 Y.-P. Pang, J. Zaidi, B. Cusack and E. Richelson, unpublished
results.
6.5, OCH2), 4.38 (2 H, t, J 6.9, CO2CH2), 7.07 (1 H, d, J 7.8,
5-ArH), 7.47 (1 H, t, J 8.5, 6-ArH), 7.57 (1 H, s, 3-ArH) and
7.79 (1 H, d, J 8.5, 7-ArH); δC(75.46 MHz; CDCl3) 21.8, 22.1,
22.2, 67.3, 70.1, 70.4, 100.9, 109.5, 113.0, 123.4, 127.6, 140.5,
147.9, 155.4, 162.4 and 165.9.
7 P. G. McDougal, J. G. Rico, Y.-I. Oh and B. D. Condon, J. Org.
Chem., 1986, 51, 3388.
8 B. Cusack and E. Richelson, J. Recept. Res., 1993, 13, 123.
9 B. Cusack, T. Stanton and E. Richelson, Eur. J. Pharmacol., 1991,
206, 339.
10 P. J. Munson and D. Rodbard, Anal. Biochem., 1980, 107, 220.
11 J. A. Gilbert, C. J. Moses, M. A. Pfenning and E. Richelson,
Biochem. Pharmacol., 1986, 35, 391.
12 N. M. Gray, M. S. Dappen, B. K. Cheng, A. A. Cordi, J. P.
Biesterfeldt, W. F. Hood and J. B. Monahan, J. Med. Chem., 1991,
34, 1283.
Acknowledgements
Support by the Mayo Foundation for Medical Education and
Research and in part by a United States Public Health Service
Grant MH27692-18A2 (E. Richelson and Y.-P. Pang) is greatly
acknowledged. We thank Dr Abdul H. Fauq of the Mayo
Foundation for helpful discussions and L. M. Benson of the
Mayo Foundation for the mass spectral analysis.
13 D. S. Dodd, A. P. Kozikowski, B. Cusack and E. Richelson, Bioorg.
Med. Chem. Lett., 1994, 4, 1241.
References
1 Y. P. Pang, J. Zaidi, A. P. Kozikowski, B. Cusack and E. Richelson,
J. Comput.-Aided Mol. Des., 1994, 8, 433.
2 B. Cusack, E. Richelson, Y. P. Pang, J. Zaidi and A. P. Kozikowski,
Mol. Pharmacol., 1993, 44, 1036.
3 A. P. Kozikowski, D. S. Dodd, J. Zaidi, Y. P. Pang, B. Cusack and
E. Richelson, J. Chem. Soc., Perkin Trans. 1, 1995, 21, 1615.
Paper 7/00894E
Received 7th February 1997
Accepted 11th March 1997
2088
J. Chem. Soc., Perkin Trans. 1, 1997