Synthesis of N-benzylated-2-aminoquinolines as ligands for the Tec SH3 domain

Article abstract of DOI:10.1016/j.bmcl.2005.09.073

In recent work, we have been developing 2-aminoquinolines as ligands for Src Homology 3 (SH3) domains, so far the only reported examples of small-molecule ligands for these domains. In this paper, we report the synthesis of a series of N-benzylated-2-amin

Full text of DOI:10.1016/j.bmcl.2005.09.073

Bioorganic & Medicinal Chemistry Letters 16 (2006) 387–390
Synthesis of N-benzylated-2-aminoquinolines as ligands for the
Tec SH3 domain
a,b
a
b
a,
*
Steven R. Inglis, Rhiannon K. Jones, Grant W. Booker and Simon M. Pyke
a
Discipline of Chemistry, School of Chemistry and Physics, The University of Adelaide, SA 5005, Australia
Discipline of Biochemistry, School of Molecular and Biomedical Science, The University of Adelaide, SA 5005, Australia
b
Received 7 April 2005; revised 23 September 2005; accepted 27 September 2005
Available online 2 November 2005
Abstract—In recent work, we have been developing 2-aminoquinolines as ligands for Src Homology 3 (SH3) domains, so far the
only reported examples of small-molecule ligands for these domains. In this paper, we report the synthesis of a series of N-benzy-
lated-2-aminoquinolines by reductive amination of aryl aldehydes with 2-aminoquinoline. These ligands bound the SH3 domain
with ca. one and a half to twofold reduced affinity relative to 2-aminoquinoline; however, some evidence was found to suggest that
the benzylic substituents made new contacts with the SH3 domain surface. These results provide useful SAR information that may
assist in future ligand design.
Ó 2005 Elsevier Ltd. All rights reserved.
Src Homology 3 (SH3) domains are small, non-catalytic
protein–protein interaction domains that bind to pro-
line-rich peptides and feature within a range of impor-
tant cell signaling events and other biological processes
A
D196
R1
O
N211
O
T192
HO Me
Q190
O
D212
H
H
2
H N
O
N
N
R2
OH
Me
O
1
2
(
see, Mayer or Zarrinpar for reviews). These domains
O
N
L213
W215
HN
have long been targets for the development of potential
therapeutics, however until recently, no examples of
entirely non-peptide ligands for the domain had been
reported. Previously, we have reported that 2-amino-
quinolines bind the Tec SH3 domain with weak to
NH
N
Me
N
H
2
H
H214
F189
LEFT
RIGHT
B
R1
R2
3
compds
Kd (µM)
moderate affinity. The mechanism of the binding of 2-
aminoquinoline 1 has been well studied and involves.
p–p stacking between the quinoline ring and tryptophan
1
2
3
125
380
H
Me
H
H
CH C H -p-OMe
2 2 3
CH O(CH ) OH 81
2
6
4
2
15 (W215) and a salt bridge between the protonated li-
gand and aspartate 196 (D196), both highly conserved
residues in the proline-rich peptide binding site of the
Tec SH3 domain (Fig. 1A).
rotamer A
rotamer B
C
rotation
R
H
N
H
N
H
N
H
N
R
The highest affinity 2-aminoquinolines identified so far
are all 6-substituted ligands with bulky hydrophobic
substituents that are thought to make contacts with
regions on the ÔrightÕ side of the ligand binding site as
illustrated in Figure 1A. However, currently little infor-
mation is known about the types of ligand/protein con-
tacts that may be formed between ÔleftÕ surface residues,
O
O
O
O
D196
D196
3
Figure 1. (A) Model for mechanism of binding of 2-aminoquinoline to
the Tec SH3 domain, and regions adjacent to both the ÔleftÕ and ÔrightÕ
sides of the binding site, predicted to make contacts with substituents
attached to the 2-aminoquinoline nucleus. (B) Summary of some
ligand binding results reported previously. [Ligands 1 and 2, see Ref. 3;
ligand 3; see Ref. 4]. (C) Chemical considerations that assist in the
explanation for why N-substituted-2-aminoquinolines bind the SH3
domain with reduced affinity.
Keywords: Src Homology 3 domain; 2-Aminoquinoline; Ligand.
*
Corresponding author. Tel.: +61 8 8303 5358; fax: +61 8 8303
358; e-mail: simon.pyke@adelaide.edu.au
4
0
960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.073

388
S. R. Inglis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 387–390
and ligand-substituents placed at other parts of the 2-
aminoquinoline platform, for example the amino group
as illustrated in Figure 1A. Substitution on the amino
10 was isolated in 10% yield and its complete charac-
terization was possible.
nitrogen atom of 2-aminoquinoline with a methyl group
3
The formation of this structure most likely occurs due to
nucleophilic attack by cyanide (from sodium cyano-
borohydride) on the imine intermediate 11, leading to
the formation of the new cycle. The formation of the
structure was accompanied by a very large downfield
(
ligand 2) resulted in a ca. threefold reduction in affinity
Fig. 1B). This may be explained by considering the
(
rotamers that exist around the HRN–C(1) bond, and
the formation of the salt bridge involved in the binding
of 2-aminoquinolines to the SH3 domain (Fig. 1C). In 1,
1
change in the H NMR chemical shift for H8 of the
(
when R = H), both rotamers can be involved in the for-
quinoline ring in 2-aminoquinoline 1 from ca. 7.8 to
9.1 ppm in 10 (H9 in 10). This is thought to be a result
of an anisotropic deshielding effect provided by the par-
mation of an ÔidealÕ salt bridge with D196. But when
R = Me (as in 2), only one rotamer interacts favorably
with D196. Thus, there is an entropic cost associated
with the binding of 2 to the SH3 domain, leading to
lower affinity. However, as part of our synthetic investi-
gations into new methods for the preparation of
2
tial sp character of the N–C bond at the amino group.
The assignment of the signal at 9.1 ppm to H9 was
confirmed by the presence of a cross-peak between the
1
1
1-amino protons and H9 in the [ H, H] ROESY NMR
spectrum.
2
enzyl)aminoquinolines were prepared and one of these,
-aminoquinolines, some 6-substituted-2-(4-methoxyb-
compound 3, was tested for binding to the Tec SH3 do-
4
main. The affinity of this ligand (Fig. 1B) was sugges-
Only one example of this heterocyclic skeleton has been
6
previously reported, however, no spectral data were
tive that the penalty for substitution on the amino
nitrogen atom was less severe when the substituent
was a 4-methoxybenzyl group, than when the substitu-
ent was a methyl group alone. Therefore, this prompted
us to further investigate the influence of N-benzylation
of 2-aminoquinoline on its SH3 binding affinity, and this
is the focus of the present study.
provided to support its structure. Thus, in order to
further validate the formation of the structure, 10 was
explicitly synthesized by an independent method. Initial-
ly, it was attempted to prepare 10 by adapting the one-
pot titanium tetraisopropoxide assisted method, but
instead using sodium cyanide rather than sodium cyano-
borohydride. However, a complex mixture of products
was isolated after workup in this case. Instead, as illus-
trated in Scheme 2, the imine intermediate 11 was
prepared first by heating the aldehyde with 2-amino-
quinoline in toluene at reflux. This imine was then treat-
ed with sodium cyanide in methanol at ca. 50 °C to form
10 in 33% overall yield from 3-fluorobenzaldehyde. A
small amount (4%) of a second imine 12 was also isolat-
ed following chromatography with silica gel, indicating
that some hydrolysis of imine 11 occurred during the
second step, and the reformed aldehyde was then
condensed with the amine 10 to form 12.
A one-pot method for the synthesis of N-benzylated-2-
aminoquinolines from aryl aldehydes and 2-aminoquin-
5
oline was sought. Thus, the method of Mattson
involving titanium tetraisopropoxide assisted reductive
alkylation of amines was adapted for use with 2-ami-
noquinoline and aryl aldehydes as illustrated in Scheme
1
. By this method, the appropriate aldehyde was stirred
with 2-aminoquinoline in titanium tetraisopropoxide
for ca. 1 h, prior to the addition of sodium cyanoboro-
hydride in ethanol. In cases where the starting
aldehydes were solids, additional titanium tetraisoprop-
oxide and/or THF were added to the mixture to assist
in stirring. Using this approach, the N-benzylated
derivatives 4–9 were prepared in low to moderate yields
As has been previously demonstrated in the synthesis
4
of 6-substituted-2-(4-methoxybenzyl)aminoquinolines,
N-benzylated-2-aminoquinolines
could
also
be
(
3–46%) (Scheme 1). The poor yields were in part
synthesized from 2-chloroquinolines by treatment with
the appropriate benzylamine. Thus, in order to
demonstrate an alternative and much improved method
for the synthesis of the desired N-benzylated-
attributed to the formation of a by-product each time,
the imidazo[1,2-a]quinolin-1-ylamine derivatives (see
the 3-fluorophenyl derivative 10 in Scheme 1). These
by-products were generally not isolated in a pure form,
but in the case of the 3-fluorophenyl derivative 10, pure
a
N
N
N
NH
2
a
1
F
Ar
+
11
N
NH
2
N
N
N
N
H
1
b
4
4
5
6
7
8
Ar = Ph, R = H
Ar = Ph, R = 2-F
Ar = Ph, R = 3-F
Ar = Ph, R = 3-OH
Ar = Ph, R = 4-OH
H N
Ar
2
5
3
+
F
R
N
N
N
N
1
0
Ar = Ph,
R = 3-F
Ar =
6
2
1
N
2
H N
F
F
9
Ar =
NH
1
2
10
i
Scheme 1. Reagents and conditions: (a)—(1) ArCHO, Ti( OPr)
THF), rt. (2) NaBH CN, EtOH. (Experimental details are provided
as Supplementary data.)
4
,
Scheme 2. Reagents and conditions: (a) 3-Fluorobenzaldehyde/
toluene/D; (b) NaCN/MeOH/ca. 50 °C. (Experimental details are
provided as Supplementary data.)
(
3

S. R. Inglis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 387–390
389
2
-aminoquinolines in the current study, the derivative 4
The SAR information provided above confirms that the
larger benzyl substituents of ligands 4–9 are better tol-
erated than the methyl substituent of ligand 2, suggest-
ing that a new lipophilic contact is formed that mediates
the improvement in affinity. This lipophilic contact is
able to partially offset the entropic costs that are likely
to be associated with the ca. threefold reduction in affin-
ity due to N-methylation of 2-aminoquinoline (dis-
cussed above and see Fig. 1c). Of the N-benzylated
ligands, the variation in the affinities may be explained
by the different substituents on the phenyl ring. The
small inductively withdrawing fluoro substituents of li-
gands 5 and 6 have little influence on the affinity, as
does the 3-hydroxy substituent of 7, suggesting that
substitution at these positions with simple groups is well
tolerated. The differences in the electronic character of
the phenyl rings of ligands 5, 6, and 7 do not appear
to significantly alter the affinities. However, substitution
with a hydroxyl group at the 4-position as in 8 leads to
a small reduction in affinity, suggesting that substitution
at this position is less well tolerated, possibly due to
either steric or electronic penalties. The reduced affinity
of ligand 9 that has a 5-membered electron-rich aryl
group may be a result of the smaller size of the ring,
and hence a reduced lipophilic contact area.
was also prepared in 97% yield by treatment of 2-chloro-
quinoline 13 with benzylamine at ca. 140 °C for 30 h as
illustrated in Scheme 3.
The N-benzylated-2-aminoquinolines 4–9 were all tested
for binding to the Tec SH3 domain using NMR chemi-
1
15
cal shift perturbation, with [ H, N] HSQC experiments
1
using N labeled Tec SH3 protein, to obtain K values.
5
d
The use of this method in the current ligand develop-
3
ment studies has been documented. Compounds 4–9
all bound the Tec SH3 domain with approximately
one and a half to twofold reduced affinity relative to
2
-aminoquinoline 1 (K s ca. 180–300 lM for 4–9,
d
K = 125 lM for 1, Table 1) (see for 4 and 7 in
Fig. 2). However, all of these ligands bound with im-
proved affinity relative to the N-methylated-2-amino-
quinoline derivative 2 (K = 380 lM) (Table 1). In the
case of ligands 4 and 6, the improvement in affinity
was approximately twofold relative to 2.
d
d
Ph
a
N
N
N
Cl
H
1
3
4
Scheme 3. Reagents and conditions: (a) benzylamine/ca. 140 °C.
The above results also suggest that the strongly electro-
negative fluorine atoms of ligands 5 and 6 do not have a
significant impact on the affinity of the ligands. It may
also be concluded that the hydroxyl groups on the phen-
yl rings of ligands 7 and 8 are not involved in ÔidealÕ
hydrogen bonding, as significantly greater improve-
ments in affinity would have been expected if this were
the case.
d
Table 1. Equilibrium binding dissociation constants (K ) for N-
benzylated-2-aminoquinoline derivatives with the Tec SH3 domain,
determined by NMR chemical shift perturbation experiments, and
d
comparison with K s from previous studies
a
Compound
d
K (lM)
4
5
193(±15)
208(±28)
177(±34)
234(±55)
292(±85)
285(±46)
380(±40)
125(±24)
1
Chemical shift mapping of all residues whose H (H–N)
6
resonances were altered by at least 0.1 ppm at the max-
imum concentration of ligand 4 revealed a similar bind-
ing footprint to 2-aminoquinoline 1, confirming that
both ligands bind at the same location of the SH3 do-
main. Furthermore, additional changes in chemical shift
were observed for residues alanine 191 (A191, upfield)
and glutamine 190 (Q190, downfield), ÔleftÕ of the 2-ami-
noquinoline binding site, attributed to anisotropic
shielding and deshielding effects, respectively, provided
by the phenyl ring of 4 (see Supplementary data). This
provides additional evidence for the positioning of the
phenyl ring and suggests that it indeed interacts with
ÔleftÕ surface residues. An obvious candidate residue in
the region likely to mediate the interaction is the phenyl
ring of phenylalanine (F189) (Fig. 1A), which may form
a p–p stack with the phenyl ring of the benzyl substitu-
ent of the ligand. Implicating F189 in the interaction
also assists in the explanation for why the penalty for
substituting the amino nitrogen atom with a methyl
group (as in ligand 2) is greater than when bulkier sub-
stituents are present (as in ligands 4–9). The lipophilic
contact area of 2 is considerably smaller, resulting in a
very small, if any, lipophilic interaction.
7
8
9
b
b
2
1
a
Quoted values are means ± standard deviation over residues whose
1
H (H–N) chemical shifts were altered by at least 0.1 ppm at the
maximum concentration of ligand.
b
3
As reported in Inglis et al.
Figure 2. Equilibrium binding of N-benzylated-2-aminoquinolines 4
and 7, and comparison with 2-aminoquinoline 1 and the N-methylated
3
derivative 2 (from previous studies) to the Tec SH3 domain as studied
In summary, a small series of N-benzylated-2-aminoquin-
5
by NMR spectroscopy. Overlays of isotherms obtained from indepen-
1
dent experiments, represented by the change in H (H–N) chemical for
olines was synthesized by adapting a reported method
for the titanium tetraisopropoxide assisted reductive
the indole H–N proton of tryptophan 215 (W215e1).

390
S. R. Inglis et al. / Bioorg. Med. Chem. Lett. 16 (2006) 387–390
alkylation of amines with carbonyl compounds. This
method poorly suited the synthesis of the desired amines
presented here, because of low yields. However, the low
yields were partly accounted for by the formation of the
corresponding imidazo[1,2-a]quinolin-1-ylamine by-
products, an uncommon heterocyclic scaffold, in yields
of ca. 10% each time the reaction was performed. N-Ben-
zylated-2-aminoquinolines can instead be prepared in
much improved yields by synthesis from 2-chloroquino-
lines by treatment with benzylamines, as was demonstrat-
ed by the alternative synthesis of derivative 4.
primary amino group, involved in the important salt
bridge with D196, could be retained.
Supplementary material
The experimental details of synthetic procedures pre-
sented in this paper together with chemical shift map-
ping of ligands 1 and 4 onto the SH3 fold are
provided as Supplementary data. Supplementary data
associated with this article can be found, in the online
version, at doi:10.1016/j.bmcl.2005.09.073.
Although all the compounds 4–9 bound the Tec SH3 do-
main with ca. one and a half to twofold reduced affinity
relative to 2-aminoquinoline 1, they all bound with im-
proved affinity relative to the N-methylated-2-amino-
quinoline compound 2. This suggests that the bulkier
benzyl substituents make new lipophilic contacts, but
the overall affinity of the ligand is lower than 2-amino-
quinoline 1, probably as a result of substitution on the
amino group. This provides useful SAR information
and may assist in future ligand design. Particularly, it
would be of interest to prepare 2-aminoquinolines with
bulky lipophilic substituents that may target the ÔleftÕ
surface residues in a similar fashion to the N-benzylated
ligands, but instead attach the substituent at other
points of the 2-aminoquinoline nucleus, for example
the 3- or 4-position of the quinoline ring. Thereby, the
References and notes
1. Mayer, B. J. J. Cell Sci. 2001, 114, 1253.
2
3
4
. Zarrinpar, A.; Bhattacharyya Roby, P.; Lim Wendell, A.
ScienceÕs STKE [online] 2003, www.stke.org/cgi/content/
full/sigtrans; 2003/179/re8.
. Inglis, S. R.; Stojkoski, C.; Branson, K. M.; Cawthray, J.
F.; Fritz, D.; Wiadrowski, E.; Pyke, S. M.; Booker, G. W.
J. Med. Chem. 2004, 47, 5405.
. Inglis, S. R.; Jones, R. K.; Fritz, D.; Stojkoski, C.; Booker,
G. W.; Pyke, S. M. Org. Biomol. Chem. 2005, 3, 2543.
5. Mattson, R. J.; Pham, K. M.; Leuck, D. J.; Cowen, K. A. J.
Org. Chem. 1990, 55, 2552.
6. Lee, C. S.; Hashimoto, Y.; Shudo, K.; Nagao, M. Hetero-
cycles 1984, 22, 2249.