Organic Letters
Letter
Trimeric turn 1 was first elongated using terephthalic acid
monomers to form pentamer 2. In order to substantially
enhance face-to-face intramolecular π−π stacking, it was
anticipated that larger units than simple benzene derivatives
would be necessary. For this purpose, we introduced a new
monomer, 6-amino-4-isobutoxy-7-methoxy-2-quinolinecarbox-
ylic acid. As described in the Supporting Information, its
synthesis follows routes developed for other amino-2-quinoline-
carboxylic acid monomers.5 The main feature of the new
monomer is that the relative orientation of its 6-amine and 2-
acid functions will confer a linear structure to oligoamide
segments in which it is inserted.6 Elongation of pentamer 2
proceeded via the saponification of its terminal esters and the
subsequent coupling of two quinoline monomers or two
quinoline dimers to yield 3 and 4, respectively.
Figure 2. Schematic representation of 1H−1H correlations observed in
400 MHz ROESY-2D spectra of nonamer 4 at 25 °C in CDCl3. Blue
arrows represent correlations within the hairpin turn; pink arrows
represent interstrand correlations; other correlations are represented
by gray arrows.
The chemical shift values of the Hint protons was monitored
in CDCl3 as a probe of the conformation of the new oligomers
in this solvent (Figure 1), as they were found to undergo
same aromatic ring such as, for example, 7-methoxy protons
and the proton in position 5 of equivalent quinoline rings
(OMe/H7 and OMe/H10 correlations in Figure 2). In contrast,
no such correlations were observed in the spectra of heptamer
3. Thus, the greater flexibility of turn 1 as compared to turn 5
can be compensated by extensive intramolecular π−π
interactions between long linear strands. All observed
correlations, and in particular OMe/H7 and OMe/H10 (Figure
2), are consistent with an antiparallel orientation of the two
strands of 4, with each ring head-to-tail with respect to the ring
on which its stacks, as observed previously with multistranded
sheets having shorter linear segments.3 Yet the existence of
conformers with a parallel orientation of the two strands of 4
cannot be completely ruled out from solution studies. Indeed,
the interstrand NOE correlations expected in a parallel
arrangement may not be distinguished from intrastrand
correlations, unless the two strands are strongly offset in
which case interstrand correlations may match with those
expected in an antiparallel arrangement of the two strands.
Spectra recorded over a wide temperature range (−50 to 40
°C) did not reveal any major change (Figure S2), suggesting
that the same structure prevails over this range. Even at low
temperature, the diastereotopicity of side chain isobutoxy
protons does not appear in the spectrum, indicating the fast
rotational dynamics of each linear strand with respect to the
other. As a final point worth noting, it is unclear whether the
Figure 1. Part of the 300 MHz 1H NMR spectra in CDCl3 at 25 °C of
(a) hairpin turn trimer 1; (b) hairpin turn pentamer 2; (c) hairpin turn
heptamer 3; (d) hairpin turn nonamer 4; and (e) hairpin turn trimer 5.
Signals of Hext are marked with white circles. Signals of Hint are marked
with full black circles. Stars indicate signals belonging to an impurity.
significant variations while δHext remained essentially un-
changed. In rigid turn 5, this resonance is strongly upfield
shifted (δHint = 4.91 ppm), while the absence of a methyl group
in 1 makes its structure more flexible and significantly reduces
intramolecular ring current effects (δHint = 6.68 ppm). Upon
elongation of trimer 1 into pentamer 2 and heptamer 3, no
major change is observed; Hint resonances actually undergo
minor downfield shifts. In contrast, a strong upfield shift is
observed between heptamer 3 (δHint = 6.80 ppm) and nonamer
4 (δHint = 6.08 ppm, Δδ = 0.72 ppm). The effect on δHint of the
additional quinoline monomers at the end of the linear strands
from 3 to 4 is very remote and suggests a better defined
conformation of 4 leading to stronger ring current effects upon
Hint. The nonlinear trend of δHint values from 1 to 4 is indicative
of cooperative effects.7
A series of multidimensional NMR experiments (HSQC,
HMBC, TOCSY, and ROESY) allowed us to fully assign the
1H NMR spectrum of 4. As shown in Figure 2, ROESY
experiments revealed contacts that unambiguously establish a
folded two-stranded hairpin structure of 4 in solution. In
particular interstrand (as opposed to intrastrand) correlations
are demonstrated when they occur between protons that are
too distant to establish a contact if they would belong to the
H
int resonance in the folded conformation of 4 remains at lower
field compared to the case of 5 because the latter is still better
organized, because the former adopts a conformation at the
turn in which ring current effects are weaker, or because ring
currents are simply less intense in the former due to the
different substitution patterns of its para-phenylenediamine
rings.
Solid state investigations were undertaken and fully
confirmed solution data. Pentamer 2 was crystallized with
two independent molecules in the asymmetric unit with both
showing divergent linear segments with no apparent stacking
interactions between the para-phenylenediamine units or the
terminal terephthalic acid units (Figure 3a, 3b). In contrast, the
crystal structure of 4 (Figure 3c, 3d) revealed a folded structure
with extensive face-to-face π−π overlap between its two linear
segments. The dinitrobenzene turn is perpendicular to the
linear strands, which are found to be in an antiparallel
orientation, presumably favored by local dipole−dipole
interactions. Consistent with solution data, quinoline 7-
methoxy protons are found to be much closer (3.7 Å) to
protons in position 5 of the quinoline stacked above them than
2327
dx.doi.org/10.1021/ol500512f | Org. Lett. 2014, 16, 2326−2329