Abiotic metallofoldamers as electrochemically responsive molecules

Article abstract of DOI:10.1021/ja050886c

Described are the design, synthesis, and study of nonbiological molecules based on salophen and salen ligands that fold into single-stranded helices in the presence of either Ni(II) or Cu(II). X-ray diffraction studies show that the materials fold into he

Full text of DOI:10.1021/ja050886c

Published on Web 07/12/2005
Abiotic Metallofoldamers as Electrochemically Responsive
Molecules
Fan Zhang, Shi Bai,† Glenn P. A. Yap,† Vinod Tarwade, and Joseph M. Fox*
Contribution from the Brown Laboratories, Department of Chemistry and Biochemistry,
UniVersity of Delaware, Newark, Delaware 19716
Received February 10, 2005; E-mail: jmfox@udel.edu
Abstract: Described are the design, synthesis, and study of nonbiological molecules based on salophen
and salen ligands that fold into single-stranded helices in the presence of either Ni(II) or Cu(II). X-ray
diffraction studies show that the materials fold into helical structures in the solid state, and a series of NMR
studies provide strong evidence that the folded structures are conserved in solution. Metal coordination is
required for folding, as NMR and X-ray show that the free ligands do not adopt helical structures. Two of
the racemic metallofoldamers spontaneously resolve during crystallization from CHCl3/acetonitrile, and CD
spectroscopy and optical rotation show that the resolved, crystalline materials racemize quickly when
dissolved at 5 °C. This shows that the secondary structures can reorganize easily and can, therefore,
provide the basis for responsive materials. By comparison, an analogue from enantiomerically pure (R,R)-
(-)-trans-cyclohexanediamine showed a strong CD signal and a large specific rotation. Electrochemical
experiments show that a structural reorganization occurs upon metal-centered reduction of a Cu(II)-containing
foldamer. When the reduction is carried out in the presence of coordinating ligands, it is proposed that
apical binding of those ligands gives square pyramidal complexes. Semiempirical (AM1) calculations support
that the helical structure would be disrupted by the reduction to Cu(I) with concomitant reorganization to a
square pyramidal complex.
Scheme 1
Introduction
In this contribution, we describe the synthesis and charac-
terization of single-stranded abiotic foldamers1-3 that adopt
helical conformations in response to metal complexation. These
metallofoldamers are derivatives of salen and salophen ligandss
frameworks that are well established in diverse fields, such as
enantioselective catalysis,4 liquid crystals,5,6 and nonlinear
optical materials.7,8 Electrochemistry can be used to reversibly
alter the coordination environment for one of the metallofol-
damers, with the implicit consequence that the entire secondary
structure is altered, as well. This provides a basis for correlating
the aforementioned materials’ properties to electrochemistry and
other types of external switching mechanisms (e.g., photochem-
istry) that can alter metal geometry. The concept is illustrated
in Scheme 1.
The interplay between metal geometry, secondary structure,
and function is well established in biology and has, conse-
quently, served as the inspiration for the design of “foldamers”1,2s
unnatural molecules that fold into compact secondary structures.
Foldamers have been broadly classified as mimics of peptides1,2,9
and oligonucleotides10 and the abiotic analogues of each.11,12
† Questions regarding NMR and X-ray should be addressed to Dr. Bai
(bais@udel.edu) and Dr.Yap (gpyap@chem.udel.edu), respectively.
(1) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173-180.
(2) (a) Hill, D. J.; Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem.
ReV. 2001, 101, 3893-4012.
(3) We use Moore’s system of nomenclature to distinguish peptidomimetic
foldamers from their abiotic analogues. Paraphrasing from pages 3910 and
3944 of ref 2: peptidomimetic foldamer research is defined as a ‘top-down’
design approach that involves structural variations of parent chain molecules
known to undergo folding reactions. Research on their abiotic analogues
employs a ‘bottom-up’ design approach, where analogous architectures to
biomacromolecules are obtained by backbones that bear little resemblance
to natural chains.
(8) For materials whose NLO properties can be switched by electrochemistry:
(a) Powell, C. E.; Humphrey, M. G.; Cifuentes, M. P.; Morrall, J. P.; Samoc,
M.; Luther-Davies, B. J. Phys. Chem. A 2003, 107, 11264-11266 and
references therein. For a review that covers photoswitchable NLO materi-
als: (b) Delaire, J. A.; Nakatani, K. Chem. ReV. 2000, 100, 1817-1846.
(9) Reviews and lead references to peptidomimetics: (a) Porter, E. A.; Wang,
X.; Lee, H.-S.; Weisblum, B.; Gellman, S. H. Nature 2000, 404, 565. (b)
Seebach, D.; Matthews, J. L. J. Chem. Soc., Chem. Commun. 1997, 2015-
2022. (c) Degrado, W. F.; Schneider, J. P.; Hamuro, Y. J. Peptide Res.
1999, 54, 206-217. (d) Kirshenbaum, K.; Zuckermann, R. N.; Dill, K. A.
Curr. Opin. Struct. Biol. 1999, 9, 530-535. (e) Stigers, K. D.; Soth, M. J.;
Nowick, J. S. Curr. Opin. Chem. Biol. 1999, 3, 714-723.
(4) Larrow, J. F.; Jacobsen, E. N. Top. Organomet. Chem. 2004, 6, 123-152.
(5) Lead references to liquid crystalline Salen metal complexes: (a) Abe, Y.;
Nakabayashi, K.; Matsukawa, N.; Iida, M.; Tanese, T.; Sugibayashia, M.;
Ohta, K. Inorg. Chem. Commun. 2004, 7, 580-583. (b) Paschke, R.;
Balkow, D.; Sinn, E. Inorg. Chem. 2002, 41, 1949-1953. (c) Miyamura,
K.; Mihara, A.; Fujii, T.; Gohshi, Y.; Ishii, Y. J. Am. Chem. Soc. 1995,
117, 2377-2378. (d) Serrette, A.; Carroll, P. J.; Swager, T. M. J. Am.
Chem. Soc. 1992, 114, 1887-1889.
(6) For materials that can switch liquid crystalline materials: Feringa, B. L.;
van Delden, R. A.; Koumura, N.; Geertsema, E. M. Chem. ReV. 2000, 100,
1789-1816.
(10) Reviews and lead references to oligonucleotide mimetics: (a) Eschenmoser,
A.; Loewenthal, E. Chem. Soc. ReV. 1992, 21, 1-16. (b) Herdewijn, P.
Biochim. Biophys. Acta 1999, 1489, 167-179. (c) Nielsen, P. E.; Haaima,
G. Chem. Soc. ReV. 1997, 26, 73-78.
(7) Di Bella, S. Chem. Soc. ReV. 2001, 30, 355-366.
9
10590
J. AM. CHEM. SOC. 2005, 127, 10590-10599
10.1021/ja050886c CCC: $30.25 © 2005 American Chemical Society

Electrochemically Responsive Abiotic Metallofoldamers
A R T I C L E S
Scheme 2
Mimics of oligonucleotides that bind metals are well-known
for abiotic systems,11 but analogues of peptides are rare12,13 and
consist mainly of molecules in which the metal serves as a
template for a helical secondary structure. Thus, most abiotic
mimics of metallopeptides bear analogy to the structure in Figure
1a,13d in which the metal coordination sphere is inherently
helical.13 In contrast, we are not aware of abiotic materials of
the type outlined in Figure 1b, in which metal coordination
nucleates the formation of a nonbiological single-stranded helix.
That is, the metal coordination sphere is not inherently chiral,
but instead causes a series of cooperative, noncovalent interac-
tions that ultimately result in a folded structure. The closest
precedent of such a material has been described by Moore, in
which an oligophenyleneethynylene with radially oriented nitrile
ligands binds two separate silver atoms.13 That material can be
considered to be a hybrid of the extremes in Figure 1.14 Another
related material has been reported by Lee and co-workers, in
which the metal is a structural unit of the backbone of a single-
stranded abiotic foldamer.13f On the basis of small-angle X-ray
data, it was proposed that the secondary structure could be
altered by varying the counterion of the metal.13f
Figure 1. Classification of abiotic analogues of metallopeptides as (a)
templated and (b) nucleated secondary structures.
In this contribution, we show that Ni(II) and Cu(II) complexes
2a and 2b, unlike the free ligand 1, adopt well-defined helical
structures in the crystalline state and in solution (Scheme 2).
Remarkably, both 2a and 2b undergo dynamic resolution (to
>94% ee) upon crystallization; crystal growth experiments
produced a single crystal that racemizes immediately when
redissolved at 5 °C. Despite their rapid racemization, a single
conformation is observed by NMR, and NOE shows that the
structure is very similar to that of the crystal. Thus, 2a and 2b
are fluxional molecules in solution whose thermodynamically
most favored conformations are helical. Smaller analogues of
1 have also been characterized crystallographically and by NMR
spectroscopy. Their Ni complexes are also helical, but the free
ligands are not. Cyclic voltamograms and modeling experiments
show that electrochemical reduction of 2b in a coordinating
environment is followed by a change in the coordination
geometry at the metal center that consequently alters the
secondary structure of the entire molecule. Salen and salophen
metal complexes have been applied in diverse areas of materials
science, and the incorporation of salen and salophen structures
into chiral polymers is known to produce materials with
exceptional chiroptical properties.15-17 For complexes such as
2b, the chiral order in the folded state should correlate to useful
properties (e.g., 2nd order NLO; induction of chiral nematic
liquid crystalline phases) that can be switched on/off by folding/
unfolding. Toward such applications, this work shows that
electrochemistry can be used to reversibly switch the structure
of Cu complex 2b from a thermodynamically helical state [as
tetracoordinate Cu(II)] to a nonhelical state [as pentacoordinate
Cu(I)].
(11) Reviews and lead references to abiotic single-stranded foldamers: (a)
Moore, J. S. Acc. Chem. Res. 1997, 30, 402-413. (b) Gin, M. S.;
Yokozawa, T.; Prince, R. B.; Moore, J. S. J. Am. Chem. Soc. 1999, 121,
2643-2644. (c) Stone, M. T.; Fox, J. M.; Moore, J. S. Org. Lett. 2004, 6,
3317-3320. (d) Hamuro, Y.; Geib, S. J.; Hamilton, A. D. J. Am. Chem.
Soc. 1997, 119, 10587-15093. (e) Hamuro, Y.; Hamilton, A. D. Bioorg.
Med. Chem. 2001, 9, 2355-2363. (f) Gong, B. Chem.sEur. J. 2001, 7,
4336-4342. (g) Yang, X.; Yuan, L.; Yamato, K.; Brown, A. L.; Feng,
W.; Furukawa, M.; Zeng, X. C.; Gong, B. J. Am. Chem. Soc. 2004, 126,
3148-3162. (h) Huc, I. Eur. J. Org. Chem. 2004, 17-29. (i) Jiang, H.;
Dolain, C.; Leger, J.-M.; Gornitzka, H.; Huc, I. J. Am. Chem. Soc. 2004,
126, 1034-1035. (j) Zych, A. J.; Iverson, B. L. J. Am. Chem. Soc. 2000,
122, 8898-8909. (k) Berl, V.; Huc, I.; Khoury, R. G.; Lehn, J. M. Chem.s
Eur. J. 2001, 7, 2798-2809. (l) Cuccia, L. A.; Lehn, J.-M.; Homo, J.-C.;
Schmutz, M. Angew. Chem., Int. Ed. 2000, 39, 233-237. (m) Zhang, W.;
Horoszewski, D.; Decatur, J.; Nuckolls, C. J. Am. Chem. Soc. 2003, 125,
4870-4873. (n) Wu, Z.-Q.; Jiang, X.-K.; Zhu, S.-Z.; Li, Z.-T. Org. Lett.
2004, 6, 229-232. (o) Preston, A. J.; Fraenkel, G.; Chow, A.; Gallucci, J.
C.; Parquette, J. R. J. Org. Chem. 2003, 68, 22-26. (p) Gawronski, J.;
Gawronska, K.; Grajewski, J.; Kacprzak, K.; Rychlewska, U. J. Chem. Soc.,
Chem. Commun. 2002, 582-583.
(12) Reviews and lead references to abiotic multiply stranded foldamers: (a)
Lehn, J.-M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevrier, B.; Moras,
D. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 2565-2569. (b) Piguet, C.;
Bernardinelli, G.; Hopfgartner, G. Chem. ReV. 1997, 97, 2005-2062. (c)
Albrecht, M. Chem. ReV. 2001, 101, 3457-3498. (d) Berl, V.; Huc, I.;
Khoury, R. G.; Krische, M. J.; Lehn, J. M. Nature 2000, 407, 720-723.
(e) Archer, E. A.; Gong, H.; Krische, M. J. Tetrahedron 2001, 57, 1139-
1159. (f) Gong, B. Synlett 2001, 5, 582-589.
(13) Lead references and reviews of abiotic single-stranded foldamers that bind
metals: (a) Prince, R. B.; Okada, T.; Moore, J. S. Angew. Chem., Int. Ed.
1999, 38, 233-236. (b) Mizutani, T.; Yagi, S.; Morinaga, T.; Nomura, T.;
Takagishi, T.; Kitagawa, S.; Ogoshi, H. J. Am. Chem. Soc. 1999, 121, 754-
759. (c) Constable, E. C. Tetrahedron 1992, 48, 10013-10059. (d)
Constable, E. C.; Drew, M. G. B.; Forsyth, G.; Ward, M. D. J. Chem.
Soc., Chem. Commun. 1988, 1450-1451. (e) Ho, P. K.-K.; Cheung, K.-
K.; Peng, S.-M.; Che, C.-M. J. Chem. Soc., Dalton Trans. 1996, 1411-
1417. (f) Kim, H.-J.; Zin, W.-C.; Lee, M. J. Am. Chem. Soc. 2004, 126,
7009-7014.
Results and Discussion
Design and Synthesis. The rationale for salen and salophen
metal complexes that form single-stranded helical foldamers is
(15) Dai, Y.; Katz, T. J. J. Org. Chem. 1997, 62, 1274-1285.
(16) (a) Zhang, H.-C.; Huang, W.-S.; Pu, L. J. Org. Chem. 2001, 66, 481-487.
(b) Pu, L. Chem. ReV. 1998, 98, 2405-2494.
(17) Tanaka, M.; Fujii, Y.; Okawa, H.; Shinmyozu, T.; Inazu, T. Chem. Lett.
1987, 1673-1674.
(14) The exterior turns of the helical structure described by Moore and co-
workers (ref 13a) are templated by each of the silver atoms, whereas the
central turn of the structure is controlled by noncovalent interactions.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10591

A R T I C L E S
Scheme 3
Zhang et al.
outlined in Figure 2. Our hypothesis was that a stable helix
would result for square planar metal complexes of ligand 1 as
a result of four six-membered hydrogen bonds between amide
donors and phenolic acceptors (Figure 2a). The helical con-
former would be further reinforced by a series of aromatic
π-stacking interactions between every fourth aromatic ring in
the sequence (as depicted in Figure 2b). The core of the
secondary structure could be thought of as a chiral analogue of
the macrocycle 317 (Figure 2c), which adopts a planar structure
with six-membered ring H-bonds.
Ligand 1 can be synthesized as shown in Scheme 3. Our
approach makes use of Stanetty’s protocol18 for making 7-ben-
zofuran carboxylic acid and precedent for preparing salicylal-
dehydes with amide functionality through ozonolysis of ben-
zofurans.19 The advantage of this route is that 5 can be prepared
in multigram quantities, and that the amide bond-forming
reactions are straightforward. Thus, while more expedient
syntheses of 1 can be envisioned, the synthesis in Scheme 3
was ideal from a discovery perspective. The metal complexes
2a and 2b were prepared in good yields (Scheme 2, 73% for
2a, 75% for 2b) by combination of 1 with Ni(OAc)2 and Cu-
(OAc)2, respectively.
Model Compounds. Before the preparation of 2a and 2b,
we first prepared the smaller models 11 (Scheme 4) and 14
(displayed in Figure 7). The syntheses mirrored those of 2a and
2b. For example, the reaction of p-toluidine with 5 and
subsequent ozonolysis gave 10, which condensed with Ni(OAc)2
to give 11. A similar sequence transforms 13 to 14 (structures
displayed in Figure 7). Like the larger analogues 2a and 2b,
11 and 14 adopt helical secondary structures, but their metal-
free precursors do not. These compounds establish the gener-
ality of helical folding for a class of metal complexes. Further,
the X-ray structure of ligand 10 provides structural support
Figure 3. Molecular diagram for ligand 10 prepared in Chem3D using
crystallographic coordinates. A thermal ellipsoid plot is available in the
Supporting Information.
Figure 2. Design of metal-dependent helical foldamers.
9
10592 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Electrochemically Responsive Abiotic Metallofoldamers
A R T I C L E S
Figure 4. Molecular diagrams for nickel complex 2a in Chem3D prepared from the crystallographic coordinates. The n-hexyl chains have been deleted for
clarity of presentation. A thermal ellipsoid plot is available in the Supporting Information.
Scheme 4
for our claim that such ligands do not fold if the metal is absent.
X-ray quality crystals of 10 could be grown directly in the
reaction mixture from which it was prepared (EtOH). The
structure of 10 is shown in Figure 3. Although we were unable
to grow X-ray quality crystals of 11, solution NMR studies (vide
infra) show that it is a helix in solution, whereas the structure
of free-ligand 10 is not helical in solution or in the crystal.
Crystallographic and Racemization Studies of 2a and 2b.
X-ray diffraction of both 2a and 2b revealed nearly identical
structures for the two metal complexes of ligand 1. Two
perspectives of the structure of 2a are displayed in Figure 4,
and the structure of 2b is displayed in Figure 5.
For a foldamer to be useful as a responsive material, it is
necessary that its secondary structure be able to form and reform
easily. Because we were able to grow large crystals of
enantiomerically pure 2a and 2b, it was possible to test the
stability of those helical molecules toward racemization in
solution. Both 2a and 2b spontaneously resolve upon crystal-
lization; crystals of 2a were grown from CHCl3/CH3CN and
crystals of 2b from CHCl3/hexanes. Interestingly, a crystal of
compound 2a that was grown from a different solvent system
(MeOH/CHCl3) formed racemic crystals in a centrosymmetric
space group instead, although the structure of the individual
molecules was still helical.
example, the bands for 12 that we assign as d-π* charge-
transfer transitions20b (418 nm, ∆ꢀ -23; 444 nm, ∆ꢀ -18) are
approximately 3 times larger than those previously reported for
N,N′-bis(salicylidene)-(R,R)-cyclohexane-1,2-diaminonickel-
(II) (413 nm, ∆ꢀ -8). We submit that the large CD signals
provide evidence that the structure of 12 is helical. By analogy,
strong circular dichroisms have previously been assigned to the
d-π* charge-transfer transitions of polymeric chiral salophens
derived from helicenes (∆ꢀ ∼ 100)15 and binaphthyls (∆ꢀ ∼
45).16 The even larger dichroisms for the polymeric materials
are partly due to increased conjugation through the phenylene
(as opposed to a cyclohexane) and to additional conjugation
through the helicenes for the former materials.
To determine the stability of metallofoldamers toward race-
mization, we chose a large (∼2 mm) crystal of 2a that we
determined to be single by polarized light optical microscopy.
From the X-ray diffraction data obtained from an appropriately
dimensioned section of the single crystal, it was determined that
2a crystallized in the noncentrosymmetric space group P212121
and, therefore, spontaneously resolved. From a refinement of
the Flack parameter based on anomalous dispersion effects, the
data crystal was 94% enantiomerically pure. The remaining
pieces from the original single crystal were then dissolved in
CH2Cl2 (3.5 × 10-4 M) at 5 °C, and the CD spectrum was
recorded. Data acquisition was completed within 5 min of the
sample preparation, and no signal could be detected. A similar
experiment conducted at room temperature at the polarimeter
showed no rotation at the D-line of sodium. The conclusion is
that solutions of 2a racemize quickly at 5 °C, and the implication
is that geometrical changes at the metal center are capable of
reorganizing the entire secondary structure. When an identical
series of experiments were performed with 2b having a crystal
phase isomorphous to that of 2a and with a refined 99%
enantiomeric purity, the conclusion was the same: the helix
had racemized at 5 °C before we were able to record a CD
spectrum. The implication is that the thermodynamic structures
of 2a and 2b are helical in solution. A useful analogy can be
drawn to the helicenes,21 which also represent a class of
thermodynamically helical structures that racemize.
Nonracemic 2a and 2b would be expected to possess
significant chiroptical properties. As a point of reference, we
prepared chiral salen derivative 12 (Figure 6) from 8, (R,R)-
(-)-trans-1,2-cyclohexanediamine, and Ni(OAc)2. Complex 12
has a strong CD spectrum (Figure 6) and a large specific rotation
([R]25 -730°). We note that the strength of the circular
D
dichroisms for 12 is significantly larger than that of N,N′-bis-
(salicylidene)-(R,R)-cyclohexane-1,2-diaminonickel(II).20 For
(18) (a) Stanetty, P.; Pu¨rstinger, G. J. Chem. Res. (M) 1991, 78. (b) Stanetty,
P.; Koller, H.; Pu¨rstinger, G. Monatshefte fu¨r Chemie 1990, 121, 883-
891.
(19) Toyota, E.; Itoh, K.; Sekizaki, H.; Tanizawa, K. Bioorg. Chem. 1996, 24,
150-158.
(20) (a) Downing, R. S.; Urbach, F. L. J. Am. Chem. Soc. 1970, 92, 5861-
5865. (b) Bosnich, B. J. Am. Chem. Soc. 1968, 90, 627-632. (c) Costes,
J. P.; Dominguez-Vera, J. M.; Laurent, J. P. Polyhedron 1995, 14, 2179-
2187.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10593

A R T I C L E S
Zhang et al.
Figure 5. Molecular diagram of copper complex 2b depicted with ellipsoids
at 30% probability. Solvent molecules and hydrogen atoms are omitted for
clarity.
Figure 6. CD spectrum of 12 (2.2 × 10-4 M in CH2Cl2).
Solution NMR Studies: Chemical Shift Anisotropy as a
Measure of Helicity. NMR studies evidence that Ni complex
2a and smaller analogues 11 and 14 are helical in solution,
whereas their free ligands are not. One way that this could be
demonstrated was by directly comparing the 1H chemical shifts
of the aromatic protons of the metal complexes to free ligands.
For example, if Ni complex 14 were to adopt a helical
conformation, then it would be expected that hydrogens of the
terminal tolyl rings would resonate at a substantially higher field
than the analogous protons on reference compound 13. Chemical
shift anisotropy22 of this type is well-known for helical aromatic
molecules,21 and it has been used to provide evidence for the
helical solution structure of several types of peptidomimetic
foldamers, including oligoanthranilimides11d and oligo-m-
phenyleneethynylenes.13a Indeed, chemical shift analysis does
suggest that the solution structures of Ni complexes 11 and 14
are helical, as the protons of the tolyl group resonate at
substantially higher field than those of 13. For 14, The ∆δ values
are 0.57 and 0.42 ppm for the aromatic resonances, and ∆δ )
0.16 ppm for the methyl resonances. In comparison, the
corresponding hydrogens on the salicyl rings of 13 and 14 have
chemical shifts that nearly coincide. Thus, the large changes in
chemical shift must be attributed to the underlying ring currents
rather than to an electronic effect upon metal binding since metal
coordination should perturb the electronics of the salicyl
hydrogens more strongly than those of the tolyl. The aromatic
hydrogens and methyls on the tolyl group of nonracemic Ni
Figure 7. Chemical shift anisotropy as a measure of helicity.
complex 11 have similar chemical shifts to 14 and resonate at
high field when compared to the corresponding hydrogens of
13. The ∆δ values are 0.54 and 0.40 ppm for the aromatic
hydrogens and 0.13 for the methyl groups. Again, the perturba-
tion of the salicyl aromatic hydrogens is minor.
NMR spectral studies on the larger structure 2a strongly
suggest that the solution structure is also helical. This was
1
demonstrated in part by comparing the H NMR spectra of 2a
with reference compounds 6, 7, 8, and 11. Unambiguous
chemical shift assignments for each of the 14 different hydrogens
and 22 different carbons of the core structure of 2a were made
by an extensive series of NMR studies (HMBC, HSQC, and
COSY). Similarly, rigorous assignments were made for each
of the model compounds 6, 7, 8, and 11. Nine hydrogens in 2a
were found to resonate at significantly higher field when
compared to the reference compounds (Figure 8). Anisotropic
effects are observed both at the core and at the ends of the
structure. The hydrogens on the terminal methoxy groups of
2a are shifted upfield by ∼0.6 ppm in comparison to 8, and the
hydrogens bound to the imines (H21) are shifted upfield by 0.43
ppm relative to 11. A table is provided in the Supporting
1
Information that compares all of H NMR chemical shifts for
(21) Laarhoven, W. H.; Prinsen, W. J. C. Top. Curr. Chem. 1984, 125, 63-
130.
2a and 6-8 (those for 11 are in Figure 8) and supports the
assertion that the solution structure of 2a is a helix.
(22) Abraham, R. J.; Fell, S. C. M.; Smith, K. M. Org. Magn. Reson. 1977, 9,
367 and references therein.
9
10594 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Electrochemically Responsive Abiotic Metallofoldamers
A R T I C L E S
Figure 10. Selected trans-helix NOEs that were observed for 2a.
would be expected that upon heating that the 1H NMR spectrum
of 1 might coalesce. We heated a solution of 1 in DMSO-d6 to
100 °C and observed only minor changes in the 1H NMR
spectrum. Optical rotation and circular dichroism experiments
(vide supra) have shown that complex 2a racemizes quickly in
solution at 5 °C, and in this context, it is difficult to rationalize
why the free ligand would adopt secondary structures with such
unusual stability. A more likely explanation is that the free-
ligand 1srich in both H-bond donors and acceptorssforms an
aggregate in nonpolar solvents. Consistent with this interpreta-
Figure 8. Comparison of the 1H NMR chemical shifts of 2a with 8 or 11.
The chemical shifts of hydrogens with blue labels were compared to 8,
and hydrogens with red labels were compared to 11. The ∆δ values refer
to upfield chemical shifts. H10, H18-H20, and H22 did not show significant
differences when compared to these standards. A table with all of the
chemical shift data is displayed in the Supporting Information.
1
tion, the polar solvent DMF-d8 disrupts the aggregate. The H
NMR spectrum in that solvent shows g70% of a single
compound that we assign as monomeric 1.
NOESY Studies of 2a. That Ni complex 2a adopts a helical
structure in solution is further supported by the phase sensitive
NOESY23 spectra, which show nuclear Overhauser effects
between the protons on the terminal anisoyl groups and those
on the internal salophen. The observed trans-helix NOE signals
are shown in Figure 10. As expected for a molecule of this size,
the cross-peaks in the NOESY spectra have the opposite sign
with respect to the diagonalsan indication that the cross-peaks
arise from positive NOE enhancements. The cross-peaks (H3-
H22, H4-H22, H4-H24, H5-H20, H5-H22, H6-H18, H6-
H19, and H6-H20) shown in Figure 11 qualitatively confirm
the close proximities between terminal anisoyl and salophen
protons. The cross-peaks (OCH3-H18, OCH3-H19, OCH3-
H20, OCH3-H22, and OCH3-H24) between methoxy protons
on the anisoyl group and salophen protons further support the
assignment of a helical structure in solution.
Conformational analysis using NOE can be especially chal-
lenging24 for small molecules. External relaxations (often termed
“leakages”) introduce additional relaxation mechanisms that may
interfere with the NOE measurements. Leakage is particularly
problematic for small molecules, even when precautionary
measures (choosing shorter mixing times, carefully degassing
the solution) are taken. The high conformational flexibility of
small molecules often makes estimation of the internuclear
distances impossible. An implication for foldamers is that the
observation of NOE signals can be taken as an indicator of a
stable secondary structure.
Figure 9. (a) 1H NMR spectrum (CDCl3) of ligand 1 (>70% pure). (b) 1H
NMR spectrum of 2a (CDCl3). The complexity of the NMR spectrum of 1
is attributed to intermolecular aggregation.
It was not possible to directly compare free-ligand 1 with 2a
1
because 1 has an extremely complex H NMR spectrum in
CDCl3 or CD2Cl2 (the solvents in which extensive NMR
characterization was performed for 2a). The 1H NMR spectrum
of 1 provides a stark contrast to the crisp spectrum of 2a (Figure
9). Despite the complexity of the NMR spectrum, the formation
of the ligand was confirmed by the observation of a prominent
parent peak (1043.3 ) M + Na) by ESI MS. It was considered
that the complex NMR spectrum is observed for 1 because it
adopts several secondary structures that do not interconvert on
the NMR time scale. If this interpretation were correct, then it
(23) Macura, S.; Huang, Y.; Suter, D.; Ernst, D. D. J. Magn. Reson. 1981, 43,
259-281.
(24) Claridge, T. D. W. High-Resolution NMR Techniques in Organic Chemistry;
Pergamon: Amsterdam, 1999.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10595

A R T I C L E S
Zhang et al.
Table 1. Internuclear Distances of Ni Complex 2a from NOESY
and X-ray Crystallography
proton pair
distance (NOE)a
Å
distance (X-ray)c
Å
H3-H22
4.60
3.94
4.10
3.98
4.38
4.55
4.52
4.28
5.08b
5.21b
4.17b
4.95b
5.10b
4.45
H4-H22
3.82
4.64
3.43
4.17
4.11
3.94
3.80
H4-H24
H5-H20
H5-H22
H6-H18
H6-H19
H6-H20
OCH3-H18
OCH3-H19
OCH3-H20
OCH3-H22
OCH3-H24
4.66, 6.24, 5.76 (5.25)d
5.15, 6.55, 5.92 (5.67)d
4.71, 4.73, 5.77 (4.93)d
4.30, 4.95, 3.66 (4.08)d
6.65, 5.78, 4.58 (5.12)d
a Distances obtained using the intensity ratio method or eq 1; an estimated
error of 0.3 Å is expected in case of a 50% inaccuracy25d of the intensity
ratio. b The cross-peak intensity of methoxy protons was divided by 3 since
diagonal peaks of H18-H24 were used in the calculations. c X-ray distances
are obtained by simply adding hydrogen in the X-ray structures, realizing
that these distances merely serve as the benchmarks to validate NOE
distances. d The rapid internal rotation of the methyl group, on the NMR
time scale, dictates that only the time average of interproton distance is
meaningful. The numbers in parentheses are calculated from the X-ray
distance of these protons based on r-6 -1/6 and represent the mean values
of three proton positions. This mean value is heavily weighted on the shorter
distances because of a r-6 dependence of the NOE distances.
that the molecular tumbling is isotropic and interproton vectors
possess the same correlation time, the interproton distances of
the terminal anisoyl protons and those on the internal salophen
can be calculated from a single NOESY spectrum with a
correlation time of 5.4 × 10-11 s. The assumption should be
valid if 2a adopts a stable helix. The interproton separations
listed in the table are the pairs between the terminal anisoyl
aromatic or methoxy protons and the internal salophen protons.
The NOE interproton distances are listed in Table 1 and are
compared to the distances obtained from the X-ray crystal-
lographic studies described in the previous section. In general,
the distances as measured by X-ray are in close agreement to
those measured by the NOE experiment. In accord with the
conclusion from the NMR anisotropy studies, the results of the
NOE analysis strongly suggest that the helical conformation of
2a in crystal is preserved in solution.
Figure 11. NOESY spectrum of 2a.
Quantitative NOE measurements were conducted on 2a using
the intensity ratio method,25 which extracts internuclear distances
from the intensity ratio of cross and diagonal peaks from a single
NOESY spectrum. The intensity ratio method is experimentally
much more straightforward than the “NOE build-up” method,26
which requires a series of NOESY spectral acquisitions as a
function of mixing time. Both methods have been shown to be
in excellent agreement for the evaluation of distances in small
molecules.25e The interproton separation, r, is expressed in terms
of the molecular correlation time, τc, and the mixing time, τmix
,
as follows (eq 1):25d,e
1/6
0.2 γ4p2(µ0/4π)2 τmix
6τc
Electrochemical Studies. Cyclic voltammograms (CV) of
2b in THF and CH2Cl2 are displayed in Figure 12. In THF, the
CV shows a single, reversible one-electron transfer at E1/2 )
r )
- τc
(1)
2
2
(
)
Iaa + Iab
1 + 4ω τc
ln
{
}
[
]
Iaa - Iab
-0.63 V that was unaltered after three repeated cycles. A similar
CV was recorded in CH2Cl2 with E1/2 ) -0.75 V. After
correcting for the proportional shift of the ferrocenium/ferrocene
couple (0.801 V in THF versus 0.701 in CH2Cl2.), a small (20
mV) cathodic shift is observed when the solvent is changed
from THF to CH2Cl2. These CVs are indicative of metal-
centered Cu(II)/Cu(I) redox events for a Cu-salophen deriva-
tive.28 The shift to a more positive potential as compared to
that of unsubstituted Cu-salophen (E1/2 ) -1.11 V) is an
expected consequence of substituting the salophen nucleus by
Iaa and Iab are intensities of diagonal and cross-peaks, respec-
tively, and ω is the proton Larmor frequency. Using eq 1, the
molecular correlation time, τc, of Ni complex 2a is calculated
to be 5.4 × 10-11 s from an interproton distance of 2.8 Å for
ortho-substituted benzene protons24 H18-H19 (see Figure 10
for atom labels). The experimental mixing time of 0.8 s and
the NOE cross and diagonal peak intensities of H18 and H19
were used in these calculations. If the assumption27 is made
(25) (a) Bodenhausen, G.; Ernst, R. R. J. Am. Chem. Soc. 1982, 104, 1304-
1309. (b) Macura, S.; Farmer, B. T.; Brown, L. R. J. Magn. Reson. 1986,
70, 493-499. (c) Esposito, G.; Carver, J. A.; Boyd, J.; Campbell, I. D.
Biochemistry 1987, 26, 1043-1050. (d) Esposito, G.; Pastore, A. J. Magn.
Reson. 1988, 76, 331-336. (e) Reggelin, M.; Hoffmann, H.; Ko¨ck, M.;
Mierke, D. F. J. Am. Chem. Soc. 1992, 114, 3272-3277.
(27) (a) Keepers, J. W.; James, T. L. J. Magn. Reson. 1984, 57, 404-426. (b)
Andersen, N. H.; Eaton, H. L.; Lai, X. Magn. Reson. Chem. 1989, 27,
515-528.
(28) (a) Patterson, G. S.; Holm, R. H. Bioinorg. Chem. 1975, 4, 257-275. (b)
Rohrbach, D. F.; Heineman, W. R.; Deutsch, E. Inorg. Chem. 1979, 18,
2536-2542. (c) Doine, H.; Stephens, F. F.; Cannon, R. D. Inorg. Chim.
Acta 1983, 75, 155-157. (d) Zanello, P.; Cinquantini, A. Transition Met.
Chem. 1985, 10, 370-374.
(26) Kumar, A.; Wagner, G.; Ernst, R. R.; Wu¨rhrich, K. J. Am. Chem. Soc.
1981, 103, 3654-3658.
9
10596 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Electrochemically Responsive Abiotic Metallofoldamers
A R T I C L E S
Figure 12. Cyclic voltammograms of 2b (1 × 10-3 M) with nBu4NPF6
(0.1 M) as supporting electrolyte. A platinum electrode was used as working
electrode, together with a platinum wire auxiliary electrode. The reference
Ag/Ag+ electrode is filled with the supporting electrolyte solution (0.1 M
nBu4NPF6). Ferrocene was used as an internal standard (1 × 10-3 M), and
potentials were rescaled to E0(Fc/Fc+) ) 0.801 V versus NHE (THF) and
to E0(Fc/Fc+) ) 0.701 V versus NHE (CH2Cl2).
electron-withdrawing groups that are in conjugation with the
phenolic oxygens.29 In contrast to the reversible behavior
observed in THF or CH2Cl2, quasireversible behavior was
observed when the CV of 2b was measured in CH3CN (Figure
13). Thus, while a single peak (Epc ) -0.75 V) was observed
during the cathodic sweep, two waves (Epa ) -0.60 and -0.16
V) were observed during the anodic sweep. With the E1/2 of
ferrocenium/ferrocene as the reference point, the Epc for 2b in
acetonitrile is anodically shifted by 80 mV compared to the Epc
of 2b in THF or CH2Cl2. The CV was also measured for a
solution of 2b in THF containing 5% N,N-4-(dimethylamino)-
pyridine (DMAP), and irreversible behavior was noted. Thus,
while the peak observed during the cathodic sweep was similar
to that observed in pure THF, a large shift (from -0.70 in THF
to -0.26 V in THF/DMAP) was observed during the anodic
sweep.
Figure 13. Cyclic voltammograms of 2b (1 × 10-3 M) in acetonitrile (top)
and in 5% DMAP/THF (bottom). nBu4NPF6 (0.1 M) was the supporting
electrolyte. A platinum electrode was used as working electrode, together
with a platinum wire auxiliary electrode. The reference Ag/Ag+ electrode
is filled with the supporting electrolyte solution (0.1 M nBu4NPF6). Ferrocene
was used as an internal standard (1 × 10-3 M), and potentials were rescaled
to E0(Fc/Fc+) ) 0.801 V versus NHE (THF) and to E0(Fc/Fc+) ) 0.651
V versus NHE (CH3CN).
While a tetrahedral geometry is most usual for tetracoordinate
Cu(I), the rigidity of the salophen ligand precludes tetrahedral
coordination. In the late 1970s, Gagne´ characterized and studied
a number of complexes of Cu(I) with tetradentate ligands that
enforce near-planar coordination30 and found that Cu(I) com-
plexes of such ligands sometimes adopt distorted square planar
geometries.30d,e Alternatively, Cu(I) complexes of this type can
form five-coordinate, square pyramidal structures (e.g., crys-
tallographically characterized 15,30 Figure 14) when they are
generated in the presence of monodentate ligands, such as carbon
monoxide, isonitriles, phosphates, phosphines, N-methylimida-
zole, or acetonitrile.30 Although binding of external ligands to
Cu(I)-salen or Cu(I)-salophen has not been investigated
extensively, the binding of carbon monoxide to these complexes
was studied and found to give a square pyramidal complex,30a
and apical solvent coordination has been suggested for elec-
Figure 14. Square pyramidal complexes of Cu(I) and Zn(II) from
tetradentate, macrocyclic ligands.
trochemically generated Cu(I)-salen complexes in DMF.31
Apical coordination of solvent is also observed in crystallo-
graphically characterized Zn(II)-salen complexes (e.g., 16,32
Figure 14), which are expected to be isostructural to their Cu-
(I) analogues.
We propose that Cu(I)-2b is tetracoordinate when generated
electrochemically in CH2Cl2. This interpretation is consistent
with the reversible nature of the CV data, the expectation that
CH2Cl2 would be a poor ligand, and the small solvent effect
upon changing to THF. Solvent coordination cannot be com-
(29) Zolezzi, S.; Spodine, E.; Decinti, A. Polyhedron 2002, 21, 55-59.
(30) (a) Gagne´, R. R.; Allison, J. L.; Ingle, D. M. Inorg. Chem. 1979, 18, 2767-
2774. (b) Gagne´, R. R.; Allison, J. L.; Gall, R. S.; Koval, C. A. J. Am.
Chem. Soc. 1977, 99, 7170-7178. (c) Gagne´, R. R. J. Am. Chem. Soc.
1976, 98, 6709-6710. (d) Gagne´, R. R.; Koval, C. A.; Smith, T. J. J. Am.
Chem. Soc. 1977, 99, 8367-8368. (e) Gagne´, R. R.; Allison, J. L.; Lisensky,
G. C. Inorg. Chem. 1978, 17, 3563-3571.
(31) The role of the solvent as a ligand in electrochemical reduction of Cu-
salophen has been pointed out by Deutsch,28b who compared the reductions
of Cu(II) and Co(II) Schiff base complexes in DMF and noted the “similar
dependencies of the Cu(I)/C(II) and Co(I)/Co(II) couples on equatorial
ligand structure (with the common axial ligand DMF)”.
(32) Reglinski, J.; Morris, S.; Stevenson, D. E. Polyhedron 2002, 21, 2175-
2182.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10597

A R T I C L E S
Zhang et al.
Figure 15. Calculated (AM1) low energy conformer of a square pyramidal complex of 2b. The model implies that apical coordination of any ligand (i.e.,
CH3CN, DMAP, or THF) will disrupt folding on that face. The apical ligand for the calculation is THF.
Figure 16. Model for quasireversible redox behavior of 2b in the presence of the external ligand 4-(dimethylamino)pyridine. Ligand association occurs
after electrochemical reduction and disrupts the helical secondary structure of the complex. Electrochemical oxidation and subsequent ligand dissociation
regenerate the helical Cu(II) complex. Quasireversible behavior is also observed when CH3CN is the external ligand.
pletely ruled out for THF, but the reversible nature of the CV
and the small magnitude of the anodic shift relative to CH2Cl2
imply that the binding constant for THF is small. However, the
effect of switching solvent to acetonitrile is significant and
provides strong evidence that acetonitrile binds after the
electrochemical reduction step. We propose that the resulting
complex is pentacoordinate and structurally analogous to 15 and
16. During the anodic sweep, more than one mechanism is in
operation. We attribute the peak at -0.60 V to solvent
dissociation prior to oxidation, whereas the peak at -0.16 V is
assigned to the electrochemical oxidation of the acetonitrile
complex. For the latter mechanism, dissociation of acetonitrile
should occur rapidly upon oxidation; we find no change in the
UV-vis absorption spectrum of Cu(II)-2b when the solvent
is changed from THF to THF/acetonitrile. In 5% DMAP/THF,
the CV is irreversible. Again, we propose five-coordinate
binding in analogy to 15 and 16, and that the binding occurs
after the electrochemical reduction. We attribute the peak at
-0.26 mV to the electrochemical oxidation of the pentacoor-
dinate complex; dissociation of DMAP does not take place prior
to oxidation. As is the case for the acetonitrile complex, we
expect that ligand dissociation is rapid upon reoxidation; we
find no difference between the UV-vis absorption spectrum
of Cu(II)-2b in pure THF versus 5% DMAP/THF.
Computational Studies. Apical coordination to 2b in a
manner analogous to that of 15 and 16 would completely block
one face of the salophen and, by necessity, disrupt the helix on
that face. Because the electrochemical experiments provide
direct evidence for ligand binding, they also provide evidence
for reversible helix folding and unfolding. To gain insight into
the structure of the pentacoordinate species, we modeled the
structure of the Cu(I) form of 2b with THF as an apical ligand.
The core structure of the salophen and the oxygen of the THF
were held fixed during the calculations, and the metal atom was
replaced by a stationary “dummy” atom to hold the place of
copper. Calculations were conducted at the semiempirical (AM1)
level. Some of the starting geometries for the semiempirical
calculations were manually selected, logical starting points that
could maximize noncovalent interactions. We also selected
starting points for the AM1 calculation by taking the lowest
energy structures from a molecular mechanics (MM3) Monte
Carlo multiple minimum search. All of the local minima that
we identified from the AM1 calculations have structures in
which helicity is disrupted on the face to which the apical ligand
9
10598 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Electrochemically Responsive Abiotic Metallofoldamers
A R T I C L E S
is bound. In the lowest energy configuration that we calculated
(Figure 15), the top face of the molecule is extended and
stabilized by a new intramolecular hydrogen bond between the
carbonyl of the internal amide and the N-H of the external
amide. While THF was the apical ligand for this computational
study, it is apparent that apical binding of any ligand (e.g.,
DMAP or acetonitrile) would disrupt the helix on that face.
Thus, the redox behavior of metallofoldamer 2b can be
correlated to a dramatic change in secondary structure in
coordinating environments. Figure 16 summarizes the structural
reorganization that takes place upon redox cycling. Future
studies will be directed toward altering materials’ properties
(e.g., NLO, liquid crystallinity) in response to reversible
perturbations of foldamer structure.
proposed that apical binding of those ligands gives square
pyramidal complexes. Semiempirical (AM1) calculations sup-
port that the helical structure would be disrupted by the reduction
to Cu(I) with concomitant reorganization to a square pyramidal
complex.
Acknowledgment. The University of Delaware is thanked
for financial support of this work. F.Z. is grateful for support
received through the competitive University Graduate Fellow-
ship program at UD. We thank Charles Riordan (UD) and Bill
Dougherty (UD) for assistance in obtaining and interpreting CV
data, and Olga Dmitrenko for assistance with calculations at
the AM1 level.
Supporting Information Available: A full experimental
Conclusions
description of the preparation and characterization of all new
1
compounds is described. H NMR and 13C NMR spectra are
In summary, we have designed and synthesized nonbiological
molecules based on salophen and salen ligands that can fold
into single-stranded helices in the presence of either Ni(II) or
Cu(II). That the structures are folded is strongly supported by
X-ray diffraction and NMR studies. Metal coordination is
required for folding, as these same techniques show that the
free ligands do not adopt helical structures. Two of the racemic
metallofoldamers spontaneously resolved during crystallization.
CD spectroscopy and optical rotation show that both molecules
racemize quickly at 5 °C, showing that the secondary structures
can reorganize easily and can, therefore, provide the basis for
responsive materials. Electrochemical experiments show that
a structural reorganization occurs upon metal-centered reduc-
tion of a Cu(II)-containing foldamer. When the reduction
is carried out in the presence of coordinating ligands, it is
displayed for all new compounds. A table compares the chemical
shifts of 2a, 8, 7, and 6 to provide a measure of helicity of 2a.
COSY, HSQC, and HMBC data are provided for compounds
6, 7, 8, 2a, and 11. HSQC and HMBC data are provided for
compound 12. The ESI MS spectrum of crude 1 is displayed.
The CV of 2b after four repeated cycles is displayed. UV-vis
spectra are displayed for compounds 2a, 2b, 11, 12, and 14.
The CD spectrum of 14 is displayed. X-ray crystallographic
data are provided for compounds 2a, 2b, and 10. Also displayed
are photographs of crystals of 2a and 2b with transmitted light
and under cross-polarized transmitted light. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA050886C
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10599