Extremely strong and readily accessible AAA-DDD triple hydrogen bond complexes

Article abstract of DOI:10.1021/ja067410t

Extremely high association constants (Ka's) of up to 2 × 10⁷ M^-1 in CH₂Cl₂ at room temperature are measured for chemically stable AAA-DDD and AA-DDD complexes which featurenovel and readily accessible multiple h

Full text of DOI:10.1021/ja067410t

Published on Web 12/29/2006
Extremely Strong and Readily Accessible AAA-DDD Triple Hydrogen Bond
Complexes
Smilja Djurdjevic, David A. Leigh,* Hamish McNab,* Simon Parsons, Gilberto Teobaldi, and
Francesco Zerbetto*
School of Chemistry, UniVersity of Edinburgh, The King’s Buildings, West Mains Road, Edinburgh EH9 3JJ, U.K.,
and Dipartimento di Chimica “G. Ciamician”, UniVersita` di Bologna, V. F. Selmi 2, 40126 Bologna, Italy
Received October 16, 2006; E-mail: david.leigh@ed.ac.uk; h.mcnab@ed.ac.uk; francesco.zerbetto@unibo.it
The development of multipoint hydrogen bonding motifs that
form complexes with high stability and selectivity is important both
for the understanding of biology and in the design of new materials.1
There is a particular lack of building blocks that can be used to
form acceptor, acceptor, acceptor-donor, donor, donor (AAA-
DDD) hydrogen bonding patterns, believed to be the strongest
contiguous triple hydrogen bond arrangement as a result of multiple
favorable secondary electrostatic interactions.2 Murray and Zim-
merman provided the first experimental example of such a system
when they reported that the Ka for complex 1‚2 is >105 M-1 in
1
CDCl3, as evidenced by H NMR spectroscopy (Figure 1).3 They
also found, however, that 1‚2 is chemically unstable, and the
presence of 1,8-bis(dimethylamino)naphthalene (proton sponge) was
required to prevent hydride shift from C-4 of 2 to C-10 of 1 during
their binding experiments.3,4 No attempt to quantify the Ka beyond
the limit measurable by NMR methods was reported and since these
important and seminal studies relatively little progress5 has been
made in developing less reactive AAA-DDD systems. Here we
report extremely high association constants for chemically stable
AAA-DDD and AA-DDD complexes that feature the novel and
Figure 1. AAA-DDD (1‚2,3 3‚2, and 3‚5) and AA-DDD (4‚2 and 6‚23)
readily accessible multiple hydrogen bond acceptors 3 and 4 (Figure
1).
We wondered whether the chemical stability of the Zimmerman
heterocomplexes and their 1:1 stability constants (Ka’s) in CDCl3 or CH2-
Cl2 at room temperature. Repetitions of the binding experiments for each
of 3‚2, 3‚5, and 4‚2 gave Ka’s within 10% of the values shown (the error
AAA unit might be improved by extending the anthyridine aromatic
framework. Accordingly, a pentacene analogue, 3, was prepared
in only two steps by the diiodination of 2,6-diaminopyridine
followed by a double Suzuki coupling with 2-formylphenyl boronic
acid and spontaneous cyclization and aromatization (Scheme 1).
A modified approach yielded the equivalent AA system, 4, the key
step being flash vacuum pyrolysis (FVP) of an oxime (Scheme 1,
step v). Single crystals suitable for X-ray analysis were obtained
for each of 3 and 4 from saturated CH2Cl2/MeOH solutions. The
solid state structures (Figure 2) confirmed the molecular geometries
and provided data regarding the acceptor heteroatom separations
for computer modeling of contiguous H-bond arrays with various
prospective H-bond donors.
in data-fitting for each run was <1%). Inset: In the absence of an additional
H-bonding partner, 2 exists in a 2:1 ratio of 1,4-dihydro/3,4-dihydro
tautomers at millimolar concentrations in CDCl3 at room temperature.3
Scheme 1. Synthesis of AAA and AA Systems 3 and 4a
a Reagents and conditions: (i) N-Iodosuccinamide, DMF, 86%; (ii)
2-formylphenyl boronic acid, Pd(PPh3)4, Cs2CO3, dioxane/water (1:1), 80%;
(iii) 2-formylphenyl boronic acid, Pd(PPh3)4, K2CO3, dioxane/water (1:1),
80%; (iv) MeONH2‚HCl, EtOH, 96%; (v) FVP (furnace temperature ) 700
°C, inlet temperature ) 182 °C, p ) 4.8 × 10-2 Torr, 10 min), 75%.
Experimentally, we first examined the ability of 3 to form
1
complexes with DDD partners 26 and 5 in CDCl3 by H NMR
spectroscopy, using a standard titration method7 under conditions
where the self-association of each component was negligible (Kdimer
< 20 M-1). To assess the effect of the extended aromatic system
on binding other than chemical reactivity, we also determined the
Ka of 4‚2 to compare with 6‚23 (Ka ) 3 × 103 M-1 in CDCl3).
Plots of the chemical shifts of the amino/hydroxyl groups of 2 or
5 versus the [DDD]/[AA or AAA] ratio for 4‚2 and 3‚5 showed
typical 1:1 binding isotherms (Figure 3; confirmed by Job plots,
see Supporting Information), and the data were computationally
matched to the best-fitting association constant: 4‚2, Ka ) 8 ×
104 M-1; 3‚5, Ka ) 2.4 × 104 M-1.8 However, the Job plot for 3‚2
showed a 2:1 complex at millimolar concentrations (see Supporting
Information), and curve-fitting suggested at least one association
constant beyond the range that could be reliably determined by
our NMR experiments, consistent with the Ka > 105 M-1 previously
reported3 for 1‚2.
Some of the results and observations from the 1H NMR binding
experiments deserve further comment. First, hydroxyl groups are
much poorer H-bond donors than amides, anilines, or pyrrole-like
NH’s,9 and the hydroxyl protons of 5 are also involved in
9
476
J. AM. CHEM. SOC. 2007, 129, 476-477
10.1021/ja067410t CCC: $37.00 © 2007 American Chemical Society

C O M M U N I C A T I O N S
solution of 2 (10-8 M) to 3 (initial concentration 1 × 10-9 M) and
monitoring the increase in fluorescence intensity at 410 nm (Figure
4b). Curve-fitting gave a Ka for 3‚2 of 2 × 107 M-1. A Job plot
confirmed the 1:1 stoichiometry (Figure 4c).
Geometry optimization and frequency calculations were carried
out on 3‚2, both in vacuum and in CH2Cl2 solution, at the B3LYP/
6-31G* level using the Gaussian03 program11 (see Supporting
Information). In the isolated molecules approximation the binding
free energy was underestimated by ∼10%, while in solution it was
overestimated by ∼25%. Both types of calculations showed an
extremely large electrostatic contribution to complex formation. The
simulations also suggest that the AAA-DDD complex is near
planar, particularly in solution: a tilt angle of ∼5° between the
planes of 2 and 3 in CH2Cl2 (∼21° in vacuum) provides the
optimum H-bonding arrangement and the strongest AAA-DDD
interaction.
Figure 2. X-ray crystal structures of (a) 3 and (b) 4 (C red, N blue, H
white). Nitrogen-nitrogen distances: (a) N13-N14 2.294 Å; N1-N14
2.290 Å and (b) N1-N12 2.300 Å (see Supporting Information).
In conclusion, heterocycles 3 and 4 are novel, readily accessible,
and chemically stable AA and AAA hydrogen bonding units that
form extremely strong supramolecular complexes with DDD
partners. The importance of secondary electrostatic interactions in
contiguous multipoint hydrogen bonding arrays is well-illustrated
by comparison of the relative binding strengths of AAA-DDD
complex 3‚2 (Ka ) 2 × 107 M-1 in CH2Cl2 at room temperature)
and the previously reported12 ADA-DAD complex between
1-butylthymine and 2,6-dibutyramidopyridine (Ka ) 90 M-1 in
CDCl3 at room temperature).
Figure 3. Binding isotherms using the change in chemical shift (∆δ) of
(a) the amino NH2 groups of 2 (10-4 M) upon addition of 4 and (b) the
hydroxyl groups of 5 (10-3 M) upon addition of 3. The red lines indicate
best-fitting Ka’s.8
Supporting Information Available: Experimental procedures and
spectral data for 3 and 4 and complexes 3‚2, 3‚5, and 4‚2, details of
X-ray analysis of 3 and 4, including cif files, and additional experi-
mental details on computational and complexation studies. This material
is available free of charge via the Internet at http://pubs.acs.org.
References
(1) (a) Zimmerman, S. C.; Corbin, P. S. Struct. Bonding (Berlin) 2000, 96,
63-94. (b) Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W.; Sijbesma, R.
P. Chem. ReV. 2001, 101, 4071-4097. (c) Prins, L. J.; Reinhoudt, D. N.;
Timmerman, P. Angew. Chem., Int. Ed. 2001, 40, 2383-2426. (d)
Schmuck, C.; Wienand, W. Angew. Chem., Int. Ed. 2001, 40, 4363-
4369. (e) Sherrington, D. C.; Taskinen, K. A. Chem. Soc. ReV. 2001, 30,
83-93. (f) Sijbesma, R. P.; Meijer, E. W. Chem. Commun. 2003, 5-16.
(g) Zimmerman, S. C.; Park, T. Polym. Prepr. 2005, 42, 1159-1160.
(2) (a) Jorgensen, W. L.; Pranata, J. J. Am. Chem. Soc. 1990, 112, 2008-
2010. (b) Pranata, J.; Wierschke, S. G.; Jorgensen, W. L. J. Am. Chem.
Soc. 1991, 113, 2810-2819.
Figure 4. (a) UV/vis spectra in CH2Cl2 at 293 K upon the addition of 2
(0f1.2 equiv) to 3 (1 × 10-5 M). The arrow indicates the change in
absorption at 390 nm with increasing 2. (b) Fluorescence intensity at 410
nm in CH2Cl2 at 293 K upon the addition of 2 (0f3 equiv) to 3 (1 × 10-9
M). (c) Job plot under similar conditions to (b).
(3) (a) Murray, T. J.; Zimmerman, S. C. J. Am. Chem. Soc. 1992, 114, 4010-
4011. (b) Zimmerman, S. C.; Murray, T. J. Tetrahedron Lett. 1994, 35,
4077-4080.
(4) The reductive instability of 1 can also be overcome by using a protonated
2,6-aminopyridine derivative as the DDD partner. See: Bell, D. A.;
Anslyn, E. V. Tetrahedron 1995, 51, 7161-7172.
(5) (a) Sugiyama, Y.; Adachi, K.; Kawata, S.; Kumagai, H.; Inoue, K.; Katada,
M.; Kitagawa, S. CrystEngComm 2000, 2, 174-176. (b) Sugiyama, Y.;
Adachi, K.; Kabir, M. K.; Kitagawa, S.; Suzuki, T.; Kaizaki, S.; Kawata,
S. Mol. Cryst. Liq. Cryst. 2002, 379, 419-424. (c) Adachi, K.; Sugiyama,
Y.; Yoneda, K.; Yamada, K.; Nozaki, K.; Fuyuhiro, A.; Kawata, S.
Chem.sEur. J. 2005, 11, 6616-6628.
intramolecular H-bonding. It is therefore somewhat remarkable that
the Ka (CDCl3, room temperature) for 3‚5 is sub-millimolar.10
Second, the use of 2 in the binding experiments is complicated by
its tautomerism (see Figure 1 inset). Murray and Zimmerman
reported3 that 10 equiv of 1 was required to fully convert 2 into
the 1,4-dihydro form involved in DDD H-bonding. In contrast, in
our NMR titration experiments only 0.5 equiv of 3 proved sufficient
to convert the initial 2:1 ratio of the 1,4-dihydro/3,4-dihydro forms
of 2 to >98:2 (see Supporting Information). A further indication
of the powerful hydrogen bond accepting ability of these new
heterocycles is seen in the direct comparison of the AA-DDD
complexes in CDCl3 at room temperature; 4‚2 is at least 20 times
more strongly bound than 6‚2 (Figure 1).
We next investigated the binding in complex 3‚2 by UV/vis and
fluorescence spectroscopy. Upon addition of 2 to 3 (ca. 10-5 M,
CH2Cl2, 293 K), the absorption intensity at 395 nm increased with
a clear isosbestic point at 390 nm, suggesting a 1:1 binding mode
in this concentration range (Figure 4a). Fluorescence titrations (3
has a fluorescence quantum yield of 0.94 in CH2Cl2, while 2 is
nonfluorescent) were performed in CH2Cl2 at 293 K by adding a
(6) Compound 2 was prepared according to Murray, T. J.; Zimmerman, S.
C. Tetrahedron 1995, 51, 635-648.
(7) Connors, K. A. Binding Constants: The Measurement of Molecular
Complex Stability; Wiley-Interscience: New York, 1987.
(8) GAs-Fit (www.djurdjevic.org.uk/software/GAsFit): A custom-written
program, suitable even for large binding constants, that uses an evolution-
ary algorithm to solve the standard equations for titration methods (see
ref 7). In tests, for data in the Ka 102-105 M-1 range, GAs-Fit gave similar
results to the widely used binding constant determination program available
from H.-J. Schneider’s group (www.uni-saarland.de/fak8/schneider/Links/
download.html).
(9) Hunter, C. A. Angew. Chem., Int. Ed. 2004, 43, 5310-5324.
(10) There is no evidence of deprotonation of 5 by 3 in the UV/vis spectra.
(11) Frisch, M. J.; et al. Gaussian 03, revision C.02; Gaussian, Inc.: Wall-
ingford, CT, 2004.
(12) Hamilton, A. D.; Van Engen, D. J. Am. Chem. Soc. 1987, 109, 5035-5036.
JA067410T
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 477