Peptide-based molecular shuttles

Full text of DOI:10.1021/ja971224t

11092
J. Am. Chem. Soc. 1997, 119, 11092-11093
Peptide-Based Molecular Shuttles
Alexander S. Lane, David A. Leigh,* and Aden Murphy
Department of Chemistry, UniVersity of Manchester
Institute of Science and Technology, SackVille Street
Manchester M60 1QD, United Kingdom
-d
ReceiVed April 17, 1997
ReVised Manuscript ReceiVed September 6, 1997
The inherent restrictions in rotational and translational
freedom imposed on the components of mechanically-inter-
locked molecules1 make them particularly attractive architectures
for precisely controlling the positioning of functional units/
substituents with the possibility of switching their relative
separation and orientation.2,3 Such control has been elegantly
demonstrated through π-donor-acceptor interactions in the
“molecular shuttles”2 initially developed by Stoddart et al. and
metal ion coordination in the catenates3a-c and pseudorotaxanes3d
prepared by Sauvage and co-workers. However, although hy-
drogen bonding has been used to synthesize a variety of simple
and polymeric rotaxanes,4 translocation of macrocycles between
specific sites (“stations”) in the threads of hydrogen bond-
assembled systems has not previously been demonstratedsi.e.,
such rotaxanes have not been elaborated into shuttles.
Figure 1. Controlling the position of the macrocycle in multistationed
peptide-based molecular shuttles. In CDCl3 or CD2Cl2 the macrocycle
shuttles between degenerate hydrogen-bonding stations A and A′. In
DMSO-d6 the macrocycle is located at polarphobic station B.
regions (to provide distinct polarphobic stations) in a thread
could provide a simple route to multistationed, hydrogen bond-
assembled molecular shuttles whose structure and dynamics are
governable by judicial choice of their operating environment
(Figure 1).
Compounds 1a-c were synthesized in two steps (Scheme
1) from a commercially available glycylglycine ethyl ester salt
to give threads containing two identical (i.e. degenerate) peptide
stations A and A′ separated by a third, lipophilic, station B,
which in the case of 1c contained a sulfur atom to allow further
derivatization. Treatment of each thread with 4 equiv of
isophthaloyl dichloride and p-xylylenediamine (Et3N, CHCl3)
gave a mixture of unconsumed thread, the corresponding [2]-
rotaxane (2a-c, produced in 30, 28, and 36% yields, respec-
tively), small amounts of the [3]rotaxanes (3, 2, and 4%), and
the octaamide [2]catenane,7 all of which could be conveniently
separated by column chromatography.
The solvent dependent translational isomerism of the [2]-
rotaxanes is apparent from comparison of the 1H NMR spectra
of the rotaxanes and threads in DMSO-d6 and CDCl3 (e.g., 1a
and 2a, Figure 2). In CDCl3 the macrocycle shuttles between
the two degenerate hydrogen bonding (peptide) stations A and
A′. The occupied A/A′ station protons (Ha-e, Figure 2b) are
shielded by the xylylene rings of the macrocycle and experience
significant shifts consistent with the four-point intercomponent
hydrogen bonding motif shown in Scheme 1, but reduced in
magnitude compared to those found4i for simple glycylglycine
rotaxanes as a result of the shuttle spending only half its time
Peptide rotaxanes4i can be prepared by the condensation of
appropriate aromatic diacid chlorides and benzylic diamines in
the presence of dipeptide derivatives which template the for-
mation of benzylic amide macrocycles around them. In non-
polar solvents the hydrogen bonding motif responsible for pep-
tide rotaxane formation is maintained, but in polar solvents the
intramolecular hydrogen bonding between isophthaloyl5 ben-
zamide macrocycles and the peptide is “switched off” leading
to nonspecific location of the macrocycle along the peptide back-
bone. Medium effects6 often influence the stability of inter-
molecular interactions6a and have previously been shown to alter
intramolecular translational isomerism in catenanes.6b,c It
therefore seemed feasible that combining two (or more)
hydrogen bonding peptide units with additional lipophilic
(1) (a) Schill, G. Catenanes, Rotaxanes and Knots; Academic Press: New
York, 1971. (b) Amabilino, D. B.; Stoddart, J. F. Chem. ReV. 1995, 95,
2725-2828. (c) Sauvage, J.-P. Acc. Chem. Res. 1990, 23, 319-327. (d)
Gibson, H. W.; Bheda, M. C.; Engen, P. T. Prog. Polym. Sci. 1994, 19,
843-945.
(2) For state-of-the-art accounts see: (a) Anelli, P. L.; Asakawa, M.;
Ashton, P. R.; Bissell, R. A.; Clavier, G.; Go´rski, G.; Kaifer, A. E.; Langford,
S. J.; Mattersteig, G.; Menzer, S.; Philp, D.; Slawin, A. M. Z.; Spencer,
N.; Stoddart, J. F.; Tolley, M. S.; Williams, D. J. Chem. Eur. J. 1997, 3,
1113-1135 and references therein. (b) Benniston, A. C. Chem. Soc. ReV.
1996, 25, 427-435.
(3) (a) Amabilino, D. B.; Sauvage, J.-P. Chem. Commun. 1996, 2441-
2442. (b) Ca´rdenas, D. J.; Livoreil, A.; Sauvage, J.-P. J. Am. Chem. Soc.
1996, 118, 11980-11981. (c) Baumann, F.; Livoreil, A.; Kaim, W.;
Sauvage, J.-P. Chem. Commun. 1997, 35-36. (d) Collin, J.-P.; Gavina˜, P.;
Sauvage, J.-P. Chem. Commun. 1996, 2005-2006.
at each station. Protons on the lipophilic station B (Hf-i
)
experience little shielding in CDCl3 compared to the analogous
protons on the thread, indicating that the macrocycle spends
no appreciable time on station B but just passes through on its
way between A and A′. The simplicity of the rotaxane spectrum
(one set of resonances for the station A and A′ protons and HE
appearing as a doublet rather than the ABX system observed
in unsymmetrical peptide rotaxanes) shows that both spinning
of the macrocycle about the thread and shuttling of the
macrocycle between stations A and A′ is rapid on the NMR
time scale at room temperature. However, cooling below the
coalescence temperature for the shuttling process (Figure 2c)
freezes the macrocycle at a single peptide station (A or A′) and
allows observation of discrete resonances for both occupied and
unoccupied stations (unprimed and primed labels, respectively).
(4) (a) Mock, W. L.; Irra, T. A.; Wepsiec, J. P.; Adhya, M. J. Org. Chem.
1989, 54, 5302-5308. (b) Gibson, H. W.; Marand, H. AdV. Mater. 1993,
5, 11-21. (c) Kolchinski, A. G.; Busch, D. H.; Alcock, N. W. J. Chem.
Soc., Chem. Commun. 1995, 1289-1291. (d) Ashton, P. R.; Glink, P. T.;
Stoddart, J. F.; Tasker, P. A.; White, J. P.; Williams, D. J. Chem. Eur. J.
1996, 2, 729-736. (e) Whang, D.; Jeon, Y.-M.; Heo, J.; Kim, K. J. Am.
Chem. Soc. 1996, 118, 11333-11334. (f) Whang, D.; Kim, K. J. Am. Chem.
Soc. 1997, 119, 451-452. (g) Vo¨gtle, F.; Dunnwald, T.; Schmidt, T. Acc.
Chem. Res. 1996, 29, 451-460. (h) Johnston, A. G.; Leigh, D. A.; Murphy,
A.; Smart, J. P.; Deegan, M. D. J. Am. Chem. Soc. 1996, 118, 10662-
10663. (i) Leigh, D. A.; Murphy, A.; Smart, J. P.; Slawin, A. M. Z. Angew.
Chem., Int. Ed. Engl. 1997, 36, 728-732. As Gibson has pointed out
[Gibson, H. W. Rotaxanes. In Large Ring Molecules; Semlyen, J. A., Ed.;
J. Wiley and Sons: New York, 1996] intercomponent hydrogen bonding
interactions were probably important in the synthesis of the early rotaxanes
prepared by Zilkha et al. [Agam, G.; Gravier, D.; Zilkha, A. J. Am. Chem.
Soc. 1976, 98, 5206-5214. Agam, G.; Zilkha, A. J. Am. Chem. Soc. 1976,
98, 5214-5216].
1
Calculations8 based on variable-temperature H NMR data
give a ∆Gq for shuttling in halogenated solvents (CDCl3 or CD2-
(6) (a) For examples based on pseudorotaxanes see: Asakawa, M.; Iqbal,
S.; Stoddart, J. F.; Tinker, N. D. Angew. Chem., Int. Ed. Engl. 1996, 35,
976-978 and references therein. (b) Ashton, P. R.; Blower, M.; Philp, D.;
Spencer, N.; Stoddart, J. F.; Tolley, M. S.; Ballardini, R.; Ciano, M.; Balzani,
V.; Gandolfi, M. T.; Prodi, L.; McLean, C. H. New J. Chem. 1993, 17,
689-695. (c) Leigh, D. A.; Moody, K.; Smart, J. P.; Watson, K. J.; Slawin,
A. M. Z. Angew. Chem., Int. Ed. Engl. 1996, 35, 306-310.
(5) The same is not true for pyridine-2,6-dicarbamidobenzylic macro-
cycles where intercomponent hydrogen bonding is maintained in polar
solvents [ref 4i and: Johnston, A. G.; Leigh, D. A.; Nezhat, L.; Smart, J.
P.; Deegan, M. D. Angew. Chem., Int. Ed. Engl. 1995, 34, 1212-1216].
(7) Johnston, A. G.; Leigh, D. A.; Pritchard, R. J.; Deegan, M. D. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1209-1212.
S0002-7863(97)01224-9 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 45, 1997 11093
Scheme 1a
Figure 3. Controlling the dynamics of macrocycle translation in
multistationed peptide-based molecular shuttles. Addition of CD3OD
a (i) Ph2CHCOCl, Et3N, CHCl3, 97%. (ii) diol, (Bu2ClSn)2O, ∆,
PhMe, 86% 1a, 74% 1b, 80% 1c. (iii) Isophthaloyl dichloride,
p-xylylenediamine, Et3N, CHCl3, 30% 2a, 28% 2b, 36% 2c.
q
to a solution of 2c in CD2Cl2 decreases ∆Gshuttling by up to 2.3 kcal
mol-1. Tosyl imination of 2c provides a chemically-reversible steric
barrier to shuttling.
Hf-i on the lipophilic station B are shielded by as much as 0.85
ppm with respect to the thread, with those in the center of the
station experiencing the greatest shifts. This, of course, contrasts
to the behavior of simple peptide rotaxanes where no specific
polarphobic station is present when addition of DMSO leads to
nonspecific location of the macrocycle along the thread.
In addition to being able to control the position of the
macrocycle in peptide-based molecular shuttles, the sensitivity
of the intercomponent hydrogen bond strengths can be exploited
to control the rates of shuttling in nonpolar solvents (Figure
3). Addition of small quantities (0.1-5% by volume) of CD3-
OD to solutions of 2a-c in CDCl3 or CD2Cl2 reduces the
strengths of the intramolecular hydrogen bonds, decreasing
q
∆Gshuttling and speeding up the shuttling of the macrocycle
between the two peptide stations. Increases in shuttling rate
occur after addition of as little as 0.1% CD3OD (which doubles
the rate of shuttling of 2c in CD2Cl2 from 8 s-1 to 16 s-1 at
203 K) and continue up to 5% CD3OD, which reduces
∆Gshuttlingq of 2c in CD2Cl2 by 2.3 kcal mol-1, equivalent to an
increase in the rate of shuttling from 8 s-1 to 2300 s-1 at 203
K (62000 s-1 to >3 million s-1 at 298 K).9
Chemical derivatization was also investigated as a means of
governing the dynamics of these molecular shuttles (Figure 3).
Somewhat surprisingly, oxidation (MCPBA, CH2Cl2) of 2c to
the corresponding sulfoxide (0 °C, 1 h, 81%) or sulfone (room
temperature, 6 h, 92%) has no effect on the rate of shuttling.
Clearly, once the hydrogen bonding of the macrocycle to an
individual station is broken, the distance it must travel to the
next station is much more influential on the rate of shuttling
than steric barriers that are not quite large enough to completely
prevent passage of the macrocycle. The shuttling process in
2c can be stopped, however, by introduction (TsNCl, Bu4NOH,
CH2Cl2, 2 h, 93%)10 of a bulky tosyl group orthogonal to the
vector of the thread. Shuttling between the peptide stations can
subsequently be reestablished by reduction (P4S10, CH2Cl2, 4
h, 100%)11 of the imine back to the sulfide, 2c.12
Figure 2. 300-MHz 1H NMR spectra of thread and shuttle (0.02 mmol
dm-3): (a) 1a CDCl3, 298 K; (b) 2a CDCl3, 298 K; (c) 2a CDCl3, 234
K; (d) 1a DMSO-d6, 298 K; (e) 2a DMSO-d6, 298 K.
Cl2) of 11.2 ( 0.3 kcal mol-1 for 2a, 12.4 ( 0.3 kcal mol-1
for 2b, and 10.9 ( 0.3 kcal mol-1 for 2c, which correspond to
shuttling rates of 37000, 5200, and 62000 s-1, respectively, at
298 K. The degeneracy of A and A′ means that once the
intercomponent hydrogen bonding is broken shuttling of the
macrocycle between them is essentially a one-dimensional
diffusion process along the vector of the thread. Thus the
increased activation energy cost (1.2 kcal mol-1) for 2b over
2a actually reflects the increased distance that the shuttle must
travel between the peptide stations!
Acknowledgment. This work was supported by the EPSRC IPS
managed program.
In response to a major change in the polarity of the
environment of the shuttles (changing from halogenated solvents
to DMSO-d6),9 the macrocycle stops shuttling between stations
A and A′ and instead spends nearly all of its time on the
lipophilic station B (Figure 1). This is evident from comparison
of the H NMR spectra of the thread 1a and rotaxane 2a in
DMSO-d6 (Figure 2, parts d and e). The protons associated
Supporting Information Available: Experimental details for the
synthesis of 2a, 2b, and 2c; NMR data for solvent dependent
translational isomerism of 2b and stopping of shuttling by imination
of 2c (8 pages). See any current masthead page for ordering and
Internet access instructions.
1
JA971224T
with the peptide units (Ha-e) undergo negligible shielding while
(9) Similar decreases in ∆Gq (up to 2.1 and 2.3 kcal mol-1, respectively)
were found for addition of 0.1-5% CD3OD to 2a and 2b. Addition of more
than 5% CD3OD begins to decrease the integrity of the positioning of the
macrocycle at just the peptide stations. In addition to DMSO-d6, the
macrocycle is positioned primarily at the polarphobic station in tetra-
methylurea, tetramethylene sulfone, and dimethylformamide.
(10) Yamamoto, T.; Yoshida, D.; Hojyo, J.; Terauchi, H. Bull. Chem.
Soc. Jpn. 1984, 57, 3341-3342.
(8) The rates of shuttling were calculated at the coalescence temperature
(Tc) by the standard NMR method [Sandstro¨m, J. Dynamic NMR Spectro-
scopy; Academic Press: London, 1982] and at temperatures lower than Tc
by selective inversion recovery (SIR) experiments where an 180° pulse is
given to a resonance and the rate of transfer to the signal it is in slow
exchange with (i.e. the rate of macrocycle shuttling) is determined directly.
The Eyring equation was used to estimate ∆Gq and extrapolate the rate of
shuttling to other temperatures [Sutherland, I. O. Annu. Rep. NMR Spectrosc.
1971, 4, 71-235].
(11) Still, I. W. J.; Turnbull, K. Synthesis 1978, 540-541.
(12) For the first example of a polymeric hydrogen bond-assembled
molecular shuttle see: Gong, C.; Gibson, H. W. Angew. Chem. In press.