Highly substituted ter-cyclopentanes as receptors for lipid A [1]

Full text of DOI:10.1021/ja003712y

5810
J. Am. Chem. Soc. 2001, 123, 5810-5811
extensive molecular mechanics simulations,11 which suggested
that such structures would be conformationally rigid, and able to
present functionality in a spatially defined manner. For example,
molecular models of 4 docked to lipid A suggested good surface
complimentarity between the two structures. It was anticipated
that variation of the acyl groups (X, Y) would allow for tuning
of the structure to optimize affinity; in particular, we anticipated
that employing amino acids as the “Y” groups would improve
the solubility of 4 in water, while positioning amino groups
optimally for hydrogen bonding to occur between 4 and the
phosphate, amide, and ester moieties of 2.
Highly Substituted ter-Cyclopentanes as Receptors
for Lipid A
Robert D. Hubbard,‡ Scott R. Horner,§,† and
Benjamin L. Miller*,†,‡
Department of Chemistry, Department of Biophysics, and
The Center for Future Health, UniVersity of Rochester
Rochester, New York 14627
ReceiVed October 18, 2000
ReVised Manuscript ReceiVed January 18, 2001
The design and synthesis of simple organic receptors capable
of high-affinity binding to carbohydrates in aqueous solution
presents a substantial challenge to our understanding of molecular
structure and the processes which underly molecular recognition.1
In essence, a simple sugar like glucose (1) is no more than an
oriented collection of hydroxy groups, differing little from bulk
water. Because of this, the vast majority of receptors for
carbohydrates reported thus far have been studied only in organic
solvents,2 since competition between the solvent and the targeted
carbohydrate for binding is not as much of a problem. Of the
carbohydrate receptors examined in aqueous solution thus far,
the best bind simple sugars with affinities (dissociation constants,
KD) in the millimolar range.3
As part of a continuing program in the development of new
methods for the recognition of biologically significant structures,4
we became interested in the possibility of employing highly
substituted, stereoregular oligomers of cyclopentane as receptors
for simple carbohydrates, liposaccharides, and oligosaccharides.
In particular, lipid A (2) represented an attractive target for
binding, both because of its structural features and due to its
biological role as the conserved portion of lipopolysaccharide
(LPS). LPS is a primary constituent of the outer cellular membrane
of Gram-(-) bacteria,5 and is perhaps better known as bacterial
endotoxin, the causative agent of sepsis.6 Antibiotics such as the
polymyxin family of cyclic peptides7 (i.e. polymyxin B, 3) derive
their effectiveness against sepsis through an ability to bind to
and neutralize bacterial endotoxin, primarily mediated by interac-
tions with lipid A.8 However, polymyxin is problematic as a
therapeutic agent both because of its complex structure (rendering
it impractical to synthesize in quantity) and because of its side
effects (acute renal toxicity,9 among others). Therefore, the
development of new lipid A-binding, endotoxin-neutralizing
compounds is a problem of potential medical significance10 as
well as a fundamental problem in molecular recognition. Our
contention that highly substituted oligocyclopentanes such as 4
might be attractive receptors for carbohydrates was based on
Our synthesis of 4 was designed with an eye toward both
producing its three rings and 10 stereocenters in as few steps as
possible and developing methodology that would be amenable
to both large-scale and combinatorial library synthesis. As shown
in Scheme 1, we began by converting norbornene (5) to the known
cis-cyclopentane-1,3-dialdehyde12 (6) via Sharpless dihydroxyl-
ation13 followed by sodium periodate-metiated oxidative cleavage
of the diol.14 Subsequent treatment of the dialdehyde with
Horner-Wadsworth-Emmons reagent 7 under standard condi-
tions (potassium tert-butoxide as base in THF) provided the bis
R,â-unsaturated ester 8 in 90% yield and >20:1 EE:EZ selectivity.
Our initial attempts to employ 8 as the dienophile in a bidirectional
Diels-Alder reaction with cyclopentadiene using standard Lewis
acid catalysts were unsuccessful; however, a hybrid 0.5:0.05
AlCl3-Al(CH3)3 catalyst system we have previously described15
provided for the smooth conversion of 8 to double Diels-Alder
cycloadduct 9 in 87% yield and 94:6 endo,endo-endo,exo
selectivity.
‡ Department of Chemistry.
With all of the carbocyclic skeleton and stereogenic centers of
4 installed, it remained for us to set the peripheral functionality.
Reduction of the cycloadduct 9 with LiAlH4 provided a diol in
53% yield, which was derivatized with benzoyl chloride to yield
§ Department of Biophysics.
† The Center for Future Health.
(1) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. 1999, 38, 2978-
2996. The synthesis of carbohydrate mimetics (as opposed to receptors) has
also been an active area of research; for a review, see: Sears, P.; Wong, C.-
H. Angew. Chem., Int. Ed. 1999, 38, 2300-2324.
(10) For other examples of lipid A-binding compounds and their develop-
ment as therapeutic agents, see: (a) Rustici, A.; Velucchi, M.; Faggioni, R.;
Sironi, M.; Ghezzi, P.; Quataert, S.; Green, B.; Porro, M. Science 1993, 259,
361-365. (b) Vaara, M.; Porro, M. Antimicrob. Agents Chemother. 1996,
40, 1801-1805. (c) Li, C.; Budge, L. P.; Driscoll, C. D.; Willardson, B. M.;
Allman, G. W.; Savage, P. B. J. Am. Chem. Soc. 1999, 121, 931-940.
(11) Structures were constructed, minimized, and docked using Macromodel
6.0 (Schroedinger, Inc.) and the AMBER* force field; details of our
computational studies of the conformational properties of substituted oligo-
cycloalkanes will be disclosed in due course.
(2) For a recent example, see: Mazik, M.; Bandmann, H.; Sicking, W.
Angew. Chem., Int. Ed. 2000, 39, 551-554.
(3) (a) Kobayashi, K.; Asakawa, Y.; Kato, Y.; Aoyama, Y. J. Am. Chem.
Soc. 1992, 114, 10307. (b) Yanagihara, R.; Aoyama, Y. Tetrahedron Lett.
1994, 35, 9725. (c) Poh, B.-L.; Tan, C. M. Tetrahedron 1993, 49, 9581.
(4) Klekota, B.; Miller, B. L. Tetrahedron 1999, 55, 11687-11697.
(5) Young, L. S.; Martin, W. J.; Meyer, R. D.; Weinstein, R. J.; Anderson,
E. T. Ann. Intern. Med. 1977, 86, 456-471.
(6) (a) Raetz, C. R. H. Annu. ReV. Biochem. 1990, 59, 129-170. (b)
Ulevitch, R. J.; Tobias, P. S. Curr. Opin. Immunol. 1994, 6, 125-130.
(7) Chapman, T. M.; Golden, M. R. Biochem. Biophys. Res. Commun. 1972,
46, 2040-2047.
(12) Wiberg, K. B.; Saegebarth, K. A. J. Am. Chem. Soc. 1957, 79, 2822-
2824.
(13) Becker, H.; Soler, M. A.; Sharpless, K. B. Tetrahedron 1995, 51,
1345-1376.
(8) (a) Morrison, D. C.; Jacobs, D. M. Immunochemistry 1976, 13, 813-
818. (b) David, S. A.; Balasubramanian, K. A.; Mathan, V. I.; Balaram, P.
Biochim. Biophys. Acta 1992, 1165, 145-152.
(14) 6 has also been prepared from norbornene via ozonolysis (Trautmann,
W.; Musso, H. Chem. Ber. 1981, 114, 982-989); however, in our hands this
procedure was generally lower yielding than the dihydroxylation-oxidative
cleavage method.
(9) (a) Craig, W. A.; Turner, J. H.; Kunin, C. M. Infect. Immun. 1974, 10,
287. (b) D. Rifkind J. Bacteriol. 1967, 93, 1463. Kunin, C. M. J. Infect.
Diseases 1970, 121, 55-64.
(15) Hubbard, R. D.; Miller, B. L. J. Org. Chem. 1998, 63, 4143-4146.
10.1021/ja003712y CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/26/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 24, 2001 5811
Scheme 1a
Figure 1. Job plot of 14 and lipid A.
solutions of 13 or 14. While some absorbance changes were
observed in both cases, these were very weak (on the order of
10% overall), and did not display a standard saturation profile.20
This suggests that while there may be some interaction between
12-14 and the phospholipid tails of lipid A, it is likely
nonspecific, and only weakly contributes to the overall affinity
constant. To ensure that interactions were not simply due to the
amino acid-derived functionality of 12-14, we also examined
the binding of lipid A to the methyl ester of tryptophan. In aqueous
solution, we saw a linear increase in the tryptophan indole ring
absorbance on addition of increasing concentrations of lipid A.
However, this increase was not saturable at the concentrations
tested (>20 µM in 2), indicating a strictly nonspecific interaction.
Finally, further verification of a strong interaction between 14
and lipid A was provided by titrating diphosphoryl lipid A into
a 1.5 mM solution of 14 in 95% D2O, 5% DMSO, and recording
a Conditions: (a) K3Fe(CN)6, K2OsO4‚2H2O, quinuclidine, K2CO3,
t-BuOH-H2O 1:1; (b) NaIO4, THF-H2O 3:1, 0 °C to room temperature
(63%); (c) (iPrO)2P(O)CH2C(O)CH2CH2Ph, t-BuOK, THF, 0 °C to room
temperature, 2 h, then: 7, -78 to 4 °C (90%); (d) (CH3)3Al (0.05 equiv)
10 min, then AlCl3 (0.50 equiv), CH2Cl2, 0 °C, 10 min, then: cyclopen-
tadiene (10 equiv), 4 °C (87%); (e) LiAlH4 (6.0 equiv), THF, room
temperature; (f) Bz-Cl (3.6 equiv), Et3N (4.0 equiv), CH2Cl2, room
temperature (50%, two steps); (g) O3, CH3OH/Ch2Cl2 1:1, -78 °C, then
NaBH4 (10 equiv), 0 °C to room temperature; (h) Boc-Gly-OH, Boc-
Phe-OH, or Boc-Trp-OH (5.6 equiv), DCC (6.4 equiv), DMAP (3.7
equiv), CH2Cl2-DMF 7:3, room temperature, 18 h; (i) CF3CO2H-
Et3SiH-CH2Cl2 20:10:70, 12 h.
1
the 500 MHz H NMR spectrum at each point. Addition of as
bis-benzoate 10 (94%). Ozonolysis of 10 followed by a reductive
workup with NaBH4 gave a tetrol, 11, in 60% yield. This com-
pound was acylated with N-Boc-glycine, N-Boc-phenylalanine,
or N-Boc-tryptophan16 using DCC as the coupling reagent in the
presence of DMAP. Deprotection of the amino acids to provide
12, 13, or 14 was accomplished by using standard conditions.
Lipid A is a very weak absorber of ultraviolet light; therefore,
we measured the affinities of 12-14 for lipid A by UV titration.
Both 13 and 14 bind to lipid A with dissociation constants of
comparable magnitude (587 and 592 nM, respectively, in
phosphate-buffered saline, pH 7.4) to the reported polymyxin-
lipid A interaction.17 Providing further indication of the importance
of the amino acid side chains of 4 in the recognition process, a
saturation point for binding of lipid A to 12 was not reached at
concentrations below the critical micelle concentration (cmc) for
lipid A,18 although changes in the UV absorbance clearly indicated
the presence of some interaction.
few as 0.1 molar equiv of diphosphoryl lipid A causes the appear-
ance of a new peak downfield in the aromatic region of the
spectrum, and subtle changes in the fine structure of resonances
throughout the spectrum. As additional amounts of lipid A were
added, resonance intensity decreased, culminating in substantial
broadening of all resonances and the appearance of a flocculent
precipitate in the NMR tube. While these data unfortunatey
suggest that it will not be possible to obtain high-resolution
solution structural information about the complex formed between
14 and lipid A in water, they are consistent with results obtained
by others in the examination of lipid A-binding compounds.21
In conclusion, the highly substituted oligocyclopentanes pre-
sented herein represent a new class of compounds which
demonstrate high levels of affinity for lipid A in aqueous solution.
Efforts to examine the solution structures of 13 and 14, and toward
the development of a more complete structural understanding of
the factors allowing 13 and 14 to function as high-affinity
receptors for lipid A, are currently underway. Given the clinical
importance of lipopolysaccharide as the causative agent of
bacterial sepsis, the ability of 12-14 to neutralize bacterial
endotoxin is currently under examination.
The method of continuous variations19 was employed to further
examine the binding of 14 to lipid A in aqueous solution. Two
inflection points are clearly observable for 14 binding to 2 (Figure
1), the first indicating the formation of a 1:1 14-2 complex. The
presence of the second inflection point, indicative of a 1:2 14-2
complex, is somewhat puzzling; however, it is possible that this
is in part due to changes in the aggregation state of 2 as its
concentration is increased over the course of the experiment. To
further clarify the mode of interaction between 12-14 and lipid
A, we titrated dipalmitoyl phosphatidylcholine (DPPC) into
Acknowledgment. This research was supported by a grant from the
W. M. Keck foundation. The authors gratefully acknowledge Professors
Thomas R. Krugh and Joseph P. Dinnocenzo for enlightening discussions
during the course of this work.
Supporting Information Available: Selected spectra for UV titration
data, spectra, NMR titration of 2 into 14, and Skatchard analyses (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
(16) We chose phenylalanine and tryptophan to derivatize 11 because of
the well-known preponderance of aromatic residues in the substrate binding
sites of carbohydrate-binding proteins; for example, see: Quiocho, F. A.;
Spurlino, J. C.; Rodseth, L. E. Structure 1997, 5, 997-1015.
(17) David, S. A.; Bechtel, B.; Annaiah, C.; Mathan, V. I.; Balaram, P.
Biochim. Biophys. Acta 1994, 1212, 167-175.
(18) (a) Hofer, M.; Hampton, R. Y.; Raetz, C. R. H.; Yu, H. Chem. Phys.
Lipids 1991, 59, 167-181. (b) Aurell, C. A.; Wistrom, A. O. Biochem.
Biophys. Res. Commun. 1998, 253, 119-123.
(19) (a) Job, P. Compt. Rend. 1925, 180, 928. (b) Blanda, M. R.; Horner,
J. H.; Newcomb, M. J. Org. Chem. 1989, 54, 4626.
JA003712Y
(20) An accurate assessment of the affinity (if any) of 13 or 14 for DPPC
is complicated in part by the extraordinarily low critical micelle concentration
of DPPC in water (46 nM at 20 °C); see: Smith, R.; Tanford, C. J. Mol. Biol.
1972, 67, 75-83.
(21) David, S. A.; Bechtel, B.; Annaiah, C.; Mathan, V. I.; Balaram, P.
Biochim. Biophys. Acta 1994, 1212, 167-175.