Lingling Peng1, Fanyang Mo, Qinghai Zhang. 1. Department of Integrative Structural and Computational Biology, The Scripps Research Institute , La Jolla, California 92037, United States.
Abstract
We describe cholate-based cage amphiphiles with a unique architecture that combines elements of structural rigidity and flexibility. The cage compounds are built by extending and bridging three polar chains underneath the concave steroid rings of cholate and capping with another rigid, symmetrically trifunctionalized cyanuric acid moiety. The connecting chains are varied to include, for instance, oligo(ethylene glycol) or chains containing 1,2,3-triazole units to present flexibility in the chemical and structural space and potentially deliver functional molecules for molecular recognition applications.
We describe cholate-based cage amphiphiles with a unique architecture that combines elements of structural rigidity and flexibility. The cage compounds are built by extending and bridging three polar chains underneath the concave steroid rings of cholate and capping with another rigid, symmetrically trifunctionalized cyanuric acid moiety. The connecting chains are varied to include, for instance, oligo(ethylene glycol) or chains containing 1,2,3-triazole units to present flexibility in the chemical and structural space and potentially deliver functional molecules for molecular recognition applications.
Cholic acid is a principal component
of bile acids found in the bile of mammals and other vertebrates that
emulsify dietary fats to promote their digestion and absorption. A
representative of facial amphiphiles, cholate features three small
α-hydroxy groups positioned on one side of the rigid, slightly
concave steroid surface. Natural bile acids and their analogue molecules
have long been used pharmacologically for the dissolution of gallstones[1,2] and in membrane biochemistry for their capacity to partition into
cellular membranes and to solubilize lipids and membrane proteins.[3,4] Additionally, considerable interest has focused on the design of
bile acid-templated “smart materials”, in forms such
as tandem oligomers or dendrimers, for the purpose of delivering biologics,
drugs, and other guest molecules in solution or across cell membranes.[5−17] Of these designs, macrocyclic architectures of cholic/deoxycholic
acid have been created via connections through the terminal 24-carboxylate
and the hydroxyl groups (or modifier groups). There is a case report
in which small cage compounds are built through cyclodimerization
of chenodeoxycholic acid.[18]While
many molecular designs based on bile acids take advantage
of their unique facial amphiphilicity and assembly properties, novel
compounds can also be conceived by appending additional polar groups
beneath the steroid scaffold to increase facial amphiphilicity and
introduce additional functions.[19−22] We report here the synthesis of size-tunable cage
compounds starting from a single cholate molecule and bridging three
polar chains that extend from the 3α,7α,12α-OH groups
(Figure 1). Our exploration of such a molecular
design originated with our study of facial amphiphiles for membrane
protein structural biology applications. Our group has recently developed
new steroid- and peptide-based facial amphiphiles to sequester the
hydrophobic surfaces of integral membrane proteins.[23−25] Variation of
polar segments of the cholate-based amphiphiles significantly impacts
membrane protein crystal growth, potentially by mediating surface
contacts of protein–detergent complexes (PDCs). To diversify
the polar functionalities of facial amphiphiles and in an attempt
to restrict the flexibility of these polar segments which may favor
ordered PDC assembly during crystal growth, we set out to investigate
the possibility of bridging all three flexible polar chains of cholate-derived
amphiphiles in cagelike macrocycles. To our knowledge, this type of
structure has not been explored in the literature.
Figure 1
Drawings of (a) cholic
acid structure, (b) cholic acid derived
facial amphiphiles with three introduced polar side chains, and (c)
a new cage compound design in which the flexible side chains underneath
the steroid ring of cholate are bridged.
Drawings of (a) cholic
acid structure, (b) cholic acid derived
facial amphiphiles with three introduced polar side chains, and (c)
a new cage compound design in which the flexible side chains underneath
the steroid ring of cholate are bridged.To cyclize the polar residues attached to −OH positions
of cholate, we attempted a strategy of joining the chains in one step
using a trifunctionalized moiety, such as cyanuric acid (Figure 2a). Initial success was achieved by treatment of
the bromide substituent of cholane precursor 1, which
we used in previous amphiphile synthesis,[24] with 4 equiv of DBU and 15 equiv of cyanuric acid. The desired cage
molecule 2 was isolated in 23% yield, along with 18%
of a dicyanuric substituted byproduct 3 in which two
chains were joined with one cyanuric unit (judged by mass and NMR
spectroscopic analyses). The exact structure of this byproduct was
not assigned, being inessential to the present report. We did not
detect tricyanuric acid substituted product in this reaction by mass
spectroscopic analysis of the crude products, suggesting that intramolecular
SN2 attack of the bromo chain by the first attached cyanuric
acid moiety can effectively compete with intermolecular reactions.
Decreasing the stoichiometry of cyanuric acid to 1.5 equiv suppressed
the formation of 3 and slightly increased the yield of 2 (26%). The identity of 2 was confirmed by molecular
mass, NMR spectroscopic analyses, and its X-ray crystal structure
(Figure 2a), which shows extreme rigidity due
to the constraint of the short linkers between the top steroid and
bottom cyanuric ring structures. Layered scaffold interactions in
the crystal packing of 2 were observed, with one layer
contributed by head-to-tail hydrophobic interaction between steroid
moieties and the other layer from stacking of apposing cyanuric moieties
(Figure 2b).
Figure 2
(a) Synthesis of cage compound 2 by bridging the three
flexible side chains on the steroid rings of cholate with cyanuric
acid in one step. Ball-and-stick drawing of the X-ray crystal structure
of 2 is shown with C atoms in gray, O atoms in red, and
N atoms in blue. (b) Crystal packing of 2 viewed along
the a-axis of the unit cells (black box) displaying
hydrophobic steroid–steroid interactions as well as interactions
between cyanuric acid moieties.
(a) Synthesis of cage compound 2 by bridging the three
flexible side chains on the steroid rings of cholate with cyanuric
acid in one step. Ball-and-stick drawing of the X-ray crystal structure
of 2 is shown with C atoms in gray, O atoms in red, and
N atoms in blue. (b) Crystal packing of 2 viewed along
the a-axis of the unit cells (black box) displaying
hydrophobic steroid–steroid interactions as well as interactions
between cyanuric acid moieties.Although typically difficult to synthesize, covalent cage
compounds
are fascinating molecules with many unique applications such as in
molecular recognition, controlled substrate binding/release, and use
as nano reaction vessels.[26−37] Given the extreme structural rigidity of compound 2 and its very small cage cavity, it is an intriguing idea that larger
cage compounds might be readily built from cholate for adaptable cage
size. Here, we extended the synthesis of cholate-templated cages by
attaching oligo(ethylene glycol) (PEG) chains to make molecules analogous
to 2 and to cryptands[38,39] (Scheme 1). We started the synthesis from the 3α,7α,12α-tri(2-hydroxyethoxy)cholane 4, which was conveniently prepared from cholic acid in three
steps.[24] Universal tosylation of the OH
groups gave 5. The reaction of 5 with NaH-deprotonated
ethylene glycol, diethylene glycol, or triethylene glycol elongated
the PEG chains in specific manner. After conversion of triols 6 to tribromides 7, treatment with cyanuric acid
in the presence of DBU was successful in giving the cyclized cage
products 8 in 25–64% yield. We note that the cyclization
yield was highest for linkers containing two ethylene glycol units
(compound 8a), possibly due to relieved strain for cage
formation compared to cyclization of 2, and the more
favorable proximity of the three chains compared to cyclization of
larger cage compounds (8b and 8c). This
series of cage compounds showed a range of internal cavity volumes,
thus conferring suitability for potential encapsulation of guest molecules
of different sizes (Figure 3).
Scheme 1
Synthesis
of PEG-Linked Cage Compounds
Figure 3
Computational space-filled structures of 2 and 8a–c (from left to right) showing increasing
volume of the cage cavity along with decreased structural rigidity.
Computational space-filled structures of 2 and 8a–c (from left to right) showing increasing
volume of the cage cavity along with decreased structural rigidity.The above cage compounds were
barely soluble in water, although
they could be effectively solubilized in mixed detergent solutions
for potential studies in aqueous solutions. Removal of the 24-caboxylate
group in these molecules, following our previous design of facial
amphiphiles, obviously contributed to their water insolubility. We
envision that it may be desirable to retain the terminal carboxylic
acid to alleviate solubility concerns and to serve as an additional
functionalization site. For these reasons, we synthesized compound 17 as well (Scheme 2). Briefly, allylation
of the three hydroxyl groups of cholic acid and the subsequent reduction
of carboxylic acid to hydroxyl were carried out in one pot to give 9 in 95% yield. After benzyl protection of the 24-OH group
(10), ozonolysis of the alkenes and NaBH4 reduction
of the resulting aldehydes were performed to give 11 in
a combined yield of 93%. Subsequent transformations from 11 to 15 were similar to the route described in Scheme 1. Finally, benzyl deprotection (16)
by hydrogenation and oxidation of the resulting OH group to restore
the carboxylic acid (17) using Jones reagent were conducted.
Like cholate, the salt form of compound 17 exhibited
the expected solubility in H2O (up to 5% w/v) and formed
micelles. Its critical micelle concentration (CMC) was determined
as 0.52 mM (0.054% w/v) using a hydrophobic dye (orange OT) solubilization
assay.[40]
Scheme 2
Synthesis of Cage 17 Bearing the Terminal Carboxylic
Acid Group
To further illustrate
the versatility of cholate-templated cage
synthesis and possibly deliver different functional molecules, we
also used the popular Cu(I)-catalyzed azide–alkyne Huisgen
cycloaddition, the so-called click chemistry,[41] for efficient linker conjugation. The corresponding 1,2,3-triazole
units have been documented to function as H-bond donors and acceptors,[42] bind halide anions,[43,44] and act as metal-binding ligands.[45] From
the previous trimesylate intermediate 18 described in
Scheme 1, we conveniently made the triazide
fragments 19 by reaction with NaN3. These
azides 19a–c, respectively containing
one, two, and three ethylene glycol units, were coupled with tripropargyl-substituted
cyanuric acid 20 in the presence of CuSO4 and
the Cu(II)-reducing sodium ascorbate. As a result, all click reactions
led to the corresponding cage products 21 in good yields
(66–74%).Finally, we show that
these cage compounds, depending on the functional
groups in the linker region, can bind different substrates. For example,
addition of KPF6 to a solution of 8a or 8b in acetone-d6 led to 1H NMR chemical shifts in a region (∼3–4 ppm) inclusive
of proton signals for the PEG linkers (Figure 4a,b). In contrast, titration of KPF6 to the smallest cage 2 solution had no effect on its NMR spectrum. These results
support the binding of K+ in the medium-sized cage cavity
in 8a or 8b but not in 2. Similarly,
we observed apparent chemical shifts on the 1H NMR spectra
of the triazole-containing cage 21a upon addition of
fluoride, chloride, bromide, or iodide anions in the forms of tetrabutylammonium
salt in acetone-d6. We show in Figure 4c the singlet proton peaks on each triazole ring
of 21a, which are well separated from all other upfield
peaks (<5 ppm). The degree of chemical shift changes also varied
in each case, roughly following an order of Cl– >
Br– > F– ∼ I–, likely indicative of a level of selectivity. Future investigations
of the selectivity and binding modes and affinities of this class
of compounds for different substrates are warranted.
Figure 4
1H NMR spectra
in acetone-d6 showing that cage compounds
(a) 8a, (b) 8b, and (c) 21a bind potassium (KPF6, 10 equiv
in a and b) and halide (Bu4NX salt, 5 equiv in c) ions,
respectively.
1H NMR spectra
in acetone-d6 showing that cage compounds
(a) 8a, (b) 8b, and (c) 21a bind potassium (KPF6, 10 equiv
in a and b) and halide (Bu4NX salt, 5 equiv in c) ions,
respectively.In summary, we synthesized
a series of novel cage molecules based
on the facial amphiphilic cholate template. Structural diversity was
achieved by selecting linkers with different functional groups and
at various lengths. These cage compounds could be made soluble in
aqueous solution to enhance their functional utility. Preliminary
studies have demonstrated that these cage molecules could bind metal
or halide ions. We envision that novel cage compounds can also be
prepared with other linkers and/or different cap molecules other than
the cyanuric acid unit employed herein. Thus, the cholic acid platform
provides a versatile and easily accessible platform for generating
a new class of covalent cage amphiphiles to potentially recognize
a variety of substrates, beyond the scope of ionic compounds studied
herein.
Experimental Section
Synthesis of 3α,7α,12α-Tris(2-bromoethoxy)cholane
(1)
To a solution of 3α,7α,12α-tri(2-hydroxyethoxy)cholane[24] (4, 2.0 g, 3.91 mmol) in dry THF
(20 mL) was added triethylamine (3.95 mL, 27.37 mmol) with stirring.
Methanesulfonyl chloride (1.81 mL, 23.46 mmol) was then added to the
mixture dropwise at 0 °C. The reaction was continued with stirring
at rt and completed in 4 h as judged by TLC analysis. Water was added
to quench the reaction, and the solution was extracted with EtOAc
(50 mL × 3). The combined organic phases were washed with brine,
dried over Na2SO4, filtered, and then concentrated
in vacuo. The residue, without purification, was dissolved in acetone
(20 mL). LiBr (3.4 g, 39.1 mmol) was then added at rt, and the suspension
was heated at 50 °C overnight. After simple workup, the mixture
was purified by silica gel chromatography (eluent: hexanes/EtOAc =
20:1 to 10:1) to give 1 as a pale yellowish solid (1.64
g, 60% over two steps): 1H NMR (600 MHz, CDCl3) δ 3.86–3.82 (m, 2H), 3.78–3.76 (m, 2H),), 3.57–3.55
(m, 1H), 3.50–3.43 (m, 8H), 3.34–3.32 (m, 1H), 3.17–3.13
(m, 1H), 2.24–2.18 (m, 1H), 2.11–2.08 (m, 2H), 1.99–1.94
(m, 1H), 1.85–1.79 (m, 2H), 1.73–1.65 (m, 5H), 1.57–1.53
(m, 1H), 1.49–1.45 (m, 1H), 1.39–1.14 (m, 8H), 1.06–0.84
(m, 12H), 0.65 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 81.5, 80.3, 76.6, 69.1, 68.7, 68.2, 46.38, 46.36, 42.4,
41.9, 39.7, 38.4, 35.8, 35.3, 35.2, 35.0, 31.3, 31.2, 31.0, 29.4,
28.0, 27.8, 27.6, 23.5, 23.4, 23.0, 19.4, 18.1, 14.7, 12.6; HRMS (ESI-TOF)
calcd for C30H51Br3O3 [M
+ Na]+ 719.1280, found 719.1281.
General Procedure for the
Preparation of Cage Compounds 2 and 8a–c
To a solution of cholic acid (10 g,
24.5 mmol) in dry THF
(150 mL) was slowly added NaH (60% in mineral oil, 14.7 g, 37.5 mol)
at 0 °C. The reaction was stirred at rt for 0.5 h before allyl
iodide (22.4 mL, 245 mmol) was added. The reaction was continued at
rt for 4 h and again cooled to 0 °C, and an additional 50 mL
of dry THF was added. To this solution was slowly added LiAlH4 (2.8 g, 73.5 mmol) in portions. The reaction temperature
was allowed to increase to rt and stirred overnight. The reaction
was carefully quenched by the addition of 3 mL of H2O at
0 °C, followed by the addition of 6 mL of 10% NaOH and 9 mL of
H2O. The solid precipitate was filtered and washed with
EtOAc. The organic solution was concentrated in vacuo, and the residue
was purified by silica gel chromatography (eluent: hexanes/EtOAc =
5:1 to 2.5:1) to give 9 as a white solid (11.8 g, 94%): 1H NMR (500 MHz, CDCl3) δ 5.96–5.84
(m,3H), 5.29–5.17 (m, 3H), 5.12–5.04 (m, 3H), 4.07–4.03
(m, 2H), 3.98–3.97 (m, 2H), 3.77–3.73 (m, 1H), 3.70–3.66
(m, 1H), 3.61–3.54 (m, 2H), 3.53–3.51 (m, 1H), 3.30–3.29
(m, 1H), 3.15–3.09 (m, 1H), 2.28–2.21 (m, 1H), 2.19–2.11
(m, 2H), 2.02–1.96 (m, 1H), 1.83–0.94 (m, 19H), 0.91–0.90
(m, 4H), 0.87 (s, 3H), 0.64 (s, 3H); 13C NMR (125 MHz,
CDCl3) δ 136.1, 135.9, 116.3, 115.6, 80.8, 79.2,
74.9, 69.4, 69.3, 68.8, 63.6, 46.5, 46.4, 42.6, 42.0, 39.8, 35.7,
35.4, 35.0, 31.8, 29.5, 28.9, 28.1, 27.7, 27.5, 23.3, 23.0, 17.8,
12.6; MS (ESI) calcd for C33H54O4 [M + Na]+ 537.4, found 537.4.
Synthesis of Compound 10
To a solution
of 9 (3 g, 5.82 mmol) in dry DMF (15 mL) was added NaH
(60% in mineral oil, 702 mg, 17.46 mmol) at 0 °C. The mixture
was stirred at rt for 0.5 h. Then benzyl bromide (1.08 mL, 8.76 mmol)
was added to the solution at 0 °C. The reaction was heated at
60 °C with stirring for 3 h. The reaction was quenched with saturated
NH4Cl after being cooled to 0 °C. The solution was
extracted with EtOAc (60 mL × 3). The combined organic phases
were washed with brine, dried over Na2SO4, filtered,
and concentrated in vacuo. The residue was purified by silica gel
chromatography (eluent: hexanes/EtOAc = 5:1 to 1:4) to give 10 as a yellow oil (3.9 g, 95%): 1H NMR (500 MHz,
CDCl3) δ 7.32–7.31 (m, 4H), 7.26–7.23
(m, 1H), 5.96–5.85 (m, 3H), 5.30–5.18 (m, 3H), 5.13–5.04
(m, 3H), 4.48 (s, 2H), 4.08–4.03 (m, 2H), 3.99–3.97
(m, 2H), 3.78–3.74 (m, 1H), 3.71–3.67 (m, 1H), 3.53–3.51
(m, 1H), 3.42 (t, J = 5.0 Hz, 2H), 3.31–3.30
(m, 1H), 3.15–3.09 (m, 1H), 2.29–2.13 (m, 3H), 2.03–1.97
(m, 1H), 1.84–0.94 (m, 19H), 0.91–0.90 (m, 4H), 0.87
(s, 3H), 0.64 (s, 3H); 13C NMR (125 MHz, CDCl3): δ = 138.8, 136.04, 136.03, 135.9, 128.4, 127.6, 127.4, 116.2,
115.5, 80.8, 79.2, 74.9, 72.8, 71.2, 69.4, 69.3, 68.7, 46.5, 46.4,
42.6, 42.0, 39.8, 35.6, 35.4, 35.1, 35.0, 32.3, 28.9, 28.0, 27.7,
27.5, 26.5, 23.3, 23.0, 17.8, 12.6; HRMS (ESI-TOF) calcd for C40H60O4 [M + H]+ 605.4570,
found 605.4564.
Synthesis of Compound 11
Ozone was bubbled
into a solution of 10 (3.9 g, 6.45 mmol) in CH2Cl2 (50 mL) and MeOH (25 mL) at −78 °C until
a blue color persisted. Excess ozone was removed with oxygen flow
until the blue color disappeared. Me2S (2.4 mL, 32.25 mmol)
was then added. The solution was stirred for another 10 min before
the addition of NaBH4 (1.47 g, 38.7 mmol) in 10 mL of 10%
NaOH solution. The reaction was allowed to slowly increase to rt.
Upon completion, the reaction was carefully quenched with 10% HCl
solution (to pH 5–6) at 0 °C to eliminate excess NaBH4. The mixture was extracted with CH2Cl2 (100 mL × 3). The combined organic phases were dried over Na2SO4, filtered, and concentrated in vacuo. The residue
was purified by silica gel chromatography (eluent: CH2Cl2/MeOH = 25:1 to 12:1) to give 11 as a colorless
oil (3.7 g, 93%): 1H NMR (500 MHz, CDCl3) δ
7.30–7.29 (m, 4H), 7.24–7.22 (m, 1H), 4.46 (s, 2H),
3.74–3.51 (m, 14H), 3.40 (t, J = 5.0 Hz, 2H),
3.30–3.28 (m, 2H), 3.18–3.09 (m, 2H), 2.22–2.14
(m, 1H), 2.12–2.05 (m, 2H), 1.94–1.88 (m, 1H), 1.85–1.60
(m, 8H), 1.54–1.16 (m, 9H), 1.09–0.97 (m, 2H), 0.93–0.89
(m, 4H), 0.86 (s, 3H), 0.63 (s, 3H); 13C NMR (125 MHz,
CDCl3) δ 138.7, 128.2, 127.5, 127.4, 80.9, 79.6,
76.3, 72.8, 70.9, 70.2, 69.6, 69.1, 62.0, 61.9, 61.8, 46.6, 46.2,
42.5, 41.7, 39.5, 35.4, 35.1, 34.9, 34.7, 32.2, 28.8, 27.8, 27.5,
27.0, 26.3, 23.2, 22.9, 22.7, 18.0, 12.5; HRMS (ESI-TOF) calcd for
C37H60O7 [M + H]+ 617.4412,
found 617.4414.
Authors: Neil P Grimster; Bernhard Stump; Joseph R Fotsing; Timo Weide; Todd T Talley; John G Yamauchi; Ákos Nemecz; Choel Kim; Kwok-Yiu Ho; K Barry Sharpless; Palmer Taylor; Valery V Fokin Journal: J Am Chem Soc Date: 2012-04-06 Impact factor: 15.419
Authors: Sung Chang Lee; Brad C Bennett; Wen-Xu Hong; Yu Fu; Kent A Baker; Julien Marcoux; Carol V Robinson; Andrew B Ward; James R Halpert; Raymond C Stevens; Charles David Stout; Mark J Yeager; Qinghai Zhang Journal: Proc Natl Acad Sci U S A Date: 2013-03-11 Impact factor: 11.205