Implementing the catch-and-release concept into a simple method for regioselective cyclodextrin modification

Article abstract of DOI:10.1021/ol400641n

Differentiation of specific positions in axial-symmetrical cyclodextrins (CDs), exhibiting a dense display of identical functional groups, is challenging. A novel strategy toward this goal that exploits a solid matrix to display the complementary reagent functionalities sufficiently far from each other to prevent CDs from reacting through more than one site is reported. Using a catch-and-release process based on the Staudinger reaction, the utility of this concept to easily produce complex CD functionalization patterns in one pot and without any purification step is demonstrated.

Full text of DOI:10.1021/ol400641n

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Implementing the “Catch-and-Release”
Concept into a Simple Method for
Regioselective Cyclodextrin Modification
Iris Pflueger and Juan M. Benito*
ꢀ
Institute ofChemicalResearch (IIQ), CSIC - Universidad de Sevilla, AmericoVespucio49,
E-41092 Sevilla, Spain
juanmab@iiq.csic.es
Received March 11, 2013
ABSTRACT
Differentiation of specific positions in axial-symmetrical cyclodextrins (CDs), exhibiting a dense display of identical functional groups, is challenging.
A novel strategy toward this goal that exploits a solid matrix to display the complementary reagent functionalities sufficiently far from each other to
prevent CDsfrom reacting through more thanone siteis reported. Using a“catch-and-release”processbasedonthe Staudingerreaction, theutilityof
this concept to easily produce complex CD functionalization patterns in one pot and without any purification step is demonstrated.
The unique molecular inclusion capabilities of cyclo-
dextrins (CDs) have been exhaustively exploited by the
pharmaceutical, cosmetic, and food industries.1 Chemical
tailoring of native hexa-, hepta-, and octameric CDs (R-,
β-, and γ-CD, respectively) has already benefited many
current applications.2 But the development of the myriad
of potential applications envisioned for functionalized
CDs3 faces the enormous challenge of selectively manip-
ulating the dense display of identical functional groups of
CDs. Toward this end, only a handful of regioselective
chemical manipulation schemes are available.4 The most
successful ones take advantage of the scaffolding or inclu-
sion capabilities of the CD host to force a preferred
approach and orientation of the CD-modifying reagent.
In this way, the reaction outcome can be significantly
restricted as compared to that statistically expected. The
level of refinement achieved with this kind of approach is
(4) (a) Khan, A. R.; Forgo, P.; Stine, K. J.; D’Souza, V. T. Chem. Rev.
1998, 98, 1977–1996. (b) Pearce, A. J.; Sinay, P. Angew. Chem., Int. Ed.
2000, 39, 3610–3612. (c) Lecourt, T.; Pearce, A. J.; Herault, A.; Sollogoub,
M.; Sinay, P. Chem.;Eur. J. 2004, 10, 2960–2971. (d) Balbuena, P.; Lesur,
ꢀ
D.; Gonzalez Alvarez, M. J.; Mendicuti, F.; Ortiz Mellet, C.; Garcıa
(1) (a) Cyclodextrins and their complexes; Doziuk, H., Ed.; Wiley-
VCH: Weinheim, 2006. (b) Cylodextrins in pharmaceutics, cosmetics, and
biomedicine: Current and future industrial applications; Bilensoy, E., Ed.;
Wiley-VCH: Weinheim, 2011.
ꢀ
ꢀ
Fernandez, J. M. Chem. Commun. 2007, 3270–3272. (e) Guieu, S.;
Sollogoub, M. J. Org. Chem. 2008, 73, 2819–2828. (f) Guieu, S.;
Sollogoub, M. Angew. Chem., Int. Ed. 2008, 47, 7060–7063. (g) Sollogoub,
M. Eur. J. Org. Chem. 2009, 1295–1303. (h) Rawal, G. K.; Rani, S.; Ward,
S.; Ling, C.-C. Org. Biomol. Chem. 2010, 9, 171–180. (i) Law, H.; Benito,
ꢀ
(2) For selected contributions, see: (a) Benito, J. M.; Gomez-Garcıa,
ꢀ
M.; Ortiz Mellet, C.; Baussanne, I.; Defaye, J.; Garcıa Fernandez, J. M.
J. Am. Chem. Soc. 2004, 126, 10355–10363. (b) Ortega-Caballero, F.;
Rousseau, C.; Christensen, B.; Petersen, T. E.; Bols, M. J. Am. Chem.
ꢀ
J. M.; Garcıa Fernandez, J. M.; Jicsinszky, L.; Crouzy, S.; Defaye, J. J.
Phys. Chem. B 2011, 115, 7524–7532. (j) Gramage-Doria, R.; Rodriguez-
Lucena, D.; Armspach, D.; Egloff, C.; Jouffroy, M.; Matt, D.; Toupet, L.
ꢀ
Soc. 2004, 126, 10355–10363. (c) Guieu, S.; Zaborova, E.; Bleriot, Y.;
Poli, G.; Jutand, A.; Madec, D.; Prestat, G.; Sollogoub, M. Angew.
€
Chem.;Eur. J. 2011, 17, 3911–3921. (k) Faugeras, P.-A.; Boens, B.;
€
Chem., Int. Ed. 2010, 49, 2314–2318. (d) Grunstein, D.; Maglinao, M.;
Elchinger, P.-H.; Brouillette, F.; Montplaisir, D.; Zerrouki, R.; Lucas,
R. Eur. J. Org. Chem. 2012, 4087–4105. (l) Ghosh, R.; Zhang, P.; Wang,
A.; Ling, C.-C. Angew. Chem., Int. Ed. 2012, 51, 1548–1552. (m) Jouffroy,
M.; Gramage-Doria, R.; Armspach, D.; Matt, D.; Toupet, L. Chem.
Commun. 2012, 48, 6028–6030. (n) Zaborova, E.; Guitet, M.; Prencipe, G.;
Kikkeri, R.; Collot, M.; Barylyuk, K.; Lepenies, B.; Kamena, F.;
Zenobi, R.; Seeberger, P. H. J. Am. Chem. Soc. 2011, 133, 13957–13966.
(3) (a) Villalonga, R.; Cao, R.; Fragoso, A. Chem. Rev. 2007, 107,
3088–3116. (b) van de Manakker, F.; Vermonden, T.; van Nostrum,
C. F.; Hennink, W. E. Biomacromolecules 2009, 10, 3157–3175. (c)
Wenz, G.; Han, B. H.; Muller, A. Chem. Rev. 2006, 106, 782–817. (d)
ꢀ
ꢀ
Bleriot, Y.; Menand, M.; Sollogoub, M. Angew. Chem., Int. Ed. 2013, 52,
ꢀ
639–644. (o) Balbuena, P.; Gonc-alves-Pereira, R.; Jimenez Blanco, J. L.;
Garcıa-Moreno, M. I.; Lesur, D.; Ortiz Mellet, C.; Garcıa Fernandez,
ꢀ
ꢀ
Ortiz Mellet, C.; Garcıa Fernandez, J. M.; Benito, J. M. Chem. Soc. Rev.
2011, 40, 1586–1608.
J. M. J. Org. Chem. 2013, 78, 1390–1403.
r
10.1021/ol400641n
XXXX American Chemical Society

paradigmatically illustrated with a recent contribution from
Sollogoub and co-workers,4n who showed how DIBAL-H
can be used as a molecular scalpel to sequentially rip off
selected benzyl protecting groups (among 18 similar ones!) to
obtain a series of triply dissymmetrized RCD derivatives.
These approaches have been mastered to furnish extra-
ordinary selectivities, but their application is often limited
to tight-matching CD-reagent pairs. Thus, development of
tools to widen the accessible CD architectural scope would
result in an excellent complement for the above strategies.
Orientating the CD-reagent pair and preventing a second
reagent species to react the same CD molecule is crucial for
the success of these strategies. We reasoned that, while
executing this control is inherently difficult in solution, it
might be easier with a reagent displayed on a solid support.
Indeed, this has been shown to be successful to achieve
monofunctionalization of 2-nm gold nanoparticles, only
slightly larger than βCD.5 More recently, Di Fabio and
co-workers have demonstrated that solid-supported reagents
canbeusedtograftasinglelabelontoaseriesofCDscaffolds
via phosphodiester linkages.6 Only moderate yields were
reported, probably due to either a suboptimal reagent display
on the solid support or reaction conditions. Moreover, the
precise functionalization site (primary or secondary rim) was
not disclosed and final HPLC purification was required.
Herein, we have refined this concept into a flexible
synthetic tool for site selective CD functionalization avoiding
purification steps. For such a purpose, we have implemented
an experimental setup permitting (i) the covalent capture
of a fully symmetric CD derivative through a single position
by a solid-supported reagent, (ii) eventual “on-bead” orthog-
onal elaboration of the remaining functional groups,
and (iii) a final chemoselective release of the dissymmetric
CD conjugate from the solid support in a sort of one-pot
“catch-and-release”7 process (Scheme 1).
Scheme 1. Schematic Representation of the Solid-Support-
Assisted “Catch-and-Release” Protocol for Site Selective CD
Functionalization
alkyne cycloaddition (CuAAC),10 which would enable
“on-bead” derivatization of the remaining functional groups.
To optimize the experimental conditions, we first studied
iminophosphorane formation in solution with the model
monosaccharide azide 111 (a surrogate of the βCD heptaazide
3)12 and a set of aromatic phosphines (Scheme 2). Reaction
with triphenylphosphine (TPP) at rt indicated relatively fast
kinetics (t1/2 = 20ꢀ30 min, 31P NMR-monitoring; Figure 1,
O series) in all tested solvents (DCM, DMF, and 1,4-
dioxane). The reaction with phosphine 4,13 with less elec-
tron-donating capabilities due to the presence of the carbo-
xymethyl handle, was slower (t1/2 = 80 min in 1,4-dioxane;
Figure 1, 4 series). Conversely, electron-rich phosphine 613
fully recovered the performance of TPP, being the most
appropriate choice to build on the solid support. Moreover,
the reaction rate could be significantly increased at 40 °C
(t1/2 = 6 min in DMF; Figure 1, [ series) with negligible
iminophosphorane hydrolysis after several hours provided
anhydrous conditions were preserved. The optimal perfor-
mance together with good resin swelling capabilities sup-
ported the choice of DMF for the assays on the solid support.
Coupling of carboxylic acid 5 onto glycine-loaded ami-
The chemical reactivity for the catch-and-release process
is not a trivial choice. In this case the versatility of the
Staudinger reaction8 between organic azides and phos-
phines has been considered. Staudinger reaction occurs in
virtually any solvent to chemoselectively furnish imino-
phosphorane species in mild conditions and good yields.
The “captured” iminophosphorane might be later re-
leased from the solid matrix chemoselectively in the form
of e. g. an amine, iso(thio)cyanate or (thio)urea. In addition,
iminophosphorane reactivity9 isorthogonal toother azide-
involving reactions, such as the Cu(I)-catalyzed azideꢀ
nomethylated polystyrene (AM-PS, 0.39 mmol gꢀ1 14
)
using conventional peptide coupling reagents furnished
3
the resin-supported phosphine 7, stable in the absence of
moisture and oxygen. AM-PS is not a rigid matrix, but even
at this relatively high loading, and assuming a swelling in the
5ꢀ10 mL gꢀ1 range, the average distance between reactive
(5) (a) Sung, K.-M.; Mosley, D. W.; Peelle, B. R.; Zhang, S.;
Jacobson, J. M. J. Am. Chem. Soc. 2004, 126, 5064–5065. (b) Worden,
J. G.; Shaffer, A. W.; Huo, Q. Chem. Commun. 2004, 518–519.
(6) Di Fabio, G.; Malgieri, G.; Isernia, C.; D’Onofrio, J.; Gaglione,
M.; Messere, A.; Zarrelli, A.; De Napoli, L. Chem. Commun. 2012, 48,
3875–3877.
3
centers is estimated to be ca. 4 nm.15 This is 3-fold larger than
the maximum distance between the azido groups in βCD 3.16
(11) Bueno Martınez, M.; Zamora Mata, F.; Ugalde Donoso, M. T.;
ꢀ
(12) Fulton, D. A.; Pease, A. R.; Stoddart, J. F. Isr. J. Chem. 2000, 40,
325–333.
(13) See SI for synthetic details.
(7) Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach,
A. G.; Longbottom, D. A.; Nesi, M.; Scott, J. S.; Storer, I. R.; Taylor,
S. J. J. Chem. Soc., Perkin Trans. 1 2000, 3815–4195.
(8) (a) van Berkel, S. S.; van Eldijk, M. B.; van Hest, J. C. M. Angew.
Chem., Int. Ed. 2011, 50, 8806–8827. (b) Saxon, E.; Bertozzi, C. R.
Science 2000, 287, 2007–2010. (c) Nilsson, B. L.; Kiessling, L. L.; Raines,
R. T. Org. Lett. 2000, 2, 1939–1941.
(9) Palacios, F.; Alonso, C.; Aparicio, D.; Rubiales, G.; de los Santos,
J. M. Tetrahedron 2007, 63, 523–575.
(10) Meldal, M.; Tornøe, C. W. Chem. Rev. 2008, 108, 2952–3015.
Galbis Perez, J. A. Carbohydr. Res. 1992, 230, 191–196.
(14) Purchased from NovaBiochem.
(15) The interphosphine distance was estimated supposing an even
distribution in fully swollen resins. From the total number of functional
groups in a given resin volume, the volume per phosphine group, and
therefore the average distance between them, was calculated.
B
Org. Lett., Vol. XX, No. XX, XXXX

Scheme 2. Structure of the Azides 1ꢀ3 and Phosphines 4, 6, and
7 Investigated To Optimize the Staudinger Reaction
Figure 1. Iminophosphorane formation kinetics from azide 1
with different phosphines monitored by 31P NMR. In a typical
experiment, 50 μmol of phosphine in the indicated solvent were
treated under N2 with 2 equiv of 1 at either rt or 40 °C.
Scheme 3. Schematic Representation of the Solid-Support-
Assisted Synthesis of Monoamines 8ꢀ10
The reaction of 1with resin 7in DMF is slow at rt (t1/2 =8h),
but practical rates were obtained at 40 °C (Figure 1, 9 series).
Indeed, treating resin 7 with 2 equiv of azide 1 overnight
at 40 °C in dry DMF under N2 resulted in virtually complete
consumption of the supported phosphine. Premature
hydrolysis of the iminophosphorane (resin washings only
contained the excess of azide) was not detected. Final
hydrolysis with 9:1 DMF/water overnight at 40 °C afforded
6-aminoglucoside 8 in 89% yield (Scheme 3).17
R,R0-Trehalose diazide 213 and βCD heptazide 3 ap-
peared to be more demanding substrates. Only minute
amounts were captured by resin 7 under the same condi-
tions. Higher “catch” temperatures increased the reactivity
but reduced the selectivity. Even diazide 2, featuring a
maximum interazide distance of 7.5 A,16 was doubly bound
˚
by two phosphine moieties, despite the much larger average
resin interphosphine distance. Probably a temperature-
induced increase in resin backbone mobility reduced the
effective interphosphine distance. A similar effect was
observed with more flexible supports (e.g., PEGA resin).18
Rewardingly, microwave (μw) irradiation revealed a
completely different scenario. The catch efficiency in-
creased up to 90% (estimated from the recovered azide)
after irradiating resin 7 swollen in a DMF solution of either
2or 3at60°C(2 ꢁ 10 min, 40 W) underN2. We hypothesize
that the larger μw-absorption capability of DMF (as
compared to nonpolar PS-resin) could selectively increase
reagent diffusion and reactivity vs resin backbone mobility.
Hydrolytic release of the supported iminophosphoranes
afforded the target monoreduced adducts (52 and 79% for
9 and 10, respectively) virtually devoid of over-reduced
adducts as denoted by ESI-MS and RP-HPLC (see Sup-
porting Information (SI)).
The flexibility of iminophosphorane chemistry9 was
also exploited to produce alternative CD functionalization
patterns (Scheme 4). Aza-Wittig-type reaction of the
resin-bound CD iminophosphorane with CS2 furnished
the isothiocyanate 11 in 86% yield. Despite the fact that a
clean HPLC trace is observed, a minor peak for diisothio-
cyanate conjugates was identified by ESI-MS (see SI).
(16) Calculated by molecular modelling using ChemBio3D
(ChemBioOffice package, CambridgeSoft).
(17) Analytical and spectroscopic data match with those of an
authentic sample prepared in solution. See SI for details.
(18) Meldal, M. Tetrahedron Lett. 1992, 33, 3077–3080.
Org. Lett., Vol. XX, No. XX, XXXX
C

Scheme 4. Selectively Functionalized βCD Derivatives 11ꢀ13
Obtained from Solid-Supported CD Iminophosphorane
Figure 2. ESI-MS spectrum (top) and HPLC trace (bottom) of
the crude bifunctional CD 13 obtained from 3.
the corresponding urea 12 in 76% yield with undetectable
traces of difunctionalized adducts (Scheme 4).
To push forward this strategy and take advantage of
the greater selectivity of 7-dil, on-bead manipulation of the
solid-supported CD has been considered. The iminopho-
sphorane tether should, in principle, remain unaffected by
the conditions for CuAAC. Indeed, treating the supported
CD iminophosphorane with an excess of a terminal alkyne
(e.g., N-Boc propargylamine) in the presence of catalytic
CuI, followed by CS2-mediated release, afforded the bi-
functional CD derivative 13 in 59% yield. The NMR
spectrum of this conjugate is far too complex to assess
purity and even identity, but the RP-HPLC trace and
ESI-MS spectra confirmed both (Figure 2).
Production of such CD functionalization patterns, despite
being feasible, is far from obvious using conventional solution
phase techniques. Herein we demonstrate that the “catch-
and-release” concept can be implemented with relative ease
into a powerful tool for selective manipulation of macro-
molecular topology. Only βCD manipulation is described
herein, but it is reasonable to suppose that this methodology
would be suited for other cyclodextrins and macromolecular
scaffolds. Room for improvement is still large, for example,
considering more rigid solid supports. Work in this direction
is currently in progress in our laboratories.
Probably the interphosphine distance in the matrix is not
sufficient to completely prevent doubly capturing a certain
amount of CD 3. Curiously, such over-reactivity was not
observed in the case of amine 10, despite the similar catch
conditions. We reasoned that iminophosphorane hydro-
lysis may occur at a slower rate due to shrinking of the
PS resin in the presence of water, thereby increasing the
chance for doubly linked CD units to remain bound to
the matrix. Conversely, the CS2-DMF cocktail swells the
PS resin much better, facilitating the release of all bound
material (including mono- and bis-iminophosphoranes).
In order to minimize this drawback, we prepared a resin
with a larger interphosphine distance. Thus, AM-PS resin
was acylated with a 9:1 mixture of Boc- and Fmoc-
protected glycine and the Boc groups were transformed
into inert acetyl groups. Phosphine was then installed onto
the remaining amino groups (see SI), resulting in a 10-fold
Acknowledgment. The Spanish Ministerio de Econo-
mıa y Competitividad (contract number CTQ2010-15848),
the Fondo Europeo de Desarrollo Regional (FEDER),
and the Fondo Social Europeo (FSE) are thanked for
funding. I.P. is a JAE fellow (CSIC).
Note Added after ASAP Publication. Scheme 4 was
corrected on April 26, 2013.
Supporting Information Available. Spectroscopic and
analytical data, and detailed experimental procedures.
This material is available free of charge via the Internet at
http://pubs.acs.org.
diluted matrix (7-dil) with a ca. 40 μmol gꢀ1 loading and
3
7ꢀ9 nm interphosphine average distance.15 To probe the
performance of this support, CD 3 was sequentially loaded
on and cleaved by 3-methoxypropylisothiocyanate to re-
lease a carbodiimide that in situ added water to furnish
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX