Homo- and hetero-difunctionalized β-cyclodextrins: Short direct synthesis in gram scale and analysis of regiochemistry

Article abstract of DOI:10.3762/bjoc.15.66

The regioselective difunctionalization of cyclodextrins (CDs) leading to derivatives amenable to further transformations is a daunting task due to challenging purification and unambiguous characterization of the obtained regioisomers with similar physico-

Full text of DOI:10.3762/bjoc.15.66

Homo- and hetero-difunctionalized β-cyclodextrins: Short
direct synthesis in gram scale and analysis of regiochemistry
Gábor Benkovics1, Mihály Bálint1, Éva Fenyvesi1, Erzsébet Varga1, Szabolcs Béni2,
Konstantina Yannakopoulou*3 and Milo Malanga*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2019, 15, 710–720.
1CycloLab, Cyclodextrin Research and Development Laboratory Ltd.,
llatos út 7, Budapest, H-1097, Hungary, 2Department of
Pharmacognosy, Semmelweis University, Budapest, H-1085 Üllői út
26, Hungary and 3Institute of Nanoscience and Nanotechnology
National Center for Scientific Research “Demokritos”, Patr. Gregoriou
E & 27 Neapoleos str., Aghia Paraskevi Attikis 15341, Greece
doi:10.3762/bjoc.15.66
Received: 07 November 2018
Accepted: 28 February 2019
Published: 18 March 2019
Associate Editor: H. Ritter
Email:
Konstantina Yannakopoulou* - k.yannakopoulou@inn.demokritos.gr;
Milo Malanga* - malanga@cyclolab.hu
© 2019 Benkovics et al.; licensee Beilstein-Institut.
License and terms: see end of document.
* Corresponding author
Keywords:
azido-tosyl-cyclodextrin; diazido-cyclodextrin;
hetero-difunctionalization; homo-difunctionalization; regioselectivity
Abstract
The regioselective difunctionalization of cyclodextrins (CDs) leading to derivatives amenable to further transformations is a
daunting task due to challenging purification and unambiguous characterization of the obtained regioisomers with similar physico-
chemical properties. The primary-side homo-difunctionalization of β-CD can lead to three regioisomers, while the hetero-difunc-
tionalization can generate three pairs of pseudoenantiomers. Previously, approaches with several synthetic steps, expensive
reagents, high purification demands and low yields of the products have been employed. Herein we present direct, short and effi-
cient primary-side difunctionalization strategies featuring reproducibility, ease of product purification, scalability of the reactions
and versatility of the substituents introduced. Specifically, the prepared ditosylated β-CDs were separated using preparative
reversed-phase column chromatography and their structures were elucidated by NMR experiments. Azidation led to the correspond-
ing pure diazido regioisomers. Direct monotosylation of 6-monoazido-β-CD or monoazidation of the single regioisomers 6A,6X-
ditosyl-β-CDs afforded hetero-difunctionalized 6A-monoazido-6X-tosyl-β-CDs in significant yields. Overall, the single regio-
isomers, 6A,6X-ditosyl-, 6A,6X-diazido- and 6A-monoazido-6X-monotosyl-β-CD were prepared in one or two steps and purified in
multigram scale thus opening the way towards further selective and orthogonal functionalizations of β-CD hosts.
Introduction
Cyclodextrins (CDs) are cyclic oligomers of α-D-glucopyra- lecular chemistry due to their ability to form host–guest inclu-
nose (Figure 1 illustrates the heptamer, β-CD) that have at- sion complexes [1]. A selective functionalization of these cyclic
tracted worldwide interest in various fields of applied supramo- oligosaccharides can remarkably improve their complexing
710

Beilstein J. Org. Chem. 2019, 15, 710–720.
ability and enables their application as artificial enzymes [2-4], In order to achieve homo-difunctionalization of the primary
chiral resolving agents [5], stimuli-responsive materials [6], side, two synthetic approaches can be applied: (i) the direct
molecular sensors [7] or bioactive hosts with significant difunctionalization, based on the regioselective installation of
emerging applications [8,9].
designed disulfonyl capping moieties to the CD core and their
subsequent substitution by the desired functional groups or
(ii) the indirect difunctionalization based on the regioselective
removal of protecting groups from a previously perfunctional-
ized CD derivative.
The first approach was developed by Fujita et al. by intro-
ducing customized capping reagents for β-CD [10]. Tabushi and
co-workers developed a series of capping agents providing
6A,6B-, 6A,6C-, or 6A,6D-selectivity on β-CD (Scheme 1)
[11,12]. The cap could be subsequently removed by a suitable
nucleophilic substitution to afford the homo-disubstituted deriv-
atives, e.g., 6A,6X-diazido-β-CDs can be prepared using sodi-
Figure 1: Schematic representation of β-CD with glucopyranose atom
numbering and with alphabetic labeling of the seven glucopyranose
subunits.
There exist reliable experimental protocols for the selective um azide in N,N-dimethylformamide (DMF) and moderate
mono- and persubstitution of native CDs that are now consid- heating (Scheme 1).
ered as established procedures. However, the introduction of
two identical (homo-difunctionalization) or two different Although capping reagents are selective in disubstitution and
(hetero-difunctionalization) functional groups at defined posi- this methodology revolutionized CD difunctionalization, their
tions on the CD macrocycle in an efficient, reproducible and application has many serious drawbacks such as unwanted over-
up-scalable process is still a very challenging task.
substitution (multiple capping [13,14]), connection of two or
Scheme 1: Syntheses of 6A,6X-diazido-β-CDs as reference compounds using the “capping” literature method [11,12].
711

Beilstein J. Org. Chem. 2019, 15, 710–720.
more β-CD molecules through intermolecular disulfonyl bridges resorting to expensive capping agents and produce scalable,
and partial hydrolysis of the capping agent during the work-up. amenable to chromatographic separation and reproducible
Consequently, a chromatographic purification is essential for difunctionalization methods. Moreover, the direct NMR spec-
the isolation of the capped β-CD derivative or the substitution troscopic analysis would provide structure elucidation without
products obtained by cap displacement which generally leads to any reference material, time-consuming chemical conversions
low yields of the isolated pure compounds.
or enzymatic degradation.
Alternatively, the preparation of homo-difunctionalized CDs is Hetero-difunctionalized CDs are very challenging to prepare
based on selective deprotection of persilylated or peralkylated due to the fact that in addition to regioisomers, pseudoenan-
CDs. This approach, developed by Sinay et al. [15] and studied tiomers are unavoidably formed that have the same substitution
in depth on perbenzylated CDs, was extended up to hexa- pattern, but mirror-image relationship between the arrange-
heterodifferentiation of α-CD by Sollogoub and co-workers ments of substituents [19]. This phenomenon has an amplified
[16]. The regioselective DIBAL deprotection has also been effect in applications which are based on stereoselective interac-
applied on the primary side of 6-persilylated-2,3-permethylated tions of CDs with guest molecules (chiral separations, asym-
or 6-persilylated-2,3-perbenzylated CDs by Ling et al. [17] and metric catalysis, enzyme mimics) [20]. However, if the target is
on the secondary side of permethylated CDs by Sollogoub and only the side-selectivity of difunctionalization, pseudoenan-
Zhang et al. [18]. The selective deprotection strategy is advanta- tiomeric mixtures of the regioisomers can be used [21]. For
geous in terms of applicability and product yields, however, CD-based multifunctional drug carriers, incorporating two dif-
compared to the direct difunctionalization approach, it requires ferent groups, for example a targeting unit and a prodrug, pseu-
two extra synthetic steps that make the atom economy of this doenantiomeric purity is not required, but the side-selective
approach very low and the entire synthetic procedure time substitution has to be ensured. This has led us to develop a
consuming.
versatile and simple synthetic route towards difunctionalized
β-CD, carrying non-identical functional groups on the primary
Structure elucidation of difunctionalized CDs is a challenging side. Three different types of reactions were investigated to
task, irrespective of the strategy used. In the direct difunctional- obtain 6A-monoazido-6X-monotosyl-β-CD (a key intermediate
ization approach, the disulfonyl-capped CD has to be converted for the preparation of various hetero-difunctionalized β-CDs),
to 6A,6X-tert-butylsulfenyl-β-CD, in which the carbon atoms having three possible regioisomers and consequently three pairs
C1, C4 and C6 of the glucopyranose rings initially bearing the of pseudoenantiomers. The first two reactions were based on
cap, experience remarkable remote substituent effects detectable step-wise substitution of the primary rim with different reac-
in the 13C NMR spectra [8]. This effect is the largest in the case tants, while in the third type the pure regioisomers of 6A,6X-
of the AB-substituted compounds and because it decreases with ditosylated β-CDs were used as starting materials and sodium
the distance between the substituents, it is barely observable for azide was the limiting reagent. In this latter case, single regio-
the AD substitution. For the unambiguous verification of regio- isomers of hetero-difunctionalized β-CD, i.e., bearing orthogo-
chemistry, however, conversion of the capped β-CDs to the cor- nal functional groups available for further manipulations, were
responding diphenylthio derivatives is needed, followed by readily obtained in good yields.
Taka-amylase enzymatic degradation and sodium borohydride
Results and Discussion
reduction. Only the in-depth analysis of the NMR and MS spec-
Direct homo-difunctionalization of β-CD on
the primary side
tra of the reduced tri- and disaccharides containing the
phenylthio moieties can reveal the structure of the starting
disulfonyl-capped CD [16]. In the indirect approach using the
selective DIBAL deprotection, the regiochemical investigation
is similarly laborious. In the case of β-CD the formed 6A,6D-
deprotected product cannot be identified directly through
conventional NMR techniques because the low symmetry of the
compound causes extensive overlapping in the 1H NMR spec-
tra. The multistep “hex-5-enose” degradation method has to be
used instead to determine the substituted positions and identify
the regioisomer [15].
The idea was to replace the regioselective capping agents with a
less selective reagent while still producing a chromatographical-
ly separable mixture of regioisomers and to use a robust and
reliable preparative column chromatography (PCC) method for
their separation. p-Toluenesulfonyl chloride (p-TsCl), an easily
accessible and inexpensive reagent was chosen, which has
showed high selectivity towards the primary rim in monosubsti-
tution of β-CD [22,23]. Our assumption was that targeting the
disubstituted product with the same tosylating agent would
preserve the side-selectivity and significantly reduce the num-
ber of possible regioisomers (6A,6B-, 6A,6C- and 6A,6D-ditosyl-
β-CD). Earlier works of Fujita et al. [24] and of our research
The aims of this work were to develop and evaluate a short and
direct strategy for the selective modification of β-CD without
712

Beilstein J. Org. Chem. 2019, 15, 710–720.
group [25] have shown that regioisomers of CD derivatives 6A,6X-diazido compounds (Figure 2) revealed that ditosylation
bearing multiple bulky hydrophobic substituents such as two in pyridine gave three regioisomers, (6A,6B-; 6A,6C- and 6A,6D-
tosyl groups or one single cinnamyl moiety can be separated in ditosyl-β-CDs) in a ratio 6A,6B:6A,6C:6A,6D = 35:31:34 (reac-
larger quantities. This approach was also supported by the fact tion 1, Scheme 2), while the Cu(II)-mediated ditosylation in
that in the large – kilogram – scale production of the key water/acetonitrile (ACN) mixture gave only two positional
synthon 6-monotosyl-β-CD besides the desired product, over- ditosyl-β-CD isomers in a ratio 6A,6C:6A,6D = 58:42, (reaction
tosylation occurs giving a mixture of regioisomers of ditosy- 2, Scheme 2). Tosylation in pyridine is known to be catalyzed
lated and tritosylated β-CDs. The ditosylated fraction repre- by the oriented inclusion of the aromatic heterocycle into the
sents a significant amount (10–20%) of the crude product, cavity of the β-CD [26], which leads to the activation of all pri-
which has to be separated either by selective crystallization or mary side OH groups equally, i.e., the presence of the substitu-
by chromatography.
ent on the glucopyranose unit A does not influence the substitu-
tion on any glucopyranose unit X. Therefore, the formation of
The first concern was to verify that using the two most com- all possible regioisomers is statistical and no regioselectivity is
monly used procedures for 6-monotosylation, the substitution observed. On the other hand, tosylation under aqueous basic
by the second tosyl moiety is still selective for the primary rim conditions has a different reaction mechanism in which p-TsCl
of β-CD. For a straightforward identification, the authentic occupies the CD cavity prior to the reaction [23]. This orienta-
diazido compounds with known regiochemistry were synthe- tion of the first tosyl group has a great impact on the substitu-
sized using the appropriately spaced disulfonate capping agent, tion of the second tosyl moiety, which can react only with the
followed by azide opening of the cap and by chromatographic more distant glucose units, therefore only AC and AD disubsti-
purification of the diazidated fractions [11,12] (reference reac- tution takes place. The detailed study of the reaction mecha-
tions 1–3, Scheme 1). Direct ditosylation reactions (reactions 1 nisms is out of the scope of this paper, however, the different
and 2, Scheme 2) were performed next using conditions which mechanisms and reaction intermediates are likely responsible
have been proved to be side-selective for 6-monotosylation of for the distinct outcome of the two ditosylation reactions and for
β-CD [22,23].
the observed partial regioselectivity in reaction 2 (Scheme 2).
After the work-up and separation of the ditosylated fractions Additionally, a Vilsmeier–Haack/Appel-type iodination reac-
from the reaction mixtures, a part of the products was con- tion was performed with β-CD, using triphenylphosphine
verted to the diazido compounds. The comparison of the HPLC (PPh3) and iodine (I2) in DMF (reaction 3, Scheme 2). This
retention times of the compounds with those of the reference reaction is known to be selective for the primary side of CDs
Scheme 2: Syntheses of homo-difunctionalized β-CDs using different reaction conditions.
713

Beilstein J. Org. Chem. 2019, 15, 710–720.
Figure 2: HPLC chromatograms of the authentic 6A,6X-diazido-β-CDs with known regiochemistry (references 1–3, Scheme 1) and of the diazido-β-
CDs prepared through ditosylation in pyridine (reaction 1, Scheme 2), through Cu(II)-mediated ditosylation in H2O/ACN mixture (reaction 2,
Scheme 2) and through a Vilsmeier–Haack/Appel-type iodination (reaction 3, Scheme 2).
[27] and therefore it can be also used as an alternative control these key intermediates and their in-depth spectroscopic charac-
reaction for side-selective disubstitution. The formed mixture of terization by NMR. For the reversed-phase HPLC chro-
6-iodo-β-CDs differing in the degree of substitution (DS) was in matograms optimized for the preparative separation of the
one-pot transformed to a mixture of 6-azido-β-CDs. The crude 6A,6D- and 6A,6C-ditosyl-β-CDs prepared in reaction 2 and for
product contained monoazido-, diazido- and triazido-β-CDs, the isolation of the 6A,6B-ditosyl-β-CD prepared in reaction 1,
from which the diazido-β-CD fraction was isolated in 55% yield see Supporting Information File 1, Figures S18–S19).
using reversed-phase PCC and water/methanol gradient elution.
The HPLC analysis of the diazidated fractions surprisingly In summary, the direct ditosylation in environmentally friendly
showed only two components, having identical retention times aqueous medium affords the 6A,6C- and 6A,6D-ditosyl-β-CDs,
as those of 6A,6C-diazido and 6A,6D-diazido derivatives (refer- from which the 6A,6C- and 6A,6D-diazides are obtained in
ence 2 and reference 3, respectively, in Scheme 1). According overall 29% and 21% yield, respectively. If direct iodination is
to this observation, under the used reaction conditions in the used the final, readily separable 6A,6C- and 6A,6D-diazides are
Vilsmeier–Haack/Appel-type iodination the 6A,6B-substitution obtained in 26% and 28% yield, respectively, in a one pot
does not take place and only the 6A,6D- and 6A,6C-substituted process. Both approaches give the desired diazides in much
products are formed in area percentage ratio 52:48, respectively higher yields than when using the capping reagents (5% and
(reaction 3, Scheme 2). This outcome might be attributed to the 12%, respectively).
bulkiness of the halogenating agent (triphenylphosphine–iodine
Analysis of regiochemistry of homo-
adduct) and to the mild reaction conditions used (i.e., a relative-
difunctionalized β-cyclodextrins by full NMR
spectral assignment
ly low temperature for the iodination), under which the substitu-
tion on adjacent glucose units is not favored.
NMR structural analysis was performed on the ditosyl deriva-
tives, precursors of the corresponding diazido compounds.
There are two factors that warrant large signal dispersion in the
1H NMR spectra of ditosyl (but not the diazido) compounds in
deuterated water: the departure from C7 molecular symmetry
that lifts the chemical equivalence of the glucopyranose units
and the ability of the tosyl group to form intra- or intermolecu-
lar inclusion complexes, as documented for 6-monotosyl-β-CD
[28] and the consequent local magnetic anisotropy effects on
the β-CD protons induced by the tosyl group confined in the
β-CD cavity. The strategy for NMR resonance assignment and
The inspection of the reversed-phase HPLC chromatograms of
reactions 1–3 (Figure 2) reveals that the individual diazido-β-
CD regioisomers have significant retention time differences if
water/acetonitrile gradient elution is applied. These differences
can be even more enhanced using methanol/water elution. The
high separation selectivity achieved for ditosyl regioisomers
allowed the method transfer from the HPLC columns to prepar-
ative columns, therefore larger quantities (gram scale) of the
pure regioisomers of homo-difunctionalized β-CDs became
readily available, allowing further chemical transformations on
714

Beilstein J. Org. Chem. 2019, 15, 710–720.
glucopyranose sequence analysis is discussed in detail in Sup- as in gg, but considerably large JH6’-H5 (≈7 Hz) vicinal cou-
porting Information File 1. Briefly, the well-resolved anomeric pling constants and gives rise to two signals: a doublet for H6
resonances allowed the identification of the spin system of each and an apparent doublet of doublets for H6’. The tg conforma-
glucopyranose unit among which the tosylated ones can be tion is the least populated in solution and is considered as unfa-
identified using the resonance frequencies of H6,6’ as an entry vorable (≈0% population) [30], especially if the substituent is
point. Three staggered conformers with respect to the bulky as in the present case. In the spectrum of 6A,6D-ditosyl-β-
TsO–C(6)–C(5)–H(5) dihedral angle (see Figure 3d for CD (Figure 3b, lower spectrum) the tosyl-substituted unit A
numbering) are possible in glucopyranose structures: gauche- presumably adopts a gg conformation as both H6,6’ protons
gauche (gg, 180°), gauche-trans (gt, −60°) and trans-gauche (tg, give rise to one doublet (yellow labelled, 4.28 ppm), whereas in
+60°) [29].
unit D (orange labelled) gt is the preferred conformation
because a doublet of doublets (H6’D, 4.16 ppm) and a doublet
The gg orientation, where both H6 and H6’ are rotated toward (H6D, 4.05 ppm) are observed. The gt seems also to be the
the cavity interior and the tosyl group is turned outwards, corre- predominant conformation in solution for both tosyl substitu-
sponds to large geminal JH6-H6’ (11.5 Hz) and very small ents in 6A,6C-ditosyl-β-CD (Figure 3b, middle spectrum) and
vicinal JH6-H5 ≈ JH6’-H5 (<1.5 Hz) coupling constants resulting 6A,6B-ditosyl-β-CD (Figure 3b, upper spectrum) as revealed by
in one H6,6’ doublet with somewhat broad components. The gt the patterns of the signals due to H6,6’. In summary, the orien-
orientation, on the other hand, is associated with small JH6-H5, tations of the tosyl groups as defined by the staggered confor-
Figure 3: NMR spectral regions of the three ditosyl regioisomers in D2O (500 MHz). The signals of the tosylated glucopyranose units are indicated in
yellow and orange in each spectrum: 1H NMR spectrum of all ditosyl derivatives indicating (a) the anomeric proton (H1) region (5.20 to 4.70 ppm) and
(b) the H6,6’-OTs proton region (4.30 to 3.90 ppm). (c) 2D TOCSY NMR spectrum of 6A,6D-ditosyl-β-CD: starting from each H1 resonance, identifica-
tion of the signals that belong to the same spin system (dotted lines) is possible leading to recognition and assignment of each of the glucopyranose
units. (d) Scheme for the clockwise connectivity in 6A,6D-ditosyl-β-CD.
715

Beilstein J. Org. Chem. 2019, 15, 710–720.
mations about the C5–C6 bond are generally of gt type, except Reactions 6, 7 and 8 gave the hetero-difunctionalized products
for one case, the gg-oriented H6,6’A-tosyl group in 6A,6D- that will be discussed in the next paragraph, reactions 9 and 10
ditosyl-β-CD. This implies that the group is rotated completely (from differently prepared starting materials) expectedly gave
outside its cavity, whereas in all other ditosyl derivatives posi- mixtures of diazidated products while reactions 11, 12 and 13
tioning of the tosyl group partially over the cavity is preferred. gave the single diazidated isomers (Scheme 3).
In all cases, however, the tosyl groups display ROESY cross-
peaks with the CD cavity protons (Supporting Information The reversed-phase HPLC chromatograms of azido-tosylated
File 1, Figure S11). The cross-peaks indicate the formation of CDs were comparable for reactions 4 and 5 in Scheme 3
intermolecular as well as intramolecular (self-inclusion) com- (Figure 4). Surprisingly, in both cases four well-separated com-
plexes in situ in D2O. This conclusion was confirmed by the ad- ponents were observed, instead of the expected three peaks of
dition of 1-adamantanecarboxylic acid into a solution of 6A,6D- the corresponding three regioisomers. The almost identical
ditosyl-β-CD. The 2D-ROESY spectra obtained (Supporting pattern of the two chromatograms indicates that these two
Information File 1, Figure S12) show absence of the formerly distinct reaction conditions give the desired hetero-difunctional-
observed correlation signals between cavity and tosyl protons ized products with comparable regiochemistry.
and emergence of new such signals between the cavity and the
adamantyl group protons. Thus, the latter group completely The unexpected appearance of one additional component in
displaces the tosyl moieties from its own and other CD cavities. both reactions could be attributed to secondary side substitution
although the plausibility that under both conditions only one
The regioisomeric diazido-β-CD products do not show suffi- secondary-side substituted product would form is very low.
cient dispersion of the 1H NMR signals to permit detailed as- It is more likely that the additional peak in both cases is the
signments, justified by the very small size of the azido groups result of the separation of one pseudoenantiomer pair by
and their inability to form efficient inclusion complexes. More- reversed-phase HPLC owing to in situ formation of pseudo-dia-
over, in the 13C NMR spectra only one C6-N3 signal is ob- stereoisomeric inclusion complexes with stationary-phase com-
served at ≈51 ppm in each case that resonates at very similar ponents.
frequencies in the 6A,6C- and 6A,6D-isomers (Δδ = 0.003 ppm),
the one of the A,B-regioisomer being more deshielded The symmetry and the 1:1 area ratio of the first pair of eluting
(Δδ ≈ 0.026 ppm) than the others (Supporting Information components in both chromatograms is an indication that these
File 1, Figure S14). The 6A,6B-regioisomer is also the least peaks might indeed belong to separated pseudoenantiomer
water-soluble. Consequently, structural verification was solely pairs. In order to prove this latter assumption and to obtain
based upon the analysis of the ditosyl precursors, as discussed hetero-difunctionalized β-CDs as single regioisomers, the
above.
6A,6D-, 6A,6C- and 6A,6B-ditosylated regioisomers prepared in
reaction 2 and reaction 1, respectively (Scheme 2), and separat-
ed by PCC, were converted to the corresponding 6A-mono-
Hetero-difunctionalized β-cyclodextrins
The hetero-difunctionalization was performed as a stepwise azido-6X-monotosyl-β-CD using NaN3 in defect, thus replacing
introduction of an azido and a tosyl moiety into the primary rim only one tosyl group in the molecule. This method afforded
of β-CD, in order to obtain a versatile intermediate which 47% conversion of starting material to the desired hetero-
allows independent manipulation of two orthogonal functions difunctionalized product, which consequently resulted in a
on the β-CD scaffold. Although the selectivity of the direct dito- single regioisomer and a 1:1 mixture of two pseudoenantiomers
sylation has been analyzed in the previous section, the side- (the clockwise and the counter-clockwise monoazidated prod-
selectivity for the tosylation of 6-monoazido-β-CD under uct). The 6A,6B- and the 6A,6C-ditosylates gave the hetero-
various conditions needs to be addressed. The tosylation of difunctionalized product as an inseparable mixture of the two
6-monoazido-β-CD was attempted for the first time using pseudoenantiomers (reaction 6 and reaction 7, Figure 4) while
TsCl either in pyridine or in a H2O/ACN mixture in the pres- 6A,6D-ditosylate gave a 1:1 mixture of pseudoenantiomers,
ence of copper(II) sulfate (reactions 4 and 5, respectively, baseline-separated by reversed-phase HPLC method (reaction 8,
Scheme 3).
Figure 4).
The product formation in both cases was ascertained by direct- With reactions 6, 7 and 8 all the possible regioisomers of
phase TLC, 1H NMR and reversed-phase HPLC. The mono- 6A-monoazido-6X-monotosyl-β-CD were prepared and due to
azido-monotosylated fraction was isolated using reversed-phase their different HPLC retention times, they can be used as refer-
PCC with water/methanol gradient elution in 35% yield for ence compounds to evaluate the substitution pattern in stepwise
reaction 4 and in 60% yield for reaction 5 (Scheme 3).
hetero-difunctionalizations (in reactions 4 and 5, Figure 4). The
716

Beilstein J. Org. Chem. 2019, 15, 710–720.
Scheme 3: Syntheses of 6A-monoazido-6X-monotosyl-β-CDs using starting materials obtained from different reaction conditions and their conversion
to 6A,6X-diazido-β-CDs.
comparison of the HPLC retention times of the components in part (6D-monoazido-6A-monotosyl-β-CD) using reversed-phase
reactions 4 and 5 with the retention times of the single regio- HPLC (Figure 5a), the separation of the azido-tosylated prod-
isomers prepared in reactions 6, 7 and 8 revealed that the first ucts was further investigated on CD-Screen stationary phase
pair of eluting components in both reactions are the separated [31], tailored to separate CD derivatives. This type of chroma-
pseudoenantiomers of the 6A,6D- and 6D,6A-substituted prod- tography separates CD derivatives based on their ability to form
uct, the third eluting peak belongs to the mixture of 6A,6C- and inclusion complexes with nitrophenol moieties, attached to the
6C,6A-substituted compound and the last eluting peak can be at- stationary phase.
tributed to the 6A,6B- and 6B,6A-substituted azido-tosylated-β-
CD. Since all the components of azido-tosylated products pre- This inclusion phenomenon clearly improved the separation of
pared in reaction 4 and 5 were identified and assigned as regio- the azido-tosylated regioisomers and the use of CD-Screen
isomers of 6A-monoazido-6X-monotosyl-β-CD, it can be con- column allowed complete or partial resolution of all the three
cluded that in both reaction conditions the introduction of the pairs of pseudoenantiomers (Figure 5b).
tosyl moiety is selective for the primary side.
Although the isolation of pure pseudoenantiomers of hetero-
Encouraged by the fact that 6A-monoazido-6D-monotosyl-β-CD difunctionalized CDs has not been the aim of the present study,
was baseline separated from its pseudoenantiomeric counter- inclusion-assisted chromatography clearly improved the resolu-
717

Beilstein J. Org. Chem. 2019, 15, 710–720.
Figure 4: Reversed-phase HPLC chromatograms of 6A-monoazido-6X-monotosyl-β-CDs prepared through reactions 4–8.
Figure 5: HPLC separation of regioisomers and pseudoenantiomers of 6A-monoazido-6X-monotosyl-β-CD prepared in reaction 4 on reversed-phase
stationary phase (a) and on CD-Screen stationary phase (b).
tion of pseudoenantiomeric pairs (Figure 5), therefore, with Figure 2), allowed for the evaluation of the regiochemical
further optimization this chromatographic method can be outcome of reactions 4–8 and unambiguously proved the side-
applied also for preparative purposes.
selectivity.
As a final proof for the side-selectivity of stepwise hetero- Reaction 9 gave three components in 35:31:34 area ratio (reac-
difunctionalizations, part of the azido-tosylates prepared in tion 9, Figure 6), having identical retention times to those of the
reactions 4–8 were transformed to the diazido-β-CDs (reactions three reference compounds (references 1, 2 and 3, Scheme 1).
9–13, Scheme 3). As the reversed-phase HPLC method was This leads us to the conclusion that the tosylation of
found to be powerful in separating the regioisomers of 6A,6X- 6-monoazido-β-CD in pyridine is selective for the primary rim
diazido-β-CDs, the comparison of retention times of the diazido and gives the three regioisomers with equal probability. Reac-
compounds prepared in reactions 9–13 (Figure 6) with those of tion 10 also gave only three components in 37:32:31 area ratio
the corresponding reference compounds (reference 1–3, (reaction 10, Figure 6), identified as 6A,6D-, 6A,6C- and 6A,6B-
718

Beilstein J. Org. Chem. 2019, 15, 710–720.
Figure 6: Reversed-phase HPLC chromatograms of 6A,6X-diazido-β-CDs prepared in reactions 9–13.
diazido-β-CD, which definitely proves that tosylation of regarding the impact of reaction conditions on the regioselectiv-
6-monoazido-β-CD under Cu(II)-assisted aqueous conditions is ity of disubstitution are:
also a side-selective process, with a slight preference towards
• The ditosylation of β-CD in pyridine is a primary-side process
that generates all the three theoretical regioisomers without
regioselectivity.
the formation of the 6A,6D–6D,6A-disubstituted product.
Conclusion
In the development of a direct strategy for β-CD difunctional-
ization, several focal points were clarified and important goals • The Cu(II)-mediated ditosylation of β-CD in aqueous solution
were accomplished. The three 6A,6X-ditosyl and three 6A,6X- is primary-side oriented and partially regioselective process as
diazido-β-CD regioisomers as well as the 6A-monoazido-6X- only two regioisomers of the ditosyl-β-CD are formed during
monotosyl-β-CD derivatives were firstly prepared in a multi- the process, 6A,6C- and 6A,6D-, with the 6A,6C-regioismer
gram scale using preparative column chromatographic purifica- being favored.
tion. These compounds are key intermediates for the straightfor-
ward regiospecific preparation of a large variety of new β-CD • The preparation of the primary-side diazido-β-CDs through a
difunctional and potentially bimodal derivatives, due to the one pot Vilsmeier–Haack/Appel-type iodination reaction also
orthogonality and versatility of the tosyl and azido functions. generates only two regioisomers, 6A,6C- and 6A,6D-, in higher
Furthermore, the presence of the tosyl groups conveniently yields than above and with the 6A,6D-regioisomer being slightly
rendered the spectra amenable to detailed NMR analysis. The favored.
unambiguous determination of the regiochemistry of difunction-
alized β-CDs was consequently accomplished for the first time • The monotosylation of 6-monoazido-β-CD (both in pyridine
without the need for chemical modification, enzymatic degrada- and Cu(II)-mediated in aqueous solution) is a primary-side
tion or reference material. Moreover, HPLC methods based on process that leads to the theoretical three couples of pseu-
reversed-phase elution were ad-hoc developed to quantify the doenantiomers.
ditosyl and diazido-β-CDs. Finally, the inclusion-assisted sepa-
ration of regioisomers and pseudoenantiomers of 6A-mono-
Supporting Information
azido-6X-monotosyl-β-CD by aid of a CD-Screen stationary
phase was realized for the first time, opening the way to the
Supporting Information File 1
preparative separation of these versatile derivatives.
Experimental details and compounds characterization.
[https://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-15-66-S1.pdf]
Besides the above successful preparation, purification and char-
acterization of difunctional CDs, the main specific conclusions
719

Beilstein J. Org. Chem. 2019, 15, 710–720.
17.Ghosh, R.; Zhang, P.; Wang, A.; Ling, C.-C. Angew. Chem., Int. Ed.
2012, 51, 1548–1552. doi:10.1002/anie.201105737
18.Xiao, S.; Yang, M.; Sinaÿ, P.; Blériot, Y.; Sollogoub, M.; Zhang, Y.
Eur. J. Org. Chem. 2010, 1510–1516. doi:10.1002/ejoc.200901230
19.Sollogoub, M. Synlett 2013, 24, 2629–2640.
Acknowledgements
The support of the program FP7-PEOPLE-ITN-2013, project
No 608407 (Cyclon Hit) is gratefully acknowledged. This work
was supported by the János Bolyai Research Scholarship of the
Hungarian Academy of Sciences and by the ÚNKP-18-4-SE-
121 Bolyai+ New National Excellence Program of the Ministry
of Human Capacities (S. Beni) as well as by National Research,
Development and Innovation Office (NRDI K_17 125093).
doi:10.1055/s-0033-1339877
20.Guieu, S.; Zaborova, E.; Blériot, Y.; Poli, G.; Jutand, A.; Madec, D.;
Prestat, G.; Sollogoub, M. Angew. Chem., Int. Ed. 2010, 49,
2314–2318. doi:10.1002/anie.200907156
21.Stephenson, R. J.; Wolber, F.; Plieger, P. G.; Harding, D. R. K.
Aust. J. Chem. 2016, 69, 328–335. doi:10.1071/ch15460
22.Matsui, Y.; Yokoi, T.; Mochida, K. Chem. Lett. 1976, 5, 1037–1040.
doi:10.1246/cl.1976.1037
ORCID® iDs
Mihály Bálint - https://orcid.org/0000-0001-6970-3562
Éva Fenyvesi - https://orcid.org/0000-0003-0609-7416
Szabolcs Béni - https://orcid.org/0000-0001-7056-6825
Konstantina Yannakopoulou - https://orcid.org/0000-0002-0725-4906
Milo Malanga - https://orcid.org/0000-0001-7952-0268
23.Law, H.; Benito, J. M.; García Fernández, J. M.; Jicsinszky, L.;
Crouzy, S.; Defaye, J. J. Phys. Chem. B 2011, 115, 7524–7532.
doi:10.1021/jp2035345
24.Fujita, K.; Matsunaga, A.; Imoto, T. Tetrahedron Lett. 1984, 25,
5533–5536. doi:10.1016/s0040-4039(01)81618-x
25.Benkovics, G.; Hodek, O.; Havlikova, M.; Bosakova, Z.; Coufal, P.;
Malanga, M.; Fenyvesi, E.; Darcsi, A.; Beni, S.; Jindrich, J.
Beilstein J. Org. Chem. 2016, 12, 97–109. doi:10.3762/bjoc.12.11
26.De Rango, C.; Charpin, P.; Navaza, J.; Keller, N.; Nicolis, I.; Villain, F.;
Coleman, A. W. J. Am. Chem. Soc. 1992, 114, 5475–5476.
doi:10.1021/ja00039a097
References
1. Szejtli, J. Chem. Rev. 1998, 98, 1743–1754. doi:10.1021/cr970022c
2. Takashima, Y.; Osaki, M.; Ishimaru, Y.; Yamaguchi, H.; Harada, A.
Angew. Chem., Int. Ed. 2011, 50, 7524–7528.
doi:10.1002/anie.201102834
27.Gadelle, A.; Defaye, J. Angew. Chem. 1991, 103, 94–95.
doi:10.1002/ange.19911030121
3. Yuan, D.-Q.; Kitagawa, Y.; Aoyama, K.; Douke, T.; Fukudome, M.;
Fujita, K. Angew. Chem., Int. Ed. 2007, 46, 5024–5027.
doi:10.1002/anie.200701156
28.Djedaini-Pilard, F.; Gosnat, M.; Steinbruckner, S.; Dalbiez, J. P.;
Crini, G.; Perly, B.; Gadelle, A. In Proceedings of the IX International
Symposium on Cyclodextrins, Santiago de Compostella, May 31–June
3, 1998; Labandeira, J. J.; Vila-Jato, J. L., Eds.; Springer
Science+Business Media: Dordrecht, 1999; pp 73–76.
29.Rao, V. S. R. Conformation of Carbohydrates; Harwood Academic
Publishers: Amsterdam, The Netherlands, 1998; pp 65–68.
30.Piras, L.; Theodossiou, T. A.; Manouilidou, M. D.; Lazarou, Y. G.;
Sortino, S.; Yannakopoulou, K. Chem. – Asian J. 2013, 8, 2768–2778.
doi:10.1002/asia.201300543
4. Dong, S. D.; Breslow, R. Tetrahedron Lett. 1998, 39, 9343–9346.
doi:10.1016/s0040-4039(98)02160-1
5. Řezanka, P.; Navrátilová, K.; Řezanka, M.; Král, V.; Sýkora, D.
Electrophoresis 2014, 35, 2701–2721. doi:10.1002/elps.201400145
6. Yao, H.; Qi, M.; Liu, Y.; Tian, W. Chem. – Eur. J. 2016, 22, 8508–8519.
doi:10.1002/chem.201601142
7. Malanga, M.; Darcsi, A.; Balint, M.; Benkovics, G.; Sohajda, T.; Beni, S.
Beilstein J. Org. Chem. 2016, 12, 537–548. doi:10.3762/bjoc.12.53
8. Bom, A.; Bradley, M.; Cameron, K.; Clark, J. K.; van Egmond, J.;
Feilden, H.; MacLean, E. J.; Muir, A. W.; Palin, R.; Rees, D. C.;
Zhang, M. Q. Angew. Chem., Int. Ed. 2002, 41, 265–270.
doi:10.1002/1521-3773(20020118)41:2<265::aid-anie265>3.0.co;2-q
9. Matsuo, M.; Togawa, M.; Hirabaru, K.; Mochinaga, S.; Narita, A.;
Adachi, M.; Egashira, M.; Irie, T.; Ohno, K. Mol. Genet. Metab. 2013,
108, 76–81. doi:10.1016/j.ymgme.2012.11.005
31.Szemán, J.; Csabai, K.; Kékesi, K.; Szente, L.; Varga, G.
J. Chromatogr. A 2006, 1116, 76–82.
doi:10.1016/j.chroma.2006.03.019
License and Terms
10.Tabushi, I.; Shimokawa, K.; Fujita, K. Tetrahedron Lett. 1977, 18,
1527–1530. doi:10.1016/s0040-4039(01)93093-x
This is an Open Access article under the terms of the
Creative Commons Attribution License
11.Tabushi, I.; Nabeshima, T.; Fujita, K.; Matsunaga, A.; Imoto, T.
J. Org. Chem. 1985, 50, 2638–2643. doi:10.1021/jo00215a008
12.Tabushi, I.; Yamamura, K.; Nabeshima, T. J. Am. Chem. Soc. 1984,
106, 5267–5270. doi:10.1021/ja00330a039
(http://creativecommons.org/licenses/by/4.0). Please note
that the reuse, redistribution and reproduction in particular
requires that the authors and source are credited.
13.Tabushi, I.; Yuan, L. C.; Shimokawa, K.; Yokota, K.-i.; Mizutani, T.;
Kuroda, Y. Tetrahedron Lett. 1981, 22, 2273–2276.
doi:10.1016/s0040-4039(01)92908-9
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
14.Gramage-Doria, R.; Rodriguez-Lucena, D.; Armspach, D.; Egloff, C.;
Jouffroy, M.; Matt, D.; Toupet, L. Chem. – Eur. J. 2011, 17, 3911–3921.
doi:10.1002/chem.201002541
(https://www.beilstein-journals.org/bjoc)
15.Pearce, A. J.; Sinaÿ, P. Angew. Chem., Int. Ed. 2000, 39, 3610–3612.
doi:10.1002/1521-3773(20001016)39:20<3610::aid-anie3610>3.0.co;2-
v
The definitive version of this article is the electronic one
which can be found at:
16.Guieu, S.; Sollogoub, M. J. Org. Chem. 2008, 73, 2819–2828.
doi:10.1021/jo7027085
doi:10.3762/bjoc.15.66
720