Synthesis of glycosyl derivatives as dopamine prodrugs: Interaction with glucose carrier GLUT-1

Article abstract of DOI:10.1039/b212066f

Glucosyl dopamine (DA) derivatives may represent a new class of DA prodrugs that would interact with glucose transporter GLUT-1, present in the blood-brain barrier, and generate DA in the brain. Therefore, compounds bearing the sugar moiety linked to eith

Full text of DOI:10.1039/b212066f

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of glycosyl derivatives as dopamine prodrugs:
interaction with glucose carrier GLUT-1†
Caridad Fernández,a Ofelia Nieto,a José Angel Fontenla,c Emilia Rivas,c
María L. de Ceballos*b and Alfonso Fernández-Mayoralas*a
a Instituto de Química Orgánica General, CSIC, Juan de la Cierva, 3, 28006 Madrid, Spain.
E-mail: iqofm68@iqog.csic; Fax: ϩ34 91 564 48 53
b Instituto Cajal, CSIC, Doctor Arce 37, 28002 Madrid, Spain
c Departamento de Farmacología, Facultad de Farmacia,
Universidad de Santiago de Compostela, Campus Sur s/n, 15782 Santiago de Compostela,
Spain
Received 9th December 2002, Accepted 29th January 2003
First published as an Advance Article on the web 7th February 2003
Glucosyl dopamine (DA) derivatives may represent a new
class of DA prodrugs that would interact with glucose
transporter GLUT-1, present in the blood–brain barrier,
and generate DA in the brain. Therefore, compounds bear-
ing the sugar moiety linked to either the amino group or
the catechol ring of DA through amide, ester, carbamate,
peptide or glycosidic bonds were synthesized. The behavior
of the compounds as prodrugs was monitored in different
media and the affinity of the glycoconjugates for the
glucose carrier GLUT-1 using human erythrocytes was also
studied. Most of the compounds were markedly stable in
buffer and plasma, and several compounds released DA
when incubated with brain extracts and the rate was related
to the bond linking DA with glucose. The new glucosyl
conjugates substituted at the C-6 position of the sugar were
more potent inhibitors of glucose transport when com-
pared to C-1 and C-3 substituted derivatives. This work
provides structure–activity information about the inter-
action of substituted glucose with the GLUT-1 transporter.
brain through the neutral amino acid carrier where it is
enzymatically converted into DA. However upon continuous
levodopa treatment a variety of problems may appear (e.g.
fluctuations in the clinical response and onset of involuntary
movements).9,10
We have recently initiated studies on a new approach to
deliver DA into the central nervous system (CNS) by linking
the neurotransmitter to a sugar molecule so that the resulting
glycoconjugate may cross the BBB using the glucose carrier
GLUT-1.11 Once transported into the CNS the prodrug would
be hydrolyzed to release DA. In order to reach and provide the
brain with the neurotransmitter the glycoconjugates must fulfil
several requirements. They must be stable in the periphery,
should be recognized by the carrier GLUT-1 and transported
into the brain, and, finally, be hydrolyzed by the action of
brain enzymes to release DA. Some authors have studied this
approach to deliver peptide drugs12 and cytotoxic agents13,14
into the CNS.
In the present work we describe the synthesis and biological
activities of new glycoconjugates in which DA is attached to
different positions of the glucose molecule using a variety of
linkages susceptible of enzymatic hydrolysis. A series of com-
pounds had the amino group of DA linked to C-6, C-3 and
C-1 of the sugar through a succinyl linker (compounds 1–3,
Fig. 1) or a carbamate bond (4–8). In another series, the sugar
was attached to the phenol group of DA through glycosidic
(9 and 10) and ester (11–13) bonds. In order to know their
properties as prodrugs, in vitro stability studies of the com-
pounds have been performed in plasma and rat brain
extracts. Finally, the affinity of the glycoconjugates for the
glucose carrier GLUT-1 using human erythrocytes was studied
as well.
Introduction
The blood–brain barrier (BBB) provides an efficient protection
of the brain, maintaining the extracellular concentrations of
ions, neurotransmitters and growth factors. Indeed hydrophilic
compounds are excluded by the BBB and only small lipophilic
molecules may cross it. This ability relies on the structure of
the brain vessels composed of endothelial cells joined by tight
junctions, lined with astrocyte projections. Only a few sub-
stances necessary for the brain’s proper function are able to
cross this barrier.1–3 Thus, brain transport of essential hydro-
philic nutrients is mediated by a number of specific carriers,
which are located in the plasma membrane of the BBB endo-
thelial cells.4 This is the case for glucose, the main energy source
of the brain, whose passage is facilitated by the glucose carrier
GLUT-1.5–7
However, the presence of the BBB hinders pharmacological
delivery to the brain thus making difficult the treatment of
certain diseases with drugs.3 This is the case with Parkinson’s
disease, which is characterized by the dramatic reduction of
dopamine (DA) levels in basal ganglia as a consequence of
the degeneration of DA neurons in the substantia nigra. The
use of DA itself to treat the disease is precluded by the inability
of this neurotransmitter to cross the BBB. The most widely
employed treatment of this disorder is levodopa,8 the immedi-
ate DA precursor. Levodopa is actively transported into the
Results and discussion
Synthesis of glycoconjugates
The synthesis of compounds 1–3, 9 and 10 has been previously
described.11 Compounds 4–7, with a carbamate bond linking
the sugar and the neurotransmitter were synthesized by
activation of the corresponding hydroxyl group of the sugar
with N,N-carbonyldiimidazole (CDI) followed by coupling
with the amine group of DA. Thus, the partially benzylated
glucoses having free hydroxyl HO-6, HO-3 or HO-1 (com-
pounds 17, 21, and 24, Scheme 1) were first obtained by
selective protection/deprotection steps.11,15
The reaction of 17 with CDI was performed in dioxane to
give 18 in excellent yield (94%). The imidazolylcarbonyl
derivative thus obtained was treated with DA derivative 1911 in
the presence of Et3N afford the protected glycoconjugate 20 in
† Electronic supplementary information (ESI) available: experimental
details for the preparation of all derivatives and biological assays. See
http://www.rsc.org/suppdata/ob/b2/b212066f/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 7 6 7 – 7 7 1
767

View Article Online
Table 1 Half-life (hours) of glycoconjugates in different media
Incubation media
Compound
Type of bond between sugar and DA
Buffer (pH 7.0)
Plasma
Brain extract
1
Ester–amide
Carbamate
Carbamate
Peptide–carbamate
Glycosidic
Glycosidic
Ester
>48
>48
>48
30
>48
>48
0.08
2.5
3
3
1.3
24
>48
>48
—
0.1
2
21
22
12
30
134
67
—
0.4
3
4
6,7a
8
9
10
11 ؉ 12
13
Ester
—
Dopamine
27
a A 2 : 1 mixture of 6 and 7 was used.
Fig. 1 Glycosyl dopamine derivatives.
87% yield. Removal of the benzyl groups by hydrogenolysis led
to 4. A similar procedure was applied to the alcohol 21, but
using THF instead of dioxane in the reaction with CDI. The
subsequent condensation with 19 followed by debenzylation
furnished 5.
In the case of the activation of alcohol 24 with CDI we faced
the problem of the stereoselectivity at the anomeric center.16–18
We tried the reaction varying solvent (dioxane, THF, CH2Cl2,
and Et2O), temperature (80 ЊC, rt, and Ϫ20 ЊC), and in the
presence or absence of a base, giving in all cases anomeric
mixtures of imidazolylcarbonyl derivatives 25 (α) and 27 (β)
with the α-anomer predominating. The best result was obtained
in Et2O at room temperature to afford a mixture of 25 and 27 in
excellent yield and good α-stereoselectivity (98%, α : β, 80 : 20).
Treatment of 25 with 19 gave the protected glycoconjugate 26
from which benzyl groups were removed by hydrogenolysis to
give 6.
activation of hydroxyl groups through the formation of stannyl
ethers19 was used. Our working hypothesis was that coordin-
ation of the tin atom with the ring oxygen would be more
favorable in the β- than in the α-configured tin ether, allowing
the subsequent reaction to proceed selectively towards the
formation of the anomeric carbamate with β-configuration.
The reaction of 24 with (Bu3Sn)2O at 110 ЊC followed by treat-
ment with CDI at room temperature, gave an anomeric mixture
of imidazolylcarbonyl derivatives 25 (α) and 27 (β) in high
yield (94%) with the β-anomer as the major compound (α : β,
30 : 70). This mixture without further separation was coupled
with the amine 19 and, then chromatographed to give com-
pound 28. Benzyl groups were removed by hydrogenation to
afford 7.
The synthesis of the glycoconjugate containing the phenyl-
alanine unit (compound 8) was performed using the inter-
mediate 30, previously obtained by treatment of 19 with N-Boc
protected amino acid 29 (90%). Deprotection of the Boc group
For the stereoselective synthesis of the β-anomer the
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 7 6 7 – 7 7 1
768

View Article Online
Scheme 1 (a) TrCl (1.5 equiv), DMAP (0.2 equiv), pyridine, 60 ЊC;
(b) BnBr (3.3 equiv), NaH (3.7 equiv), DMF; (c) p-TsOH (0.2 equiv),
CH2Cl2–MeOH; (d) N,N-carbonyldiimidazole (1.2 equiv), dioxane;
(e) 19 (0.85 equiv), Et3N, THF, 80 ЊC; (f ) H2, Pd/C, AcOEt–MeOH–
toluene 4 : 3 : 3; (g) N,N-carbonyldiimidazole (1.2 equiv), THF;
(h) N,N-carbonyldiimidazole, Et2O; (i) (Bu3Sn)2 (1 equiv), toluene,
120 ЊC, 2 h, then, N,N-carbonyldiimidazole (2 equiv), rt.
Scheme 2 (a) 19 (1 equiv), 29 (1 equiv), Et3N (4 equiv), BOP
(1.1 equiv), THF; (b) TFA, CH2Cl2; (c) 18 (1.2 equiv), Et3N, THF,
80 ЊC; (d) H2, Pd/C, AcOEt–MeOH–toluene 4 : 3 : 3; (e) 33 (1 equiv),
34 (0.66 equiv), diisopropylcarbodiimide (DIIC) (1.3 equiv), CH2Cl2.
media, together with that of DA. For comparative purposes the
previously reported11 results of 1, 9, and 10 are also included
in the Table. Considering the different linkages established
between sugar and DA, the latter should be released after
hydrolysis of ester (11, 12, and 13), amide (1), glycosidic (9, and
10), or carbamate (4, 6, and 8) bonds.
Most of the compounds were more stable in buffer than DA,
with the exception of compounds 11, 12, and 13 which showed
labile phenol ester bonds. The monoesters 11 and 12 were
hydrolyzed more rapidly than diester 13, indicating that the
unsubstituted phenol group must accelerate the hydrolysis of
the adjacent ester by anchimeric assistance during the reaction.
In the case of the succinamate 1 hydrolysis took place first on
the ester bond, to give a succinamic acid which accumulated
during the incubation period without producing DA. The high
stability of the amide bond in 1 is in accordance with literature
data on other prodrugs containing this type of bond.20 Among
the glycoconjugates containing a carbamate bond (4, 6, 7, and
8), the compounds with the linkage at the anomeric position
of the glucose (6,7) exhibited the highest rate of hydrolysis
in 30 followed by coupling with the glucose derivative 18 gave
32 (78%), which after debenzylation afforded 8.
Finally, the phenol esters 11–13 were synthesized from the
acid 33, previously obtained by reaction of 17 with succinic
anhydride (Scheme 2). Esterification of 33 with the N-Cbz pro-
tected DA 34 gave the diester 37 and a mixture of monoesters
35 and 36. Hydrogenolysis of the benzyl groups on 37 gave
target 13, and the same treatment on 35/36 afforded a mixture
of monoesters 11 and 12.
Stability studies
To determine the stability in physiological media and the ability
to deliver DA after brain uptake, some of the compounds were
incubated in buffer (pH 7.0), rat plasma, and rat brain extracts.
The progress of the incubations were monitored by HPLC
(UV detector, λ = 280 nm) analyzing the disappearance of the
glycoconjugate and the formation of DA. Table 1 reports the
half-life (t ) of the compounds in the different incubation
½
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 7 6 7 – 7 7 1
769

View Article Online
Table 2 IC50 values of the glycoconjugates for the inhibition of
glucose uptake
in all the media. Carbamate 8, having an internal phenylalanine
unit, was hydrolyzed in the presence of α-chymotrypsin
from bovine pancreas (30 units per mL, t 6.5 h). This result
½
Compound
Linkage at the hexose
IC50/mM
was expected given the primary specificity of this protease
for an aromatic amino acid at the amino-terminal unit side
of the scissile bond. On the other hand, the disappearance
of 4 and 8 in plasma and brain extracts was not accompanied
by DA formation. This suggests that 4 and 8 may be cleaved
in a different way from hydrolysis of carbamate bond, which
does not release DA (for instance, oxidation of the catechol
ring).
The glycosides 9 and 10 showed an extraordinary stability in
plasma. After 48 h of incubation no appreciable transformation
of 9 and 10 was observed. Their high stability in this medium
is probably due to the protection of the phenolic groups from
oxidation to quinones by the glycosyl substituents, together
with the fact that in the blood low levels of β-glucosidases
are found. Indeed, β-glucosides of catecholamines occur in
nature,21 and such a conjugation has been interpreted as a
protection from oxidation to quinones by the ubiquitous
phenoloxidases in tissues. In brain extract, the glycosides were
slowly hydrolyzed to give DA.
3
C-1 (β)
C-1 (α)
C-1 (β)
C-1 (β)
C-3
C-3
C-6
C-6
C-6
C-6
C-6
>100.0
50.0
31.1
>100.0
60.3
51.2
12.1
1.5
2.2
76.8
6
7
9
5
3-MG
1
4
8
13
2
>100.0
The inhibition constant Ki was determined with the best
inhibitor, carbamate 4. The Ki (0.791 0.058 mM) for 4 was
almost 15 times lower than the KM (11.0 mM) of -glucose.
Therefore, it appears that 4 binds more effectively to the carrier
than glucose itself.
Glucose uptake inhibition was dependent on the nature of
the sugar since the galactose derivative 2 was a poor inhibitor
(IC50 > 100.0 mM). It has been suggested that HO-4 of glucose
establishes a hydrogen bond with the transporter. Given that
galactose is the epimer at the C-4 position of glucose, this inter-
action may not be possible.23
In conclusion, anomeric carbamates 6 and 7, glycosides 9 and
10, and diester 13 released DA after incubation in brain extract,
in the following stability order: ester ӷ carbamate > glycoside.
In the case of the compounds 4 and 8 exhibiting a carbamate
linkage, formation of DA was not observed after their incu-
bation in plasma and brain extract.
Conclusions
For the targeted delivery of dopamine into the CNS using the
glucose transport system, the glycosyl dopamine derivatives
must show affinity for the GLUT-1 carrier and a good stability
balance in the periphery and the brain. The results of glucose
uptake inhibition by the glycoconjugates indicate that except
for glucose derivatives substituted at position C-6, all other
modifications of the sugar gave compounds with moderate
or no binding to the carrier. Glycosides 9 and 10, although
fulfilling the prodrug criteria since they exhibited an extra-
ordinarily high stability in plasma and a sustained release of
DA in brain extract, can be considered unsuitable candidates
due to their lack of affinity for GLUT-1. On the other hand,
glycoconjugates with an ester linkage (10–12) were found too
labile and may be hydrolyzed in the periphery. Carbamate
derivatives 4, 6, and 7 seem to be the prodrugs of choice. Com-
pound 4 showed the best affinity for GLUT-1, even higher than
glucose itself. However, when 4 was incubated in brain extract
no DA was observed, probably as a consequence of oxidation
of the catechol ring occurring more rapidly than hydrolysis
of the carbamate bond. Better results gave the anomeric
carbamate 6 and 7. They showed a moderate affinity for
GLUT-1, adequate stability in plasma and released DA when
incubated in brain extract.
Inhibition of glucose uptake by glycosyl dopamine derivatives
Next we examined the ability of the glycoconjugates to inhibit
glucose transport using human erythrocytes, which express the
same GLUT-1 transporter present in the BBB. Glucose uptake
by erythrocytes was determined using 14C-labeled glucose
following a published procedure.14,22
In Table 2 the values of IC50 obtained for the different glyco-
conjugates are shown. Monoesters 11 and 12 were not assayed
due to their low stability. For comparative purposes the 3-O-
methyl--glucose (3-MG) was also tested.
Compounds with the substitution at the C-1 position of
glucose were moderate or weak inhibitors depending on both
the anomeric configuration and the type of linker used. The
β-configured carbamate 7 was a better inhibitor than the α-
anomer 6, and both were more effective than glucopyranoside 9
and succinamate 3.
Previous observations suggest that increasing the substituent
size at the C-3 position of glucose caused a reduction in
affinity.23 However, compound 5, in which DA is forming
a carbamate linkage with the hydroxyl HO-3 of glucose
showed an IC50 value of 60.3 mM, similar to that obtained
with the less bulky 3-MG. Nevertheless, the presence of an
aromatic ring in 5 could produce additional interactions with
the protein.
Glucose derivatives substituted at the C-6 position showed
the highest affinities for the GLUT-1. Thus, 6-O-carbamates 4
and 8 had IC50 values of 1.5 and 2.2 mM, respectively, and the
succinamate 1 had a value of 12.1 mM. The diester 13, however,
was a poor inhibitor of glucose uptake probably due to its
large size. The high affinities of the C-6 modified glucoses are
consistent with a proposed model23,24 of glucose transport that
involves a hydrophobic site near the 6-position of bound
glucose, which can accommodate relatively bulky and apolar
substituents. In our compounds there could be a stabilizing
interaction of the aromatic ring of the DA unit with a hydro-
phobic residue in the active site of the protein. Moreover,
the difference of inhibitory effect between carbamate 4 and
succinamate 1 suggests that the distance of the aromatic ring to
the C-6 position of the sugar is important for activity.
In conclusion, this work presents efficient synthetic routes to
new glycosyl dopamine derivatives and has provided structure–
activity data about the interaction of substituted glucose to the
transporter GLUT-1 that may pave the way for drug delivery
for the treatment of CNS or other diseases.
Experimental
Detailed information on the preparation of all derivatives and
the biological assays is deposited as Electronic Supplementary
Information (ESI).†
Acknowledgements
Financial support by Ministerio de Ciencia y Tecnología
(grant BQU2001-1503 and fellowship to C.F.) is gratefully
acknowledged.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 7 6 7 – 7 7 1
770

View Article Online
14 C. Uriel, M.-J. Egron, M. Santarromana, D. Scherman,
References
K. Antonakis and J. Herscovici, Bioorg. Med. Chem., 1996, 4,
2081–2090.
1 Y. Madrid, L. F. Langer, H. Brem and R. Langer, Adv. Pharmacol.,
1991, 22, 299–324.
2 Y. Takakura, K. L. Audus and R. Borchardt, Adv. Pharmacol., 1991,
22, 137–165.
3 K. L. Audus, P. J. Chikhale, D. W. Miller, S. E. Thompson and
R. T. Borchardt, Adv. Drug Res., 1992, 23, 1–63.
4 I. Tamai and A. Tsuji, J. Pharm. Sci., 2000, 89, 1371–1388.
5 H. Lund-Andersen, Physiol. Rev., 1979, 59, 305–349.
6 S. A. Baldwin, Biochim. Biophys. Acta, 1993, 1154, 17–49.
7 M. Mueckler, Eur. J. Biochem., 1994, 219, 713–725.
8 A. Barbeau, Trends Pharmacol. Sci., 1981, 2, 297–299.
9 C. D. Marsden and J. D. Parkes, Lancet, 1976, 1, 292–296.
10 D. B. Calne, New Engl. J. Med., 1984, 310, 523–524.
11 C. Fernández, O. Nieto, E. Rivas, G. Montenegro, J. A. Fontenla
and A. Fernández-Mayoralas, Carbohydr. Res., 2000, 327, 353–365.
12 R. Polt, F. Porreca, L. Z. Szabò, E. J. Bilsky, P. Davis,
T. J. Abbruscato, T. P. Davis, R. Horvath, H. I. Yamamura and
V. J. Hruby, Proc. Natl. Acad. Sci. USA, 1994, 91, 7114–7118.
13 T. Halmos, M. Santarromana, K. Antonakis and D. Sherman,
Eur. J. Pharmacol., 1996, 318, 477–484.
15 E. Decoster, J.-M. Lacombe, J. L. Strefler, B. Ferrari and
A. A. Pavia, J. Carbohydr. Chem., 1983, 2, 329–341.
16 T. Kappes and H. Waldmann, Carbohydr. Res., 1998, 305, 341–349.
17 D. B. A. de. Bont, R. G. G. Leenders, H. J. Haisma, I. Van der
Meulen-Muileman and H. W. Scheeren, Bioorg. Med. Chem., 1997,
5, 405–414.
18 R. G. G. Leenders, K. A. A. Gerrits, R. Ruijtenbeek,
H. W. Scheeren, H. J. Haisma and E. Boven, Tetrahedron Lett., 1995,
36, 1701–1704.
19 S. David and S. Hanessian, Tetrahedron, 1985, 41, 643–663.
20 F. M. Menger and M. J. Rourk, J. Org. Chem., 1997, 62, 9083–9088.
21 T. L. Hopkins, T. D. Morgan, D. D. Mueller, K. B. Tomer and
K. J. Kramer, Insect Biochem. Mol. Biol., 1995, 25, 29–37.
22 A. G. Lowe and A. R. Walmsley, Biochim. Biophys. Acta, 1986, 857,
146–154.
23 J. E. G. Barnett, G. D. Holman and K. A. Munday, Biochem. J.,
1973, 131, 211–221.
24 J. E. G. Barnett, G. D. Holman, R. A. Chalkley and K. A. Munday,
Biochem. J., 1975, 145, 417–429.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 7 6 7 – 7 7 1
771