Rigidified multivalent lactose molecules and their interactions with mammalian galectins: A route to selective inhibitors

Article abstract of DOI:10.1039/b210923a

New and rigid multivalent lactose molecules were prepared. The structures contain lactose-2-aminothiazoline units at the periphery that were formed from a cyclisation of the thiourea sulphur onto the triple bond of the spacer. The lactosides were evaluate

Full text of DOI:10.1039/b210923a

View Article Online / Journal Homepage / Table of Contents for this issue
Rigidified multivalent lactose molecules and their interactions with
mammalian galectins: a route to selective inhibitors
Ioannis Vrasidas,a Sabine André,b Paola Valentini,a Corina Böck,b Martin Lensch,b
Herbert Kaltner,b Rob M. J. Liskamp,a Hans-J. Gabiusb and Roland J. Pieters*a
a Department of Medicinal Chemistry, Utrecht Institute for Pharmaceutical Sciences,
Utrecht University, PO Box 80082, NL-3508 TB Utrecht, The Netherlands
b Institut für Physiologische Chemie, Tierärztliche Fakultät,
Ludwig-Maximilians-Universität München, Veterinärstr. 13, D-80539 München, Germany
Received 2nd November 2002, Accepted 9th January 2003
First published as an Advance Article on the web 7th February 2003
New and rigid multivalent lactose molecules were prepared. The structures contain lactose-2-aminothiazoline units at
the periphery that were formed from a cyclisation of the thiourea sulphur onto the triple bond of the spacer. The
lactosides were evaluated as inhibitors against lectin binding in a solid phase inhibition assay. In this assay the
glycoprotein asialofetuin was immobilized onto the surface of microtiter plate wells, mimicking cell surface
presentation, while mammalian galectins-1, -3 or -5 were in solution. Between the three galectins, the folding pattern
and sequence are closely related but the topology of presentation of the carbohydrate recognition domains differs.
Strong multivalency effects were observed for the tetravalent lactoside in the inhibition of galectin-3 binding with
enhancements of almost 4300-fold compared to lactose. Remarkable selectivity was obtained in the inhibition since
relative potencies of the tetravalent lactoside with the proto type galectins-1 and -5 did not exceed a factor of 143
relative to lactose. The binding of the lactosides to galectin-3 was also studied by fluorescence spectroscopy with all
components in solution. These studies showed no multivalency effects in the inherent binding affinities.
tandem repeat type family are both divalent cross-linkers,
Introduction
whereas galectin-3 is established by the carbohydrate recog-
Throughout nature protein(lectin)–carbohydrate interactions
niton domain linked to a stalk section.7 The orientation of the
play a variety of important roles.1 With the continuous
appearance of more and more discoveries, for example in the
role of lectins as cell adhesion molecules, it is becoming increas-
binding sites in galectins is such that effective chelation by a
spaced divalent carbohydrate is not likely.9 Many of the bio-
logical effects of galectins are believed to be due to their ability
ingly clear that interference with these interactions has great
to act as cross-linkers.10 Our efforts towards rigidification of the
potential in medicine.2 One hurdle to be overcome on the
spacers should be viewed as affecting the cross-linking, rather
than aiming at chelation of the binding sites. The effects of
rigidification are more often than not disadvantageous, but
when the proper geometries can be created large benefits can be
expected. Notably, the topological factor, properly exploited,
might lead to galectin-type selective compounds. In this
respect, comparative analysis of proto type galectins (in our
study, galectins-1 and -5) against galectin-3 is a point of inter-
way to therapeutics is the weak millimolar affinity typically
encountered between monovalent carbohydrates and their
receptor proteins. Looking at biological systems several
research groups have taken the lesson to heart that higher
affinities can be attained by linking sugars together to produce
multivalent systems that can exhibit greatly improved affinities.
The area of multivalency has been extensively reviewed3 and
while the benefits are universally accepted it is often not yet
clear what the origin of the enhanced affinities is. In some
est. In fact, galectin-3 is known to be a monomer in solution but
aggregates when bound to surface ligands and does so with
cases the chelate effect and favorable entropic contributions are
positive cooperativity.11 Thus, we performed this analysis with
probably involved4 while in others aggregation into a lattice-
the aim of finding out whether strong and selective inhibition
of a distinct galectin could be detected. Such a selectivity might
like network plays a role.5
Despite the fact that the causes of multivalency effects are
find various uses since galectins have a characteristic profile
largely unknown it is worthwhile, for the long-term goal of
of biomedical functions. One example was found in a colon
optimising increases in binding affinities, to make variations in
model, where liver metastasis was correlated with galectin-3
the spacers between linked carbohydrate units. To this end we
expression. Another was discovered in a neuroblastoma
have started work to make rigidified versions of previously
model, where galectin-1 induced non-classical apoptosis of the
reported lactose dendrimers.6 We report herein on the synthesis
tumor cells akin to mechanisms of spontaneous regression
of these rigidified multivalent lactose systems and their
while galectin-3 blocks this effect. A third example can be the
biological evaluation in solid phase binding assays of three
observation that in glioblastoma cells galectin-1 is a major
mammalian galectins. We deliberately focused on this lectin
factor for the invasive behavior.12 As noted above, the topology
class instead of commonly studied plant agglutinins to add an
of the multivalent carbohydrate–lectin binding is assumed to be
immediate medical perspective. Galectins are animal Ca2ϩ
independent carbohydrate binding proteins with a β-galactos-
ide specificity, that share a jelly roll like folding pattern.7 They
crucial for important signalling phenomena.
are involved in cell–cell and cell–matrix adhesion, cell migration
and growth regulation with relevance to inflammation and
tumor spread. Because of their involvement in these important
processes the galectins are attractive target proteins for syn-
thetic ligand design.2d,8 For multivalency effects the topological
display of the binding sites is of interest. The proto type and
Results and discussion
Synthesis
We previously prepared and tested lactose dendrimers based
on a flexible 3,5-di-(2-aminoethoxy)benzoic acid branching
unit and evaluated them up to the third generation (octavalent)
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, 8 0 3 – 8 1 0
803

View Article Online
in inhibition studies with galectins and non-homologous
galactose specific carbohydrate binding proteins.13 We then
started to introduce the more rigid propargylic amine spacers
into the dendritic branching unit. In the case of a monovalent
version, as we showed recently,14 reaction of lactose isothio-
cyanate led to an initially formed thiourea-linked sugar, which
underwent a facile conversion to a 2-aminothiazoline moiety.
This was due to attack of the thiourea sulfur onto the triple
bond and resulted in compound 1 (Fig. 1). The structure of this
compound was identified by several two-dimensional NMR
techniques, and the compound turned out to be a remarkably
strong binder of the cholera toxin B-subunit.
were coupled by means of the BOP coupling reagent16 and gave
the protected scaffold 7, which after treatment with TFA
yielded the tetraamino compound 8, in its poly TFA salt form.
The stage was now set for the coupling of the di- and tetra-
amine salts 5 and 8 to acetylated lactose β-isothiocyanate
(Scheme 2). The coupling of 5 was performed in CH2Cl2 in the
presence of iPr2NEt. In order to facilitate the cyclisations14 of
the intermediate to the 2-aminothiazoline rings, the crude
product was exposed to acetic acid in CH2Cl2. The desired
product 9 was isolated by column chromatography from a
mixture of two other byproducts. These two compounds were
identified by mass spectrometry as oxidation products of 9
containing one or two additional oxygen atoms, presumably
attached to sulfur of the 2-aminothazoline ring. In a similar
way 8 was treated with lactose β-isothiocyanate in acetonitrile
and iPr2NEt to yield, after work up and chromatographic puri-
fication, the tetravalent lactose derivative 11. Characterization
1
of 9 and 11 was difficult by H-NMR spectroscopy due to the
large broadening of the signals, consistent with the reduced
mobility of the rigid arms linking the sugars. However, crucial
signals in the H-NMR spectra at δ 6.6 ppm representing the
Fig. 1 Structure of monovalent lactoside 1.
1
bridging olefinic protons were observed for both compounds.
Characteristic signals of the 2-aminothiazoline moiety of 9 and
both the 2-aminothiazoline unit and the triple bonds of 11
could clearly be identified in the 13C-NMR spectra. Further
confirmation of the structure assignment came from the use
of HSQC and HMBC techniques and gave correlations also
observed for 1.14 ESI mass spectra from 9 and 11 showed the
[M ϩ 2H]2ϩ (100%) ion at m/z = 799.95 and [M ϩ 3H]3ϩ (100%)
ion at m/z = 1125.40, respectively, along with some galactose
cleavage as observed before in the mass spectra.6 Both divalent
9 and tetravalent 11 were deprotected with base using the
same conditions that were used to protect the flexible analogues
14 and 15 (see next section).6 The NMR spectra of 10 and 12
in D2O were even broader than at the protected stage. In
the 13C-NMR spectra only signals of the galactose units
were visible, indicating this to be the most mobile part of the
molecule. Such a selective broadening has also been observed
in lactose polymers.9 However, ESI mass spectra showed the
correct masses for 10 and 12 and indicated complete deprotec-
tion. Again, the [M ϩ 2H]2ϩ (100%) and [M ϩ 3H]3ϩ (100%)
ions for 10 and 12 respectively were observed. From
our experiments with 1 we knew that the basic deprotection
conditions do not affect the heterocyclic ring.14
The extension of this chemistry to di- and tetravalent systems
(Scheme 1) started with diiodide 2. This compound was reacted
with Boc-protected propargylamine
3
under modified
Sonogashira reaction conditions,15 and yielded the fully pro-
tected branching unit 4. Part of this material was exposed to a
CH2Cl2–TFA 5 : 1 mixture to yield the N-deprotected com-
pound 5. Another part of 4 was exposed to basic conditions
in order to hydrolyse the methyl ester, which led to carboxylic
acid 6. The two differently protected building blocks 5 and 6
Assays
The inhibitory potency of the prepared lactose-containing
compounds was determined with the use of a solid phase assay.
In this assay the biotin-labeled lectins were allowed to bind to
the carbohydrate moieties of a glycoprotein coated onto the
surface of microtiter plate wells (the matrix) in competition
with our lactosides.10a,13 This design deliberately resembles
properties of a model cell surface. The inhibition assay was
performed with the galectins-1, -3, and -5. The comparative
analysis of the homologous members of the galectin family
required that their purity had to be ascertained rigorously.
Thus, we examined their biochemical characteristics, i.e.
molecular weight and isoelectric point, after purification by two
dimensional gel electrophoretic analysis. As shown in Fig. 2,
the three galectin preparations were free of contamination.
The molecular weights calculated from their electrophoretic
mobilities were close to the theoretical values based on the
amino acid sequence. Occurrence of shifts of about 0.5–1 kDa
from theoretical to experimental data, as seen for galectin-5,
are known from other galectins, for example the chicken
liver galectin CG-16 is referred to as a 16 kDa protein based
on electrophoretic mobility despite its molecular mass of
14,976 Da.17 Presence of isoelectric variants was only observed
for bovine galectin-1. In this case, bovine heart and spleen
Scheme 1 Reaction conditions: a) Pd(), CuI, NEt3, CH3CN, 2 d
(65%); b) TFA, CH2Cl2, 20 min (quant.); c) Tesser’s base, 14 h (quant.);
d) BOP, iPr2NEt, CH3CN, 14 h (71%).
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
804

View Article Online
Scheme 2 Reaction conditions: a) first iPr2NEt, CH2Cl2, then AcOH, CH2Cl2, 14 h (65%) for 9; iPr2NEt, CH3CN, 14 h (33%) for 11; b) NaOMe,
MeOH, 2 h (70%) for 10; dioxane, MeOH, 1  NaOH, 3 h (94%) for 12.
asialofetuin (ASF) which presents a heterogeneous glycan popu-
lation. It contains three N-glycosylation sites with complex
type triantennary chains either with an N-acetyllactosamine
terminus (74%) or with the Galβ1,3GlcNAc isomer in the outer
Manα1,3-arm (9%) and a biantennary chain (17%). These N-
glycan branches are known to be potent ligands for galectins in
cross-linking experiments.20 Three O-linked chains, present as
well, primarily with the Galβ1,3GalNAcα-disaccharide,20 also
contribute to ASF binding.21
In the assay the following inhibitory carbohydrates were
evaluated: galactose, lactose, ASF and compounds 1, 10 and 12.
IC50 values are listed in Table 1. From the table it can be seen
that for homodimeric galectin-1, a potent agent of haemag-
glutination, no significant multivalency effect was observed,
except for ASF. For the other proto type galectin, galectin-5,
however, a gradual increase in potency was detected while going
from 1 to 10 to 12. The tetravalent lactoside 12 showed a 21-
fold increase in potency per lactose when compared to the
most relevant monovalent reference: compound 1. Large multi-
valency effects, even better than for ASF, were observed in the
experiments with galectin-3, with enhancements, relative to
lactose, of almost 4300-fold for 12 and an IC50 of 70 nM. In
comparison to 1 the multivalency effect of 12 is a factor of
75 per monovalent unit while for 10 this number is 53.
Fig. 2 Two dimensional gel electrophoresis analysis after silver
staining of the three tested galectins showing purity in the cases of
bovine galectin-1 (black arrowhead), rat galectin-5 (open arrow), and
mouse galectin-3 (open arrowhead). Human galectin-7 (black arrow)
was added as a further control to underscore the resolving power of the
method. Occurrence of isoelectric variants was detected for bovine
galectin-1.
galectin-1 obtained by chromatography on ASF-Sepharose had
already shown pI heterogeneity in the detected range.18,19 The
matrix coated on the microtiter well was the glycoprotein
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
805

View Article Online
Table 1 Determination of the IC50 values and the relative inhibitory capacity (relative potency, rel. pot.)a of lactosides and a glycoprotein (ASF) in
a solid phase assay with surface immobilized ASF and three different labeled galectins in solution
Matrix probe:
1 µg ASF galectin-1
1 µg ASF galectin-3
1 µg ASF galectin-5
(10 µg mlϪ1
)
(30 µg mlϪ1
)
(15 µg mlϪ1
)
Lactose units
per molecule
Inhibitor
IC50/µM
Rel. pot.
IC50/µM
Rel. pot.
IC50/µM
Rel. pot.
Galactose
1
1
9
1
2
4
70000
4000
1.2
436
151
0.06
1
3200 (354)
9.2 (9.2)
26.5 (13.3)
22.5 (5.6)
400
300
0.08
21
0.2
0.07
0.75
1
3750 (417)
14 (14)
1500 (750)
4286 (1071)
7000
300
0.34
174.5
15.1
2.1
0.04
1
882 (98.0)
1.7 (1.7)
19.9 (9.9)
143 (35.7)
Lactose
ASF
1
10
12
178
a The numbers in parentheses express the relative potency of each lactose unit in the more than univalent inhibitor compared to carrier-free lactose.
Table 2 Effect of lactosides in the solid phase inhibition assay with galectin-3 (see Table 1) compared with the apparent Kd’s of the same lactosides
to galectin-3 as determined by a fluorescence binding assaya
Solid phase inhibition assay
matrix: 1 µg ASF
Fluorescence titration
Lactose units
per molecule
Carbohydrate
IC50/µM
rel. pot.
Kd/µM
rel. pot.
Lactose
1
1
2
4
1
2
4
300
n.d.
260
23
21
0.2
0.07
1
—
1.2 (0.6)
13 (3.3)
14 (14)
1500 (750)
4286 (1071)
618
162
79
31
50
30
14
1
4 (4)
8 (4)
20 (5)
12 (12)
21 (10)
44 (11)
13
14
15
1
10
12
a The numbers in parentheses express the relative potency of each lactose unit in the more than univalent inhibitor compared to carrier-free lactose.
In order to investigate in more detail whether or not these
multivalency phenomena were direct consequence of
potent on a per lactose basis: about 11-fold more potent than
lactose. In order to verify whether or not these results were
influenced by a different concentration of galectin-3 used in
both assays (0.86 µM (30 µg mLϪ1) in the solid phase inhibition
assay and 0.57 µM in the fluorescence titration) we repeated the
fluorescence titration of 12 with a galectin-3 concentration
of 1.5 µM. In the fluorescence titration this had no significant
effect as at both galectin-3 concentrations Kd’s were
comparable.
a
enhanced affinity for the galectin-3 monomers in solution, we
performed fluorescence titration studies. Galectin-3 contains
a single conserved tryptophan unit in the binding site whose
fluorescence spectrum is affected by ligand binding.22 Fig. 3
shows an example of a titration experiment. Besides the rigidi-
fied compounds we also tested their more flexible counterparts
13–15 (Fig. 4).6,13 In the solid phase assay 14 and 15 were not
particularly effective inhibitors. Table 2 shows all compounds
as evaluated in both solid phase inhibition tests (IC50) and
fluorescence titrations (Kd). The large differences in efficacy
between 14 and 15 versus 10 and 12 in the solid phase inhibition
studies is the most striking observation. In the fluorescence
assay both spacered monovalent lactosides 13 and 1 showed
improved binding relative to lactose, 4-fold and 12-fold respec-
tively. For 1 this was also the case in the solid phase assay,
where a 14-fold enhancement relative to lactose was seen. In the
fluorescence titrations no multivalency effects were seen. In the
flexible spacer series all three compounds 13–15 exhibited a
relative potency per lactose of around 4. In the rigid spacer
series all three compounds 1, 10 and 12 were roughly equally
Conclusion
In the present paper a pronounced effect of spacer rigidification
of a glycodendrimer on its selectivity for lectin binding was
demonstrated. Equally important, a striking difference between
multivalency effects observed in a common solid phase ELISA
type inhibition assay, a model for cell surface binding, and a
direct binding assay in solution using fluorescence spectroscopy
was observed. For the rigid tetravalent 12 an apparent Kd for
galectin-3 of 14 µM was determined. However, 12 led to a
50% inhibition of galectin-3 binding to matrix ASF at a con-
centration of 70 nM, a concentration at which very little of this
compound is associated to galectin-3 in solution according
to the fluorescence binding assay. In the literature reports
have appeared that emphasise that the type of assay can be
an important factor in the size of the multivalency effects
measured. Not all assays measure the same phenomena.3g,5b
Toone et al. and Brewer et al. have demonstrated this with the
binding of mannose dendrimers to the tetrameric plant lectin
ConA and noted sizeable effects they attributed to aggregation
phenomena.5c,23 Another illustration of the discrepancy
between assays in different set-ups was recently shown with a
divalent ligand binding to the Shiga-like toxin of pathogenic
E. coli. A mass spectrometric analysis gave only a 6-fold multi-
valency enhancement, while an enzyme-linked immunosorbent
assay showed a 40-fold enhancement.24 The explanation of the
enhanced inhibitory power of 12 for galectin-3 likely results
from the aggregation behaviour of the chimeric galectin-3 on a
Fig. 3 Fluorescence titration of murine galectin-3 by 12.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
806

View Article Online
Fig. 4 Structures of lactosides 13–14.
surface presented ligand. Corroborating previous work on
centration. Moreover, studies with a related lectin showed no
positive cooperativity of laminin binding by rat galectin-3,11b
Hughes et al.25 have concluded from careful analysis of SPR
studies with hamster galectin-3 that initially the carbohydrate
recognition domain (CRD) of the galectin binds to the carbo-
hydrate but that this is followed by a recruitment of additional
lectin molecules from free solution forming lectin aggregates on
the surface via protein–protein interactions. This recruitment
is primarily mediated by contacts between the N-terminal
collagen like domains of galectin-3. Fittingly, presence of this
domain is essential for positive cooperativity and potent bio-
medical effects.11b,12b,25 It could explain that despite the fact that
only a small portion of 12 was bound to the CRD (as indicated
by fluorescence) at 70 nM, a large portion of galectin-3 was
made unavailable for firmly binding to the ASF matrix by addi-
tional aggregation. Observations made by mass spectrometry
are consistent with this scenario. Free galectin-3 is present as a
monomer,26 while the addition of equimolar (10 µM) amounts
of a multivalent lactose derivative leads to a total dis-
appearance of the monomer signal, presumably due to the fact
that large aggregates are formed.27 The fact that the galectin-3
in the solid phase assay was biotinylated and non-biotinylated
in the fluorescence assay can not be fully excluded as the cause
of the discrepancy between the assays, however biotinylation
was performed in the presence of a saturating lactose con-
notable differences between biotinylated and non-biotinylated
forms.35 Additional studies will have to shed more light on
galectin-3 behaviour in the presence of multivalent carbo-
hydrates, which is relevant to its biological activities. Regarding
the capacity of cells to express more than one galectin,28 our
studies showed that selective blocking of one particular galectin
is possible with synthetic compounds exposing the same head
group. It is a salient finding of this study that this topological
parameter already engendered a marked selectivtiy. It can
further be increased by exploiting subtle inter-galectin and
inter-lectin family differences in ligand specificity.8,29 Such
selective compounds may prove to be valuable research tools in
further determining the biological roles of the galectins and
may also provide an entry into drug development, a promising
perspective especially for galectin-3 and its role in promoting
metastasis and angiogenesis.30
Experimental
General remarks
Chemicals were obtained from commercial sources and used
without any further purification unless stated otherwise. The
solvents CH2Cl2, MeOH and dioxane were purchased from
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
807

View Article Online
Biosolve, the Netherlands. The solvents CH2Cl2 and MeOH
were stored on molecular sieves (4 Å and 3 Å respectively). The
base iPr2NEt was distilled from ninhydrin and KOH. All the
solvents used for the Sonogashira reaction were dried with
molecular sieves and degassed using nitrogen and the reaction
was performed under an argon stream in the dark. Column
chromatography was performed on Merck Kieselgel 60 (40–63
µm). For neutralization Dowex 50 × 8 (Hϩ-form; 20–50 mesh)
disaccharides. The solid phase assays on a matrix established by
the glycoprotein, which resembles a cell surface, were performed
and quantitatively evaluated, as described previously.13,34 In
brief, 1 µg asialofetuin in 50 µl 20 mM phosphate-buffered
saline (pH 7.2) was coated for 12 h at 4 ЊC to the surface of
each well in an ELISA plate. Following blocking of remaining
protein-binding sites with buffer containing 1% bovine serum
albumin and thorough washing a solution containing the
galectin and a glycodendrimer which had been preincubated
for 30 min at room temperature was transferred to the wells.
Probing of the extent of specifically bound marker by strepta-
vidin peroxidase and chromogenic substrate followed after the
incubation with the galectin containing solution for 1 h at 37 ЊC
and washing steps to remove any traces of galectin completely.
The assay conditions had been selected on the basis of a
prior experimental series defining an optimal carbohydrate
dependent response in this system.
1
purchased from Fluka, was used. H-NMR (300 MHz) and
13C-NMR (75 MHz) spectra were recorded on a Varian G-300
spectrometer. Chemical shifts are reported in ppm relative to
CHCl3 (δ = 7.26 and δ = 77.0 respectively) or DMSO-d6 (δ1H =
2.49) or to CD3OD (δ 13C = 49). Coupling constants (J) are
given in Hz. Assignments were given and the numbering
schemes of the compounds can be seen in both Scheme 1 and
Scheme 2. In addition to that the carbons of the 2-aminothiazo-
line ring carry the index ‘h’ as a subscript to indicate this fact.
Electrospray ionization (ESI) mass spectrometry was carried
out with a Shimadzu LCMS QP-8000 single quadrupole bench-
top mass spectrometer (m/z range < 2000), coupled with a
QP-8000 data system. Elemental analyses were carried out at
Kolbe Mikroanalytisches Laboratorium (Mülheim an der Ruhr,
Germany). For the fluorescence measurements a Fluorolog-3
FL 3–21 spectrofluorometer was used, connected to a PC with
the program Datamax for WindowsTM 2.00.
3,5-Bis(3-tert-butoxycarbonylamino-prop-1-ynyl)benzoic acid
methyl ester (4)
To a suspension of 3,5-diiodobenzoic acid methyl ester 2
(7.25 g, 18.7 mmol), tetrakis(triphenylphosphine)palladium()
(2.16 g, 1.87 mmol) and copper() iodide (356 mg, 1.87 mmol)
in acetonitrile (80 mL) was added simultaneously a solution of
prop-2-ynylcarbamic acid tert-butyl ester 3 (9.26 g, 56.1 mmol,
94% pure) in acetonitrile (10 mL) and Et3N (13 mL, 93.5
mmol). The reaction mixture was stirred at room temperature.
After 14 h a homogeneous solution was obtained but according
to the TLC (EtOAc–hexanes 1 : 1) the reaction was not
completed. After stirring for an additional 24 h a precipitate
appeared and complete conversion of 2 was observed. The
precipitate was filtered off and washed with cold acetonitrile
and hexanes (Ϫ20 ЊC) to yield 5.4 g (65%) of a colourless solid.
(Found: C 65.04; H 6.75; N 6.26. Calc. for C24H30N2O6:
C 65.14; H 6.83; N 6.33); δH(300 MHz; CDCl3) 8.00 (2 H, d,
J 1.8 2/6-H), 7.60 (1 H, s, 4-H), 4.78 (2 H, br, NH), 4.15 (4 H, d,
J 5.1, CH2), 3.92 (3 H, s, CH3O), 1.47 (18 H, s, CH3Boc); δC(75.4
MHz; CDCl ) 165.5 (C᎐OI), 155.3 (C᎐OBoc), 138.3 (4-C), 132.2
Galectins and solid phase inhibition assay
The three galactoside specific lectins were purified from
aqueous extracts (as material sources, we used a bovine heart
obtained from a local slaughterhouse in the case of galectin-1
and bacteria within recombinant production in the cases of
galectins-3 and -5) by affinity chromatography on lactosylated
Sepharose 4B, derived from ligand coupling to divinyl sulfone
activated resin, as the crucial step.31 For recombinant expres-
sion of murine galectin-3 we used the plasmid prCBP35, kindly
provided by J. L. Wang (East Lansing, MI, USA), and E. coli
JA221 cells, as described previously.13 The cDNA of rat
galectin-5 was cloned from total RNA of rat kidney isolated
using an RNAeasy kit (Qiagen, Hilden, Germany) and was sub-
jected to reverse transcription using 200 U SuperscriptTMII
RnaseHϪ Reverse Transcriptase (Invitrogen/Life Technologies,
Karlsruhe, Germany). PCR amplification was directed by
the sense primer 5Ј-CATGCCATGGCTTCCTTCAGCAC-
CGAGA-3Ј with an internal NcoI restriction site (underlined)
and the antisense primer 5Ј-CGGCATGGTAGGTCTCCA-
CGTGTGTT-3Ј with an internal BamHI restriction site
(underlined) based on the published sequence of this galectin.32
The introduction of the described NcoI cleavage site led to a
Ser2–Ala2 (S2A) substitution in the protein. Digestion with
NcoI/BamHI of the amplicon was followed by its insertion into
the procaryotic expression vector pQE60, and recombinant
production was carried out in TB (terrific broth) medium
(Roth, Karlsruhe, Germany) at 30 ЊC following induction with
0.5 mM isopropyl-β--thiogalactoside using E. coli M15-
[pREP4] cells from Qiagen (Hilden, Germany). The isolated
lectins were checked for purity and molecular characteristics by
two dimensional gel electrophoresis using the IPGphorTM iso-
electric focusing system and the Hoefer SE600 standard vertical
unit (16 × 18 cm) with ImmobilineTM dry strips (pH 3–10,
linear) from Amersham Biosciences (Freiburg, Germany). The
silver staining kit for proteins from this supplier and computer
assisted analysis of the staining profile using a commercial set
of standards as internal controls for calibration enabled calcu-
lation of molecular weights and isoelectric points. Biotinylation
of the lectins was carried out using the N-hydroxysuccinimide
ester derivative of biotin under activity preserving con-
ditions.33,34 After being prepared as described previously,13
ASF was further subjected to β-elimination by treatment with
0.2 M NaOH for 4 h at 45 ЊC to remove the three O-glycan
᎐
᎐
3
(2/6-C), 130.4 (1-C), 123.4 (3/5-C), 87.2 (Cb), 81.0 (Ca), 80.0
(CBoc), 52.3 (CH3O), 30.9 (CH2), 28.2 (CH3Boc); m/z (ESI) 465.15
(M ϩ Na)ϩ.
TFA-salt of 3,5-bis(3-aminoprop-1-ynyl)benzoic acid methyl
ester (5)
To a solution of 4 (500 mg, 1.1 mmol) in CH2Cl2 (10 mL)
trifluoroacetic acid (TFA) (2 mL) was added and the reaction
mixture stirred at room temperature for 20 minutes. After
the solvent was removed in vacuo and the TFA co-evaporated
with CH2Cl2 the product was obtained as an oil in quanti-
tative yield. δH(300 MHz; D2O) 7.93 (2 H, d, J 1.1, 2/6-H),
7.65 (1 H, s, 4-H), 3.89 (4 H, s, CH2), 3.75 (3 H, s, CH3O);
δ (75.4 MHz; D O) 170.2 (C᎐OI), 141.7 (4-C), 135.9 (2/6-C),
᎐
C
2
133.3 (1-C), 125.0 (3/5-C), 87.3 (Cb), 84.4 (Ca), 55.8 (CH3O),
32.4 (CH2).
3,5-Bis(3-tert-butoxycarbonylaminoprop-1-ynyl)benzoic acid (6)
Compound 4 (2 g, 4.5 mmol) was dissolved in a mixture of
dioxane (119 mL) and methanol (42.5 mL). To that solution
4 M NaOH (8.5 mL) were added. After stirring at room
temperature for 14 h, the pH was adjusted to approximately 2
with 1 M KHSO4 and the solvents were evaporated in vacuo. To
the residue CH2Cl2 (100 mL) was added and the solution was
washed with brine (2 × 75 mL) and dried (Na2SO4). Acid 6 was
isolated as a white powder in quantitative yield. δH(300 MHz;
CDCl3) 8.05 (2 H, br, 2/6-H), 7.60 (1 H, s, 4-H), 4.85 (2 H, br,
NH), 4.15 (4 H, br, CH2), 1.48 (18 H, s, CH3Boc); m/z (ESI)
339.40 (M Ϫ 2CH᎐C(CH ) ϩ Na)ϩ 395.37 (M Ϫ CH᎐C(CH )
᎐
᎐
3
2
3 2
ϩ Na)ϩ 451.52 (M ϩ Na, 100%)ϩ.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
808

View Article Online
Protected tetraamine (7)
(all C᎐OOAc), 166.9, 166.6 (C᎐OI ϩ C᎐OII), 158.3 (2-C ), 140.3,
᎐
᎐
᎐
h
138.6, 137.0, 134.3, 132.3, 130.5, 126.1, 125.2, 123.1 (br, all i-C
ϩ 5-Ch), 117.1 (Cd), 100.9 (Gal–Canom), 86.8, 82.5–81.3 (Ca ϩ Cb
ϩ Gal–Canom), 75.8, 74.5–73.5, 71.0, 70.9, 70.2, 70.0, 68.9, 68.5,
68.0, 66.4 (all Lac–CH), 61.8, 60.6 (2 × Lac–CH2), 60.5–59.5
(4-Ch), 52.4 (CH3O), 30.4 (Cc), 20.9, 20.8, 20.7, 20.6, 20.5
iPr2NEt (118 µL, 675 µmol) was added to a suspension of 5
(42.3 mg, 90 µmol), 6 (75 mg, 175 µmol) and BOP (88 mg, 198
µmol) in dry acetonitrile (5 mL) to give a homogeneous solu-
tion. The reaction mixture was stirred at room temperature for
14 h and the obtained precipitate was filtered and washed with
ice cold acetonitrile yielding 7 (68 mg, 71%) as a white solid.
(Found: C 67.86; H 6.27; N 7.83; Calc. for C60H66N6O12: C
67.78; H 6.26; N 7.90); δH(300 MHz; DMSO-d6) 9.23 (2 H, br,
NHCO), 7.89 (6 H, s, 2/6-H, 2Ј/6Ј-H), 7.72 (s, 1H, 4-H), 7.53
(2 H, s, 4Ј-H), 7.40 (4 H, br, NHBoc), 4.32 (4 H, d, J 4.8, CH2),
3.99 (8 H, d, J 5.4, CH2Ј), 3.85 (3 H, s, CH3O), 1.38 (36 H, s,
CH3Boc). δC(75.4 MHz; 43% CD3OD–CD3COCD3,) 166.4
(C᎐OI), 166.0 (C᎐OII), 156.9 (C᎐OBoc), 138.9, 137.5, 135.7,
(all CH3OAc); m/z (ESI) 772.85 (M Ϫ Gal ϩ 2H ϩ 2Na)4ϩ
,
844.00 (M ϩ 4H)4ϩ, 919.55 (M Ϫ 2Gal ϩ H ϩ 2Na)3ϩ, 1014.95
(M Ϫ Gal Ϫ CH2CO ϩ H ϩ 2Na)3ϩ, 1029.50 55 (M Ϫ Gal ϩ
H ϩ 2Na)3ϩ, 1125.40 (M ϩ 3H, 100%)3ϩ, 1110.95 (M Ϫ
CH2CO ϩ 3H)3ϩ, 1542.65 (M ϩ 2Na)2ϩ, 1687.10 (M ϩ 2H)2ϩ
.
Tetravalent lactose (12)
᎐
᎐
᎐
Aqueous NaOH (1 M, 0.095 mL) was added to a solution of
11 (55 mg, 16.3 µmol) in dioxane (1.33 mL) and MeOH
(0.475 mL). The resulting suspension was agitated at room
temperature for 3 h. A homogeneous solution was obtained
after neutralization with Dowex resin. After removal of the
resin by filtration the solvents were evaporated in vacuo. The
material was dissolved and lyophilized from H2O to yield 12 as
a white solid (33.7 mg, 94%). m/z (ESI) 625.35 (M Ϫ 2Gal ϩ
132.7, 132.0, 130.9 (1-C ϩ 2/6-C ϩ 4-C ϩ 1Ј-C ϩ 2Ј/6Ј-C ϩ
4Ј-C), 124.9 (3/5-C ϩ 3Ј/5Ј-C), 89.2, 88.5 (Cb ϩ CbЈ), 81.0, 80.9
(Ca ϩ CaЈ), 79.8 (CBoc), 52.8 (CH3O), 31.0 (Cc), 28.5 (CH3Boc).
Tetra-TFA-salt 8
Boc deprotection of compound 7 (60 mg, 56 µmol) was per-
formed using the procedure described for 5 to afford 8 as
a colorless oil in quantitative yield. Compound 8 was used
without any further purification and characterization for the
synthesis of 11.
3H)3ϩ, 678.80 (M Ϫ Gal ϩ 3H)3ϩ, 732.8 (M ϩ 3H, 100%)3ϩ
,
1017.90 (M Ϫ Gal ϩ 2H)2ϩ, 1099.00 (M ϩ 2H)2ϩ
.
Fluorescence titrations
Protected divalent lactose (9)
A solution (1000 µL) of galectin-3 (0.57 µM) in PBS buffer
(Na2HPO4 8.1 mM, KH2PO4 1.5 mM, pH 7.2, 0.15 M NaCl)
was placed in a cuvette (1 × 1 cm). The fluorescence spectrum
was recorded between 300 nm and 500 nm, using an irradiation
wavelength of 282 nm, and the temperature was maintained at
25 ЊC. Incremental additions (> 10 datapoints) of a solution
containing the lactose (derivative) and galectin-3 (0.57 µM)
were added until at least 75% saturation was achieved. The
fluorescence spectra were recorded by scans of 1 nm secϪ1 and
the data were stored. An additional spectrum of a reference
cuvette with the same amount of the ligand but no galectin-3
was taken for each datapoint. These spectra were subtracted
from those in which galectin-3 was present to correct for some
inherent ligand fluorescence. The samples were left to incubate
for about 5 min after each addition before the measurements
were started. Fluorescence data at 350 nm, 340 nm and 335 nm
for lactose derivatives 1, 10 and 12 respectively were used to
determine the apparent dissociation constant by fitting them
to a one-site binding model.
A suspension of 5 (219 mg, 466 µmol) and Lac(OAc)7NCS
(920 mg, 1.4 mmol) in dry CH2Cl2 (5 mL) was treated with
iPr2NEt (245 µL, 1.4 mmol) and stirred at room temperature
overnight. The reaction mixture was washed with 1 M KHSO4
(twice), 0.5 M NaOH (twice) and brine, dried (Na2SO4) and
concentrated to dryness. The residue was redissolved in 6 mL
CH2Cl2 and was treated with acetic acid (7 mL) at room tem-
perature for 14 h. The crude mixture was purified by column
chromatography (4% MeOH–CH2Cl2) and afforded product 9
in 65% (446 mg) yield. δC(75.4 MHz; 4% AcOH–CDCl3) 170.8,
170.6, 170.4, 170.2, 170.1, 169.9, 169.3 (all C᎐OOAc), 167.6
᎐
(2-C ), 166.1 (C᎐O), 133.2 (5-C ), 136.3, 131.1, 126.8 (1-C ϩ
᎐
h
h
2/6-C ϩ 3/5-C ϩ 4-C), 120.2 (Cd), 100.6 (Gal–Canom), 83.9 (Glc–
Canom), 75.5, 74.1, 72.7, 70.9, 70.8, 69.0, 66.7 (all Lac–CH), 61.9,
60.7 (2 × Lac–CH2), 59.8 (4-Ch), 52.4 (CH3O), 20.3, 20.2, 20.1,
20.0, 19.9 (all CH3OAc); m/z (ESI) 655.75 (M Ϫ Gal ϩ 2Na)2ϩ
,
778.60 (M Ϫ CH2CO ϩ 2H)2ϩ, 799.95 (M ϩ 2H, 100%)2ϩ
.
Divalent lactose (10)
Acknowledgements
NaOMe (30 wt% solution in MeOH, 46 µL, 238 µmol) was
added to a solution of 9 (27 mg, 17 µmol) in dry MeOH
(1.5 mL). The resulting suspension was agitated at room tem-
perature for 2 h. The reaction mixture was neutralized with
Dowex resin. After addition of water (1 mL) and removal of the
resin by filtration, the solvents were evaporated in vacuo. The
material was dissolved and lyophilized from H2O to yield 12 mg
(70%) of 10 as a yellowish solid. m/z (ESI) 505.15 (M ϩ 2H,
100%)2ϩ, 1010.35 (M ϩ H)ϩ.
The excellent technical assistance by B. Hofer and L. Mantel
and the generous financial support of the Royal Netherlands
Academy of Arts and Sciences (KNAW), the EC program for
high level conferences (BIOPATH) and the Wilhelm Sander-
Stiftung (Munich, Germany) are gratefully acknowledged.
References
1 (a) S. J. Williams and G. J. Davies, Trends Biotechnol., 2001, 19,
356–362; (b) H.-J. Gabius, S. André, H. Kaltner and H.-C. Siebert,
Biochim. Biophys. Acta, 2002, 1572, 165–177.
2 (a) H. Kaltner and B. Stierstorfer, Acta Anat., 1998, 161, 162–179;
(b) E. E. Simanek, G. J. McGarvey, J. A. Jablonowski and
C.-H. Wong, Chem. Rev., 1998, 98, 833–862; (c) J. C. McAuliffe
and O. Hindsgaul, Frontiers Mol. Biol., 2000, 30, 249–285;
(d ) H. Rüdiger, H.-C. Siebert, D. Solís, J. Jiménez-Barbero,
A. Romero, C.-W. von der Lieth, T. Diaz-Mauriño and H.-J. Gabius,
Curr. Med. Chem., 2000, 7, 389–416.
3 (a) Y. C. Lee, FASEB J., 1992, 6, 3193–3200; (b) N. V. Bovin and
H.-J. Gabius, Chem. Soc. Rev., 1995, 24, 413–421; (c) R. Roy, Curr.
Opin. Struct. Biol., 1996, 6, 692–702; (d ) M. Mammen, S.-K. Chio
and G. M. Whitesides, Angew. Chem., Int. Ed., 1998, 37, 2754–2794;
(e) L. L. Kiessling, L. E. Strong and J. E. Gestwicki, Ann. Rep. Med.
Protected tetravalent lactose (11)
To a suspension of 8 (63 mg, 56 µmol) in acetonitrile (5 mL),
LacNCS (152 mg, 224 µmol) was added and the reaction
mixture treated at room temperature with at least 5 equiv. iPr2-
NEt until a basic (pH 8) homogeneous solution was obtained.
After stirring at room temperature for 14 h the mixture was
concentrated in vacuo. The residue was dissolved in CH2Cl2 and
washed with 1 M KHSO4 (twice), 0.5 M NaOH (twice) and
brine, dried (Na2SO4) and concentrated to dryness. Compound
11 was isolated in 33% yield (62 mg) after purification by
column chromatography (3% MeOH–7% acetone–CH2Cl2).
δC(75.4 MHz; CDCl3) 170.6, 170.3, 170.2, 170.1, 169.7, 169.2
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
809

View Article Online
Chem., 2000, 35, 321–330; ( f ) H.-J. Gabius, Anat. Histol. Embryol.,
2001, 30, 3–31; (g) T. K. Dam and C. F. Brewer, Chem. Rev., 2002,
102, 387–430; (h) J. J. Lundquist and E. J. Toone, Chem. Rev., 2002,
102, 555–578; (i) T. K. Lindhorst, Top. Curr. Chem., 2002, 218,
201–235.
4 P. I. Kitov, J. M. Sadowska, G. Mulvey, G. D. Arnstrong, H. Ling,
N. S. Pannu, R. J. Read and D. R. Bundle, Nature, 2000, 403,
669–672.
5 (a) S. M. Dimick, S. C. Powell, S. A. McMahon, D. N. Moothoo,
J. H. Naismith and E. J. Toone, J. Am. Chem. Soc., 1999, 121,
10286–10296; (b) T. K. Dam, R. Roy, S. K. Das, S. Oscarson and
C. F. Brewer, J. Biol. Chem., 2000, 275, 14223–14230; (c) T. K. Dam,
R. Roy, D. Pagé and C. F. Brewer, Biochemistry, 2002, 41,
1351–1358.
19 H. Ahmed, N. E. Fink, J. Pohl and G. R. Vasta, J. Biochem., 1996,
120, 1007–1019.
20 D. Gupta, H. Kaltner, X. Dong, H.-J. Gabius and C. F. Brewer,
Glycobiology, 1996, 6, 843–849.
21 In order to decipher whether the three O-glycan disaccharides could
also serve as ligands under these experimental conditions, we
removed them chemically by β-elimination and performed solid
phase assays to determine the contribution of the O-glycans to the
overall binding. In the solid phase assays, binding to ASF was
reduced for the lectins. In detail, by about 30% in the cases of
galectins-1, -3 and -5.
22 (a) S. Jia and J. L. Wang, J. Biol. Chem., 1988, 263, 6009–6011;
(b) J. Seetharaman, A. Kanigsberg, R. Slaaby, H. Leffler, S. H.
Barondes and J. M. Rini, J. Biol. Chem., 1998, 273, 13047–13052.
23 J. B. Corbell, J. L. Lundquist and E. J. Toone, Tetrahedron:
Asymmetry, 2000, 11, 95–111.
6 I. Vrasidas, N. J. de Mol, R. M. J. Liskamp and R. J. Pieters, Eur. J.
Org. Chem., 2001, 4685–4692.
7 (a) H.-J. Gabius, Eur. J. Biochem., 1997, 243, 543–576; (b) H. Lis and
N. Sharon, Chem. Rev., 1998, 98, 637–674; (c) R. Loris, Biochim.
Biophys. Acta, 2002, 1572, 198–208; (d ) G. A. Rabinovich,
N. Rubinstein and M. A. Toscano, Biochim. Biophys. Acta, 2002,
1572, 274–284.
8 P. Sörme, Y. Qian, P.-G. Nyholm, H. Leffler and U. J. Nilsson,
ChemBioChem, 2002, 3, 183–189 and references cited therein.
9 See for a study with galectin-3 and lactose polymers: N. L. Pohl and
L. L. Kiessling, Synthesis, 1999, 1515–1519.
24 E. N. Kitova, P. I. Kitov, D. R. Bundle and J. S. Klassen,
Glycobiology, 2001, 11, 605–611.
25 E. A. M. Barboni, S. Bawumia and R. C. Hughes, Glycoconjugate
J., 1999, 16, 365–373.
26 J. Kopitz, S. André, C. von Reitzenstein, K. Versluis, H. Kaltner,
R. J. Pieters, K. Wasano, I. Kuwabara, F.-T. Liu, M. Cantz,
A. J. R. Heck and H.-J. Gabius, submitted.
27 K. Versluis, A. J. R. Heck, unpublished observation.
28 (a) H. Lahm, S. André, A. Hoeflich, J. R. Fischer, B. Sordat,
H. Kaltner, E. Wolf and H.-J. Gabius, J. Cancer Res. Clin. Oncol.,
2001, 127, 375–386; (b) A. Danguy, I. Camby and R. Kiss, Biochim.
Biophys. Acta, 2002, 1572, 285–293; (c) U. Wollina, T. Graefe,
S. Feldpappe, S. André, K. Wasano, H. Kaltner, Y. Zick and
H.-J. Gabius, J. Cancer Res. Clin. Oncol., 2002, 128, 103–110.
29 (a) D. Solís, A. Romero, H. Kaltner, H.-J. Gabius and T. Diaz-
Mauriño, J. Biol. Chem., 1996, 271, 12744–12748; (b) S. André,
B. Frisch, H. Kaltner, D. L. Desouza, F. Schuber and H.-J. Gabius,
Pharm. Res., 2000, 17, 985–990; (c) N. Ahmad, H.-J. Gabius,
H. Kaltner, S. André, I. Kuwabara, F.-T. Liu, S. Oscarson,
T. Norberg and C. F. Brewer, Can. J. Chem., 2002, 80, 1096–1104;
(d ) A. M. Wu, J. H. Wu, M.-S. Tsai, J.-H. Liu, S. André, K. Wasano,
H. Kaltner and H.-J. Gabius, Biochem. J., 2002, 367, 653–664.
30 P. Nangia-Makker, J. Conklin, V. Hogan and A. Raz, Trends Mol.
Med., 2002, 8, 187–192.
10 (a) S. André, P. J. C. Ortega, M. A. Perez, R. Roy and H.-J. Gabius,
Glycobiology, 1999, 9, 1253–1261; (b) C. F. Brewer, Biochim. Biophys.
Acta, 2002, 1572, 255–262.
11 (a) L. G. Frigeri, M. W. Robertson and F.-T. Liu, J. Biol. Chem.,
1990, 265, 20763–20769; (b) S. M. Massa, D. N. W. Cooper,
H. Leffler and S. H. Barondes, Biochemistry, 1993, 32, 260–267.
12 (a) R. S. Bresalier, N. Mazurek, L. R. Sternberg, J. C. Byrd,
C. K. Yunker, P. Nangia-Makker and A. Raz, Gastroenterology,
1998, 115, 287–296; (b) J. Kopitz, C. von Reitzenstein, S. André,
H. Kaltner, J. Uhl, V. Ehemann, M. Cantz and H.-J. Gabius, J. Biol.
Chem., 2001, 276, 35917–35923; (c) I. Camby, N. Belot, F. Lefranc,
N. Sadeghi, Y. de Launoit, H. Kaltner, S. Musette, F. Darro,
A. Danguy, I. Salmon, H.-J. Gabius and R. Kiss, J. Neuropathol.
Exp. Neurol., 2002, 61, 585–596.
13 S. André, R. J. Pieters, I. Vrasidas, H. Kaltner, I. Kuwabara,
F.-T. Liu, R. M. J. Liskamp and H.-J. Gabius, ChemBioChem, 2001,
2, 822–830.
14 I. Vrasidas, J. Kemmink, R. M. J. Liskamp and R. J. Pieters, Org.
Lett., 2002, 4, 1807–1808.
31 H.-J. Gabius, Anal. Biochem., 1990, 189, 91–94.
32 M. A. Gitt, M. F. Wiser, H. Leffler, J. Herrmann, Y.-R. Xia,
S. M. Massa, D. N. W. Cooper, A. J. Lusis and S. H. Barondes,
J. Biol. Chem., 1995, 270, 5032–5038.
15 K. Sonogashira, Y. Tohda and N. Hagihara, Tetrahedron Lett., 1975,
16, 4467–4470.
33 H.-J. Gabius, B. Wosgien, U. Brinck and A. Schauer, Pathol. Res.
Pract., 1991, 187, 839–847.
34 (a) S. André, C. Unverzagt, S. Kojima, X. Dong, C. Fink, K. Kayser
and H.-J. Gabius, Bioconjugate Chem., 1997, 8, 845–855;
(b) C. Unverzagt, S. André, J. Seifert, S. Kojima, C. Fink,
G. Srikrishna, H. H. Freeze, K. Kayser and H.-J. Gabius, J. Med.
Chem., 2002, 45, 478–491.
16 BOP
=
benzotriazol-1-yloxytris(dimethylamino)phosphonium
hexafluorophosphate.
17 M. Schneller, S. André, J. Cihak, H. Kaltner, H. Merkle,
G. J. Rademaker, J. Haverkamp, J. Thomas-Oates, U. Lösch and
H.-J. Gabius, Cell. Immunol., 1995, 166, 35–43.
18 A. De Waard, S. Hickman and S. Kornfeld, J. Biol. Chem., 1976,
251, 7581–7587.
35 O. E. Galanina, H. Kaltner, L. S. Khraltsova, N. V. Bovin and
H.-J. Gabius, J. Mol. Recognit., 1997, 10, 139–147.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 8 0 3 – 8 1 0
810