Immobilisation on polystyrene of diazirine derivatives of mono- and disaccharides: Biological activities of modified surfaces

Article abstract of DOI:10.1016/S0968-0896(01)00172-9

The potential of surface glycoengineering for biomaterials and biosensors originates from the importance of carbohydrate protein interactions in biological systems. The strategy employed here utilises carbene generated by illumination of diazirine to achieve covalent bonding of carbohydrates. Here, we describe the synthesis of an aryl diazirine containing a disaccharide (lac-tose). Surface analysis techniques [X-ray photoelectron spectroscopy (XPS) and time of flight secondary ion mass spectroscopy (ToF-SIMS)] demonstrate its successful surface immobilisation on polystyrene (PS). Results are compared to those previously obtained with an aryl diazirine containing a monosaccharide (galactose). The biological activity of galactose- or lactose-modified PS samples is studied using rat hepatocytes, Allo A lectin and solid-phase semi-synthesis with α-2,6-sialyltransferase. Allo A shows some binding to galactose-modified PS but none to lactose-modified surfaces. Similar results are obtained with rat hepatocytes. In contrast, sialylation of lactose-modified PS is achieved but not with galactose-modified surfaces. The different responses indicate that the biological activity depends not only on the carbohydrate per se but also on the structure and length of the spacer. Copyright

Full text of DOI:10.1016/S0968-0896(01)00172-9

Bioorganic & Medicinal Chemistry 9 (2001) 2943–2953
Immobilisation on Polystyrene of Diazirine Derivatives of Mono-
and Disaccharides: Biological Activities of Modified Surfaces
Y. Chevolot,a,* J. Martins,b N. Milosevic,c D. Leonard,a,y S. Zeng,d M. Malissard,d
E.G. Berger,d P. Maier,c H.J. Mathieu,a D.H.G. Croutb and H. Sigriste
a
´ ´ ´ ´
Departement des Materiaux, LMCH, Ecole Polytechnique Federale de Lausanne (EPFL), CH-1015 Lausanne-EPFL, Switzerland
bDepartment of Chemistry, University of Warwick, CV4 5AL Coventry, UK
cInstituteofToxicology,SwissFederalInstituteofTechnologyofZurich(ETH),Schorenstrasse16,CH-8603Schwerzenbach,Switzerland
¨
dInstitute of Physiology, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland
¨
¨
eCSEM (Centre Suisse d’Electronique et de Microtechnique SA), Jaquet-Droz 1, CH-2007 Neuchaˆtel, Switzerland
Received 2 March 2001; accepted 16 May 2001
Abstract—The potential of surface glycoengineering for biomaterials and biosensors originates from the importance of carbohy-
drate–protein interactions in biological systems. The strategy employed here utilises carbene generated by illumination of diazirine
to achieve covalent bonding of carbohydrates. Here, we describe the synthesis of an aryl diazirine containing a disaccharide (lac-
tose). Surface analysis techniques [X-ray photoelectron spectroscopy (XPS) and time of flight secondary ion mass spectroscopy
(ToF-SIMS)] demonstrate its successful surface immobilisation on polystyrene (PS). Results are compared to those previously
obtained with an aryl diazirine containing a monosaccharide (galactose). The biological activity of galactose- or lactose-modified
PS samples is studied using rat hepatocytes, Allo A lectin and solid-phase semi-synthesis with a-2,6-sialyltransferase. Allo A shows
some binding to galactose-modified PS but none to lactose-modified surfaces. Similar results are obtained with rat hepatocytes. In
contrast, sialylation of lactose-modified PS is achieved but not with galactose-modified surfaces. The different responses indicate
that the biological activity depends not only on the carbohydrate per se but also on the structure and length of the spacer. # 2001
Elsevier Science Ltd. All rights reserved.
Introduction
with biomolecules and living cells.1,2 Plasma membranes
display arrays of receptors that interact with specific
ligands. Among these interactions, carbohydrate–pro-
tein interactions play a key role in a number of reg-
ulatory processes in cells.3À6 Thus, oriented surface
immobilisation of carbohydrates (‘surface glyco-engin-
eering’) allows one to mimic the specific interactions
that occur naturally at cell surfaces and can create a
matrix on which specific cellular functions of cultured
cells might be investigated.
Recently developed biomaterials aim to mimic biological
surfaces to obtain defined and controlled interactions
Abbreviations: MAD, 4-(2,5-dioxo-2,5-dihydropyrrol-1-yl)-N-[3-(3-tri-
fluoro-diazirin-3-yl)phenyl]butyramide; MAD-Gal, 4-(3-d-galactopyr-
anosylsulfanyl-2,5-dioxopyrrolidiny-1-yl)-N-(3-(3-trifluoromethyl-3H-
diazirin-3-yl)phenyl]butyramide; ToF-SIMS, time-of-flight secondary
ion mass spectrometry; XPS, X-ray photoelectron spectroscopy;
mamu, milli atomic mass unit, MS; UHV, ultra high vaccum; PS,
polystyrene; Allo A, allomyrina dichotoma; RCA, ricinus communis
agglutinin; PBS, phosphate buffer saline; BSA, bovine serum albumin;
CMP-Neu 5 Ac, cytidine 50-monophospho-N-acetylneuramic acid;
NR, neutral red; MTT, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphe-
nyltetrazolium bromide; CMF, crude membrane fraction; EROD,
ethoxyresorufin de-ethylase; ASGP-R, asialoglycoprotein receptor;
PVLA, poly (vinbenzyl-b-d-lactonamide); DCC, N,N0-dicyclohexyl-
carbodiimide
Carbohydrate immobilisation has been achieved by
various techniques7,8 (copolymerisation, coupling with
divinyl sulphone, coupling to CNBr-activated sub-
strates, reductive amination, bisorane method, etc.).
Photo-immobilisation provides a versatile tool with
respect to the substrate (organic and inorganic) and
allows one to easily create microdomains of biorecogni-
tion with addressable printing, using mask-assisted
lithography techniques. The photoreagents most often
used for photo-immobilisation of biomolecules are
*Corresponding author. Tel.: +41-21-693-4124; fax: +41-21-693-
3946; e-mail: ian.chevolot@epfl.ch
yCurrent address: GE Plastics, Analytical Technology, Plasticslaan 1,
PO Box 117, 4600 AC Bergen op Zoom, The Netherlands.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00172-9

2944
Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
arylazides, trifluoromethyl-aryl diazirines and benzo-
phenones.9 These reagents generate very reactive inter-
mediates upon light activation. Their interaction with
the support material leads to the formation of covalent
bonds.9 Thus, the synthesis of molecules containing
carbohydrate and photoactivatable domains makes it
possible to achieve covalent binding of biologically
active carbohydrates.
lysosomes.14À17 Moreover, it was considered of interest
to develop a photoactivatable precursor for the semi-
synthesis of more complex carbohydrates. The transfer
of sialic acid by a-2,6-sialyltransferase18 to galactose
residues to a solid support was studied to develop our
approach to solid phase synthesis.
The present paper describes the synthesis of a 6-(b-d-
galactopyranosyl-(1!4)-b-d-glucopyranosylsulfanyl)hexa
noic acid [3-(3-trifluoromethyl-3H-diazirin-3-yl)phenyl]-
amide 8 (Scheme 2). Its immobilisation on polystyrene
(the material most often used in biological assays) is
demonstrated with surface analysis spectroscopic meth-
ods XPS and ToF-SIMS. The biological activity of the
polystyrene-modified surfaces is probed with the lectin
Allo A, primary rat hepatocytes, and a-2,6-sialyl-
transferase and compared with MAD-Gal modified
polystyrene.10
The synthesis and characterisation of an aryl diazirine
containing a galactose photoreagent [4-(3-d-galactopyr-
anosylsulfanyl-2,5-dioxopyrrolidin-1-yl)-N-(3-(3-trifluoro-
methyl-3H-diazirin-3-yl)phenyl]butyramide (MAD-Gal)]
1 was previously reported (Scheme 1).10 Using a mask-
ing technique, a specific pattern of immobilised carbo-
hydrate was laid down on diamond and confirmed with
surface analysis techniques [X-ray photoelectron spec-
troscopy (XPS) and time-of-flight secondary ions mass
spectrometry, (ToF-SIMS)]. However, the biological
activity of the modified surface remained to be probed.
Accordingly, the objective of the present study is to
investigate the biological activity of the photo-immobi-
lised galactose and to compare it with the properties of
surface-photo-immobilised lactose prepared in an ana-
logous manner. These engineered surfaces are expected
to be of interest for the design of biosensor devices. The
biological functionality of the surfaces is tested with a
defined substance (Allo A lectin) and with intact cells
(rat hepatocytes), both of which were expected to bind
to galactose residues. Allo A lectin was reported to have
a specific affinity for galactose residues especially as part
of lactose (3.1 mM inhibits haemagglutinin activity) and
O-nitrophenyl b-d-galactopyranoside (inhibitory at
12.5 mM).11 Hepatocytes express on their surface the
asialoglycoprotein receptor which is responsible for the
clearance of abnormal galactose-terminated serum
glycoproteins.12,13 Most serum glycoproteins carry
terminal sialic acid residues and a penultimate galactose
residue. When desialylated, the exposed galactose resi-
dues of the glycoprotein can interact with the
asialoglycoprotein receptor, initiating removal of the
glycoprotein from the circulation by endocytosis. Sub-
sequently, the incorporated protein is hydrolysed in
Results
Synthesis of 6-(ꢀ-D-galactosyl-(1!4)-ꢀ-D-glucopyrano
sylsulfanyl)hexanoic acid
diazirin-3yl)-phenyl]amide 8
[3-(3-trifluoromethyl-3H-
Lactose octa-acetate 5 was first coupled to 6-mercapto-
hexanoic acid 2 to give 6-(2,3,4,6-tetra-O-acetyl)-(1!4)-
b-d-2,3,6-triacetyl-glucopyranosylsulfanyl)hexanoic acid
6 using BF3–Et2O. It was then linked to 3-(3-tri-
fluoromethyl-3H-diazirin-3-yl)-phenylamine 3 via amide
formation to give the protected lactose derivative 7.
Deprotection of derivative 7 gave compound 8 with
30% yield.
The overall yield in the synthesis was 4.3%. In Figure 1
is shown the UV absorption spectrum of the product
exhibiting characteristic diazirine adsorption band at
357 nm.19 Figure 1 further illustrates conversion of the
diazirine into carbene, as the absorbance decreased with
increasing light exposure at 357 nm, demonstrating
exposure time-dependent photoactivation.
Scheme 1. Structure of MAD-Gal and synthesis scheme of the crosslinker 4 by the condensation of 6-mercaptohexanoic and with 3-(3-tri-
fluoromethyl-3H-diazirin-3-yl)phenylamine 3.

Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
2945
.
Scheme 2. Reagents: (i) BF3 Et2O; (ii) ClCO2Et/Et3N; (iii) NaOMe/MeOH. Lactose aryl diazirine synthesis is illustrated. Lactose octaacetate 5 was
first coupled to 6-mercaptohexanoic acid 2 to give 6-(2,3,4,6-tetra-O-acetyl)-(1-4)-d-2,3,6-triacetyl-glucopyranosylsulfanyl)hexanoic acid 6 using
BF3–Et2O. It was then linked to 3-(3-trifluoromethyl-3H-diazirin-3-yl)-phenylamine 3 via amide formation to give the protected lactose derivative 7.
ised intensity (Table 1) was eight times higher on sample
C than on sample B. The residual level of molecules
observed on sample B by ToF-SIMS is below the XPS
instrument detection limit. Other characteristic ions
such as CFÀ3 confirm the immobilisation of MAD-Gal
similarly to what has already been observed in previous
studies on diamond.22,23 For lactose aryl diazirine, fluor-
ine atomic percentage and FÀ normalised intensity (Table
2) illustrated in the same way the successful immobilisa-
tion of the molecule at the surface of polystyrene.
Figure 2 shows that ToF-SIMS FÀ normalised intensity
and XPS fluorine atomic percentage increased as a
function of the concentration of MAD-Gal solution
used for surface immobilisation on polystyrene. The
intensity of FÀ and F atomic% are related to the surface
density of photobonded molecules, and thus MAD-Gal
density increased with the concentration of applied
Figure 1. The absorbance of the diazirine functions at 357 nm is dis-
played as a function of irradation time. It illustrates conversion of the
MAD-Gal solution. This observation confirms previous
results obtained with photobonded 4-(2,5-dioxo-2,5-
dihydropyrrol-1-yl)-N-[3-(3-trifluoromethyl-3H-diazirin-3-
yl)phenyl]butyramide (MAD) on silicon nitride.21
diazirines into carbenes as the absorbance diminished with increasing
light exposure at 357 nm, demonstrating exposure time-dependent
photoactivation.
Surface analysis
Binding of Allo A lectin
Light activation at 357 nm of the diazirine leads to a
reactive carbene and a covalent bond is generated with
the surface provided that close contact between the sur-
face and the carbene is obtained. Reaction of aryl dia-
zirine containing molecules with various substrates has
previously been described.19À23 In the case of a diamond
substrate, fluorine indicates the presence of MAD-
Gal.23 It is thus expected that immobilisation of MAD-
Gal or lactose aryl diazirine on polystyrene should
result in an increase of the surface fluorine content.
Galactose residue availability at the surface of poly-
styrene was probed with the b-galactose specific lectin
Allo A. Biotinylated lectin was used. The amount of
surface lectin was measured with 35S streptavidin.
Retained radio-activity relates to the amount of radio-
activity in the radiolabelled streptavidin deposited. This
represents an arbitrary standard but is used to indicate
the relative levels of radioactivity remaining after the
various treatments of the modified surface. The results
are illustrated in Figures 3 and 4 for MAD-Gal and
lactose modified polystyrene surfaces, respectively. For
the two sets of data, the percentages of radioactivity for
samples A are similar (1.2 and 2.5%). For samples B,
the signal remains very low, even though it is slightly
higher in the case of MAD-Gal.
In Table 1, XPS atomic percentages of samples at dif-
ferent steps of MAD-Gal surface modification on poly-
styrene are displayed. The significant detection of
oxygen on sample A can be explained by the g-ray ster-
ilisation of the sample.24 Fluorine was only detected on
sample C with XPS indicating that molecules are at the
surface after photoactivation and the final washing. The
result is confirmed with ToF-SIMS where FÀ normal-
In the case of MAD-Gal (Fig. 3), radioactivity mea-
surements of sample C demonstrated that Allo A lectin

2946
Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
had a higher binding for the surface grafted with galac-
tose residues than polystyrene (A) and that this affinity
is a function of MAD-Gal surface density. Further-
more, inhibition with asialofetuin of lectin binding on
immobilised galactose was observed (Fig. 3I). It con-
firms that this interaction was specific and not owing to
a variation in physisorption (related to a change in sur-
face energy).
In the case of lactose modified polystyrene surfaces (Fig.
4), the radioactivity only reaches the value of approxi-
mately 3%, which is comparable to sample B in the case
of MAD-Gal experiments but almost 7 times lower than
for the corresponding sample C. Surface analysis
demonstrated that the surface density of the two mol-
ecules were almost similar. Thus, no specific affinity was
demonstrated for lactose modified polystyrene surfaces.
Table 1. XPS atomic percentages and ToF-SIMS relative intensities
for the different steps of MAD-Gal immobilisation on PS
At percentages (%)
Sample A
Sample B
Sample C
N
F
S
C
O
bdl
bdl
bdl
bdl
0.28Æ0.26
0.45Æ0.16
bdl
bdl
93.7Æ0.6
bdl
94.1Æ1.0
95.5Æ1.2
6.1Æ0.6
5.9Æ1.0
3.8Æ0.8
ToF-SIMS
Binding of primary cultured rat hepatocytes
Corrected total
intensity (counts)Â104
FÀ normalised
intensity %
110.4Æ1.8
0.9Æ0.13
106.4Æ12.9 132.2Æ15.7
On the cellular level, accessibility of the galactose and
lactose residues was tested with primary cultured rat
hepatocytes. Binding to surface galactose residues did
26Æ4
205Æ53
1.2Æ0.5
CFÀ3 normalised
intensity %
0.008Æ0.003
0.2Æ0.08
XPS and ToF-SIMS analysis of MAD-Gal grafted PS. Atomic per-
centages of the elements were observed with XPS. bdl means below
detection limit. Three areas were measured per sample. Sample A cor-
responds to the plain polystyrene, sample B corresponds to the surface
on which MAD-Gal was deposited and washed without light activa-
tion and sample C is the surface on which MAD-Gal was deposited,
light activated and washed. MAD-Gal is expected to be observed only
on sample C. Fluorine and nitrogen are signatures of the molecule and
are only observed on sample C. ToF-SIMS normalised intensity of
selected ions characteristic of MAD-Gal molecule (FÀ and CF3À) are
also displayed. Normalised intensity is the ratio of the absolute inten-
sity of the selected ion over the total corrected intensity which is the
total intensity minus HÀ intensity. Four areas were measured per
sample. MAD-Gal was expected to be observed only on sample C.
Characteristic ions were observed on sample C, not observed on sam-
ple A while sample B exhibited slightly higher intensities than sample
A.
Figure 2. FÀ normalised intensity values [absolute intensity of FÀ/
(total intensityÀHÀ intensity)] and fluorine atomic percentages are
displayed as a function of the MAD-Gal concentration used for
immobilisation. The intensity of surface characteristics signatures
increases with increasing concentration of MAD-Gal.
Table 2. XPS atomic percentages and ToF-SIMS relative intensities
for the different steps of lactose aryl diazirine immobilisation on PS
At percentages (%)
Sample A
Sample B
Sample C
N
F
S
C
O
bdl
bdl
bdl
0.34Æ0.20
bdl
bdl
0.51Æ0.49
1.18Æ0.4
0.18Æ0.16
97.79Æ0.36 94.53Æ1.27 88.77Æ0.54
2.27Æ0.47
5.13Æ1.08
9.35Æ0.49
ToF-SIMS
Corrected total
intensity (counts)Â104
FÀ normalised
intensity %
16.3Æ4.5
1.10
49.6Æ5.4
43.7Æ10.7
2.72Æ0.42 63.17Æ21.78
0.20Æ0.08 0.65Æ0.07
CFÀ3 normalised
intensity %
0.13
XPS and ToF-SIMS analysis of lactose aryl diazirine grafted PS. XPS
atomic percentages are displayed for samples A, B and C. Sample A
corresponds to as received polystyrene. Sample B corresponds to
polystyrene on which a methanolic solution of lactose aryl diazirine
was deposited, and no light activation was performed before the final
washing. This surface should be similar to sample A. Sample C corre-
sponds to polystyrene on which a methanolic solution of lactose aryl
diazirine was deposited and light activation was performed before the
final washing. The molecule should be present at the surface of the
material. Three areas were analysed per sample. ToF-SIMS normal-
ised intensities of sample A, B and C are also displayed. Normalised
values were calculated by dividing the absolute intensity of secondary
ions by the corrected total intensity (=total intensity from which
intensity of HÀ was subtracted). Three areas were analysed per sample.
Lactose aryl diazirine observed only on sample C.
Figure 3. Binding of biotinylated Allo A lectin to MAD-Gal deriva-
tised PS surface. The derivatised PS surfaces were incubated with Allo
A lectin and then extensively washed to remove physisorbed lectins.
[
35S] streptavidin was added and the surfaces were again washed to
remove excess streptavidin. Finally, radioactivity was measured by
scintillation counting. The radioactivity is higher on sample C and
increases with the concentration of the MAD-Gal solution (C, C1 and
C2 corresponding to 0.25, 0.025 and 0.0025 mM, respectively) used for
immobilisation. This illustrates the higher binding of streptavidin to
the surface C. I corresponds to incubation of MAD-Gal grafted PS
with asialofetuin and Allo A lectin. 100% corresponds to the total
added radioactive streptavidin.

Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
2947
not affect xenobiotic metabolism or viability as indi-
Enzym-catalysed sialylation of immobilised galactose
cated by high EROD activities and protein content (as
an indicator of number of attached cells). Both are
comparable to those obtained with the reference sub-
strate CMF (data not shown). Figure 5 shows that NR
uptake of hepatocytes cultured on MAD-Gal coated
polystyrene surfaces increased in proportion with the
density of galactose molecules on the surface. MTT
formation (not shown) and NR uptake increased up to
a concentration of 2.5 mM galactose and showed no
saturation effect (Fig. 5). NR uptake was reduced, when
lectins (RCA) were added most probably because RCA
competes with asialoglycoprotein receptors for terminal
galactose residues of MAD-Gal (Fig. 5).
Experiments in solution (data not shown) demonstrated
that transferase activity was 30 times slower with MAD-
Gal than with the natural substrate while it was only 2
times slower with N-acetyllactosamine pNp than with
lactose aryl diazirine. It was assumed that this difference
was due to the succinimidyl group of MAD-Gal.
Accordingly solid phase semisynthesis experiments were
carried out on lactose modified polystyrene only, as
illustrated in Figure 7. As shown in Figure 7, transfer of
radioactive NeuAc to immobilised lactose proceeded
linearly up to 4 h under the experimental conditions.
Control including zero time incubation and a no
enzyme (uncontrolled attachment due to CMP-Neu5Ac
decomposition) control afforded background incor-
poration level only (approx. 20 cpm).
In the case of lactose modified polystyrene, no increase
in the NR uptake was observed (Fig. 6).
Discussion
Initially, it was planned to synthesise the crosslinker 4 by
condensation of 6-mercaptohexanoic acid with 3-(3-tri-
fluoromethyl-3H-diazirin-3-yl)phenylamine 3 (Scheme
1). This crosslinker 4 would have a thiol function for
Figure 4. Binding of biotinylated Allo A lectin to aryl lactose deriva-
tised PS surfaces. Derivatised surfaces were incubated with Allo A
lectin and then extensively washed to remove physisorbed lectins. [35S]
streptavidin was incubatedand the surfaces were again rinsed to
remove excess streptavidin. Finally, radioactivity was measured by
scintillation counting. The radioactivity is higher on sample C but in
comparison with Figure 3, it can be concluded that no specific affinity
was observed in this case.
Figure 6. Primary rat hepatocytes were incubated on PS modified with
aryl lactose diazirine. No effect of the surface was observed.
Figure 5. NR uptake is related to the lysosomal activity. The NR
uptake decreased with the decreasing of surface galactose residues (C0,
C, C1 and C2 corresponding to 2.5, 0.25, 0.025 and 0.0025 mM,
respectively). The NR uptake was similar to that of surface A (plain
polystyrene) when the surface were incubated with a lectin (RCA) that
protects the galactose residues. Hepatocytes possess higher NR uptake
on asialofetuin coated surfaces than on surface A, but lower than on
MAD-Gal grafted PS. 100% corresponds to surface A.
Figure 7. PS modified with aryl lactose diazirine was incubated with a-
2,6-sialyltransferase. Incorporation of radioactive NeuAc was linear
for 4 h.

2948
Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
coupling with oligosaccharides and a diazirine function
for carbene-mediated surface photo-immobilisation.
The biological availability of the terminal galactose
residues was tested with Allo A lectin and primary rat
hepatocytes. Allo A showed affinity for MAD-Gal
modified PS. Although, Yamashita et al.11 and Umetsu
et al.31 demonstrated that hemagglutination was not
inhibited by simple carbohydrates such as d-glucose,
d-galactose, Adachi et al.32 reported a similar test with
galactose coated polystyrene surfaces [thin film coating
with a lactonamide containing polymer (PVLA)]. In a
competitive inhibition study with asialofetuin
(1 mg mlÀ1), they obtained 70% inhibition. With MAD-
Gal coated polystyrene, 100% inhibition was obtained
using the same concentration of asialofetuin. This sug-
gests that in this case, either the galactose surface den-
sity or the binding of Allo A for surface grafted
galactose was lower.
However, model reactions with aniline demonstrated
that coupling of lactose octa-acetate with 6-mercapto-
hexanoic acid phenylamide was not successful with BF3-
Et2O (60% recovery of the starting thiol) using
conditions that have been successful in other
cases.25,26
The successful procedure investigated is illustrated in
Scheme 2. Condensation of a carboxylic acid with the
amine requires the activation of the carboxylic group.27
Activation of thioglycoside 6 with N,N0-dicyclohexyl-
carbodiimide (DCC)/hydroxy benzotriazole was not
successful. The lack of reaction may have been owing to
steric hindrance by the 3-(trifluoromethyl)-diazirinyl
group or to the inductive effect of the 3-trifluoromethyl-
3H-diazirin-3-yl group since DCC/hydroxy benzo-
triazole activation gave an 80% yield of the condensa-
tion product with aniline (data not shown).
In the case of lactose modified PS, Allo A did not exhi-
bit affinity for the terminal galactose. Therefore, the
difference in lectin binding could be attributed to dif-
ferences in affinity of the lectin for the different terminal
galactose residues at the surface.
According to Collioud et al.,19 4-maleimidobutyric acid
can be successfully coupled to 3-(3-trifluoromethyl-3H-
diazirin-3-yl)phenylamine using mixed anhydride acti-
vation of the carboxylic acid with ethyl chloroformate.
However, the coupling reaction of the thioglycoside 6
with 3-(3-trifluoromethyl-3H-diazirin-3-yl)phenylamine
under the same conditions failed. Nevertheless, at a
lower temperature (À65 ꢀC) and with three equivalents
of triethylamine, the reaction proceeded successfully to
give the amide 7. Deprotection of 7 gave 8 with a lower
yield than expected. This could suggest that either some
material is lost by adsorption on Celite or/and Amber-
lyst or that 7 may contain impurities.
On the cellular level, an effect was also observed only in
the case of MAD-Gal modified surfaces. MTT forma-
tion and NR uptake were increased. NR uptake showed
to be dependent upon galactose surface density. This
effect on NR uptake needs further explanation which
may relate to an unknown intracellular effect of immo-
bilized asialoglycoprotein receptors on lysosomal activ-
ity as well as MTT formation. ASGP-R
(asialoglycoprotein receptor) found as surface receptors
of hepatocytes is expressed on their sinusoidal mem-
branes33 when the cells are kept at high cell density34
similar to the ones chosen in the present investigation. It
might well be, that binding of ASGP-R to galactose
induces endocytosis and degradation processes (lyso-
somal activity) comparable to the effects obtained by
asialoglycoproteins in an intact organ. This would
explain why galactose coating increases NR uptake
without altering the viability of hepatocytes.
In conclusion, 6-(b-d-galactosyl-(1!4)-b-d-glucopyr-
anosylsulfanyl)hexanoic acid [3-(3-trifluoromethyl-3H-
diazirin-3yl)-phenyl]amide 8 was successfully synthe-
sised, albeit in low yield. Better yield may be achieved
using other synthetic routes (by activation of the glyco-
syl donor with trichloroacetimidate28,29 or by silylation
followed by coupling using ZnI230).
No effect was observed for lactose modified surfaces.
The literature is well documented on ligands that have
been used for the study of the asialoglycoprotein
receptor.35À41 The structure of the lactose aryl diazirine
should not impair this interaction according to the lit-
erature. In consequence, it remains to be clarified why
hepatocytes interacted with MAD-Gal modified poly-
styrene, while they did not with lactose modified poly-
styrene.
Upon light activation of the diazirine, a reactive carbene
is generated and may lead to covalent bonding to the
substrate if close contact is allowed. Fluorine can be
related to the presence of the molecule. XPS atomic
percentages and ToF-SIMS relative intensities (Tables 1
and 2) demonstrate the successful grafting of the two
molecules to PS surface.
The extent of grafting of the two molecules is of the
same order of magnitude and is estimated between 1012
and 1013 molecules cmÀ2 (corresponding to the detec-
tion limit of XPS). According to ToF-SIMS FÀ nor-
malised intensity, MAD-Gal grafting is higher than
lactose aryl diazirine while, according to XPS, it is the
opposite. However, owing to the complexity of second-
ary ion creation (possible difference in matrix effects),
one should rather consider XPS data to compare graft-
ing extents, which are then only slightly higher for lac-
tose aryl diazirine.
In the case of enzym-catalysed sialylation, incorpora-
tion of NeuAc was observed for only lactose modified
surfaces and proceeded linearly up to 4 h. Based on the
estimation of 1012–1013 molecules per cm2, the estimated
incorporation yield for 6 h incubation is in the range of
4–40%. According to Yamada et al.,42 the extent of
galactosylation is dependent upon the primer arm in
polymer supported semisynthesis with bovine galactosyl
transferase. These authors reported a yield of 30% in
the case of short-arm primer and a nearly quantitative
yield with longer-armed primer for 24 h incubation

Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
2949
times.42 In our case, the incorporation (on a linear
approximation) was lower than the one observed by
Yamada et al. The difference might be related to the spacer
structure. Accordingly, further investigation should focus
on the spacer design to increase the transferase efficiency.
The efficiency of the transfer reaction may also be ham-
pered by increasing concentration of CMPwhich inhibits
sialyltransferase by competitive inhibition.43
performed with ammonia as the ionising gas on a Ner-
mag R-10-10 quadrupole spectrometer. Electrospray
ionisation was also used. The analyses were carried out
on a VG Platform (micromass) single quadrupole MS in
the negative mode, courtesy of the University of Bern.
The samples were dissolved in MeOH/H2O (1/1, v/v)
containing 1% of triethylamine. The major ions were all
singly charged.
Accurate mass spectra of carbohydrate derivatives were
obtained using a Fourier transform ion cyclotron mass
spectrometer (Bioapex, Bruker Daltonics, USA) equip-
ped with a 94 Tesla magnet (Magnex, Abingdon, UK).
1H NMR and 13C NMR spectra were obtained using
either an AC-P-200 spectrometer (4.7 T) or an AC-P-
250 (5.9 T) spectrometer equipped with an auto sam-
pler. CDCl3 and CD3OD (Glaser, Switzerland) were
used as solvents.
Conclusions
Diazirine-containing photoreagents containing mono-
and disaccharides were successfully synthesised and
immobilised on polystyrene. Biological activity of the
modified polystyrene was demonstrated with Allo A
lectin, primary rat hepatocytes and a-2,6-sialyltransfer-
ase. Different responses of lectin Allo A and hepato-
cytes to the galactose and lactose modified surfaces
indicate that the biological activity depends not only on
the carbohydrate but also on the mode of biomolecule
presentation (structure of the spacer). Improvement of
the biological interactions will require more complex
structures (cluster, spacer design, oligosaccharides).
XPS analysis was carried out under UHV conditions
(10À9 Torr) using a PHI 5500 system equipped with a
hemispherical analyser and a non-monochromatised Mg
Ka (14.0 kV, 350 W) radiation source. The analysed area
was 0.12 mm2, allowing the acquisition of three spectra
on each sample. Spectra were taken at a 45ꢀ take-off
angle. Pass energies used for survey scans and high
resolution elemental scans were 93.5 and 23.5 eV,
respectively. Spectra were referenced in the C1s spec-
trum to C–H/C–C at 285 eV (PS). The spectrometer was
calibrated by using Cu 2p3/2, Ag 3d5/2 and Au 4f7/2 at
932, 368, 84 eV, respectively. Data reduction was per-
formed by applying the PHI-Access Software 5.2.44
Absolute areas were calculated using a parabolic inte-
gration routine. Background was calculated using a
trapezoid area after subtracting a Shirley–Sherwood
background. The corrected area was normalised to time
and to the unit step to give the intensity in counts sÀ1
eVÀ1. Instrumental sensitivity factors (PHI sensitivity
factors) were 0.296, 0.711, 0.477, 1.000, 0.339 for carbon
(C1s), oxygen (O1s), nitrogen (N1s), fluorine (F1s) and
silicon (Si2p), respectively.44 Atomic percentages were
calculated using the above sensitivity factors Sx and the
relation: X%=(Ix/Sx)/(Æ Ii/Si). Sample charge-up was
observed for polystyrene substrate and was compen-
sated by low energy electrons (maximum of 12 eV).
Experimental
Materials
Biotinylated Allo A lectins (Allo A-I MW 65000, Allo
A-II MW 66500 g/mol) were purchased from Oxford
Glycosystems (Oxford, UK). Radiolabelled [35S] strep-
tavidin from Amersham, UK, (Mr 60,000 g molÀ1, spe-
cific radioactivity 1.35 TBq mmolÀ1
, radioactive
concentration 3.70 MBq mlÀ1) was used to detect lectin
binding. Radiolabelled (14C) CMP-Neu5Ac (Amer-
sham, UK) was used to probe the enzymatic reaction.
The sialic acid was labelled at positions 4, 5, 6, 7 and 9.
(3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) was purchased from Fluka (Switzer-
land). CYP1A1 was purchased from Amersham Life
Science (UK). Liquid scintillation measurements were
performed with ultima gold scintillation liquid (Pack-
ard) on a Tri-Carb 2300tr scintillation beta counter
from Packard (Switzerland). Dry toluene was purchased
from Fluka (Switzerland).
The ToF-SIMS system used in this study was a Trift 1
mass spectrometer from PHI-EVANS & Associates
(described in detail elsewhere45,46) equipped with a
pulsed FEI Ga+ ion gun operated at 15 kV. Sample
surfaces were biased Æ3 kV with respect to the groun-
ded extraction electrode for positive and negative mode
SIMS, respectively. The 800–1700 pA DC ion beam
current was pulsed at 5 kHz repetition rate (8 ns pulse
width, measured with the unbunched beam). The ana-
lysed area was estimated to 84Â84 mm2. The analysis
was performed under so-called ‘static’ conditions with
an ion dose on the order of 1012 ions cmÀ2. All spectra
were collected using high mass resolution conditions
(bunched beam). The mass resolution obtained for Si+/À
on a Si wafer was m/Ám>3600 in the positive mode
and >2400 in the negative mode. Charge compensation
was performed with a pulsed 20 eV electron gun.
Falcon multiwell (non tissue culture treated, Heidelberg,
Germany) or ELISA plates (Nunc, Switzerland) were used
for XPS analysis, biomolecule and cell assays. These PS
dishes are g-ray sterilised by the manufacturer. The plates
were used as received. For ToF-SIMS experiments, poly-
styrene (Mr 30000 g/mol, monodispersed, kindly provided
by the polymer laboratory (LP-DMX) of EPFL) was spin
cast on a silicon wafer (5Â5 mm) from a 30 mg mLÀ1
solution (in dry toluene) in order to limit charging of the
sample during analysis. Spin casting conditions were 65 s
at 5000 rpm (acceleration 1000 r minÀ1 sÀ1).
PBS (Phosphate buffered saline, 150 mM NaCl, 50 mM
phosphate, pH 7.2–7.4) and cacodylate buffer (pH 6.8,
1 M) were prepared on site. Chemical ionisation was

2950
Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
ToF-SIMS spectra were recorded for each sample on
three different sample spots and in each mode (positive
and negative). Values displayed are mean values and
standard deviations from negative mode data. ToF-
SIMS data were analysed using Cadence software from
Charles Evans & Associates. Mass calibration was per-
formed using similar calibration peaks for a meaningful
comparison. Accurate calibration was achieved using
hydrocarbon peaks such as CHÀ, C2HÀ, C3HÀ, and
C4HÀ (in the negative mode) as well as low intensity
characteristic PDMS (polydimethylsiloxane) peaks such
as C5H15O4SiÀ3 . The standard deviation on all the
detected peaks was less than 10 milli-atomic mass unit
(mamu). It must be noted that the data treatment was
based on intensity values measured in mass windows
which were identical for all the samples. The residual
intensity measured in some cases should be considered
as noise but was not arbitrarily subtracted in the data
presentation. ToF-SIMS peak values were normalised
by dividing the absolute peak intensity of secondary
ions by the corrected total intensity, that is the total
intensity from which intensities of H+/À were sub-
tracted due to limited reproducibility of their absolute
values.
tion required two separations by flash chromatography
(CH2Cl2/MeOH 20:1 and 10:1) to give the derivative 6
(1.65 mg, 67%). dH (250 MHz; CDCl3) 1.34–1.46 (2 H,
m, CH2), 1.54–1.67 (4 H, m, 2ÂCH2), 1.96 (3H, s,
COCH3), 2.02 (3H, s, COCH3), 2.03 (6H, s, COCH3),
2.04 (3H, s, COCH3), 2.09 (3H, s, COCH3), 2.13 (3H, s,
COCH3), 2.32 (2H, t, J=7.4 Hz, CH2), 2.54–2.68 (2H,
m, CH2), 3.58–3.61 (1H, m, 50-H), 3.75 (1H, t, J=9.4,
40-H), 3.85 (1 H, t, J=6.6 Hz, 5-H), 3.98–4.19 (m, 3H,
C-6 and 6b0-H), 4.44 (1H, d, J=9.2 Hz, 1-H), 4.84–4.96
(2H, m), 5.07 (1H, dd, J=6.2 Hz), 5.18 (1H, t,
J=9.4 Hz, 40-H), 5.32 (1 H, d), dC (62.90 MHz; CDCl3)
20.37, 20.50, 20.61, 20.71, 23.96, 27.89, 29.17, 33.54,
60.68, 62.13, 66.49, 68.95, 70.15, 70.53, 70.85, 73.63,
76.04, 83.35 (1-C), 100.91 (10-C), 169.01 (COCH3),
169.59 (COCH3), 169.68 (COCH3), 170.07 (COCH3),
170.30 (COCH3), 178.35 (COCH3, or CH2COOH). IR
1665 cmÀ1. Found m/z 789.2246. (C32H46O19S+Na)+
requires m/z 789.2246.
6-(2,3,4,6-Tetra-O-acetyl)-ꢀ-D-galactopyranosyl-(1!4)-
2,3,5-triacetyl-ꢀ-D-glucopyranosylsulfanyl)hexanoic acid
[3-(3-trifluoromethyl-3H-diazirin-3-yl)phenyl]amide 7. m-
[3-(Trifluoromethyl)diazirine-3-yl]aniline 3 was stored as
N-[3-(3-Trifluoromethyl-3H-diazirin-3-yl)phenyl]forma-
mide to avoid oxidation of the amine function. N-[3-(3-
Trifluoromethyl - 3H - diazirin - 3 - yl)phenyl]formamide
(120 mg) was dissolved in methanol (3 cm3). Con-
centrated HCl (1 mL) was added. After 10 min, the
mixture was cooled to 0 ꢀC and neutralised with NaOH
(6 M, 3 mL). The aqueous phase was extracted with
ethyl ether (5Â4 mL). The combined organic phases
were washed with deionised water (5 mL) and dried
MgSO4). The solvent was evaporated under reduced
pressure to give m-[3-(trifluoromethyl)diazirine-3-yl]ani-
line 3 as a yellow oil which was used without further
purification. A solution of thioglycoside 6 (383 mg) in
dry DMF (9 mL) was cooled to À20 ꢀC in a salt-ice bath
under nitrogen. Triethylamine (209 mL) was added
dropwise. The mixture was further cooled at À60 ꢀC in a
dry ice/acetone bath. Ethyl chloroformate (80 mL in
1.1 mL of dry DMF) was added. The reaction was
allowed to proceed for 30 min. A solution of the m-[3-
(trifluoromethyl)diazirine-3-yl]aniline (above) in dime-
thylformamide (1.1 mL) was added dropwise. The mix-
ture was allowed to stand for 5 h and 35 min at room
temperature. The reaction mixture was poured on to a
mixture of NaCl solution (1 M, 40 mL) and ice and was
extracted with diethyl ether (4Â15 mL). The combined
organic phases were washed sequentially with HCl (1 M,
2Â10 mL), saturated NaHCO3 (2Â10 mL) and deion-
ised water (2Â10 mL). The organic phase was then dried
over MgSO4. The solvent was evaporated under
reduced pressure. The residual oil was purified by flash
chromatography (1.5Â16 cm) on silica gel 60 (0.015–
0.04 mm, column dimensions 1.5Â16 cm) with ethyl
acetate-hexane (1:1, 200 mL) as eluent to give the amide
7 (71 mg), Rf 0.1 [ethyl acetate/hexane (1:1)]. dH
(200 MHz; CDCl3) 1.37–1.54 (2H, m, CH2), 1.54–1.85
(4H, m, 2ÂCH2), 1.96 (3H, s, COCH3), 2.03 (3H, s,
COCH3), 2.04 (3H, s, COCH3), 2.05 (3H, s, COCH3),
2.06 (3H, s, COCH3), 2.13 (3H, s, COCH3), 2.15 (3H, s,
COCH3), 2.34 (2H, t, J=6.9 Hz, CH2), 2.54–2.72 (2H,
Methods
6-Mercaptohexanoic acid 2. A mixture of e-caprolactone
(13.6 g, 120 mmol), thiourea (8.4 g, 110 mmol) and
hydrobromic acid 48% (9.8 g, 120 mmol) was heated
(ꢁ120 ꢀC) with stirring in an oil bath for 9 h, then
cooled to room temperature. Sodium hydroxide solu-
tion (50% m/v) was added until a clear solution was
obtained. This solution was further heated (ꢁ100 ꢀC)
for 3 h, cooled to room temperature and acidified with
sulphuric acid (aqueous solution 50% v/v) until a clear
solution was obtained (pHꢁ1). The clear solution was
extracted with diethyl ether (3Âl00 cm3). The combined
organic phase was washed with water (3Âl00 cm3), dried
(MgSO4) and the solvent was evaporated under reduced
pressure. The remaining yellow oil was purified by dis-
tillation under reduced pressure (135–140 ꢀC, ꢁ1 mm
Hg) to give the acid 2 as a colourless oil (3.76 g). dH
(250 MHz; CDCl3) 1.34 (1H, t, J=7.9 Hz, SH), 1.37–
1.51 (2H, m, CH2), 1.57–1.71 (4 H, m, 2ÂCH2), 2.37 (2
H, t, J=7.3 Hz, CH2COOH), 2.53 (2 H, dt, J=7.9 Hz,
7.0, HSCH2), 10.4 (1 H, br, COOH); dC (62.90 MHz;
CDCl3) 23.92, 24.21, 27.56, 33.41 (HSCH2), 33.75
(CH2COOH), 179.92 (COOH); m/z (+CI) 166
([M+NH4]+, 1.7%), 148 ([M]+, 0.8), 130 (2.8), 115
(2.8), 102 (3.5).
6-(2,3,4,6-Tetra-O-acetyl)-ꢀ-D-galactopyranosyl-(1!4)-
2,3,6-triacetyl-ꢀ-D-glucopyranosylsulfanyl)hexanoic acid
6. To a stirred mixture of lactose octaacetate 2.72 g
(4 mmol), 6-mercaptohexanoic acid (0.48 g, 3.2 mmol)
and dry CH2Cl2 (10 mL) under N2, BF3 diethyl etherate
(0.85 g, 6 mmol) were added dropwise. The mixture was
stirred for 2 h at room temperature under nitrogen,
diluted with CH2Cl2 (200 mL) and washed with HCl
(1 M, 3Â100 mL) and water (1Â100 mL). The organic
phase was dried (MgSO4), and the solvent was evapo-
rated at reduced pressure to give a yellow oil. Purifica-

Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
2951
m, CH2), 3.58–3.68 (1H, m, C-50), 3.81 (1 H, t,
J=9.4 Hz, C-40), 3.87 (1H, t, J=6.6 Hz, C-5), 4.00–
4.22 (m, 3H, C-6 and C-6b0), 4.44 (1H, m, C-1),
4.88–5.02 (2H, m), 5.15 (1H, m), 5.21 (1H, m), 5.32
(1H, d), 7.08 (1H, m), 7.3 (1H, m), 7.5 (2H, m,
aromatic). dC (50.3 MHz; CDCl3) 20.51, 20.63,
20.79, 20.96, 24.78, 27.74, 29.21, 29.81, 37.28, 60.76,
61.96, 66.61, 69.17, 70.22, 70.75, 70.98, 73.77, 76.07,
76.39, 76.81, 77.03, 77.66, 83.66 (C-1), 101.07 (C-10),
117.36, 120.75, 121.89, 129.63, 129.92, 138.90,
168.99 (COCH3), 169.67 (COCH3), 169.87 (COCH3),
170.07 (COCH3), 170.13 (COCH3), 170.34 (COCH3),
170.72 (COCH3), 171.43 (O¼C–N). MS ESI: calcd:
948.26 m/z, exp: 948.20 m/z, additional peaks at
822 (MÀH-3 Ac), 780 (MÀH-4 Ac), 738 m/z (MÀH-
5 Ac).
grafting extent was studied with 0.25, 0.025 and
0.0025 mM solutions.
The solvent was evaporated for 2 h at 30–40 mbar
before
irradiation
of the samples (350 nm,
0.95 mW cmÀ2, 25 min). The samples were washed four
times in water (HPLC grade). Samples are referred to as
A, B and C. Sample A is the pristine polystyrene. Sam-
ple B is polystyrene with MAD-Gal deposited and
washed. This sample treatment does not include light
exposure. Physisorbed molecules should be removed by
washing. Sample C is the polystyrene surface on which
the molecules were deposited. The surface was then
photoactivated and washed. Covalently attached mol-
ecules are expected to remain bound to the surface, by
contrast with sample B.23
6-(ꢀ-D-Galactosyl-(1!4)-ꢀ-D-glucopyranosylsulfanyl)-
hexanoic acid [3-(3-trifluoromethyl-3H-diazirin-3yl)-phe-
nyl]amide 8. To a solution of the acetylated glycoside 7
(71 mg) in dry methanol (8 mL) was added dropwise,
under nitrogen, a solution of sodium methoxide in
methanol (0.12 mM, 0.6 mL, 0.6 equiv). After 9 h the
mixture was neutralised with Amberlyt 15, and filtered
over Celite. The organic solvent was evaporated under
reduced pressure to give thioglycoside 8 (14.3 mg). dH
(200 MHz; MeOH-d4) 1.49–1.69 (2H, m, CH2), 1.69–
1.89 (4H, m, 2ÂCH2), 2.46 (2H, t, J=7.2 Hz, CH2),
2.73–2.95 (2H, m, CH2), 3.44–4.09 (12H, m, carbohy-
drate, methanol), 4.39 (1H, C-10), 6.94–7.16 (1H, m,
aromatic), 7.39–7.55 (1H, m, aromatic), 7.64–7.76 (2H,
m, aromatic). dC (50.3 MHz; MeOH-d4) 26.25, 29.35,
30.70, 30.75, 37.80, 62.08, 62.5, 64.43, 70.30, 72.54,
74.10, 74.83, 77.09, 77.94, 80.4, 80.5, 87.08, 105.04,
118.63, 122.20, 122.65, 130.73. Quaternary carbons
(CF3, C¼O, C aromatic–N, C aromatic–C) are not
observed owing to the low signal/noise ratio. MS ESI
(MÀ1): Found m/z 678.1894. (C26H36F3N3O11S+Na)+
requires m/z 678.1921. TLC in CH2Cl2/MeOH (2:1) of
the deprotected compound 8 (Scheme 2) gave a single
spot on TLC (Rf 0.37).
Lectin binding to lactose or galactose modified
polystyrene
Lectin binding studies were performed on MAD-Gal (1)
or lactose aryl diazirine (8) modified ELISA plates.
Individual wells were coated with BSA (1 mg/mL,
100 mL) in PBS and 1 mL of a 500 mg mlÀ1 biotinylated
Allo A lectin (500 mg mLÀ1,, 1 mL) in PBS was added.
For inhibition tests, incubation was performed in BSA
solution (1 mg mLÀ1 in PBS, 90 mL), followed by asia-
lofetuin (10 mg mLÀ1
, 10 mL) and Allo A lectin
(500 mg mLÀ1 1 mL). The samples were incubated for 1 h
at room temperature with agitation. The samples were
washed subsequently three times with Tween (0.02% in
PBS, 100 mL, 5 min, agitation), three times with PBS
(100 mL, 5 min, agitation), and once with deionised
water (100 mL, 5 min, agitation).
Commercial [35S] radiolabelled streptavidin solution
was used to quantify surface bound biotin–Allo A. The
stock solution was diluted in water (1:100). Aliquots
(50 mL) of the solution were added to each well. The
plate was incubated for 1 h at room temperature. The
samples were washed three times with Tween 0.02% in
PBS (100 mL, 5 min, agitation), three times with PBS
(100 mL, 5 min, agitation) and once with deionised water
(100 mL, 5 min, agitation). For radioactivity measure-
ments, individual wells were placed in scintillation vials,
and dissolved in toluene (30–40 min). Scintillation fluid
(5 mL) was added and radioactivity was measured by
scintillation counting. One hundred percent of radio-
activity corresponded to the radioactivity of 50 mL of
the diluted streptavidin solution. Triplicates of samples
A, B and C were measured.
Immobilisation of MAD-Gal 1and 6-( ꢀ-D-galactosyl-
(1!4)-ꢀ-D-glucopyranosylsulfanyl)hexanoic acid [3-(3-
trifluoromethyl-3H-diazirin-3yl)-phenyl]amide 8 on poly-
styrene. Immobilisation on PS was carried out as
described elsewhere.23 Briefly, four different polystyrene
sample types were prepared for further experimental
characterisation. Volume of the deposited solution was
a function of sample area.
Droplets of a MAD-Gal or lactose aryl diazirine solu-
tion (in ethanol and in methanol, respectively) were
deposited on the PS samples. Following the experi-
ments, various PS types and various concentrations
were used: 10 mL of a solution (0.25 mM) on spin-casted
PS for ToF-SIMS; 25 mL of solution (0.25 mM) on PS
cut out from falcon well for XPS; 45 mL of solution
(0.25 mM) on individual Elisa wells for Allo A and
solid-phase semi-synthesis; 80 mL of solution (0.25 mM)
on Falcon non tissue culture treated plates for rat
hepatocytes. Moreover, in the case of XPS and Allo A,
the effect of the MAD-Gal solution concentration on
Isolation and cultures of primary rat hepatocytes
The reagents used for hepatocyte isolation and cultures
were identical to those described earlier.47 The hepato-
cytes were isolated by two-step in situ liver perfusion47
from male Sprague–Dawley rats [Zur: SD (Crl:CD1)],
(Institute of Toxicology, Swiss Federal Institute of
Technology, Zurich, Switzerland; age 54Æ2 days; 240–
¨
270 g body weight) kept under controlled environmental
conditions (12 h light–dark cycle, diet No. 3430 KLIBA,
Kaiseraugst, Switzerland, food and water ad libitum) as

2952
Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
reported.47 Isolated hepatocyte showed a viability of
>85% (assessed by Trypan Blue exclusion in samples
without bovine serum albumin). The contamination
with non parenchymal cells was less than 0.5%. The
hepatocytes were seeded into 24-well non tissue culture
plates at densities of 50Â103 cells cmÀ2, in William’s
medium E (Sigma) without phenol red, supplemented
with dexamethasone (100 nM), insulin (10 nM), l-gluta-
isopropanol and fluorescence was measured at 530 nm
(excitation) and 645 nm (emission) (Millipore Cyto-
Fluor 2350 microplate reader).
EROD assay. The enzyme activity was determined in
situ. Hepatocyte monolayers were incubated 30 min in
0.5 mL of Hank’s buffer solution supplemented with the
corresponding substrate (5 mM 5-ethoxyresorufin) and
10 mM dicumarol as described by Wortelboer et al.50
Fluorescence in the supernatant was measured after 15
and 30 min at Ex/Em: 530 nm (excitation) and 580 nm
(emission) (Millipore CytoFluor 2350 microplate
reader) and the enzyme activity was calculated.
mine (2 mM), selenium (30 nM), aprotinin (1 mg mLÀ1
)
and penicillin/streptomycin (100 IU mLÀ1). The cells
were cultured for 5 days at 37 ꢀC at 5% CO2 in air. The
influence of the surface graft on cellular functions was
assessed by comparing cellular functions in uncoated
and coated wells on the same 24-well plates. Coating
includes a well-established extracellular matrix as refer-
ence [Crude Liver Membrane Fraction (CMF) 30 mg
and 0.3 mg collagen in 0.5 mL PBS overnight at 4 ꢀC] as
described elsewhere48 and asialofetuin (1 mg in 0.5 mL
PBS overnight at 4 ꢀC) as a mimic of the galactose-
coated surface. To some galactose modified wells, addi-
tionally 5 mg of a galactose binding lectin (from Cytisus
scoparius, Sigma, St Louis, USA) was applied in 0.5 mL
PBS for 1 h at room temperature.
Incorporation of sialic acid using ꢁ-2,6-sialyltransferase
Diazirine 8 was photoimmobilised on ELISA title plates
(Nunc). The recombinant a-2,6-sialyltransferase18 from
Pichia pastoris was diluted with water (34 mL, except for
the no enzyme control in which 44 mL of water was added)
and Cacodilate buffer (5 mL). The enzyme substrate
[14C]CMP-Neu5Ac (0.925 kBqmLÀ1, 10.8 GBq mmolÀ1
)
was added and the samples were incubated for various
times at 37 ꢀC. The reaction was stopped with 100 mL of
cold water (0 ꢀC). The wells were then sequentially
washed four times with water. Radioactivity retained in
the PS wells was quantified as described above (tripli-
cate analysis).
Hepatocyte function
The binding of hepatocytes to the engineered poly-
styrene surfaces was investigated in four ways, each of
which is related to specific changes in cell function as a
result of the hepatocyte surface interactions: This
includes lysosomal activity by neutral red (NR) uptake
(neutral-red assay49), the total protein as an estimation
of attached cells (cell density, BCA assay, Pierce,
Rockford, IL, USA), the total cellular bioreductive
capacity or viability by using MTT [3-(4,5-dimethyl-
thiazol-2-yl)-2,5-diphenyltetrazolium bromide, MTT-
assay50] and an estimation of the activity of a repre-
sentative cytochrome P450 enzyme (CYP 1A1) involved
in the oxidative metabolism of xenobiotics using ethox-
yresorufin de-ethylase activity (EROD-activity50).
Acknowledgements
The authors acknowledge financial support from the
Swiss Priority Program on Material Research, Bern and
SPP-Biotech grant 002–46084 (to E.G.B.).
References and Notes
1. Ratner, B. D. J. Biomed. Mater. Res. 1993, 27, 837.
2. Ruiz, L.; Hilborn, J. H.; Leonard, D.; Mathieu, H. J. Bio-
materials 1998, 19, 987.
3. Lee, Y. C.; Lee, R. T. Accounts Chem. Res. 1995, 28, 321.
4. Petrak, K. Adv. Drug Deliv. Rev. 1994, 13, 211.
5. Rademacher, R. W.; Parekh, R. B.; Dwek, R. A. Annu.
Rev. Biochem. 1988, 57, 785.
6. Varki, A. Glycobiology 1995, 3, 97.
7. Hatanaka, K.; Takeshige, H.; Akaike, T. Carbohyd. Chem.
1994, 13, 603.
8. Hermanson, G.T.; Mallia, A. K.; Smith, P. K. In Immobi-
lized Affinity Ligand Techniques; Hermanson, G. T., Mallia, A.
K., Smith, P. K. Eds.; Academic: London, 1992; pp 137–200,
Vol. 1.
After 24 h (attachment period) and after 5 days in cul-
ture (end of the experiment), the four parameters were
determine simultaneously.
Protein content. After removal of the medium, the BCA-
solution (Pierce, Rockford, IL, USA) was applied
directly to each well for 20 min and the developing colour
(depending on the amount of total protein) was mea-
sured at 550 nm (Hamilton HR 7000 microplate reader).
MTT assay. The tetrazolium salt (1.2 mM) was applied
in culture medium to each well and incubated for 30 min
(hepatocytes) at 37 ꢀC. The medium was replaced with
ethanol [90% (v/v)] in HEPES–buffer (50 mM, pH 8.0).
The amount of soluble MTT-formazan formed was
measured at 550 nm (Hamilton HR 7000 microplate
reader).
9. Sigrist, H.; Collioud, A.; Clemence, J. F.; Gao, H.; Lugin-
buhl, R.; Sanger, M.; Sundarababu, G. Opt. Eng. 1995, 34,
¨
2339.
10. Chevolot, Y.; Bucher, O.; Leonard, D.; Mathieu, H. J.;
Sigrist, H. Bioconjugate Chem. 1999, 10, 169.
11. Yamashita, K. Method Enzymol. 1989, 179, 331.
12. Stockert, R. J.; Morell, A. G.; Ashwell, G. Targeted
Diagnosis Ther. 1991, 12, 441.
Neutral-red assay. Neutral red was applied in culture
medium to each well and incubated for 30 min (hepato-
cytes) at 37 ꢀC. After fixation of the cells and washing
with PBS, the chromophore was solubilised in 100%
13. Baenziger, J. U.; Maynard, Y. J. Biol. Chem. 1980, 255,
4607.
14. Baenziger, J. U.; Maynard, Y. J. Biol. Chem. 1980, 255,
4607.

Y. Chevolot et al. / Bioorg. Med. Chem. 9 (2001) 2943–2953
2953
15. Schwartz, A. L. Annu. Rev. Immunol. 1990, 8, 195.
16. Ashwell, G.; Harford, J. Annu. Rev. Biochem. 1982, 51,
531.
17. Kawasaki, T.; Ashwell, G. J. Biol. Chem. 1976, 251, 12.
18. Malissard, M.; Zeng, S.; Berger, E. G. Glycoconjugate J.
1999, 16, 125.
33. Massimi, M.; Falasca, L.; Felici, A.; Dini, L.; Conti-
Devirgilis, L. Bioscience Rep. 1996, 16, 477.
34. Lee, R. T. Biochemistry 1982, 21, 1045.
35. Lopina, S. T.; Wu, G.; Merrill, E. W.; Griffith-Cima, L.
Biomaterials 1996, 17, 559.
36. Biessen, E. A.; Bakkeren, H. F.; Beuting, D. M.; Kuiper,
J.; Berkel, T. J. V. Biochem. J. 1994, 229, 291.
37. Biessen, E.; Beuting, D.; Rolen, H.; Marel, G. V. D.; van
Berkel, T. J. Med. Chem. 1995, 38, 1538.
38. Lee, R. T.; Meyer, R. T.; Lee, Y. C. Biochemistry 1982,
21, 6292.
39. Shimada, K.; Kamps, J. A.; Regts, J.; Ikeda, K.; Shio-
zawa, T.; Hirota, S.; Scherphof, G. L. Biochim. Biophys. Acta
1997, 1326, 329.
40. Connolly, D. T.; Towsend, R. R.; Kawagushi, K.; Bell,
W. R.; Lee, Y. C. J. Biol. Chem. 1982, 257, 939.
41. Yamada, K.; Fujita, E.; Nishimura, S. Carbohydr. Res.
1998, 305, 443.
42. Bernacki, R. J. Eur. J. Biochem. 477, 58.
43. In Operator’s Reference Manual Part No. 625612; Perkin
Elmer: Eden Prairie, 1992.
19. Collioud, A.; Clemence, J. F.; Sanger, M.; Sigrist, H. Bio-
conjugate Chem. 1993, 4, 528.
20. Gao, H.; Sanger, M.; Luginbuhl, R.; Sigrist, H. Biosens.
Bioelectron. 1995, 10, 317.
¨
21. Gao, H.; Luginbuhl, R.; Sigrist, H. Sens. Actuator. B
¨
1997, 38–39, 38.
22. Leonard, D.; Chevolot, Y.; Bucher, O.; Sigrist, H.;
Mathieu, H. J. Surf. Interface Anal. 1998, 26, 783.
23. Leonard, D.; Chevolot, Y.; Bucher, O.; Sigrist, W.;
Haenni, H.; Mathieu, H. J. Surf. Interface Anal. 1998, 26,
793.
24. Onyiriuka, E. C. Appl. Spectrosc. 1990, 44, 808.
25. Elofsson, M.; Walse, B.; Kihlberg, J. Tetrahedron 1991,
7613.
26. de la Torre, B. G.; Torres, J. L.; Bardaji, E.; Clapes, P.;
Xaus, N.; Jorba, X.; Clavet, S.; Albericio, F.; Valencia, J. G.
Peptide Protein Res. 1990, 36, 86.
44. Franzreb, K.; Mathieu, H. J.; Landolt, D. Surf. Interface
Anal. 1995, 23, 641.
27. Bodanzky, M. In Peptide Synthesis; Bodanszky, M., Ed.;
Springer: Berlin, 1993.
45. Schueler, B. W. Microsc. Microanal. Microstruct. 1992, 3,
119.
28. Rademann, J.; Schmidt, R. R. Tetrahedron Lett. 1996, 37,
3989.
46. Milosevic, N.; Schawalder, H. P.; Maier, P. Eur. J. Phar-
macol. 1999, 368, 75.
29. Rademann, J.; Schmidt, R. R. J. Org. Chem. 1997, 62,
3650.
47. Saad, B.; Schawalder, H. P.; Maier, P. In Vitro Cell Dev.
Biol. 1993, 29A, 32.
30. Buskas, T.; Garegg, P. J.; Konrasson, P.; Maloisel, J. L.
Tetrahedron: Asymmetry 1994, 5, 2187.
31. Adachi, N.; Maruyama, A.; Ishihara, T.; Akaike, T. J.
Biomater. Sci. Polym. Ed 1994, 6, 463.
32. Braun, J. R.; Willnow, T. E.; Ishibashi, S.; Ashwell, G.;
Herz, J. J. Biol. Chem. 1996, 271, 21160.
48. Borenfreund, E.; Puerner, J. A. Toxicol. Lett. 1985, 24, 119.
49. Borenfreund, E.; Babich, H.; Martin-Alguacil, N. Toxicol.
In Vitro 1988, 2, 1.
50. Wortelboer, H. M.; Kruif, C. A. D.; Iersel, A. A. J. V.;
Falke, H. E.; Nordhoek, J.; Blaauboer, B. J. Biochem. Phar-
macol. 1990, 42, 381.