Pragmatic studies on protein-resistant self-assembled monolayers

Article abstract of DOI:10.1007/s00706-007-0584-7

The present study describes new synthetic routes to oligo(ethylene glycol)-terminated alkanethiols (OEG-ATs), starting from α,ω- dibromoalkanes, which are reacted either with OEG or with trityl mercaptan in the first step. In addition to these ether conju

Full text of DOI:10.1007/s00706-007-0584-7

Monatshefte fu¨r Chemie 138, 245–252 (2007)
DOI 10.1007/s00706-007-0584-7
Printed in The Netherlands
Pragmatic Studies on Protein-Resistant Self-Assembled Monolayers
1
2
1
1
3
¨
Christoph D. Hahn , Ali Tinazli , Martin Holzl , Christa Leitner , Filip Frederix ,
Bernd Lackner , Norbert Mu¨ller , Christian Klampfl , Robert Tampe , and Hermann J. Gruber
4
4
5
2
1;ꢀ
´
1
Institute of Biophysics, University of Linz, Linz, Austria
2
Institute of Biochemistry, Johann Wolfgang Goethe-University, Frankfurt a.M., Germany
3
IMEC, MCP-ART, Leuven, Belgium
4
Institute of Organic Chemistry, University of Linz, Linz, Austria
5
Institute of Analytical Chemistry, University of Linz, Linz, Austria
Received December 12, 2006; accepted (revised) December 20, 2006; published online January 29, 2007
# Springer-Verlag 2007
Summary. The present study describes new synthetic routes
to oligo(ethylene glycol)-terminated alkanethiols (OEG-ATs),
Introduction
Dense brushes of poly(ethylene glycol) chains (PEG)
on solid surfaces or lipid membranes have long been
known to afford high resistance to nonspecific pro-
tein adsorption [1, 2]. Long PEG chains are usually
preferred because short PEG chains require high
grafting site densities which are difficult to achieve
in practice [1, 2]. In contrast, self-assembled mono-
layers (SAMs) of oligo(ethylene glycol)-terminated
alkanethiols (OEG-ATs) are easily prepared with
high OEG density, in which case even short tri(eth-
ylene glycol) chains (EG3) afford complete protein
resistance [3, 4]. Later, it was shown that the ether
bonds between the hydrophobic tail and the polar
OEG chain could be replaced by ester or amide link-
ages (see Fig. 1) while retaining high resistance to
protein adsorption [5, 6].
Protein-resistant OEG-terminated SAMs have
widely been used for suppression of nonspecific ad-
sorption in biosensing [6–13], for passivation of the
area which surrounds the active spots of capture mole-
cules on microarrays [14, 15], as well as on nanoarrays
which had been prepared by dip-pen nanolithography
[14–16] or nanografting [17]. Moreover they represent
a clean, ultraflat surface for single molecule micro-
scopy [11, 18] or for buildup and imaging of large
nanoassemblies [19].
starting from ꢀ,!-dibromoalkanes, which are reacted either
with OEG or with trityl mercaptan in the first step. In addition
to these ether conjugates of OEG and AT, analogous ester and
amide conjugates were prepared by established procedures.
All thiols were used to form self-assembled monolayers
(SAMs) on cleaned gold surfaces and these were stored for
1–2 weeks under water at 4ꢁC before the extent of nonspecific
protein adsorption was tested with IgG, BSA, and lysozyme at
1 mg cmꢂ3 protein concentration in phosphate-buffered saline.
Under these practice-oriented testing conditions, SAMs with
tri(ethylene glycol) chains (EG3) exhibited nonsatisfactory
protein resistance, in sharp contrast to EG4 or longer OEG
chains. The effectiveness of EG3 was partially restored when
they were linked to a long acyl chain (16-mercaptohexa-
decanoic acid) instead of 12-mercaptododecane or 11-mer-
captoundecane. Furthermore it was found that (i) SAM
formation at 20 ꢁM thiol versus 500 ꢁM OEG-AT gave iden-
tical results, (ii) gel-filtered proteins were much less ad-
sorbed than the unpurified commercial products, and (iii)
the method for gold-precleaning was very critical. In con-
clusion, this study offers convenient synthetic routes to
OEG-AT and helps to choose molecules and procedures
for reliable preparation of protein-resitant SAMs with pro-
longed stability during storage.
Keywords. Adsorption; Alkanethiol; Biosensors; Mono-
layers; Proteins.
ꢀ
Corresponding author. E-mail: hermann.gruber@jku.at

246
C. D. Hahn et al.
level (by atomic force microscopy, AFM) [11, 18].
Longer OEG chains are known to form more loosely
packed SAMs on gold [20], and such soft ‘‘cushions’’
appear less favorable as support for AFM studies
on single molecules than thin, compact SAMs with
shorter OEG chains.
Results and Discussion
Synthesis of OEG-Terminated Alkanethiols
from ꢀ,!-Dibromoalkanes
The conventional synthesis route to OEG-AT starts
from 11-bromoundecene, the first step being ether
conjugation with OEG by substitution of the bromine
atom, followed by addition of thioacetic acid to the
terminal C¼C double bond which requires catalysis
by both UV light and azobisisobutyronitrile [4]. A
major advantage of this route is that it starts from
a heterobifunctional module which is commercially
available. Practical problems, however, are the slow
turnover of thioacetic acid addition and the need for a
suitable photochemical setup which is not ubiquitous.
The present study reports two alternative synthetic
schemes for OEG-AT both of which start from an
ꢀ,!-dibromoalkane. Scheme 1 [21] is a combination
of previously described synthetic steps. In the first
step, 1,11-dibromoundecane was reacted with tri(eth-
ylene glycol), followed by isolation of the asymmetric
ether conjutate [22]. In the second step, the thiol
group was introduced by reaction with thiourea, alka-
line hydrolysis, and acidic extraction [23]. Scheme 2
[24] describes the opposite strategy, i.e. replacement
Fig. 1. Molecular structures of the OEG-terminated alkane-
thiols examined in this study; for convenience, the term HS–
C11–EG3 is used for 20-sulfanyl-3,6,9-trioxa-1-icosanol (1),
HS–C12–EG3 for 21-sulfanyl-3,6,9-trioxa-1-heneicosanol
(2), HS–C12–EG4 for 24-sulfanyl-3,6,9,12-tetraoxa-1-tetra-
cosanol (3), HS–C15–COO–EG3 for 8-hydroxy-3,6-dioxaoc-
tyl 16-sulfanylhexadecanoate (4), HS–C15–CONH–EG8 for
N-(23-hydroxy-3,6,9,12,15,18,21-heptaoxatricosanyl)-16-
sulfanylhexadecanoylamide (5), and HS–C11–CONH–EG6
for N-(17-hydroxy-3,6,9,12,15-pentaoxaheptadecyl)-12-sul-
fanyldodecanoylamide (6)
The purpose of the present study was to examine
a representative variety of OEG-AT (see Fig. 1) with
respect to the following criteria: (i) High protein re-
sistance should be retained over several weeks so that
larger batches of chips can be prepared at once and
used on demand. (ii) SAMs formed at low thiol con-
centration were hoped to show full protein resistance
in order to save precious thiols. Finally, (iii) shorter
OEG chains were much preferred because a typical
application is the parallel study of protein–protein
interaction, both in the ensemble (by surface plasm-
on resonance, SPR) and on the single molecular
Scheme 1

Protein-Resistant Self-Assembled Monolayers
247
measurably reduced [4], in other words, longer
OEG chains obviously provide for more robustness
of protein resistance.
The intention of the present study was to apply
this concept for the preparation of SAMs with ex-
tended lifetime of protein resistance when stored
under water. Storage under water has the advantage
that the SAM is fully protected from atmospheric in-
fluences and is not manipulated between SAM prep-
aration and its use. By analogy with the above cited
study [4], the expected trend was clear: longer OEG
chains should provide for more robustness and long
term stability. Nevertheless it was essential to find
out which particular OEG chain length actually pro-
vides for uncompromised protein resistance on the
time scale of 2 weeks. Moreover, little was known
about the possible influence of the length of the hy-
drophobic tail and of the type of linkage between
OEG chain and alkanethiol.
Scheme 2
of bromine by a protected thiol function in the first
step [25], followed by ether conjugation with tri- or
tetra(ethylene glycol), cleavage of the S-trityl group
and reduction of the disulfidic intermediate.
We therefore selected the different types of OEG-
ATs depicted in Fig. 1 to prepare SAMs the pro-
tein resistance of which was examined by SPR after
1–2 weeks of storage under water. The chips with
the SAM-covered gold surfaces were mounted in a
BIAcore+ X setup and superfused with phosphate-
buffered saline (PBS) at a flow of 10 mm3 minꢂ1.
Although Schemes 1 and 2 start from symmetric
ꢀ,!-dibromoalkanes, the isolation of the initially
generated asymmetric intermediates from unreacted
educt and from disubstituted byproduct was straight-
forward, due to the widely differing physical proper-
ties of the components. Meanwhile, intermediate 9
proved to be a generally useful module because the
stability of the S-trityl group towards strong bases
allowed for extension of the terminal hydroxyl group
with a propionic acid function and introduction of
further coupling=ligand functions [26]. In conclu-
sion, Schemes 1 and 2 represent valuable alternatives
to the conventional synthesis route, depending on
circumstances, preferences, and the possible need
for further functional groups in the basic OEG-AT
structure. Moreover, the OEG-terminated dodeca-
nethiol derivatives 2 and 3 synthesized via the route
in Scheme 2 have not been reported before.
Protein Resistance of OEG-Terminated
SAMs after Prolonged Storage
Fig. 2. SPR measurement of nonspecific protein binding to a
SAM prepared in 20 ꢁM HS–C12–EG3 (2); goat IgG, BSA,
and hen egg lysozyme were each applied twice at 1 mg cmꢂ3
concentration in PBS; in the first sequence, monomeric pro-
teins were applied which had been isolated by gel filtration
on a Superdex+ 200 column; after washing with 0.5% SDS,
two injections of unpurified goat IgG and single injections of
BSA and lysozyme were applied; after a second wash with
0.5% SDS, purified IgG was injected again; the solid and the
dotted line reflect the SPR signal from two different spots on
the same chip which are serially passed by the same solution
with ꢃ1 s time difference
In several previous studies [3–6], SAMs from various
OEG-ATs have been examined for their protein re-
sistance. In case of unitary SAMs, i.e. SAMs with an
OEG chain on every alkanethiol, even EG2 chains
gave perfect protein resistance [4]. In mixed SAMs
consisting of simple alkanethiols and OEG-ATs,
however, longer OEG chains were shown to allow
for a progressively larger molar fraction of the OEG-
lacking alkanethiol before protein resistance was

248
C. D. Hahn et al.
corresponding to 5 and 23 RU in flow cell 1 and 2, as
compared to the initial baseline. Subsequent injec-
tion of unpurified commercial proteins (see previous
paragraph) led to dramatic nonspecific adsorption,
corresponding to þ240 and þ300 RU after the first
injection and another þ120 and þ110 RU after the
second injection of goat IgG in flow cell 1 and 2
(solid and dotted line in Fig. 2). The following injec-
tions of BSA and lysozyme had minor effects on top
of slow continuous IgG desorption. The second in-
jection of 0.5% SDS caused a perfect return of both
traces to the level obtained after the first SDS injec-
tion. In separate experiments it was shown that the
enhanced adsorption of IgG (in comparison to BSA
and lysozyme), as well as the enhanced adsorption of
commercial proteins over gel filtered proteins were
general observations and not caused by the order of
injections shown in Figs. 2 and 3 (data not shown).
The pronounced nonspecific adsorption of proteins
to the EG3-terminated SAM in Fig. 2 is in contrast
with the minimal protein adsorption on the EG4-ter-
minated SAM in Fig. 3. Together, all six injections of
monomeric proteins caused a signal increase of þ11
and þ7 RU in flow cell 1 and 2 (solid and dotted line
in Fig. 3), which was also exactly reversed in the
subsequent washing step with 0.5% SDS. Even more
surprising was the lack of adsorption with unpurified
IgG, BSA, and lysozyme in the next cycle, with the
total signal increases amounting to þ6 and þ9 RU in
the two flow cells. When compared with the signal for
a dense layer of IgG (ꢃ2500 RU) [9], the resonance
angle shift of <10 RU for the EG4-terminated SAM
corresponds to <0.4% monolayer coverage by IgG
and is almost negligible.
The protocol in Fig. 3 was applied to SAMs pre-
pared from thiols 1–6 at 500 ꢁM or at 20 ꢁM thiol
concentration. The most characteristic parameter was
the SPR angle shift observed after the first injection
of monomeric IgG (compare Fig. 2), thus this cri-
terion was used for a comparison of different SAM
types (see Table 1). The striking contrast between
high adsorption on thiol 2 (exemplified in Fig. 2)
and low adsorption on thiol 3 (depicted in Fig. 3)
was confirmed in triplicate experiments (see Table 1).
Since no comparable literature data were available
for thiols 2 (HS–C12–EG3) and 3 (HS–C12–EG4),
it was critical to also test the widely used thiol 1
(HS–C11–EG3) for protein adsorption after storage
of the SAM under water (see Table 1). Under the
testing conditions, thiol 1 showed the same increased
Fig. 3. SPR measurement of nonspecific protein binding to a
SAM prepared in 20 ꢁM HS–C12–EG4 (3); the sequence of
injections was closely similar as in Fig. 2 (see legend within
the figure)
The SPR signals from the two separate flow cells
were simultaneously recorded (solid and dotted lines
in Figs. 2 and 3), thereby monitoring protein adsorp-
tion on two different regions of the same chip. The
two flow cells were serially perfused with the same
solution, the time delay being in the order of 1 s. After
insertion of the chip, the baselines in the two flow
cells showed a minor drift but became constant with-
in 30–60min. Then, 100 mm3 volumes of protein
(1mg cmꢂ3) were injected at the same flow (see Figs.
2 and 3). In a first series, monomeric proteins (goat
IgG, BSA, lysozyme) were applied all of which had
been purified by HPLC gel filtration to eliminate
aggregates (<5% dimers and traces of oligomers, data
not shown). After a washing step with 0.5% SDS,
goat IgG, BSA, and lysozyme were injected which
had not been purified by gel filtration. Finally, another
washing step with 0.5% SDS was performed in order
to check for reversibility of protein adsorption.
Figures 2 and 3 show the nonspecific binding of
proteins to SAMs of HS–C12–EG3 (2) and HS–C12–
EG4 (3), which differ by a single ethylene glycol
unit only. Surprisingly, the nonspecific adsorption
of protein was much more pronounced on the EG3-
terminated SAM (Fig. 2), as compared to the EG4-
terminated SAM (Fig. 3). On the EG3-terminated
SAM, the first injection of monomeric IgG caused
a signal increase of þ83 and þ89 resonance units
(RU) in flow cell 1 and 2 (solid and dotted line in
Fig. 2), albeit further injections of monomeric pro-
teins (IgG, BSA, and lysozyme) caused little addi-
tional binding. An injection of 0.5% SDS removed
most of the adsorbed protein, with residual binding

Protein-Resistant Self-Assembled Monolayers
249
Table 1. Nonspecific adsorption of monomeric goat IgG at
1 mg cmꢂ3 towards SAMs prepared from the OEG-terminated
alkanethiols 1–6 shown in Fig. 1; the thiol concentration was
either 500 ꢁM or 20 ꢁM during SAM formation; the extent of
adsorption is given in RU, one RU corresponding to 0.0001ꢁ
SPR angle shift; for comparison, 25 RU corresponds to ꢃ1%
of maximal coverage by IgG [9]; the data have been averaged
from 3 to 4 chips for each SAM type, each chip yielding two
SPR traces from two different spots on the chip (solid and
dotted line in Figs. 2 and 3)
kit) were used that had been precleaned by sonication
in ethanol according to the manufacturer’s instruc-
tions, then even thiol 4 (HS–C15–CONH–EG8) gave
severe adsorption of IgG, equivalent to full monolayer
coverage (data not shown). Precleaning with piranha
and=or ozone afforded only moderate improvements,
whereas boiling in SC1 gave the satisfactory results
reported above. Unfortunately, SC1 is extremely irri-
tating and potentially toxic. Even the small amount of
fumes released before heating can cause severe irrita-
tion when working for several minutes in front of a
half way closed hood, especially with pre-sentitized
personnel. In case of an accident, repeated inhalation
of cortison for a 24h period in a hospital is indispen-
sable to exclude the development of a potentially fatal
lung edema.
Thiol
Adsorption=RU
500 ꢁM
20 ꢁM
1, HS–C11–EG3
2, HS–C12–EG3
3, HS–C12–EG4
4, HS–C15–COO–EG3
5, HS–C15–CONH–EG8
6, HS–C11–CONH–EG6
115 ꢅ 21
85 ꢅ 28
16 ꢅ 12
69 ꢅ 41
77 ꢅ 26
4 ꢅ 2
a
–
36 ꢅ 8
8 ꢅ 4
12 ꢅ 3
a
–
25 ꢅ 14
a
Not determined
Conclusions
The main intention of this study was to select types
of protein-resistant SAMs on gold that can be pre-
pared in larger batches and then consumed in the
course of 1–2 weeks. Since no relevant data were
available in the literature, a representative variety
of OEG-ATs was examined by this criterion. The re-
sults in Table 1 show that a clear distinction was
found between EG3-terminated SAMs which were
rather unsatisfactory, and SAMs with EG4 or longer
OEG chains which showed high protein resistance.
Interestingly, the most dramatic difference was found
in that pair of thiols which differed by a single eth-
ylene glycol unit only, i.e. thiol 2 and thiol 3.
Although this finding is similar to the trends pre-
viously reported for mixed SAMs of OEG-terminated
and OEG-lacking alkanethiols [4], the dramatic
superiority of EG4- over EG3-terminated SAMs after
1–2 weeks of storage could not have been antici-
pated without explicit testing.
Other findings of practical value were that (i) lon-
ger hydrophobic tails also helped to enhance long
term stability of protein resistance, up to the point
that thiol 3 proved satisfactory, in spite of its short
EG3 chain, (ii) gel-filtered proteins showed much
less adsorption than unpurified commercial proteins,
(iii) repeated challenge of a SAM with protein fol-
lowed by washing with SDS significantly improved
protein resistance, (iv) even at 20 ꢁM thiol concen-
tration SAMs with long term protein resistance could
be formed, and (v) the choice of method for gold
precleaning was most essential. Together with the
level of protein adsorption as thiol 2, from which
follows that the startling difference between thiol 2
(HS–C12–EG3) and thiol 3 (HS–C12–EG4) was sole-
ly due to the single additional ethylene glycol unit in
the latter. The reduced protein repellency of EG3-ter-
minated SAMs after storage is also reflected in the
data for thiol 4 (HS–C15–COO–EG3), taking into
account that the shortness of EG3 was partially com-
pensated for by the long hydrophobic anchor which
may confer higher stability of the SAM over time.
Unexpectedly, thiol 3 (HS–C12–EG4) was not topped
by thiols 5 (HS–C15–CONH–EG8) and 6 (HS–C11–
CONH–EG6), in spite of their longer OEG termini.
Thiol 5 was equal to thiol 3 within experimental er-
ror, while thiol 6 appeared slightly more adsorptive.
A pleasant side effect of the study was the find-
ing that SAMs formed in 20 ꢁM thiol were always
at least as good as those formed at 500 ꢁM (see
Table 1), allowing for minimization of thiol con-
sumption. In addition, it was found that the 1 mM
stock solutions of the thiols maintained full activity
over several months when stored at ꢂ25ꢁC.
Role of the Gold Precleaning Method
In the beginning of this study it was found that an es-
sential precondition for a protein resistant SAM was
precleaning of gold with boiling SC1, i.e. 2ꢄ20min
incubation at ꢃ70ꢁC in a mixture of water, 25%
ammonia, and 30% hydrogen peroxide (5=1=1,
v=v=v) [5]. When commercial BIAcore+ chips (SIA

250
C. D. Hahn et al.
(42.5 mmol) in a few cm3 water and refluxed for an additional
5 h under N2 atmosphere. The mixture was then treated with
HCl in ice-cold water, 300 cm3 CH2Cl2 were added, and the
organic layer was washed three times with water. The solvent
was evaporated and the oil-like residue was recrystallized
from ethanol giving rise to 1.41g 1 (70%). The composition
and purity were verified using 1H NMR, TLC, and mass spec-
trometry. 1H NMR (300MHz, CDCl3): ꢂ ¼ 3.75–3.5 (m, 12H),
3.4 (t, J ¼ 7 Hz, 2H), 2.5 (q, J ¼ 7 Hz, 2H), 1.5–1.1 (m, 19H)
ppm; MS (CI): m=z ¼ 337 (MHþ).
new synthetic routes in Schemes 1 and 2, the above
findings should help to reliably prepare OEG-termi-
nated SAMs for applications which require long term
stability of protein resistance.
Experimental
Water was always taken directly from a Milli-Q50 system.
Analytical grade solvents and materials were used, as long as
they were commercially available. NaH (60%, in mineral oil)
was obtained from Acros Organics. 1,11-Dibromoundecane,
1,12-dibromododecane, potassium t-butoxide, CH3COSK,
t-butyl acrylate, thiourea, tri(ethylene glyol) (for reaction with
1,11-dibromododecane), and trityl mercaptan were purchased
from Aldrich. Sephadex LH-20 was obtained from Amersham.
CH3COOH, CH3CN, CHCl3, CH2Cl2, HCl (37%), and toluene
were purchased from J.T. Baker. Tri(ethylene glycol) (for re-
action with 1,12-dibromododecane) and tetra(ethylene glycol)
were obtained from Fluka. Aqueous NH3 (25%), anhydrous
KH2PO4, H2O2 (30%, aqueous solution), I2, KCl, CH3OH,
ninhydrine, H3PO4, NaCl, NaOH, Na2SO4, TLC plates (silica
60, without fluorescent indicator), and triphenylphosphine
were purchased from Merck (Germany). Azido-EG8 was ob-
tained from Polypure (Norway). Glass substrates (D263 T
Du¨nnglas, 12 mmꢄ12 mmꢄ0.3 mm) for SPR were purchased
S-Trityl S-(12-bromododecyl) sulfide (8, C31H39BrS)
Both 14.35g 1,12-dibromododecane (44 mmol) and 2.76g tri-
tyl mercaptan (10mmol) were dissolved in 300cm3 CH3CN
under Ar atmosphere and 9.12 g K2CO3 (66 mmol) were added.
The mixture was refluxed for 23 h during which time the col-
or turned to light yellow. The solvent was removed by rotary
evaporation, the residue was dissolved in CH2Cl2 (100cm3),
and washed with 1 M HCl (50 cm3), 1 M NaOH (50 cm3), and
brine (100cm3). The organic layer was dried (Na2SO4) and fil-
tered. The filtrate was subjected to rotary evaporation and
purified by repeated recrystallization from n-hexane, yielding
1
3.12g 8 (6.0mmol). H NMR (200MHz, CDCl3): ꢂ ¼ 1.15–
1.50 (m, 18H, –(CH2)9–), 1.87 (tt, JAB ¼ JBC ¼ 6.9 Hz, 2H,
CH2(A)–CH2(B)–CH2(C)–Br), 2.15 (t, J ¼ 7.2 Hz, 2H, CH2–
S), 3.42 (t, J ¼ 6.9Hz, 2H, CH2–Br), 7.21–7.33 (m, 9H, trityl:
H3, H4, H5), 7.43 (d, J ¼ 6.9 Hz, 6H, trityl: H2, H6) ppm.
¨
from Prazisions Glas & Optik GmbH, Iserlohn, Germany. Etha-
nol (analytical grade) for SAM formation was obtained from
Roth (Germany). BSA (no. 775 827, Fraction V, fatty acid free)
was purchased from Roche (Austria). Goat IgG (I-5256), 16-
hydroxyhexadecanoic acid, lysozyme (L-6876), and Na2HPO4
were obtained from Sigma. HS–C15–COO–EG3 (thiol 4) was
available from a previous study [11]. HS–C11–CONH–EG6
(thiol 6) and HS–C15–CONH–EG8 (thiol 5) were synthesized
as described before [10], except that commercial N3–EG8
(Polypure, Norway) was used for the synthesis of thiol 5.
NMR spectra were recorded on a Bruker WM300 spec-
trometer or on a Bruker DPX200 spectrometer at 300MHz
or 200 MHz (as specified) in 5 mm dual 1H7=13C probes. Mass
spectra were measured on a Kratos MS50T spectrometer.
21-Tritylsulfanyl-3,6,9-trioxa-1-heneicosanol
(9, C37H52O4S) and 24-Tritylsulfanyl-3,6,9,12-
tetraoxa-1-tetracosanol (10, C39H56O5S)
Potassium t-butoxide (217mg, 1.93mmol) was suspended in
10cm3 tri(ethylene glycol) or tetra(ethylene glycol) and vig-
orously stirred under Ar for 15 min. After addition of 1012 mg
8 (1.93 mmol) the mixture was gently heated to 90ꢁC and re-
acted at this temperature for 5 h. The mixture was allowed to
cool to r.t., diluted with CHCl3 (30 cm3), washed with water
(2ꢄ40 cm3), dried (Na2SO4), and filtered. After evaporation
of the filtrate the residue was subjected to chromatography on
silica 60 (80 g, 3.5cm ID column) in CHCl3 yielding 672 mg
9 (1.06 mmol) or 721 mg 10 (1.22 mmol), as colorless oils
1
that gradually crystallized. H NMR (9, 200 MHz, CDCl3):
20-Bromo-3,6,9-trioxa-1-icosanol (7, C17H35BrO4)
ꢂ ¼ 1.10–1.50 (m, 18H, –(CH2)9–), 1.59 (m, 2H, CH2–CH2–
CH2–O), 2.14 (t, J ¼ 7.4 Hz, 2H, CH2–S), 3.46 (t, J ¼ 6.7 Hz,
2H, CH2–CH2–CH2–O), 3.61–3.75 (m, 12H, O–CH2–CH2–
O), 7.21–7.32 (m, 9H, trityl: H3, H4, H5), 7.43 (d, J ¼ 6.9 Hz,
Both 4.7 g tri(ethylene glycol) (31 mmol) and 0.41g NaH
(17 mmol) were dissolved in dry DMF and stirred for 30 min.
The resulting solution was treated with 20g 1,11-dibromo-
undecane (6.3mmol) and subsequently stirred for 17h. The
reaction was quenched with CH3OH and the solvent was eva-
porated. The resulting oil was dissolved in 250 cm3 CH2Cl2,
washed four times with water, and dried (MgSO4). Subsequent-
ly, the solvent was evaporated and the residue was purified by
column chromatography (silica 60, ethyl acetate) yielding 2.3 g
7 (60%), as verified by 1H NMR and mass spectrometry [22].
1
6H, trityl: H2, H6) ppm; H NMR (10, 200MHz, CDCl3):
ꢂ ¼ 1.10–1.50 (m, 18H, –(CH2)9–), 1.55 (m, 2H, CH2–
CH2–CH2–O), 2.15 (t, J ¼ 7.2 Hz, 2H, CH2–S), 3.45 (t,
J ¼ 6.7Hz, 2H, CH2–CH2–CH2–O), 3.55–3.75 (m, 16H, O–
CH2–CH2–O), 7.15–7.35 (m, 9H, trityl: H3, H4, H5), 7.45 (d,
J ¼ 6.9Hz, 6H, trityl: H2, H6) ppm.
21,210-Dithiobis(3,6,9-trioxa-1-heneicosanol)
(11, C36H74O8S2) and 24,240-Dithiobis(3,6,9,12-
20-Sulfanyl-3,6,9-trioxa-1-icosanol (HS–C11–EG3,
1, C17H36O4S)
tetraoxa-1-tetracosanol) (12, C40H82O10S2)
The corresponding trityl derivative (9 or 10, 0.5 mmol) was
dissolved in 8 cm3 CH3OH under Ar atmosphere and I2 crys-
Both 2.3 g 7 (6mmol) and 2.36 g thiourea (31 mmol) were
dissolved in ethanol and refluxed for 14 h under N2 atmo-
sphere. The resulting solution was treated with 1.7g NaOH

Protein-Resistant Self-Assembled Monolayers
251
tals (1mmol) were added. The reaction mixture was stirred
at r.t. and the reaction was monitored by TLC (chloroform=
methanol, 90=15, v=v). When no trityl derivative could be
detected anymore, the solution was diluted with 10cm3 CHCl3
and subjected to rotary evaporation. The residue was redis-
solved in CH3OH (2.5cm3) and chromatographed in the same
solvent on Sephadex LH-20 (1.5cmꢄ95cm, at a flow of
0.4 cm3 minꢂ1), yielding 0.2 mmol disulfide as colorless crys-
tals. The product was pure by TLC in ethyl acetate. The Rf of
the product (Rf ¼ 0.06) was very distinct from that of the
byproduct trityl iodide (Rf ¼ 0.40), thus flash chromatography
(e.g. in ethylacetate or CHCl3 with a low percentage of
CH3OH) should be a good alternative to chromatography on
Sephadex LH-20. 1H NMR (11, 200 MHz, CDCl3): ꢂ ¼ 1.23–
1.40 (m, 32H, –(CH2)8–), 1.50–1.75 (m, 8H, CH2–CH2–
CH2–O and CH2–CH2–CH2–S), 2.49 (t, J ¼ 6.1 Hz, 2H,
–OH), 2.69 (t, J ¼ 7.1 Hz, 4H, CH2–S), 3.47 (t, J ¼ 6.9 Hz,
4H, CH2–CH2–CH2–O), 3.55–3.80 (m, 24H, O–CH2–CH2–
O); 1H NMR (12, 200 MHz, CDCl3): ꢂ ¼ 1.20–1.40 (m, 32H,
–(CH2)8–), 1.50–1.75 (m, 8H, CH2–CH2–CH2–O and CH2–
CH2–CH2–S), 2.31 (s, 2H, –OH), 2.69 (t, J ¼ 7.2 Hz, 4H,
CH2–S), 3.45 (t, J ¼ 6.7 Hz, 4H, CH2–CH2–CH2–O), 3.55–
3.80 (m, 32H, O–CH2–CH2–O) ppm.
glass dishes by another two standard cleaning cycles. The
cleaned slides were washed in ethanol (3ꢄ), blown dry with
N2 gas, and coated by thermal evaporation at 10ꢂ4 Pa pressure
with chromium (2.8–3.0 nm at 0.05nm=sec) and gold (41 nm
at 0.6 nm=sec) in a Bal-Tec Med020 system (Baltech AG,
Liechtenstein).
SAM Formation on the Gold Surfaces
Thiol stock solutions (1mM in ethanol) were prepared in glass
vials with screw-thread open top caps and separate PTFE
septa which had been precleaned in boiling SC1 and rinsed
in water. The stock solutions were stored at ꢂ25ꢁC for up to
several months and thawed in desiccators over blue gel before
use. The gold chips were cleaned by two standard cleaning
cycles (boiling SC1 þ 5ꢄ rinsing, boiling SC1 þ 10ꢄ rins-
ing), just as described above for precleaning of the bare glass
slides. The gold chips were rinsed in ethanol (3ꢄ) and im-
mersed for at least 1 min in the same solvent (CH3CN) or
solvent mixture (CH3CN=ethanol, 1=1, v=v) which was sub-
sequently used for SAM formation. After this equilibration, the
chips were immersed in 500 ꢁM or 20 ꢁM thiol solutions
which had been prepared in weighing dishes by diluting an
ethanolic 1 mM stock solution with CH3CN. The glass dishes
were tightly closed and kept under ambient conditions for
36h. The coated chips were rinsed in the same solvent as
previously used during SAM formation (3ꢄ) and then bath-
sonicated in the same solvent in the weighing dish for no
longer than 3 min. The chips were again rinsed in the same
solvent (2ꢄ), followed by 3 rinses in ethanol, and one rinse in
water. Each chip was individually transferred into a 14cm3
Falcon tube that had been filled with water before. The water
was carefully decanted and the tube was slowly filled with
water again. All Falcon tubes of one batch were screw capped
and placed in a water-filled beaker which was placed in a bath
sonicator for no longer than 3 min. Subsequently, the water
was decanted and the Falcon tube was carefully refilled. This
step was repeated, the tubes were completely filled with water
and capped. They were kept in the dark at 4ꢁC for 1–3 weeks.
21-Sulfanyl-3,6,9-trioxa-1-heneicosanol (HS–C12–EG3,
2, C18H38O4S) and 24-Sulfanyl-3,6,9,12-tetraoxa-1-
tetracosanol (HS–C12–EG4, 3, C20H42O5S)
CH2Cl2 (3cm3) and CH3COOH (1.5cm3) were added to the
symmetric disulfide (0.1mmol 11 or 12) and the solution was
stirred under Ar atmosphere. Zinc powder (163mg, 2.5mmol)
was added against the Ar flow and the mixture was stirred
under Ar until all disulfide had been converted into free thiol
according to TLC in CHCl3=CH3OH (90=15, v=v). The sus-
pension was filtered through cotton wool, diluted with 20cm3
CHCl3, washed with 0.5% aqueous HCl (2ꢄ20cm3), dried
(Na2SO4), and filtered. The filtrate was taken to dryness and
dried at 1–10 Pa, yielding 0.18 mmol 2 or 3 as a colorless oil
that was found to be pure by TLC in the above eluent, ex-
1
cept for traces of disulfide. H NMR (2, 200 MHz, CDCl3):
ꢂ ¼ 1.25–1.50 (m, 16H, –(CH2)8–), 1.50–1.70 (m, 4H, CH2–
CH2–CH2–O and CH2–CH2–CH2–S), 2.54 (quartettoid,
J ¼ 7.3 Hz, 2H, CH2–SH), 3.47 (t, J ¼ 6.7 Hz, 2H, CH2–
CH2–CH2–O), 3.50–3.80 (m, 12H, O–CH2–CH2–O); 1H
NMR (3, 200 MHz, CDCl3): ꢂ ¼ 1.20–1.50 (m, 16H,
–(CH2)8–), 1.50–1.70 (m, 4H, CH2–CH2–CH2–O and CH2–
CH2–CH2–S), 2.53 (quartettoid, J ¼ 7.2Hz, 2H, CH2–SH),
3.45 (t, J ¼ 6.7 Hz, 2H, CH2–CH2–CH2–O), 3.50–3.80 (m,
16H, O–CH2–CH2–O) ppm.
Measumement of Protein Adsorption by Surface
Plasmon Resonance
SPR measurements were performed in a BIAcore+ X system
(BIAcore+ AB, Uppsala, Sweden) at 25ꢁC with a flow rate
of 10mm3 minꢂ1. The stored chips were washed with water
(3ꢄ), blown dry with N2 gas and mounted on the chip holder
with double sided adhesive tape. The newly inserted chip
was equilibrated with PBS (140mM NaCl, 2.7 mM KCl,
10mM Na2HPO4, and 1.8 mM KH2PO4, pH¼ 7.3 was usually
obtained without adjustment, otherwise minute amounts of
HCl or NaOH were used to adjust the pH to 7.3) for about 1 h
before measurements of protein adsorption were performed.
All injections had a volume of 100 mm3. The SDS solution
(0.5% in water) was the original ‘‘BIAdesorb solution 1’’. PBS
was used as running buffer, as well as to dissolve the commer-
cially obtained proteins (goat IgG, BSA, and hen egg lyso-
zyme, each at 1 mg cmꢂ3 protein concentration). A fraction of
these proteins had been purified by HPLC gel filtration in PBS
on a HR 10=30 column (10 mm ID, 30cm column length) of
Preparation of Gold Chips for Surface Plasmon Resonance
All glassware and tweezers were precleaned, in a closed and
well ventilated hood (!), by submerging them in or filling them
with SC1 (a 5:1:1 mixture of water, 25% NH3, and 30% H2O2,
v=v=v, see Results section for the high risks of this reagent!),
heating to 70–80ꢁC for 20min, and rinsing with water (5ꢄ).
This standard cleaning cycle (boiling SC1 þ 5ꢄ rinsing with
water) was repeated [11]. Subsequently, the glass substrates
(12 mmꢄ12 mmꢄ0.3 mm) were cleaned in the precleaned

252
C. D. Hahn et al.: Protein-Resistant Self-Assembled Monolayers
Superdex+ 200 (Amersham) at a flow rate of 0.5 cm3 minꢂ1 in
order to remove dimers and higher aggregates. The monomers
of IgG, BSA, and lysozyme eluted as expected for Mr of
150000, 66000, and 14600, as compared to BioRad gel filtra-
tion standard. The monomer fractions from many column
runs were pooled and the protein concentration was deter-
mined from UV-Vis spectrum ("280 ¼ 210000 Mꢂ1 cmꢂ1 for
IgG according to the manufacturer’s declaration, "280 ¼ 44300
and 39300 Mꢂ1 cmꢂ1 for BSA and lysozyme [27]) and diluted
to 1 mg cmꢂ3 with PBS. The typical injection sequence is
shown in Fig. 2 (see legend to Fig. 2). In each injection,
140 mm3 protein solution was pulled into the yellow tip of a
200 mm3 digital pipette, then 10mm3 air were also pulled into
the yellow tip by turning the pipette to the 150 mm3 setpoint.
This little air bubble served to clear the 100 mm3 sample loop
from buffer and thus to ensure an instantaneous rise from 0 to
1.0 mg cmꢂ3 protein concentration in each individual injection
of protein.
[8] Houseman BT, Gawalt ES, Mrksich M (2003) Langmuir
19: 1522
[9] Lahiri J, Isaacs L, Tien J, Whitesides GM (1999) Anal
Chem 71: 777
[10] Svedhem S, Hollander CA, Shi J, Konradsson P,
Liedberg B, Svensson SCT (2001) J Org Chem 66: 4494
[11] Tinazli A, Tang J, Valiokas R, Picuric S, Lata S, Piehler
´
J, Liedberg B, Tampe R (2005) Chem Eur J 11: 5249
[12] Frederix F, Bonroy K, Reekmans G, Laureyn W,
Campitelli A, Abramov MA, Dehaen W, Maes GJ (2004)
Biochem Biophys Methods 58: 67
[13] Saerens D, Frederix F, Reekmans G, Conrath K, Jans K,
Brys L, Huang L, Bosmans E, Maes G, Borghs G,
Muyldermans S (2005) Anal Chem 77: 7547
[14] Houseman BT, Mrksich M (2002) Chem Biol 9: 443
[15] Lahiri J, Ostuni E, Whitesides GM (1999) Langmuir 15:
2055
[16] Lee KB, Park SJ, Mirkin CA, Smith JC, Mrksich M
(2002) Science 295: 1702
[17] Jang CH, Stevens BD, Phillips R, Calter MA, Ducker
WA (2003) Nanoletters 3: 691
Acknowledgements
¨
[18] Gamsjager R, Wimmer B, Kahr H, Tinazli A, Picuric S,
Lata S, Tampe R, Maulet Y, Gruber HJ, Hinterdorfer P,
This work was supported by the Austrian Science Foundation
(projects P15295, N00104, and P15380) and (A.T. and R.T.) by
the Deutsche Forschungsgemeinschaft (DFG, FOR495) and
the Bundesministerium fu¨r Bildung und Forschung (BMBF,
Programm Nanobiotechnologie).
´
Romanin C (2004) Langmuir 20: 5885
¨
[19] Huisman BH, Schonherr H, Huck WTS, Friggeri A, van
Manen HJ, Menozzi E, Vancso GJ, van Veggel FCJM,
Reinhoudt DN (1999) Angew Chem Int Ed 38: 2248
[20] Harder P, Grunze M, Dahint R, Whitesides GM, Laibinis
PE (1998) J Phys Chem B 102: 426
[21] Frederix F (2004) Ph.D. thesis, K. U. Leuven, Leuven,
Belgium
References
[1] Szleifer I (1997) Curr Opin Solid State Mat Sci 2: 337
[2] Love JC, Estroff LA, Kriebel JK, Nuzzo RG, Whitesides
GM (2005) Chem Rev 105: 1103
[22] Flink S (2000) Ph.D. thesis, University of Twente,
Enschede, The Netherlands
[3] Prime KL, Whitesides GM (1991) Science 252: 1164
[4] Prime KL, Whitesides GM (1993) J Am Chem Soc 115:
10714
[23] Bader MM (1996) Phos Sulf Silicon 116: 77
[24] Leitner C, ongoing diploma thesis at University of Linz,
Linz, Austria
[5] Benesch J, Svedhem S, Svensson SCT, Valiokas R,
Liedberg B, Tengvall P (2001) J Biomater Sci Polymer
Edn 12: 581
[6] Ostuni E, Chapman RG, Holmlin RE, Takayama S,
Whitesides GM (2001) Langmuir 17: 5605
[7] Houseman BT, Mrksich M (2001) Biomaterials 22: 943
[25] De Jong MR, Huskens J, Reinhoudt DN (2001) Chem
Eur J 7: 4164
¨
[26] Holzl M, Tinazli A, Leitner C, Hahn CD, Lackner B,
Tampe R, Gruber HJ, manuscript under review at
´
Langmuir
[27] Kondo A, Higashitani K (1992) J Coll Int Sci 150: 344