Imaging the binding ability of proteins immobilized on surfaces with different orientations by using liquid crystals

Article abstract of DOI:10.1021/ja0398565

We report an investigation of the binding ability of a protein immobilized on surfaces with different orientations but in identical interfacial microenvironments. The surfaces present mixed self-assembled monolayers (SAMs) of 11 -[19-carboxymethylhexa(ethylene glycol)]undecyl-1-thiol, 1, and 11-tetra(ethylene glycol) undecyl-1-thiol, 2. Whereas 2 is used to define an interfacial microenvironment that prevents nonspecific adsorption of proteins, 1 was activated by two different schemes to immobilize ribonuclease A (RNase A) in either a preferred orientation or random orientations. The binding of the ribonuclease inhibitor protein (RI) to RNase A on these surfaces was characterized by using ellipsometry and the orientational behavior of liquid crystals. Ellipsometric measurements indicate identical extents of immobilization of RNase A via the two schemes. Following incubation of both surfaces with RI, however, ellipsometric measurements indicate a 4-fold higher binding ability of the RNase A immobilized with a preferred orientation over RNase A immobilized with a random orientation. The higher binding ability of the oriented RNase A over the randomly oriented RNase A was also apparent in the orientational behavior of nematic liquid crystals of 4-cyano-4′- pentylcyanobiphenyl (5CB) overlayed on these surfaces. These results demonstrate that the orientations of proteins covalently immobilized in controlled interfacial microenvironments can influence the binding activities of the immobilized proteins. Results reported in this article also demonstrate that the orientational states of proteins immobilized at surfaces can be distinguished by examining the optical appearances of liquid crystals.

Full text of DOI:10.1021/ja0398565

Published on Web 06/30/2004
Imaging the Binding Ability of Proteins Immobilized on
Surfaces with Different Orientations by Using Liquid Crystals
Yan-Yeung Luk,†,# Matthew L. Tingey, Kimberly A. Dickson,
†
‡
Ronald T. Raines,*,§,‡ and Nicholas L. Abbott*
,†
Contribution from the Departments of Chemical and Biological Engineering, Biochemistry, and
Chemistry, UniVersity of Wisconsin-Madison, Madison, Wisconsin 53706
Received November 28, 2003; E-mail: abott@engr.wisc.edu; raines@biochem.wisc.edu
Abstract: We report an investigation of the binding ability of a protein immobilized on surfaces with different
orientations but in identical interfacial microenvironments. The surfaces present mixed self-assembled
monolayers (SAMs) of 11-[19-carboxymethylhexa(ethylene glycol)]undecyl-1-thiol, 1, and 11-tetra(ethylene
glycol) undecyl-1-thiol, 2. Whereas 2 is used to define an interfacial microenvironment that prevents
nonspecific adsorption of proteins, 1 was activated by two different schemes to immobilize ribonuclease A
(RNase A) in either a preferred orientation or random orientations. The binding of the ribonuclease inhibitor
protein (RI) to RNase A on these surfaces was characterized by using ellipsometry and the orientational
behavior of liquid crystals. Ellipsometric measurements indicate identical extents of immobilization of
RNase A via the two schemes. Following incubation of both surfaces with RI, however, ellipsometric
measurements indicate a 4-fold higher binding ability of the RNase A immobilized with a preferred orientation
over RNase A immobilized with a random orientation. The higher binding ability of the oriented RNase A
over the randomly oriented RNase A was also apparent in the orientational behavior of nematic liquid
crystals of 4-cyano-4′-pentylcyanobiphenyl (5CB) overlayed on these surfaces. These results demonstrate
that the orientations of proteins covalently immobilized in controlled interfacial microenvironments can
influence the binding activities of the immobilized proteins. Results reported in this article also demonstrate
that the orientational states of proteins immobilized at surfaces can be distinguished by examining the
optical appearances of liquid crystals.
within the protein is unknown.5,6 In contrast, when the location
of the binding domain of a protein is known, the oriented
Introduction
The development of surface-based proteomics tools requires
general and facile methods for the immobilization of proteins
on surfaces in known orientations and without disruption of the
protein structure. Because the number of proteins presented on
a surface is small when using microarrayed protein receptors,
maximization of the binding capacity of each immobilized
immobilization of proteins is generally thought to be prefer-
7
,8
able. This notion has motivated several approaches leading
to the oriented immobilization of proteins on surfaces, including
the use of protein A (to bind the Fc region of an antibody),9
,10
1
1
the immobilization of oriented Fab′ fragment of antibodies,
1
2
the use of fusion proteins with oligohistidine or cutinase
protein molecule is central to development of sensitive protein
chips.1,2 Nonspecific adsorption of nontargeted proteins must
1
3
14
domains, as well as the use of affinity contact printing.
A major challenge underlying the use of immobilized proteins
is preserving the binding/catalytic activities of the proteins.
Various factors can lead to loss of activity, including confor-
also be minimized. Many approaches leading to the immobiliza-
tion of proteins, such as via imine formation between the lysine
side chains of a protein and aldehyde-modified glass slides,
result in the decoration of surfaces with proteins without control
of their orientations. The presentation of protein receptors with
multiple orientations can be useful when the binding domain
(
5) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760.
(6) MacBeath, G.; Koehler, A. N.; Schreiber, S. L. J. Am. Chem. Soc. 1999,
21, 7967.
7) Huang, W.; Wang, J. Q.; Bhattacharyya, D.; Bachas, L. G. Anal. Chem.
997, 69, 4601.
3,4
1
(
1
*
To whom correspondence may be addressed: abbott@engr.wisc.edu
(8) Lee, Y. S.; Mrksich, M. Trends Biotechnol. 2002, 20, S14.
(
9) Van Oss, C. J.; Van Regenmortel, M. H. V. Immunochemistry; Dekker:
New York, 1994.
or raines@biochem.wisc.edu.
†
Department of Chemical & Biological Engineering.
Department of Biochemistry.
Department of Chemistry.
Current address: Department of Chemistry, Syracuse University,
‡
(10) Turkova, J. J. Chromatogr., B 1999, 722, 11.
11) Spitznagel, T. M.; Jacobs, J. W.; Clark, D. S. Enzyme Microb. Technol.
993, 15, 916.
12) Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.;
Lan, N.; Jansen, R.; Bidlingmaier, S.; Houfek, T.; Mitchell, T.; Miller, P.;
Dean, R. A.; Gerstein, M.; Snyder, M. Science 2001, 293, 2101.
(13) Hodneland, C. D.; Lee, Y. S.; Min, D. H.; Mrksich, M. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 5048.
(14) Renault, J. P.; Bernard, A.; Juncker, D.; Michel, B.; Bosshard, H. R.;
Delamarche, E. Angew. Chem., Int. Ed. 2002, 41, 2320.
(
§
1
#
(
Syracuse, NY 13244.
(
(
(
1) Kodadek, T. Trends Biochem. Sci. 2002, 27, 295.
2) Kodadek, T. Chem. Biol. 2001, 8, 105.
3) Huang, S. C.; Caldwell, K. D.; Lin, J. N.; Wang, H. K.; Herron, J. N.
Langmuir 1996, 12, 4292.
(4) Nakanishi, K.; Muguruma, H.; Karube, I. Anal. Chem. 1996, 68, 1695.
9024
9
J. AM. CHEM. SOC. 2004, 126, 9024-9032
10.1021/ja0398565 CCC: $27.50 © 2004 American Chemical Society

Using Liquid Crystals To Image Protein Binding Ability
A R T I C L E S
mational changes (partial denaturation) induced by immobiliza-
tion, and immobilization in orientations that bury the active site.
The microenvironment of the interface can also influence the
activity of immobilized proteins. For example, surface charge
density can influence the approach of binding partners from
bulk solution, solvent structure at the interface can affect the
binding activities, and nonspecific protein adsorption or protein-
protein association can occur when the density of the im-
mobilized protein is high.8,11 Hence, lack of control over surface
Figure 1. Schematic illustration of the preparation of the A19C variant of
RNase A. Residue 19 (alanine) is replaced with a cysteine, which is
subsequently protected by reaction with DTNB.
chemistry in general leads to interfacial microenvironments that
8
can compromise the activity of immobilized proteins.
cyt b5. The authors attributed this observation to the dynamic
nature of the cyt c-cyt b5 complex and to similar binding site
Because of these issues, past studies have not provided
unambiguous evidence of the influence of protein orientation
on the binding ability of proteins immobilized on surfaces.
2
0
15-21
accessibility of both orientations of cyt b5.
In this article, we report the results of an experimental system
that permits the unambiguous identification of the influence of
the orientation of an immobilized protein on its binding ability.
To compare the binding ability of proteins on surfaces with
either random or preferred orientation, we designed an experi-
mental system that satisfied three criteria. First, the surfaces
prevented nonspecific protein adsorption, thus leading to the
exclusive immobilization of protein via a covalent linkage.
Second, the areal density of protein molecules on the surface
was designed to be constant, independent of the mode of
immobilization (random or oriented). Third, the system permit-
ted the immobilization of proteins in different orientations but
identical interfacial microenvironment so as to minimize the
influence of confounding factors (e.g., variation in solvent
structure) on the binding ability of the immobilized proteins.
Our studies are based on immobilization of ribonuclease A
For examples, Saavedra and co-workers illustrated the impor-
tance of controlling interfacial microenvironments when devel-
oping approaches for oriented immobilization of proteins.
1
5-19
By immobilizing cyt c with a monomeric cysteine residue on a
thiol-capped silane monolayer, the authors found a broad
distribution of orientations of the heme groups in cyt c due to
1
7
nonspecific adsorption of cyt c on the thiol-capped surface.
Similarly, immobilization of cyt c to a pyridyl disulfide-capped
phospholipid bilayer on a glass substrate via a disulfide bond
afforded 70% oriented cyt c (due to disulfide linkage) and 30%
nonspecifically adsorbed cyt c on these surfaces.18 Interestingly,
a recent study also by Saavedra and co-worker demonstrated
that electrostatically driven adsorption of positively charged cyt
c (horse heart cytochrome c) on a negatively charged (sulfonate)
silane-based SAM produced narrower distribution of orientations
of cyt c than site-directed covalent immobilization (disulfide
(RNase A, 13.7 kDa). Recent studies of RNase A have revealed
formation) of cyt c that is accompanied by nonspecific adsorp-
tion.16 These results demonstrate the importance of controlling
biological functions that include antitumor, immunosuppressive,
22-24
embryotoxic, and aspermatogenic activities.
In this article,
nonspecific adsorption of proteins when covalent attachment is
employed to achieve oriented immobilization of proteins.
Leckband and co-workers have also reported studies of the
influence of interfacial microenvironment on the covalent
we report the binding ability of immobilized RNase A with its
most potent known inhibitor: the ribonuclease inhibitor protein
(
RI, 50 kDa). The Kd for the RI‚RNase A complex has been
-
14
immobilization of proteins.2
0,21
measured to be in the femtomolar range (4.4 × 10
to 6.7 ×
A variant of cytochrome b5
-
14
25,26
1
0
M).
To immobilize RNase A on a surface with
(
T8C) that possessed a monomeric cysteine was covalently
coupled to the maleimide groups of supported lipid bilayers
using a linker that was either neutral or negatively charged) to
controlled orientation and without interfering with the binding
site of RNase A for RI, we prepared a RNase A variant in which
alanine 19 was replaced by a cysteine protected as a mixed
disulfide with 2-nitro-5-thiobenzoic acid (NTB) (Figure 1).2
Alanine 19 was chosen because it is in a solvent-exposed loop
that is not in the RI‚RNase A binding interface.
(
afford covalent immobilization of cyt b5 on the bilayers. The
authors found that the surface coverage of covalently im-
mobilized cyt b5 was significantly higher on the neutral surfaces
7,28
21
than the negatively charged surfaces. This work illustrates how
changes in microenvironment of an interface presenting the same
reactive group can influence the extent of immobilization of
proteins. These authors also compared the binding of cyt c to
We prepared surfaces for immobilization of RNase A by
forming mixed self-assembled monolayers (SAMS) from al-
kanethiols 1 and 2 on evaporated films of gold.
2
0
cyt b5 immobilized with two different orientations. Two
different variants of cyt b5 (T8C and T65C) biotinylated on
installed cysteine residues (8 and 65) were immobilized onto a
monolayer of streptavidin-tethered lipids to afford different
orientations of cyt b5. The authors did not observe a measurable
difference in the binding between cyt c and the two variants of
Whereas the ethylene glycol moieties provide a molecularly
(
22) Raines, R. T. Chem. ReV. 1998, 98, 1045.
(
15) Edmiston, P. L.; Lee, J. E.; Cheng, S. S.; Saavedra, S. S. J. Am. Chem.
(23) Leland, P. A.; Raines, R. T. Chem. Biol. 2001, 8, 405.
(24) Leland, P. A.; Staniszewski, K. E.; Kim, B. M.; Raines, R. T. J. Biol. Chem.
2001, 276, 43095.
Soc. 1997, 119, 560.
(
16) Du, Y. Z.; Saavedra, S. S. Langmuir 2003, 19, 6443.
17) Wood, L. L.; Cheng, S. S.; Edmiston, P. L.; Saavedra, S. S. J. Am. Chem.
Soc. 1997, 119, 571.
(
(25) Lee, F. S.; Shapiro, R.; Vallee, B. L. Biochemistry 1989, 28, 225.
(26) Vicentini, A. M.; Kieffer, B.; Matthies, R.; Meyhack, B.; Hemmings, B.
A.; Stone, S. R.; Hofsteenge, J. Biochemistry 1990, 29, 8827.
(27) Kothandaraman, S.; Hebert, M. C.; Raines, R. T.; Nibert, M. L. Virology
1998, 251, 264.
(28) Sweeney, R. Y.; Kelemen, B. R.; Woycechowsky, K. J.; Raines, R. T.
Anal. Biochem. 2000, 286, 312.
(
(
(
18) Edmiston, P. L.; Saavedra, S. S. Biophys. J. 1998, 74, 999.
19) Edmiston, P. L.; Saavedra, S. S. J. Am. Chem. Soc. 1998, 120, 1665.
20) Yeung, C.; Purves, T.; Kloss, A. A.; Kuhl, T. L.; Sligar, S.; Leckband, D.
Langmuir 1999, 15, 6829.
(21) Yeung, C.; Leckband, D. Langmuir 1997, 13, 6746.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004 9025

A R T I C L E S
Luk et al.
Figure 2. A SAM presenting a mixture of carboxylic acid groups and tetra(ethylene glycol) groups is activated by EDC/NHS to form a monolayer presenting
an NHS ester. RNase A is immobilized via the activated SAMs with either a unique orientation (A) or random orientations (B). (A) The NHS-activated
SAMs are coupled with 2-aminoethanethiol to present thiol groups on the surface. The thiol groups selectively couple with the NTB-protected cysteine
residue of A19C RNase A, forming a covalent disulfide bond to afford RNase A immobilized with a unique orientation. (B) The amino groups of lysine
residues and N-terminus of the wild-type RNase A directly couple to the NHS-activated acid on SAMs to afford immobilized RNase A without a unique
orientation.
defined interface that is known to resist nonspecific adsorption
ellipsometric measurements, however, the orientations of liquid
crystals are sensitive to the structure and organization of
molecules immobilized on surfaces. In this article, we report
experimental results that demonstrate that the orientational states
of proteins immobilized at surfaces can be distinguished by
examining the optical appearances of liquid crystals.
29,30
of proteins,
the carboxylic acids are activated in our scheme
3
8
to facilitate two different modes of protein immobilization. In
one scheme, we use a chemoselective coupling28 between a thiol
3
1,32
group presented by the surface and an activated cysteine
in
RNase A (Figure 2A). In a second scheme, the NHS-activated
SAMs are directly treated with wild-type RNase A. The lysine
residues in the RNase A directly couple with the NHS ester
presented by these surfaces to afford another type of covalent
immobilization of RNase A (Figure 2B). Because the coupling
of the surface thiol groups to the A19C RNase A is selective,
the A19C RNase A immobilized on the thiol-terminated SAMs
should assume a preferred orientation. In contrast, because each
RNase A molecule presents 10 lysine residues and an N-terminal
amino group, immobilization of wild-type RNase A on the NHS-
activated surface should afford RNase A with random orienta-
tions. Although the kinetics of the two different immobilization
schemes (Figure 2A,B) are not designed to be identical, the areal
densities of immobilized protein that result from the two
immobilization schemes are controlled by the areal density of
reactive sites on the surfaces as the reactions are driven to
completion (see below).
Experimental Section
Materials. The glass microscope slides were Fisher’s Finest,
premium grade obtained from Fisher Scientific (Pittsburgh, PA). The
nematic liquid crystal 4-cyano-4′-pentylbiphenyl (5CB), manufactured
by BDH, was purchased from EMD Chemicals (Hawthorne, NY). All
other chemicals were obtained from Aldrich Chemicals (Milwaukee,
WI) and used as received.
Preparation of A19C RNase A and Ribonuclease Inhibitor. An
RNase A variant in which alanine 19 is replaced with a cysteine residue
was produced using a published procedure and then protected by
28
reaction with 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB). The A19C
RNase A was purified by a Mono S cation-exchange FPLC column
(
Pharmacia, Piscataway, NJ) to ensure the purity of final product (data
3
9
not shown). RI was produced using a published procedure. Wild-
type RNase A (type I-A from bovine pancreas) was from Sigma
Chemical (St. Louis, MO) and was used without further purification.
Cleaning of Substrates. Microscope slides were cleaned sequentially
We assess the binding ability of RNase A immobilized with
the schemes shown in Figure 2 by using ellipsometry and the
2 4 2 2
in piranha (70% H SO , 30% H O ) and base solutions (70% KOH,
3
0% H
2
O
2
), using nitrogen to provide agitation (1 h at ∼80 °C).
33-35
orientational behavior of liquid crystals.
Measurements of
Warning: Piranha solution should be handled with extreme caution;
in some circumstances, most probably when it has been mixed with
significant quantities of an oxidizable organic material, it has detonated
unexpectedly. The slides were then rinsed thoroughly in deionized water
(18.2 MΩ-cm), ethanol, and methanol and dried under a stream of N2-
(g). The clean slides were stored in a vacuum oven at 110 °C. All
other glassware was cleaned in piranha solution prior to use.
ellipsometric thickness following each binding event provide
an optical thickness of the layer of protein bound on the
surface.36 This optical thickness is similar to that measured by
37
using surface plasmon reflectometry. The orientational be-
havior of liquid crystals on surfaces can also be used to report
the presence of proteins bound to surfaces.3
3-36
In contrast to
Uniform Deposition of Gold Films. Semi-transparent films of gold
with thicknesses of ∼100 nm were deposited onto glass slides mounted
on rotating planetaries (no preferred direction or angle of incidence)
by using an electron beam evaporator (VES-3000-C manufactured by
Tek-Vac Industries, Brentwood, NY). The rotation of the substrates
on the planetaries ensured that the gold was deposited without a
(
(
(
(
(
29) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164.
30) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714.
31) Drewes, G.; Faulstich, H. Anal. Biochem. 1990, 29, 109.
32) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82, 70.
33) Gupta, V. K.; Skaife, J. J.; Dubrovsky, T. B.; Abbott, N. L. Science 1998,
2
79, 2077.
(
(
(
34) Skaife, J. J.; Abbott, N. L. Langmuir 2000, 16, 3529.
35) Skaife, J. J.; Brake, J. M.; Abbott, N. L. Langmuir 2001, 17, 5448.
36) Luk, Y. Y.; Tingey, M. L.; Hall, D. J.; Israel, B. A.; Murphy, C. J.; Bertics,
P. J.; Abbott, N. L. Langmuir 2003, 19, 1671.
(38) Luk, Y. Y.; Abbott, N. L. Science 2003, 301, 623.
(39) Klink, T. A.; Vicentini, A. M.; Hofsteenge, J.; Raines, R. T. Protein
Expression Purif. 2001, 22, 174.
(37) Mrksich, M.; Sigal, G. B.; Whitesides, G. M. Langmuir 1995, 11, 4383.
9026 J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004

Using Liquid Crystals To Image Protein Binding Ability
A R T I C L E S
Scheme 1
preferred direction of incidence. We refer to these gold films as
AIBN (13 mg, 0.077 mmol) was irradiated in a photochemical reactor
(Rayonet reactor lamp, Southern New England Ultraviolet Co., model
no. RPR-100) for 5 h under in nitrogen (ca. 1 atm). Concentration of
the reaction mixture in vacuo, followed by flash chromatography (30:1
“
uniformly deposited gold films”. A layer of titanium (thickness ∼10
nm) was used to promote adhesion between the glass microscope slide
and the film of gold. The rates of deposition of gold and titanium were
-
7
∼0.2 Å/s. The pressure in the evaporator was less than 5 × 10 Torr
before and during each deposition. The gold source was periodically
cleaned by sequentially immersing it in aqua regia (70% HNO , 30%
CH
2 2
Cl /MeOH) gave compound C as a clear oil (0.84 g, 1.304 mmol,
1
93%). H NMR (250 MHz, CDCl
3
): δ 1.29-1.33 (m, 17H), 1.53-
3
1.58 (m, 2H), 2.00-2.04 (dd, 2H, J ) 7.02, 6.77 Hz), 2.3 (s, 3H), 2.8
(t, 2H), 3.43 (t, 2H, J ) 6.79 Hz), 3.49-3.76 (m, 24H), 4.13 (s, 2H),
4.20 (q, 2H, J ) 7.14 Hz).
HCl) and piranha solutions at 50 °C (30 min in each solution). The
cycle was repeated 3 to 4 times, rinsing between cycles in deionized
water.
1
1-Mercapto-undecyl-hexa(ethylene glycol) Acetic Acid (1). A
Oblique Deposition of Gold Films. Semi-transparent films of gold
with thicknesses of ∼13 nm were deposited onto glass microscope slides
mounted on stationary holders by using the electron beam evaporator
described above. The gold was deposited from a fixed direction of
incidence and a fixed angle of incidence of 40° (measured from the
normal of the surface). We refer to these gold films as “obliquely
deposited gold films”. A layer of titanium (thickness of ∼5 nm) was
used to promote adhesion between the glass and the film of gold.
Synthesis of Alkanethiols 1 and 2. The hexa(ethylene glycol)-
terminated alkanethiol (2) was synthesized using published procedures.40
The carboxylic acid-terminated alkanethiol (1) was synthesized by four
solution of thioacetate C (100 mg, 0.17 mmol) in MeOH (5 mL) and
a solution of NaOH (50 wt %, 1 equiv) were mixed together and stirred
under argon for 5 h at room temperature. Acetic acid was then added
to neutralize the reaction mixture. The residue product was concentrated
1
in vacuo to give thiol 1. H NMR (250 MHz, CDCl
3
): δ 1.29-1.33
(
3
m, 16H), 1.53-1.59 (m, 2H), 2.55 (t, 2H), 3.43 (t, 2H, J ) 6.79 Hz),
.49-3.76 (m, 24H), 4.13 (s, 2H). ESI calcd for [MNa] , 549.3176;
+
found, 549.2.
Preparation of Optical Cells for Surface Imaging Using Liquid
Crystals. Optical cells were fabricated from two films of gold, each
of which supported mixed SAMs that had been incubated with the
4
0-42
steps (Scheme 1) using published procedures.
1-Hexa(ethylene glycol)-1-undecene (A). 11-Bromo-1-undecene
was added to a solution of containing excess hexa(ethylene glycol)
neat, 20 equiv) and NaOH (50 wt %, 1 equiv) over 5 min. After being
proteins of interest.3
3-36
The two gold films (obliquely deposited at
1
4
0° from normal of the surface) were aligned facing each other, with
the direction of deposition of each gold film parallel to the other. The
gold films were then clipped together (binder or bulldog clips) using a
thin film of Mylar or Saran Wrap (nominal thickness ∼13 µm) to space
the two surfaces apart. A drop of 5CB heated into its isotropic phase
(
stirred at 70 °C for 12 h, the solution was mixed with hexane and water
and neutralized with HCl (1 N). The aqueous phase was then extracted
six times with hexane; the combined organic extracts were dried over
(33 °C < T < 45 °C) was then drawn by capillarity into the cavity
4
MgSO , concentrated in vacuo, and purified by flash chromatography
between the two surfaces of the optical cell. The cell was subsequently
cooled to room temperature, and the optical texture was observed with
an Olympus BX-60 polarizing light microscope (Tokyo, Japan) in
transmission mode.
1
(
ethyl acetate) to give olefin A as a clear oil. H NMR (250 MHz,
CDCl ): δ 1.26-1.29 (br s, 12H), 1.53-1.58 (qui, 2H), 2.00-2.04
3
(
2
4
dd, 2H, J ) 7.02, 6.77 Hz), 3.43 (t, 2H, J ) 6.79 Hz), 3.57-3.76 (m,
4H), 4.88-4.96 (m, 2H), 5.74-5.84 (m, 1H). ESI calcd for [MNa] ,
57.3244; found, 457.2.
+
Activation of SAMs and Binding of Proteins on SAMs. The
kinetics of the reactions that underlie the two immobilization schemes
reported in this article are expected to be different. Thus, to achieve
comparable surface area coverages of immobilized proteins via the two
schemes, we used reaction times and reagent concentrations that were
determined to drive each immobilization reaction on the surface to
completion. Procedures for activation of the carboxylic acids on SAMs
to afford NHS esters were adopted and modified from published
procedures.41 Briefly, SAMs formed from solutions of 45% 1 and 55%
Undecenyl 11-Hexa(ethylene glycol) Acetic Acid Ethyl Ester (B).
A solution of 11-hexa(ethylene glycol)-1-undecene A (0.83 g, 1.912
mmol) in dry dichloromethane (20 mL) was stirred at 0 °C for 10 min,
followed by addition of ethyl diazoacetate (0.48 mL, 3.824 mmol) and
BF
3
‚Et O (20 µL, 0.1912 mmol). After the mixture was stirred for 30
2
min at 0 °C, saturated aqueous ammonium chloride (10 mL) was added,
and the reaction mixture was placed in a separatory funnel. The organic
phase was collected, and the aqueous phase was extracted with
dichloromethane (5 × 30 mL). The combined organic phases were dried
2
(mole ratio; total concentration 1 mM in ethanol) were soaked in
solutions of EDC/NHS (200 mM/50 mM in 50/50 H O/DMF) for 40
min. For immobilization of RNase A with preferred orientation (Figure
A), the NHS-activated carboxylic acid-terminated SAMs were directly
2
4
over MgSO and concentrated in vacuo to give yellow oil. Purification
by flash chromatography using gradient elution (1:1 ethyl acetate/hexane
2
f ethyl acetate) gave 783 mg (2.018 mmol, 60%) of ester B as a clear
1
transferred to a solution of 200 mM of 2-aminoethanethiol and 20 mM
of diisopropylethylamine for 20 min. The substrates were then rinsed
with ethanol and dried under a stream of nitrogen and incubated in a
phosphate-buffered saline (PBS) containing A19C RNase A (1 µM,
pH 7.4) for 4 h. For immobilization of RNase A with random
orientations (Figure 2B), the NHS-activated SAMs were rinsed briefly
with PBS containing wild-type RNase A (1 µM, pH 8.5) and then
incubated in the PBS containing wild-type RNase A for 4 h. Binding
of RI to surface-bound RNase A (with either mode of immobilization)
was also carried out by using procedures similar to those previously
published.41 Briefly, immobilized RNase A was incubated in a Tris
oil. H NMR (250 MHz, CDCl
3
): δ 1.29-1.33 (m, 15H), 1.53-1.58
(m, 2H), 2.00-2.04 (dd, 2H, J ) 7.02, 6.77 Hz), 3.43 (t, 2H, J ) 6.79
Hz), 3.49-3.76 (m, 24H), 4.13 (s, 2H), 4.20 (q, 2H, J ) 7.14 Hz),
4
.88-4.96 (m, 2H), 5.74-5.84 (m, 1H).
Thioacetate C. A solution of olefin B (0.73 g, 1.402 mmol) in dry
THF (20 mL) containing thioacetic acid (0.13 mL, 1.96 mmol) and
(
(
(
40) Palegrosdemange, C.; Simon, E. S.; Prime, K. L.; Whitesides, G. M. J.
Am. Chem. Soc. 1991, 113, 12.
41) Lahiri, J.; Isaacs, L.; Grzybowski, B.; Carbeck, J. D.; Whitesides, G. M.
Langmuir 1999, 15, 7186.
42) Houseman, B. T.; Mrksich, M. J. Org. Chem. 1998, 63, 7552.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004 9027

A R T I C L E S
Luk et al.
Table 1. Ellipsometric Thickness (nm) and Wettability of SAMs
Presenting Thiol Groups Before and After Treatment with
Hexadecanethiol (HDT)
%
of1a
thicknessb
∆ thicknessb
2
wet by H O
before treatmentc
after treatment
37
2.4 ( 0.1
2.5 ( 0.1
3.4 ( 0.1
4.0 ( 0.1
yes
yes
no
46
37
1 ( 0.1
1.5 ( 0.1
46
no
a
b
c
P
e
r
c
e
n
t
o
f
a
l
k
a
n
e
t
h
i
o
l
1
i
n
e
t
h
a
n
o
l
i
c
s
o
l
u
t
i
o
n
.
I
n
n
a
n
o
m
e
t
e
r
s
.
Treatment: Activation with NHS/AET and reaction with hexadecanethiol.
buffer containing RI (1 µM, pH 7.4, 2 mM Tris, 10 mM EDTA, 100
mM NaCl, 10 mM DTT) for 30 min.
Ellipsometry. Ellipsometric measurements were performed to
determine the optical thicknesses of SAMs and films of proteins. The
reported optical thickness of each sample is the average of 4 to 5
measurements performed at different locations on each substrate by
using a Rudolph Auto EL ellipsometer (Flanders, NJ) at a wavelength
of 632 nm and an angle of incidence of 70°. The gold substrates used
for ellipsometric measurements were 100 nm in thickness (Au) and
were deposited without a preferred direction or angle of incidence. The
ellipsometric thicknesses of SAMs and immobilized proteins were
estimated by using a three-layer model and by assuming a refractive
index of 1.46 for both the monolayer and protein.
Figure 3. Average ellipsometric thicknesses of SAMs prepared from either
mixtures of alkanethiols 1 and 2 dissolved in ethanol or hexadecanethiol
(white bars), and average ellipsometric thicknesses of these surfaces after
activation of the acid groups of 1 by 3-aminoethanethiol (H2NCH2CH2SH)
and treatment with A19C RNase A (black bars). The composition of the
solution (percent alkanethiol 1) is indicated on the x-axis. HDT denotes
SAMs prepared from hexadecanethiol.
the immobilization of A19C RNase A (1 µM in PBS) on thiol-
terminated SAMs by ellipsometry using incubation times of 30
min, 1 h, and 4 h. We measured the ellipsometric thickness of
captured protein to be constant for incubation times greater than
Results and Discussion
30 min, consistent with a maximization of the extent of reaction
Whereas the coupling of lysine groups of a protein to NHS-
activated carboxylic acid groups on SAMs (Figure 2B) is well-
documented, the reaction between NTB-protected cysteine
groups of a protein and thiol groups of SAMs (Figure 2A) has
not been reported. Therefore, we first evaluated the reactivity
of the thiol groups on surfaces prepared by coupling of
on the surface. To ensure maximal reaction of the thiol-
terminated SAMs, we used an incubation time of 4 h for the
remainder of the experiments reported in this article.
4
1
Figure 3 shows the ellipsometric thicknesses of SAMs
prepared from various mixtures of 1 and 2 before and after
activation of the acid groups with 3-aminoethanethiol (H NCH -
2
2
2
-aminoethanethiol (H2NCH2CH2SH) to NHS-activated car-
CH SH) and treatment with the A19C RNase A. For the purpose
2
boxylic acid groups of SAMs. After treating the NHS-activated
carboxylic acid-terminated SAMs with H2NCH2CH2SH (see
Experimental Section), the SAMs were soaked in a solution of
hexadecanethiol (2 mM) and hydrogen peroxide (∼1µL/3 mL)
in ethanol for 30 min. The changes in ellipsometric thickness
and wettability by water are tabulated in Table 1.
Inspection of Table 1 reveals that treatment of the SAMs with
hexadecanethiol transforms the surface from one wet by water
to one that is not. Furthermore, reaction of SAMs prepared from
of comparison, results obtained with SAMs formed from
hexadecanethiol are also shown. Inspection of Figure 3 reveals
that the ellipsometric thickness of the SAMs increases with the
ratio of alkanethiol 1 to alkanethiol 2 in solution, indicating an
increase in the surface density of the longer alkanethiol 1.
Treatment of the mixed SAMs with A19C RNase A causes an
increase in the measured thickness. The increment of the
ellipsometric thickness after treatment of A19C RNase A also
increases with the percentage of alkanethiol 1, consistent with
a specific coupling between the A19C RNase A and the thiol
groups on SAMs (see below for additional evidence for specific
coupling). We note that the increment in ellipsometric thickness
of RNase A physically adsorbed on a hydrophobic surface, a
SAM prepared from form hexadecanethiol (HDT), is larger than
the increment of RNase A specifically immobilized on mixed
SAMs with low surface density of thiol-terminated alkanethiols
(by using low ratio of alkanethiols 1/2 in the ethanolic solution).
This result supports the proposition that the density of reactive
groups presented by these mixed SAMs controls the amount of
RNase A immobilized on the mixed SAMs.
37% of 1 and 63% of 2 with hexadecanethiol caused the
ellipsometric thickness to increase by 1 ( 0.1 nm, whereas
reaction of SAMs prepared from 46% of 1 and 54% of 2 with
hexadecanethiol caused the ellipsometric thickness to increase
by 1.5 ( 0.1 nm. This increase in ellipsometric thickness of
the monolayer upon reaction with HDT correlates with the
surface density of alkanethiol 1. These results, when combined
with the increase in the hydrophobicity of the SAMs after
treatment with hexadecanethiol, indicate that the thiol groups
on the surface are reactive and the extent of incorporation of
HDT is controlled by the density of reactive groups.
Next, we investigated the immobilization of A19C RNase A
on thiol-terminated SAMs with different surface densities of
alkanethiol 1 (Figure 2A). We note that while a native cysteine
in a protein could be used for site-specific immobilization via
disulfide formation with ligands presented on a surface, the
free thiol group in the protein can cause protein dimerization
by forming disulfide bonds between protein molecules. Hence,
we have used NTB-protected thiol groups as activated cysteines
to prevent dimerization of the protein. The NTB groups on
cysteines are good leaving groups for disulfide exchange with
the free thiol groups presented on the surface. We examined
Next, we investigated the selectivity of the immobilization
of A19C RNase A on thiol-terminated SAMs (Figure 2A).
Figure 4 shows a comparison of ellipsometric thicknesses
resulting from the immobilization of wild-type and A19C RNase
A on thiol-terminated SAMs (Figure 2A) and methyl-terminated
SAMs (CH3(CH2)15SH). Inspection of Figure 4 reveals that
while treatment of SAMs presenting thiol groups with A19C
RNase A caused an increase in the ellipsometric thickness of
0.5 ( 0.1 nm, treatment of wild-type RNase A did not cause a
measurable increase in the ellipsometric thickness of these
interfaces. In contrast, on SAMs formed from hexadecanethiol,
1
7
9028 J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004

Using Liquid Crystals To Image Protein Binding Ability
A R T I C L E S
increase in ellipsometric thickness of only 0.4 ( 0.4 nm. Control
experiments in which the immobilized wild-type RNase A was
treated with either BSA or the RI‚RNase A complex did not
afford any significant increase in ellipsometric thickness of the
interfacial layer (Table 2).
A past study by de Feijter and co-workers has derived the
relationship between the ellipsometric thickness and mass per
4
4
unit area of protein immobilized on surface to be
Γ ) <h >(<n > - n )/(dn/dc)
1
1
2
where Γ is the mass of protein per unit area, <h1> is the average
ellipsometric thickness of the protein layer, <n1> is the average
refractive index of the protein layer (1.46), n2 is the refractive
index of air (1.00), and dn/dc is the concentration dependence
of the refractive index of the protein, which was found to be a
constant of 0.184 mL/g with variation of 0.003 for a range of
proteins.
Figure 4. Selectivity of the immobilization of A19C RNase A and wild-
type RNase A on thiol-terminated SAMs. Average ellipsometric thickness
of mixed SAMs prepared from 1 and 2 and hexadecanethiols (HDT) (white
bars), average ellipsometric thickness of SAMs activated by H2NCH2CH2-
SH and treated with A19C RNase A (black bars) and wild-type RNase A
Using this formalism, we calculated the mass per unit area
of RNase A immobilized with either random or preferred
orientations (Table 2). Because the ellipsometric thickness scales
linearly with the amount of protein per unit surface, the ratio
of the increment in the optical thickness upon binding of RNase
A and RI provides the ratio of the masses of the two proteins
on theses surfaces. Inspection of Table 2 leads us to three
conclusions. First, the specific mass of RNase A immobilized
(
gray bars).
treatment of both wild-type and A19C RNase A led to the same
increase in ellipsometric thickness (0.7 nm). There are no free
thiol groups in the native RNase A. Native RNase A does
contain four disulfide bonds formed from the thiols of its eight
cysteine residues. The disulfide bonds are highly stable and
contribute to the robustness of RNase A.43 The only free thiol
group in the mutant RNase A (A19C RNase A) is due to the
introduction of the cysteine residue at position 19. Therefore,
these results indicate that the immobilization of RNase A on
the thiol-terminated SAMs is highly selective to the presence
of the site-directed incorporation of an activated cysteine group
2
on the surface (1.25 mg/m ) corresponds to an areal density of
∼
5.5 molecules within an area 10 nm × 10 nm. We estimate
native RNase A to have a dimension of approximately 4 nm ×
45
2
nm × 1 nm by previously published crystal structures. This
estimation provides a surface coverage of approximately 44%
of the saturation coverage based on an areal coverage of 5.5
(
i.e., NTB-protected cysteine in A19C RNase A) and that the
2
molecules per 100 nm . Second, when RNase A was im-
surfaces resist nonspecific adsorption of RNase A.
Immobilization of wild-type RNase A on SAMs presenting
NHS-activated carboxylic acids (Figure 2B) was carried out with
procedures similar to that previously published for other
proteins.41 The 10 lysine residues in RNase A are uniformly
distributed along the primary sequence of the protein (appearing
at positions 1, 7, 31, 37, 41, 61, 66, 91, 98, and 104 of the
1
carboxylic acid groups, these lysine residues should cause the
immobilized RNase A to assume an ensemble of up to 10
different orientations on the surface. A 4-h incubation of wild-
type RNase A on NHS-activated SAMs (formed from 46% of
mobilized with a preferred orientation, the ratio of the specific
2
masses of immobilized RI (4.25 mg/m ) and RNase A (1.25
2
mg/m ) corresponds closely to the ratio of the molecular weights
of the two proteins (RI, 50 kD; RNase A, 13kD). This result is
consistent with a one-to-one complex forming on the surface
when RNase A is immobilized with a preferred orientation.
Finally, the results presented in Table 2 suggest that the binding
capacity of the SAM decorated with oriented RNase A is greater
than that of the SAM presenting randomly oriented RNase A.
We further characterized these interfaces by examining the
orientational behavior of liquid crystals on them. Past studies
have demonstrated that the orientations assumed by liquid
24-residue sequence).22 When reacted with NHS-activated
1
and 54% of 2) afforded a layer of protein with the same optical
crystals are sensitive to the presence of proteins on interfaces
thickness (0.5 ( 0.1 nm) as that obtained by immobilization of
A19C RNase A on the thiol-terminated SAMs (Table 2). These
results indicate that the two different schemes of immobilization
reported herein can be used to immobilize the same amount of
wild-type and A19C RNase A on SAMs.
Next, we investigated the binding of RI to RNase A
immobilized on SAMs via the two schemes shown in Figure 2.
Ellipsometric measurements indicated an increase in thickness
of 1.7 ( 0.4 nm upon binding of RI to A19C RNase A
immobilized on the thiol-terminated SAMs. Treatment of
immobilized A19C RNase A with BSA or the RI‚RNase A
complex did not cause a significant increase in ellipsometric
thickness. In contrast, binding of RI to wild-type RNase A
immobilized on the NHS ester-terminated SAM caused an
when the interfaces possess nanometer-scale structure.3
3-36
Here,
we prepared nanostructured surfaces by oblique deposition of
gold films and formed SAMs on these surfaces using procedures
identical to those procedures used on uniformly deposited gold
films, as reported above. These obliquely deposited gold films
are composed of gold grains that possess both in-plane and out-
of-plane crystallographic textures as well as anisotropic topog-
raphy that can be idealized as a corrugation with an amplitude
46,47
of 1-2 nm and a wavelength of 10-40 nm.
The anisotropy
in the structure of the gold film guides the azimuthal orientation
of liquid crystals supported on these surfaces. In addition, when
decorated with SAMs, the orientations assumed by the liquid
(
44) DeFeijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 1759.
45) Wlodawer, A.; Svensson, L. A.; Sjolin, L.; Gilliland, G. L. Biochemistry
1988, 27, 2705.
(
(43) Klink, T. A.; Woycechowsky, K. J.; Taylor, K. M.; Raines, R. T. Eur. J.
Biochem. 2000, 267, 566.
(46) Skaife, J. J.; Abbott, N. L. Chem. Mater. 1999, 11, 612.
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 9029
9

A R T I C L E S
Luk et al.
Table 2. Ellipsometric Thickness (nm) of RNase A Immobilized with Preferred and Random Orientations, and RI Bound to Immobilized
RNase A
mixed SAMsa
wild-type RNase Ab
A19C RNase Ab
RI
BSA
RI‚RNase A
4.7 ( 0.4c
1.7 ( 0.4e
4.25 ( 1
3.5 ( 0.4g
0.4 ( 0.4h
1 ( 1
3 ( 0.1c
0 ( 0.1e
0 ( 0.25
3.1 ( 0.1g
0 ( 0.1h
0 ( 0.25
3.3 ( 0.1
0.3 ( 0.1
c
path A: (oriented)
2.5 ( 0.1
2.5 ( 0.1d
0 ( 0.1
3 ( 0.1d
0.5 ( 0.1
1.25 ( 0.25
e
∆
Th
f
M/A
2.6 ( 0.3
Th
M/A
0 ( 0.25
0.75 ( 0.25
g
path B: (not oriented)
3.1 ( 0.1d
0.5 ( 0.1
1.25 ( 0.25
3.2 ( 0.1
0.1 ( 0.1
h
∆
0.25 ( 0.25
a
b
Mixed SAMs formed from 1 and 2. The mole percentage of 1 in ethanol was 46%. Immobilization of A19C RNase A (1 µM) or wild-type RNase
c
A (1 µM) on mixed SAMs via path A or B (Figure 2). Binding of RI, BSA, or the RI‚RNase A complex to A19C RNase A immobilized on thiol-
terminated SAMs. Increase in ellipsometric thickness due to the binding of wild-type or A19C RNase A to the mixed SAMs. Increase in ellipsometric
d
e
f
2
thickness due to the binding of RI, BSA, or the RI‚RNase A complex to A19C RNase A immobilized on mixed SAMs. M/A (mg/m ) is the calculated
g
mass per unit area of protein on surface. Binding of RI, BSA, or the RI‚RNase A complex to wild-type RNase A immobilized on NHS-activated SAMs.
h
Increase in ellipsometric thickness due to binding of RI, BSA, or the RI‚RNase A complex to wild-type RNase A immobilized on mixed SAMs.
SAMs supported on obliquely deposited gold films. We
measured the optical appearance of a 13-µm-thick film of
nematic liquid crystal 5CB sandwiched between two surfaces,
both of which presented RNase A with either preferred or
random orientations. In the absence of RNase A immobilized
on the SAM, the liquid crystal assumed a uniform optical
appearance (Figure 6A). As the sample was rotated between
crossed polars, the intensity of light transmitted through the
sample was strongly modulated. The uniform optical appearance
and strong modulation of the brightness of the sample is
consistent with a uniform planar alignment of the liquid crystal
Figure 5. (A) Schematic illustration of the uniform alignment of liquid
(orientation parallel to the surface). We determined that the
crystals on surface possessing topography that can be idealized as a
nanometer-scale corrugation. The blue bar indicates the orientation of the
liquid crystal. (B) Protein-binding events on the surface can mask the
topography of the gold film and thus disrupt the uniform orientation of the
liquid crystal.
azimuthal orientation of the planar-aligned liquid crystal was
perpendicular to the direction of the incidence of gold during
3
6
the deposition of the gold films. In past studies, we have
demonstrated that the nanometer-scale topography of obliquely
deposited gold films causes the liquid crystal to assume this
crystal on these gold films depends on the specific functionality
49
azimuthal orientation.
3
8,47-49
and orientation of the terminal groups of the SAMs.
The response of liquid crystal to the presence of proteins
supported on obliquely deposited gold films depends on both
the nanometer-scale topography of the gold films and surface
When proteins are immobilized on SAMs supported on
obliquely deposited gold films, the presence of the proteins on
these surfaces changes the orientation of the liquid crystal
relative to that assumed by the liquid crystal in the absence of
4
4
area coverage of the proteins. Our past studies have demon-
strated that the optical appearance of a liquid crystal becomes
increasingly non-uniform within a sample with increasing
surface area coverage of protein. Figure 6B shows the optical
appearance (crossed polars) of 5CB in contact with the SAM
supporting oriented RNase A, whereas Figure 6C shows the
optical texture of 5CB in contact with randomly oriented RNase
A. Inspection of Figure 6B,C reveals that both samples contain
large areas that have a uniform optical appearance and that upon
rotation of the samples between crossed polars, optical appear-
ance of these regions changes in a manner that is indistinguish-
able from the RNase-free samples shown in Figure 6A. This
result indicates that the liquid crystals assume a largely (but
not completely; see below for a discussion of the domains and
defects in Figure 6B,C) uniform azimuthal orientation in the
presence of RNase A that is the same as that assumed by the
liquid crystal in the absence of RNase A on the surface. We
also note that the presence of the RNase A on the surfaces does
introduce domains and linear defects into the otherwise uniform
orientation of the liquid crystals.
3
3,36
the bound protein.
Past studies have identified two possible
mechanisms by which bound protein can influence the orienta-
tion of liquid crystal supported on obliquely deposited gold.
First, as mentioned above, obliquely deposited gold films
possess a nanometer-scale topography that can orient a liquid
crystal (Figure 5A); protein bound to such a surface can prevent
the uniform alignment of the liquid crystal along the corrugation
of the surface (Figure 5B). Second, because the orientations of
functional groups in SAMs supported on obliquely deposited
49
gold films can also influence the orientations of liquid crystal,
bound proteins can mask these oriented functional groups
preventing interaction between the functional groups and liquid
(
crystal), thus leading to a second mechanism by which protein
can be reported by liquid crystal on these surfaces (Figure 5B).
Both mechanisms predict that protein bound to a SAM formed
on an obliquely deposited gold film will be reported as a loss
in the uniform orientation of the liquid crystal.
In this work, we used gold films prepared by oblique
deposition onto glass slides at an angle of 40° from the normal
of the slides. We first investigated the orientational behavior
of liquid crystals in contact with RNase A immobilized on
In particular, close inspection of Figure 6 reveals differences
in the optical appearance of the liquid crystal in contact with
the surfaces that present oriented RNase A (Figure 6B) as
compared to the surfaces that present randomly oriented RNase
A (Figure 6C). On surfaces that present oriented RNase A
(
47) Follonier, S.; Miller, W. J.; Abbott, N. L.; Knoesen, A. Langmuir 2003,
1
9, 10501.
(Figure 6B), the liquid crystal possesses domains with a well-
(
(
48) Shah, R. R.; Abbott, N. L. J. Am. Chem. Soc. 1999, 121, 11300.
49) Gupta, V. K.; Abbott, N. L. Science 1997, 276, 1533.
defined optical appearance (brightness) that differs from the
9030 J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004

Using Liquid Crystals To Image Protein Binding Ability
A R T I C L E S
Figure 6. Optical images (crossed polarizers) of nematic 5CB sandwiched between nanostructured surfaces that support (A) mixed SAMs formed from 1
and 2, (B) RNase A immobilized in a preferred orientation on mixed SAMs formed from 1 and 2 (Figure 2 path A), and (C) RNase A immobilized in
random orientations on mixed SAMs formed from 1 and 2 (Figure 2, path B). The orientation of the sample relative to the polarizer is shown above the
images of the liquid crystal. The arrow indicates the direction of gold deposition. A schematic illustration of RNase A immobilized with preferred and
random orientations is shown to the right of the images of the liquid crystal.
Figure 7. Use of liquid crystals to detect RI binding to RNase A immobilized on SAMs with different modes of orientation. Optical images (crossed
polarizers) of nematic 5CB sandwiched between nanostructured surfaces that support (A) RI bound to preferentially oriented RNase A and (B) RI bound to
randomly oriented RNase A. A schematic illustration of RI binding to RNase A immobilized with preferred and random orientations is shown to the right
of the images of the liquid crystal.
appearance of the liquid crystal in the absence of RNase A.
The bright domains are encircled by line defects. The coexist-
ence of these two distinct optical domains within the liquid
crystal suggests the existence of two unique orientations of the
liquid crystal with different azimuthal directions on the surfaces
that present oriented RNase A. In contrast, on surfaces that
present randomly oriented RNase A (Figure 6C), we do not
observe distinct orientational domains of liquid crystal, but rather
a single domain containing line defects that do not form closed
loops. Past studies have reported that the orientations of terminal
functional groups of SAMs can influence the orientation liquid
crystals on obliquely deposited gold films;38,48,49 the results
presented in Figure 6 suggest that the orientations of im-
mobilized proteins may also influence the orientations of liquid
crystals supported on them.
the optical appearance of 5CB in contact with RI bound to
oriented RNase A, and Figure 7B shows the optical appearance
of 5CB in contact with RI bound to randomly oriented RNase
A. Inspection of Figure 7A reveals that 5CB assumes a non-
uniform orientation on the surfaces with RI bound to oriented
RNase A: many line defects are observed to be embedded
within the liquid crystal. When compared to Figure 6B, a
diminished modulation of the total intensity of the light
transmitted through the sample was observed when the sample
was rotated between the crossed polarizers. This result suggests
that the amount of RI bound to the oriented RNase A is
sufficient to largely mask the influence of the gold film and
SAM on the orientation of the liquid crystal. In contrast, Figure
7B reveals a more uniform orientation of 5CB supported on
the surface presenting random RNase A that was treated with
RI. The optical appearance of the liquid crystal in Figure 7A,
when compared to Figure 7B, indicates greater binding of RI
to the oriented RNase A than that to the random RNase A on
Next we measured the optical appearance of 5CB in contact
with RI bound to RNase A with preferred and uniform
orientations on these nanostructured surfaces. Figure 7A shows
J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004 9031

A R T I C L E S
Luk et al.
the surface. These differences can be quantified by the standard
deviation of the average luminosity of the optical texture of
the liquid crystal. 36 For instance, the standard deviation of the
luminosity of 5CB on the surface presenting oriented RNase A
treated with RI (Figure 7A) is 75, whereas the surface presenting
randomly oriented RNase A treated with RI caused the standard
deviation of the luminosity of 5CB to be 45 (Figure 7B). We
note here that other indices can be used to quantify the
appearance of images such as those shown in Figure 7A,B. We
do not yet know which index represents the optimal method of
characterizing these images.
We make several observations regarding the results repre-
sented above. First, we note that the measurements of the
ellipsometric thicknesses of RI bound to RNase A as well as
the orientational behavior of liquid crystals are consistent with
greater binding of RI to the surfaces presenting the RNase A
immobilized with a preferred orientation than random orienta-
tion. Because the interfacial microenvironments of the im-
mobilized RNase A were similar on the two surfaces, we
conclude that the greater binding capacity of the oriented RI is
indeed due to orientational effects and not to differences in
interfacial microenvironments.
observations suggest that the RI is reported by the liquid crystal
because it masks the influence of the structure of the underlying
gold film/SAM on the liquid crystal. This conclusion is
consistent with the results of several past studies of the binding
of proteins to SAMs supported on obliquely deposited gold
3
3-36
films.
The structure-sensitive nature of the orientational behavior
of liquid crystals is also apparent in our observations of 5CB
on surfaces presenting RNase A immobilized in preferred and
random orientations. Whereas ellipsometry reports the effective
optical thickness of the immobilized RNase A to be identical,
the orientations assumed by the 5CB on the surfaces with
oriented RNase A were different from the orientations of 5CB
observed on the surfaces with RNase A immobilized with
random orientations. In particular, we note that the liquid crystal
supported on the oriented RNase A assumed two distinct
orientations, whereas the liquid crystal supported on the RNase
A immobilized in random orientations reported the presence of
the RNase A on the surfaces as linear defects within the liquid
crystal. We also note that in a past study of the immobilization
of the cell-signaling protein (MEK-His-tag) on SAMs presenting
nitrilotriacetic acid chelated with Ni(II), the liquid crystal also
showed two distinct orientational domains on a layer of oriented
MEK-His-tag supported on an obliquely deposited gold film.52
These results, when combined, suggest that liquid crystals offer
the basis of tools for reporting the organization of proteins on
surfaces.
We note that a number of past studies have reported the
immobilization of antibodies on surfaces in preferred orienta-
tions so that the binding ability of the antigen-binding (Fab)
50
domains is optimized. These strategies include the use of a
predeposited layer of protein A or protein G that binds the Fc
region of an antibody9
,10
and the use of a predeposited
Conclusion
streptavidin layer to bind antibody (or Fab) that is conjugated
with site-specific biotin.50 When compared to immobilization
schemes that do not control the orientations of immobilized
antibodies (or Fab), oriented antibodies (or Fab) generally (but
not always) exhibit higher binding capacity than antibodies
The principal conclusions of this article are threefold. First,
by comparing two schemes that lead to the immobilization of
RNase A in the same interfacial microenvironment, one which
provides a controlled orientation of RNase A and another that
does not, we conclude that control of the orientation of an
immobilized protein receptor on a surface can have a substantial
effect on the capacity of the surface to bind a protein ligand.
By using surfaces that present the same amount of RNase A,
our results reveal that the binding capacity of RNase A without
a preferred orientation for RI is only 25% of that of RNase A
presented in a controlled orientation. Because the microenvi-
ronment of the RNase A is similar in both immobilization
schemes, we conclude that the binding capacity of the random
RNase A is likely low because the RNase A assumes orienta-
tions in which its binding face is not accessible to RI. Second,
we observed the orientations assumed by nematic 5CB on
surfaces presenting RNase A to depend on the orientation of
the immobilized RNase A. This result suggests that liquid
crystals can be used to probe the orientations of proteins at
surfaces. Third, we demonstrated that the increased binding
capacity achieved by the oriented immobilization of proteins
facilitated the use of liquid crystals to rapidly detect protein-
binding events on surfaces.
50
immobilized with random orientations. It is difficult, however,
to assign this increase in binding capacity to the effects of
orientation because the microenvironments of the surfaces that
support oriented and randomly oriented antibodies are different,
and thus variations in the degree of nonspecific adsorption,
denaturation, and interfacial solvent structure can all contribute
5
1,52
to differences in binding activities (as discussed above).
In our study, we used both ellipsometry and liquid crystals
to report the presence of the RI bound to the oriented RNase
A. Ellipsometry reports the presence of RI as the effective
optical thickness of an equivalent slab of material at the surface
that possesses a refractive index of 1.46. In contrast, the
orientations of liquid crystals at surfaces are sensitive to the
nanometer-scale structure of the surfaces. We observed the
presence of the RI bound to the RNase A to be reported as a
loss of the largely uniform orientation of the liquid crystal that
was observed in the absence of bound RI (Figure 6A). Because
the uniform orientation of the liquid crystal observed in the
absence of bound protein was caused by the structure of the
obliquely deposited gold film and SAM supported on it, our
Acknowledgment. This research was supported by funding
from the Biophotonic Partnership Initiative of NSF (ECS-
0086902), the Center for Nanostructured Interfaces (NSF-DMR
(
(
(
50) Peluso, P.; Wilson, D. S.; Do, D.; Tran, H.; Venkatasubbaiah, M.; Quincy,
D.; Heidecker, B.; Poindexter, K.; Tolani, N.; Phelan, M.; Witte, K.; Jung,
L. S.; Wagner, P.; Nock, S. Anal. Biochem. 2003, 312, 113.
51) Butler, J. E.; Ni, L.; Nessler, R.; Joshi, K. S.; Suter, M.; Rosenberg, B.;
Chang, J.; Brown, W. R.; Cantarero, L. A. J. Immunol. Methods 1992,
0
079983) at the University of Wisconsin-Madison, and Grant
GM 44783 (NIH to R.T.R.).
1
50, 77.
52) The orientation of the liquid crystals (5CB) on oriented MEK-His-tag is
documented in Figure 7 in ref 36.
JA0398565
9032 J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004