вход по аккаунту


Micron-Scale Patterning of Biological Molecules.

код для вставкиСкачать
Micron-Scale Patterning of Biological
David J o h n Pritchard, Hywel Morgan *, and Jonathan
The immobilization procedure is represented by Figure 1 and
involves the modification of a surface with avidin D, followed
by the binding of photobiotin 1 to provide a matrix onto which
molecules can be "written" by means of light and a mask.
Mark Cooper*
Techniques for controlling the architecture of immobilized
biomolecular films on surfaces have a wide range of potential
applications in biosensing, cell guidance, and molecular electronics. Previously, films with a predetermined molecular composition in the r direction have been produced by, for example,
either Langmuir-Blodgett technology or self-assembly.['I To
date. however, it has not been possible to selectively pattern
more than one type of biological molecule on the X , J ~plane of
a surface. although Fodor et al. have synthesized multiple arrays
of different oligopeptides or oligonucleotides on glass surfaces.[" We dcscribe here for the first time a method for patterning two or more functional proteins on a surface whilst minimizing nonspecific binding (NSB). Such a technology is important
in the development of niultianalyte microsensors, and we illustrate a possible application by immobilizing antibodies.
At present, patterning of molecules may be performed by
using photolithographic procedures with lift-off,[31 methods
that are not appropriate for biologically active molecules. Alternatively, photoactivation or photodeactivation of surfacebound coupling molecules has been used for immobilizing proteins on surfaces (e.g., with organosilanes or aryl
a ~ i d e s ) . ''I~ These approaches have not, however, been used
for immobilization of more than one type of functional
molecule.'' - " I
We show in this paper that it is possible to selectively attach
different biological molecules in precise areas of a surface by
using a combination of self-assembly and photactivation. The
method is based on the interaction between streptavidin (or a
deglycosylatcd form of avidin, avidin D) and the analogue 1
(counterion : citrate) of the ligand biotin. which is functionalized
with a photoactive group and is thus referred to as a photobiotin
(obtained frotn Vector Laboratories, Peterborough, UK).
Photobiotin was first developed as a nucleic acid hybridization probe."' and has also been used to label proteins."] It has
a terminal aryl azide group, which is stable in the dark, but
forms a rcactive aryl nitrene group upon exposure to light (340375 nm) with the potential to bind an organic species.[" The
photobiotin molecule used in this study incorporates a spacer
arm to minimize steric hindrance to binding. By combining the
NSB properties of avidin D with the activation of photobiotin,
we are able to produce discrete molecular patterning in the plane
of a molecular film.
_I M Cooper. Dr H. Morgan. D. J. Pritchard
Dcp;irtinent ol' Hcctronics & Electrical Engineering
Uniucraity olGlasgou. G128LT ( U K )
Tclclax I n r code + (41)330-4907
Thi\ w o r k % ; I \ wpported by Biotechnology and Biological Sciences Research
Council (BHSRC) (Grant N o GR:H31967).
i L-
Fig 1. A schematic representation of the immobilization procedure: a ) avidin D ( 0 )
with photobiotin 1 ( I ) immobilized onto a surface: b) exposure of selected areas to
light through a mask results in activation of the photobiotin molecule, and a protein
(e) In the solution is immobilized specifically; c) unbound material is removed by
washing, and the procedure repeated with a second protein (A): d) the entire surface
is exposed to light and unreacted photobiotin IS blocked with casein ( 0 )
Protein can be immobilized onto both gold and SiO, by this
method. In the case of gold, the surface was first modified with
a thiol monolayer, onto which avidin D was bound by means of
a carbodiimide coupling agent. Alternatively, avidin could be
immobilized onto SiO, by using a suitable silanization procedure and a protein coupling reagent. In either case. the avidinmodified surface was incubated in a solution of photobiotin
(Fig. 1 a), and selected areas of this surface were exposed to light,
resulting in the activation of the photobiotin molecule. The aryl
nitrene which was formed reacted with any protein present in
solution and thus immobilized it onto the surface (Fig. 1 b).
Patterns were "written" by exposing the surface to light
through a mask in the presence of a second protein (Fig. 1 c) or,
further, with a sequencr of proteins. After patterning. all unreacted photobiotin groups were neutralized by exposing the
whole surface to light in the presence of a blocking molecule
such as casein (Fig. 1 d). The technique has the advantage that
any protein that is not immobilized can be recovered, unaltered.
Biomolecular patterns were written by these techniques. The
fluorescent micrographs of two proteins, rabbit IgG and rat
IgG, coupled to SiO, are presented in Figure 2. These proteins
were immobilized onto the surface, which was then exposed to
goat anti-rabbit IgG labeled with tetramethylrhodamine isothiocyanate (TRITC) and rabbit anti-rat labeled with IgG fluorescein isothiocyanate (FITC), and the pattern was thus developed. Figure 2 shows that both antibodies are only immobilized
onto the surface where the photobiotin had been exposed to light.
The resolution of the lithographic patterning technique is limit-
Fig. 2. A photograph of SiO, patterned with two fluorescently labeled antibodies
The antibodies were immobilized by the photoactivation procedure described previously. Red fluorescence indicatcs the presence of hound rabbit IgG. and grcen
fluorescence the presence of bound rat IgG. The lines are 10 pn wide separated by
10 pm gaps. Black areas indicate minimal nonspecific binding of antibodies. The
length of the bar represents 10 pm.
ed by diffraction of light. This effect is exaggerated, since the
mask is not in direct contact with the SiO,. but is separated by
a film of buffered protein solution, estimated to be 1 pm thick.
Nevertheless, Figure 3 shows that 1.5 pm lines can be written by
using a TRITC-labeled IgG. It should be noted that in Figure 2
the TRITC lines are less red than in Figure 3 owing to the fact
that a double exposure (TRITC and FITC filters) was needed to
obtain the former. This color change is not due to NSB of the
FITC probe, as it occurs in the absence of the FITC label.
t [Sl
big 4 Responses from two gold electrodes patte~nedwith immobili~edantibod~ec
Further det& see Experimental Procedure
Experimental Procedure
Antibody patterning on a SiO, surface: Avidin D (0.2 mgmL I ) in phosphate
buffered saline (10mM phosphate buffer, 2 . 7 m ~KCI. 137mu NaCI). pH 7.4 (PBS
buffer). was immobilized onto a 1 cm2 disc of silani~edSiO, [3], which was subsequently incubated in bovine serum albumin in PBS (0.2 mgmL ’) at room temperature for 1 h. PBS wasalso used for washing between each step. The wafer was then
incubated in long-arm photobiotin I in PBS (10 FgmL I ) for 20 min in the dark,
and was covered with rabbit IgG (10 pgmL I ) . Light activation of the photobiotin
and coupling of the respective antibodies was performed with a 100 W. high-pressure mercury vapor lamp (185 mm from the surface. irradiance = 9 mWcm ’) for
15 min through a mask and a Hoya SL glass filter, which removed radiation below
300 nm. N o decrease in the activity of the polyclonal IgGs used in this work or of
ALP was observed upon exposure to the filtered light.
The patterning process was then repeated with rat IgG (10 p g m L I ) . During the
course of the light-activation experiment. the temperature of the wafer did not
increase by more than 5°C. In order to “block” unreacted photobiotin sites, the
wafer was immersed in casein in PBS ( I O m g m L ’) and was exposed to light for
15 min. The pattern was then “developed” in TRITC-labeled goat anti-rabbit IgG
in PBS (10 pgmL I ) for 60 minand then in FITC-labeled rabbit anti-rat IgG in PBS
(10 pgmL I ) for 60 min. All IgGs were polyclonal and were obtained from Sigma
Fig. 3 . Photograph ofSiO, patterned with TRITC-labeled IgG. This illustrates the
resolution obtainable with this technique. The lines w e 1.5 pm wide separated by
1.5 pm gaps, and were written with a 1.5 pm mark-spacc mask. The length of the bar
represents 1 Fm.
Nonspecific binding of antibodies to other regions of the SiO,
surface was not detectable by fluorescence microscopy. It was.
however, possible to use a sensitive electrochemical assay for
IgG labeled with an alkaline phosphatase (ALP) to estimate the
size of NSB effects on a modified gold surface (Fig. 4). The total
charge density generated as a result of NSB (trace b) was
0.43 mC crn-’, whilst that for the specific binding reaction
(trace a) was 23.9 mCcm-’; this indicates that NSB is limited to
1.8 % of total binding. Measurement of the extent of NSB with
steady-state currents (minus the background reading in buffer)
gave a similar value (1.9 YO).This value includes cross-reactivity
between different antibodies as well as NSB.
The binding of ALP-labeled IgG directly onto an avidin Dcoated gold electrode was also measured in an electrochemical
ALP assay, and the charge density was 0.17 mCcm-2, equivalent to 0.7% NSB. Such values represent a significant advance
on those presently obtainable and now allow patterning of biological molecules to be carried out at microsensor arrays. The
technique also has the advantage that it may be integrated with
lithographic processes currently in use in device fabrication in
the electronics industry.
Antibody patterning on a gold surface: Gold b a s modified with A-acetyl cysteine.
onto which avidin D was bound by means of a carbodiimide coupling agent [ l o ] .
The method for immobilimtion of IgG was. otherwise. as described previously.
Electrochemical measurements: The gold electrodes (0.2 cm diameter) were lirst
modified in 5mM N-acetyl cysteine (Sigma) iii 10mM phosphate buffer (pH 7.0) for
2 h. and were then wished in water. The A-acetyl cysteine modified electrodes were
incubated in a 1 0 % ( w h ) solution of I-ethyl-3~(3-ditnethylaminopropyl)carbodiimide (EDC) in lOmM phosphate buffer for 2 h at 25 ‘C. and were washed in water.
The EDC-activated electrodes were placed in a solution of avidin D in l O m ~
phosphate buffer (0.2 m g m L I ) for 16 h at 4 ’ C before washing. The electrodes
were then incubated in casein in 10mM phosphate buffer (0.2 mgmL I ) foi- 1 h at
room temperature. Photobiotin 1 was assembled on the avidinated gold. and both
electrodes were patterned with a goat anti-rabbit IgG. One electrode was subsequently exposed to an excess of antigen rabbit IgG (trace a). and the other to rat IgG
(trace b) Finally both electrodes were incubated in the presence of an excess of
ALP-labeled goat anti-rabbit IgG for 120min and were washed. A solution of
lOmM I-naphthyl phosphate in lOOmM tris(hydroxymethy1)aminomethane (Tris)
containing 2 . 7 m ~KCI and 1 3 7 m ~NaCl was added to the electrodes. and after
5 min the electrodes were poised at 320 mV vs. Ag;AgCI. A current was measured for 2 min owing to the electrochemical oxidation of enzymically produced
I-naphthol (which. unlike I-naphthyl phosphate. is electroactive at this potential).
When ALP-labeled goal anti-rat IgG was used in place of ALP-labeled goat antirabbit IgG. then both electrodes exhibited behavior similar to that seen in trace b.
Thisdemonstrates that nonspecific binding ofrat IgG toelectrodescoated with goat
anti-rabbit IgG is minimal.
Received: June 28, 1994
Revised version: September 5 , 1994 [Z7081 IE]
German version: Angew. Chcni. 1995. 107. 84
Keywords: antibodies
[I]P. E. L;iibinis. J. I . Hickman. M. S. Wrighton. G. M. Whitesides. S c i n i w 1989.
245.845 847
[2] S P. A. Fodoi. J. L Rcid. M. C. Pirrung, L. Stryer. A. T. Lu. D. Solas. S < . i e n c ~
1991. 2.5/. 707 773.
131 P. ('iinnolly. 1. M Cooper. G. M . Moores. J. Shen. G. Thompson. h'm~otr~di/ l l J / f l , ~ l1991.
2. IhO-lh3.
[4] C' S. Dulce!. J. H Georgei-. V. Krauthmaer. D. A. Stenger. T. L. Fare. J. M.
Caluert. . S ~ . w m c1991. 2-72. 551 554.
[5] c' R LOW. t.. (;. P. Earley. US-A 4562157. 1985.
[h] S. K . R h a t i i i . J. I.. Teixeira. M. Anderson. L. C. Shriver-Lake. J. M. Calvert.
.I.H. Cicorger. .I. J. Hickman. C. S. Dulcey. P. E. Sclioen. F. S . Ligler. A n d .
B / l ~ ~ l / <1993.
~ l l i 2/18. 197 205.
[7] A (' F:or\ter. J. L. Mclnnes. D. C. Skingle. R. H. Symons. Nu(
iyns. /.<. 745 776.
[XI E. I.>iCcj. b: N CjIRlnt. A t i U / . ~lIJ~hPI77.
1987. 163. 151 - 158
[Y] P. A . S. Sinitli iii A_irk,.vundNirrenr\. R c [ r ~ r t , . i r ~ a n r / U / i / i / ? . ( EE.d .E: V. Scrivcn). ilcadciiiic Press, London. 1984. p. 95.
[lo] S. M .C'oopei-. K (ireenough. C. J. McNeil. J. E/<~<rrou!ia/.
Clrwn. Itrrivfucroi
E/?<I o ~ h r l l / 1993. 347. 267-275.
The Structural and Thermodynamic Basis
for the Formation of Self-Assembled
Peptide Nanotubes* *
L-amino acid residues can adopt a flat ring-shaped conformation in which the backbone amide groups are approximately
perpendicular to the plane of the ring." - 'I Furthermore, it
was postulated that intermolecular hydrogen bonding and ring
stacking interactions of peptide subunits in this conformation
would be energetically favored under appropriate conditions
and would thus produce open-ended, hollow. tubular ensembles. The key structural requirement for producing a multiply
ring-stacked tubular structure was the spatial disposition of the
hydrogen-bond donor and acceptor sites of the backbone on
both faces of the (monomeric) peptide ring structure.[''- "I In
the present design. however, the cyclic peptide subunit is devoid
of hydrogen-bond donation from one Pace of the ring structure
because of selective methylation of backbone amide nitrogen
functionalities. Such a ring-shaped peptide subunit cannot participate in an extended hydrogen-bonded network and is therefore predisposed toward the dimeric, antiparallel stacked, cyclindrical structure-the key fundamental repeating structural
motif of the larger peptide nanotube counterparts. The eightresidue cyclic peptide cyclo[(-L-Phe- D-N-MeAla -)4-] (1) was
designed for the task in hand1*'] (Scheme 1). The sequence and
M. Reza Ghadiri,* Kenji Kobayashi, Juan R. Granja,
R a j K. C h a d h a , a n d D u n c a n E. McRee
Design of molecular objects with predefined shapes and functions has come to the forefront of chemistry and materials research lately." - 14] In particular, considerable effort has been
devoted toward the rational design of tubular structures because
of their potential utility in molecular transport, inclusion chemistry, chemical catalysis, and manufacture of novel optical and
electronic devices.['"241 Here, in the context of a new design,
we provide a detailed analysis of the conformation in solution
and thermodynamic basis for the peptide self-assembly process,
and present the first definitive high-resolution structural model
in support of the recently reported peptide nanotubes and
transmembrane ion channel structure^.['^- "I A cyclic peptide
with a selectively N-methylated backbone has been designed
and shown to self-assemble into a discrete soluble cylindrical
dimer in nonpolar organic solvents. Under appropriate conditions, the peptide also assembles to produce a unique porous
crystalline object having an ordered parallel array of waterfilled channels of 7--8 8, diameter, the interior amphiphilic surface of which has alternating hydrophobic and hydrophilic domains approximately every 11 A. The structural assignments
and thermodynamic characterizations are supported by a variety of H N M R techniques, FT-IR spectroscopy, and high-resolution X-ray crystallography.
Recent designs of peptide nanotubes were based on the central premise that cyclic peptides made up of alternating n- and
[*] Prof. M . R . Ghadiri, Dr. K Kobayashi. Dr. J. R. Gran,ja. Dr. R. K . Chadha,
Dr. 0 . E. McRce
Departiiieiits OF Chemistry and Molecular Biology
The S c r i p p Research Institute
10666 Noi-th Torrey Pinea Road. La Jolla. CA 92307 (USA)
Tclefax: Int code + (619)554-6656
This \cork \c;ic supported by the U.S. Office of Naval Research. We thank
0:H. Hu:iiig and M. A Case for their assistance with NMR spectroscopy, G.
Siii~diikfoi ni'iss spectrometric analysis. G. M. Morris for assistance with
coinputcr gi-aphics irepresentation?. and Professor G. Sheldrick for valuable
sugge\tions reg~rdingthe X-ray crystallographic nnalysis. K . K. acknowledges
thc Ministr? of Education. Science. and Culture of Japan for a postdoctoral
F e l l o ~ a l i i p .J. R. G. thanks NSCORT for a summer fellowship. M . R . G. is a
Sciirle Scholai (1W1 ~ 9 4 and
Alfred P. Sloan Research Fellow (1993-1995).
Scheme 1. A two-dimensional repre5entation of the chemical structure of the
peptide subunit (u and L refers to the amino acid chirality) and the self-assembled
cylindrical ensemble (for clarity only four of the eight CH,l'h groups arc represented).
the choice of amino acid residues were based o n the following
considerations: Molecular modeling indicated that methylation
of backbone amide nitrogen functionalities at all alanine
residues would be sufficient to effectively prevent one face of the
putative peptide ring structure from participating in intermolecular hydrogen-bonding and ring-stacking interactions. The peptide was also designed to have an allowed symmetry for favorable intermolecular packing interactions in the solid state thus
permitting its detailed structural characterization by X-ray crys-
Без категории
Размер файла
743 Кб
scala, micro, biological, molecules, patterning
Пожаловаться на содержимое документа