Immobilization of biomolecules onto surfaces according to ultraviolet light diffraction patterns

$Page 1: Immobilization of biomolecules onto surfaces according to ultraviolet light diffraction patterns$
Immobilization of biomolecules onto surfacesaccording to ultraviolet light diffraction

patterns

Steffen Bjørn Petersen,1,2,* Ane Kold di Gennaro,1 Maria Teresa Neves-Petersen,1

Esben Skovsen,1 and Antonietta Parracino1

1Nanobiotechnology Group, Department of Physics and Nanotechnology, Aalborg University,Skjernvej 4A, Aalborg, Denmark

2The Institute for Lasers, Photonics, and Biophotonics, University at Buffalo,The State University of New York, Buffalo, New York 14260-3000, USA

*Corresponding author: [email protected]

Received 8 July 2010; accepted 15 July 2010;posted 24 August 2010 (Doc. ID 131240); published 24 September 2010

We developed a method for immobilization of biomolecules onto thiol functionalized surfaces according toUV diffraction patterns. UV light-assisted molecular immobilization proceeds through the formation offree, reactive thiol groups that can bind covalently to thiol reactive surfaces. We demonstrate that, byshaping the pattern of the UV light used to induce molecular immobilization, one can control the patternof immobilizedmolecules onto the surface. Using a single-aperture spatial mask, combined with the Four-ier transforming property of a focusing lens, we show that submicrometer (0:7 μm) resolved patterns ofimmobilized prostate-specific antigen biomolecules can be created. If a dual-aperture spatial mask isused, the results differ from the expected Fourier transform pattern of the mask. It appears as a super-position of two diffraction patterns produced by the two apertures, with a fine structured interferencepattern superimposed. © 2010 Optical Society of AmericaOCIS codes: 070.6120, 280.1415, 310.1860, 350.6670.

1. Introduction

Several well-established techniques have alreadybeen developed for the deposition of biomoleculesonto surfaces on the microscale and the nanoscale.Techniques such as pin dispensing and piezoelectricinjectors are routinely used to dispense very smallamounts of fluids in the form of either simple spotsor more complicated microarrays [1–5]. Dropletswith diameters of 1–20 μm can now be depositedupon surfaces in a reproducible manner [6]. Evensmaller structures can be made by using nanolitho-graphic approaches, such as dip-pen nanolitho-graphy, nanoshaving, nanografting, electron-beam

lithography, and nanocontact printing. All thesemethods rely on sequential positioning.

Molecular deposition onto a surface can beachieved through noncovalent adsorption to the sur-face (physisorption) or through covalent attachmentto surfaces chemically modified with, e.g., aldehydesor activated esters, or by using photoactivatablecrosslinkers [7–9]. The different covalent and nonco-valent chemistry approaches to molecular immobili-zation were described in a recent review paper byJonkheijm et al. [10]. Each technique has its ownstrengths and weaknesses with regard to resolution,patterning speed, biocompatibility, complexity, andcost. The new photonic method of UV light-assistedmolecular immobilization (LAMI) presented here is asimple, cheap, and fast alternative for creating pat-terns of biomolecules with submicrometer resolution,which, in contrast with existing methods, is capable

0003-6935/10/285344-07$15.00/0© 2010 Optical Society of America

5344 APPLIED OPTICS / Vol. 49, No. 28 / 1 October 2010

of creating patterns such as microarrays in onesingle step.

The reaction mechanism behind the immobiliza-tion method used in this work involves breakage ofdisulfide bridge(s) mediated upon UV excitation ofaromatic residues in the molecules to be immobi-lized. UV excitation of aromatic residues induceselectron ejection from these residue side chains.The electrons can be captured by nearby disulfidebridges, leading to the formation of the disulfide elec-tron adduct RSSR:- [11]. This radical can dissociateinto free thiol radicals and thiol groups, which cansubsequently bind to thiol reactive surfaces, suchas thiol derivatized glass or gold surfaces [12–16].As UV light is used as an initiator for the immobili-zation process, molecules will only be immobilized inthe areas irradiated by UV light [15]. Thus, by shap-ing the UV light used to initiate the immobilizationprocess, well-controlled patterns of immobilized bio-molecules can be achieved.

One way to shape the UV light is to place a simpleaperture in the path of the UV light beam in order tocreate a diffraction pattern, which can then be fo-cused onto the sample using a focusing lens. Immo-bilization of biomolecules using the UV diffractionpattern created by a straight edge (e.g., a razor bladeblocking half of the beam) has already been demon-strated [12]. The obtained patterns of immobilizedproteins hadmicrometer-sized features. In this work,we will demonstrate that patterns of immobilizedproteins can be made with feature sizes in the sub-micrometer range, which approaches the diffractionlimit for the optical setup we used. We compare anddiscuss the results obtained when using single-aperture masks (pinhole and slits) with the morecomplex results obtained when using two aperturescombined in the same mask, e.g., two pinholes.

According to classic Fourier optics, a transmissionmask placed in the back focal plane of a focusing lenswill result in a light pattern in the front focal planethat is identical to theFourier transformof the spatialmask (this is often referred to as the Fourier trans-formproperties of lenses) [17]. In this paper,we createsubmicrometer resolved patterns of prostate-specificantigen (PSA), a cancermarker, immobilized on a sur-face using a beam of UV light shaped by four simplemillimeter-sized apertures (three single aperturesand a dual-aperture transmission mask) and focusedby a lens. The patterns of immobilized protein corre-spond to theUVdiffractionpatternused to induce bio-molecular immobilization when the single-aperturemasks are used. The use of dual-aperture spatialmasks leads to deviations from the expected Fouriertransform pattern of the mask. It appears as a super-position of individual diffraction patterns producedby each aperture with a fine structured interferencepattern superimposed.

The simplicity and flexibility of our new photonicimmobilization technology allows for engineeringpatterns of covalently immobilized proteins withdiffraction limited resolution.

2. Materials and Methods

A. Optical Setup

A beam of UV light (1mW average power, 280nm)was sent through a computer-controlled shutter andthen expanded by a beam expander consisting of twoquartz lenses. Afterward, the beam passed throughan iris diaphragm before being reflected 90° by amirror. The beam was then focused onto the sampleby an 18mm focal length, 0:5 in: diameter plano–concave quartz lens. The beam expander increasedthe beam diameter enough to slightly overfill theaperture of the focusing lens. One of the four maskswas inserted in the path of the excitation beam justbefore the focusing lens in order to generate the de-sired UV diffraction pattern in the focal plane of thelens (Fig. 1). The geometric center of each mask wasaligned with the center of the UV light beam. A thiolderivatized quartz slide with a thin film of the mole-cules to be immobilized was mounted on a computer-controlled three-dimensional translation stage. Toinitiate biomolecular immobilization, the translationstage moved the sample into position and the shutteropened to expose the sample for a well defined periodof time (an exposure time of 1 s was used for the re-sults presented here). The diffraction limit of our op-tical setup is approximately 0:5 μm, since we workwith a 280nm beam (5mm in diameter) and the fo-cusing lens has a focal length of 18mm.

B. Mask Design

Four simple spatial masks were created: a circularaperture, a short slit, a long slit (relative to the beamdiameter), and a double pinhole mask. To choose theproper dimensions when designing the masks, thecross-sectional intensity profile of the UV beam wasfirst evaluated. The cross-sectional beam profile wasGaussian, and the FWHM was measured by scan-ning a razor blade through the beam and recordingthe power of the unblocked part of the beam as afunction of the position of the edge of the razor. Theestimated FWHM was 4:85mm and, at 90% of themaximum intensity, the beam width was 1:89mm.The four chosen masks were then created by drillingthe desired pattern into a brass plate (12mm×12mm× 0:7mm) with a 1mm diameter drill. The cir-cular aperture was 1mm in diameter. The drilled

Fig. 1. (Color online) Immobilization setup: illumination setupused for light-assisted molecular immobilization using a UV lightdiffraction pattern.

1 October 2010 / Vol. 49, No. 28 / APPLIED OPTICS 5345

slits were 1mm wide, and 3 and 10mm long for thesmall slit and the large slit, respectively, and in bothcases they were rounded on the ends (Fig. 2). Thedouble pinhole mask has two 1mm pinholes with acenter-to-center distance of 3mm (Fig. 7).

C. Thiol Functionalization of Quartz Slides

Optically flat quartz slides were purchased from Ar-rayit microarray technology SuperClean Substrates.The slides were chemically modified by silanization.Prior to silanization, the slides were immersed in sul-furic acid for 1h and then rinsed thoroughly withwater. The clean quartz surface was then hydroxy-lated for 1h into 5 wt./vol. % potassium persulfate(K2S2O8 99%, Acros Organics 20201) in order toincrease the number of OH groups on the surface.To prepare the quartz surface for immobilization, theslides were “SH-activated” by incubating 400 μL0:3 vol:% 3-mercaptopropyl-trimethoxysilane (Merck63800) in m-xylene (99þ%, Acros Organics1808600100) on a horizontally oriented quartz slidefor 30 min at room temperature (20 °C–25 °C). Sub-sequently the surface was rinsed with pure xyleneand thoroughly rinsed with ethanol and deionizedwater. Finally, the slides were dried using com-pressed air.

D. Molecular Labeling

PSA was purchased from Research Diagnostics Inc.and labeled with the primary amine reactive dyeAlexa Fluor555 purchased from Invitrogen Molecu-lar Probes. The labeling was carried out followingInvitrogen molecular probes protocol. The degreeof labeling (DOL) was calculated as the ratio betweenthe concentration of dye molecules and protein con-

centration, measuring the Abs555nm and Abs280nm.15; 000 cm−1 M−1 was used as the molar extinctioncoefficient of AF555 at 555nm. The DOL was 3.7.The correction factor, as described in the Invitrogenprotocol, was applied. The final concentration of la-beled PSA used for immobilization was 1 μM.

E. Light-Assisted Molecular Immobilization

Prior to immobilization, a 1 μL drop of 1 μM PSA-AlexaFluor555 was placed onto the functionalizedslide and dried for 10 min. The thin film from thedried droplet was then exposed to 280nm light(1mW average power) for 1 s through one of the fourmasks. This was repeated 15 times at differentplaces within the same droplet to create an arrayof 5 × 3 immobilized patterns interspaced with100 μm. After UV exposure, the slide was immersedin 1%Mucasol for 1h (changing the Mucasol solutionevery 15 min) in order to wash away noncovalentlybound PSA from the surface.

F. Visualization of Diffraction Patterns byFluorescence Microscopy

Fluorescence microscopy was used to visualize thepatterns of immobilized (fluorescently labeled) PSAmolecules on the surface. AlexaFluor555 fluores-cence was visualized with a microscope (OlympusIX71, inverted optics), using a bandpass filter (U-MWIY2, Olympus, excitation at 545–580nm, andemission from 610nm) through an oil-immersion60× objective (NA ¼ 1:42). Snapshots were recordedwith a CCD camera (DP70, Olympus). Further inter-nal magnification of ×1:6 in the microscope allowedfor a total magnification of ×96. The whole 5 × 3 arrayof identical patterns created with each mask was vi-sualized with a Tecan LS 200 laser scanner (excitingat 532nm).

G. Protein Immunoassay

Immobilized PSAmolecules were incubated with Fabanti-PSA and immunoassays were carried out [18].The results confirmed that, after immobilization,PSA epitopes are undamaged and are recognizedby Fab anti-PSA [18].

H. Image Processing and Analysis

The fluorescence microscope images were processedand analyzed using the home-developed softwarepackage BNIP. The images of the immobilized mole-cular patterns were processed using the followingprocedure. First, the images were base plane cor-rected (i.e., the background signal in the areas be-tween the immobilized patterns was fitted andsubtracted). The resulting image was filtered usinga fast Fourier transform (FFT)-based resolution en-hancement routine.

The experimentally obtained fluorescence image ofthe immobilization pattern shown in Fig. 2(a) wasprocessed to remove some intense specks that woulddisturb simple visual evaluation The image processinvolved a multistep procedure, where the initialstep is identifying the location extent of each of

Fig. 2. Single-aperture masks used (left), false color images of theexpected diffraction patterns obtained after Fourier transformingeach mask (middle), and the experimentally obtained fluorescenceintensity images of immobilized PSA-AF555 molecules (right). (a)Mask with a small circular pinhole aperture. (b) Mask with ashort slit. (c) Mask with a long slit. It is apparent from the imagesthat the patterns of light-induced immobilized biomolecules corre-spond reasonably well with the predicted diffraction patterns usedto expose the film of fluorescently labeled biomolecules.


the specks, using a FFT high-pass filter. It is as-sumed that all specks are much smaller than the im-age features that should be preserved. Subsequently,a binary image was produced from this informationafter multiple morphological dilations were done onthe binary image. The resulting binary image is amap that will highlight all speck locations in the cor-responding original image. The morphological dila-tion will ensure that also the rim of the specks isincluded in the binary map. The complement of thebinary image wasmultiplied with the original image.The resulting product image displays the original im-age, with holes (intensity equal to zero) at the specklocations. This image was exposed to an “in-painting”procedure, which uses the intensity in the vicinity ofthe holes to estimate intensity values to be used tofill the holes. The resulting image has all specks re-moved, and the speck locations are filled with inten-sity extrapolated from the local image region aroundeach of the specks. This process works very well if thespeck size is very small compared to the image de-tails that one would like to retain. In the images dis-played in the current paper, speck size was smallcompared to all relevant feature sizes.

If a circular symmetric pattern is present in an im-age, as is the case in this paper, which presents sev-eral Airy patterns, it can be advantageous to make aspatial transform of the image into a polar form. Todo this, it is necessary to define the geometric centerfor the polar plot, e.g., the center of the Airy pattern.In the polar plot, one axis defines the angle with re-spect to the center and the ordinate axis, and theother defines the radial distance from the center.Summing up the intensity data for a given radial dis-tance for all angles allows for integration over allpoints in an Airy pattern. This results in the optimalsignal to noise in the resulting plot.

I. Simulation of the Generated Diffraction Patterns

MATLAB 7.5.0 was used to simulate the light diffrac-tion patterns created by different masks. The simu-lated diffraction light patterns obtained after Fouriertransforming each mask were compared with thepatterns of immobilized molecules created upon ex-

citation of the PSA molecules with each UV light dif-fraction pattern. A zoom of the central part of eachlight-simulated diffraction pattern is displayed tofacilitate the comparison.

3. Results

A collimated beam of UV light passed through aparticular spatial mask and was focused onto a thinfilm of PSA on a thiol functionalized quartz slide (seeFig. 1). As a result, PSAmolecules exposed by the UVlight pattern were immobilized onto the slide. Thepatterns of immobilized PSA were visualized uponimaging the fluorescence of the immobilized labeledPSA molecules after washing away noncovalentlybound PSA molecules not exposed to UV light. Thethree single-aperture masks are displayed in Fig. 2(left column) and the double-aperture mask is dis-played in Fig. 7 (left column) together with a falsecolor image of the expected diffraction pattern gener-ated in the focal plane of the lens, where the samplewas placed (middle column). The right column inboth Figs. 2 and 7 shows images of the fluorescencepattern emitted by PSA-AF555 molecules immobi-lized using the diffraction pattern generated by eachof the four masks.

Fig. 3. Fluorescence emitted by PSA-AF555 after UV-light-assisted immobilization using a pinhole as the single-aperturemask, together with the integrated fluorescence intensity profilealong the direction indicated by the arrow. The observed distancebetween the central intense peak and the first dip in the airyprofile is ∼6–6:4 μm, in good agreement with theoreticalcalculations.

Fig. 4. Fluorescence emitted by PSA-AF555 after UV-light-assisted immobilization using a small slit as the single-aperturemask, together with the integrated fluorescence intensity profilealong the direction indicated by the arrow.

Fig. 5. Fluorescence emitted by PSA-AF555 after UV-light-assisted immobilization using a large slit as the single-aperturemask together with the integrated fluorescence intensity profilesalong the directions indicated by the arrows. The smallest visibleresolved features are the fringes separated by 0:7 μm, which isclose to the diffraction limit of 0:5 μm for the optics used for theimmobilization.


It can be observed in the Fourier transform imagedisplayed in Fig. 2(a) that a small pinhole placed inthe collimated beam path creates a diffraction pat-tern of concentric rings with the highest intensityat the center and with rings of decaying intensityaround it. It can also be seen that the recorded fluor-escence from labeled PSA molecules immobilizedusing this mask display a central intense fluores-cence signal surrounded by weaker concentric fluor-escent rings, as expected. The fluorescent patterns ofimmobilized protein obtained using the short andlong slits as masks are shown in Figs. 2(b) and 2(c),respectively. Again, it can be observed that the pat-terns of immobilized PSA were very similar to theexpected UV diffraction patterns that induced mole-cular immobilization.

To highlight the characteristics of the three pat-terns of immobilized PSA obtained using the single-aperture masks, Figs. 3–5 display the integratedfluorescence intensity along the directions indicatedby the arrows. In Fig. 3, the radial intensity has beenintegrated over all angles to show the fluorescenceintensity profile as a function of the distance fromthe center of the diffraction pattern. The first dipin the Airy profile is located at ∼6–6:4 μm. In Figs.4 and 5 the fluorescence intensity profiles displaythe fluorescence intensity along the arrows inte-grated over the direction perpendicular to the arrowsin the highlighted area marked with a white box. Thesmallest well-resolved features are observed in Fig. 5(upper right vertical profile). The fringes are sepa-

rated by 0:7 μm, which is close to the diffraction limit(0:5 μm) of the optical setup used for proteinimmobilization.

Figure 6 shows a fluorescence image of a 2 × 2 ar-ray of immobilized patterns of PSA. All four patternshave been immobilized using the same mask (smallslit) to evaluate the reproducibility of our technique.All patterns are observed to be identical. Arrays ob-tained using the other three masks showed the samelevel of reproducibility (data not shown).

Figure 7 displays the dual-aperture (two pinholes)transmission mask used (left), a false color image ofthe expected diffraction pattern obtained after Four-ier transforming the mask (middle), and the experi-mentally obtained fluorescence intensity image ofimmobilized PSA-AF555 molecules (right). The pat-tern of immobilized protein is different from theexpected Fourier transform pattern of the mask. Itappears as a superposition of individual Airy pat-terns produced by each pinhole with a fine structuredinterference pattern superimposed. This interfer-ence pattern is expected from the Fourier transformpattern of the mask. The observed distance betweenthe fringes is ∼800nm (Fig. 8). The Airy pattern ob-served around both circular images agrees with thepredicted pattern for a single pinhole. To highlightthe characteristics of the pattern of immobilizedPSA obtained when using the dual-aperture mask,Fig. 8 displays the integrated fluorescence intensityalong the directions indicated by the arrows. In theregion marked with a circle, the radial intensity hasbeen integrated over all angles to show the fluores-cence intensity profile as a function of the distancefrom the center of the diffraction pattern.

4. Discussion and Conclusions

We have demonstrated that one can obtain sub-micrometer resolved protein patterns using UVlight-assisted immobilization in combination withtransmission masks and a focusing lens. The

Fig. 6. Zoom onto a 2 × 2 array of spots each showing the immo-bilization patterns obtained when carrying out light-assisted im-mobilization using a small slit as mask. Notice the high degree ofreproducibility, which is typical of this technique.

Fig. 7. Dual-aperture (two pinholes) transmission mask used(left), false color image of the expected diffraction pattern obtainedafter Fourier transforming the mask (middle), and the experimen-tally obtained fluorescence intensity image of immobilized PSA-AF555 molecules (right).

Fig. 8. Fluorescence image of the PSA-AF555 protein pattern im-mobilized using the dual-aperture mask, together with the inte-grated intensity profiles along the directions indicated by thearrows. The radial intensity has been integrated over all anglesto show the fluorescence intensity profile as a function of thedistance from the center of the diffraction pattern. The smallestvisible resolved fringes are separated by 0:8 μm.


immobilizedmolecular patterns are observed to be al-most identical to the diffraction patterns used fortheir immobilization when a single-aperture maskis used. The distance between the central intensepeak and the first dip in the Airy profile displayedin Fig. 3 is ∼6:0–6:4 μm. The theoretical expected va-lue is∼6:14 μm, according to 1:22 × λ × L=d. The exci-tation wavelength was 280nm, the distance betweenmask and image plane (L) was 3 cm, and the pinholediameter (d) was 1mm. Furthermore, the patterns ofimmobilizedmoleculeswere shown tobehighly repro-ducible for a given single-aperture mask (Fig. 6).

The density of immobilized molecules is propor-tional to the irradiance of the light used to initiatethe immobilization process. However, at very highor very low intensity areas of the diffraction patternused for immobilization, deviations from this propor-tionality are to be expected. At low intensities, it isvery likely that there will be an intensity thresholdbelow which molecules will not be immobilized [16].On the other hand, if UV light intensity is too high,saturation effects might be observed when there areonly a few vacant binding sites left at the surface. Itis also likely that the fluorophore will bleach in theregions exposed to high UV light intensities, leadingto a decrease of fluorescence yield in those areas.While these factors explain why the immobilized pat-terns (single-aperture cases) can deviate from thediffraction pattern used for the immobilization atthe regions exposed to very high or very low intensity,they do not explain how the patterns can get dis-torted, as they are in Fig. 2(c). These artifacts maybe caused by hot spots in the UV beam, slight mis-alignment of the mask relative to the beam, or thenonnegligible thickness of the mask.

As mentioned above, when using a dual-pinholemask, the pattern of immobilized protein is differentfrom the expected Fourier transform pattern of themask (Fig. 7). It appears as a superposition of twoAiry patterns produced by the pinholes with a finestructured interference pattern superimposed. TheAiry pattern observed around both circular imagesagrees with the predicted pattern for a single pin-hole. One possible reason leading to the two circularapertures is partial loss of coherence between theparts of the laser beam that passes through thetwo apertures. The longitudinal coherence dependson three components: (1) λ2=Δλ, where Δλ is thebandwidth of the laser line, (2) the duration of theshort pulses used, and (3) the presence of detectors,in this case biomolecules. Regarding point 1, we es-timate our bandwidth to be ∼12nm centered at280nm, leading to a coherence length of∼7 μm. Sincethe immobilized patterns form inside a 10–20 μm re-gion and since our image plane is placed perpendicu-lar to the optical axis and is carefully adjusted, we donot expect coherence loss due to a different distancetraveled by the photons that pass the two pinholes.With regard to point 2, coherence can also be lost ifthe 200 fs laser pulse, after passing the two pinholes,does not overlap in time and space at the image

plane. We estimate a longitudinal coherence lengthof 60 μm (velocity of light × pulse duration). Themaximum path length difference we can expect fromthe two-pinhole mask is 27 μm (due to a maximal dis-tance between pinhole edges of 1mm); thus, loss oflongitudinal coherence is not likely. With the qualityof our laser system, we assume that lateral coherenceis also present across the beam. In our experimentwe have one additional feature that can preventthe formation of interference patterns: the presenceof a photon detector, the biomolecules. When aphoton is absorbed by a biomolecule, that photonis no longer available for interference. The photonicexcitation of the molecules leads to chemical changesin the biomolecules and to subsequent bonding to thesurface. Even if the photon is reemitted, that event islikely to take place picoseconds or nanoseconds afterthe initial excitation. We therefore speculate thatmolecular absorption of photons prior to the imageplane effectively abolishes the interference processesbetween photons that pass the two apertures. How-ever, the residual coherence is still sufficient to resultin a detailed interference pattern.

With the masks used in this work, we have demon-strated that it was possible to immobilize patterns ofproteins with submicrometer spatial features in onesingle operation step,making our processmuch fasterthan competing methods. The size of the smallest ob-served features was approximately 0:7 μm, which isclose to the diffraction limit of the optical setup used.In this work, we combined knowledge about light-induced effects in proteins with some simple resultsfromclassical Fourier optics to develop a simplemeth-od to create patterns of immobilized biomoleculeswith high reproducibility on surfaces.We believe thatsuch precise patterning with submicrometer resolu-tion can be applied tomedical diagnostics, biosensors,bioelectronics, and nanotechnology.

The authors thank Hans Nilsson for the produc-tion of the masks.

References1. C. J. Strobl, Z. von Guttenberg, and A. Wixforth, “Nano- and

pico-dispensing of fluids on planar substrates using SAW,”IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 51, 1432–1436(2004).

2. H. B. Hsieh, J. Fitch, D. White, F. Torres, J. Roy, R. Matusiak,B. Krivacic, B. Kowalski, R. Bruce, and S. Elrod, “Ultra-high-throughput microarray generation and liquid dispensingusing multiple disposable piezoelectric ejectors,” J. Biomol.Sceening 9, 85–94 (2004).

3. O. Gutmann, R. Kuehlewein, S. Reinbold, R. Niekrawietz, C.P. Steinert, B. de Heij, R. Zengerle, and M. Daub, “A highlyparallel nanoliter dispenser for microarray fabrication,”Biomed. Microdev. 6, 131–137 (2004).

4. O. Gutmann, R. Kuehlewein, S. Reinbold, R. Niekrawietz, C.P. Steinert, B. de Heij, R. Zengerle, and M. Daub, “Fast andreliable protein microarray production by a new drop-in-droptechnique,” Lab Chip 5, 675–681 (2005).

5. B. R. Ringeisen, P. K. Wu, H. Kim, A. Pique, R. Y. C. Auyeung,H. D. Young, D. B. Chrisey, and D. B. Krizman, “Picoliter-scaleprotein microarrays by laser direct write,” Biotechnol. Prog.18, 1126–1129 (2002).


6. S. Nettikadan, K. Radke, J. Johnson, J. T. Xu, M. Lynch, C.Mosher, and E. Henderson, “Detection and quantification ofprotein biomarkers from fewer than 10 cells,” Mol. Cell.Proteomics 5, 895–901 (2006).

7. P. Angenendt, J. Glokler, Z. Konthur, H. Lehrach, and D. J.Cahill, “3D protein microarrays: performing multiplex immu-noassays on a single chip,” Anal. Chem. 75, 4368–4372 (2003).

8. X. H. Liu, H. K. Wang, J. N. Herron, and G. D. Prestwich,“Photopatterning of antibodies on biosensors,” Bioconjug.Chem. 11, 755–761 (2000).

9. M. C. Rodriguez and G. A. Rivas, “Label-free electrochemicalaptasensor for the detection of lysozyme,” Talanta 78, 212–216 (2009).

10. P. Jonkheijm, D. Weinrich, H. Schroder, C. M. Niemeyer, andH. Waldmann, “Chemical strategies for generating proteinbiochips,” Angew. Chem., Int. Ed. Engl. 47, 9618–9647 (2008).

11. M. T. Neves-Petersen, S. Klitgaard, T. Pascher, E. Skovsen, T.Polivka, A. Yartsev, V. Sundström, and S. B. Petersen, “Flashphotolysis of cutinase: identification and decay kinetics oftransient intermediates formed upon UV excitation of aro-matic residues,” Biophys. J. 97, 211–226 (2009).

12. E. Skovsen, M. T. Neves-Petersen, M. Duroux, A. Kold, L.Duroux, and S. B. Petersen, “Immobilizing biomolecules nearthe diffraction limit,” J. Nanosci. Nanotechnol. 9, 4333–4337(2009).

13. T. Snabe, G. A. Roder, M. T. Neves-Petersen, S. Buus, and S. B.Petersen, “Oriented coupling of major histocompatibilitycomplex (MHC) to sensor surfaces using light assisted immo-bilisation technology,” Biosens. Bioelectron. 21, 1553–1559(2006).

14. M. T. Neves-Petersen, T. Snabe, S. Klitgaard, M. Duroux, andS. B. Petersen, “Photonic activation of disulfide bridgesachieves oriented protein immobilization on biosensor sur-faces,” Protein Sci. 15, 343–351 (2006).

15. M. Duroux, E. Skovsen, M. T. Neves-Petersen, L. Duroux, L.Gurevich, and S. B. Petersen, “Light-induced immobilisationof biomolecules as an attractive alternative to microdropletdispensing-based arraying technologies,” Proteomics 7,3491–3499 (2007).

16. M. T. Neves-Petersen, M. Duroux, E. Skovsen, L. Duroux, andS. B. Petersen, “Printing novel molecular architectures withmicrometer resolution using light,” J. Nanosci. Nanotechnol.9, 3372–3381 (2009).

17. J. W. Goodman, Introduction to Fourier Optics, 3rd ed.(Roberts, 2005), pp. 103–108.

18. A. Parracino, M. T. Neves-Petersen, A. Kold di Gennaro, K.Pettersson, T. Lövgren, and S. B. Petersen, “Arraying prostatespecific antigen PSA and Fab anti-PSA using light assistedmolecular immobilization technology,” Protein Sci. 19,1751–1759 (2010).


Documents

Immobilization of biomolecules onto surfaces according to ultraviolet light diffraction patterns