Photochemical Patterning of Biological Molecules Inside a Glass

Maxim Y. Balakirev,*,†,# Ste´ phanie Porte,†,# Maud Vernaz-Gris,†,§ Michel Berger,‡ Jean-Philippe Arie´ ,§. Brigitte Fouque´ ,† and Fra...
0 downloads 0 Views 285KB Size
Anal. Chem. 2005, 77, 5474-5479

Photochemical Patterning of Biological Molecules Inside a Glass Capillary Maxim Y. Balakirev,*,†,# Ste´phanie Porte,†,# Maud Vernaz-Gris,†,§ Michel Berger,‡ Jean-Philippe Arie´,§ Brigitte Fouque´,† and Franc¸ ois Chatelain†

Laboratoire Biopuces, De´ partement Re´ ponse et Dynamique Cellulaires, and Laboratoire d'Electronique de Technologie de l'Information (LETI), Commissariat a` l'Energie Atomique, 17 rue des Martyrs, 38054 Grenoble, France, and AbAg SA, 4 Rue Pierre Fontaine, 91000 Evry, France

A simple way for photochemical patterning of biological molecules onto the inner wall of fused-silica capillary is described. The method is based on a modification of the inner capillary surface with photoactive benzophenone (BP) derivative. The UV irradiation at 365 nm of the capillary filled with a sample solution results in crosslinking of the solutes to the BP moiety via a stable covalent bond. As a proof of concept, oligonucleotides and proteins were arrayed inside the capillary using an inverted microscope as an irradiation device. We demonstrated that the capillary arrays produced in this way are functional and could be used in different bioassays including DNA hybridization, protein interaction studies, and immunoassays. Having a sensitivity comparable to the fluorophore-based assays in a planar format, the capillary array possesses several advantages including submicroliter sample volume and a short assay time. The capillary format should therefore be considered as a possible alternative to a planar format in a number of low-density array applications such as mutation detection and diagnostic immunoassays. The immobilization of biological molecules on surfaces is a critical step in many bioassays including diagnostic analysis, highthroughput screening, and bioelectronic sensing.1-6 A planar format is usually employed for molecule arrays, where the positional control (i.e., micropatterning) is achieved through robotic or photochemical addressing of the molecules and their subsequent attachment to the surface.7-10 Although very success* Corresponding author. Phone: +33 438782103. Fax: +33 438785917. E-mail: [email protected]. † Laboratoire Biopuces, De´partement Re´ponse et Dynamique Cellulaires, Commissariat a` l'Energie Atomique. ‡ Laboratoire d'Electronique de Technologie de l'Information (LETI), Commissariat a` l'Energie Atomique. § AbAg SA. # Both authors contributed equally to this work. (1) Zammatteo, N.; Hamels, S.; De Longueville, F.; Alexandre, I.; Gala, J. L.; Brasseur, F.; Remacle, J. Biotechnol. Annu. Rev. 2002, 8, 85-101. (2) Russo, G.; Zegar, C.; Giordano, A. Oncogene 2003, 22, 6497-6507. (3) Ratner, D. M.; Adams, E. W.; Su, J.; O’Keefe, B. R.; Mrksich, M.; Seeberger, P. H. Chembiochem 2004, 5, 379-382. (4) Reimer, U.; Reineke, U.; Schneider-Mergener, J. Curr. Opin. Biotechnol. 2002, 13, 315-320. (5) Xiao, Y.; Patolsky, F.; Katz, E.; Hainfeld, J. F.; Willner, I. Science 2003, 299, 1877-1881. (6) Yeo, W. S.; Mrksich, M. Angew. Chem., Int. Ed. 2003, 42, 3121-3124.

5474 Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

ful in various biological applications, these molecule arrays suffer from several drawbacks such as relatively large sample volume, inefficient molecular interactions, and limited sensitivity. Microfluidic capillary format offers several advantages relative to planar format, which include a much smaller sample volume, high surface-to-volume ratio, fast reaction kinetics, potential for automated fluid delivery, and analysis of many samples in parallel.11-14 The fused-silica capillaries are widely used in electrophoretic15 and flow-through analytical systems,16-18 but a closed geometry of the capillary limits significantly the use of conventional immobilization methods for molecule patterning. To our knowledge, the only method reported to date is a coating of the capillary with a photosensitive 2-nitro-5-[11-(trimethoxysilyl)undecyl]oxybenzylmethoxypoly(ethylene glycol) propanoate (NMPEG-silane).19 In this study, surface-bound aldehyde groups generated by patternwise irradiation of NMPEG-silane were used for immobilization of the proteins via Schiff base. A number of alternatives to the glass capillary have also been described, which include the following: (i) patterning inside plastic capillaries based on a microsyringe injection and physical adsorption of the proteins;20,21 (ii) fabrication of elastomeric microfluidic channels that surround the molecule arrays patterned on a planar surface;11-13 (7) Fodor, S. P.; Read, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Solas, D. Science 1991, 251, 767-773. (8) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1763. (9) Okamoto, T.; Suzuki, T.; Yamamoto, N. Nat. Biotechnol. 2000, 18, 438441. (10) Lange, S. A.; Benes, V.; Kern, D. P.; Horber, J. K.; Bernard, A. Anal. Chem. 2004, 76, 1641-1647. (11) Delamarche, E.; Bernard, A.; Schmid, H.; Michel, B.; Biebuyck, H. Science 1997, 276, 779-781. (12) Sia, S. K.; Linder, V.; Parviz, B. A.; Siegel, A.; Whitesides, G. M. Angew. Chem., Int. Ed. 2004, 43, 498-502. (13) Zhan, W.; Seong, G. H.; Crooks, R. M. Anal. Chem. 2002, 74, 4647-4652. (14) Cousino, M. A.; Jarbawi, T. B.; Halsall, H. B.; Heineman, W. R. Anal. Chem. 1997, 69, 544A-549A. (15) Horvath, J.; Dolnik, V. Electrophoresis 2001, 22, 644-655. (16) Holt, D. B.; Kusterbeck, A. W.; Ligler, F. S. Anal. Biochem. 2000, 287, 234-242. (17) Narang, U.; Gauger, P. R.; Kusterbeck, A. W.; Ligler, F. S. Anal. Biochem. 1998, 255, 13-19. (18) Koch, S.; Wolf, H.; Danapel, C.; Feller, K. A. Biosens. Bioelectron. 2000, 14, 779-784. (19) Ligler, F. S.; Breimer, M.; Golden, J. P.; Nivens, D. A.; Dodson, J. P.; Green, T. M.; Haders, D. P.; Sadik, O. A. Anal. Chem. 2002, 74, 713-719. (20) Misiakos, K.; Kakabakos, S. E. Biosens. Bioelectron. 1998, 13, 825-830. (21) Petrou, P. S.; Kakabakos, S. E.; Christofidis, I.; Argitis, P.; Misiakos, K. Biosens. Bioelectron. 2002, 17, 261-268. 10.1021/ac0504619 CCC: $30.25

© 2005 American Chemical Society Published on Web 07/27/2005

and (iii) incorporation of molecule-carrying microbeads inside microchannels.22,23 Here we report a method that allows direct immobilization of biological molecules onto the inner wall of fused-silica capillary via photochemical reaction with surface-bound benzophenone (BP). The BP moiety was chosen because of its remarkable chemical and photochemical robustness; this molecule is more stable in ambient light than other known photoactive compounds and can be repeatedly activated at 330-365-nm wavelengths in aqueous solution without significant loss of activity.24,25 In the presence of organic solutes, the excited triplet species of BP react almost exclusively via C-H bond insertion and form a stable covalent linkage with a target molecule.24,25 The BP-containing bifunctional reagents are widely employed as photochemical crosslinkers for studying the interaction of biological molecules.24 The immobilized BP has also been used for covalent attachment of the polymer films to a solid substrate.26 We show here that the BP moiety can be used for light-directed patterning of oligonucleotides and proteins inside the capillary and that the molecules immobilized in this way retain their biological activity and can be used in various bioassays. EXPERIMENTAL SECTION Materials. Flexible fused-silica capillaries with UV-transparent coating (TSU100375, 100 ( 0.4 µm i.d., 363 ( 10 µm o.d., 15 µm of coating thickness) were obtained from Polymicro Technologies (Phoenix, AZ). All the chemicals were of highest available purity and were used as received: 3-aminopropyl)triethoxysilane (APS, ABCR); benzophenone 4-isothiocyanate (BP-NCS), diisopropylethylamine (DIEA), sulfuric acid, hydrogen peroxide, dimethyl formamide (DMF), and Tween-20 (Sigma-Aldrich); ethanol (CarloErba,). Oligonucleotides 5′-tca agg cac tct tgc cta cgc c-3′ (derived from human K-Ras protooncogene gene), 5′-NH2-GCT TCA GCT TCG TCT CCT TG-3′ (derived from phospholipase 2A gene, P2A), 5′-NH2-CGA ACG ACC TAC ACC GAA CT-3′ (derived from green fluorescent protein gene, EGFP), complementary target oligonucleotides, and Cy3-, biotin- and FITC-labeled analogues were purchased from Eurogentec and MWG Biotech. Proteins used in this work were as follows: biotin-labeled recombinant protein G′ from Streptococcus sp., streptavidin labeled with Cy3 (Amersham) and purified bovine serum albumin fraction V (Sigma); recombinant protein-base (PB) of human adenovirus Ad3 expressed in baculovirus system (a gift from P. Fender);27 and recombinant bacterial listeriolysin O (LLO) of Listeria monocytogenes expressed in Escherichia coli (manuscript in preparation). Whole-molecule goat anti rabbit IgG fluorescein conjugate (IgG-FITC), and rabbit anti-human IgG-FITC were obtained from Sigma; rabbit anti-PB antiserum was from P. Fender; and rabbit anti-LLO antiserum was a gift from J. L. Gaillard. (22) Sato, K.; Tokeshi, M.; Kimura, H.; Kitamori, T. Anal. Chem. 2001, 73, 12131218. (23) Noda, H.; Kohara, Y.; Okano, K.; Kambara, H. Anal. Chem. 2003, 75, 32503255. (24) Turro, N. J. Modern Molecular Photochemistry; University Science Books: Mill Valley, CA, 1991. (25) Dorman, G.; Prestwich, G. D. Biochemistry 1994, 33, 5661-5673. (26) Prucker, O.; Naumann, C. A.; Ru1he, J.; Knoll, W.; Frank, C. W. J. Am. Chem. Soc. 1999, 121, 1, 8766-8770. (27) Fender, P.; Ruigrok, R. W.; Gout, E.; Buffet, S.; Chroboczek, J. Nat. Biotechnol. 1997, 15, 52-56.

Figure 1. Capillary setups. (a) Capillary mounted on the microscope stage, UV light crosses the capillary in the focus zone indicated by circle. (b) Metallic support for three capillaries adopted for slide array scanner.

Capillary Modification. A 1-m-long capillary was filled with freshly prepared piranha solution (7:3 (v/v) of concentrated sulfuric acid and 30% hydrogen peroxide) and incubated for 14 h at room temperature. Caution: piranha solution reacts violently with organic materials and should be handled carefully. The capillary was then flashed consecutively with distilled water (2 mL) and 95% ethanol (2 mL) by using a syringe pump set at 100 µL/min. Next, the capillary was treated with 3% APS solution in 95% ethanol at a constant reagent flow of 5 µL/min for 2 h. After that the capillary was flashed with ethanol (2 mL) to remove a noncovalently bound APS, dried, and cured at 115 °C for 2 h. The resultant capillary was filled with a solution of 200 mM BP-NCS and 50 mM DIEA in dry DMF and incubated for 14 h at room temperature. The constant reagent current of ∼0.1 µL/min was maintained using gravity flow. Finally, the capillary was washed with DMF (2 mL) and ethanol (2 mL) at 50 µL/min, dried, and stored at +4 °C in the dark. Irradiation Setup. The irradiation experiments were performed on an Olympus inverted microscope model IX60, equipped with a 100-W mercury arc lamp and a 365-nm interference filter. The capillary filled with sample was mounted on the microscope stage equipped with a small ruler, and the microscope objective was focused on the inner capillary surface using visible light. The patterning was performed by irradiation of the focus zone with microscope UV source at 2 mW/mm2 followed by manual translation of the stage (Figure 1a). Oligonucleotide Patterning and Hybridization. For immobilization of oligonucleotides on the BP-modified surface, the capillary was filled with a solution of oligonucleotide in SSC buffer (150 mM NaCl, 15 mM sodium citrate, pH 7.0) and a series of irradiations was performed by varying the exposure time from 1 to 120 s. Each irradiation time corresponded to a separate spot on the capillary array. Then the capillary was washed with SSC for 30 min at 100 µL/min and analyzed as described below. The surface attachment of the oligonucleotides into the capillary was shown through the use of a Cy3-labeled EGFP oligonucleotide with a confocal microscope (Leica DM IRE2). The detection of biotinylated oligonucleotides bound to the inner capillary surface Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5475

was performed after a 30-min treatment with a solution of 20 µg/ mL streptavidin-Cy3 in PBS at 3 µL/min followed by a 30-min PBS washing at 50 µL/min. In hybridization assays, the capillary was pretreated with 2 mL of washing buffer (2× SSC and 0.1% SDS) at a 100 µL/min flow rate and then filled with a solution of the complementary oligonucleotides in hybridization buffer (PBS, 0.5 M NaCl, 20 mM EDTA, 100 µg/mL salmon sperm DNA). The hybridization was performed at 42 °C for 30 min, and the capillary was washed with 2 mL of the washing buffer as described before. Protein Patterning and Interaction Assay. Recombinant protein G′ was used to address the efficiency of the protein grafting in the BP capillary. The capillary was filled with a solution of 2 mg/mL biotinylated protein G′ in PBS, and the series of irradiations was performed. The capillary was washed with PBS for 30 min at 50 µL/min and then treated with a solution of 20 µg/mL streptavidin-Cy3 or 5 µg/mL IgG-FITC in PBS for 30 min at 3 µL/min. Before scanning, the capillary was washed with PBS at 50 µL/min for 30 min. Concerning IgG-FITC, the signal-to-noise ratio could be increased by adding 1% BSA and 0.025% Tween to the washing buffer (BSA/Tween/PBS). For immunoassay, both PB and LLO antigens were arrayed in parallel at 0.2 mg/mL in four capillaries and the resulted arrays were washed with BSA/Tween/PBS for 15 min. The capillaries were then treated with anti-PB (1:10 000 dilution) or anti-LLO (1: 2000) antisera in BSA/Tween/PBS for 30 min at 3 µL/min, washed with BSA/Tween/PBS for 15 min at 50 µL/min, and treated with 5 µg/mL anti-IgG-FITC in BSA/Tween/PBS for 30 min at 3 µL/min. Finally, the capillaries were washed with BSA/ Tween/PBS for 15 min and then with PBS for 15 min at 50 µL/ min. Data Acquisition. The capillaries filled with corresponding solutions were mounted on the metallic support of a slide format (Figure 1b), and the capillary images were obtained using a fourcolor microarray scanner GeneTAC LS-IV (Genomic Solutions). The data were analyzed using the scanner software. RESULTS AND DISCUSSION Irradiation and Signal Acquisition Setups. We used standard flexible fused-silica capillaries with UV-transparent coating for array construction. The capillaries had an outer diameter of 363 µm and an internal diameter of 100 µm, providing a filling capacity of ∼76 nL/cm. The irradiation was performed on an inverted microscope equipped with a 100-W mercury arc lamp and a 365-nm interference filter. The capillary filled with sample was mounted on the microscope stage, and the patterning was performed by successive irradiation of the focus zone and lateral translation of the stage (Figure 1a). The length of the zone was determined by a slit aperture and was fixed at 160 µm. The resultant light intensity at the capillary level was measured to be ∼2mW/mm2. Fluorescent images of the capillaries were obtained using a four-color microarray scanner equipped with a microscope slide holder. To scan the capillaries, the metallic support of a slide format was fabricated with three 400 µm wide and 300 µm high grooves to hold the capillaries (Figure 1b). Surface Modification with Benzophenone and Molecule Grafting. The modification of the inner capillary surface with benzophenone was performed following a two-step procedure (Figure 2a). First, the inner surface precleaned with piranha was amino-modified with a 3% solution of APS in ethanol. Although 5476 Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

Figure 2. Photochemical patterning of the molecules inside glass capillary. (a) Scheme of the chemical modification of inner capillary surface with BP and photochemical grafting of the molecules via CH bond insertion. (b) Confocal microscope fluorescent images of the capillary rotated in a vertical plane show that the patterned Cy3-EGFP oligonucleotide is confined to the inner capillary surface. (c) Acquisition and treatment of the fluorescent signal obtained by the scanning of the capillary with slide array reader; the signal profile is shown in red and background fluorescence in black.

trifunctional silanes may result in a branched polymerization on the surface and in a microheterogeneity of the coating,28,29 they provide a high density of the functional groups on a surface. Moreover, APS coating prepared in this way appeared to be stable for at least several months and thus could be used for array assembly. To reduce surface heterogeneity, all modification steps were performed at a constant reagent flow (0.1-10 µL/min) by using the gravity flow or a syringe pump. The silanization step was followed by an ethanol washing and a curing at 115 °C for 2 h to stabilize the APS layer. Finally, the BP moiety was introduced via the reaction of the aminated surface with 200 mM BP-NCS in DMF (Figure 2a). The characterization of the modified inner surface was difficult due to the closed geometry of the capillary, though the modification steps could be followed by a gradual change in the capillarity force: water capillarity decreased significantly after the silanization step and further decreased after BP-NCS treatment reflecting an increasing hydrophobicity of the surface. The efficiency of the photochemical grafting of the molecules on the surface was also an indirect measure of the quality of BP coating (see below). A 365-nm UV light used for (28) Parikh, A. N.; Allara, D. L. J. Phys. Chem. 1994, 98, 7577-7590. (29) Brzoska, J. B.; Ben Azouz, I.; Rondelez, F. Langmuir 1994, 10, 4367-4373.

the irradiation of the capillaries is sufficient to excite BP and does not significantly affect nucleic acids and proteins. The major exited state of BP, a carbonyl-centered n, π* diradicaloid triplet, is able to cross-link with organic residues, the main reaction being the insertion into the C-H bond24,25 (Figure 2a). Usually high yields of photo-cross-linking (up to 80%) are obtained.30 The BP residue has already been used for photochemical grafting of the biological molecules on the surface, though in reported cases, the molecules in solution (and not on the surface) were labeled with BP.31,32 This approach differs significantly from the use of surface-bound BP since the molecules cross-link preferentially to each other and not with a surface. This results in the formation of polymers and decreases the efficiency of the molecule immobilization. On the other hand, the attachment of BP to the surface confines photochemical reactions to the irradiated surface area (Figure 2a) providing a high yield of photografting.26 The irradiation of the BP capillary filled with Cy3-labeled oligonucleotide resulted in the attachment of the oligonucleotide to the inner capillary surface (Figure 2b). No photografting was observed in the capillaries without BP coating, clearly demonstrating the role of BP in molecule attachment. The confocal microscopy analysis revealed cylindrical fluorescent zones of 160-190µm length (Figure 2b). Some enlargement of the zones with respect to the size of the irradiated area (160 µm) was observed with prolonged irradiation times (>2 min) and resulted most probably from the light scattering at the capillary-air and capillary-inner solution interfaces. Figure 2c shows a typical capillary image obtained with a microarray scanner. For optimal signal-to-noise ratio, the laser intensity was kept at the maximum level and the detector gain was set just high enough to barely measure the background. The signal was analyzed with a scanner software by measuring the profile of signal intensity along the x-axes perpendicular to the capillary axes (Figure 2c). The maximum intensity value that was attained at the intersection of the internal capillary wall and a scanning surface was used for calculation of signal-to-noise ratio. Oligonucleotide Patterning and Hybridization Assay. First we studied the photochemical patterning of oligonucleotides inside the BP capillary (Figure 3). Because of the possible photobleaching upon UV irradiation, the fluorophore-labeled oligonucleotides were not appropriate to follow the kinetics of immobilization. Therefore, biotinylated oligonucleotides were used. The solution of 3′-biotin 30-mer oligonucleotide coding for a region of human K-Ras protooncogene was irradiated in the capillary at different exposure times. The amount of attached oligonucleotide was then determined by reacting the biotin moiety with Cy3-labeled streptavidin. The immobilization appeared to be rapid and concentration dependent (Figure 3a). At smaller oligonucleotide concentrations, the curve levels off already at 15 s of irradiation. This was probably due to exhausting the available oligonucleotide pool at the surface proximity and not due to saturation of reactive surface sites, since an increase in oligonucleotide concentration resulted in higher (30) Prestwich, G. D.; Dorman, G.; Elliott, J. T.; Marecak, D. M.; Chaudhary, A. Photochem. Photobiol. 1997, 65, 222-234. (31) Liu, X. H.; Wang, H. K.; Herron, J. N.; Prestwich, G. D. Bioconjugate Chem. 2000, 11, 755-761. (32) Herbert, C. B.;., McLernon, T. L.; Hypolite, C. L.; Adams, D. N.; Pikus, L.; Huang, C. C.; Fields, G. B.; Letourneau, P. C.; Distefano, M. D.; Hu, W. S. Chem. Biol. 1997, 4, 731-737.

Figure 3. Oligonucleotide array inside capillary. (a) Effect of irradiation time and oligonucleotide concentration on the efficiency of oligonucleotide grafting. Biotin-labeled K-Ras oligonucleotide was irradiated for indicated time at 1 (diamonds), 50 (squares), or 100 µM (circles), and the amount of the attached oligonucleotide was determined with Cy3-streptavidin. (b) Hybridization efficiency of the immobilized oligonucleotide. EGFP oligonucleotide (600 µM) was irradiated inside the capillary for indicated time and the attached oligonucleotide was hybridized with Cy3-labeled complementary target at 10 (circles), 60 (squares), or 90 µM (diamonds). (c) Specificity of hybridization assay: the EGFP and P2A oligonucleotides were arrayed in the same capillary, and the array was hybridized with a mixture of complementary oligonucleotides labeled with Cy3 (EGFP, false colored in red) and with FITC (P2A, false colored in green). The capillary was then scanned on the slide array reader using the corresponding lasers.

grafting yield (Figure 3a). The longer irradiation times (>1 min) induced some decrease in fluorescent signal probably caused by Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5477

photochemical cleavage of BP-oligonucleotide adducts. Another explanation may be that the extensive cross-linking of the oligonucleotide induced by prolonged irradiation reduced the accessibility of the biotin moiety for Cy3-streptavidin. Similar results were obtained with oligonucleotides of different composition (sequence, GC content) and length (15-50-mer). In all cases, the immobilization yield was optimal at ∼30-60-s irradiation times. Higher yields were obtained by increasing the oligonucleotide concentration with a maximum at a submillimolar range (∼600 µM for K-Ras 30-mer, not shown). Then we examined whether the immobilized oligonucleotides are still able to hybridize with a complementary DNA. The photochemical cross-linking may damage the biological molecules and impair their functional properties. A 365-nm UV light used in this work does not affect DNA directly but can do it through photosensitizers such as BP.33 Two competing mechanisms of photosensitization have been identified. In a type I mechanism, the excited photosensitizer reacts directly with the biological molecule via charge transfer, hydrogen abstraction, or both, whereas in a type II reaction, the exited state is quenched by molecular oxygen producing singlet oxygen as reactive species.33,34 In the case of BP, the comparison of the reaction rate constants gives the ratio type I/type II >103, making the type II mechanism much less probable.35 By contrast, electron transfer from purine bases to excited triplet BP may result in the oxidation of the bases via the type I mechanism.36 Furthermore, hydrogen abstraction is the first step leading to the insertion of BP into the C-H bond and thus underlies oligonucleotide immobilization.24,25 On the other hand, since n, π* triplet of BP has been shown to react preferentially with DNA sugar backbone37 and not with the bases, the immobilization was not expected to impair significantly DNA hybridization potential. As shown in Figure 3b, the oligonucleotides immobilized inside the BP capillary could be hybridized with complementary DNA. In these experiments, the capillary with photochemically arrayed nonlabeled EGFP oligonucleotide was filled with a solution of Cy3-labeled oligonucleotide target and incubated at 42 °C. After washing, the attached target was visualized by using the laser scanner. Consistent with the results shown before (Figure 3a), the maximum fluorescent signal was observed with oligonucleotides immobilized with ∼60-s irradiation times (Figure 3b). The hybridization signal was raised with increasing target concentration and reached the maximum at the 1-10 µM range (not shown). Further increase in target concentration resulted in a decrease in signal-to-noise ratio (Figure 3b). The sensitivity limit of the hybridization assay was as low as ∼50 nM for Cy3-labeled target giving the detection limit of ∼2 × 107 molecules, comparable with a performance of other fluorescencebased microarray techniques ((2-3) × 107 molecules,38). Noncomplementary target sequences did not give any detectable fluorescent signal demonstrating the specificity of the hybridiza(33) Ravanat, J. L.; Douki, T.; Cadet, J. J. Photochem. Photobiol. B 2001, 63, 88-102. (34) Cosa, G. Pure Appl. Chem. 2004, 76, 263-275. (35) Miranda, M. A. Pure Appl. Chem. 2001, 73, 481-486.) (36) Gut, I. G.; Wood, P. D.; Redmond, R. W. J. Am. Chem. Soc. 1996, 118, 2366-2373. (37) Bosca, F.; Miranda, M. A. J. Photochem. Photobiol. B 1998, 43, 1-26. (38) Bertucci, F.; Bernard, K.; Loriod, B.; Chang, Y. C.; Granjeaud, S.; Birnbaum, D.; Nguyen, C.; Peck, K.; Jordan, B. R. Hum. Mol. Genet. 1999, 8, 17151722.

5478

Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

tion assay. To further address the specificity issue, two different oligonucleotides derived from EGFP and phospholipase 2A gene (P2A) were photochemically arrayed in the same capillary. The resultant array was then hybridized with a mixture of complementary oligonucleotides labeled with Cy3 (EGFP) or with FITC (P2A). Scanning of the capillary using two different lasers revealed that the labeled targets hybridize specifically with corresponding oligonucleotides (Figure 3c). Protein Patterning and Interaction Assay. Immobilization of proteins inside capillaries has been employed for integrated immunoassay and biosensor devices. Most of these devices are one protein-one capillary biosensors,16-18,39 where the entire internal capillary surface is covered with the protein of interest and detection occurs downstream from an assay region16-18 (flowthrough analytical system) or directly in the capillary19,39 (capillary fill device). As a result, for a multianalyte detection several capillaries have to be used. Recently, Misiakos and Kakabakos have described a method for protein patterning inside plastic capillaries based on a microsyringe injection and physical adsorption of the proteins on the internal plastic surface.20,21 Using this method, multiple protein interactions could be analyzed simultaneously in the same capillary. Ligler et al. used a light-sensible NMPEG-silane to produce a protein array inside fused-silica capillaries. Irradiation of the NMPEG-silane yields surface-bound aldehyde, which was used for protein immobilization via reductive amination.19 We studied the potential of direct BP-mediated photochemical grafting for protein patterning inside the capillary. For that reason, the biotinylated streptococcal protein G′ was chosen in the first photochemical patterning assays in the BP capillary. The functional integrity of the immobilized protein G′ can be tested by its ability to bind the constant Fc region of IgGs,40 while the biotin label allows the visualization of the attached protein by reacting with Cy3-streptavidin. Protein G′ at 2 mg/mL in PBS buffer was arrayed in two capillaries, which were then filled in parallel with 20 µg/mL Cy3-streptavidin (Str-Cy3) or 5 µg/mL fluoresceinlabeled IgG-FITC (Figure 4a). The kinetics of protein G′ immobilization was rapid: at 5 s of irradiation, the Str-Cy3 signal was at ∼70% of the maximum level and the plateau was reached at 30 s. The immobilized protein was still able to bind IgG-FITC, suggesting that at least a fraction of the protein remained functional and was not damaged by BP-mediated grafting. On the other hand, we observed significant nonspecific adsorption of IgG-FITC to the capillary wall, which resulted in a decrease in signal-to-noise ratio compared to Str-Cy3 (Figure 4a). The signalto-noise ratio can be improved by flushing the capillary array with 1% BSA and 0.025% Tween in PBS, though this also decreased the absolute signal amplitude. Finally, we developed a capillary immunoassay with two different antigens: viral protein-base (PB) and bacterial LLO. The PB is a highly antigenic capsid protein implicated in the human adenovirus Ad3 cell entry.27,41 The LLO is one of the major virulence factors of L. monocytogenes,42 a ubiquitous food-borne Gram-positive bacterium, responsible for septicemia and menin(39) Flanagan, M. T.; Sloper, A. N. U.S. Patent 5,081,012, 1992. (40) Sauer-Eriksson, A. E., Kleywegt, G. J.; Uhlen, M.; Jones, T. A. Structure 1995, 3, 265-278. (41) Wickham, T. J.; Mathias, P.; Cheresh, D. A.; Nemerow, G. R. Cell 1993, 73, 309-319.

were still measurable with higher antisera dilutions, 1:60 000 for R-PB and 1:10 000 for R-LLO. Taking into account the concentration of R-PB in stock solution, estimated to be ∼1.8 mg/mL, the detection limit of this assay is less than 200 pM. A similar value (163 pM) was reported by Whitesides et al. for fluorescent IgG immunoassay in a microfluidic chip made by soft lithography.12 The approach of this work resembles that of Ligler et al., which introduces a photosensitive NMPEG-silane for protein patterning inside the capillary.19 In both cases, the use of a photochemical addressing provides much higher resolution and density of array than does a liquid-dispensing technique.20,21 For example, we could reduce the protein bandwidth to 30 µm simply by decreasing the microscope slit aperture. However, it is worth noting that in the case of BP-modified capillary the photochemical patterning and protein immobilization are the same process, whereas with NMPEG-silane, they are performed in two steps: first, the pattern of surface-bound aldehyde groups is photogenerated and then the protein is immobilized via light-independent chemical reaction. Thus, although the use of NMPEG-silane avoids the proteins being exposed to UV irradiation, this increases significantly the time of reaction, which may be critical for the immobilization of multiple proteins.

Figure 4. Protein array inside capillary. (a) Effect of the irradiation on protein G attachment and functionality: the biotinylated protein G at 2 mg/mL was irradiated for indicated times in two capillaries and the attached protein was probed in parallel with Cy3-streptavidin (StrCy3) for biotin and with IgG-FITC for functional Fc-binding domain. (b) Specificity of the capillary immunoassay: LLO and PB antigens were irradiated for indicated times in four capillaries and the attached antigens were probed with specific rabbit antisera RLLO and RPB and anti-rabbit IgG-FITC.

gitis in immunocompromised persons. The recombinant proteins and their polyclonal rabbit antisera were used in the immunoassay. The antigens at 0.2 mg/mL were arrayed in parallel in four capillaries, and the resulting arrays were washed with 1% BSA and 0.025% Tween in PBS for 15 min. Next, the capillaries were treated successively with R-PB (1:10 000) or R-LLO (1:2000) antiserum in PBS/Tween and with 5 µg/mL anti-IgG-FITC. Figure 4b shows that the capillary array can reliably distinguish between the specific antisera. The dilutions used in this experiment were closed to standard ELISA assay dilutions (1:5000 and 1:1000 for R-PB and R-LLO, respectively). The fluorescent signals (42) Lety, M. A.; Frehel, C.; Beretti, J. L.; Berche, P.; Charbit, A. Microbiology 2003, 149, 1249-1255.

CONCLUSION We have shown here the application of surface-bound benzophenone for photochemical patterning of oligonucleotides and proteins inside fused-silica capillaries. The method is simple and requires only standard laboratory equipment that includes an inverted fluorescent microscope for array assembly and a slide array reader for data acquisition. The capillary arrays produced in this way are functional and can be used in different bioassays such as DNA hybridization, protein interaction studies, and immunoassays. The method has several advantages including submicroliter sample volume and a short assay time. The capillary format should therefore be considered as a possible alternative to the classical 2D format in a number of low-density array applications such as single-nucleotide polymorphism detection and diagnostic immunoassay. Another advantage of the capillary arrays is that they can be integrated into an analytical lab-on-a-chip microsystem that combines sample preparation, analysis, and data treatment. ACKNOWLEDGMENT We are indebted to P. Fender and G.-L. Gaillard for proteins and antibodies. We thank J. Chroboczek for critical reading of the manuscript and for helpful discussion. Received for review March 17, 2005. Accepted June 23, 2005. AC0504619

Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5479