Single-step imprinting of femtoliter microwell arrays allows digital

37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58 ..... w/w % of FDMA (0 %, 1 %, 2 %, 2.5 %, 3 %, 3.5 %, 5 %). ...
0 downloads 0 Views 7MB Size
Research Article www.acsami.org

Single-Step Imprinting of Femtoliter Microwell Arrays Allows Digital Bioassays with Attomolar Limit of Detection Deborah Decrop,† Gaspard Pardon,‡ Luigi Brancato,§ Dries Kil,§ Reza Zandi Shafagh,‡ Tadej Kokalj,† Tommy Haraldsson,‡ Robert Puers,§ Wouter van der Wijngaart,‡ and Jeroen Lammertyn*,† †

Department of Biosystems, KU LeuvenUniversity of Leuven, Willem de Croylaan 42, 3001 Leuven, Belgium Department of Micro- and Nanosystems, KTH Royal Institute of Technology, Stockholm, Sweden § Department of Electrotechnical Engineering (ESAT-MICAS), KU LeuvenUniversity of Leuven, Kasteelpark Arenberg 10, 3001 Leuven, Belgium ‡

S Supporting Information *

ABSTRACT: Bead-based microwell array technology is growing as an ultrasensitive analysis tool as exemplified by the successful commercial applications from Illumina and Quanterix for nucleic acid analysis and ultrasensitive protein measurements, respectively. High-efficiency seeding of magnetic beads is key for these applications and is enhanced by hydrophilic-in-hydrophobic microwell arrays, which are unfortunately often expensive or labor-intensive to manufacture. Here, we demonstrate a new single-step manufacturing approach for imprinting cheap and disposable hydrophilic-inhydrophobic microwell arrays suitable for digital bioassays. Imprinting of arrays with hydrophilic-in-hydrophobic microwells is made possible using an innovative surface energy replication approach by means of a hydrophobic thiol-ene polymer formulation. In this polymer, hydrophobic-moiety-containing monomers self-assemble at the hydrophobic surface of the imprinting stamp, which results in a hydrophobic replica surface after polymerization. After removing the stamp, microwells with hydrophobic walls and a hydrophilic bottom are obtained. We demonstrate that the hydrophilic-in-hydrophobic imprinted microwell arrays enable successful and efficient self-assembly of individual water droplets and seeding of magnetic beads with loading efficiencies up to 96%. We also demonstrate the suitability of the microwell arrays for the isolation and digital counting of single molecules achieving a limit of detection of 17.4 aM when performing a streptavidin−biotin binding assay as model system. Since this approach is up-scalable through reaction injection molding, we expect it will contribute substantially to the translation of ultrasensitive digital microwell array technology toward diagnostic applications. KEYWORDS: microwell array, femtoliter droplets, digital bioassay, lab-on-chip, polymer microfluidics, surface energy replication, thiol-ene-epoxy polymer, OSTE+

1. INTRODUCTION Isolation and digital detection of single molecules enable nextgeneration analytical tools for genomics,1−5 proteomics, cell analysis,6 and medical diagnostics.7−10 On the basis of this principle of isolating and counting single molecules, bead-based digital bioassays11−13 reach an attomolar limit of detection (LOD). In the following, we briefly introduce bead-based digital bioassays, discuss some of the most important recent developments in the area, and introduce the structure of the manuscript. In bead-based digital bioassays, target analytes are digitally detected through confinement of single targets in a microwell array using magnetic beads and a subsequent fluorescent amplification reaction.10,14 Magnetic beads are first functionalized with bioreceptors, such as antibodies or aptamers, to capture target molecules from a sample. Adding an excess of beads to the sample containing a low concentration of target © 2017 American Chemical Society

molecules results in a Poisson distributed capture of target molecules by the beads, meaning that either one or no target molecule will be captured on each bead. Subsequently, the captured target molecules are labeled with an enzyme that can produce a fluorescent signal. The magnetic beads with the built immunocomplex are then seeded in the microwells with single bead resolution. Efficient seeding of magnetic beads and simultaneous retaining of femtoliter droplets are accomplished using microwells with a hydrophilic bottom and hydrophobic interwell surface. This hydrophilic-in-hydrophobic nature allows effective embedding of individual magnetic beads with fluid droplets in each microwell when shuttling a liquid sample over the array. Finally, a fluorescent signal is generated only Received: December 1, 2016 Accepted: March 7, 2017 Published: March 7, 2017 10418

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces

Figure 1. Graphical representation of the hydrophilic-in-hydrophobic microwell array imprinting method and a bead-based digital bioassay. A) A PDMS stamp, containing a micropillar array is fabricated using standard soft-lithography techniques. Silanized glass slides are spin coated with the hydrophobic OSTE+ formulation and subsequently, the PDMS stamp is pressed in the liquid polymer to imprint the microwells. The hydrophobic monomers self-assemble at the interface and after UV-curing they covalently bind to the polymer matrix. Finally, the PDMS stamp is removed, revealing hydrophilic microwells within the hydrophobic polymer surface. B) As much magnetic beads as possible are seeded in a microwell array with single bead resolution. A droplet of a fluorogenic substrate is transported over the surface for printing single droplets generating a detectable fluorescent signal.

preparation of 106 dome-shaped droplets that can be individually accessed and manipulated using a micropipette.20,21 Recently, Kan et al.22 reported on the DVD-based fabrication method of microwell array discs using injection molding of cyclic olefin polymer. This microwell array disc addressed the limitations of the fiber-optic microwell arrays, facilitating the establishment of a commercial SiMoA platform (Quanterix corporation). For the SiMoA HD-1 Analyzer, Quanterix entered into a collaboration with Sony DADC BioSciences for the development and manufacturing of the Simoa discs (smart consumables).14,23 Despite the low-cost and high throughput manufacturing,23 the hydrophobic-in-hydrophobic microwell arrays of the Simoa discs only allow loading efficiencies of 40−50%.22 These low loading efficiencies are attributed to the gravitational bead seeding approach where magnetic beads are delivered fluidically and allowed to settle in the microwells via gravity. Witters et al.16 described a digital microfluidic platform with a microwell array for printing single magnetic beads using electrowetting-on-dielectric actuation of droplets. In their platform, 62 500 hydrophilic-in-hydrophobic femtoliter-sized microwells (4.5 and 3 μm width and depth, respectively) were patterned in a Teflon-AF layer of a double plated digital microfluidic chip. Because of this plate configuration, a permanent magnet could be placed below the array to assist seeding of magnetic beads in the microwells with single bead resolution, achieving bead-loading efficiencies as high as 99%. Despite the good feature quality and high loading efficiencies, the modified dry lift-off technique24 they developed is rather

inside those wells where an enzyme-labeled target molecule is present (“fluorescent” wells), and, provided that the microwell array surface has a low autofluorescence, fluorescent signals can be imaged for subsequent analysis. As a consequence, this digital principle allows counting of the number of “fluorescent” wells that linearly correlates with the bulk concentration of the target.11,14−16 As an example, digital ELISA was used for precise measurements of prostate specific antigen11 and Influenze A Nucleoprotein,15 achieving an LOD of 50 aM and 4 ± 1 fM, respectively. Rissin et al.11,17 were the first to develop this bead-based method for digital detection. They established a microwell array platform on an optical fiber bundle that is loaded with individual magnetic beads using centrifugal forces, reaching loading efficiencies of around 60%. However, the use of optical fiber glass bundles is expensive and the fabrication method uses aggressive chemicals. On the basis of these fiber-optic microwell arrays, Illumina developed a BeadArray platform enabling high throughput nucleic acid analysis.1−5,18 Using an evaporationbased bead seeding mechanism, a bead retention of >97% was achieved. However, dehydration is not compatible with proteins or enzymes since it induces protein denaturing and loss of enzyme activity. Noji et al.19 proposed a simple method using PDMS microwell arrays and a planar sealing glass substrate for performing digital bioassays. This mechanical sealing of the microwell array, however, requires high precision and skills. They also developed a method for the simultaneous 10419

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces

stamp. The polymer is thereafter cross-linked by UVillumination. This step ensures that the hydrophobic monomers are covalently bound with the OSTE+ matrix. After peeling-off the PDMS stamp, uniform microwells remain in the hydrophobic polymer layer, with glass as the microwell bottom. At this point, a digital bioassay can be performed (Figure 1B). First magnetic beads, that have captured single enzymelabeled target molecules, are seeded in the microwell array effectively isolating the individual target molecules into femtoliter reaction chambers. Subsequently, individual droplets of a fluorogenic substrate are printed on top of the beads such that a fluorescent signal can be generated in those wells that accommodate a magnetic bead containing an enzyme-labeled target molecule. The scheme of this process is presented in Figure 1B.

complex, labor-intensive and is hard to automate through standard MEMS processing technologies. The key hardware enabler of bead-based digital bioassays are arrays enclosing ∼104−105 femtoliter (fL)-sized microwells that each can accommodate one magnetic bead. Most common large-scale manufacturing methods for lab-on-chip devices rely on injection molding of thermoplastics. However, to obtain hydrophilic-in-hydrophobic microwells, the bottom surface of each microwell must be modified to become hydrophilic. Such surface energy patterning is not trivial, as it commonly requires complex postmanufacturing surface modifications. Recently, UV-curable thiol-enes have attracted attention as an interesting material for the production of lab-on-chip devices, compatible with both lab routines and adaptable to large-scale production.25 Specifically, off-stoichiometric thiol-ene and offstoichiometric thiol-ene-epoxy (here referred to as OSTE and OSTE+, respectively) polymers with well-controlled chemistry and modularity offer many advantages, including UV-photo structuring, UV-patternable surface modifications, tunable mechanical properties and direct low-temperature dry bonding.25,26 Pardon et al.27,28 modified a naturally hydrophilic OSTE polymer such that it can simultaneously and spontaneously replicate the surface energy of its mold surface during polymerization. This new development has inspired a novel fabrication method for disposable microwell arrays, making microwell array technology more accessible. In this work, we first describe the fabrication of arrays with hydrophilic-in-hydrophobic microwells using a novel single-step fabrication procedure by stamp molding of a low autofluorescent, leach-free, hydrophobic OSTE+ polymer formulation that does not require postmanufacturing surface modifications. We describe how the microwell imprinting method was validated for its effective replication quality and high well-size uniformity, and how the addition of fluorinated methacrylate monomers (FDMA) to the OSTE+ polymer formulation was investigated for its ability to effectively replicate hydrophobic surface energies. In the following, we demonstrate seeding of magnetic beads with high loading efficiencies (∼96%) and digital detection of single biotinylated β-galactosidase (BβG) enzyme molecules, as a well characterized model system, in the microwell array platform reaching a LOD of 17.4 aM.

3. EXPERIMENTAL SECTION 3.1. Materials. Silicon wafers (3 in.), hexamethyldisilazane (HDMS), AZ 6632 (1.2-μm grade) positive photoresist and AZ 726 MIF developer were bought from MicroChemicals GmbH (Ulm, Germany). PDMS (Sylgard 184 silicone elastomer kit) and Z-6030 silane were purchased from Dow Corning (California, U.S.A.). The fluoroalkyl silane, Dynasylan, was provided by Evonik (Essen, Germany). OSTEmerX Crystal Clear (322−40) was purchased from Mercene Laboratories AB (Stockholm, Sweden). Borosilicate glass microscope slides were bought from VWR (Leuven, Belgium). SigmaAldrich (Bornem, Belgium) was a supplier of the following items: fluorogenic substrate fluorescein di(β-D-galactopyranoside) (FDG), bovine serum albumin (BSA), biotinylated-B-galactosidase (BβG) from Escherichia coli, 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10-heptadeca-fluorodecyl methacrylate (FDMA), Poly(ethylene glycol) methacrylate, Tween-20, and MgCl2. Lodestar superparamagnetic beads with diameter of 2.7 μm and streptavidin functionaliztion were obtained from Agilent technologies. PlusOne Drystrip Coverfluid oil was obtained from GE Healthcare (The Netherlands). 3.2. Silicon Master and PDMS Stamp Fabrication. A PDMS stamp was fabricated with a range of micropillar diameters (2 to 100 μm) according to the following protocol. A silicon (Si) wafer was first primed with adhesion promotor HMDS by spin coating 5 s at 500 rpm and 10 s at 3000 rpm. Immediately after, a ∼1.2 μm layer AZ 6632 positive photoresist was applied by spin coating for 30 s at 3000 rpm followed by baking at 120 °C for 1 min. Next, the photoresist was treated with UV-light (12 mW/cm2) through a chrome-on-glass mask, developed for 45 s with AZ 726 MIF Developer and postbaked for 1 h at 100 °C. Subsequently, a Bosch silicon deep reactive ion etching (DRIE) process29 was used for transferring the microwell arrays in the Si wafers (see Table S1 of the Supporting Information, SI, for process parameters). After DRIE etching, the remaining photoresist was removed by sonicating the wafers for 10 min in acetone, followed by rinsing with acetone, isopropanol, and water, respectively. Prior to soft replication into a PDMS stamp, all Si masters were piranha cleaned using 3:1 mixture of sulfuric acid supplemented with 30% hydrogen peroxide in order to remove remaining debris and subsequently spin coated with fluoroalkyl silane for 15 s at 3000 rpm to prevent adhesion of cured PDMS to the master. Liquid PDMS prepolymer, with a mass ratio base:cross-linker of 5:1, was stirred for 2 min and degassed for 30 min at 0.005 mPa prior to pouring onto the Si masters. After pouring onto the Si master and a second degassing step of 30 min at 0.005 mPa, the PDMS was cured at 60 °C for 6 h. The obtained PDMS slabs with micropillars (Figure 1A-1) were peeled off from the Si masters and cut into appropriate pieces. The microwell array and micropillar array features of respectively the Si master and PDMS stamp were investigated using scanning electron microscopy (SEM). The PDMS stamp was sputter coated (Quorum Q150T ES, Quorum Technologies, East Sussex, U.K.) with a thin gold layer to avoid electrostatic charging. 3.3. Modified OSTE+ Polymer Formulation for Replication of Hydrophobic Surface Energies. In this work, OSTE+ was exploited

2. CONCEPT OF MICROWELL ARRAY MANUFACTURING To develop and deploy sensitive digital bioassays, disposable hydrophilic-in-hydrophobic microwell arrays are required, as schematically illustrated in Figure 1A. A reusable PDMS stamp, containing micropillar arrays, is prepared by conventional PDMS casting (Figure 1A-1) for imprinting of microwell arrays (Figure 1A-2). A hydrophobic OSTE+ polymer formulation is prepared, spin-coated onto a glass substrate, imprinted in the liquid state with the PDMS stamp and polymerized with UVlight. To achieve this single-step manufacturing, a hydrophobic OSTE+ polymer formulation was specifically developed to enable replication of hydrophobic surface energies. In this formulation, FDMA monomers are added to the OSTE+ matrix to allow replication of hydrophobic surface energies by molecular self-assembly at the mold interface. When the PDMS stamp makes contact with the polymer layer, the hydrophobic FDMA monomers, freely diffusing in the yet uncured polymer, self-organize at the hydrophobic stamp surface, effectively replicating the hydrophobicity of the PDMS 10420

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces for preparing the hydrophobic polymer formulation. The OSTE+ polymer matrix was prepared by mixing the OSTEmerX Crystal Clear (322−40) compounds A and B with a mass ratio of 1.1:1. Subsequently, the OSTE+ polymer matrix was modified with FDMA monomers. First, the FDMA solution was preliminary diluted to a 15 w/w % solution in toluene for mixing with the OSTE+ prepolymer matrix to obtain final FDMA concentrations ranging from 0 to 5 w/w % of FDMA (0%, 1%, 2%, 2.5%, 3%, 3.5%, and 5%). Water contact angles were measured on the surface of both nonmodified OSTE+ and hydrophobic OSTE+ cured against either unstructured PDMS or borosilicate glass slides. For these measurements, a goniometer (Theta Lite, Biolin Scientific, Sweden) was used. 3.4. Investigation of the Autofluorescence of Hydrophobic OSTE+. The autofluorescence of hydrophobic OSTE+ surfaces was examined using fluorescence microscopy (Nikon TiEclipse, Japan) and compared to that of hydrophobic OSTE, Teflon-AF and glass surfaces. Fluorescence images were taking using three different excitation/ emission filter channels (DAPI ex340−380/em435−485; FITC ex465−495/em515−555; TRITC ex540−575/em605−655) and analyzed using ImageJ software and JMP Pro12 statistical software. 3.5. Stamp-Molding of Microwell Arrays. Figure 1A-2 schematically illustrates the microwell array imprinting method. First, acetone, isopropanol, and deionized water were used to clean borosilicate microscope glass slides (5 min each). Subsequently, the glass slides were submerged in a methacrylate silane solution (5% silane in methanol, 10 min) and baked (120 °C, 10 min). Thereafter, a 3 μm hydrophobic OSTE+ layer was spin-coated for 30 s at 3000 rpm on the glass slides. Next, the PDMS stamp, containing micropillar arrays, i.e., inverted microwell array structures, was gently pushed into the 3 μm polymer film, followed by 120 s of UV-curing at 12 mW/cm2 for (OAI, U.S.A.). After UV-treating, the stamp was peeled off followed by curing at room temperature overnight. The manufactured microwell arrays were investigated using bright field microscopy and SEM. 3.6. Study of Hydrophilic Character of the Microwell Bottoms. The absence of polymer squeeze-film at the microwell bottoms was investigated using fluorescence microscopy. A 50 μL droplet of a 1 μM rhodamine B solution was incubated for 2 h on top of the microwell array and subsequently rinsed with deionized water. Fluorescence images were made both before and after incubation with rhodamine using the TRITC ex540−575/em605−655 filter channel and the images were analyzed using ImageJ software. 3.7. Printing of Single Water Droplets in the Microwell Array. The hydrophilic nature of the microwell bottoms was also confirmed by printing single water droplets. A droplet of dyed water was dragged over the microwell array, with wells of 50 μm diameter, and a movie was taken of the self-assembly of single water droplets using a stereomicroscope. Similarly, femtoliter-sized droplets were generated in a microwell array with wells of 5 μm diameter. A 30 μL droplet of a 1 nM fluorescein solution was first pipetted atop of the microwell array and then covered with 180 μL of PlusOne Drystrip Coverfluid oil to prevent evaporation. Subsequently, the fluorescein droplet was pulled away from the microwell array to print and seal fluorescein in femtoliter-sized droplets. Fluorescent images of the fluorescein femtoliter droplets were acquired using an inverted fluorescence microscope (Nikon TiEclipse, Japan). 3.8. Printing of Single Superparamagnetic Beads in the Microwell Array. Superparamagnetic Lodestar beads of 2.7 μm diameter were manually loaded in the microwell array using a pipet tip. First, a 15 μL 1× PBS buffer droplet, suspended with beads (2.5 × 105 beads) was pipetted on top of the microwell array. Subsequently, this droplet was manually shuttled over the array. This was accomplished by dragging the droplet 12 times back and forth with the pipet tip, while keeping a magnet (NdFeB, 12.7 N, 4 × 4 mm2, Supermagnete, Gottmadingen, Germany) underneath the array substrate to assist capturing of the beads in the microwells. One back and forward movement was called a seeding cycle. The seeding performance was analyzed using bright field microscopy and the number of seeded beads were counted after each cycle using the NIS-elements software of Nikon. Using this superparamagnetic bead seeding principle, digital

detection of biotin−streptavidin interaction was performed to examine the suitability of the imprinted microwell arrays for digital bioassay implementation. The experimental work flow is illustrated in Figure S1 and described in detail in section 2.1 of the SI.

4. RESULTS AND DISCUSSION 4.1. Silicon Master and PDMS Stamp Fabrication. The microwell and micropillar features of respectively the Si master and PDMS stamp were examined using SEM-imaging and revealed smooth in Si etched microwells and subsequent transferring through PDMS molding. Figure 2A,B shows that

Figure 2. SEM images of A) Si wafer containing a microwell array with ∼4 μm diameter; B) cleaved Si wafer, under an angle of 45°, showing microwells of 3.5 μm height; C) PDMS stamp, under an angle of 60°, containing micropillar structures of ∼4 μm diameter replicated from the Si master; D) PDMS stamp, under an angle of 60°, showing micropillars of ∼3 μm height.

the microwell dimension in the Si master is ∼3 μm deep, after the DRIE-process. High aspect ratio dry etching processes, like DRIE, require an etch mask of a specifically tailored photoresist (e.g., AZ 6632 photoresist) to achieve steep sidewalls and an elevated softening point. Moreover, the cyclic nature of the Bosch DRIE-process provided accurate control of the anisotropy by alternating etching and depositing phases, resulting in vertically straight etched microwells with small sidewall undulations (Figure 2B). Short cycling times (see process parameters in Table S1) limited this scalloping to a maximum amplitude of 50 nm29 which is preferred for replication processes via molding because they form a critical factor for unmolding. Overly large undulations in the Si master would cause breakage of the PDMS pillars during demolding from either the silicon master or the hydrophobic OSTE+ replica. Also, the dual curing property of OSTE+ benefits successful demolding since the polymer is not yet completely hard after the first curing. This allows minor bending of the PDMS stamp during demolding, preventing breakage of the micropillars. ImageJ was used for the analysis of the microwell array diameter distribution and resulted for the Si master as well as for the PDMS stamp in an average microwell diameter of 3.7 ± 0.3 μm and 3.7 ± 0.7 μm, respectively. These results indicate successful molding of PDMS micropillars without breakage of the pillars during demolding (see Figure 2C and D). 10421

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces 4.2. Modified OSTE+ Polymer Formulation for Replication of Hydrophobic Surface Energies. The imprinting of hydrophilic-in-hydrophobic microwell arrays was achieved by adding hydrophobic FDMA monomers to the OSTE+ polymer matrix. The use of hydrophobic monomers was inspired by the surface energy mimicking technology earlier reported.27,28 However, instead of using different surface energy mimicking monomers, only one hydrophobic monomer was used to transform the naturally hydrophilic OSTE+ polymer into a hydrophobic polymer capable of replicating hydrophobic surface energies. The principle of surface energy mimicking remains the same, with the polymer becoming hydrophobic when cured against a hydrophobic surface, such as PDMS. Compared to OSTE,27−29 OSTE+ is a thiol-ene-epoxy polymer in which the extra epoxy monomer enables full monomer cross-linking, despite the offstoichiometric ratio that exists between the thiol and ene monomers. This full cross-linking precludes presence of leachable monomers in the final polymer. Moreover, the dual-cure (UV light + heat) of OSTE+ supports using the epoxy or thiol groups for low-temperature bonding during the second curing step, enabling integration with a supporting substrate or additional microfluidic layer. The property of the hydrophobic OSTE+ polymer formulation to replicate hydrophobic surface energies was characterized with contact-angle measurements. Different FDMA concentrations ranging from 0 to 5 w/w%, were tested. Figure 3A shows the contact angle of hydrophobic OSTE+ when cured against PDMS (ΘPDMS = 113°) as a function of the FDMA concentration (w/w %). For FDMA concentrations between 1 and 5 w/w% the contact angle of the hydrophobic OSTE+ surfaces, cured

against PDMS, was above 100° (Figure 3A). A hydrophilic contact angle of 87.01° was obtained when curing OSTE+, without the addition of FDMA monomers (0%), against the PDMS mold. Curing of a 3 w/w % hydrophobic OSTE+ layer against hydrophilic borosilicate glass, results in a hydrophilic contact angle of 83.4°, whereas curing against hydrophobic PDMS results in a hydrophobic contact angle (Figure 3B). This is contrary to the results obtained for nonmodified OSTE+, which shows contact angles of only 87° and 70° when cured against PDMS and glass, respectively.28 These results indicate that self-assembly of hydrophobic FDMA monomers at the mold interface is responsible for replication of hydrophobic surface energy of the mold. This can be further explained by diffusion of FDMA monomers within the hydrophobic OSTE+ prepolymer matrix during UV-curing and the self-assembly of these FDMA monomers at the prepolymer/hydrophobic PDMS interface, which is driven by surface energy minimization. However, the FDMA monomers do not organize themselves at the interface when curing hydrophobic OSTE+ against a glass surface (or any other hydrophilic surface), resulting in a naturally hydrophilic OSTE+ surface. The results of these contact angle measurements demonstrate the successful ability of hydrophobic OSTE+ to mimic the hydrophobic surface energy of the PDMS mold. A 2.5 w/w % concentration of FDMA was found to give the highest contact angle with a high reproducibility and was therefore used for all subsequently imprinted microwell arrays. 4.3. Investigation of the Autofluorescence of Hydrophobic OSTE+. Low autofluorescence of the microwell array surface is a requirement for obtaining detectable fluorescent signals in a digital bioassay. Therefore, the autofluorescence of hydrophobic OSTE+ surfaces was analyzed using fluorescence microscopy with three different excitation/emission filters. The graph in Figure 4 shows the fluorescence intensity of four

Figure 4. Autofluorescence of glass, Teflon-AF, hydrophobic OSTE, and hydrophobic OSTE+ surfaces using three different excitation/ emission filters (UV, blue, and green excitation wavelength). The errors are standard deviations (n = 3). Threshold letters (A and B) indicate statistically significant different surfaces within a specific filter channel group (significance level at α = 0.05). Figure 3. A) Contact angle measurements of hydrophobic OSTE+ when cured against PDMS as a function of the FDMA concentration mixed in the OSTE+ polymer matrix. The standard deviation was calculated based on three independent substrate measurements. B) The static contact angle of a water droplet on the modified hydrophobic OSTE+ surfaces, after molding against borosilicate glass and PDMS. The hydrophobic OSTE+ inherits PDMS hydrophobicity while remaining hydrophilic against glass.

different surfaces for the three fluorescence filter channels. Using JMP Pro12 software, a one-way ANOVA statistical test was performed followed by a Tukey multiple comparison test to determine statistically significant differences between the means of the four different surfaces within each filter channel group. Bars with different letters indicate surfaces within a specific filter channel that significantly differ from each other 10422

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces

substrate and the PDMS stamp pillars), was verified using rhodamine adsorption and fluorescence microscopy. The selective absorption of rhodamine molecules to OSTE+ surfaces, but not to glass allows for discrimination between different surfaces.26 Figure 6A,B shows fluorescence images taken before and after rhodamine incubation. As illustrated in Figure 4, cured hydrophobic OSTE+ shows limited autofluorescence when excited with a wavelength of 540 nm (green/red filter channel). This low autofluorescence is shown on the left images of Figure 6A,B for 5 and 50 μm wells, respectively. After adsorption of rhodamine, fluorescence can be seen on the top surface and on the vertical walls of microwells, while the microwell bottoms show no fluorescence (Figure 6A,B, right images). This absence of fluorescence on the bottom side of microwells is probably due to being free of hydrophobic OSTE+. Using ImageJ, fluorescence intensity profiles were made before and after rhodamine incubation across a 50 μm well (Figure 6C). Figure S3 shows the profile of fluorescence intensity across a 5 μm well before and after incubation of rhodamine. These results indicate the absence of hydrophobic OSTE+ in the microwell bottoms, i.e., no formation of polymer squeeze-film during imprinting. The absence of this squeeze-film is attributed to the stamping nature of our imprinting method (e.g., the polymer is squeezed from under the micropillars). Moreover, because of oxygen diffusion through the PDMS stamp, the polymerization rate in the squeezed-film region is effectively reduced. As a result, polymerization of thin squeezed-film is prevented and the uncured polymer can be washed away, leaving a hydrophilic glass surface exposed at the microwell bottoms. 4.6. Printing of Single Droplets in the Microwell Array. Printing of individual water droplets confirmed the hydrophilicity of the microwell bottoms. Dye solution was transported in a droplet across the microwell array and visualized under a stereomicroscope. Figure 7A illustrates the printing of individual droplets in 50 μm diameter microwells. A section of a 62 500-microwell array of 5 μm diameter is shown in the fluorescence image of Figure 7B after printing of femtoliter fluorescein droplets. 4.7. Printing of Single Superparamagnetic Beads in the Microwell Array. By transporting a droplet suspended with magnetic beads as detailed in section 3.8, beads get trapped inside the wells. One backward and forward movement is called a seeding cycle. Trapping of the beads inside the microwells is achieved due to the (i) hydrophilic nature of the microwell bottom surface, (ii) the microwell geometry, and (iii) the magnetic force exerted on the magnetic beads. Excess of beads that remain on the array surface after seeding are dragged away by the receding droplet meniscus. Figure 8A depicts bright field microscopy image of a microwell array filled with magnetic particles. As demonstrated in the graph of Figure 8B, there was a positive correlation between the percentage of microwells capturing an individual bead and the number of seeding cycles. Already after 8 seeding cycles a loading efficiency of 96% ± 1% was obtained. In the next 4 seeding cycles, the loading efficiency only increased slightly, up to 97% ± 1% in the 12th cycle. The ability to perform digital bioassays in the imprinted microwell arrays was tested by capturing single BβG enzymes on streptavidin-coated beads and subsequently seeding of beads in the microwell array for visualization. Digital detection of streptavidin−biotin interactions, using β-galactosidase as a reporter enzyme, was specifically selected as a model bioassay because it was also used in previously published

(see Figure 4). On the basis of this statistical analysis, TeflonAF surfaces have as low autofluorescence as glass surfaces for all three the filter channels. For hydrophobic OSTE, the analysis indicated significantly higher autofluorescence compared to glass for all three the filter channels whereas compared to Teflon-AF only for the UV/Blue and Green/Red filter. In contrast, the autofluorescence of hydrophobic OSTE+ only appeared to be significantly higher compared to glass and Teflon-AF surfaces when using the UV-excitation filter. Typically, digital bioassays are performed using a fluorescein substrate that has an excitation and emission peak at 494 and 515 nm, respectively. This excitation and emission peak fall within the range of the Blue/Green filter channel (FITC ex465−495/em515−555). Since the autofluorescence of hydrophobic OSTE+ is as low as for glass and Teflon-AF when using this Blue/Green filter channel, no interference will occur with the fluorescent signal coming from the fluorescein substrate. 4.4. Stamp-Molding of Microwell Arrays. The in hydrophobic OSTE+ imprinted microwell arrays were investigated using two types of microscopy: bright field and SEM. Both imaging approaches confirmed the performance of the proposed imprinting method by revealing excellent features of the microwell arrays. High uniformity in the size of microwells, and a limited number of defects was visible in the bright field image (Figure 5A). ImageJ was used for analysis of the

Figure 5. Microwell arrays imaged with A) bright field and B) SEM microscopy, showing microwells with a diameter of 4 and 5 μm, respectively. (It is important to note that electrostatic charging hindered the quality of the SEM image, which is attributed to the lack of an electrically conductive coating on the polymer.)

microwell diameter distribution and gave for a 4 μm microwell array an average diameter of 4.0 ± 0.3 μm. The diameter of the PDMS micropillars used for imprinting the microwell array was 3.7 ± 0.7 μm, as determined in section 4.1. The diameter of the imprinted microwells is slightly bigger than that of the PDMS micropillars due to the stamping method that slightly squeezes the micropillars against the glass substrate. SEM images (Figure 5B) further confirm excellent dimensional and topological aspects of the investigated arrays with 5 μm diameter of imprinted microwells. 4.5. Study of Hydrophilic Character of the Microwell Bottoms. Loading of magnetic beads in a microwell array with high efficiency requires hydrophilic-in-hydrophobic microwells, meaning that the microwells have hydrophilic bottoms and a hydrophobic surrounding surface. The absence of a polymerized hydrophobic OSTE+ squeeze-film (i.e., the hydrophobic OSTE+ polymer that remains trapped between the glass 10423

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces

Figure 6. Fluorescence microscopy images of A) 5 μm and B) 50 μm microwell arrays before and after rhodamine incubation (left and right, respectively). No autofluorescence of the hydrophobic OSTE+, is visible before incubation with rhodamine. After incubation with rhodamine, fluorescence is visible only on the top surface and on the vertical walls of microwells (bright ring structures). C) A fluorescence intensity profile through a microwell with 50 μm diameter after rhodamine incubation reveals minor fluorescence at the glass bottom of the microwell, whereas large adsorption is detected on both sidewalls and top surface of the hydrophobic OSTE+.

5. CONCLUSIONS We have combined a stamp-molding technique with a novel hydrophobic OSTE+ polymer formulation to manufacture hydrophilic-in-hydrophobic microwell arrays in a single step while simultaneously replicating the hydrophobic surface energy of the stamp mold. The obtained microwell arrays are shown to have appropriate dimensions and excellent topological features. Attributed to local oxygen polymerization inhibition in the squeezed-film regions, hydrophilic microwell array bottoms are obtained. Due to the hydrophilic-inhydrophobic nature of the microwell arrays, successful printing of 62 500 femtoliter droplets was successfully demonstrated. The imprinted microwell arrays do not only allow for printing femtoliter droplets but also for the simultaneous seeding of individual magnetic particles in the microwell array with a seeding efficiency of 96% ± 1%. Using biotinylated-BβG and streptavidin coated magnetic beads, we demonstrated the capability of the microwell arrays for performing digital bioassays. The developed hydrophobic replication imprinting method can be used as a fast, inexpensive and out-of-cleanroom way of making microwell array devices. Moreover, to enable increased manufacturing throughput of our microwell arrays, we are currently adapting the recently published reaction injection molding of OSTE+ polymer31 to accommodate for the fabrication of hydrophilic-in-hydrophobic microwells. This

Figure 7. A) Microwell array (wells have a 50 μm diameter) imaged with stereomicroscopy. Individual water droplets are printed while moving a dyed water droplet across the array. B) Array with 62 500microwells (5 μm diameter) with printed femtoliter droplets of fluorescein as imaged on the fluorescence microscope.

platforms.16,30 The obtained results of this digital bioassay, shown in Figure S2, are discussed in section 2.2 of the SI. A linear dynamic range was obtained between 100 aM and 100 fM with a calculated LOD of 17.4 aM. This obtained linear dynamic range is comparable to earlier reported results for streptavidin−biotin BβG digital bioassays by Witters et al.16 (10 aM to 90 fM) and Zhang et al.30 (846 aM to 423 fM).

Figure 8. Seeding magnetic beads in the microwell array. A) Bright field microscopy image of a microwell array seeded after 10 seeding cycles (60× magnification); B) Percentage of seeded beads as a function of the seeding cycle. 10424

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces

Detection via Single Molecule Arrays Towards Early Stage Cancer Monitoring. Sci. Rep. 2015, 5, 11034−11041. (8) Wu, D.; Milutinovic, M. D.; Walt, D. R. Single Molecule Array (Simoa) Assay with Optimal Antibody Pairs for Cytokine Detection in Human Serum Samples. Analyst 2015, 140, 6277−6282. (9) Wilson, D. H.; Hanlon, D. W.; Provuncher, G. K.; Chang, L.; Song, L.; Patel, P. P.; Ferrell, E. P.; Lepor, H.; Partin, A. W.; Chan, D. W.; Sokoll, L. J.; Cheli, C. D.; Thiel, R. P.; Fournier, D. R.; Duffy, D. C. Fifth-Generation Digital Immunoassay for Prostate-Specific Antigen by Single Molecule Array Technology. Clin. Chem. 2011, 57, 1712− 1721. (10) Walt, D. R. Protein Measurements in Microwells. Lab Chip 2014, 14, 3195−3200. (11) Rissin, D. M.; Kan, C. W.; Campbell, T. G.; Howes, S. C.; Fournier, D. R.; Song, L.; Piech, T.; Patel, P. P.; Chang, L.; Rivnak, A. J.; Ferrell, E. P.; Randall, J. D.; Provuncher, G. K.; Walt, D. R.; Duffy, D. C. Single-Molecule Enzyme-Linked Immunosorbent Assay Detects Serum Proteins at Subfemtomolar Concentrations. Nat. Biotechnol. 2010, 28, 595−599. (12) Burger, R.; Reith, P.; Kijanka, G.; Akujobi, V. I.; Abgrall, P.; Ducrée, J. Array-Based Capture, Distribution, Counting and Multiplexed Assaying of Beads on a Centrifugal Microfluidic Platform. Lab Chip 2012, 12, 1289−1295. (13) Chang, L.; Rissin, D. M.; Fournier, D. R.; Piech, T.; Patel, P. P.; Wilson, D. H.; Duffy, D. C. Single Molecule Enzyme-Linked Immunosorbent Assays: Theoretical Considerations. J. Immunol. Methods 2012, 378, 102−115. (14) Wild, D., John, R.; Sheehan, C. The Immunoassay Handbook Theory and Applications of Ligand Binding, 4th ed.; Elsevier: London, 2013; Chapter 2.13, pp 223−242. (15) Leirs, K.; Kumar, P. T.; Decrop, D.; Pérez-Ruiz, E.; Leblebici, P.; Van Kelst, B.; Compernolle, G.; Meeuws, H.; Van Wesenbeeck, L.; Lagatie, O.; Stuyver, L.; Gils, A.; Lammertyn, J.; Spasic, D. Bioassay Development for Ultrasensitive Detection of Influenza A Nucleoprotein Using Digital ELISA. Anal. Chem. 2016, 88, 8450−8458. (16) Witters, D.; Knez, K.; Ceyssens, F.; Puers, R.; Lammertyn, J. Digital Microfluidics-Enabled Single-Molecule Detection by Printing and Sealing Single Magnetic Beads in Femtoliter Droplets. Lab Chip 2013, 13, 2047−2054. (17) Rissin, D. M.; Walt, D. R. Digital Concentration Readout of Single Enzyme Molecules Using Femtoliter Arrays and Poisson Statistics. Nano Lett. 2006, 6, 520−523. (18) Nesterov-Mueller, A.; Maerkle, F.; Hahn, L.; Foertsch, T.; Schillo, S.; Bykovskaya, V.; Sedlmayr, M.; Weber, L. K.; Ridder, B.; Soehindrijo, M.; Muenster, B.; Striffler, J.; Bischoff, F. R.; Breitling, R.; Loeffler, F. F. Particle-Based Microarrays of Oligonucleotides and Oligopeptides. Microarrays 2014, 3, 245−262. (19) Rondelez, Y.; Tresset, G.; Tabata, K. V.; Arata, H.; Fujita, H.; Takeuchi, S.; Noji, H. Microfabricated Arrays of Femtoliter Chambers Allow Single Molecule Enzymolgoy. Nat. Biotechnol. 2005, 23, 361− 365. (20) Sakakihara, S.; Araki, S.; Iino, R.; Noji, H. A Single-Molecule Enzymatic Assay in a Directly Accessible Femtoliter Droplet Array. Lab Chip 2010, 10, 3355−3362. (21) Kim, S. H.; Iwai, S.; Araki, S.; Sakakihara, S.; Iino, R.; Noji, H. Large-Scale Femtoliter Droplet Array for Digital Counting of Single Biomolecules. Lab Chip 2012, 12, 4986−4991. (22) Kan, C. W.; Rivnak, A. J.; Campbell, T. G.; Piech, T.; Rissin, D. M.; Mösl, M.; Peterça, A.; Niederberger, H.-P.; Minnehan, K. A.; Patel, P. P.; Ferrell, E. P.; Meyer, R. E.; Chang, L.; Wilson, D. H.; Fournier, D. R.; Duffy, D. C. Isolation and Detection of Single Molecules on Paramagnetic Beads Using Sequential Fluid Flows in Microfabricated Polymer Array Assemblies. Lab Chip 2012, 12, 977−985. (23) Wilson, D. H.; Rissin, D. M.; Kan, C. W.; Fournier, D. R.; Piech, T.; Campbell, T. G.; Meyer, R. E.; Fishburn, M. W.; Cabrera, C.; Patel, P. P.; Frew, E.; Chen, Y.; Chang, L.; Ferrell, E. P.; von Einem, V.; McGuigan, W.; Reinhardt, M.; Sayer, H.; Vielsack, C.; Duffy, D. C. The Simoa HD-1 Analyzer: A Novel Fully Automated Digital

provides a path toward automated and high throughput fabrication of microwell arrays paving the way toward the commercial application of digital bioassay technology.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b15415. Table containing Bosch silicon DRIE process parameters, a schematic illustration and detailed description of the digital bioassay work flow, fluorescence images and results obtained for the digital detection of BβG, and a fluorescence intensity profile across a 5 μm microwell (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (J.L.). ORCID

Deborah Decrop: 0000-0002-4948-9981 Wouter van der Wijngaart: 0000-0001-8248-6670 Jeroen Lammertyn: 0000-0001-8143-6794 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The research leading to these results has received funding from the Research Foundation−Flanders (FWO G086114N and G080016N), De Vlaamse Liga tegen Kanker (EXM C8744-A), the KU Leuven (OT 13/058, C32/15/005 and IOF KP/12/ 009) and EU project Norosensor (FP7-NMP-2013-SMALL7604244). D.D. is financially supported by the Agency for Innovation by Science and Technology in Flanders (IWT 121615). We would also like to acknowledge Maarten Roeffaers, from the Department of Microbial and molecular systems, Centre for Surface Chemistry and Catalysis, KU Leuven, for the SEM equipment.



REFERENCES

(1) Fan, J.-B.; Gunderson, K. L.; Bibikova, M.; Yeakley, J. M.; Chen, J.; Garcia, E. W.; Lebruska, L. L.; Laurent, M.; Shen, R.; Barker, D. Illumina Universal Bead Arrays. Methods Enzymol. 2006, 410, 57−73. (2) Gunderson, K. L.; Kruglyak, S.; Graige, M. S.; Garcia, F.; Kermani, B. G.; Zhao, C.; Che, D.; Dickinson, T.; Wickham, E.; Bierle, J.; Douchet, D.; Milewski, M.; Yang, R.; Siegmund, C.; Haas, J.; Zhou, L.; Oliphant, A.; Fan, J.-B.; Barnard, S.; Chee, M. S. Decoding Randomly Ordered DNA Arrays. Genome Res. 2004, 14, 870−877. (3) Miller, M. B.; Tang, Y.-W. Basic Concepts of Microarrays and Potential Applications in Clinical Microbiology. Clin. Microbiol. Rev. 2009, 22, 611−633. (4) Fan, J.-B.; Hu, S. X.; Craumer, W. C.; Barker, D. L. BeadArraybased Solutions for Enabling the Promise of Pharmacogenomics. BioTechniques 2005, 39, 583−588. (5) Kuhn, K.; Baker, S. C.; Chudin, E.; Lieu, M.-H; Oeser, S.; Bennett, H.; Rigault, P.; Barker, D.; McDaniel, T. K.; Chee, M. S. A Novel, High-Performance Random Array Platform for Quantitative Gene Expression Profiling. Genome Res. 2004, 14, 2347−2356. (6) Kumar, P. T.; Vriens, K.; Cornaglia, M.; Gijs, M.; Kokalj, T.; Thevissen, K.; Geeraerd, A.; Cammue, B. P. A.; Puers, R.; Lammertyn, J. Digital Microfluidics for Time-Resolved Cytotoxicity Studies on Single Non-Adherent Yeast Cells. Lab Chip 2015, 15, 1852−1860. (7) Schubert, S. M.; Arendt, L. M.; Zhou, W.; Baig, S.; Walter, S. R.; Buchsbaum, R. J.; Kuperwasser, C.; Walt, D. R. Ultra-Sensitive Protein 10425

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426

Research Article

ACS Applied Materials & Interfaces Immunoassay Analyzer with Single-Molecule Sensitivity and Multiplexing. J. Lab. Autom. 2016, 21, 533. (24) Jackman, R. J.; Duffy, D. C.; Cherniavskaya, O.; Whitesides, G. M. Using Elastomeric Membranes as Dry Resists and for Dry Lift-Off. Langmuir 1999, 15, 2973−2984. (25) Carlborg, C. F.; Haraldsson, T.; Ö berg, K.; Malkoch, M.; van der Wijngaart, W. Beyond PDMS: Off-Stoichiometry Thiol-Ene (OSTE) Based Soft Lithography for Rapid Prototyping of Microfluidic Devices. Lab Chip 2011, 11, 3136−3147. (26) Pardon, G.; Saharil, F.; Karlsson, J. M.; Supekar, O.; Carlborg, C. F.; van der Wijngaart, W.; Haraldsson, T. Rapid Mold-Free Manufacturing of Microfluidic Devices with Robust and Spatially Directed Surface Modifications. Microfluid. Nanofluid. 2014, 17, 773− 779. (27) Pardon, G.; Haraldsson, T.; van der Wijngaart, W. Surface Energy Micropattern Inheritance from Mold to Replica. MEMS USA 2014, 96−99. (28) Pardon, G.; Haraldsson, T.; van der Wijngaart, W. Simultaneous Replication of Hydrophilic and Superhydrophobic Micropatterns through Area-Selective Monomers Self-Assembly. Adv. Mater. Interfaces 2016, 3, DOI: 160040410.1002/admi.201600404. (29) Weng, K.-Y.; Wang, M.-Y.; Tsai, P.-H. Planarize the Sidewall Ripples of Silicon Deep Reactive Ion Etching. NSTI-Nanotech Technol. Proc. 2004, 473−476. (30) Zhang, H.; Nie, S.; Etson, C. M.; Wang, R. M.; Walt, D. R. OilSealed Femtoliter Fiber-Optic Arrays for Single Molecule Analysis. Lab Chip 2012, 12, 2229−2239. (31) Sandström, N.; Shafagh, R. Z.; Vastesson, A.; Carlborg, C. F.; van der Wijngaart, W. Reaction Injection Molding and Direct Covalent Bonding of OSTE+ Polymer Microfluidic Devices. J. Micromech. Microeng. 2015, 25, 75002−75014.

10426

DOI: 10.1021/acsami.6b15415 ACS Appl. Mater. Interfaces 2017, 9, 10418−10426