Low melting point agarose as a protection layer in photolithographic patterning of aligned binary proteins

Abstract

A novel photolithography method to build aligned patterns of two different proteins is presented. Chessboard patterns of 125 µm × 125 µm squares are constructed on a silicon dioxide substrate, using standard photoresist chemistries in combination with low-temperature oxygen plasma etching. Low-melting-point agarose (LMPA) is used to protect underlying protein layers and, at the appropriate stage, the digestive enzyme GELase (EPICENTRE) is used to selectively remove the prophylactic LMPA layers. Two antibodies, mouse-IgG and human-IgG, were immobilized and patterned by this procedure. The patterned antibodies maintained the specificity of their antigen-antibody binding, as demonstrated by fluorescence microscopy. In addition, normalized fluorescence intensity profiles illustrate that the patterned proteins layers are uniform (standard deviations below 0.05). Finally, a trypsin activity test was conducted to probe the effect of the patterning protocol on immobilized enzymes; the results imply that this photolithographic process using LMPA as a protection layer preserves 70% of immobilized enzyme activity.

---

1. Introduction

Over the last decade, bio-microelectromechanical systems (Bio-MEMS) and lab-on-a-chip devices have been developed from synergetic applications of microsystem technology and biotechnology.^1,2 The design and fabrication of biologically functional surfaces are important steps in the production of such microdevices and this provides a substantial driving force to develop novel surface nanotechnologies, including protein patterning methods.³ Selectively modified surfaces, with biologically functional components, are essential working elements in biosensors^4,5 and devices for cell manipulation and tissue engineering.^6–8

Two prominent methods for producing patterned surface modifications are photolithography and soft lithography. Soft lithography, also known as micro-contact printing, employs a pre-formed stamp or mold, often made from poly(dimethlysiloxane) (PDMS), to transfer patterns onto a substrate.⁹ While this technology has many promising uses,^9,10 soft lithography is not especially suited to applications where precise alignment is required to position a sequence of surface modifications and the stamping process can result in a loss of biological function in the deposited layer.¹⁰ A three-dimensional, PDMS-based microfluidic system has recently been demonstrated to pattern multiple proteins on the same surface.¹¹ A parylene mask is used to transfer fine features.¹² This is not a compact method and the peel-off process of the parylene mask may limit the resultant protein pattern; for example, the chessboard protein patterns presented in this paper cannot be formed by the PDMS molding and peel-off method.

Standard photolithography, on the other hand, is a well-established technique for the bulk manufacture of integrated circuits. In photolithography, a pattern defined by a photomask is transferred to an overlying layer of photoresist, and from the photoresist to underlying layers. The method provides for high resolution and precise feature alignment, but the conventional solutions used to develop and strip the photoresist are largely incompatible with biological molecules and adversely affect their function.¹³

One approach to circumvent this problem has been reported by Ryan et al.,¹⁴ who designed a photomask that acts as an area-selective filter at different wavelengths of light. Alignment is achieved by photo-cleaving selected regions of a layer of photosensitive molecules immobilized on a gold surface and the method has been used to pattern multiple self-assembled monolayers. Rather sophisticated chemistry is required to implement the process so the method is not yet generically ready for engineering applications.

Still another strategy, and one that we employ here, is to place an intermediate layer between the photoresist and underlying biological species so as to protect their function. Sorribas et al.¹⁵ have used a disaccharide, sucrose, in dehydrated form to protect proteins against oxygen plasma etching. However, sucrose does not provide a robust protection for creating multiple immobilized proteins due to its poor adhesion to protein layers and high solubility in aqueous solutions.¹⁵

In this work, we have used low-melting-point agarose (LMPA) as a protective layer. There are several advantages to this approach. First, LMPA has a low gel point, usually room temperature. LMPA solutions can thus be spin-coated onto immobilized protein layers at temperatures below 37 °C, which prevents protein from denaturing. Second, LMPA adheres well to the protein layer and is very stable in aqueous and organic solutions. LMPA gel thus provides robust and effective protection of underlying proteins during subsequent additions of other protein species to neighboring substrate.¹⁶ Third, feature sizes as fine as 2 µm can be patterned with agarose gels,¹⁶ which suffices for different proteomics, cell to cell interaction and tissue engineering applications. Fourth, GELase, a commercially available substance that contains a unique β-agarose digestive enzyme, breaks LMPA down into water-soluble oligosaccharide fragments. The products of the enzymatic digestion are readily washed away by standard buffer solutions at neutral pH and room temperature (e.g. pH 7.4, 25 °C). This enzymatic degradation of agarose is very specific, so it does not affect the immobilized protein layers and the removal process has a high recovery rate for biological molecules. For example, 100% recovery of DNA from LMPA gel matrices has been reported,¹⁷ indicating that the LMPA/GELase system is compatible with biological compounds.

The presentation of this novel LMPA patterning technique is organized as follows. First the LMPA coating and enzymatic removal processes are characterized. Next, we describe the LMPA photolithography process that we employed to align patterns of two different immobilized antibodies, mouse-IgG and human-IgG, on the same silicon dioxide surface. We demonstrate by fluorescence immuno-labeling that the immobilized antibodies remain active and selectively bind their corresponding conjugate antigens.¹⁸ Finally, the effect of coating and removing the LMPA protective layer was investigated by the immobilization of a proteolytic enzyme, trypsin, on silicon dioxide. The overall process with the LMPA layer was found to preserve about 70% of the activity of the immobilized trypsin.

2. LMPA process characterization

LMPA solutions were prepared by dissolving the LMPA powder (Invitrogen) at 65–70 °C in deionized (DI) water. The LMPA solution forms a gel at a temperature that depends on the LMPA concentration. That transition to gelatinous state was monitored by the viscosity change of the LMPA solution using a glass stirrer in a beaker. The low gelling point of the LMPA, in contrast to other types of agarose, allows uniform spin coating at 35 °C and transition to gel at 25 °C for two commonly used concentrations, 0.5% and 1%. Room-temperature spin coating is desirable because the underlying (immobilized) proteins denature at elevated temperatures. A pre-cleaned silicon dioxide substrate was used to investigate the thickness of the gelled LMPA layer. LMPA-coated substrates were patterned using standard photolithography and low temperature oxygen plasma etching.¹⁹ After stripping the photoresist in acetone, the topology of the gelled LMPA patterns was measured with a profilometer. The thickness of the resultant LMPA gel is shown in Fig. 1 as a function of spin coater RPM for LMPA solution concentrations of 0.5 and 1% (wt/vol). The roughness of the coated LMPA gel is found to be about 5 nm, which is within the resolution of the profilometer. At 1000 rpm and 1% concentration, the gelled LMPA layer is 60 nm thick, which is sufficient to protect most immobilized biological molecules (e.g. antibodies or enzymes).

Fig. 1 LMPA layer thickness as a function of coater spinning speed for two common solution concentrations (0.5 and 1.0%). Coating applied at 35 °C. Gel formation at 25 °C. Roughness within the range of the resolution of the profilometer = 5 nm. Each data point is based on 5 samples.

To investigate gel removal rates, the patterned LMPA was broken down (etched) into water-soluble oligosaccharides with 2 ml of a solution of the agarose enzyme with a concentration of 1 U µl⁻¹ GELase, in 25× reaction buffer at 45 °C, according to the manufacturer (EPICENTRE) protocol. The digestion products were readily carried away by the buffer solution. The thickness of the gelled LMPA layer is shown in Fig. 2 as a function of etching time. Removal rates were about 1.5 nm/min for gels formed from 0.5 and 1.0% LMPA solutions. A 90 nm-thick gel, the thickest in our test, was removed enzymatically within 1 hour.

Fig. 2 Isothermal etching of LMPA with GELase. Data shown are LMPA layer thickness [nm] versus etching time [min] for two separate concentrations of agarose solution (0.5%, 1.0%). Etching temperature at 45 °C; GELase concentration is 1 U µl⁻¹. Roughness within the range of the resolution of the profilometer = 5 nm. Each data point is based on 5 samples.

3. Heterogeneous protein patterned surfaces

Except where noted, Sigma Aldrich was the source of chemicals used for these studies. The starting substrates were 2″ <110> N-type silicon wafers. A layer of 3000 Å thick oxide was thermally grown on the silicon wafers. The surface of each silicon substrate was first cleaned with a 1 ∶ 1 ∶ 5 ratio of 29% ammonium hydroxide, 30% hydrogen peroxide, and DI water for 15 minutes at 80 °C in order to remove the organic contaminants. The initial cleaning was followed by surface treatment with a 1 ∶ 1 ∶ 6 ratio of 50% hydrogen chloride, 30% hydrogen peroxide, and DI water for 15 minutes at 80 °C to maximize the hydroxyl groups on the surface. The oxide surfaces were then rinsed with DI water and dried with nitrogen gas.¹⁵

After standard cleaning and surface treatment, the hydroxyl groups of the oxide surface are silanated in a 1% (vol/vol) 3-aminopropyltriethoxysilane (APTES)-acetone solution for 15 minutes under magnetic stirring (Fig. 3(a)). The substrates were rinsed carefully under acetone and dried in the furnace at 50 °C for 15 minutes. Next, the APTES surfaces were activated with 2% (vol/vol) glutaraldehyde in 1 × phosphate buffer saline (PBS) at pH 7.4 for 2 hours, promoting a Schiff-base reaction between the amine and aldehyde groups. Recombinant protein A from Staphylococcus aureus (Zymed Lab Inc.), at a concentration of 550 µg ml⁻¹ in 1 × PBS, was then incubated on the activated surface at 4 °C for 3 hours (Fig. 3(b)). Mouse immunoglobulin G (IgG) from Serum, 100 µg ml⁻¹, was incubated on the protein A (Fig. 3c), thus completing the first immunoassay process. Next, 1%-LMPA solution was spin coated on the substrates at 1000 rpm for protection of the first immuno-active surface during the patterning process (Fig. 4(a)). The LMPA coating gels at room temperature after 5 minutes. Photoresist (AZ 1512) was then spin coated, at 4000 rpm, on the substrate. A soft bake procedure was carried out at 45 °C for 20 minutes in a furnace. Following an exposure to UV light for 6 seconds using a contact aligner (AB-Manufacturing), the pattern was developed in MIF 300 solution for about 20 seconds; the substrate temperature had not exceeded the room temperature of 25 °C throughout the exposure and development process. The photolithography step was completed by hard-baking the substrates in an oven furnace at a temperature of 45 °C for 30 minutes (Fig. 4(b)). Low temperature oxygen plasma etching was used for the final pattern transfer from the photoresist to the first immobilized protein layer (mouse IgG). The substrates were then etched in an oxygen plasma etcher (PLASMA-PREEN II-862) with discharge conditions of 300 W and 28 kPa. The process was carried out for 3 minutes in an interval of 30 seconds to prevent excessive heating, which can denature the immobilized protein. The temperature in the reaction chamber measured immediately after each interval was found to be about 40 °C. After stripping the photoresist by acetone, with the LMPA gel protecting the first immobilized antibody pattern (Fig. 4(c)), the exposed silicon dioxide area was subjected to the immunoassay described earlier (Fig. 3(a)–(c)). In this process, though, human IgG molecules from serum, in a concentration of 100 µg ml⁻¹, were incubated on the surface not protected by LMPA. Finally, the protective LMPA layer was removed overnight with 1 U µl⁻¹ GELase in 25 × reaction buffer at 45 °C. Following the standard GELase protocol, the enzymatic reaction is optimized at a temperature between 42 °C to 47 °C. The reaction products were washed away with 1 × PBS. As a result of the process, two different antibodies were immobilized on the oxide surface with desired heterogeneous patterns, defined by a photomask (Fig. 4(d)).

Fig. 3 Schematic representation of the immunoassay on a silicon dioxide surface: (a) silanization of the surface with APTES, (b) incubation of protein A layer, and (c) incubation of mouse IgG antibody or human IgG antibody.

Fig. 4 Cross sections for patterning two different biologically-functional layers on the same surface: (a) first immunoassay for mouse IgG with LMPA coating as a protection layer, (b) photoresist patterning, (c) oxygen plasma etching followed with photoresist strip, (d) second immunoassay for human IgG and removal of protective LMPA pattern using GELase.

4. Antibody-antigen binding with immobilized IgG

Conjugate antibodies, tagged with fluorescent dyes, were used to examine the effect of the patterning protocol on the antibody-antigen binding of immobilized mouse-IgG and human-IgG. CY3-anti-mouse IgG (H + L) and FITC-anti-human IgG (H + L) (Jackson ImmunoResearch), at a concentration of 100 µg ml⁻¹, in 1% bovine serum albumin (BSA) 1 × PBS were incubated on the patterned substrates in the dark for 30 minutes. Note that the conjugate antibodies used here have been exposed to IgG from other species and show minimum cross-reactivity. The patterned substrate used in the study was formed in the manner shown in Fig. 4(a)–(d), so immobilized mouse-IgG regions were assembled first and protected by an LMPA layer.

The patterned substrate was excited in a fluorescence microscope (Nikon, E80I), initially with a CY3 filter. The region of immobilized mouse IgG was found to emit red radiation, as shown in Fig. 5(a). Subsequently, the substrate was excited with an FITC filter. In this case, the human-IgG regions were seen to emit green radiation, as is evident in Fig. 5(b). Intensity profiles were measured along line segments AA′ and BB′ of Figure 8; the normalized results of these measurements are shown in Fig. 6(a)–(b). The spatial distribution of the intensity profile for CY3-anti-mouse IgG region shows a standard derivation of 0.0487. The standard deviation of the FITC-anti-human IgG profile was 0.0440. Thus the immobilized antibody that was protected under the LMPA layer exhibits only a 10% increase in non-uniformity.

Fig. 5 Fluorescence excitation of protein pattern: (a) under Cy3 filter excitation, (b) under FITC filter excitation.

Fig. 6 Intensity profile of the proteins pattern shown in Fig. 5: (a) for the cross section (A–A′) with excitation under a CY3 filter, and (b) for the cross section (B–B′) with excitation under a FITC filter.

In summary, these results from the LMPA protocol for protein patterning indicate that each of the regions is uniform. Moreover, the patterned species retain their specific biological function; in terms of antigen-antibody specificity, no significant distinctions are evident between the first and second immobilized antibody.

5. Trypsin activity test

A trypsin activity test was used to probe the loss of biological activity that might result from surface immobilization followed by LMPA coating and etching. The enzyme trypsin catalyzes the hydrolysis of the amide bond in the reagent N-benzoyl-DL-arginine-4-nitronanilide (DL-BAPNA) to produce p-nitroaniline. This enzymatic reaction can be monitored with a spectrophotometer (UV160, UV/VIS visible recording spectrophotometer, Shimadzu) by measuring the absorbance of the reaction solution at 405 nm. A standard absorbance curve, showing the temporal evolution of the reaction, is first obtained by a BAPNA solution, which was prepared by dissolving 0.5 mM BAPNA in a buffer solution of 50 mM Tris, HCl, 20 mM CaCl₂ at pH 8.4.²⁰ 2% N-Dimethylformamide (DMF) was used to dissolve the BAPNA powder in advance to enhance its solubility in the buffer. Bovine pancreatic trypsin (dialyzed and lyophilized) was then introduced to the buffer at a free concentration of 0.01 g ml⁻¹. Trypsin was immobilized on silicon dioxide surfaces by the procedure given in Section 3. In this case, trypsin (1 mg ml⁻¹ in 1× PBS, pH 7.4) was substituted for protein A and IgG. Oxide surfaces derivatized with immobilized trypsin were then incubated in the BAPNA solution. The temporal evolution of the 405 nm-absorbance is plotted in Fig. 7. The activity of the immobilized trypsin with respect to BAPNA can be calculated in relation to the absorbance curve for known concentration of free trypsin in the standard absorbance curve. Using the molecular mass and molecular spherical diameter of trypsin,²¹ the density of active immobilized trypsin without LMPA process was estimated to be 38.3 ng cm⁻²; decrease in enzymatic activity after LMPA coating and digestion was correlated to give an apparent density of 25.9 ng cm⁻². These results imply that the protocol for LPMA coating and digestion has diminished the enzymatic activity by about 30%.

Fig. 7 Visible light absorbance (405 nm) versus time [h] on surfaces coated with trypsin and incubated in a calibrated BAPNA solution. Each data point is based on 5 samples.

6. Conclusions

A robust photolithography technique has been developed for direct patterning of two different proteins in immediate proximity on a silicon dioxide surface. The technique makes use of LMPA to protect underlying protein layers and oxygen-plasma etching to expose native silicon dioxide substrate. Protective LMPA is removed with the agarase enzyme GELase. LMPA spin coating and removal processes were characterized. Several physical properties make LMPA suitable for protecting immobilized bio-species. Its low gelling point allows uniform spreading across the substrates without denaturing the proteins. The coating thickness is in the order of tens of nanometres, which suffices to protect the immobilized bio-molecular species used in this work. It also allows reasonable bio-etching time, typically within several hours, by GELase. Two different antibodies were patterned on silicon dioxide surfaces to demonstrate this technique for immobilizing multiple proteins. Following the first immunoassay, the immobilized antibody (mouse IgG) was protected by LMPA gel. After photoresist patterning and low temperature oxygen plasma etching, the second antibody (human IgG) was immobilized on exposed oxide regions. Fluorescence microscopy revealed that the patterned antibodies were active in as much as they specifically bind their conjugates, anti-mouse IgG and anti-human IgG, respectively. Both antibody patterns also exhibited reasonable uniformity, with a standard deviation of less than 0.05 in their normalized intensity profile. The recovery of enzyme activity after the LMPA protection and removal was accessed indirectly by performing the same assay with the enzyme trypsin. The apparent density of the active trypsin immobilized substrates was measured by the rate of hydrolysis of DL-BAPNA. A comparison of the relative enzymatic activity of immobilized trypsin with and without the coating process allowed an estimation of recovery of activity of 70%.

Acknowledgements

This work was supported by the Multidisciplinary Projects in Life Sciences Initiative, BIO5 Institute, the University of Arizona.

References

1. D. R. Reyes, D. Iossifidis, P. A. Auroux and A. Manz, Micrototal analysis systems. 1. introduction, theory, and technology, Anal. Chem., 2002, 74(12), 2623–2636.
2. P. A. Auroux, D. Iossifidis, D. R. Reyes and A. Manz, Micrototal analysis systems. 2. Analytical standard operations and application, Anal. Chem., 2002, 74(12), 2637–2652.
3. A. S. Blawas and W. M. Reichert, Protein patterning, Biomaterials, 1998, 19, 595–609.
4. E. Katz and I. Willner, Probing biomolecular interactions at conductive and semiconductive surfaces by impedance spectroscopy: routes to impedimetric immunosensors, DNA-sensors, and enzyme biosensors, Electroanalysis, 2003, 15, 913–47.
5. N. V. Zaytseva, V. N. Goral, R. A. Montagna and A. J. Baeumner, Development of a microfluidic biosensor module for pathogen detection, Lab Chip, 2005, 5, 805–811.
6. H. Andersson and A. Van den Berg, Microfabrication and microfluidics for tissue engineering: state of the art and future opportunities, Lab Chip, 2004, 4, 98–103.
7. X. Jiang, D. A. Bruzewicz, A. P. Wong, M. Piel and G. M. Whitesides, Directing cell migration with asymmetric micropatterns, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 975–978.
8. N. Wang, E. Ostuni, G. M. Whitesides and D. E. Ingber, Micropatterning tractional forces in living cells, Cell Motil. Cytoskeleton, 2002, 52, 97–106.
9. R. S. Kane, S. Takayama, E. Ostuni, D. E. Ingber and G. M. Whitesides, Patterning Proteins and Cells using Soft Lithography, Biomaterials, 1999, 20, 2363–2376.
10. J. Tien, Y. Xia and G. M. Whitesides, Microcontact Printing of SAMs, in Thin Films, ed. A. Ulman, Academic Press, New York, 1998, vol. 24, pp. 227–253.
11. D. T. Chiu, N. L. Jeon, S. Huang, R. Kane, C. J. Wargo, I. S. Choi, D. E. Ingber and G. M. Whitesides, Patterned deposition of cells and proteins onto surfaces by using three-dimensional microfluidic systems, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 2408–2413.
12. K. Atsuta, H. Suzuki and S. Takeuchi, Micro patterning of active proteins with perforated PDMS sheets (PDMS sieve), Lab Chip, 2004, 4, 333–336.
13. A. W. Flounders, D. L. Brandon and A. H. Bates, Patterning of immobilized antibody layers via photolithography and oxygen plasma exposure, Biosens. Bioelectron., 1997, 12, 447–456.
14. D. Ryan, B. A. Parviz, V. Linder, V. Semetey, S. K. Sia, J. Su, M. Mrksich and G. M. Whitesides, Patterning multiple aligned self-assembled monolayers using light, Langmuir, 2004, 20, 9080–9088.
15. H. Sorribas, C. Padeste and L. Tiefenauer, Photolithographic generation of protein micropatterns for neuron culture applications, Biomaterials, 2002, 23, 893–900.
16. M. Mayer, J. Yang, I. Gitlin, D. H. Gracias and G. M. Whitesides, Micropatterned agarose gels for stamping arrays of proteins and gradients of proteins, Proteomics, 2004, 4, 2366–2376.
17. http://www.epibio.com/item.asp?id=297.
18. L. M. Lee, R. L. Heimark, J. C. Baygents and Y. Zohar, Self-aligned immobilization of proteins utilizing PEG patterns, Nanotechnology, 2006, 17, S29–S33.
19. L. M. Lee, R. L. Heimark, J. C. Baygents and Y. Zohar, Photolithographic patterning of multi proteins on a single chip using low-melting-point agarose as a protection layer, Proceedings of the 9th International Conference on Miniaturized Chemical and BioChemical Analysis Systems, Boston, USA, 2005, pp. 764–766.
20. M. Nouaimi, K. Moschel and H. Bisswanger, Immobilization of trypsin on polyester fleece via different spacers, Enzyme Microb. Technol., 2001, 29, 567–574.
21. J. F. Diaz and K. J. Balkus, Jr., Enzyme immobilization in MCM-41 molecular sieve, J. Mol. Catal. B: Enzym., 1996, 2, 115–126.

---