Micrometer and Nanometer Scale Photopatterning of Proteins on Glass Surfaces by Photo-degradation of Films Formed from Oligo(Ethylene Glycol) Terminated Silanes
- First Online:
- Cite this article as:
- Tizazu, G., el Zubir, O., Patole, S. et al. Biointerphases (2012) 7: 1. doi:10.1007/s13758-012-0054-6
Exposure of films formed by the adsorption of oligo(ethylene glycol) (OEG) functionalized trichlorosilanes on glass to UV light from a frequency-doubled argon ion laser (244 nm) causes photodegradation of the OEG chain. Although the rate of degradation is substantially slower than for monolayers of OEG terminated thiolates on gold, it is nevertheless possible to form micrometer-scale patterns by elective adsorption of streptavidin to exposed regions. A low density of aldehyde functional groups is produced, and this enables derivatization with nitrilotriacetic acid via an amine linker. Complexation with nickel enables the site-specific immobilization of histidine-tagged yellow and green fluorescent proteins. Nanometer-scale patterns may be fabricated using a Lloyd’s mirror interferometer, with a sample and mirror set at right angles to each other. At low exposures, partial degradation of the OEG chains does not remove the protein-resistance of the surface, even though friction force microscopy reveals the formation of patterns. At an exposure of ca. 18 J cm−2, the modified regions became adhesive to proteins in a narrow region ca. 30 nm (λ/8) wide. As the exposure is increased further the lines quickly broaden to ca. 90 nm. Adjustment of the angle between the sample and mirror enables the fabrication of lines of His-tagged green fluorescent protein at a period of 340 nm that could be resolved using a confocal microscope.
Protein patterning is important in many applications, from fundamental studies of cellular interactions [1, 2, 3, 4, 5, 6, 7], molecular recognition , and other phenomena, to the development of biosensors for the detection and diagnosis of disease [9, 10]. However, the control of the spatial organization of proteins remains challenging. The greater part of the literature on bionanofabrication has addressed the patterning of oligonucleotides [11, 12, 13], and there is much less work that has addressed the fabrication of protein nanostructures [14, 15, 16, 17, 18, 19, 20, 21, 22, 23]. Similarly, while there is a mature market for chip-based DNA analysis systems, the protein chip remains at a very early stage of development. Undoubtedly, the significant difficulties associated with the formation of protein patterns are an important factor in both cases. Proteins can explore large conformational spaces, so they adhere strongly to most surfaces. This renders both spatial and orientational control of adsorption difficult because, proteins may adhere to regions, where they are not required and because, they may become kinetically trapped in an unfavorable orientation, even where a linking group is available at the surface to promote site-specific attachment. Moreover, proteins can undergo facile conformational change after adsorption to surfaces, rendering control of conformation difficult.
The most important challenge in organizing proteins at interfaces is the control of non-specific adsorption. The most widely reported approaches employ poly(ethylene glycol)  and oligomeric ethylene glycol [1, 2, 25, 26, 27], and related materials such as oligo(ethylene glycol) (OEG) methacrylate brushes [28, 29, 30] and plasma-polymerized tetraglyme . However, there is a growing group of other materials, such as zwitterionic poly(methacroyl phosphorylcholine) brushes [32, 33, 34, 35], that have also been found to exhibit strong resistance to non-specific adsorption. Self-assembled monolayers of OEG terminated alkylthiolates adsorbed on gold surfaces are exceptionally resistant to protein adsorption, and have been used previously to fabricate protein patterns [1, 2, 3]. They are conveniently patterned by a variety of methods, including micro-contact printing [1, 2] and, at smaller length-scales, dip-pen nanolithography , near-field lithography [18, 20, 36, 37], electron beam lithography [38, 39, 40], and nanoimprinting . However, gold has a marked propensity to quench fluorescence, and attenuates transmitted light. Given the significance of optical methods for the characterization of biological systems, these are significant limitations. Glass is a very important substrate for biologists, but comparatively speaking, much less attention has been given to the precise patterning of proteins on glass surfaces .
Oligo(ethylene glycol) materials have been demonstrated previously to be susceptible to photodegradation at 244 nm. Ducker et al.  found that triethylene glycol functionalized SAMs of alkylthiolates on gold were rapidly degraded by UV exposure to yield protein-binding aldehyde functional groups. Subsequently, plasma-polymerized tetraglyme , poly(oligoethylene methacrylate) polymer brushes  and monolayers [20, 43] have been modified by photodegradation, to facilitate protein patterning. In this paper, we present a simple method for patterning proteins on glass, based on spatially selective photodegradation of a protein-resistant film formed by the adsorption of 2-[methoxy(polyethylenoxy)propyl] trichlorosilanes to yield protein-binding regions. Films of this adsorbate have been patterned previously using conventional photolithography involving exposure and development of a photoresist ; here we demonstrate direct patterning by photodegradation and also the fabrication of nanometer scale patterns by interferometric lithography. The use of processes that do not require photoresist is likely to be important in the development of strategies for the fabrication of multicomponent protein nanopatterns redeposition and development of photoresist in successive steps would damage proteins attached to the surface. Thus, there is a need for simple methods that involve direct conversion of protein-resistant materials to protein-binding ones.
At the nanometer scale, protein patterning remains challenging. Moreover, there is also an urgent need for methods that enable control of biomolecular organization at high resolution over macroscopic areas [15, 17]. Previously, successful approaches to patterning over large areas have included the use of self-assembling arrays of polymer-stabilized metal nanoparticles , and micro-contact printing . Here, we report an alternative approach that is compatible with the rapid patterning of large areas. In interferometric lithography (IL) [45, 46, 47, 48], two or more coherent laser beams interfere to produce a periodic intensity pattern that is used to expose a photosensitive material. The interference pattern has a period λ/2nsinθ, where λ is the wavelength of the excitation beam, n is the refractive index of the medium and 2θ is the angle between the two interfering beams. The exposure process is rapid and by changing the geometry and number of interfering laser beams, it is possible to change quickly and readily the nature of the resulting interference pattern. Macroscopic areas may be exposed in a maskless process using, other than a UV laser, only very simple apparatus. The method is potentially very attractive for many laboratories that do not possess extensive infrastructure for nanofabrication. As an example, we demonstrate herein the fabrication of 30 nm lines of streptavidin at a period of 180 nm over a region ca. 1 cm2 in area.
All glassware and substrates were cleaned before using piranha solution [a mixture of 30 % H2O2 and 98 % concentration sulfuric acid (both purchased from Fisher Scientific, Loughborough, UK)] in the ratio 3:7 for 45 min. (Caution: piranha solution is a strong oxidizing agent and has been known to detonate spontaneously upon contact with organic material; it should be handled with extreme caution). After treatment with piranha solution, glassware, and substrates were rinsed with copious amounts of deionized water (Elga Pure Nanopore, 18.2 MΩ) and dried in an oven. Silane films were formed by immersing clean substrates in a 1 mmol dm−3 solution of 2-[methoxy(polyethylenoxy)propyl] trichlorosilane (OEG-silane), CH3O(C2H4O)6–9(CH2)3Cl3Si (Gelest, Maidstone, UK), henceforth OEG-silane, in dry toluene under nitrogen.
The gene sequence of yellow fluorescent protein (YFP) was amplified by PCR from pCS2-Venus vector (a kind gift from Dr. Atsushi Miyawaki, RIKEN Brain Science Institute, Japan). The resulting Nde I/Bam HI fragment was cloned into a pET14b expression vector (Novagen). Introducing the combined F64L, S65T, V68L, S72A, M153T, V163A, S175G, and A206K mutations into the YFP gene resulted in enhanced green fluorescent protein (GFP) gene . Both His6-YFP and His6-GFP proteins were produced by heterologous expression in E. coli (BL21); cells were grown to an O.D680 of 0.6 at 37 °C then induced using IPTG (0.4 mm) for 12 h at 25 °C. Pelleted cells (19,000×g/20 min) were lysed by sonication, and the resulting lysate was clarified by a further spin (33,000×g/30 min). Both His-tagged fluorescent proteins were purified to homogeneity from clarified lysate using a Chelating Sepharose Fast Flow Ni–NTA gravity flow column (GE Healthcare) as detailed in the manufacturer’s instructions. Protein purity was assessed by gel electrophoresis (SDS–PAGE).
Alexa Fluor 647-conjugated streptavidin and FluoSpheres® NeutrAvidin® labeled particles, 0.04 μm, yellow-green fluorescent (505/515) 1 % were purchased from Invitrogen, Paisley, UK. Both were dissolved in 0.1 M PBS at a concentration of 10 μg mL−1. The samples were immersed in these solutions for 2 h, washed with PBS solution and deionized water and gently dried with nitrogen, before being characterized using confocal microscopy.
Photopatterning was carried out using a Coherent Innova 300C FreD frequency-doubled argon ion laser (Coherent UK, Ely, UK), emitting at 244 nm with a maximum power of 100 mW. The beam was expanded using a divergent lens to irradiate an area ca. 0.8 cm in diameter. Micrometer-scale patterns were fabricated by exposing the sample through an electron microscope grid (Agar, Cambridge, UK).
Surface analysis by X-ray photoelectron spectroscopy (XPS) was carried out using a Kratos Axis Ultra X-ray photoelectron spectrometer (Kratos, Manchester, UK), equipped with a monochromatic Al-Kα X-ray source operating with a base pressure of 10−10 mbar. For high resolution scans, an electron energy analyzer pass energy of 80 eV was used. Data were processed using CasaXPS software (http://www.casaxps.com/). All binding energies were referenced to the main hydrocarbon peak, set at a binding energy of 285 eV. Peaks were fitted with a linear background, using symmetrical Voigt type functions, with contributions from both Gaussian and Lorentzian profiles (typically in the proportion 90 % Gaussian to 10 % Lorentzian). The full width at half maximum (FWHM) values of the peaks within a given spectrum were constrained to be identical and were found to be 1.31 ± 0.02 eV. Contact mode topographical and friction imaging and tapping mode characterization were performed using a Multimode atomic force microscope with a Nanoscope IIIA controller (Veeco, Santa Barbara, USA). Silicon nitride probes with nominal spring constants of 0.06 and 0.12 Nm−1, and tip radii of between 20 and 60 nm (Veeco), were used for contact mode measurements. For tapping mode, Si probes (Veeco, TESP) were used.
Fluorescence images of the His-tagged GFP and YFP patterns were obtained on a Delta Vision RT Microscope (Applied Precision, Issaquah, WA). The samples were placed on a microscope slides followed by a mount in Citifluor as an antifade reagent (glycerol-PBS solution, AF1) (Citifluor Ltd., London, United Kingdom). A 60× UPLSAPO magnification oil immersion lens with a numerical aperture of 1.42 was used for imaging. A small drop of immersion oil (Type F, Olympus) was placed on the slide in the center of the illuminated area. A bandpass filter (FITC Ex-490/20) was used to excite the GFP on the surface and the fluorescence was collected with band pass emission filter (FITC Em-528/38). The fluorescence images were analyzed using Image J version 1.46J image processing software.
3 Results and Discussion
3.1 UV Photodegradation
Following exposure to UV light, the most noticeable change in the C1s spectrum was a substantial decrease in the size of the ether component. This was attributed to degradation of the OEG group. The total area of the C1s peak envelope was found to have reduced, too, consistent with the removal of material from the surface. A very small peak was observed at ca. 288 eV, corresponding to an aldehydic carbon atom. These changes are consistent with photodegradation of the OEG group, and are qualitatively similar to those observed by Ducker et al.  in studies of OEG-terminated alkylthiolate SAMs on gold.
These data, taken together, suggest that the degradation process for the OEG-silane films is qualitatively similar to that which occurs for OEG-terminated SAMs on gold, but that the rate of change is significantly smaller for the OEG-silane films. The reason for this is not clear. It has previously been suggested that hot electron formation at gold surfaces triggers the photo-oxidation of alkylthiolate SAMs during exposure to UV light [51, 52]; it is possible that hot electrons also accelerate degradative processes in the OEG chain. An alternative explanation is that the different optical properties of silicon dioxide, compared to gold, lead to an altered rate of modification.
3.2 Surface Patterning
Figure 3b shows a friction force microscopy (FFM) image acquired simultaneously with the topographical image in Fig. 3a. The exposed regions (squares) exhibit brighter contrast than the masked regions (bars), indicating that photodegradation causes an increase in the friction force measured by the probe and hence a larger coefficient of friction. The frictional interaction between the probe and the surface results from both adhesive contributions (shearing) and conformational changes induced by the interaction of the AFM probe with the surface under loading (molecular plowing). Contributions to friction from both shearing and plowing are likely to be increased following degradation of the surface, the creation of new oxygen-containing functional groups will lead to increased adhesion, and the degradation of the OEG chains will increase the degree of variation in chain length thus creating a less closely packed surface expected to deform more easily under loading. Both mechanisms are expected to lead to an increase in energy dissipation at the tip-sample contact and an increase in the coefficient of friction. However, it is notable that the contrast difference between the exposed and masked regions in Fig. 3 is quite clear, under conditions that yield only modest changes in the surface composition as determined by XPS. This suggests that the dominant effect on the friction force may be the change in molecular organization, rather than the change in composition (i.e., that it may result largely from the change in the relative contribution from shearing following photodegradation).
Photodegradation of the OEG chains leads to the introduction of aldehyde groups, albeit at low density. These may be exploited to enable site-specific binding of proteins. Following exposure, samples were immersed in a solution of aminobutyl nitrilotriacetic acid (AB-NTA). The amine group of the aminobutyl linker reacts with aldehydes at pH 5 to form an imine bond. Subsequent complexation of NTA with Ni2+ yields a surface that can bind histidine-labeled proteins strongly. Figure 5d shows a tapping mode AFM topographical image of a sample treated in this way, following incubation in a solution of His-tagged YFP. The protein has attached densely to the exposed regions (squares), but not to the areas that were masked during exposure (bars). The strong intrinsic fluorescence of YFP enables the patterns to be characterized by confocal microscopy. Fluorescence images (Fig. 5e, f) exhibited bright fluorescence from exposed regions, and dark contrast in the masked areas, confirming the spatial selectivity of the adsorption of the protein.
Figure 6c shows a tapping mode AFM topographical image of a sample similar to that shown in Fig. 6a, b following immersion in a solution of streptavidin (1 μg mL−1) in phosphate buffered saline for 30 min. Even at comparatively low magnification, fine lines of protein may be observed running vertically down the imaged region. Small bright features observed to be distributed randomly across the sample surface are attributed to salt crystals (samples exposed to buffer solutions often exhibit small amounts of particulate matter, likely due to deposition of ions from the buffer). Figure 6d shows the region marked with a box in Fig. 6c at higher magnification. The lines of adsorbed protein are much more clearly evident. The FWHM of the protein lines was only 30 nm, ca. λ/8 and substantially smaller than the period and the FWHM of the interference pattern. A small number of dark features are observed in Fig. 6d; these are attributed to defects in the siloxane film.
The observation of such narrow protein lines is surprising, given the comparatively broad features that were observed by FFM prior to immersion of the sample in the protein solution. FFM measures changes in surface friction that are probably attributable to changes in the mode of adsorbate conformational deformation under loading through the AFM probe; the FFM images do not measure the adhesiveness of the surface toward protein adsorption. The most likely explanation for the observation of such narrow protein features in Fig. 6d is that while the interferometric exposure causes degradation of the OEG terminal groups over a linear region area some 200 nm wide (and macroscopic in length), the greater part of the exposed area remains resistant to protein adsorption, probably, because only a small number of OEG groups have been removed. The low degree of order in the siloxane film, and the comparatively lengthy OEG chain used here (6–9 ethylene glycol monomers), confer sufficient conformational freedom for intact chains to cover defects induced by photodegradation and protein-binding aldehyde functional groups, except in the areas that have been exposed to the highest intensities of UV irradiation, close to the maxima in the interference pattern. We conclude that there is a threshold dose above which the OEG-silane film becomes adhesive toward proteins, conferring an additional degree of control in the nanopatterning process and enabling the fabrication of structures significantly smaller than one half the period of the interference pattern.
An increase in the exposure time from 6 to 10 min (18–30 J cm−2) yielded an increase in the FWHM of the streptavidin structures from 30 to 90 nm (Fig. 6e). The substantial increase in line width accompanying this modest increase in exposure supports the hypothesis above, that an effective non-linearity is produced in the process by the existence of a “threshold” density of ethylene glycol units at which protein-resistance is lost. From Fig. 6c, d, we conclude that much of the surface has received an exposure below that required to exceed the threshold, while in Fig. 6e a much larger fraction of the surface has received adequate exposure for protein-resistance to be lost.
Finally, nanopatterning was also combined with site-specific immobilization. In order to enable subsequent characterization of the samples using fluorescence microscopy, the angle between the sample and mirror was adjusted to 20°, increasing the period to 340 nm. After exposure to a dose of 58 J cm−2, the sample was derivatized with AB-NTA, and complexed with Ni2+, following the same procedure used to prepare the samples shown in Fig. 5d–f. The sample was then incubated in a solution of His-tagged GFP for 20 h. Despite the small period, the nanoscopic lines are clearly evident in Fig. 6f. The protein features appear broadened, because the line width (ca. 170 nm) is smaller than the resolution limit of the microscope. However, the contrast difference is clear.
Previously, interferometric methods were applied without success to the patterning of poly(oligoethylene methacrylate) polymer brushes by UV degradation. The success in the present case is attributed to the much thinner resist used. However, the slower rate of modification may also be important in the present case, because the more rapidly modified polymer brushes are apt to become more easily over-exposed, leading to broader protein-adhesive bands.
The principal drawback to the methodology described here is the comparatively slow rate of modification of the siloxane film, as reflected, for example, in the XPS data in Fig. 2. Moreover, to yield the narrowest protein structures, the exposure time must be finely judged to avoid either causing too little modification for extensive protein adsorption or causing modification over a broader region (ca. 90 nm for sample and mirror set at right-angles). However, the simplicity of the process is very attractive, and the capacity to fabricate protein nanopatterns over a macroscopic area, on a convenient substrate such as glass, is extremely attractive for fundamental studies of biological behavior at interfaces. The protein-resistance of the unmodified OEG-silane films seems to be extremely good, yielding extremely clear fluorescence contrast even for a protein like YFP, with a high level of intrinsic fluorescence. The simple process described here offers significant scope for adaptation in future work to enable the construction of more complex nanostructured assemblies of biomolecules.
Exposure of films formed by the adsorption of OEG terminated silanes onto glass may be photodegraded by exposure to UV light with a wavelength of 244 nm. The ether component in the C1s spectrum and the total amount of carbon in the film both decrease with exposure. A small aldehyde peak is observed in the C1s region after UV modification. The coefficient of friction, measured by FFM, was found to increase following UV exposure. The rate of change is significantly slower than is observed for OEG terminated monolayers on gold. However, patterning may be accomplished easily by exposing samples to UV light through a mask. Fluorescence microscopy data demonstrate that proteins adsorb non-specifically to exposed regions, but not to unmodified regions providing a simple one-step route to the creation of protein micropatterns. Attachment of NTA, by imine bond formation from an amine linker to the exposed regions, enables site-specific attachment of His-tagged proteins. Using a Lloyd’s mirror interferometer, it was possible to fabricate nanopatterns over an area 8 mm in diameter with a high degree of perfection in a simple, rapid process. As an example, 30-nm wide (λ/8) lines of streptavidin were formed by immersion of the samples in a solution of the protein.
GT, DGL, and GJL thank EPSRC (grants EP/D064767/1 and GR/C523857/1) for financial support. GJL, SP, and CNH gratefully acknowledge financial support from the Biotechnology and Biological Sciences Research Council (BBSRC). OEZ thanks the Libyan Government for a Research Scholarship. GPL thanks the Office of Naval Research (Grant no. N00014-10-1-0907) and the National Science Foundation, Research Triangle MRSEC (Grant no. DMR-1121107) for financial support. The Wellcome Trust (Grant Nos. GR077544AIA and WT093134MA) is thanked for support of the Sheffield Light Microscopy Facility.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
This article is published under license to BioMed Central Ltd. Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.