Sol–gel films for DNA microarray applications

Materials Letters 60 (2006) 1833 – 1838 www.elsevier.com/locate/matlet Sol–gel films for DNA microarray applications K. Saal a,b,⁎, T. Tätte a,b , I. Tulp b , I. Kink a , A. Kurg c , U. Mäeorg b , A. Rinken b , A. Lõhmus a a Institute of Physics, University of Tartu, 142 Riia St., 51014 Tartu, Estonia b Institute of Organic and Bioorganic Chemistry, University of Tartu, 2 Jakobi St., 51014 Tartu, Estonia c Institute of Molecular and Cell Biology, Estonian Biocentre, 23 Riia St., 51010 Tartu, Estonia Received 8 September 2005; accepted 9 December 2005 Available online 4 January 2006 Abstract Sol–gel derived (3-aminopropyl)trimethoxysilane–tetramethoxysilane ((CH3O)3SiCH2CH2CH2NH2–Si(OCH3)4) hybrid films are shown to have properties that make the films suitable for DNA microarray applications. The essential characteristics of the films are discussed on the basis of binding of aminated 25-mer oligonucleotide DNA to the films via 1,4-phenylenediisothiocyanate linkering. The binding of DNA onto the films is shown to depend on films' composition having an optimum where the binding is substantially superior compared to commercial analogues. The relevant properties of the films are characterized by AFM, FTIR and MALDI-TOF MS measurements. © 2005 Elsevier B.V. All rights reserved. Keywords: Sol–gel preparation; Thin films; Surfaces; DNA microarrays 1. Introduction its molecule [2] and terminal functionality [2,3], and also on trace quantities of water in reaction medium and on the DNA microarrays are devices displaying specific oligonu- substrate [4,5]. In search for a robust procedure for cleotides or longer DNA fragments attached in two-dimensional reproducible fabrication of silane coatings we have proposed order onto activated solid surface [1]. DNA microarrays permit an alternative silanization technique that was less dependent on the analysis of gene expression and DNA sequence variation in humidity and nature of the substrate and enabled to prepare a massively parallel format. The physical and chemical nature of homogeneous and smooth aminosiloxane surfaces [6]. This the substrate on which the reactions are performed is one of the was achieved by dip coating of mica substrate with partially key factors influencing the quality and reproducibility of the pre-polymerized (3-aminopropyl)trimethoxysilane (APTMS) results. Among many different types of substrates for DNA sol, followed by its gelation in humid air. Still, the films did microarray analysis, the most common chemical treatments not feature prolonged stability in water, which was probably provide chemically reactive amine or aldehyde groups prepared caused by low rate of cross-linking between individual by silanization. Despite being widely used, silane-treated slides siloxane molecules. In the present study we focus on lack the desired reproducibility—a fact that has driven a fabrication of (3-aminopropyl)trimethoxysilane–tetramethoxy- constant search for chemically alternative techniques rather than silane (APTMS–TMOS) hybrid films in search for new and improvements of silanization protocols. improved substrates for potential use in DNA microarray The structure of silane layer formation on the substrate analyses. The ability of the films to bind 25-mer oligonucle- surface heavily depends on the nature of silane, i.e. length of otide DNA is discussed in comparison with their commercial analogues (SAL-1 slides, Asper Biotech Ltd. [7]). The characteristics of the films are investigated by FTIR spectros- copy (Fourier transform infrared spectroscopy), MALDI-TOF ⁎ Corresponding author. Institute of Physics, University of Tartu, 142 Riia St., mass spectrometry (matrix assisted laser desorption ionization- 51014 Tartu, Estonia. time of flight mass spectrometry) and AFM (atomic force E-mail address: [email protected] (K. Saal). microscopy) measurements. 0167-577X/$ - see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.matlet.2005.12.035 1834 K. Saal et al. / Materials Letters 60 (2006) 1833–1838 2. Experimental acetone, methanol and ethanol (Naxo Ltd) and centrifuged at 280 ×g for 3 min. 2.1. Cleaning of glass slides before silanization 1 part of Cy3 3′labelled 5′-aminomodified 25-mer oligonu- cleotide DNA (a type of fluorescent-labelled oligonucleotide In order to exclude the possible effects of impurities on DNA, MWG-Biotech) was mixed with 100 parts of unlabelled reproducibility of coupling of silane and subsequently DNA to 25-mer oligonucleotide DNA (MWG-Biotech). The mixture glass surface the slides were subjected to cleaning procedure was spotted to glass slides with a spotter (Virtek CWP) in 100, developed in Asper Biotech Ltd. Glass slides (75 × 25 × 1 mm, 80, 50, 30, 10, 3, 1, 0.4 and 0.1 μM series, respectively. For Waldemar Knittel Glasbearbeitungs GmbH and Co KG) were dilutions Genorama Spotting Solution Type I (Asper Biotech sonicated for 10 min in 0.5% aqueous Alconox solution (Sigma- Ltd.) was used. The spotted slides were incubated in humid air Aldrich Co), washed thoroughly with distilled water and at 37 °C for 2 h and subsequently treated with ammonia vapour sonicated for 10 min in acetone (Naxo Ltd, analytical grade). for 1 h in order to block residual aminoreactive groups, washed Thereafter the slides were gently shaken for 1 h in 3 M NaOH thoroughly with hot distilled water and wiped dry by solution in 1 : 1 v / v mixture of water/95% ethanol (Naxo Ltd, centrifugation at 280 ×g for 3 min. analytical grade) and thoroughly washed with distilled water. For comparison, SAL-slides (prepared by incubation of Finally, the water was expelled by centrifugation of slides at cleaned glass slides in 2% w / w APTMS solution in 95% w / 280 ×g (Jouan CR422) for 3 min and the slides were stored in w acetone/water for 2 min and activated with phenylenedii- clean box until usage. sothiocyanate) were processed in the similar manner. The fluorescence of DNA spots was detected using 2.2. Preparation of APTMS–TMOS films ScanArray 5000 (Perkin-Elmer Inc.) microarray scanner. The spots were analyzed with Genorama Genotyping Software 4.0 APTMS and TMOS (both Sigma-Aldrich Co) were mixed at (Asper Biotech Ltd). molar ratios 0 : 1, 1 : 10, 1 : 5, 1 : 3, 1 : 1, 3 : 1, 5 : 1, 10 : 1, and 1 : 0, respectively. Then, at room temperature and constant stirring, a 2.4. Spectroscopic measurements mixture of water/methanol (Naxo Ltd, analytical grade) was added dropwise to the mixture of silanes. The final molar ratio Chemical groups and transformations during heating process of (APTMS + TMOS) / H2O / MeOH was kept as 1 : 2 : 2. In the (see 2.2) on silanized slides were analyzed by FTIR case of pure TMOS (APTMS–TMOS 0 : 1) the mixture of spectroscopic measurements. IR spectra were measured with water / methanol was acidified with concentrated HCl, so that Perkin-Elmer PC 16 FTIR spectrometer. A conventional the final molar ratio of TMOS / H 2 O / MeOH / HCl was Perkin-Elmer equipment was used for preparation of KBr 1 : 2 : 2 : 0.005. The mixtures were stirred until they turned to pellets (ø 12 mm) by compressing spectroscopically pure KBr highly viscous spinnable matter (ca. 30 min). Then, the powder under 10 tons of weight. Freshly prepared pellets polymerisation reaction was suppressed by introducing cold were coated with solutions of pre-polymerized precursors of dry methanol to the mixture of silanes, thus making up the final pure and mixed APTMS and TMOS in methanol and post- molar ratio of (APTMS + TMOS) / MeOH 1 : 7. The final treated as described in 2.2. product was kept sealed at 4 °C as stock solutions. MALDI TOF mass spectrometry measurements were For silanization of glass slides the stock solutions were performed with an instrument designed at the National Institute diluted 40 times with dry methanol and subsequently the slides of Chemical and Biological Physics of Estonia using 1,8,9- were dipped in these solutions. Thereafter the slides were kept trihydroxyanthracene (dithranol, Sigma Aldrich Co) as matrix. in open air (relative humidity 30%) for 48 h, and subsequently the temperature was raised to 140 °C (0.3 °C/min) for 12 h. 2.5. AFM measurements The thickness of prepared films was estimated with AFM (see 2.5). Freshly prepared films were scraped with a pin and The topographic features of APTMS–TMOS-films were processed as described above (unprocessed films are soft and investigated with an atomic force microscope SMENA-B (NT- there is possible to avoid harming the underlying glass surface MDT) working in semi-contact mode in air (at 20 °C, relative when scraping). The AFM measurements were performed at humidity 30%) using ultrasharp non-contact “Golden” silicon edges of the scratches. The thickness of the films was in the cantilevers NSG11 (NT-MDT). Different locations typically order of 30 nm. spanning over several square cm were scanned with different resolutions on each sample for reliable characterization of a 2.3. Immobilization of 25-mer oligonucleotide DNA onto sample. silanenized slides and DNA spot analysis 3. Results and discussion The silanized slides were gently shaken in 0.2% w / w 1,4- phenylenediisothiocyanate (Sigma-Aldrich Co) solution in 3.1. Immobilization of DNA to APTMS–TMOS films 10% w / w pyridine/dimethylformamide (Fluka, analytical grade) for 2 h, which activated the slides for immobilization A series of APTMS–TMOS films were prepared by variation of the of DNA. Then the slides were thoroughly washed with relative content of two silanes and measured their ability to bind 25-  K. Saal et al. / Materials Letters 60 (2006) 1833–1838 1835 maximal florescence signal within minimal area—an optimal mixture should be selected. Analogous aminopropyl embedded silica films were used for further studies on their potential use in DNA microarray analysis performing Arrayed Primer Extension (APEX) reactions mutation analysis [8]. Detailed description of microarray results is described in Ref. [9]. The layer of DNA on APTMS–TMOS 1 : 1 film was uniform showing some roughness only on nanometer scale, which allowed also clearly visualize the edge of the DNA spot (Fig. 2a). Note that smoothness of the surface, its stability and presence of sufficient reactive groups for the immobilization is a critical condition for higher resolution visualization of biomolecules. The surface of the SAL-slide had higher surface roughness both inside and outside the spot area (Fig. 2b), which could originate either from topographical features of the glass surface or immobilized siloxane clusters. Fig. 1. The binding curves of 25-mer oligonucleotide DNA to APTMS–TMOS hybrid films, normalized to the signal of 80 μM DNA spot on the SAL-slide. 3.2. Spectroscopic data of APTMS–TMOS precursors and films Each data point corresponds to an average fluorescence intensity of 16 independent spots. The inset next to legends on the plot shows the typical The formation of precursors and films was studied by FTIR and fluorescence images (300 × 300 μm2) of the 50 μM spots on the corresponding MALDI TOF mass spectrometry. It was observed that unbaked films films. had relatively strong absorption at 3342–3420 cm− 1, a band that corresponds to OH stretching of SiOH, CH3OH and H2O (Fig. 3). After heating at 140 °C the intensity of this absorption decreased mer oligonucleotide DNA (Fig. 1). In the case of APTMS–TMOS 0 : 1 substantially and starting from the APTMS–TMOS 1 : 1 film two film no binding was detected, which was because of the absence of well defined signals appeared at 3366 and 3284 cm− 1. These bands isothiocyanate groups on the surface. It indicated also that the non- correspond to the antisymmetric and symmetric stretching of NH2 specific binding of aminated DNA to the film was very low. The group, respectively. Surprisingly, these two absorptions were not binding of DNA to APTMS–TMOS 1 : 10 and 1 : 5 films was also low detected even in pure APTMS precursor. The reason could be an remaining on the level of 10% of the commercial SAL film binding. overlap with strong vibration of OH bond or formation of hydrogen Further increase of the content of APTMS in the mixture gave bond between NH2 and SiOH groups before the film formation, which considerable rise in the amount of immobilised DNA, and with the is in agreement with the proposed mechanism of formation of APTMS equimolar mixture the signal achieved 140% level of the SAL-glass layer on silica [10]. (Fig. 1). Similar high binding was achieved also in the case of The most intriguing region is between 1700 and 900 cm− 1. The APTMS–TMOS 3 : 1 films, but increase of APTMS excess to 5 fold or spectra of unbaked films clearly showed strong signals at 1580 cm− 1 higher led to diffuse DNA spots, which binding efficiencies could not that corresponds to N–H scissoring vibration and at 1486 cm− 1 that is be reliably obtained. The latter conforms with expectations that low believed to correspond to symmetric NH+3 deformation mode partly fraction of TMOS in the mixture causes slower and lower cross-linking superimposed by CH2 bending [11]. After heating of the precursor between aminosiloxane oligomers and therefore the formed film is not films the strong signal at 1486 cm− 1 disappeared and the medium stable in aqueous solutions. bending signal of CH2 of the usual value of 1476 cm− 1 was detected. It is important to note that the dimensions of the DNA spots The change of absorption bands corresponding to N–H and CH2 decreased with the increase of the amount of APTMS used for the films vibrations at 1580 and 1486 cm− 1, respectively (Fig. 3), is supposedly (Fig. 1., inset). The size of the spots correlates with the wettability of caused by decomposition of the relatively labile H-bonding network the surface, which is determined by the amount of hydrophobic between SiOH and NH2 groups. As a result of this process the degree of aminopropyl groups. Therefore, for the best practical conditions— polymerization increases and 3D structure is formed. At 1630 cm− 1 Fig. 2. Semi-contact mode AFM images of DNA spots on APTMS–TMOS 1 : 1 film (a) and on SAL slide (b); scan range 8.5 × 8.5 μm2. The scale bars on the right correspond to the line profiles drawn on the images to illustrate sharp edges of DNA spots. 1836 K. Saal et al. / Materials Letters 60 (2006) 1833–1838 Fig. 3. FTIR spectra of APTMS–TMOS hybrid films at 3 characteristic APTMS / TMOS mixtures. The pre-polymerized APTMS–TMOS mixtures were gelled on KBr pellets and spectra recorded before and after baking the pellets at 140 °C. Several well-known absorption peaks are marked. Intensity is in arbitrary units. only a weak shoulder in spectra of precursors as well as baked films As it can be seen in Fig. 3, two well-defined bands at 1034 and 1122 was detected. This signal did not change during heating and we could cm− 1 appeared in the spectra of baked APTMS–TMOS films. Similar not assign this to NH+3 deformation as it was proposed earlier [11]. structure at 1055 and 1088 cm− 1 in the spectra of (CH3)2Si Fig. 4. MALDI TOF mass spectrum of APTMS–TMOS 1 : 1 hybrid material.  K. Saal et al. / Materials Letters 60 (2006) 1833–1838 1837 (OCH2CH3)2 has been assigned to the linear and cyclic forms of 20 nm/μm (Fig. 5b). The surfaces of APTMS / TMOS 1 : 5 and 1 : 3 siloxane polymer, respectively [[12] cited therein]. The relative shift films were similar to 1 : 10 film, but the surface line profiles can be also explained by structural differences of initial monomers. indicated to substantially lower deviations in height (Fig. 5c and d). MALDI TOF mass spectra revealed that APTMS–TMOS hybrid Starting from APTMS / TMOS 1 : 1 film the surface profiles ranged materials had molar masses in range of up to 1500 amu (Fig. 4). The between two nanometers, thus showing practically featureless mass range detected did not significantly depend on the relative topography in 1 μm2 scale (Fig. 5e–i). compositions of APTMS / TMOS in the mixtures. As expected, the It has been shown that in basic medium alkoxysilanes polymerise to spectra had very complicated structure that contained different three-dimensional nanosized siloxane particles, whereas in acidic “families“ of oligomers. For pure pre-polymerized APTMS it was medium linear molecules are formed [13]. This result is confirmed estimated that such mass distribution corresponded to the oligomers since the polymerisation of TMOS was carried out in acidic medium containing approximately up to 12 monomers. where spinnable viscous material was formed and the corresponding film showed relatively uniform surface with no evidence of grainy 3.3. Surface of APTMS–TMOS films texture (Fig. 5a). When no acid as catalyst was used white powder-like material precipitated, indicating to the formation of branched siloxane APTMS–TMOS 0 : 1 film exhibited a uniform and smooth surface particles. On the other hand, the polymerisation of the mixture of (average vertical difference 5 nm per 1 μm scan) (Fig. 5a). APTMS / APTMS / TMOS can be considered as base-catalysed process because TMOS 1 : 10 film showed surface consisting of grains with several to a of amino groups of APTMS. Thus, in great excess of TMOS silane hundred nanometers in diameter and average height distribution of polymerises to nanosized granules that are clearly evident in Fig. 5b–d. Fig. 5. Semi-contact mode AFM images of APTMS–TMOS hybrid films; scan range 1 × 1 μm2. The scale bar on the right corresponds to the line profiles drawn in the middle of the scan. AVE represents the average height difference in scan profile. High-resolution inset (200 × 200 nm2) in (g) illustrates granule-like structure of the surface (see text). 1838 K. Saal et al. / Materials Letters 60 (2006) 1833–1838 At higher APTMS content the granules appear smaller probably Science Foundation Nanotribology Programme. I.K. acknowl- because of the steric hindrance of aminopropyl groups that diminishes edges support by EC FW5 “Centres of Excellence” programme the condensation to progress in all directions (see inset in Fig. 5g). In (ICA1-1999-70086). The authors acknowledge Asper Biotech between these extremes of APTMS / TMOS ratios observed we believe Ltd. for helpful cooperation. this steric hindrance to remain sufficient for disabling the growth of siloxane particles big enough to precipitate, which makes them use of References as precursors for film making. Consequently, the APTMS / TMOS hybrid films can be thought of as consisting of densely packed [1] M. Schena, D. Shalon, R.W. Davis, Science 270 (1995) 467. nanosized siloxane particles, and the dimensions as well as the rate of [2] K. Bierbaum, M. Kinzler, C. Wöll, M. Grunze, G. Hähner, S. Heid, F. cross-linking between individual “building blocks” are determined by Effenberger, Langmuir 11 (1995) 512. the ratio of APTMS / TMOS. [3] T.J. Horr, P.S. Arora, Colloids Surf., A Physicochem. Eng. Asp. 126 (1997) 113. [4] M. Hu, S. Noda, T. Okubo, Y. Yamaguchi, H. Komiyama, Appl. Surf. Sci. 4. Conclusions 18 (2001) 307. [5] M.E. McGovern, K.M.R. Kallury, M. Thompson, Langmuir 10 (1994) 3607. 1. It was shown that APTMS–TMOS hybrid films have [6] T. Tätte, K. Saal, I. Kink, A. Kurg, R. Lõhmus, U. Mäeorg, M. Rahi, A. potential as substrates for immobilisation of aminated Rinken, A. Lõhmus, Surf. Sci. 532–535 (2003) 1085. DNA via 1,4-phenylenediisothiocyanate linkering. [7] For details see http://www.asperbio.com. 2. The ratio of APTMS to TMOS chosen in precursor synthesis [8] A. Kurg, N. Tonisson, I. Georgiou, J. Shumaker, J. Tollett, A. Metspalu, determines the number of amino groups on the films' Genetic Invest. 4 (1) (2000) 1. [9] K. Saal, M. Plaado, I. Kink, A. Kurg, V. Kiisk, J. Kožvnikova, U. Mäeorg, surfaces available for immobilization of DNA and shape of A. Rinken, I. Sildos, T. Tätte, A. Lõhmus, Biological and Bio-Inspired the formed spots. The maximal binding of DNA was Materials and Devices (Mater. Res. Soc. Symp. Proc. 873E, Warrendale, achieved on the APTMS–TMOS 1 : 1 and 3 : 1 films. PA), 2005, K9.3. 3. Films with higher APTMS molar content (starting from [10] T. Jesionovski, A. Krzistafkiewicz, Appl. Surf. Sci. 172 (2001) 18. [11] I. Shimizu, H. Okabayashi, K. Taga, E. Nishio, Vibr. Spectrosc. 14 (1997) APTMS–TMOS 5 : 1 film) were not stable in aqueous 113. medium. The instability was probably caused by relatively [12] Z. Zhang, B.P. Gorman, H. Dong, R.A. Oronzco-Teran, D.W. Mueller, R.F. low content of cross-linking agent TMOS. Reidy, J. Sol-Gel Sci. Technol. 28 (2003) 159. [13] S. Sakka, H. Kozuka, J. Non-Cryst. Solids 100 (1988) 142. Acknowledgements This work was supported by the Estonian Science Founda- tion grants no. 5015, 5545, 5467 and 6492, and by European