- Email: [email protected]
Improved surface sensing of DNA on gas-etched porous silicon
Sensors and Actuators B 120 (2006) 220–230
Improved surface sensing of DNA on gas-etched porous silicon D.C. Tessier, S. Boughaba ∗ , M. Arbour, P. Roos, G. Pan IatroQuest Corporation, 1000 Chemin du Golf, Verdun, QC H3E 1H4, Canada Received 30 March 2005; received in revised form 14 January 2006; accepted 5 February 2006 Available online 15 March 2006
Abstract A new gas-based etching process has been used to produce porous silicon (PSi) samples as sensor substrates. The as-etched porous material demonstrated an intense photoluminescence signal at an average wavelength of 658 nm (σ = 1%). The photoluminescence signal, with a peak wavelength in the range 530–540 nm, was found to increase with the length and concentration of unmodified oligonucleotide probes on surfaces coated with either amino-propyl trimethoxysilane (APTMS) or glycidoxy-propyl trimethoxysilane (GPTMS). The large surface area that characterized this material significantly increased the binding of DNA probes and led to increased sensitivity in target detection. Using fluorescently-labeled DNA, the target capture for sub-micromolar detection on this novel sensor substrate was demonstrated to be 100 times higher than that obtained on non-etched silicon. In a 1 h assay, we were able to detect specifically approximately 3 fmole of a 25mer target oligonucleotide per mm2 of substrate. © 2006 Elsevier B.V. All rights reserved. Keywords: Porous silicon; Etching; Photoluminescence; DNA; Silane; Sensors
1. Introduction Accurate and sensitive DNA detection tools have become pivotal to address fundamental questions relating to gene expression and regulatory biology [1–4], but also for disease diagnosis [5,6], drug discovery [7] and the identification of pathogens present in environmental samples [8]. This growing need to have access to specific DNA sequence information has prompted the development of faster, more sensitive and affordable nucleic acid biosensors [9]. Optical DNA detection methods include the use of various intercalating dyes [10,11] or rely on spectral interference of reflected white light [12,13]. Amperometric or electrochemical biosensors use a variety of electrode types functionalized with DNA to monitor enzymatic reactions occurring upon hybridization of the desired target [14,15]. Some electrical biosensors are based on monitoring the intrinsic molecular charge of DNA on field-effect sensors [16] or upon hybridization of a target molecule on PNA probes immobilized on silicon nanowires [17]. Other detection technologies utilize functionalized gold nanoparticles to detect hybridization events optically [18] or
∗
Corresponding author. Tel.: +1 613 995 7752. E-mail address: [email protected] (S. Boughaba).
0925-4005/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2006.02.002
electrically following silver enhancement [19]. Many of the DNA detection strategies deployed today are based on their hybridization to probe molecules immobilized to solid surfaces [20]. The choice of the material and its chemical functionalization is of utmost importance to produce a sensitive and reliable biosensor. Silicon has been the premier material driving the microelectronics revolution due to its unique electrical, chemical, and mechanical properties. Because of its indirect energy band-gap, bulk silicon is a poor light emitter and this has restricted its use in light-emitting devices and optoelectronics. Porous silicon (PSi), on the other hand, has been demonstrated to yield efficient visible light emission at room temperature [21–23]; such a behavior is primarily attributed to the electron confinement in the nanocrystals that constitute the porous structure [21,24]. Since 1990, the prospect of extending the use of silicon to lightemitting devices and optoelectronics has triggered numerous investigations on PSi and its photo- or electroluminescence properties [21,25]. Furthermore, the large fraction of voids inside PSi results in a very large specific surface area, on the order of a few hundred m2 /cm3 [26,27]. Such considerable surface area, which provides a large number of potential binding sites, has generated significant interest in PSi for chemical and biosensing applications [21,28–33]. As such, PSi is used as a transducer whose photoluminescence, refractive index, or electrical conductivity
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
221
Fig. 1. Schematic of the gas etching setup.
properties, are modulated upon exposure to target molecules. Starodub et al. [34] were able to detect as little as 10 ng/ml of myoglobin using inherent properties of PSi both as a matrix and as a transducer of light. In this article, we report on the characterization of PSi substrates, fabricated using a new gas etching method, as an excellent surface enhancing material for the development of a sensitive DNA sensor. This new etching technique was developed as part of a novel biosensing platform, Bio-AlloyTM1 , which attempted to make use of the unique photoluminescence properties of PSi to detect label-free target agents in real-time. Chemically functionalized PSi was validated as a substrate with increased surface area and sensitivity for DNA detection. 2. Experimental 2.1. Fabrication of porous silicon by gas etching The gas etching technique consists of exposing silicon samples to a mixture of oxygen (O2 ) and nitrogen dioxide (NO2 ) gases and an acid vapor. The experimental setup is schematically presented in Fig. 1. 4 mm × 4 mm silicon samples were obtained from the dicing of 1 0 0 boron-doped p-type wafers (Shin-Etsu Handotai), whose electrical resistivity was 20 cm. The samples were loaded onto a tray, which was mounted at the bottom of a chamber. After installing a gas distribution plate, whose role was to improve the uniformity of the gas flow, the chamber was hermetically sealed. The chamber, tray, and distribution plate were made of chemically inert TeflonTM2 . Pure oxygen (99.995%, Air Liquide) was flowed through a heated scrubber containing HF (47–51%, TraceMetal grade, Fisher Scientific), then merged with a flow of diluted nitrogen dioxide (2% in air, MEGS) before entering the etching chamber. The
1 2
Bio-Alloy is a registered trademark of IatroQuest Corporation. Teflon is a registered trademark of DuPont.
outlet of the chamber was connected to a scrubber containing a 2 N sodium hydroxide (NaOH) solution made from electrolytic pellets (certified ACS grade, Fisher Scientific). The role of the NaOH solution was to neutralize the HF. The O2 and NO2 gas flow rates could be varied in the range of 100–500 ml/min and 10–50 ml/min, respectively, using flowmeters (Cole-Parmer), while the HF scrubber could be kept at room temperature or heated up to 70 ◦ C. The NO2 cylinder was heated at its base to a temperature of 40 ◦ C to avoid accumulation of nitrogen dioxide at its bottom and to enhance the mixing of NO2 and air. Similarly, the stainless steel tubing connecting the NO2 cylinder to the chamber was heated to a temperature of 30 ◦ C to avoid condensation of NO2 on the tubing wall. For all the investigations reported in this article, samples were etched for 30 min. Following the etching process, the samples were rinsed using ethyl alcohol (95%, undenatured grade, Commercial Alcohols); the substrates were dipped in ethyl alcohol for 5 min, then removed and left to dry in a nitrogen environment for about 30 min. Prior to gas etching, the 4 mm × 4 mm silicon samples were first cleaned using RCA-type hydrogen peroxide mixtures [35], etched in a 5% HF solution for 2 min, rinsed in deionized water for 5 min, then oxidized at room temperature in a 3 SLM flow rate of ozone (O3 ) gas (280 g/N m3 concentration in nitrogen) for 5 min. At the end of the oxidation step, the samples were immediately loaded in the chamber and the etching was started. This cleaning sequence allowed a stringent control of the sample surface, yielding a hydrophilic surface. Spectroscopic ellipsometry measurements (Scientific Computing International, FilmTek 2000SE) revealed an oxide film thickness comparable to that of silicon native oxide (<2 nm). The morphology of the porous layers was investigated by scanning electron microscopy (SEM, Hitachi, S-4700), while their photoluminescence properties were characterized using a photo-detection system (Fluorosense, PDS) whose excitation source was a diode laser (λ = 470 nm), with an on-sample spot size of about 2 mm. A 500 nm longpass filter was inserted at the entrance of the spectrometer to eliminate any contribu-
222
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
tion to the signal from the source. Moreover, the thickness and porosity of the layers were simultaneously evaluated by spectral reflectance in the range 400–900 nm (Scientific Computing International, FilmTek 2000M) using an on-sample probing beam of 10–20 m diameter. SCI’s generalized dispersion formula was used to model fitted values of the dielectric function ε(λ); silicon and void were selected as the constituents of the porous layers. 2.2. Oxidation of gas-etched porous silicon The surface of as-etched luminescent PSi is hydrogenpassivated [21]. Silanization of such a surface requires the formation of silanols, i.e. Si–OH intermediate bonds. In order to provide hydroxyl group coverage of the surface of gasetched samples, room-temperature ozone oxidation was performed after the ethanol rinsing/drying step. In this method, samples were exposed for 30 min, at room temperature, to a 3 SLM flow rate of high-purity O3 + N2 mixture produced by an O3 generator (MKS, AX8560). The oxidation was carried out in a hermetically sealed chamber, and the O3 concentration was 280 g/N m3 . The photoluminescence properties of the oxidized porous layers were investigated using the PDS, while their porosity and thickness were evaluated by spectral reflectance. For the latter, silicon, silicon dioxide, and void were selected as the constituents of the layers. Fourier transform infrared (FTIR) spectroscopy analyses were performed, in the 400–4000 cm−1 range and transmission mode, on the oxidized porous samples to determine the type of chemical bonds present (Bio-Rad, FTS3000).
and with different 5 and/or 3 terminal modifications and were all purified by HPLC (Integrated DNA Technologies). BACA-AS (25mer) IQT2 (75mer)
IQT4 (25mer) IQT15 (70mer)
IQT15M1 (70mer)
IQT15M2 (70mer)
IQT15M3 (70mer)
IQT20 (70mer)
IQT39 (25mer):
5 -C6-amino-CCCTCGAGATTCCATGCCTATTACA-3 5 -C6-amino-TGTAATAGGCATGGAATCTCGAGGG-TGTAATAGGCATGGAATCTCGAGGG-TGTAATAGGCATGGAATCTCGAGGG-3 5 -CCCTCGAGATTCCATGCCTATTACA-3 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGGGCTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (complementary to IQT20) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGGACTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for a central point mutation) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGAACTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for two central mutations) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGAAATGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for three central mutations) 5 -CATCTGGCGTAAGAGTGCCAACAATGCTTGGCCAGCCCACCATGGCAAACTAATATAGCTTGCTCCAAGT-3 (complementary to IQT15) 5 -ATGCTTGGCCAGCCCACCATGGCAA6-FAM-3 (complementary to the central portion of IQT15)
2.5. Immobilization of DNA 2.3. Silanization In order to produce a bio-compatible material, the surface of the PSi samples needed to be chemically modified to allow for the immobilization of DNA recognition elements. Oxidized porous substrates were silanized in the liquid phase for 1 h in the presence of 0.1% amino-propyl trimethoxysilane (APTMS) or glycidoxy-propyl trimethoxysilane (GPTMS) (Sigma-Aldrich) diluted in 100% toluene. Following this incubation, the substrates were rinsed once with toluene, once with 1:1 toluene:ethanol and twice with ethanol. The silanized substrates were then dried at 110 ◦ C for 10 min and stored in a dry nitrogen atmosphere until use. 2.4. List of oligonucleotides Here is the list of all the oligonucleotide sequences that were used in the present study. While they differed primarily in their length, some contained mutations to demonstrate singlenucleotide discrimination of the assay. Bold characters indicate the 25-base region of sequence homology/complementarity between IQT15, IQT20 and IQT39. Italic and underlined characters indicate the location of the point mutations between IQT15 and IQT15M1, M2 and M3, and their complementary target. Oligonucleotides were purchased in a variety of lengths
In general, 50 M dilutions of the oligonucleotide capture probes were prepared in water immediately prior to use. A volume of 2 l of spotting solution (∼100 pmole DNA) was applied to the center of the silanized substrates to eventually cover most of the 4 mm × 4 mm surface. Spotted substrates were incubated in microtiter plates at room temperature for 2 h; 1 h covered and 1 h uncovered to dry. UV cross-linking was carried out at 600 mJ in a UV Stratalinker1800TM (Stratagene). The substrates were then washed four times with 200 l of H2 O and air-dried for about 15 min. 2.6. Target hybridizations Hybridizations were usually carried out by placing the probeimmobilized substrates in individual wells of microtiter plates. The substrates were covered with 50 l of a target solution in 5× SSC (75 mM sodium citrate + 750 mM NaCl). The microtiter plates were then incubated at 37 ◦ C for 1 h with agitation (60 rpm). Following hybridization, the remaining target solution was removed by aspiration and the substrates were washed three times with 200 l of 1× SSC (15 mM sodium citrate + 150 mM NaCl), three times with 200 l of 0.1× SSC (1.5 mM sodium citrate + 15 mM NaCl), air-dried, and scanned for fluorescence.
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
2.7. Complex media A culture of Escherichia coli K12 was grown to saturation in LB broth overnight at 37 ◦ C. After centrifugation, the cell pellet was resuspended in the original culture volume of 10× SSC, boiled for 10 min and sonicated at maximum power in a sonicating bath (Branson, 2510) for 30 s. The lysate was mixed with an equal volume of target DNA. Purified E. coli genomic DNA was purchased (United States Biochemicals) and diluted to 42 g/ml in 5× SSC with the target to evaluate the impact of high DNA content on the efficiency of hybridization. Approximately 1 g of local soil was mixed with 1 ml of 50 mM ethylenediamine tetraacetic acid (EDTA, Sigma-Aldrich) in water, pH 8.0, and then boiled for 5 min. The mixture was centrifuged at 17000 × g for 2 min and the supernatant “dirty” water was used to dilute the fluorescent target. All complex mixtures were boiled for 5 min and cooled on ice for 5 min immediately prior to hybridization with the substrates. 2.8. Photoluminescence and fluorescence measurements At certain steps, the PSi substrates were monitored for fluorescence (6-FAMTM ; Abmax , 495 nm, Emmax , 520 nm) or photoluminescence using the photo-detection system (Fluorosense,
223
PDS). The emission spectra were acquired in the 400–900 nm range with an integration time of 200–3000 ms. 3. Results 3.1. Gas etching of porous silicon samples Fig. 2 presents scanning electron micrographs of a sample etched using O2 and NO2 flow rates of 100 and 10 ml/min, respectively. Etching resulted in the formation of a porous layer, which consisted of islands (Fig. 2 (b–e)) embedded in a matrix of different morphology (Fig. 2 (f)). The islands are likely related to the condensation of reactant-laden drops onto the surface during the etching process. Such a condensation could nucleate at surface defects, such as particulates. The island morphology presented in Fig. 2 is representative of all samples etched at various conditions. The density and size of the islands were, however, found to be affected by the O2 :NO2 flow rate ratio for a given O2 flow rate; the island size increased when decreasing the O2 :NO2 ratio, while the island density decreased concurrently. Fig. 3 presents scanning electron micrographs of two samples etched at O2 :NO2 ratios of 20 and 6.7. In this case, the average island diameter was found to be of 71 ± 17 m and 160 ± 25 m for a ratio of 20 and 6.7, respectively. The corresponding density of
Fig. 2. Scanning electron micrographs of a silicon sample etched using O2 and NO2 flow rates of 100 and 10 ml/min, respectively. HF scrubber kept at room temperature. (a) Low magnification of entire 4 mm × 4 mm sample; (b–e) increasing magnifications of the surface inside an island; and (f) surface between islands.
224
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 3. Scanning electron micrographs of silicon samples etched using O2 and NO2 flow rates of (a–c) 200 and 10 ml/min, respectively, and (d–f) 200 and 30 ml/min, respectively. HF scrubber kept at room-temperature. Micrographs (c) and (f) represent the areas between islands.
islands was determined to be 115 and 43 mm−2 , respectively. The size of the islands is directly related to the size of the drops, which depends on the surface tension and liquid drop coalescence driven by surface tension [36]. Such dependence directly relates to the chemical composition of the drops, which varies when changing the O2 :NO2 flow rate ratio. As illustrated in Fig. 3, the morphology inside the islands was also affected when changing the O2 :NO2 ratio; a more pronounced pyramidal texture, related to the anisotropic etching of silicon, was observed. Conversely, the morphology of the matrix surrounding the islands was not affected by changes in the O2 :NO2 flow rate ratio. Regardless of the flow parameters, the morphology of the porous matrix surrounding the islands was found to be different from that inside the islands. A smoother surface was observed, revealing an isotropic etching, which is characteristic of a process occurring in the vapor phase. Two concurrent etching modes could thus be considered to describe the gas etching process: a liquid-based one, occurring within the areas covered by condensed drops, and a gas phase-based one, taking place between the drops. Spectral reflectance measurements revealed an average porosity, i.e. void fraction, of 88% (σ = 3.1%) inside the islands and 79% between the islands (σ = 3.5%) for all gas flow rates. No dependence of the porosity on the etching conditions was observed. The spectral reflectance measurements also permitted the determination of the porous layer thickness within and between islands. The thickness was found to have an average value of 1230 nm (σ = 16%) inside the islands, and 301 nm (σ = 22%) in the matrix. Such a large difference in thickness was confirmed by focused ion beam cross-sections (Hitachi, FB2000A). The disparity in layer thickness and porosity between the two areas further substantiates the assumption of two different etching modes. Fig. 4 (a) shows the photoluminescence spectrum of a sample etched using O2 and NO2 flow rates of 400 and 20 ml/min, respectively. Intense PL in the red portion of the visible light
spectrum was obtained. Fig. 5 presents the photoluminescence peak intensity and wavelength as a function of O2 and NO2 gas flow rates. For each set of flow rates, the average of four samples was calculated. The average peak wavelength was determined to be 658 nm (σ = 1%). There was no dependence of the average peak wavelength on the processing conditions, in agreement with the fact that comparable porosity was obtained for all conditions. Furthermore, no dependence of the PL peak intensity on the gas flow rate was observed; the average PL peak intensity was found to be 8677 counts/s (2603 counts for an integration time of 300 ms), with a standard deviation of 11%. For all the investigations reported in the subsequent sections, porous samples were fabricated using O2 and NO2 flow rates of 300 and 30 ml/min, respectively, and an HF solution (in scrubber) kept at a temperature of 50 ◦ C.
Fig. 4. Photoluminescence spectra measured on a sample etched using O2 and NO2 flow rates of 400 and 20 ml/min, respectively, (a) after etching and (b) after oxidation. HF scrubber kept at temperature of 50 ◦ C. Integration time during PL measurements, 300 ms after etching and 3 s after oxidation.
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 5. Average photoluminescence peak intensity and wavelength as a function of the O2 and NO2 flow rates. HF scrubber kept at room temperature. For each set of flow rates, the average of four samples was calculated. Integration time during PL measurements, 300 ms.
3.2. Oxidation of gas-etched porous silicon samples Oxidized samples were characterized in terms of PL properties, surface chemistry, and composition. For all etching conditions, severe attenuation of the PL emission was observed as illustrated in Fig. 4 (b). Such attenuation in PL is attributed to the poor electronic quality of the interface between the remnant silicon and SiO2 , i.e. to trap states at the interface that are efficient luminescence quenchers [22]. The quenching of the original PL on oxidized samples is beneficial with respect to fluorescence-based detection of specifically immobilized DNA or biomolecules. The original PL spectra, obtained after etching, had broad peaks (Fig. 4 (a)) originating from a distribution of the size of the crystallites inside the porous layers [22]. Such a large luminescence bandwidth (∼120 nm), if present, would impact fluorescence signals related to biological detection events, even if the signal collection was carried out within a narrow wavelength window through the use of bandpass filters. An FTIR spectrum representative of all oxidized samples is presented in Fig. 6. The spectrum exhibited absorption bands
Fig. 6. FTIR spectrum measured on an oxidized porous sample. Etching was performed using O2 and NO2 flow rates of 400 and 20 ml/min, respectively, with the HF scrubber kept at a temperature of 50 ◦ C.
225
Fig. 7. Photoluminescence spectra of gas-etched porous samples following silanization with 1% APTMS or GPTMS. Integration time, 3 s.
at 1640 and 3400 cm−1 , which are characteristic of hydroxyl groups. It also revealed the presence of residual Si–H bonds through the presence of an absorption band at 2260 cm−1 . Furthermore, the spectrum was characterized by a siloxane band at 1105 cm−1 , with a shoulder at 1076 cm−1 , attributed to Si–O–Si asymmetric stretching vibrational mode. Other absorption bands were observed at 816, 883, and 970 cm−1 ; they were assigned to Si–O–Si bending mode, a combination of symmetric Si–H deformation and Si–O stretching in Si–OH groups, and Si–O stretching mode in either negatively charged non-bridging oxygen (Si–O− ) or Si–OH groups, respectively. The determination of the silicon dioxide, silicon, and void fractions by spectral reflectance revealed that the porous layers were only partially oxidized, in agreement with the FTIR results. The average ratio of silicon to silicon dioxide fractions after oxidation was determined to be 0.64 ± 0.02 inside the islands and 0.62 ± 0.11 between the islands. 3.3. Effect of silanization on photoluminescence The effect of silanes on the photoluminescence properties of Bio-AlloyTM samples was thoroughly investigated. Two classes of silanes were used to immobilize DNA on the nanostructured substrates: amino-containing silanes, typically APTMS, for electrostatic interaction with the phosphate backbone of DNA, and the epoxy-containing silane GPTMS, for covalent attachment of amino-modified or unmodified DNA to the substrates. Silane coupling experiments revealed the importance of the nature of the silane on the resulting photoluminescence of PSi. Photoluminescence spectra of APTMS and GPTMS-treated samples are presented in Fig. 7. Intense photoluminescence, with a peak wavelength of 530–540 nm, was observed after silanization with APTMS, whereas very little photoluminescence, with a peak wavelength in the range 620–660 nm, could be observed after silanization with GPTMS. The weak photoluminescence using GPTMS could be attributed to residual red photoluminescence following the oxidation step. It was, however, possible to artificially generate photoluminescence in the green portion
226
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
of the light spectrum, at a peak wavelength of 533 nm, on oxidized porous samples treated with 0.1% GPTMS by spotting and baking the amino acid l-lysine or poly-l-lysine. Water and l-norleucine, a non-natural amino acid lacking the -amino group of l-lysine did not produce this green photoluminescence (data not shown). This green photoluminescence might originate from the covalent bonds created between the epoxy group of the GPTMS and the primary amines of l-lysine and polyl-lysine. These amino-containing, poly-l-lysine green photoluminescent surfaces were later shown to immobilize unmodified fluorescently-labeled DNA to the same extent as poly-l-lysinecoated glass slides [37]. 3.4. Probe immobilization and effect on photoluminescence In order to characterize the binding of DNA probes to the surface of PSi samples, 25mer oligonucleotides carrying fluorescent tags (TAMRA) at their 3 -end were applied to silanized substrates; a range of APTMS and GPTMS concentrations were evaluated to maximize the immobilization of the probes. Concurrently, for the same silane concentrations, photoluminescence was monitored using unlabeled oligonucleotides. Maximum loading and retention of labeled probes was achieved between 0.001 and 1% silane coating concentrations, whereas maximum photoluminescence of the PSi substrates exposed to unlabeled probes was measured between 0.01 and 0.1% silane coating concentrations, with a maximum peak wavelength in the range 530–540 nm (data not shown). Similar DNA binding and photoluminescence characteristics were observed with PSi substrates silanized with other amino-containing organosilanes. Consistent photoluminescence was demonstrated following immobilization of unmodified probes on silanized surfaces. The intensity of this photoluminescence correlated with the size and the amount of immobilized DNA. Fig. 8 illustrates the photoluminescence associated with the immobilization of 25 and 100 pmole of 25mers compared to 75mers on 0.01% APTMS (Fig. 8 a) and to a lesser extent on 0.1% GPTMS after baking (Fig. 8 b). Increasing amounts of fluorescently-labeled probes were immobilized on samples functionalized with 1% APTMS or GPTMS. These experiments revealed that covalent attachment to GPTMS-coated porous surfaces retained almost twice as much probe as the electrostatic interaction on APTMS. Consequently, GPTMS was selected for the subsequent immobilization and hybridization studies. 3.5. Target hybridizations on GPTMS Unmodified oligonucleotides of different lengths were spotted on GPTMS-treated porous samples and dried before hybridization with 25mer fluorescent targets. The effect of UV cross-linking of 25 and 70mer probes on hybridization efficiency was studied. It was found that an irradiation energy ranging from 0 to 1200 mJ had no significant effect on the hybridization to 25mer probes. On the other hand, at least four-fold increase in fluorescence signal intensity was obtained with 70mer probes when they were exposed to 600 mJ of UV light compared to no treatment, indicating a more efficient retention of probes
Fig. 8. Photoluminescence of 25 and 75mer probes on 0.01% APTMS (a) and 0.1% GPTMS (b). Two microliter of 12.5 or 50 M IQT4 or BACA-AS or IQT2 DNA were spotted and treated as described in Section 2. Integration time, 3 s.
on the porous surface. This positive effect of UV cross-linking on hybridization to longer probes was also observed on mixed amine/epoxy surfaces [38]. Under optimal hybridization conditions for both 25 and 70mer probes, we found that the longer probes yielded three to five times more fluorescence signal. This has also been reported by other groups [39]. The effect of probe loading and its impact on the efficiency of hybridization was also evaluated. Fig. 9 shows that spotting 500 M of a fluorescent 25mer probe (1 nmole or 4 × 1013 molecules/mm2 ) was not sufficient to saturate the surface of the porous sample. This probe density is 100-fold higher than 3 × 1011 molecules/mm2 reported for the immobilization of 12mer oligonucleotides on flat silica APTMS-treated surfaces [40]. Maximum target hybridizations were nonetheless obtained with 50–100 M spotting, i.e. 100–200 pmole of probe (Fig. 9). This probe-dependent levelling off in target hybridizations is supported by a published study on epoxy-silanized flat surfaces [41], and by the demonstration from other groups that dense physical packing of probes on solid phases dramatically reduces the efficiency and the kinetics of target hybridization [42,43]. It was of interest to determine the effect of temperature on the kinetics of hybridization. Three temperatures were selected: room temperature (20–22 ◦ C), 37 ◦ C, and 65 ◦ C. Specific hybridizations occurred within 5 min at all three temperatures, and the intensities continued to increase for up to about
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 9. Effect of probe loading on the efficiency of hybridization. Fluorescentlylabeled probes were spotted in parallel on etched and non-etched substrates and washed as described in Section 2. Hybridization of 0.5 M IQT39 target on complementary 70mer probes (IQT15) on 0.1% GPTMS-coated substrates in 5× SSC at 65 ◦ C for 1 h. Background and non-specific signals were measured to be less than 25 fluorescent counts. Integration time, 500 ms.
227
Fig. 11. Single base-pair discrimination in hybridization using 0.5 M IQT21 target on 70mer probes containing a fully complementary central portion to the target (IQT15), single mismatch (IQT15M1), two base-pair mismatch (IQT15M2) and three base-pair mismatch (IQT15M3) on 0.1% GPTMS samples. Hybridizations were carried out in 5× SSC in a thermocycler under the conditions described in the figure. Integration time, 500 ms.
2 h as illustrated in Fig. 10. Hybridizations at 65 ◦ C were significantly accelerated compared to room temperature and 37 ◦ C. Although it was possible to demonstrate highly specific and sensitive hybridizations using fluorescently-tagged targets, we were not able to produce similar photoluminescence responses under these optimized conditions. Slight variations of this protocol were used to discriminate single-nucleotide differences between probes. Mutated 70mer sequences containing single, double and triple base-pair mismatches were differentiated from wild-type using a 25mer labeled target complementary to the central segment of these probes (see Section 2). Samples were incubated in 5× SSC in a microtiter plate, inside a thermocycler for approximately 40 min, then washed and scanned for associated fluorescence (Fig. 11). The results showed a five-fold increase in specificity for the fulllength probe as compared to a single base-pair mutant probe.
Further addition of 15% formamide to the 5× SSC concomitant with a reduction of the final assay temperature to 60 ◦ C improved the discrimination for mismatched probes to more than 10-fold in less than 15 min. The sensitivity of the assay was evaluated by determining the lower limit of detection of a 25mer fluorescent target. The lower limit of detection was defined as the lowest target concentration detectable having more than two standard deviations above average background fluorescence, as measured from pools of non-specific hybridizations. Although the assay exhibited specific detection of target sequences from a 500 nM hybridization mixture in as little as 5 min at room temperature (Fig. 10), it was possible to increase the sensitivity of the assay by 500-fold (to 1 nM) by increasing the assay temperature to 65 ◦ C for 1 h (Fig. 12). The results displayed a linear response over approximately two logs. Above 100 nM of target, most of the sam-
Fig. 10. Effect of temperature on hybridization efficiency. Hybridization of 0.5 M IQT39 target on complementary 70mer probes (IQT15) on 0.1% GPTMS-coated substrates. Targets were diluted in 5× SSC and hybridized to the substrates at room temperature, 37 or 65 ◦ C for 5 min to 4 h. Background and non-specific signals were determined to be less than 25 fluorescence units. Integration time, 100 ms.
Fig. 12. Determination of the lower limit of detection and comparison between etched and non-etched silicon surfaces for identical hybridization assays (IQT15 probe, IQT39 target, 5× SSC, 65 ◦ C, 1 h). The dashed line represents the same hybridization assay carried out on etched substrates in the presence of a boiled and sonicated lysate of E. coli. Background and non-specific signals were measured to be less than 25 fluorescence units.
228
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 13. Hybridization of 0.5 M IQT39 target in complex media on complementary and non-complementary 70mer probes on 0.1% GPTMS-coated substrates. Unless otherwise stated in Section 2, hybridization mixtures were prepared by combining three parts of the indicated media with one part of 20× SSC to which was added a negligible volume of target. All complex mixtures were boiled for 5 min and cooled on ice for 5 min before hybridization at 65 ◦ C for 1 h. Integration time, 200 ms.
ples exhibited saturated signals for the integration time used to acquire the fluorescent data. The same experiment was repeated on non-etched silicon samples to compare the sensitivity of the assay between porous and flat silicon surfaces. Non-etched silicon samples were washed and immediately oxidized before silanization with 0.1% GPTMS; thus repeating the same oxidation procedure as for the porous samples. Maximum signals obtained on non-etched samples were 100-fold lower than those obtained on etched samples, with severe consequences on specificity of the assay. In fact, etched samples displayed specific/nonspecific signal ratios higher than 100 compared to ratios smaller than 3 for non-etched samples using 100 nM target. The presence of a boiled and sonicated E. coli lysate did not significantly affect the performance of the assay (Fig. 12, dashed line). The assay was tested in a variety of complex media to assess the robustness of the assay and the stability of the PSi samples. Comparable signal intensities were obtained after hybridizing 0.5 M of the IQT39 target (6-FAM-labeled 25mer target) in the presence of 42 g of E. coli genomic DNA, soil-drained “dirty” water, rabbit serum, human urine, a buccal swab suspension or lake water compared to deionized water (Fig. 13). To avoid possible degradation of DNA by nucleases, 50 mM EDTA was added to the soil-drained water. The only test medium which significantly affected the sensitivity and specificity of the assay was the urine sample. The stability of GPTMS-silanized and DNA-spotted porous samples was tested over a period of 1 month. In a first experiment, silanized porous samples were stored at room temperature, in a dry nitrogen environment, for a given period of time before immobilization of the probe. A complete hybridization assay was then performed and results were compared to other storage time points (Fig. 14). The fluorescence response of GPTMSsilanized material increased over the first 15 days and returned to its original level after 1 month without affecting specificity. In a second experiment, GPTMS-silanized porous samples pre-
Fig. 14. Ageing study of PSi samples. Gas-etched substrates were silanized with 0.1% GPTMS and stored desiccated under N2 atmosphere for the indicated times. At these time points, complementary and non-complementary 70mer probes were immobilized, UV cross-linked to the substrates, and hybridized with 0.5 M IQT39 in 5× SSC at 65 ◦ C for 1 h. Integration time, 100 ms.
spotted with 100 pmole of probe were stored under identical conditions and tested by hybridization at the same time points. The fluorescence response of pre-spotted samples did not change in the first week but steadily decreased over the following 3 weeks to reach ∼33% efficiency after 1 month. Specificity was not affected by long-term storage of pre-spotted samples indicating minimal degradation of capture probes on the nanostructured samples. 4. Discussion and conclusions The development of a new gas etching technique for nanostructuring of silicon resulted in the fabrication of porous samples with well-controlled structure and reproducible photoluminescence properties. Room-temperature ozone oxidation produced samples with an attenuated photoluminescence emission. A high coverage of the surface by hydroxyl groups was obtained even though some Si–H groups were still present on the oxidized surface. The silane chemistry required to bind the biological recognition elements to the samples plays a significant role in the generation of photoluminescence in the green portion of the visible light spectrum. Nitrogen-containing organosilanes were required to obtain the photoluminescence. Additional studies, not reported here, have shown that the likely origin of this photoluminescence is the generation of a fluorophore species on the surface by nitrogen-containing molecules, primarily organosilanes, coming into close proximity to silicon atoms on the sample surface. Several-fold increases in photoluminescence were consistently observed on gas-etched porous samples following immobilization of oligonucleotides on amino-containing silanes and lower fluorescence intensities were observed on GPTMSfunctionalized samples. We have developed a simple and robust fluorescent model assay to demonstrate sensitive and specific oligonucleotide capture. Photoluminescence was monitored at each step of the assay in order to validate process reproducibility, stability, and assay
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
performance. This data proved to be valuable in establishing interpretations relating to the underlying principles of the BioAlloyTM platform. Longer oligonucleotide probes (i.e. 70 bases compared to 25 bases) were best immobilized by drying and UV cross-linking to GPTMS-functionalized samples and were consistently better at capturing 25mer targets. Detection of such single-labeled fluorescent oligonucleotide targets displayed more than 100fold specificity in hybridizations at room temperature in as little as 5 min. Sensitivity was estimated at 1 nM, i.e. 3 fmole, in a 1 h assay per mm2 . Single-nucleotide discrimination was also achieved in less than 15 min at 60 ◦ C with fluorescent oligonucleotide targets. Gas-etched Bio-AlloyTM samples were demonstrated to operate in a variety of complex media. The development of a fluorescence assay as well as other studies we conducted indicate the potential for high-surface area chemically-functionalized PSi to be a sensitive, selective, reproducible and stable DNA biosensor substrate. Acknowledgements The authors would like to thank IatroQuest’s R&D and Pilot Line team members for their contribution to the advancement and better understanding of Bio-AlloyTM . Partial funding for this study was provided by In-Q-Tel. References [1] R.A. Young, Biomedical discovery with DNA arrays, Cell 102 (2000) 9–15. [2] M.O. Noordewier, P.V. Warren, Gene expression microarrays and the integration of biological knowledge, Trends Biotechnol. 10 (2001) 412–415. [3] T.R. Hughes, D.D. Shoemaker, DNA microarrays for expression profiling, Curr. Opin. Chem. Biol. 1 (2001) 21–25. [4] F.C. Holstege, E.G. Jennings, J.J. Wyrick, T.I. Lee, C.J. Hengartner, M.R. Green, T.R. Golub, E.S. Lander, R.A. Young, Dissecting the regulatory circuitry of a eukaryotic genome, Cell 95 (1998) 717–728. [5] D. Gerhold, T. Rushmore, C.T. Caskey, DNA chips: promising toys have become powerful tools, Trends Biochem. Sci. 5 (1999) 168–173. [6] J. Marx, Disease genes clarify cholesterol trafficking, Science 22 (2000) 2227–2229. [7] P.A. Clarke, R. te Poele, R. Wooster, P. Workman, Gene expression microarray analysis in cancer biology, pharmacology, and drug development: progress and potential, Biochem. Pharmacol. 62 (2001) 1311–1336. [8] J. Zhou, Microarrays for bacterial detection and microbial community analysis, Curr. Opin. Microbiol. 6 (2003) 288–294. [9] J. Wang, From DNA biosensors to gene chips, Nucleic Acids Res. 28 (2000) 3011–3016. [10] B. Meric, K. Kerman, D. Ozkan, P. Kara, A. Erdem, O. Kucukoglu, E. Erciyas, M. Ozsoz, Electrochemical biosensor for the interaction of DNA with the alkylating agent 4,4 -dihydroxy chalcone based on guanine and adenine signals, J. Pharm. Biomed. Anal. 30 (2002) 1339–1346. [11] K. Dor´e, S. Dubus, H.A. Ho, I. Levesque, M. Brunette, G. Corbeil, M. Boissinot, G. Boivin, M.G. Bergeron, D. Boudreau, M. Leclerc, Fluorescent polymeric transducer for the rapid, simple, and specific detection of nucleic acids at the zeptomole level, J. Am. Chem. Soc. 126 (2004) 4240–4244. [12] S. Chan, Y. Li, L.J. Rothberg, B.L. Miller, P.M. Fauchet, Nanoscale silicon microcavities for biosensing, Mater. Sci. Eng. C15 (2001) 277–282. [13] V.S.-Y. Lin, K. Motesharei, K.-P.S. Dancil, M.J. Sailor, M.R. Ghadiri, A porous silicon-based optical interferometric biosensor, Science 278 (1997) 840–843.
229
[14] E. Kim, K. Kim, H. Yang, T.Y. Kim, J. Kwak, Enzyme-amplified electochemical detection of DNA using electrocatalysis of ferrocenyl-tethered dendrimer, Anal. Chem. 75 (2003) 5665–5672. [15] H. Xie, C. Zhang, Z. Gao, Amperometric detection of nucleic acid at femtomolar levels with a nucleic acid/electrochemical activator bilayer on gold electrode, electronic detection of DNA by its intrinsic molecular charge, Anal. Chem. 76 (2004) 1611–1617. [16] J. Fritz, B.E. Cooper, S. Gaudet, K.P. Sorger, S.R. Manalis, Electronic detection of DNA by its intrinsic molecular charge, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 14142–14146. [17] J.-I. Hahm, C.M. Lieber, Direct ultrasensitive electrical detection of DNA sequence variations using nanowire nanosensors, Nano Lett. 4 (2004) 51–54. [18] J.J. Storhoff, D.A. Lucas, V. Garimella, P.Y. Bao, R.U. M¨uller, Homogeneous detection of unamplified genomic DNA sequences based on colorimetric scatter of gold nanoparticles probes, Nat. Biotechnol. 22 (2004) 883–887. [19] A. Cs´aki, P. Kaplanek, R. M¨oller, W. Fritzche, The optical detection of individual DNA-conjugated gold nanoparticle labels after metal enhancement, Nanotechnology 14 (2003) 1262–1268. [20] E.S. Lander, Array of hope, Nat. Genet. 21 (Suppl.) (1999) 3–4. [21] O. Bisi, S. Ossicini, L. Pavesi, Porous silicon: a quantum sponge structure for silicon based optoelectronics, Surf. Sci. Rep. 38 (2000) 1–126. [22] A.G. Cullis, L.T. Canham, P.D.J. Calcott, The structural and luminescence properties of porous silicon, J. Appl. Phys. 82 (1997) 909–965. [23] L.T. Canham, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 249–255 (chapter 9.1). [24] P.M. Fauchet, Photoluminescence and electroluminescence from porous silicon, J. Lumin. 70 (1996) 294–309. [25] L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997. [26] L.T. Canham, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 83–88 (chapter 2.1). [27] R. Herino, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 89–96 (chapter 2.2). [28] M.J. Sailor, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 364–370 (chapter 12.4). [29] M.P. Stewart, J.M. Buriak, Chemical and biological applications of porous silicon technology, Adv. Mater. 12 (2000) 859–869. [30] S. Chan, S.R. Horner, P.M. Fauchet, B.L. Miller, Identification of gram negative bacteria using nanoscale silicon microcavities, J. Am. Chem. Soc. 123 (2001) 11797–11798. [31] C. Baratto, G. Faglia, G. Sberveglieri, Z. Gaburro, L. Pancheri, C. Oton, L. Pavesi, Multiparametric porous silicon sensors, Sensors 2 (2002) 121–126. [32] M. May, Building a better biosensor, The Scientist 18 (2004) 36. [33] L. De Stefano, I. Rendina, L. Moretti, A.M. Rossi, A. Lamberti, P. Arcari, in: P. Siciliano (Ed.), Sensors for Environmental Control, World Scientific, Singapore, 2003, pp. 46–50. [34] V.M. Starodub, L.L. Fedorenko, A.P. Sisetskiy, N.F. Starodub, Control of myoglobin level in a solution by an immune sensor based on the photoluminescence of porous silicon, Sens. Actuators 58 (1999) 409–414. [35] D.C. Burkman, D. Deal, D.C. Grant, C.A. Peterson, in: W. Kern (Ed.), Handbook of Semiconductor Wafer Cleaning Technology—Science, Technology and Applications, William Andrew Publishing/Noyes, 1993, pp. 111–151 (chapter 3). [36] M. Wu, T. Cubaud, C.-M. Ho, Scaling law in liquid drop coalescence driven by surface tension, Phys. Fluids 16 (2004) L51–L54. [37] J.L. DeRisi, V.R. Iyer, P.O. Brown, Exploring the metabolic and genetic control of gene expression on a genomic scale, Science 278 (1997) 680–686. [38] S.-H. Chiu, M. Hsu, W.-C. Ku, C.-Y. Tu, Y.-T. Tseng, W.-K. Lau, R.-Y. Yan, J.-T. Ma, C.-M. Tzeng, Synergistic effects of epoxy- and amine-silanes on microarray DNA immobilization and hybridization, Biochem. J. 374 (2003) 625–632. [39] N. Zammatteo, L. Jeanmart, S. Hamels, S. Courtois, P. Louette, L. Hevesi, J. Remacle, Comparison between different strategies of covalent attachment of DNA to glass surfaces to build DNA microarrays, Anal. Biochem. 280 (2000) 143–150.
230
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
[40] Y. Belosludtsev, Y. Belosludtsev, B. Iverson, S. Lemeshko, R. Eggers, R. Wiese, S. Lee, T. Powdrill, M. Hogan, DNA microarrays based on noncovalent oligonucleotide attachment and hybridization in two dimensions, Anal. Biochem. 292 (2001) 250–256. [41] D.R. Call, D.P. Chandler, F. Brockman, Fabrication of DNA microarrays using unmodified oligonucleotide probes, Biotechniques 30 (2001) 368–379. [42] V. Chan, J.D. Graves, E.S. McKenzie, The biophysics of DNA hybridization with immobilized oligonucleotide probes, Biophys. J. 69 (1995) 2243–2255. [43] A.W. Peterson, R.J. Heaton, R.M. Georgiadis, The effect of surface probe density on DNA hybridization, Nucleic Acids Res. 29 (2001) 5163–5168.
Biographies Daniel C. Tessier holds an MScA in Microbiology and Immunology from McGill University, Montreal, Canada (1985). He is currently the R&D manager of the Genomics Group at IatroQuest Corporation, leading activities in the fields of genomics and biosensors.
Samir Boughaba holds a doctorate degree in Materials Science and Engineering from Institut National Polytechnique, Grenoble, France (1993). He is currently the R&D manager of the Advanced Materials Group at IatroQuest Corporation, leading activities on the development of materials and processing techniques for biosensing applications. M´elanie Arbour holds an MSc in Biomedical Sciences from Universit´e de Montr´eal, Montreal, Canada (2001). She is currently a scientist, within the Genomics Group of IatroQuest Corporation, working on the development of DNA biosensors. Her interests are molecular biology and genomics. Pieter Roos holds a BSc in Biochemistry/Microbiology from Larenstein International Agriculture College, Wageningen, The Netherlands (1992). During the course of this study, he held a position of senior scientist within the Genomics Group of IatroQuest Corporation. His interests are analytical chemistry and molecular biology. Guan Pan holds a doctorate degree in Materials Science and Engineering from The University of Tokyo, Japan (1992). During the course of this study, he held a position of senior scientist within the Advanced Materials Group of IatroQuest Corporation, investigating the oxidation of porous silicon.
Improved surface sensing of DNA on gas-etched porous silicon D.C. Tessier, S. Boughaba ∗ , M. Arbour, P. Roos, G. Pan IatroQuest Corporation, 1000 Chemin du Golf, Verdun, QC H3E 1H4, Canada Received 30 March 2005; received in revised form 14 January 2006; accepted 5 February 2006 Available online 15 March 2006
Abstract A new gas-based etching process has been used to produce porous silicon (PSi) samples as sensor substrates. The as-etched porous material demonstrated an intense photoluminescence signal at an average wavelength of 658 nm (σ = 1%). The photoluminescence signal, with a peak wavelength in the range 530–540 nm, was found to increase with the length and concentration of unmodified oligonucleotide probes on surfaces coated with either amino-propyl trimethoxysilane (APTMS) or glycidoxy-propyl trimethoxysilane (GPTMS). The large surface area that characterized this material significantly increased the binding of DNA probes and led to increased sensitivity in target detection. Using fluorescently-labeled DNA, the target capture for sub-micromolar detection on this novel sensor substrate was demonstrated to be 100 times higher than that obtained on non-etched silicon. In a 1 h assay, we were able to detect specifically approximately 3 fmole of a 25mer target oligonucleotide per mm2 of substrate. © 2006 Elsevier B.V. All rights reserved. Keywords: Porous silicon; Etching; Photoluminescence; DNA; Silane; Sensors
1. Introduction Accurate and sensitive DNA detection tools have become pivotal to address fundamental questions relating to gene expression and regulatory biology [1–4], but also for disease diagnosis [5,6], drug discovery [7] and the identification of pathogens present in environmental samples [8]. This growing need to have access to specific DNA sequence information has prompted the development of faster, more sensitive and affordable nucleic acid biosensors [9]. Optical DNA detection methods include the use of various intercalating dyes [10,11] or rely on spectral interference of reflected white light [12,13]. Amperometric or electrochemical biosensors use a variety of electrode types functionalized with DNA to monitor enzymatic reactions occurring upon hybridization of the desired target [14,15]. Some electrical biosensors are based on monitoring the intrinsic molecular charge of DNA on field-effect sensors [16] or upon hybridization of a target molecule on PNA probes immobilized on silicon nanowires [17]. Other detection technologies utilize functionalized gold nanoparticles to detect hybridization events optically [18] or
∗
Corresponding author. Tel.: +1 613 995 7752. E-mail address: [email protected] (S. Boughaba).
0925-4005/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2006.02.002
electrically following silver enhancement [19]. Many of the DNA detection strategies deployed today are based on their hybridization to probe molecules immobilized to solid surfaces [20]. The choice of the material and its chemical functionalization is of utmost importance to produce a sensitive and reliable biosensor. Silicon has been the premier material driving the microelectronics revolution due to its unique electrical, chemical, and mechanical properties. Because of its indirect energy band-gap, bulk silicon is a poor light emitter and this has restricted its use in light-emitting devices and optoelectronics. Porous silicon (PSi), on the other hand, has been demonstrated to yield efficient visible light emission at room temperature [21–23]; such a behavior is primarily attributed to the electron confinement in the nanocrystals that constitute the porous structure [21,24]. Since 1990, the prospect of extending the use of silicon to lightemitting devices and optoelectronics has triggered numerous investigations on PSi and its photo- or electroluminescence properties [21,25]. Furthermore, the large fraction of voids inside PSi results in a very large specific surface area, on the order of a few hundred m2 /cm3 [26,27]. Such considerable surface area, which provides a large number of potential binding sites, has generated significant interest in PSi for chemical and biosensing applications [21,28–33]. As such, PSi is used as a transducer whose photoluminescence, refractive index, or electrical conductivity
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
221
Fig. 1. Schematic of the gas etching setup.
properties, are modulated upon exposure to target molecules. Starodub et al. [34] were able to detect as little as 10 ng/ml of myoglobin using inherent properties of PSi both as a matrix and as a transducer of light. In this article, we report on the characterization of PSi substrates, fabricated using a new gas etching method, as an excellent surface enhancing material for the development of a sensitive DNA sensor. This new etching technique was developed as part of a novel biosensing platform, Bio-AlloyTM1 , which attempted to make use of the unique photoluminescence properties of PSi to detect label-free target agents in real-time. Chemically functionalized PSi was validated as a substrate with increased surface area and sensitivity for DNA detection. 2. Experimental 2.1. Fabrication of porous silicon by gas etching The gas etching technique consists of exposing silicon samples to a mixture of oxygen (O2 ) and nitrogen dioxide (NO2 ) gases and an acid vapor. The experimental setup is schematically presented in Fig. 1. 4 mm × 4 mm silicon samples were obtained from the dicing of 1 0 0 boron-doped p-type wafers (Shin-Etsu Handotai), whose electrical resistivity was 20 cm. The samples were loaded onto a tray, which was mounted at the bottom of a chamber. After installing a gas distribution plate, whose role was to improve the uniformity of the gas flow, the chamber was hermetically sealed. The chamber, tray, and distribution plate were made of chemically inert TeflonTM2 . Pure oxygen (99.995%, Air Liquide) was flowed through a heated scrubber containing HF (47–51%, TraceMetal grade, Fisher Scientific), then merged with a flow of diluted nitrogen dioxide (2% in air, MEGS) before entering the etching chamber. The
1 2
Bio-Alloy is a registered trademark of IatroQuest Corporation. Teflon is a registered trademark of DuPont.
outlet of the chamber was connected to a scrubber containing a 2 N sodium hydroxide (NaOH) solution made from electrolytic pellets (certified ACS grade, Fisher Scientific). The role of the NaOH solution was to neutralize the HF. The O2 and NO2 gas flow rates could be varied in the range of 100–500 ml/min and 10–50 ml/min, respectively, using flowmeters (Cole-Parmer), while the HF scrubber could be kept at room temperature or heated up to 70 ◦ C. The NO2 cylinder was heated at its base to a temperature of 40 ◦ C to avoid accumulation of nitrogen dioxide at its bottom and to enhance the mixing of NO2 and air. Similarly, the stainless steel tubing connecting the NO2 cylinder to the chamber was heated to a temperature of 30 ◦ C to avoid condensation of NO2 on the tubing wall. For all the investigations reported in this article, samples were etched for 30 min. Following the etching process, the samples were rinsed using ethyl alcohol (95%, undenatured grade, Commercial Alcohols); the substrates were dipped in ethyl alcohol for 5 min, then removed and left to dry in a nitrogen environment for about 30 min. Prior to gas etching, the 4 mm × 4 mm silicon samples were first cleaned using RCA-type hydrogen peroxide mixtures [35], etched in a 5% HF solution for 2 min, rinsed in deionized water for 5 min, then oxidized at room temperature in a 3 SLM flow rate of ozone (O3 ) gas (280 g/N m3 concentration in nitrogen) for 5 min. At the end of the oxidation step, the samples were immediately loaded in the chamber and the etching was started. This cleaning sequence allowed a stringent control of the sample surface, yielding a hydrophilic surface. Spectroscopic ellipsometry measurements (Scientific Computing International, FilmTek 2000SE) revealed an oxide film thickness comparable to that of silicon native oxide (<2 nm). The morphology of the porous layers was investigated by scanning electron microscopy (SEM, Hitachi, S-4700), while their photoluminescence properties were characterized using a photo-detection system (Fluorosense, PDS) whose excitation source was a diode laser (λ = 470 nm), with an on-sample spot size of about 2 mm. A 500 nm longpass filter was inserted at the entrance of the spectrometer to eliminate any contribu-
222
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
tion to the signal from the source. Moreover, the thickness and porosity of the layers were simultaneously evaluated by spectral reflectance in the range 400–900 nm (Scientific Computing International, FilmTek 2000M) using an on-sample probing beam of 10–20 m diameter. SCI’s generalized dispersion formula was used to model fitted values of the dielectric function ε(λ); silicon and void were selected as the constituents of the porous layers. 2.2. Oxidation of gas-etched porous silicon The surface of as-etched luminescent PSi is hydrogenpassivated [21]. Silanization of such a surface requires the formation of silanols, i.e. Si–OH intermediate bonds. In order to provide hydroxyl group coverage of the surface of gasetched samples, room-temperature ozone oxidation was performed after the ethanol rinsing/drying step. In this method, samples were exposed for 30 min, at room temperature, to a 3 SLM flow rate of high-purity O3 + N2 mixture produced by an O3 generator (MKS, AX8560). The oxidation was carried out in a hermetically sealed chamber, and the O3 concentration was 280 g/N m3 . The photoluminescence properties of the oxidized porous layers were investigated using the PDS, while their porosity and thickness were evaluated by spectral reflectance. For the latter, silicon, silicon dioxide, and void were selected as the constituents of the layers. Fourier transform infrared (FTIR) spectroscopy analyses were performed, in the 400–4000 cm−1 range and transmission mode, on the oxidized porous samples to determine the type of chemical bonds present (Bio-Rad, FTS3000).
and with different 5 and/or 3 terminal modifications and were all purified by HPLC (Integrated DNA Technologies). BACA-AS (25mer) IQT2 (75mer)
IQT4 (25mer) IQT15 (70mer)
IQT15M1 (70mer)
IQT15M2 (70mer)
IQT15M3 (70mer)
IQT20 (70mer)
IQT39 (25mer):
5 -C6-amino-CCCTCGAGATTCCATGCCTATTACA-3 5 -C6-amino-TGTAATAGGCATGGAATCTCGAGGG-TGTAATAGGCATGGAATCTCGAGGG-TGTAATAGGCATGGAATCTCGAGGG-3 5 -CCCTCGAGATTCCATGCCTATTACA-3 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGGGCTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (complementary to IQT20) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGGACTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for a central point mutation) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGAACTGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for two central mutations) 5 -ACTTGGAGCAAGCTATATTAGTTTGCCATGGTGAAATGGCCAAGCATTGTTGGCACTCTTACGCCAGATG-3 (similar to IQT15 except for three central mutations) 5 -CATCTGGCGTAAGAGTGCCAACAATGCTTGGCCAGCCCACCATGGCAAACTAATATAGCTTGCTCCAAGT-3 (complementary to IQT15) 5 -ATGCTTGGCCAGCCCACCATGGCAA6-FAM-3 (complementary to the central portion of IQT15)
2.5. Immobilization of DNA 2.3. Silanization In order to produce a bio-compatible material, the surface of the PSi samples needed to be chemically modified to allow for the immobilization of DNA recognition elements. Oxidized porous substrates were silanized in the liquid phase for 1 h in the presence of 0.1% amino-propyl trimethoxysilane (APTMS) or glycidoxy-propyl trimethoxysilane (GPTMS) (Sigma-Aldrich) diluted in 100% toluene. Following this incubation, the substrates were rinsed once with toluene, once with 1:1 toluene:ethanol and twice with ethanol. The silanized substrates were then dried at 110 ◦ C for 10 min and stored in a dry nitrogen atmosphere until use. 2.4. List of oligonucleotides Here is the list of all the oligonucleotide sequences that were used in the present study. While they differed primarily in their length, some contained mutations to demonstrate singlenucleotide discrimination of the assay. Bold characters indicate the 25-base region of sequence homology/complementarity between IQT15, IQT20 and IQT39. Italic and underlined characters indicate the location of the point mutations between IQT15 and IQT15M1, M2 and M3, and their complementary target. Oligonucleotides were purchased in a variety of lengths
In general, 50 M dilutions of the oligonucleotide capture probes were prepared in water immediately prior to use. A volume of 2 l of spotting solution (∼100 pmole DNA) was applied to the center of the silanized substrates to eventually cover most of the 4 mm × 4 mm surface. Spotted substrates were incubated in microtiter plates at room temperature for 2 h; 1 h covered and 1 h uncovered to dry. UV cross-linking was carried out at 600 mJ in a UV Stratalinker1800TM (Stratagene). The substrates were then washed four times with 200 l of H2 O and air-dried for about 15 min. 2.6. Target hybridizations Hybridizations were usually carried out by placing the probeimmobilized substrates in individual wells of microtiter plates. The substrates were covered with 50 l of a target solution in 5× SSC (75 mM sodium citrate + 750 mM NaCl). The microtiter plates were then incubated at 37 ◦ C for 1 h with agitation (60 rpm). Following hybridization, the remaining target solution was removed by aspiration and the substrates were washed three times with 200 l of 1× SSC (15 mM sodium citrate + 150 mM NaCl), three times with 200 l of 0.1× SSC (1.5 mM sodium citrate + 15 mM NaCl), air-dried, and scanned for fluorescence.
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
2.7. Complex media A culture of Escherichia coli K12 was grown to saturation in LB broth overnight at 37 ◦ C. After centrifugation, the cell pellet was resuspended in the original culture volume of 10× SSC, boiled for 10 min and sonicated at maximum power in a sonicating bath (Branson, 2510) for 30 s. The lysate was mixed with an equal volume of target DNA. Purified E. coli genomic DNA was purchased (United States Biochemicals) and diluted to 42 g/ml in 5× SSC with the target to evaluate the impact of high DNA content on the efficiency of hybridization. Approximately 1 g of local soil was mixed with 1 ml of 50 mM ethylenediamine tetraacetic acid (EDTA, Sigma-Aldrich) in water, pH 8.0, and then boiled for 5 min. The mixture was centrifuged at 17000 × g for 2 min and the supernatant “dirty” water was used to dilute the fluorescent target. All complex mixtures were boiled for 5 min and cooled on ice for 5 min immediately prior to hybridization with the substrates. 2.8. Photoluminescence and fluorescence measurements At certain steps, the PSi substrates were monitored for fluorescence (6-FAMTM ; Abmax , 495 nm, Emmax , 520 nm) or photoluminescence using the photo-detection system (Fluorosense,
223
PDS). The emission spectra were acquired in the 400–900 nm range with an integration time of 200–3000 ms. 3. Results 3.1. Gas etching of porous silicon samples Fig. 2 presents scanning electron micrographs of a sample etched using O2 and NO2 flow rates of 100 and 10 ml/min, respectively. Etching resulted in the formation of a porous layer, which consisted of islands (Fig. 2 (b–e)) embedded in a matrix of different morphology (Fig. 2 (f)). The islands are likely related to the condensation of reactant-laden drops onto the surface during the etching process. Such a condensation could nucleate at surface defects, such as particulates. The island morphology presented in Fig. 2 is representative of all samples etched at various conditions. The density and size of the islands were, however, found to be affected by the O2 :NO2 flow rate ratio for a given O2 flow rate; the island size increased when decreasing the O2 :NO2 ratio, while the island density decreased concurrently. Fig. 3 presents scanning electron micrographs of two samples etched at O2 :NO2 ratios of 20 and 6.7. In this case, the average island diameter was found to be of 71 ± 17 m and 160 ± 25 m for a ratio of 20 and 6.7, respectively. The corresponding density of
Fig. 2. Scanning electron micrographs of a silicon sample etched using O2 and NO2 flow rates of 100 and 10 ml/min, respectively. HF scrubber kept at room temperature. (a) Low magnification of entire 4 mm × 4 mm sample; (b–e) increasing magnifications of the surface inside an island; and (f) surface between islands.
224
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 3. Scanning electron micrographs of silicon samples etched using O2 and NO2 flow rates of (a–c) 200 and 10 ml/min, respectively, and (d–f) 200 and 30 ml/min, respectively. HF scrubber kept at room-temperature. Micrographs (c) and (f) represent the areas between islands.
islands was determined to be 115 and 43 mm−2 , respectively. The size of the islands is directly related to the size of the drops, which depends on the surface tension and liquid drop coalescence driven by surface tension [36]. Such dependence directly relates to the chemical composition of the drops, which varies when changing the O2 :NO2 flow rate ratio. As illustrated in Fig. 3, the morphology inside the islands was also affected when changing the O2 :NO2 ratio; a more pronounced pyramidal texture, related to the anisotropic etching of silicon, was observed. Conversely, the morphology of the matrix surrounding the islands was not affected by changes in the O2 :NO2 flow rate ratio. Regardless of the flow parameters, the morphology of the porous matrix surrounding the islands was found to be different from that inside the islands. A smoother surface was observed, revealing an isotropic etching, which is characteristic of a process occurring in the vapor phase. Two concurrent etching modes could thus be considered to describe the gas etching process: a liquid-based one, occurring within the areas covered by condensed drops, and a gas phase-based one, taking place between the drops. Spectral reflectance measurements revealed an average porosity, i.e. void fraction, of 88% (σ = 3.1%) inside the islands and 79% between the islands (σ = 3.5%) for all gas flow rates. No dependence of the porosity on the etching conditions was observed. The spectral reflectance measurements also permitted the determination of the porous layer thickness within and between islands. The thickness was found to have an average value of 1230 nm (σ = 16%) inside the islands, and 301 nm (σ = 22%) in the matrix. Such a large difference in thickness was confirmed by focused ion beam cross-sections (Hitachi, FB2000A). The disparity in layer thickness and porosity between the two areas further substantiates the assumption of two different etching modes. Fig. 4 (a) shows the photoluminescence spectrum of a sample etched using O2 and NO2 flow rates of 400 and 20 ml/min, respectively. Intense PL in the red portion of the visible light
spectrum was obtained. Fig. 5 presents the photoluminescence peak intensity and wavelength as a function of O2 and NO2 gas flow rates. For each set of flow rates, the average of four samples was calculated. The average peak wavelength was determined to be 658 nm (σ = 1%). There was no dependence of the average peak wavelength on the processing conditions, in agreement with the fact that comparable porosity was obtained for all conditions. Furthermore, no dependence of the PL peak intensity on the gas flow rate was observed; the average PL peak intensity was found to be 8677 counts/s (2603 counts for an integration time of 300 ms), with a standard deviation of 11%. For all the investigations reported in the subsequent sections, porous samples were fabricated using O2 and NO2 flow rates of 300 and 30 ml/min, respectively, and an HF solution (in scrubber) kept at a temperature of 50 ◦ C.
Fig. 4. Photoluminescence spectra measured on a sample etched using O2 and NO2 flow rates of 400 and 20 ml/min, respectively, (a) after etching and (b) after oxidation. HF scrubber kept at temperature of 50 ◦ C. Integration time during PL measurements, 300 ms after etching and 3 s after oxidation.
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 5. Average photoluminescence peak intensity and wavelength as a function of the O2 and NO2 flow rates. HF scrubber kept at room temperature. For each set of flow rates, the average of four samples was calculated. Integration time during PL measurements, 300 ms.
3.2. Oxidation of gas-etched porous silicon samples Oxidized samples were characterized in terms of PL properties, surface chemistry, and composition. For all etching conditions, severe attenuation of the PL emission was observed as illustrated in Fig. 4 (b). Such attenuation in PL is attributed to the poor electronic quality of the interface between the remnant silicon and SiO2 , i.e. to trap states at the interface that are efficient luminescence quenchers [22]. The quenching of the original PL on oxidized samples is beneficial with respect to fluorescence-based detection of specifically immobilized DNA or biomolecules. The original PL spectra, obtained after etching, had broad peaks (Fig. 4 (a)) originating from a distribution of the size of the crystallites inside the porous layers [22]. Such a large luminescence bandwidth (∼120 nm), if present, would impact fluorescence signals related to biological detection events, even if the signal collection was carried out within a narrow wavelength window through the use of bandpass filters. An FTIR spectrum representative of all oxidized samples is presented in Fig. 6. The spectrum exhibited absorption bands
Fig. 6. FTIR spectrum measured on an oxidized porous sample. Etching was performed using O2 and NO2 flow rates of 400 and 20 ml/min, respectively, with the HF scrubber kept at a temperature of 50 ◦ C.
225
Fig. 7. Photoluminescence spectra of gas-etched porous samples following silanization with 1% APTMS or GPTMS. Integration time, 3 s.
at 1640 and 3400 cm−1 , which are characteristic of hydroxyl groups. It also revealed the presence of residual Si–H bonds through the presence of an absorption band at 2260 cm−1 . Furthermore, the spectrum was characterized by a siloxane band at 1105 cm−1 , with a shoulder at 1076 cm−1 , attributed to Si–O–Si asymmetric stretching vibrational mode. Other absorption bands were observed at 816, 883, and 970 cm−1 ; they were assigned to Si–O–Si bending mode, a combination of symmetric Si–H deformation and Si–O stretching in Si–OH groups, and Si–O stretching mode in either negatively charged non-bridging oxygen (Si–O− ) or Si–OH groups, respectively. The determination of the silicon dioxide, silicon, and void fractions by spectral reflectance revealed that the porous layers were only partially oxidized, in agreement with the FTIR results. The average ratio of silicon to silicon dioxide fractions after oxidation was determined to be 0.64 ± 0.02 inside the islands and 0.62 ± 0.11 between the islands. 3.3. Effect of silanization on photoluminescence The effect of silanes on the photoluminescence properties of Bio-AlloyTM samples was thoroughly investigated. Two classes of silanes were used to immobilize DNA on the nanostructured substrates: amino-containing silanes, typically APTMS, for electrostatic interaction with the phosphate backbone of DNA, and the epoxy-containing silane GPTMS, for covalent attachment of amino-modified or unmodified DNA to the substrates. Silane coupling experiments revealed the importance of the nature of the silane on the resulting photoluminescence of PSi. Photoluminescence spectra of APTMS and GPTMS-treated samples are presented in Fig. 7. Intense photoluminescence, with a peak wavelength of 530–540 nm, was observed after silanization with APTMS, whereas very little photoluminescence, with a peak wavelength in the range 620–660 nm, could be observed after silanization with GPTMS. The weak photoluminescence using GPTMS could be attributed to residual red photoluminescence following the oxidation step. It was, however, possible to artificially generate photoluminescence in the green portion
226
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
of the light spectrum, at a peak wavelength of 533 nm, on oxidized porous samples treated with 0.1% GPTMS by spotting and baking the amino acid l-lysine or poly-l-lysine. Water and l-norleucine, a non-natural amino acid lacking the -amino group of l-lysine did not produce this green photoluminescence (data not shown). This green photoluminescence might originate from the covalent bonds created between the epoxy group of the GPTMS and the primary amines of l-lysine and polyl-lysine. These amino-containing, poly-l-lysine green photoluminescent surfaces were later shown to immobilize unmodified fluorescently-labeled DNA to the same extent as poly-l-lysinecoated glass slides [37]. 3.4. Probe immobilization and effect on photoluminescence In order to characterize the binding of DNA probes to the surface of PSi samples, 25mer oligonucleotides carrying fluorescent tags (TAMRA) at their 3 -end were applied to silanized substrates; a range of APTMS and GPTMS concentrations were evaluated to maximize the immobilization of the probes. Concurrently, for the same silane concentrations, photoluminescence was monitored using unlabeled oligonucleotides. Maximum loading and retention of labeled probes was achieved between 0.001 and 1% silane coating concentrations, whereas maximum photoluminescence of the PSi substrates exposed to unlabeled probes was measured between 0.01 and 0.1% silane coating concentrations, with a maximum peak wavelength in the range 530–540 nm (data not shown). Similar DNA binding and photoluminescence characteristics were observed with PSi substrates silanized with other amino-containing organosilanes. Consistent photoluminescence was demonstrated following immobilization of unmodified probes on silanized surfaces. The intensity of this photoluminescence correlated with the size and the amount of immobilized DNA. Fig. 8 illustrates the photoluminescence associated with the immobilization of 25 and 100 pmole of 25mers compared to 75mers on 0.01% APTMS (Fig. 8 a) and to a lesser extent on 0.1% GPTMS after baking (Fig. 8 b). Increasing amounts of fluorescently-labeled probes were immobilized on samples functionalized with 1% APTMS or GPTMS. These experiments revealed that covalent attachment to GPTMS-coated porous surfaces retained almost twice as much probe as the electrostatic interaction on APTMS. Consequently, GPTMS was selected for the subsequent immobilization and hybridization studies. 3.5. Target hybridizations on GPTMS Unmodified oligonucleotides of different lengths were spotted on GPTMS-treated porous samples and dried before hybridization with 25mer fluorescent targets. The effect of UV cross-linking of 25 and 70mer probes on hybridization efficiency was studied. It was found that an irradiation energy ranging from 0 to 1200 mJ had no significant effect on the hybridization to 25mer probes. On the other hand, at least four-fold increase in fluorescence signal intensity was obtained with 70mer probes when they were exposed to 600 mJ of UV light compared to no treatment, indicating a more efficient retention of probes
Fig. 8. Photoluminescence of 25 and 75mer probes on 0.01% APTMS (a) and 0.1% GPTMS (b). Two microliter of 12.5 or 50 M IQT4 or BACA-AS or IQT2 DNA were spotted and treated as described in Section 2. Integration time, 3 s.
on the porous surface. This positive effect of UV cross-linking on hybridization to longer probes was also observed on mixed amine/epoxy surfaces [38]. Under optimal hybridization conditions for both 25 and 70mer probes, we found that the longer probes yielded three to five times more fluorescence signal. This has also been reported by other groups [39]. The effect of probe loading and its impact on the efficiency of hybridization was also evaluated. Fig. 9 shows that spotting 500 M of a fluorescent 25mer probe (1 nmole or 4 × 1013 molecules/mm2 ) was not sufficient to saturate the surface of the porous sample. This probe density is 100-fold higher than 3 × 1011 molecules/mm2 reported for the immobilization of 12mer oligonucleotides on flat silica APTMS-treated surfaces [40]. Maximum target hybridizations were nonetheless obtained with 50–100 M spotting, i.e. 100–200 pmole of probe (Fig. 9). This probe-dependent levelling off in target hybridizations is supported by a published study on epoxy-silanized flat surfaces [41], and by the demonstration from other groups that dense physical packing of probes on solid phases dramatically reduces the efficiency and the kinetics of target hybridization [42,43]. It was of interest to determine the effect of temperature on the kinetics of hybridization. Three temperatures were selected: room temperature (20–22 ◦ C), 37 ◦ C, and 65 ◦ C. Specific hybridizations occurred within 5 min at all three temperatures, and the intensities continued to increase for up to about
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 9. Effect of probe loading on the efficiency of hybridization. Fluorescentlylabeled probes were spotted in parallel on etched and non-etched substrates and washed as described in Section 2. Hybridization of 0.5 M IQT39 target on complementary 70mer probes (IQT15) on 0.1% GPTMS-coated substrates in 5× SSC at 65 ◦ C for 1 h. Background and non-specific signals were measured to be less than 25 fluorescent counts. Integration time, 500 ms.
227
Fig. 11. Single base-pair discrimination in hybridization using 0.5 M IQT21 target on 70mer probes containing a fully complementary central portion to the target (IQT15), single mismatch (IQT15M1), two base-pair mismatch (IQT15M2) and three base-pair mismatch (IQT15M3) on 0.1% GPTMS samples. Hybridizations were carried out in 5× SSC in a thermocycler under the conditions described in the figure. Integration time, 500 ms.
2 h as illustrated in Fig. 10. Hybridizations at 65 ◦ C were significantly accelerated compared to room temperature and 37 ◦ C. Although it was possible to demonstrate highly specific and sensitive hybridizations using fluorescently-tagged targets, we were not able to produce similar photoluminescence responses under these optimized conditions. Slight variations of this protocol were used to discriminate single-nucleotide differences between probes. Mutated 70mer sequences containing single, double and triple base-pair mismatches were differentiated from wild-type using a 25mer labeled target complementary to the central segment of these probes (see Section 2). Samples were incubated in 5× SSC in a microtiter plate, inside a thermocycler for approximately 40 min, then washed and scanned for associated fluorescence (Fig. 11). The results showed a five-fold increase in specificity for the fulllength probe as compared to a single base-pair mutant probe.
Further addition of 15% formamide to the 5× SSC concomitant with a reduction of the final assay temperature to 60 ◦ C improved the discrimination for mismatched probes to more than 10-fold in less than 15 min. The sensitivity of the assay was evaluated by determining the lower limit of detection of a 25mer fluorescent target. The lower limit of detection was defined as the lowest target concentration detectable having more than two standard deviations above average background fluorescence, as measured from pools of non-specific hybridizations. Although the assay exhibited specific detection of target sequences from a 500 nM hybridization mixture in as little as 5 min at room temperature (Fig. 10), it was possible to increase the sensitivity of the assay by 500-fold (to 1 nM) by increasing the assay temperature to 65 ◦ C for 1 h (Fig. 12). The results displayed a linear response over approximately two logs. Above 100 nM of target, most of the sam-
Fig. 10. Effect of temperature on hybridization efficiency. Hybridization of 0.5 M IQT39 target on complementary 70mer probes (IQT15) on 0.1% GPTMS-coated substrates. Targets were diluted in 5× SSC and hybridized to the substrates at room temperature, 37 or 65 ◦ C for 5 min to 4 h. Background and non-specific signals were determined to be less than 25 fluorescence units. Integration time, 100 ms.
Fig. 12. Determination of the lower limit of detection and comparison between etched and non-etched silicon surfaces for identical hybridization assays (IQT15 probe, IQT39 target, 5× SSC, 65 ◦ C, 1 h). The dashed line represents the same hybridization assay carried out on etched substrates in the presence of a boiled and sonicated lysate of E. coli. Background and non-specific signals were measured to be less than 25 fluorescence units.
228
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
Fig. 13. Hybridization of 0.5 M IQT39 target in complex media on complementary and non-complementary 70mer probes on 0.1% GPTMS-coated substrates. Unless otherwise stated in Section 2, hybridization mixtures were prepared by combining three parts of the indicated media with one part of 20× SSC to which was added a negligible volume of target. All complex mixtures were boiled for 5 min and cooled on ice for 5 min before hybridization at 65 ◦ C for 1 h. Integration time, 200 ms.
ples exhibited saturated signals for the integration time used to acquire the fluorescent data. The same experiment was repeated on non-etched silicon samples to compare the sensitivity of the assay between porous and flat silicon surfaces. Non-etched silicon samples were washed and immediately oxidized before silanization with 0.1% GPTMS; thus repeating the same oxidation procedure as for the porous samples. Maximum signals obtained on non-etched samples were 100-fold lower than those obtained on etched samples, with severe consequences on specificity of the assay. In fact, etched samples displayed specific/nonspecific signal ratios higher than 100 compared to ratios smaller than 3 for non-etched samples using 100 nM target. The presence of a boiled and sonicated E. coli lysate did not significantly affect the performance of the assay (Fig. 12, dashed line). The assay was tested in a variety of complex media to assess the robustness of the assay and the stability of the PSi samples. Comparable signal intensities were obtained after hybridizing 0.5 M of the IQT39 target (6-FAM-labeled 25mer target) in the presence of 42 g of E. coli genomic DNA, soil-drained “dirty” water, rabbit serum, human urine, a buccal swab suspension or lake water compared to deionized water (Fig. 13). To avoid possible degradation of DNA by nucleases, 50 mM EDTA was added to the soil-drained water. The only test medium which significantly affected the sensitivity and specificity of the assay was the urine sample. The stability of GPTMS-silanized and DNA-spotted porous samples was tested over a period of 1 month. In a first experiment, silanized porous samples were stored at room temperature, in a dry nitrogen environment, for a given period of time before immobilization of the probe. A complete hybridization assay was then performed and results were compared to other storage time points (Fig. 14). The fluorescence response of GPTMSsilanized material increased over the first 15 days and returned to its original level after 1 month without affecting specificity. In a second experiment, GPTMS-silanized porous samples pre-
Fig. 14. Ageing study of PSi samples. Gas-etched substrates were silanized with 0.1% GPTMS and stored desiccated under N2 atmosphere for the indicated times. At these time points, complementary and non-complementary 70mer probes were immobilized, UV cross-linked to the substrates, and hybridized with 0.5 M IQT39 in 5× SSC at 65 ◦ C for 1 h. Integration time, 100 ms.
spotted with 100 pmole of probe were stored under identical conditions and tested by hybridization at the same time points. The fluorescence response of pre-spotted samples did not change in the first week but steadily decreased over the following 3 weeks to reach ∼33% efficiency after 1 month. Specificity was not affected by long-term storage of pre-spotted samples indicating minimal degradation of capture probes on the nanostructured samples. 4. Discussion and conclusions The development of a new gas etching technique for nanostructuring of silicon resulted in the fabrication of porous samples with well-controlled structure and reproducible photoluminescence properties. Room-temperature ozone oxidation produced samples with an attenuated photoluminescence emission. A high coverage of the surface by hydroxyl groups was obtained even though some Si–H groups were still present on the oxidized surface. The silane chemistry required to bind the biological recognition elements to the samples plays a significant role in the generation of photoluminescence in the green portion of the visible light spectrum. Nitrogen-containing organosilanes were required to obtain the photoluminescence. Additional studies, not reported here, have shown that the likely origin of this photoluminescence is the generation of a fluorophore species on the surface by nitrogen-containing molecules, primarily organosilanes, coming into close proximity to silicon atoms on the sample surface. Several-fold increases in photoluminescence were consistently observed on gas-etched porous samples following immobilization of oligonucleotides on amino-containing silanes and lower fluorescence intensities were observed on GPTMSfunctionalized samples. We have developed a simple and robust fluorescent model assay to demonstrate sensitive and specific oligonucleotide capture. Photoluminescence was monitored at each step of the assay in order to validate process reproducibility, stability, and assay
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
performance. This data proved to be valuable in establishing interpretations relating to the underlying principles of the BioAlloyTM platform. Longer oligonucleotide probes (i.e. 70 bases compared to 25 bases) were best immobilized by drying and UV cross-linking to GPTMS-functionalized samples and were consistently better at capturing 25mer targets. Detection of such single-labeled fluorescent oligonucleotide targets displayed more than 100fold specificity in hybridizations at room temperature in as little as 5 min. Sensitivity was estimated at 1 nM, i.e. 3 fmole, in a 1 h assay per mm2 . Single-nucleotide discrimination was also achieved in less than 15 min at 60 ◦ C with fluorescent oligonucleotide targets. Gas-etched Bio-AlloyTM samples were demonstrated to operate in a variety of complex media. The development of a fluorescence assay as well as other studies we conducted indicate the potential for high-surface area chemically-functionalized PSi to be a sensitive, selective, reproducible and stable DNA biosensor substrate. Acknowledgements The authors would like to thank IatroQuest’s R&D and Pilot Line team members for their contribution to the advancement and better understanding of Bio-AlloyTM . Partial funding for this study was provided by In-Q-Tel. References [1] R.A. Young, Biomedical discovery with DNA arrays, Cell 102 (2000) 9–15. [2] M.O. Noordewier, P.V. Warren, Gene expression microarrays and the integration of biological knowledge, Trends Biotechnol. 10 (2001) 412–415. [3] T.R. Hughes, D.D. Shoemaker, DNA microarrays for expression profiling, Curr. Opin. Chem. Biol. 1 (2001) 21–25. [4] F.C. Holstege, E.G. Jennings, J.J. Wyrick, T.I. Lee, C.J. Hengartner, M.R. Green, T.R. Golub, E.S. Lander, R.A. Young, Dissecting the regulatory circuitry of a eukaryotic genome, Cell 95 (1998) 717–728. [5] D. Gerhold, T. Rushmore, C.T. Caskey, DNA chips: promising toys have become powerful tools, Trends Biochem. Sci. 5 (1999) 168–173. [6] J. Marx, Disease genes clarify cholesterol trafficking, Science 22 (2000) 2227–2229. [7] P.A. Clarke, R. te Poele, R. Wooster, P. Workman, Gene expression microarray analysis in cancer biology, pharmacology, and drug development: progress and potential, Biochem. Pharmacol. 62 (2001) 1311–1336. [8] J. Zhou, Microarrays for bacterial detection and microbial community analysis, Curr. Opin. Microbiol. 6 (2003) 288–294. [9] J. Wang, From DNA biosensors to gene chips, Nucleic Acids Res. 28 (2000) 3011–3016. [10] B. Meric, K. Kerman, D. Ozkan, P. Kara, A. Erdem, O. Kucukoglu, E. Erciyas, M. Ozsoz, Electrochemical biosensor for the interaction of DNA with the alkylating agent 4,4 -dihydroxy chalcone based on guanine and adenine signals, J. Pharm. Biomed. Anal. 30 (2002) 1339–1346. [11] K. Dor´e, S. Dubus, H.A. Ho, I. Levesque, M. Brunette, G. Corbeil, M. Boissinot, G. Boivin, M.G. Bergeron, D. Boudreau, M. Leclerc, Fluorescent polymeric transducer for the rapid, simple, and specific detection of nucleic acids at the zeptomole level, J. Am. Chem. Soc. 126 (2004) 4240–4244. [12] S. Chan, Y. Li, L.J. Rothberg, B.L. Miller, P.M. Fauchet, Nanoscale silicon microcavities for biosensing, Mater. Sci. Eng. C15 (2001) 277–282. [13] V.S.-Y. Lin, K. Motesharei, K.-P.S. Dancil, M.J. Sailor, M.R. Ghadiri, A porous silicon-based optical interferometric biosensor, Science 278 (1997) 840–843.
229
[14] E. Kim, K. Kim, H. Yang, T.Y. Kim, J. Kwak, Enzyme-amplified electochemical detection of DNA using electrocatalysis of ferrocenyl-tethered dendrimer, Anal. Chem. 75 (2003) 5665–5672. [15] H. Xie, C. Zhang, Z. Gao, Amperometric detection of nucleic acid at femtomolar levels with a nucleic acid/electrochemical activator bilayer on gold electrode, electronic detection of DNA by its intrinsic molecular charge, Anal. Chem. 76 (2004) 1611–1617. [16] J. Fritz, B.E. Cooper, S. Gaudet, K.P. Sorger, S.R. Manalis, Electronic detection of DNA by its intrinsic molecular charge, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 14142–14146. [17] J.-I. Hahm, C.M. Lieber, Direct ultrasensitive electrical detection of DNA sequence variations using nanowire nanosensors, Nano Lett. 4 (2004) 51–54. [18] J.J. Storhoff, D.A. Lucas, V. Garimella, P.Y. Bao, R.U. M¨uller, Homogeneous detection of unamplified genomic DNA sequences based on colorimetric scatter of gold nanoparticles probes, Nat. Biotechnol. 22 (2004) 883–887. [19] A. Cs´aki, P. Kaplanek, R. M¨oller, W. Fritzche, The optical detection of individual DNA-conjugated gold nanoparticle labels after metal enhancement, Nanotechnology 14 (2003) 1262–1268. [20] E.S. Lander, Array of hope, Nat. Genet. 21 (Suppl.) (1999) 3–4. [21] O. Bisi, S. Ossicini, L. Pavesi, Porous silicon: a quantum sponge structure for silicon based optoelectronics, Surf. Sci. Rep. 38 (2000) 1–126. [22] A.G. Cullis, L.T. Canham, P.D.J. Calcott, The structural and luminescence properties of porous silicon, J. Appl. Phys. 82 (1997) 909–965. [23] L.T. Canham, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 249–255 (chapter 9.1). [24] P.M. Fauchet, Photoluminescence and electroluminescence from porous silicon, J. Lumin. 70 (1996) 294–309. [25] L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997. [26] L.T. Canham, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 83–88 (chapter 2.1). [27] R. Herino, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 89–96 (chapter 2.2). [28] M.J. Sailor, in: L. Canham (Ed.), Properties of Porous Silicon, INSPEC, London, 1997, pp. 364–370 (chapter 12.4). [29] M.P. Stewart, J.M. Buriak, Chemical and biological applications of porous silicon technology, Adv. Mater. 12 (2000) 859–869. [30] S. Chan, S.R. Horner, P.M. Fauchet, B.L. Miller, Identification of gram negative bacteria using nanoscale silicon microcavities, J. Am. Chem. Soc. 123 (2001) 11797–11798. [31] C. Baratto, G. Faglia, G. Sberveglieri, Z. Gaburro, L. Pancheri, C. Oton, L. Pavesi, Multiparametric porous silicon sensors, Sensors 2 (2002) 121–126. [32] M. May, Building a better biosensor, The Scientist 18 (2004) 36. [33] L. De Stefano, I. Rendina, L. Moretti, A.M. Rossi, A. Lamberti, P. Arcari, in: P. Siciliano (Ed.), Sensors for Environmental Control, World Scientific, Singapore, 2003, pp. 46–50. [34] V.M. Starodub, L.L. Fedorenko, A.P. Sisetskiy, N.F. Starodub, Control of myoglobin level in a solution by an immune sensor based on the photoluminescence of porous silicon, Sens. Actuators 58 (1999) 409–414. [35] D.C. Burkman, D. Deal, D.C. Grant, C.A. Peterson, in: W. Kern (Ed.), Handbook of Semiconductor Wafer Cleaning Technology—Science, Technology and Applications, William Andrew Publishing/Noyes, 1993, pp. 111–151 (chapter 3). [36] M. Wu, T. Cubaud, C.-M. Ho, Scaling law in liquid drop coalescence driven by surface tension, Phys. Fluids 16 (2004) L51–L54. [37] J.L. DeRisi, V.R. Iyer, P.O. Brown, Exploring the metabolic and genetic control of gene expression on a genomic scale, Science 278 (1997) 680–686. [38] S.-H. Chiu, M. Hsu, W.-C. Ku, C.-Y. Tu, Y.-T. Tseng, W.-K. Lau, R.-Y. Yan, J.-T. Ma, C.-M. Tzeng, Synergistic effects of epoxy- and amine-silanes on microarray DNA immobilization and hybridization, Biochem. J. 374 (2003) 625–632. [39] N. Zammatteo, L. Jeanmart, S. Hamels, S. Courtois, P. Louette, L. Hevesi, J. Remacle, Comparison between different strategies of covalent attachment of DNA to glass surfaces to build DNA microarrays, Anal. Biochem. 280 (2000) 143–150.
230
D.C. Tessier et al. / Sensors and Actuators B 120 (2006) 220–230
[40] Y. Belosludtsev, Y. Belosludtsev, B. Iverson, S. Lemeshko, R. Eggers, R. Wiese, S. Lee, T. Powdrill, M. Hogan, DNA microarrays based on noncovalent oligonucleotide attachment and hybridization in two dimensions, Anal. Biochem. 292 (2001) 250–256. [41] D.R. Call, D.P. Chandler, F. Brockman, Fabrication of DNA microarrays using unmodified oligonucleotide probes, Biotechniques 30 (2001) 368–379. [42] V. Chan, J.D. Graves, E.S. McKenzie, The biophysics of DNA hybridization with immobilized oligonucleotide probes, Biophys. J. 69 (1995) 2243–2255. [43] A.W. Peterson, R.J. Heaton, R.M. Georgiadis, The effect of surface probe density on DNA hybridization, Nucleic Acids Res. 29 (2001) 5163–5168.
Biographies Daniel C. Tessier holds an MScA in Microbiology and Immunology from McGill University, Montreal, Canada (1985). He is currently the R&D manager of the Genomics Group at IatroQuest Corporation, leading activities in the fields of genomics and biosensors.
Samir Boughaba holds a doctorate degree in Materials Science and Engineering from Institut National Polytechnique, Grenoble, France (1993). He is currently the R&D manager of the Advanced Materials Group at IatroQuest Corporation, leading activities on the development of materials and processing techniques for biosensing applications. M´elanie Arbour holds an MSc in Biomedical Sciences from Universit´e de Montr´eal, Montreal, Canada (2001). She is currently a scientist, within the Genomics Group of IatroQuest Corporation, working on the development of DNA biosensors. Her interests are molecular biology and genomics. Pieter Roos holds a BSc in Biochemistry/Microbiology from Larenstein International Agriculture College, Wageningen, The Netherlands (1992). During the course of this study, he held a position of senior scientist within the Genomics Group of IatroQuest Corporation. His interests are analytical chemistry and molecular biology. Guan Pan holds a doctorate degree in Materials Science and Engineering from The University of Tokyo, Japan (1992). During the course of this study, he held a position of senior scientist within the Advanced Materials Group of IatroQuest Corporation, investigating the oxidation of porous silicon.