|Home | About | Journals | Submit | Contact Us | Français|
Fabrication of microarray devices using traditional glass slides is not easily adaptable to integration into microfluidic systems. There is thus a need for the development of polymeric materials showing a high hybridization signal-to-background ratio, enabling sensitive detection of microbial pathogens. We have developed such plastic supports suitable for highly sensitive DNA microarray hybridizations. The proof of concept of this microarray technology was done through the detection of four human respiratory viruses that were amplified and labeled with a fluorescent dye via a sensitive reverse transcriptase PCR (RT-PCR) assay. The performance of the microarray hybridization with plastic supports made of PMMA [poly(methylmethacrylate)]-VSUVT or Zeonor 1060R was compared to that with high-quality glass slide microarrays by using both passive and microfluidic hybridization systems. Specific hybridization signal-to-background ratios comparable to that obtained with high-quality commercial glass slides were achieved with both polymeric substrates. Microarray hybridizations demonstrated an analytical sensitivity equivalent to approximately 100 viral genome copies per RT-PCR, which is at least 100-fold higher than the sensitivities of previously reported DNA hybridizations on plastic supports. Testing of these plastic polymers using a microfluidic microarray hybridization platform also showed results that were comparable to those with glass supports. In conclusion, PMMA-VSUVT and Zeonor 1060R are both suitable for highly sensitive microarray hybridizations.
Plastic microfluidic systems represent promising technologies for the miniaturization and cost-effective mass production of medical diagnostic devices. These devices can be used to replace existing assays or implement novel assays for the detection of specific nucleic acids, proteins, or antibodies. One widely used approach for the detection of microbial pathogens is based on the detection of their genomic material using DNA microarray hybridization. Our group has previously developed a centrifuge-based microfluidic platform in a compact-disk format for rapid DNA microarray hybridization on standard glass slides (20). Considering that fabrication of microarray devices using the traditional glass support is not well suited for integration into a low-cost, portable micro-total-analysis system (μ-TAS), there is great interest in the development of polymeric materials showing a high hybridization signal-to-background ratio. Plastic material suitable for use in DNA/RNA-based molecular diagnostic applications must (i) be inexpensive, (ii) possess excellent micromachining properties, (iii) be suitable for mass production, (iv) be able to withstand the high temperatures required during PCR thermocycling, and (v) possess excellent optical properties, such as low intrinsic fluorescence background and high optical transparency to the excitation and detection wavelengths employed in microarray technologies.
Plastic material selection for DNA microarray hybridization, capture probe immobilization, and surface activation have been investigated previously (2, 4-6, 9, 10, 13, 19, 25). Most studies dealing with hybridization on plastic supports reported only data for the hybridization of DNA to target complementary oligonucleotides (4-6, 10, 13, 19, 25). Therefore, their usefulness for molecular diagnostic purposes, in which the target nucleic acids are mainly PCR-amplified genetic materials, has not been demonstrated.
PCR-amplified bacterial and human genomic DNAs have both been used in plastic microarray hybridizations (11, 15, 17, 22). However, these hybridizations were carried out using complex detection methods based on the use of radioactive isotopes, enzymatic assays, or biotin-labeled targets, which are incompatible with rapid and simple molecular diagnostics. Furthermore, these investigators have not reported any data on analytical sensitivity. Others have undertaken the chemical functionalization of plastics, allowing the use of fluorescently labeled amplicons. However, their detection technology using biochannel microarrays showed poor analytical sensitivities (approximately 10,000 PCR-amplified targets) (11).
We recently identified two suitable commercially available plastics, PMMA [poly(methylmethacrylate)]-VSUVT and Zeonor 1060R, and developed methods for surface chemistry and covalence immobilization of DNA oligonucleotides, allowing ultrasensitive detection of fluorescently labeled nucleic acid targets on plastic microarrays (2). Efficient chemical functionalization of plastic substrates for the immobilization of DNA probes was achieved.
In the present study, we have developed plastic supports for highly sensitive (≤100 copies of PCR-amplified targets) DNA microarray hybridizations. The proof of concept of this plastic biochip technology was done through sensitive detection of several important human respiratory viral pathogens for molecular diagnostic purposes. The performance of hybridization with plastic supports made of PMMA-VSUVT or Zeonor 1060R was compared with that of high-quality glass-slide microarrays by using both passive and active (microfluidic) hybridization systems. This plastic microarray hybridization technology may be integrated into disposable microfluidic devices, thereby paving the way toward a fully integrated μ-TAS for sensitive nucleic acid detection.
Influenza A virus (IAV) (CCRI-15590, vaccine strain A/Panama/5502/98), enterovirus (EV) (CCRI-15531; clinical sample from 2004 to 2005), respiratory syncytial virus (RSV) (CCRI-15552; clinical sample from 2004 to 2005), and metapneumovirus (MPV) (CCRI-15597; clinical sample from 1997) were cultured on the MDCK (ATCC CCL-34), HT-29 (ATCC HTB-38), HEP-2 (ATCC CCL-23), and LLC-MK2 (ATCC CCL-7) cell lines, respectively. Viral RNAs were purified from cell culture supernatants by using the MagaZorb RNA isolation kit (Cortex Biochem, Concord, MA) on a KingFisher ML instrument (Thermo Scientific, Waltham, MA). The concentration of the purified viral RNA was determined by measurement of the optical density at 260 nm in an Ultrospec 2000 spectrophotometer (Pharmacia Biotech, Piscataway, NJ), and RNA quality was assessed with both a 2100 Bioanalyzer system and an RNA 6000 nano-LabChip (Agilent, Palo Alto, CA).
For sensitivity assays, the target genes were cloned using the TOPO TA cloning system (Invitrogen, Burlington, Ontario, Canada) (Table (Table1).1). RNA transcription was performed using the Ampliscribe T7-Flash transcription kit (Epicentre, Madison, WI). The RNA transcripts were purified and quantified, and their quality was assessed, as described above for the viral RNA.
Reverse transcriptase PCR (RT-PCR) primers and microarray capture probes (Tables (Tables1,1, ,2,2, and and3)3) were designed using alignments of multiple sequences analyzed with the Seqlab Editor program from the GCG Wisconsin package (version 10.3; Accelrys). OLIGO software (version 6.1; Molecular Biology Insights, Cascade, CO) was used to check the compatibility of the primers. Multiplex RT-PCR using primers labeled with Cy3 at their 5′ ends (Table (Table2)2) was performed using the OneStep RT-PCR kit (Qiagen, Mississauga, Ontario, Canada). Purified RNA transcripts were used for analytical sensitivity assays, while viral RNAs purified from cell culture supernatants were used for microfluidic microarray hybridizations and slide stability tests. One microliter of purified viral RNA or of purified RNA transcripts per RT-PCR was used. Cycling conditions on a PTC-200 thermal cycler were as follows: 30 min at 50°C for reverse transcription; 15 min at 95°C for reverse transcriptase inactivation and DNA polymerase activation; and 45 cycles of 15 s at 95°C for denaturation, 10 s at 54°C for annealing, and 25 s at 72°C for extension.
Polymer-based slides of PMMA-VSUVT and Zeonor 1060R were fabricated via injection molding (2, 10, 19, 25). Carboxylic acid groups were generated at the surfaces of PMMA-VSUVT slides by immersing the slides in a saturated aqueous solution of sodium hydroxide for at least 10 h. This surface treatment effectively induces hydrolysis of the methyl ester groups at the surfaces of the PMMA slides. On the other hand, carboxylic acid groups were formed at the surfaces of Zeonor 1060R slides via an oxidative process in the presence of ozone for 11 min. The oxidation was carried out with an ozone generator from Ozomax, Inc. (Ozo 2VTT), set at 50% power and an oxygen flow rate of 3.5 standard cubic feet per hour (2, 10, 19, 25).
Carboxylic acid groups were activated using a 0.417 M solution of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and a 0.173 M solution of N-hydroxysuccinimide (NHS) in phosphate-buffered saline solution for 45 min prior to arraying with amino-modified probes. During the activation step, the carboxylic acids react with NHS to form the corresponding NHS esters, which in turn react efficiently with the amino groups from the DNA probes to form an amide linkage.
Oligonucleotide capture probes (Table (Table3)3) targeting the four selected viruses were arrayed onto PMMA-VSUVT and Zeonor 1060R plastic slides as well as onto high-quality commercial glass slides (Aldehyde Plus arraying slides; Genetix, Boston, MA). Chemically activated plastic slides of PMMA-VSUVT and Zeonor 1060R were used to print the control and virus-specific capture probes. Oligonucleotide probe solutions at 60 μM in water were diluted twofold in Array-IT Micro Spotting Solution Plus (MSSP) (Telechem International, Sunnyvale, CA). Oligonucleotide probes were spotted onto both glass and plastic slides using a Virtek SDDC2 arrayer equipped with SMP2 pins (Telechem International, Sunnyvale, CA). After spotting, the slides were incubated at 23°C under 60% humidity for 4 h. The glass slides were subsequently washed for 5 min in a solution containing 0.2% NaBH4, 20% ethanol, 0.8% NaCl, 0.02% KCl, 0.144% Na2HPO4, and 0.024% KH2PO4 (pH 7.4). Plastic slides were washed in the same washing solution without NaBH4. Glass and plastic slides were both rinsed five times with ultrapure water.
Cy3-labeled RT-PCR amplicons (5 μl) were mixed with 15 μl of hybridization buffer (8.4% NaCl, 12.5% NaH2PO4·H2O, 3.0% EDTA, 0.04% polyvinylpyrrolidone, and 40% formamide). Passive hybridization was performed for 1 h at room temperature (23°C) by using 20-μl (15- by 13-mm) Hybri-well hybridization chambers (Sigma-Aldrich). Amplicons produced by RT-PCR were denatured by heating at 95°C for 5 min in hybridization buffer. The slides were subsequently washed with 0.2× SSPE (0.03 mol/liter NaCl, 2 mmol/liter NaH2PO4·H2O, 0.2 mmol/liter EDTA [pH 7.4]) containing 0.1% sodium dodecyl sulfate for 5 min and then rinsed with 0.2× SSPE for 5 min under continuous shaking (300 rpm).
Conditions for microfluidic flowthrough microarray hybridization using a compact-disk platform were those previously described (20) except that the hybridization time at room temperature was reduced from 5 to 3 min. Fluorescent images of all slides were obtained with a ScanArray 4000 XL microarray scanner (GSI Lumonics/Packard Biochips, Billerica, MA), and data were analyzed with GenePix Pro (version 6.0; MDS Analytical Technologies, Downingtown, PA). All microarray hybridization experiments were validated using an amino-modified 20-mer control oligonucleotide (A-S-bbc1 [5′-AGGATAGGCAGACCATACTC-3′]) hybridized to a fully complementary Cy3-labeled 20-mer target oligonucleotide (C3-bbc1a [5′-Cy3-GAGTATGGTCTGCCTATCCT-3′]). Arrays with a median signal of 5,000 fluorescence units or more for the control oligonucleotide were considered valid, while arrays with lower signals were discarded. All hybridization signals were corrected for background (the median background was 100 fluorescence units or fewer) and were then expressed as a percentage of the control oligonucleotide signal. This normalization allowed the removal of variations in fluorescence signal intensities resulting from potential slide-to-slide heterogeneity. Negative controls consisted of spotted MSSP.
Genomic DNA from Homo sapiens (Hsap-11) as well as from five common respiratory bacterial pathogens—Haemophilus influenzae (ATCC 9006), Streptococcus pneumoniae (ATCC 700673), Legionella pneumophila (ATCC 33152), Staphylococcus aureus (ATCC BAA-40), and Moraxella catarrhalis (ATCC 43628)—were used to test the specificities of the RT-PCR primers and capture probes by using the RT-PCR assay and the microarray hybridization procedure described above. Bacterial DNA was extracted using the Gnome DNA isolation kit (Qbiogene), while human DNA was isolated using the Magnesil KF genomic system (Promega, Madison, WI) on a KingFisher ML instrument (Thermo Scientific, Waltham, MA). Ten nanograms of purified bacterial or human genomic DNA per RT-PCR was used.
Zeonor 1060R slides were employed for testing the stabilities of the chemically activated plastic surfaces. The NHS esters can undergo hydrolysis, and therefore it is important to test their stability during storage. Freshly coated slides were stored in a desiccator for 2, 4, 6, 8, or 10 weeks prior to spotting with IAV capture probes and testing using passive microarray hybridization.
The cloned target genes for each of the four viruses were used to perform analytical sensitivity assays. One hundred, 500, and 1,000 copies of RNA transcripts from each virus were amplified by RT-PCR and then hybridized under passive conditions on PMMA-VSUVT, Zeonor 1060R, or glass printed supports. The amplicons generated from each concentration of viral RNA transcript were detected efficiently (Fig. (Fig.11 and and2).2). The laser scanner photomultiplier (PMT) gain, with fixed laser power at 80%, ranged from 69 to 77% for glass, from 67 to 77% for PMMA-VSUVT, and from 67 to 76% for Zeonor 1060R. The quality of the fluorescence signals for the PMMA-VSUVT and Zeonor 1060R supports was comparable to that obtained with the high-quality glass slide tested (Fig. (Fig.1).1). Zeonor 1060R and glass supports showed similar fluorescence signal intensities for the probes targeting EV, RSV, and MPV, while the intensity of the signal for the IAV probe was twofold higher with Zeonor 1060R than with glass (Fig. (Fig.2).2). Fluorescence signals were slightly stronger with the PMMA-VSUVT support.
Signal-to-background ratios for hybridizations performed with amplicons generated from 100 copies of RNA transcripts for each target virus were similar for all three supports (Fig. (Fig.3).3). The analytical sensitivity for passive microarray hybridization ranged from 25 to 100 copies of viral RNA transcripts per RT-PCR (Table (Table33).
The microarray of capture probes on Zeonor 1060R slides targeting the four respiratory viruses was used to verify probe specificity using passive hybridization. Selected capture probes were all specific to their respective target viruses, and no cross-hybridization was observed with human DNA or with DNAs from common respiratory bacterial pathogens (data not shown). The hybridization signals for the microarrays on plastic slides were similar to those measured with the commercial glass slides.
Microfluidic microarray DNA hybridizations on the compact-disk platform using the plastic and glass supports showed that purified target viral RNAs for each of the four viruses were efficiently PCR amplified and detected by their specific capture probes. The hybridization signal intensities were comparable to those obtained under the passive hybridization conditions (Fig. (Fig.44 and and5).5). Normalized fluorescence signals with the microfluidic platform showed that the signal intensities obtained with PMMA-VSUVT slides were slightly stronger than those obtained with the Zeonor 1060R or glass slides (Fig. (Fig.5),5), as observed with passive hybridization experiments (Fig. (Fig.22).
Stability tests of functionalized Zeonor 1060R slides with NHS ester groups on their surfaces showed that chemically activated slides are stable during a period of at least 10 weeks upon storage in a desiccator (Fig. (Fig.6).6). Freshly activated slides served as controls, and the fluorescent signals were compared with those from slides that had surface chemicals aging from 2 to 10 weeks. The fluorescence intensities ranged from 96.8% to 103.9% during the 10-week period. Our data confirmed that this small variation was associated with slide-to-slide heterogeneity.
Microarray technologies have become a tool of choice for multiparametric diagnostic assays and for gene expression profiling (7, 14, 21, 24). The ability to perform sensitive and reproducible microarray hybridizations depends to a great extent on the quality of the microarray substrate. Microbiological diagnostic assays on glass slide microarrays for highly sensitive and specific detection of hepatitis B virus, hepatitis C virus, and human immunodeficiency virus type 1 in human blood samples have been demonstrated recently (8). Nonetheless, plastic materials are likely more promising than glass supports for the future of microarrays dedicated to microbiological diagnostics, because the former have excellent microfabrication properties and are more easily amenable to integration into low-cost, portable microanalysis systems. The development of microanalysis systems for the molecular detection of infectious disease agents requires the development of assays that are both rapid and highly sensitive. However, the high autofluorescence of standard polymeric materials is not compatible with a highly sensitive detection method that employs fluorescent targets. Plastic microarray hybridizations compatible with fluorescent dyes have been developed previously (5, 6, 13). However, these investigators reported only hybridizations with labeled oligonucleotides, and the specific hybridization signal-to-background ratios were relatively low. Others have avoided this issue by using nonfluorescent detection technologies such as radiolabeling (13, 22) or enzymatic reactions (17), but none of these detection technologies fulfill the diagnostic requirement in terms of both rapidity and safety.
In this work, we report ultrasensitive microarray hybridization of fluorescently labeled PCR-amplified target DNA using two high-quality plastic substrates for molecular diagnostic purposes. Both plastics (PMMA-VSUVT and Zeonor 1060R) were previously selected for their premium optical properties and low autofluorescence background (2). They have been specifically functionalized in order to reduce background hybridization and increase specific hybridization signals. We observed that the hybridization background signals were as low as, or even lower than, those measured with commercially available high-quality functionalized microarray glass slides. The amplicons generated via multiplex RT-PCR from as few as 25 to 100 viral particles of IAV, EV, RSV, or MPV could be detected and discriminated, respectively, on plastic microarrays using either 1-h passive hybridization or 3-min microfluidic hybridization without any amplicon purification step. Specific fluorescence signals were nonambiguous and were as high as (Zeonor 1060R), or higher than (PMMA-VSUVT), those obtained with a glass substrate. Similar results were obtained using the microfluidic platform (20). To our knowledge, this is by far (at least 100-fold) the lowest detection limit achieved with DNA microarray hybridizations on plastic supports. Indeed, most published studies of microarray hybridizations on plastic supports did not report any analytical sensitivity data (2, 5, 6, 9, 10, 12, 13, 16-18, 22, 23). Lenigk et al. (11) have described microfluidic hybridization of PCR-amplified, Cy- labeled staphylococcal DNA onto a polycarbonate microarray, detecting a minimum of 10,000 genome copies. This analytical sensitivity may be insufficient for reliable detection of microbial pathogens, which can be found at lower clinically relevant loads. For example, infections associated with viral loads as low as 100 viruses per ml of respiratory tract sample may be encountered (1, 3). A highly sensitive hybridization on a plastic microarray using DNA capture probes spotted onto microfabricated micropillars has been described recently (19). In this system, the background fluorescence from the area surrounding the pillars is rejected through confocal detection. An excellent signal-to-background ratio was obtained for hybridizations using labeled and reverse-transcribed total human RNA, but the fabrication and functionalization of those micropillars are far more complex than those of flat plastic substrates such as those described in the present study.
High background fluorescence on plastic supports is often due to nonspecific adsorption of the labeled target onto the substrate (4, 13, 19). In the present study, we report high signal-to-background ratios. On average, these ratios reached 250 for hybridization using 5 nM labeled oligonucleotide and 137 for hybridization of nonpurified labeled amplicons generated from 100 viral genome copies. This means that the background fluorescence represents as little as 0.4% or 0.7% of the signal, respectively. By comparison, Fixe et al. (5), using 40 times more labeled oligonucleotides (0.2 μM) on PMMA slides, reported a signal-to-background ratio of 90.
We previously described a centrifuge-based microfluidic system for rapid and efficient microarray hybridization on standard glass slides (20). In the present study, we obtained high fluorescence signal intensities and low background with both plastic and glass slides using this 3-min microfluidic microarray hybridization system, in which hybridizations are performed at room temperature (23°C) using nonpurified PCR products. The only other reported study of room temperature hybridization on plastic supports was performed using a 16-h protocol, and the hybridization signal-to-background ratios obtained were low (approximately 10) (4). Others have performed hybridization experiments on plastic supports at 30 to 65°C, requiring special heating devices to carry out the hybridizations (5, 6, 15, 17, 23, 25). In addition, stability tests demonstrated that when stored in a vacuum desiccator, the chemically activated plastic slides were stable for at least 10 weeks.
In conclusion, we have developed two functionalized plastic supports (PMMA-VSUVT and Zeonor 1060R) that allow efficient DNA microarray hybridizations at room temperature, with signal-to-background ratios and analytical sensitivities comparable to those obtained with high-quality standard glass slides. These plastic supports, which were shown to be highly stable, have been validated for microbiological diagnostic applications with both passive hybridization and active hybridization in a microfluidic system. Considering that both plastics are suitable for low-cost fabrication, this plastic microarray technology is promising for the development of fully integrated medical diagnostic devices, such as portable μ-TASs.
This work was supported by grants from Genome Canada and Génome Québec.
We thank Xavier Bouhy for viral culture preparations and Marie-Jeanne Fiola for microarray printing.
Published ahead of print on 10 September 2008.