Indian Journal of Medical Microbiology Medknow Publications on behalf of Indian Association of Medical Microbiology ISSN: 0255-0857 EISSN: 1998-3646 Vol. 26, Num. 1, 2008, pp. 13-20

Indian Journal of Medical Microbiology, Vol. 26, No. 1, January-March, 2008, pp. 13-20

Special Article

Fabrication and evaluation of a sequence-specific oligonucleotide miniarray for molecular genotyping

Iqbal J, Hanel F, Ruryk A, Limmon GV, Tretiakov A, Durst M, Saluz HP

Department of Pathology and Microbiology, Nebraska Medical Center, Omaha, NE
Correspondence Address:Department of Pathology and Microbiology, Nebraska Medical Center, Omaha, NE, jiqbal@unmc.edu

Date of Submission: 27-Jul-2007
Date of Acceptance: 22-Sep-2007

Code Number: mb08003

Abstract

Keywords: DNA microarray, genotyping, type-specific oligonucleotide hybridization

Past studies have shown clearly that molecular diagnostics has made a significant impact on disease management. ^[1] Novel molecular assays may significantly enhance our ability to quantitate genetic information about pathogens. Therefore, it is important to investigate different genetic analysis procedures being used in clinical laboratories. Clinical molecular assays should precisely reveal the identity of the genome because it is also relevant to the biological and clinical behavior of the disease.

Epidemiological studies have increasingly shown a strong association of human papillomaviruses (HPVs) with cervical cancers and benign lesions. ^[2] Genital HPV infections appear to be the most common sexually transmitted disease and a majority of them are induced by two HPV types: 6 and 11. ^[3] In addition, a group of at least 15-18 HPV subtypes (e.g. HPV 16, 18, 31, 33 and 45) are detected in about 80% of all cervical cancers and most precancerous lesions (cervical intraepithelial neoplasias, CIN 2/3). ^[4],[5] However, the pap smear test or hybrid capture II assay (HCA-II) which is the common screening program for HPV-related infection is limited in its ability by the number of identified subtypes. ^[6] HPV subtype detection requires the amplification of the E1 or L1 open reading frame (ORF) of the genome and gene identification by sequencing, restriction endonuclease digestion or by hybridization using sequence-specific radiolabeled probes. ^[7],[8]

Recognizing the potential of the DNA microarray technology, ^[9] we studied the fundamental features of this technology as a potential genotyping tool and evaluated its significance in molecular subtyping of HPV. The fundamental principle involves the highly selective nature of the complementary DNA double-helix formation and therefore provides a precise analytical power. ^[10] In our study, the DNA microchip assay was optimized to detect a single mismatch and various specificity and sensitivity issues were addressed in miniaturized conditions. To assess the application of this simplified assay, the procedure was also employed for HPV genotyping using a 149 basepair (bp) region of L1 ORF ^[11] from different cell lines and clinical specimens. The HPV subtypes were identified by hybridization with a fabricated oligonucleotide array, immobilized with DNA sequence specific for different HPV subtypes. This simplified procedure can be easily adapted for the identification of other pathogens with high sensitivity, specificity and at a reasonably low cost.

Materials and Methods

Preparation of aldehyde substrates for DNA immobilization

To prepare the chemically active substrates for DNA microarrays, glass slides-Borofloat-33, (Schott Inc, Jena, Germany) were cleaned in 5N NaOH for 2 h and rinsed vigorously with fresh distilled water. The washed slides were then exposed to a 1% ethanolic solution of 3-aminopropyl-trimethoxysilane (Sigma, Germany) for 30 min at room temperature with subsequent drying by centrifugation (1000xg ) for 3 min and baking at 110 °C for 30 min. The baked slides were incubated in 1% glutaraldehyde solution (Sigma, Germany) for 4-5 h at 3-4 °C. The coated slides were rinsed with fresh distilled water 5-6 times, dried by centrifugation and used immediately or stored in a light-protected vacuum chamber. The aldehyde coating was checked by a test print and hybridization to determine the slide batch quality. The slides were prepared in batches of 25. For comparative analysis of substrate coating, poly-L-lysine- or epoxy-coated slides were either purchased commercially or prepared as per the protocol described on the website: http://cmgm.stanford.edu/pbrown/protocols.

Critical factors for DNA binding to chip substrate

Spotting solutions

Several oligonucleotides [Table - 1] were suspended in a number of spotting solutions, e.g. (1) Commercial spotting solution (TeleChem International, Inc.), (2) 150 mm Sodium phosphate (pH 7.4), (3) 150 mm Sodium phosphate + 0.01% sodium dodecyl sulfate (SDS), (4) 50% dimethylsulfoxide (DMSO), (5) 5% DMSO + 5X sodium chloride and sodium citrate (SSC), (6) 5% DMSO + 150 mm sodium phosphate, (7) 5X SSC, (8) 5X SSC + 0.1% SDS, (9) 3X SSC + 0.01% SDS, (10) 150 mm sodium sulfate + 4% dextran, (11) 5X SSC + 4% dextran, (12) 4% dextran + 0.1% SDS, (13) 0.5 m 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer + 0.01% SDS, (14) 5X phosphate-buffered saline (PBS), (15) 5X PBS + 0.01% SDS, (16) 5X SSC + 1N NaOH, (17) 150 mm Sodium phosphate + IN NaOH, (18) 5X SSC + 4% sodium sulfate, (19) 150 mm sodium phosphate + 4% sodium sulfate and (20) tris (trishydroxymethylaminomethane)-ethylene diamine tetraacetic acid (EDTA), TE. The oligonucleotides were dissolved in above buffers and hybridized with the labeled probe for comparative analysis.

Optimal concentrations

The minimum concentration of the DNA oligonucleotide was determined by printing a range of oligonucleotide concentrations (10, 20, 40 and 80 pmol/µL) on the slides prepared as described above. The printing was performed with spotting pins CPM2 (Telechem, CA, USA) in a microarrayer (Gene Machines, CA, USA) at a relative humidity of 50-70%. The distance between the centers of the spots was 300-400 µm. After printing, the slides were kept in a humid chamber overnight. Printing was carried out in five replicates for each oligonucleotide.

Oligonucleotide sequences

To standardize the specificity and sensitivity issues of hybridization in a miniaturized assay, a set of two perfectly matched (PM) or two single mismatch (SMM) oligonucleotides complementary to each of the probes (at two different positions) were immobilized on a microarray [Table - 1]. All the oligonucleotides were synthesized by Interactive-Biotech (Ulm, Germany) with modified 5´ amino groups and with a 15-mer spacer of oligo-dT immediately following the amino group [Table - 1]. Two of the above oligomers were also synthesized without a spacer molecule for a comparative study. All the oligonucleotides were purified by high performance liquid chromatography (HPLC) and 100 mm solutions in distilled water stored as aliquots. The oligomers were diluted to a final concentration of 40 µm in 3.5X SSC in a 96-microwell plate in a total volume of 10 µL. To evaluate the microarray assay for HPV genotyping, two sets of oligonucleotides (30-mer and 15-mer) were immobilized on aldehyde substrates through the amino-linked ends [Table - 2].

Prehybridization and quality control

Printed oligonucleotide slides incubated in a humidified chamber overnight were used for prehybridization processing. The slides were washed with 0.2% SDS followed by rinsing and incubation in distilled water at 95-100°C for 2 min. They were subsequently incubated in sodium borohydride solution (1.3 g NaBH ₄ dissolved in 375 mL phosphate-buffered saline and 125 mL pure ethanol) for 15 min and washed with ethanol and distilled water. The slides were air-dried and stored in the dark at 25°C. The quality of immobilization or the efficiency of binding was checked by hybridization with a randomly labeled (Cy3) 9-mer oligonucleotide for 2-3 h at room temperature.

Preparation of labeled probes

Plasmids

Four different viral ORFs were cloned for standardization including the optimization of the interaction between the probe and target oligonucleotides on the DNA microarray. These viral genes were herpes simplex virus (HSV) DNA polymerase (UL 0.3-0.46 KB), HCMV DNA polymerase (UL 0.54-1.4 kB), DNA phosphotransferase (UL 0.97-1.1 kB), HPV-16 L 1 (150 bp) and HPV-18 L1 (150 bp). All clones were confirmed by sequencing prior to the generation of labeled probes.

Cell lines and clinical specimens

In addition to the above plasmids, genomic DNA was obtained from the HPV subtype-positive cell lines SiHa, MS751, HeLa and CaSKi and from the clinical samples as described previously by Dürst et al . ^[12] Clinical samples were obtained from the department of Gynecology and Obstetrics, Jena and had previously undergone HPV testing by HC-II (Digene Co, USA) and other standardized genotyping procedures including DNA sequencing of L1 ORF.

Various sizes of labeled probes ranging from 0.15-1.4 kB were generated by polymerase chain reaction (PCR) from the plasmids whereas a 149 bp fragment was generated from HPV-positive cell lines and clinical specimens. Internal amplification controls were included in clinical specimens to distinguish between true viral negatives containing human DNA and false viral negatives in samples that either lacked human DNA or that contained PCR inhibitors. β-globulin primers were used in a parallel reaction to check a 268 bp DNA fragment. PCR was carried out with PWO polymerase (Boehringer, Mannheim, Germany) and Amplitaq (Perkin Elmer, USA). Briefly, PCR reactions were carried out in 50 µL containing 0.01-0.1 µg template DNA from plasmids or 5-10 µL of crude suspensions from cervical specimens. Reaction mixtures contained the UL54 primer set (5´-GTCCCAGTCAG GA CGGCGT-3´ and Cy3-5´-CGAGAG CAC CGA GAC GCG-3´) or the UL97 primer set (5´-GCTGCAGAAGCTGCTCATC -3´ and Cy3-5´-GACACGAGGA CATCTTGG-3´) or the GP5⁺/GP6⁺ primer set (5-TTTGTTACTGGGTAGATACTAC-3 and Cy3-5´-GAAAAATAAA CTGTAA ATCATATTC-3) primers (50 pmoles each), PCR buffer; 2% DMSO and 200 µm dNTPs with the addition of 1-2.5 U of thermostable DNA polymerase (Amplitaq Perkin Elmer, USA). The standard PCR reaction was carried out in a volume of 25 µL for clinical samples and cell lines. The PCR products were purified with PCR purification kits (Qiagen, Germany), vacuum concentrated to 10 µL volume and used for hybridization. The PCR product was checked on agarose gels.

Hybridization, washing and scanning

The hybridization buffers tested were 3X SSC (containing 0.01% SDS) or 100 mm sodium phosphate buffer (pH 7.4) (Na ₂ HPO ₄ + NaH ₂ PO ₄ , 4:1), with competitor DNA, polyA RNA or yeast t-RNA (1 µg for each 10 µL of hybridization volume). In addition, the labeled probes were resuspended in UniHyb ^TM hybridization solution (Telechem, CA, USA) for comparative analysis and hybridization was carried out as per manufacturer′s instructions under a silanized cover slip (Hybri-slips, Sigma, Germany) for 3-4 h. The slides were washed in 2X SSC with 0.2% SDS for 15 min, 0.2X SSC for 5 min, 0.1X SSC for 2 min, dried by centrifugation for 5 min at 1500xg and finally air-dried for 15 min. Fluorescent signals from the microarray were scored using a ScanArray-3000 machine (GSI Lumonics, CA, USA). Scoring of the hybridization results was done by uploading the scanned image tiff file into the AIDA-analysis program (Raytest quantitation Package, Germany) and each feature was examined for fluorescence intensity. The background was calculated around each spot at a distance of two pixels from the recognized edge of the spot.

Results

Critical factors for optimization of DNA chip assay

(A) Substrate chemistry. The function of the DNA microarray is highly dependent on the quality of the solid support and we selected the aldehyde substrate for its cost-effectiveness and ease in preparation and handling. Aldehyde functional groups proved to be efficient substrates for the attachment of amino-linked 15-mer oligonucleotides as compared to lysine- and epoxy-based chemistries under our specified conditions. In addition, we observed that the oligo-dT ₍₁₅₎ spacer used increased the proximity of the probe molecule to the target. ^[13] Of significance was the observation that there were no specific hybridization signals with oligonucleotides directly immobilized without any spacer molecule (Supplemental [Figure - 1]).

(B) Spotting solutions and other technical factors. The purpose of using a spotting solution is to increase the DNA′s interaction with the chemically active substrate and to stabilize the DNA on the solid surface. We tested different compositions of spotting solutions and the results of the hybridization quality are summarized in [Figure - 1]. The commercial spotting solution gave comparable results with 3.5X SSC. The addition of 0.01% SDS in SSC increased the spotting size and helped in forming spherical spots. On the other hand, the use of sodium sulfate in spotting solutions decreased the spotting size and often did not yield any uniformly sized spots. We identified an optimal spotting solution to be 3.5X SSC + 0.01% SDS for both aldehyde and lysine substrates.

(C) Optimal concentration of immobilized (target) DNA. Target DNA molecules should be in excess, so that the labeled probe is direct measure of molecular abundance In our experimental procedure, 40-100 µm target DNA did not show any significant difference in signal intensity for hybridization between the target and probe sequences. These results were consistent with different kinds of substrates (aldehyde, lysine or epoxy). Nevertheless, a minimum target concentration of (40-60 pmol/µl) was observed to give an optimal hybridization signal (high signal-to-noise ratio) with different batches of slides and substrates [Figure - 2].

DNA spotting performed under lower humidity (< 50%) conditions decreases the spot size and also results in asymmetrical hybridization signals. The signals were often confined to the outer corner of the spot of a target molecule while at higher humidity (70-80%), the spot size increased nearly to an optimal diameter of 100-200 µm and the signal intensity was observed to be distributed over the whole DNA spot [Figure - 3]. The correct size and orientation of the spots are also essential for image data extraction.

Hybridization temperatures 5-10 °C below the T _m of the immobilized oligonucleotide gave specific hybridization signals in a time period of 2-3 h (15-mers at 40-42 °C, 20-mers at 44-46 °C and 30-mers at 46-48 °C). Hybridizations carried out under silanized cover slips gave good signal intensity with the least background. Washing followed by quick drying by centrifugation was also found to improve the quality of the image results.

(D) Single nucleotide mismatches discrimination. The arrays consisting of 15-30 bases require relatively stringent hybridization conditions to maintain a high degree of discrimination. Therefore, to optimize the conditions for single nucleotide mismatch discrimination on a solid surface under miniaturized conditions, the hybridizations of 15-mer target oligonucleotides were performed with short probes generated from plasmids [Table - 1]. For every labeled probe, four 15-mer target oligonucleotides were immobilized with two of them having one nucleotide mismatch in the middle. The mean signal intensity from the spots with a perfect match (PM) vs a single mismatch (SMM) gave> 2-fold discrimination [Figure - 4]A, B. The results were consistent with other oligonucleotides (UL54 ORF) and their respective probes. The nonspecific hybridization signal intensity assessed from unrelated oligonucleotides showed> 8-fold difference between a perfect match and an unrelated sequence.

Hybridization with HPV DNA microarray

Thirty-mer (30-mer) oligonucleotides of L1 ORF were selected from eight different HPV types (HPV 6, 11, 16, 18, 31, 33, 35 and 39) ^[7] [Table - 2]. The selected genotypes represent the major HPV subtypes prevalent globally including low-risk types HPV (6 and 11), intermediate risk (33, 35 and 39) and high-risk types (16, 18 and 31). ^[3],[4] We used Cy3 end-labeled Gp6⁺ primer and unlabeled GP5⁺ for PCR amplification because Cy3 is not only more stable than Cy5 but also the primer set amplifies a short 149 bp fragment that is easy to amplify and suitable for specific hybridization under stringent conditions. Besides, this primer set has a broad spectrum for HPV genotyping. ^[11] To evaluate the HPV microarray, labeled probes were sequentially generated from plasmids containing L1 ORF of HPV 16 and HPV 18, HPV-positive cell lines (CaSki, SiHa, HeLa and MS751) and well-characterized clinical samples. The results were obtained from two to three independent experimental conditions carried out on either single slides or two different ones. The quality of DNA from all the clinical samples was tested with b-globulin. In our controlled experiments, probes generated from the plasmids gave a highly specific (~> 5-fold on average) hybridization pattern as observed on a microarray [Figure - 5]A. Probes were generated from the genomic DNA of four HPV-positive cell lines including HeLa (10-50 copies of HPV 18), SiHa (1-2 copies of HPV 16), CaSki (500-600 copies of HPV 16) and MS751 (HPV 45 as negative control). Sequence-specific hybridization was observed from each cell line and no hybridization was observed from MS751, which carries HPV45. The sensitivity data from the above experiments revealed that even 1-2 viral copies per cell can be detected as in the case of the SiHa cell line in as little as 5-10 µL of hybridization volume.

In addition, to validate the hybridization results from several target sequences, the 30-mers of HPV 6, 16 and 18 were dissected into three 15-mers so that they all had similar GC content [Table - 3]. The use of multiple oligonucleotide target sequences for each genotype is advantageous in increasing the discrimination capabilities of the assay. Suitable sequence-specific hybridization was observed from three different features [Figure - 6]A, B. The 15-mers confirmed specific discrimination between individual subtypes. To test the possibility of using the procedure for obtaining a HPV profile in clinical samples, 24 clinical isolates carrying HPV-positive DNA were also tested on a microarray. Two to four independent hybridizations were evaluated from each sample. The genotype was determined if the target specific for a given subtype displayed >2-fold signal-to-noise ratio. All the HPV 16 (3/3), HPV 18 (2/2) and HPV 6 (2/2) samples showed a clear signal distinction upon hybridization with both 15-mers and 30-mers. Additionally, HPV 31 (1/2), HPV 35 (2/2), HPV 11 (1/3) and HPV 33 (2/2) samples were clearly discriminated from the other subtypes on the 30-mer oligo chips (examples shown in [Figure - 6]C). In some samples, we observed hybridizations with multiple sequences in a single assay (e.g. A clinical sample showed a specific hybridization signal from the HPV 6 as well as HPV 31 target sequence, similarly two other samples showed hybridization with HPV 11 and 35 target sequences). Thus, multiple infections can be observed on a single platform. The specimens carrying HPV DNA other than the target sequences present on the microarray showed no hybridization to any of the target sequences, thus acting as negative controls. Therefore, we were able to evaluate a sensitive and specific method of detection of pathogens in a clinical isolate with the microarray assay.

In conclusion, our study addressed fabrication and sensitivity issues of targeted DNA arrays and highlighted the need for further improvements for the detection of other pathogens and the need to make the use of microarray screening programs more feasible in diagnostic laboratories.

Discussion

The genome analysis of microbes in clinical specimens in a simple and rapid way can directly influence the diagnosis and management of infections by reducing the unnecessary therapy. ^[14],[15] However, due to the heterogeneous diversity associated with various pathogens, subspecies or subtype characterization of genome is essential for proper diagnosis. The detection of HPV DNA with amplification still remains the favorite tool for its sensitivity and specificity. However, due to the heterogeneous diversity of the HPV family, the use of type-specific primer PCR is impractical and uneconomical. In the past, HPV subtype identification has been performed by DNA type specific hybridization, ^[16] sequencing ^[17] and by restriction fragment length polymorphism. ^[18] The reliability of different methods is compromised, when comparing them in different laboratories. ^[19],[20] Recently, the DNA chip assay has been introduced in HPV detection using different DNA microarray platforms, ^{[21],[22],[23],[24],[25]} including few commercial assays e.g. PapilloCheck (Greiner Bio-One, Germany), PreTect HPV proofer (NorChip; Norway) ^[26] and HPVDNAChip (Biomedlab Co., South Korea). ^[27] To substantiate and further extend the investigation, we studied some of the technical and specificity and sensitivity issues of DNA oligo-array for genotyping. We made a comprehensive analysis of important issues and discussed their impact in the genotyping procedure.

The DNA chips have chemically active groups, such as aldehyde, poly L-lysine, epoxy, or silane deposited on a non porous glass surface on which DNA molecules are immobilized. Different substrate chemistries are in focus and methods are being developed for efficient binding of oligonucleotides to different kinds of substrate. ^[28],[29] In our hands, the aldehyde reactive surface showed maximum binding capacity for amino linked oligonucleotides. In addition, use of dT ₍₁₅₎ spacer overcomes the steric hindrance of the surface structure and solvent layer due to its charge, hydrophobicity, degree of solvation. ^[13] Other critical factors include composition of spotting solution, humidity and hybridization temperature play an essential and significant role in optimal performance of a microarray assay.

The specificity of single mismatch discrimination in microarray assay provides significant evidence for using this procedure for HPV genotyping. In our mini array, eight HPV genotypes were selected They represent major HPV subtypes prevalent globally, including low-risk types HPV (6 and 11); intermediate risk (33, 35 and 39) and high-risk types (16,18 and 31). ^[3],[4] The broad spectrum of HPV genotypes by suing GP5⁺/GP6⁺ primer set for amplification, combined sequence specificity of microarray assay makes this attractive tool for diagnostic procedures. However, the origin of the labeled probe remains a critical criterion for specificity. The probes generated from cloned genes or cell lines provide specific signal to noise ratio, whereas the probes from the pathological samples may results in cross hybridization. However, oligonucleotides with high diversity in their sequences can essentially overcome cross hybridization in genotyping assays as shown in HPV microarray.

In our HPV mini array we were able to detect specific genotypes with high accuracy, by both a 15-mer and a 30-mer target oligonucleotide, therefore demonstrating the potential of this technology for broad diagnostic procedures.

In conclusion, our study addressed fabrication and sensitivity issues of targeted DNA array assay and further improvements need to be implemented for the detection of other pathogens and make the use of microarray screening programs more feasible in diagnostic laboratories.

Acknowledgement

References

Copyright 2008 - Indian Journal of Medical Microbiology

The following images related to this document are available:

Photo images