Nucleic acid isolation. Total nucleic acid was purified using the automated NucliSens extraction system (BioMerieux, Durham, North Carolina). Following the manufacturer's instructions, 100 μl of each specimen was added to tubes containing 900 μl of prewarmed NucliSens lysis buffer and incubated at 37°C for 30 min with intermittent mixing. Fifty microliters of silica suspension provided in the extraction kit was added to each tube and mixed. The mixtures were then transferred to a nucleic acid extraction cartridge and loaded onto the extractor workstation for processing. Approximately 50 μl of total nucleic acid eluate was recovered.
Amplification. For the culture supernatants, 450 ng of nucleic acid was used as input for the amplification protocol. In parallel, 50 ng of HeLa cell RNA was used as a positive amplification control and water was used for a negative control. Samples were amplified using a random-primer protocol as described by Wang et al. (2002), with the following modifications: first- and second-strand synthesis were primed using primer-A (5′-GTTTCCCAGTCACGATCNNNNNNNNN) followed by PCR amplification using primer-B (5′-GTTTCCCAGTCACGATC) for 40 cycles. Aminoallyl-dUTP was incorporated into the PCR product using an additional 20 cycles of thermocycling. A detailed protocol is available as Protocol S1.
Microarray hybridization and analysis. DNA microarrays were printed and hybridized essentially as described by Wang et al. (2002), with the following modifications: for array printing, a single-defined 70mer (spike-70) was mixed with each viral oligonucleotide in a 1:50 ratio. Array hybridizations used Cy5-labeled amplified probe from either virally infected cultures or controls (mock-infected culture, HeLa RNA, or water); a reference signal for every spot on each array was generated by using a Cy3-labeled version of the reverse complement of spike-70. Oligonucleotides were assessed by Cy5 intensity. Oligonucleotides from the astrovirus and coronavirus families that passed a conservative, arbitrarily set cutoff of (Cy5infection-Cy5mock) > 1500 intensity units are listed in Table 1. Additional oligonucleotides from these families and their homology to the SARS coronavirus are listed in Table S1. Array data has been deposited in the Gene Expression Omnibus (GEO) database (accession number GSE546). A complete list of the viral oligonucleotide sequences on the microarray is also available as Table S2.
Conventional PCR using array element sequences. PCR primers were designed by aligning the hybridizing oligonucleotides (Oligo IDs 15081544_766 and 12175745_728) to the IBV genome (Fwd: 5′-TGTTTTGGAATTGTAATGTGGAT; Rev: 5′-TACAAACTACCTCCATTACAGCC) and selecting stretches of near-identity. Primer-B-amplified material was used as the template for 35 cycles of thermocycling using the following program: 94°C for 30 s, 56°C for 30 s, and 72°C for 60 s.
Direct sequence recovery from the microarray. Amplified viral sequences hybridized to individual microarray spots were recovered by scraping a 100 μm area of the microarray using a tungsten wire probe (Omega Engineering, Inc.) mounted on a micromanipulator while visualized by fluorescence microscopy (Nikon TE300). Recovered material was PCR amplified using primer-B, cloned into pCR2.1TOPO (Invitrogen), and sequenced. A detailed protocol is available as Protocol S2.
Shotgun sequencing. Primer-B-amplified nucleic acid (see above) was cloned in pCR2.1TOPO, plated on 2xYT/kan plates, and grown overnight at 37°C. White colonies were picked into 384-well plates containing 2xYT/kan plus 8% glycerol and incubated overnight at 37°C. DNA was purified by magnetic bead isolation. DNA sequencing involved adding 3 μl of water to each bead pellet, followed by 3 μl of Big Dye terminator (v3.1) sequencing cocktail, and incubation for 35 cycles of 95°C for 5 s, 50°C for 5 s, and 60°C for 2 min. Reaction products were ethanol precipitated, resuspended in 25 μl of water, and loaded onto the ABI 3730xl sequencer. The resulting sequence reads were trimmed to remove primer sequences from the RT-PCR step and then assembled by Phrap (P. Green, unpublished data). Resulting contigs were screened by blast to remove any contigs with high human or monkey sequence similarity. The remaining contigs were edited to high quality, making any obvious joins. (Sequences are available as Data S1.)