Virus samples were collected from the different geographical regions of Korea and obtained from the Korean Agricultural Culture Collection (KACC). All viral infections in this study were confirmed by RT-PCR. Samples of microarray analysis were isolated from various host plants; Kyuri green mottle mosaic virus (KGMMV) isolated from cucumber, Tobacco rattle virus (TRV) isolated from tobacco, Cowpea mild mottle virus (CPMMV) isolated from pea, Zucchini yellow mosaic virus (ZYMV) isolated from zucchini, Melon necrotic spot virus (MNSV) isolated from melon, Tomato bushy stunt virus (TBSV) isolated from tomato, Potato leafroll virus (PLRV) isolated from potato, Tobacco ringspot virus (TRSV) isolated from soybean, Alfalfa mosaic virus (AMV) isolated from goosefoot, Cucumber mosaic virus (CMV) isolated from pepper and Tomato spotted wilt virus (TSWV) isolated from pepper.

All of the probes were designed on the basis of several public databases such as NCBI and DPVWeb by using the CLC Main Workbench 6 (CLC bio, USA) according to the manufacturer’s instructions. Each probe was forty-five to fifty nucleotides in length. The orientation of the probe sequences is sense. Thus, cDNA, which was synthesized from total RNA of a plant infected with a plant virus, can be detected by the LSON chip containing the probes. GC content of the probes was ranged from forty to fifty, and Tm (melting temperature) value was ranged from forty-five to eighty. In addition, in order to avoid secondary structures of the probes being printed on the chip, the values of the self-annealing and the secondary structure were ranged no more than forty-five, and the value of the self-end annealing was ranged no more than fifteen.

Microarray slides were printed by the GenoCheck Co. (Ansan, Korea) with a Pix5500 Arrayer (Cartesian technologies co.). Corning GAPS II slides are 25 × 75 mm glass slides coated with an aminosilane surface chemistry that enables the even immobilization of DNA. Manufactured through a proprietary process, GAPS II coated slides offer unparalleled performance in terms of uniform spot morphology, high DNA retention for maximum signal strength, and improved sensitivity through ultra-low fluorescent background. Each slide contained a total of 3,978 features spotted in duplicate to reduce positional bias of the fluorescence readout, and divided up into 16 blocks (4 × 4 blocks) of 17 rows and 15 columns each. Printed slides were exposed to UV light for cross-linking and blocked in a succinic anhydride/sodium borate solution. Excess DNA probe material was then washed off in a water bath at 95°C. Blocked slides were dehydrated in a bath of 100% ethanol, centrifuged, and stored in a dark cool dry location before use.

Total RNA from infected and healthy plants, used in each hybridization event, was extracted using the Tri-reagent (MRC Reagents, Inc., OH, USA) as described by the manufacturer. RNA concentration and quality was determined spectrophotometrically by NanoDrop (NanoDrop Technologies, Wilmington, DE, USA), The A260/280 ratios of all RNAs were above 1.9. The reverse transcription (RT) reactions were performed in a 30 μl volume containing 10 μg of total RNA with 2uM of N25 random primers, 500 μM dATP, 500 μM dCTP, 500 μM dGTP, 200 μM dTTP, 100 μM cy3-dUTP (targeted for virus-infected sample) or 100 μM cy5-dUTP (targeted for healthy control sample) and 300 U of RevertAid H Minus Reverse Transcriptase (Invitrogen, CA) in 1X first-strand buffer. After incubation at 42°C for 2 hours, the samples were incubated at 70°C for 15 minutes and cDNA was purified using the PCR Purification Kit (Bioneer, Deajeon, Korea) according to the manufacturer’s protocol, quantified by NanoDrop (NanoDrop Technologies, Wilmington, DE, USA). Oligonucleotide-printed slides were outfitted with a MAUI Mixer DC hybridization lid (BioMicro, Salt Lake City, Utah) prior to hybridization. Each Cy3-labelled cDNA (target virus) probe was paired with Cy5-labelled cDNA probe (healthy plant) and hybridized to LSON chip using a Maui Hybridization System (BioMicro Systems Inc). The fluorescently labeled cDNA probes were resuspended in 2X hybridization buffer (25% formamide, 5X SSC, 0.1% SDS and Bromophenol blue), denatured at 95°C for 5 minutes, chilled on ice for 2 minutes and applied to the microarray. Hybridization was performed on MAUI Hybridization System (BioMicro Systems Inc.) for 16 hours at 42°C. After hybridization, the slides were washed twice with 2X SSC plus 0.1% sodium dodecyl sulfate for 5 minutes at 42°C, twice with 0.1X SSC for 1 minute at room temperature and rinsed once with distilled water. After being dried by centrifugation, the microarrays were scanned.

The hybridized slides were scanned with a GenePix 4200A scanner (Axon Instrument, Union City, CA) containing a 532 nm laser for Cy3 and a 635 nm laser for Cy5 fluorescence measurement, and for image analysis GenePixpro 6.0 software was used. The level of PMT power was selected to balance the background with the level of spots detection on a slide. The hybrid spots were then identified on the array and located using a Gal file as feature indicator, so that the measurements of fluorescence and local background fluorescence for each spot could be recorded.

As a step towards developing an efficient plant virus detection chip, a small-scaled oligonucleotide microarray was constructed as a prototype for high-throughput detection (Supplementary Fig. 1). This microarray chip was designed with probes specific for the 30 plant virus species, most of which are economically important viruses in Korea. The validation of the prototype oligonucleotide microarray was performed by hybridizing labeled cDNA from plants infected with 30 viruses from different genera. Most of the virus probes were highly specific and were able to identify virus samples to be tested (data not shown). Based on these results, a LSON chip was developed for the detection of known viruses and for the monitoring of unknown viruses.

The LSON chip which is capable of detecting most of plant viruses with known sequence information has been designed. This oligonucleotide chip contained 3,978 probes in duplicate potentially capable of detecting more than 500 plant-infecting viruses in a single experiment which includes satellite RNAs, and viroids (Table 1). Probes were spotted in 16 blocks (4 × 4 blocks) in duplicate on each slide. Individual probes are arranged in 15 × 17 grid in one block for a total of 249 or 248 spots. Each block includes three positive and negative control (buffer only) probes with virus-specific probes (Fig. 1L). Positive control probes were prepared by mixing all of the probe.

Total RNA was extracted from virus-infected plants and healthy plants, labeled with Cy3 and Cy5 and used for microarray analysis. The green spots labeled with Cy3 only indicated the detection of virus in the test plant. Probes homologous to plant genes hybridize to cDNA from both the virus-infected plants and healthy plants, thus these spots appear yellow.

The validation of the LSON chip was accomplished by hybridization of labeled cDNA from plants infected with 11 viruses from different genera. An example of 11 hybridizations on the LSON chip with 3,978 probes included is shown in Fig. 1. This LSON chip was hybridized with labeled cDNA from KGMMV in the genus Tobamovirus (A), TRV in the genus Tobravirus (B), CPMMV in the genus Carlavirus (C), ZYMV in the genus Potyvirus (D), MNSV in the genus Carmovirus (E), TBSV in the genus Tombusvirus (F), PLRV in the genus Luteovirus (G), TRSV in the genus Nepovirus (H), AMV in the genus Alfamovirus (I), CMV in the genus Cucumovirus (J) and TSWV in the genus Tospovirus (K) infected plants. Each virus species provided a powerful specific hybridization. Cucumber sample infected from KGMMV showed a hybridization pattern to the ORSV spots belonging to the same genus with KGMMV. Through the hybridization pattern to the virus belonging to the same genus, unexpected viruses can be predict the virus genus. Nine out of sixteen oligo- nucleotide probes designed for TRV did not produce a reaction signal, while seven probes exhibit a high signal strength. The probes that showed negative signal in TRV specific probes are response of different strains of TRV. This observation suggests that genotyping of several strains of TRV may be possible. In all hybridization experiments, the background signal was low which indicated a high quality of hybridization. Good quality hybridization signals show that the LSON chip has a great potential for accurate detection of various viruses.

In nationwide surveys of viral diseases of soybean (Glycine max) in Korea, various soybean samples were collected and analyzed using electron microscopy and RT-PCR with primers specific to several legume-infecting viruses. A virus with isometric particles was detected from a soybean sample showing yellow mosaic on the leaves in Sangju County of South Korea in 2005 (Nam et al., 2011a). No virus infection was detected in this samples using RT-PCR with soybean virus specific primers which was previously reported in Korea. The LSON chip developed in this study was used to identify this virus sample. Specific probes of Peanut stunt virus (PSV) are indicated with the red rectangles in block 5 (Fig. 2). The green spots in red rectangles indicate the detection of PSV in the soybean. PSV in the genus Cucumovirus of the family Bromoviridae is an economically important pathogen and occurs worldwide in legume plants. Using a LSON chip this way would enable the detection of unexpected interactions such as infection of a new host.

Sensitivity of a microarray is generally dependent on virus concentration in host cells and hybridization efficiency. So, it is hard to determine the exact sensitivity of a microarray. Pepper mild mottle virus (PMMoV) belongs to the plant virus genus, Tobamovirus which accumulated to high titers in plants. To test the sensitivity of the LSON chip with PMMoV, plant sap was extracted from both PMMoV-infected and healthy plants in RNA extraction buffer. Plant sap of PMMoV-infected plants was diluted with sap of healthy plants 1:10, 1:100, 1:1,000, 1:5,000 and 1:10,000 and hybridized to the microarray together with a healthy control (Fig. 3A-D). For a comparison with RT-PCR, plant material from the PMMoV infected plants was diluted 1:10, 1:100, 1:1,000, 1:5,000 and 1:10,000 and tested by RT-PCR (Fig. 3E). Comparison of the array detection of PMMoV with RTPCR showed that RT-PCR was more sensitive method for detecting virus up to the 1:10,000 dilutions, the microarray experiments revealed the presence of PMMoV in dilutions up to 1:1,000 (Fig. 3).