A two-genome microarray for the rice pathogens Xanthomonas oryzae pv. oryzae and X. oryzae pv. oryzicola and its use in the discovery of a difference in their regulation of hrp genes

Quality of the Xo array

We used a diagnostic method developed by Rocke [29] to examine the contribution of different factors to measured differences in signal intensity in 6 hybridizations (3 replicates with dye-swaps) comparing Xoo gene expression in PSB vs. XOM2 using the Xo array. This method employs an alysis o f va riance (ANOVA). ANOVA can be used to calculate the gene expression changes in replicated array experiments and to correct systematic errors [30]. Factors examined were treatment, dye, sample, and error (Fig. 1). The treatment factor (red line) had the largest effect whereas the dye, sample, and error effects were much smaller. This analysis indicates that the measured significant changes in gene expression are due to the treatment and not to variability of other parameters.

The non-specific background error rate was assessed using 100 replicate spots of an oligonucleotide probe corresponding to the non-Xo gene encoding hygromycin phosphotransferase (hph). Across 6 hybridizations each using cDNA prepared from Xoo or Xoc RNA (3 replicates for each pathovar, each with a dye swap), error rates for positive artifacts were 0.00–1.00% and 1.00–2.00% respectively. That is, less than 2% of the hph gene oligonucleotides showed over two-fold differences in signal intensity. Error rates associated with non-spotted controls (632 empty spots) were similarly low, 0.00–0.94% and 0.00–1.10% following hybridization with Xoo- and Xoc-derived cDNA, respectively (Additional File 1).

Optimization of hybridization temperature and sample amounts

Application and validation of the array to identify Xoo and Xoc genes differentially expressed in a rich vs. a minimal medium

To identify differentially expressed genes, the LMGene Package [29] was used. The resulting list of genes with significantly different expression between the two growth conditions was then refined using a false discovery rate (FDR) of 5% and a fold-change minimum of 1.75 (log2ratio value > 0.8), resulting in 247 genes for Xoo and 39 genes for Xoc. Additional File 2 provides a complete list of the differentially expressed Xoo and Xoc genes, sorted according to functional category and fold-change in expression (log2ratio).

To validate these results, semi-quantitative RT-PCR was used to independently assess expression levels for 16 Xoo and 16 Xoc genes selected arbitrarily from the list (genes and primer sets used are given in Table 3, and semi-quantitative RT-PCR results are shown in Additional file 3). RNA samples that were used in the microarray experiment as well as RNA samples extracted from three additional replicate sets of cultures were used as templates. There was good correlation between the semi-quantitative RT-PCR and the microarray results (correlation coefficients were 0.8225 and 0.7791 for Xoo and Xoc genes, respectively, Fig. 2). Although the amplitude of gene expression fold change between the two techniques is different, as might be expected since semi-quantitative RT-PCR is not a reliable measure of quantitative differences, the general trend of gene expression is consistent. For additional verification, we performed quantitative RT-PCR on 5 genes from Xoo and 2 from Xoc. In each case the results verified the expression patterns observed using semi-quantitative RT-PCR (Additional file 4).

GeneIDa	Primer Sequence
XOO4289	F 5' ACA TCG CCG ATA ATT TCC AG 3' R 5' CGC AAC ACC TTG TAC TCG AC 3'
XOO4035	F 5' GGT CTT CGG ATC GTC AAC AT 3' R 5' GAT CAG AAA GCC GAT CTT GC 3'
XOO1994	F 5' GTT GGA GCA CAC CAT GAA AG 3' R 5' GGT ACA GCT CCA GAC CGA TG 3'
XOO2803	F 5' CTG TTC CAA GCA GAC CCT GT 3' R 5' CAC GAT GGG AAA CCT GAA AC 3'
XOO0424	F 5' CGG CTG AAG AAC TAC GCT TC 3' R 5' CTT GGT CAG CTC GTT GAT GA 3'
XOO0423	F 5' CGA AGA AGG CCT CTA CAT GG 3' R 5' CGA AGA AGG CCT CTA CAT GG 3'
XOO0076	F 5' GTG CCA CGT TGA AGT CAA GA 3' R 5' CTC ACT TAA TTC GCG CTT CC 3'
XOO1379	F 5' GCG ATA CCA GTC CAG GAT GT 3' R 5' CTT TTC CTC GTT GCA CTG GT 3'
XOO0094	F 5' CAC CTA CGG CTT TGT CTG GT 3' R 5' CAT TGC CAA ATG TGT TGG AG 3'
XOO0770	F 5' ATC GGC AGG TCG TAC TTG AT 3' R 5' GTC AGA CCC TGC TGT TCT CC 3'
XOO0282	F 5' CTG ATG AAT GAG CCT CAC GA 3' R 5' GAT TCC ATG TAG CCC AGC AT 3'
XOO2163	F 5' AAC GGT AGA ACT TGC CAT CG 3' R 5' AAC CTG GAC ATC CTG GAC AT 3'
XOO2757	F 5' AGC GCA GTC GCT TAC CTT C 3' R 5' GCA TAC GAC GAC GAC TAC GA 3'
XOO1664	F 5' CAC GCG TCT ACT GGG AAG AT 3' R 5' AAC ACG TCA TAC AGC GCA TC 3'
XOO4468	F 5' ACG ATT TCG ACC TGG ACC AC 3' R 5' ACA AGG ACG CCG AAA AGA T 3'
XOO1458	F 5' CCA GCG TTC CAT CAC TAC G 3' R 5' AGG GTA ATT AAC CGG CTT CG 3'
XOCORF1456	F 5' AAT GAC AAT GAG GGC ATC AA 3' R 5' ACT GAT TTG CGT TGT CGT TG 3'
XOCORF3144	F 5' GCA GAC GTT CGA CAC TTT CA 3' R 5' GCC TGT GTC TGC GAC TTG TA 3'
XOCORF3137	F 5' ACG ACC GTA TCC AAC CAG AC 3' R 5' AAC ATG CTG CGG ATT TCT TC 3'
XOCORF2869	F 5' AGT CGT TCG TAC CAG CCA TC 3' R 5' GCT CAC CTC CTG CTT GTA GC 3'
XOCORF0857	F 5' GCC AGC TTG AAA GTC AGC TC 3' R 5' CAT TTG CAG CAT TGG TGA AG 3'
XOCORF0690	F 5' TTC CTT TTC GCC TGG AGT T 3' R 5' TTC ATC GAC ACC GTC ATT G 3'
XOCORF4060	F 5' AAG TCA GTC CCG GTC AAG GT 3' R 5' ATT CCT CCA CCA TCT CGT TG 3'
XOCORF0488	F 5' GAC GTT CCG ACC AAT CTG TT 3' R 5' CTG CCC GAT CTT GAT CAT CT 3'
XOCORF2820	F 5' AAG ACC ACC CAC AAG CTG TT 3' R 5' GAC CCT TGA TGG TCA GAT CG 3'
XOCORF4434	F 5' AAT CTG GCC AAC GTC GAT AC 3' R 5' AGC TGG ATC ATT TTC CAC CA 3'
XOCORF0026	F 5' ATG GTG GAA AGC CTC AAC AC 3' R 5' GCC AGG ATA TTG GTC TGG AA 3'
XOCORF0775	F 5' AAA CTC TCG TGC TTG GTG CT 3' R 5' CAG CGT ATT CGT AGG TGA CG 3'
XOCORF4022	F 5' CAG CAT TCG CTG AAG GAA CT 3' R 5' AAA TAC GGC ACC TTG TGC TC 3'
XOCORF3816	F 5' TAT ACT GGT CGC TGC TGG TG 3' R 5' CGG TAA GTC ACC TCG TAG CC 3'
XOCORF1384	F 5' CCA AGA TCC GCA AGA AGA AG 3' R 5' GGA TCA GCT TTT CGA TCT GC 3'
XOCORF2448	F 5' GCT CAC TTA ATT CGC GCT TC 3' R 5' AAC GAG CTG CTT AGC GTT GT 3'

^aThe gene ID is according to the primary annotation obtained from the Comprehensive Microbial Resource. XOO designations represent IDs of genes from Xanthomonas oryzae pv. oryzae KACC 10331 and XOCORF designations represent IDs of genes from Xanthomonas oryzae pv. oryzicola BLS256.

Xoo and Xoc genes differentially expressed in XOM2 relative to PSB

Of the differentially expressed genes, 106 Xoo genes were up-regulated and 141 were down-regulated in XOM2 as opposed to PSB. For Xoc, only 28 and 11 genes were up-regulated and down-regulated, respectively, in XOM2 (Additional File 2). These genes represent 17 functional categories, based on the TIGR annotation for the genomes available through the Comprehensive Microbial Resource [5] (Fig. 3). The Xoo genes up-regulated in XOM2 encode primarily hypothetical proteins (29.2%) and proteins involved in cellular processes (22.6%); most of the down-regulated Xoo genes encode hypothetical proteins (34.8%) or proteins involved in signal transduction (5.0%), DNA metabolism (5.0%), mobile and extra-chromosomal element functions (4.3%), or transport and binding (12.1%). Xoc genes expressed at a higher level in XOM2 relative to PSB, as in Xoo, encode hypothetical proteins (28.6%) and proteins involved in cellular processes (21.4%). In contrast to Xoo however, Xoc genes down-regulated in XOM2 predominantly encode proteins involved in protein synthesis (45.5%).

Bacterial strains, growth conditions, and media

Xanthomonas oryzae pv. oryzae strain PXO99^A (Philippine race 6 provided by Jan Leach) and X. oryzae pv. oryzicola strain BLS256 were used for these experiments. Cells were grown at 28°C with shaking at 200 rpm, in nutrient-rich PSB (10 g/liter of peptone, 10 g/liter of sucrose, 1 g/liter of L-glutamic acid, monosodium salt; [28]). For experiments testing the effects of the modified minimal medium, XOM2, bacterial cells were cultured in PSB until OD600 equaled 0.2, washed twice, and then immediately transferred into XOM2 for 16 hrs. XOM2 [18] consists of 0.18% xylose sugar, 670 μM D, L-methionine, 10 mM sodium L(+)-glutamate, 14.7 mM KH₂PO₄, 40 μM MnSO₄, 240 μM Fe(III). EDTA and 5 mM MgCl₂, pH 6.5. Cells were washed twice prior to being harvested.

RNA preparation

RNA was isolated using TRIzol^® reagent (Invitrogen, Carlsbad, CA, U.S.A.). The RNA samples were treated with 10 units of DNaseI (Invitrogen, Carlsbad, CA, U.S.A) for 30 min at room temperature, followed by column purification using the RNeasy midi kit (Qiagen, Germantown, MD, U.S.A.). The quality of RNA was determined by carrying out gel electrophoresis on a 1% agarose gel and was verified visually by using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, U.S.A.). The quantity of total RNA was determined by measuring the absorbance at 260 nm and 280 nm. In addition, the level of protein contamination in the RNA was measured by using A260/A280 ratio.

cDNA generation and labeling

cDNA was generated by using SuperScript™ III First-Strand kit and following the manufacturer's protocol (Invitrogen, Carlsbad, CA, U.S.A.). Twenty micrograms of high quality RNA was used mixed with random hexamers that were used as primers for cDNA generation and the mixture then preheated at 70°C for 15 mins. Primers were annealed to total RNA and extended with a labeling mixture consisting of 6.0 μl of 5× buffer, 2.0 μl of 0.1 M DTT, 1.0 μl of RNasin, 2.0 μl of SuperScript III reverse transcriptase, 2.0 μl of 25× dNTP-allyl-amino (aa) dUTP mixture (final concentration 0.5 mM each of dATP, cCTP, and dGTP, 0.35 mM aa-dUTP, 0.15 mM dTTP) at 25°C for 10 min followed by 2 h at 42°C. The RNA template was hydrolyzed using 3 μl of 2.5 N NaOH (37°C, 15 min) followed by neutralization with 15 μl of 2 M HEPES. Unincorporated primers and nucleotides were removed using the Zymo research kit according to the manufacturer's protocol (Zymo research, Genetix, UK) and the purified amino allyl-modified cDNA was resuspended in 60 μl of 50 mM sodium bicarbonate (pH 9.0). The amino allyl-modified cDNA was used to resuspend lyophilized Cy3 or Cy5 and incubated for 1 hr at room temperature in the dark. The reaction was quenched by adding 15 μl of 4 M hydroxylamine (15 min, room temperature in the dark). The dye-coupled cDNA was then purified by using the Zymo research kit (Zymo research, Genetix, UK).

Oligo design

Our goal was to design a complete set of oligos that would uniformly detect gene-specific expression patterns for both Xoo and Xoc. To achieve the highest standard of uniformity, sensitivity and specificity for the Xo array it was necessary to utilize an optimized oligo design software that integrates the whole Xo gene set in its computation instead of considering each gene individually in a batch design mode. The Xo array carries 2 copies of the combined oligonucleotide set chosen by PICKY, the most efficient software developed to date, for this task [55].

The combined oligo set was designed based on the following steps, using PICKY for each oligo selection step: 1) The gene sets of Xoo and Xoc were combined with an additional hygromycin phosphotransferase gene and given to PICKY as a whole to design the shared oligos, i.e., oligos that can target at least one Xoo and one Xoc gene; 2) Xoo and Xoc genes targeted by the chosen shared oligos were removed from their individual gene sets and served as the second round design nontargets, i.e., genes that should be avoided by any PICKY designed oligo; 3) PICKY was then used to design oligos that can identify the remaining Xoo genes, using the earlier removed Xoo genes and the hygromycin phosphotransferase gene as nontargets; 4) Similarly, PICKY was used again to design oligos that can identify the remaining Xoc genes, using the removed Xoc genes and the hygromycin phosphotransferase gene as nontargets; 5) Finally, all oligos designed in steps 1, 3 and 4 were merged together to form the combined 4,675 oligo set.

Genome sequences and primary annotation of Xoo and Xoc were retrieved from the Comprehensive Microbial Resource [5] version 2.3 on December 22, 2005. The annotation is provided as Additional files 5 and 6. For Xoo KACC10331, these data are also available from the NCBI GenBank, under accession NC_006834. For Xoc BLS256, the finished genome sequence is also available from the NCBI Genbank, under accession AAQN01000001, but at the time of writing, the annotation has not yet been accessioned in that database.

Oligo synthesis

Oligos used for spotting were synthesized by Integrated DNA Technology [56]. As a control, an oligo was designed by PICKY to detect hph but not Xo genes. This control was included in 2 well positions on each 384-well oligo synthesis plate except the last plate, which had only one hph position. Because the positions of this control were randomized across plates, the oligo also served as a check for array printing when labelled and hybridized against the array.

Spotting

Microarrays were prepared at the ArrayCore Microarray Facility at the University of California, Davis [57]. Oligonucleotides were suspended in 1× Nexterion Spot solution at a final concentration of 20 μM and spotted onto aminosilane coated glass slides (Schott-Nexterion, USA). Oligonucleotides were spotted using a Lucidea Array Spotter (Amersham) in a humidity controlled spotting chamber (70%) at room temperature. Microarrays were deposited using 190 μm column and row pitches, and spot diameters averaged 80 μm under these conditions. After spotting, slides were allowed to sit at 70% humidity overnight at room temperature to maximize oligo binding. Microarrays were allowed to dry at ambient conditions and stored in the dark under argon at room temperature until use. Slides and spotting plates were tracked using the array spotter's built-in barcode reader and the information was used to generate the gene array layout file of the spotted 5 k Xo oligonucleotide microarray.

Amino-blocking pre-treatment

Prior to the hybridization process, oligo-spotted slides were pre-treated with a blocking step that removes unbound DNA-molecules and buffer substances from the slides by extensive washing in order to avoid any interference with subsequent hybridization experiments. Spotted slides were incubated in amino blocking solution (5 g succinic anhydride in 315 ml n-methylpyrrolidone, 35 ml 0.2 M sodium borate pH 8.0) at room temperature for 15 mins. Slides were placed in 0.1% SDS solution 20 sec and washed with nanopure water 20 sec two times and then were transferred into the sodium borohydride block solution to undergo a sodium borohydride pre-treatment.

Sodium borohydride pre-treatment

To minimize non-specific autofluorescence from the spotted material [58], slides were placed into a block solution containing 2× SSC, 0.05% SDS, 0.25% NaBH₄ (Biochemical Technologies, USA) and incubated at 42°C for 20 min. Slides were transferred to 1× SSC for 5 min at room temperature and then sequentially washed with vigorous stirring using fresh 1× SSC (3 × 5 min, room temperature), 0.2× SSC (4 × 2 min, room temperature), and Nanopure water (1 × 2 min, room temperature). Slides were spin-dried (1000 rpm, 10 min) and stored under argon until use.

Hybridization and scanning

Labeled probes were evaporated in a vacuum centrifuge on aqueous setting at 60°C to a volume of approximately 2–3 μl. Evaporated probes were then resuspended in 100 μl of a salt based hybridization solution (Ocimum Biosolutions) at room temperature. All hybridization and scanning steps were performed in a hepa and carbon filtered clean room at the ArrayCore Microarray Facility at University of California, Davis [57]. Hybridization occurred on a Tecan HS 4800 hybridization station. To block non-specific hybridization, a pre-hyridization buffer (5× SSPE, 6 M Urea, 0.5% Tween-20, 10× Denhardt's solution) was applied to the slides at 50°C and agitated for 15 minutes on the medium setting. Labeled probes were denatured by heating the mixture at 95°C for 3 mins and then snap-cooling on ice for 30 seconds. Probes were applied into the injector to hybridize with printed slides. Samples were hybridized for 16 hours at 42°C, Following hybridization, the slides were consecutively washed at 37°C with three salt based buffers of increasing stringency (2× SSC, 0.1% SDS, 1.0× SSC, and 0.5× SSC). Each buffer wash step was repeated twice, with a soak time of one minute followed by a one minute wash. A final wash step with water was performed. Following the final wash, slides were dried under a constant stream of N₂ at 30°C. Slides were kept under N₂ until scanning.

Capture of raw data

Hybridized microarray slides were imaged using a GenePix 4000B dual laser microarray scanner (Axon Instruments, USA) at 5 μm resolution. Slides were imaged using 100% laser power for both lasers (532 nm and 635 nm) and scanned twice using the high PMT and low PMT settings. All images were processed using GenePix software (Axon Instruments, USA) for element identification and quantification. The metadata associated with the hybridizations, along with the "raw" intensities obtained from the GenePix quantitation.

Validation of expression patterns of candidate genes using semi-quantitative RT-PCR and quantitative RT-PCR

For the first-strand cDNA synthesis, 100 ng of mRNA was reverse-transcribed in a total volume of 20 μl that contained 50 ng of random hexamer, 2.5 mM dNTP, 40 unit of RNaseOUT™, and 200 units of SuperScript™ III reverse transcriptase (the latter two components from Invitrogen, Carlsbad, CA) in reaction buffer supplied by the manufacturer. The reaction mixtures were incubated in 25°C for 10 min, then 4°C for 60 min. PCRs were performed in 50 μl reactions (containing 0.1 μl aliquots of the respective cDNA reaction mixture, 0.2 μM of gene-specific primers, 10 mM dNTPs, 1 unit of Taq DNA polymerase (Invitrogen), and 10× Taq buffer supplied by the manufacturer). Each reaction included an initial 5-min denaturation at 94°C, followed by 22 to 30 cycles of PCR (94°C, 45 sec; 60°C, 45 sec; 72°C, 45 sec), and a final 10 min at 72°C. Afterward, 20 μl of each reaction mixture was separated on a 1.0% agarose gel. (The primers used for semi-quantitative RT-PCR are described in Table 3). 16S ribosomal RNAs was used as controls for semi-quantitative RT-PCR. Primer sets used for semi-quantitative RT-PCR were designed using Primer3 [59]. Sequences of the primers used are shown in Table 3. Visualized band intensities of semi-quantitative RT-PCR products on the EtBr-stained agarose gels were transformed to digital values using Totallab TL100 software (Nonlinear Dynamics Ltd.). Log2 transformation was applied to digital band intensity values using the same mathematic transformation equations that had been applied to the microarray data. Fold changes from the microarray experiments were plotted against those from the semi-quantitative RT-PCR experiments.

For quantitative RT-PCR, cDNA generated as described in semi-quantitative RT-PCR was used as template for quantitative RT-PCR. Five Xoo and 2 Xoc genes were tested by using 50 μl reaction mixed with SYBR^® Green PCR Master Mix kit (Applied Biosystems, CA, USA.) and following the protocol as provided by manufacturer. Each reaction included an initial ramping 2 min 50°C, activation 10 mins 95°C, and then followed by 40 cycles of PCR (95°C, 15 sec; 60°C, 60 sec). The amount of Xoo and Xoc tested genes from PSB and XOM2 cultures were quantitated by calculating from each corresponding standard curves.

Accession numbers

All of the microarray data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) under accession number GSE 9658.