flowers were tagged on the day of anthesis (0 days post anthesis (dpa)). Developing bolls from the first fruiting positions closest to the stem were collected at 8 dpa between 8 a.m. and 10 a.m. to eliminate diurnal effects. Bolls were collected from different plants, and immediately mingled, frozen in liquid Nitrogen, and then stored in 50-ml polypropylene tubes at -80°C
Growth protocol
G. hirsutum L. cv. TM1, also known as upland, was grown in the green house under a 30/21°C day/night temperature regime in a randomized block design
Extracted molecule
total RNA
Extraction protocol
Total RNA was extracted using the modified hot borate method (Wan and Wilkins, 1994; Wilkins and Smart, 1996; Arpat et al., 2004)
Label
Alexa Fluor® 555 and Alexa Fluor® 647 (Dye Swap included)
Label protocol
Fluorescent-tagged amino-allyl labeled cDNA probes were generated by reverse transcription of cotton fiber total RNA (30 μg) spiked with reference mRNA spike mix (10, 50, 200, 400, 1000, and 2000 pg Alien Spike mRNAs 1 through 6, respectively, Stratagene), 5g poly-T21VQ primers (Operon), 0.2 mM each of dATP, dCTP, and dGTP (Sigma), and a 2:3 ratio of 0.2 mM aminoallyl-dUTP/dUTP mix (Sigma) in a total volume of 40 μl incubated at 42°C for 4 hours using SuperScript II (Invitrogen). Aminoallyl-labeled cDNA was purified using the QiaQuick PCR purification kit (Qiagen), and labeled with Cy3 or Cy5 monofunctional dyes (Amersham) according to manufacturer’s instructions. Labeled cDNA were quality controlled by measuring absorbance at 260, 280, 555, and 650. For each probe, the frequency of incorporation (FOI) of dye molecules expressed as the number of dye molecules per 1000 bps and concentration of cDNA were calculated and probes with an FOI < 20 were discarded.
Channel 2
Source name
Developing Cotton fiber at 5 days post anthesis (dpa)
flowers were tagged on the day of anthesis (0 days post anthesis (dpa)). Developing bolls from the first fruiting positions closest to the stem were collected at 5 dpa between 8 a.m. and 10 a.m. to eliminate diurnal effects. Bolls were collected from different plants, and immediately mingled, frozen in liquid Nitrogen, and then stored in 50-ml polypropylene tubes at -80°C
Growth protocol
G. hirsutum L. cv. TM1, also known as upland, was grown in the green house under a 30/21°C day/night temperature regime in a randomized block design
Extracted molecule
total RNA
Extraction protocol
Total RNA was extracted using the modified hot borate method (Wan and Wilkins, 1994; Wilkins and Smart, 1996; Arpat et al., 2004)
Label
Alexa Fluor® 555 and Alexa Fluor® 647 (Dye Swap included)
Label protocol
Fluorescent-tagged amino-allyl labeled cDNA probes were generated by reverse transcription of cotton fiber total RNA (30 μg) spiked with reference mRNA spike mix (10, 50, 200, 400, 1000, and 2000 pg Alien Spike mRNAs 1 through 6, respectively, Stratagene), 5g poly-T21VQ primers (Operon), 0.2 mM each of dATP, dCTP, and dGTP (Sigma), and a 2:3 ratio of 0.2 mM aminoallyl-dUTP/dUTP mix (Sigma) in a total volume of 40 μl incubated at 42°C for 4 hours using SuperScript II (Invitrogen). Aminoallyl-labeled cDNA was purified using the QiaQuick PCR purification kit (Qiagen), and labeled with Cy3 or Cy5 monofunctional dyes (Amersham) according to manufacturer’s instructions. Labeled cDNA were quality controlled by measuring absorbance at 260, 280, 555, and 650. For each probe, the frequency of incorporation (FOI) of dye molecules expressed as the number of dye molecules per 1000 bps and concentration of cDNA were calculated and probes with an FOI < 20 were discarded.
Hybridization protocol
Printed slides were pre-hybridized in a 50-ml solution of 5X SSC, 0.1% SDS, and 0.1 µg/µl BSA in a sealed coplin jar for 45 min @ 42ºC, followed by two washes of 10 sec each in filtered ddH2O, and dried with filtered compressed air. The final hybridization solution (50 µl) contained amino-labeled cDNA in 30% formamide, 5X SSC, 0.1% SDS, and 0.1 µg/µl sheared salmon sperm DNA and was applied to the slide surface using mSeries LifterSlips (Erie Scientific). After incubating at 42ºC for 16-20 hours in a humidified hybridization chamber (Genetix), the LifterSlips were removed by dipping slides in 2X SSC @ 42 °C. Slides were washed sequentially at 42ºC in 2X SSC, 0.2% SDS for 5 min, 0.5X SSC for 3 minutes, 4 times in 0.1X SSC for 1 minute per wash, and 0.05X SSC for 10 seconds, followed by a dip in ddH2O and drying with filtered compressed air.
Scan protocol
Hybridized slides were scanned using the 428 Array Scanner (Affymetrix) followed by feature extraction to quantify signal intensities using ImaGene 4.2 software (BioDiscovery).
Description
PDF files of detailed experimental protocols for RNA isolation (SOP MA002 v1.1), amino-allyl labeling (SOP MA003 v2.1) and microarray hybridization (SOP MA004 v2.1) can be accessed at Cotton Functional Genomics Center website.
Data processing
Microarray images were obtained using the 428 Array Scanner (Affymetrix) and signal intensities were quantified with ImaGene 4.2 (BioDiscovery). Low quality spots (dust particles, scratches, high background, or irregularly shaped spots) were manually flagged following visual inspection. Signal intensity was corrected based on blanks (N=274), and flagged spots with a background corrected intensity smaller than the average plus three standard deviations (>99.99% confidence interval) using a custom PERL script were excluded from further analysis. On average, 91.3% of all genes were present for analysis. Using the GNU statistical computing language R and LIMMA, linear models were applied to the data (Smyth, 2005). To remove any systematic intensity or space-dependent differences between Cy3 and Cy5 channels, data were normalized using a within-slide space-dependent global intensity (print-tip-lowess) normalization function provided in the R software package. The distribution of up and down regulated genes in data sets from hybridizations between distant developmental stages was not symmetrical as assumed for global normalization. To therefore analyze asymmetric results, reference calibration spots, which were spiked-in at a ratio of 1:1 in both channels, were weighted to 10 before initiating normalization procedures. Normalized data was fit to a linear model using the least squares method with an inter-duplicate correlation of 0.75 for duplicated spots. Results were interpreted using the 5 dpa time point as the common reference. The significance of differential expression between developmental stages was determined using the empirical Bayes method to shrink the gene-wise sample variances towards a common value. Resulting statistics were corrected for multiple-testing using Benjamini and Hochberg false discovery rates (Benjamini and Hochberg, 1995) and significant changes in gene expression were identified at p <0.05. Significance test was performed for each time point after 5 dpa relative to all preceding time points.
