The goal of the experiment:	Quadruple 9-mer-Based Protein Binding Microarray with DsRed Fusion Protein
Experiment Description:	The protein binding microarray (PBM) was introduced to determine if a DNA motif interacts with proteins in a genome-wide manner. We facilitated the PBM technology using a DsRed fluorescent protein and a concatenated sequence of oligonucleotides. The PBM was designed in such a way that target probes were synthesized as quadruples of all possible 9-mer combinations, permitting unequivocal interpretation of the cis-acting elements. The complimentary DNA strands of the features were synthesized with a primer and DNA polymerase on microarray slides. Proteins were labeled via N-terminal fusion with DsRed fluorescent protein. The PBM presented herein confirmed the well-known DNA binding sequences of Cbf1 and CBF1/DREB1B, and it was also applied to elucidate the unidentified cis-acting element of the OsNAC6 rice transcription factor.
Keywords:	protein binding microarray (PBM), transcription factor (TF), DsRed, binding motif.
Experimental factors:	different transcription factors.
The origin of each biological sample:	Cbf1 (S. cerevisia; Genbank accession number NC_001142), DREB1B (A. thaliana, Genbank accession number NM_118681) and OsNac6  (O. sativa; Genbank accession number NM_001051551)
Person Last Name:	Kim
Person First Name:	Yeon-Ki
Person Email:	kim750a11@gmail.com
Person Phone:	82-31-330-6193
Person Fax:	82-31-321-6355
Person Address:	GreenGene Biotech Inc., Myongji University, Yongin 449-728, Korea
Person Affiliation:	Director, Genomics & Genetics Institute
Manipulation of biological samples and protocols used:	Three proteins were expressed in Escherichia coli strain BL21-CodonPlus (Stratagene, La Jolla, CA, USA), respectively. Overnight cultured cells were inoculated in fresh liquid LB medium, grown at 37 ‘C to an OD600 of 0.6 and induced with 1 mM isopropyl モ-D-1-thiogalactopyranoside (IPTG) at 25 ‘C for 5 h. Cell pellets were obtained by centrifugation at 4‘C for 5 min at 5,000 g, resuspended and washed with cold PBS buffer including a protease inhibitor cocktail (Roche, Basel, Switzerland). Cell pellets were collected by centrifugation, resuspended in 5 ml of cold PBS buffer containing a protease inhibitor cocktail and sonicated until lysis for 5 min at 45 sec intervals on ice. The supernatant soluble fractions were retained after centrifugation at 4 ‘C for 30 min at 9,000 g. Proteins were enriched using TALON resins (Clontech) adapted with immobilized metal affinity chromatography (IMAC) according to the manufacturer¨s protocols. The purified protein fractions were collected in a volume of 500 レl and the protein concentrations were then determined.
The complementary DNA strand was synthesized as in a previous report. Reaction solution containing 40 レM dNTP (Takara, Shiga, Japan), 1.6 レM Cy5-dUTP (GE Healthcare, Giles, UK), 1 レM 5¨-CTG CAC TAG GTG ACT CCG-3¨ primer (Bioneer, Deajon, Korea), 1X ThermoSequenase buffer and 40 U ThermoSequenase (USB, Cleveland, Ohio, USA) was prepared. A custom-designed protein binding microarray (Agilent) was combined with the reaction solution in a hybridization chamber (Agilent) according to the manufacturer¨s protocol. The assembled hybridization chamber was incubated at 85 ‘C for 10 min and then at 60 ‘C for 90 min. The microarray was washed in phosphate buffered saline (PBS)?0.01 % Triton X-100 at 37 ‘C for 1 min, PBS?0.01 % Triton X-100 at 37 ‘C for 10 min, PBS at room temperature for 3 min and dried by centrifugation at 500 g for 2 min. The double-stranded microarray was scanned to verify successful synthesis.
The double-stranded microarray was blocked with PBS-2 % BSA (Sigma, St. Louis, MO, USA) for 1 h and then washed with PBS?0.1 % Tween-20, PBS?0.01 % Triton X-100 and PBS for 1 min. A protein binding mixture containing 200 nM protein in PBS-2 % BSA, 50 ng/レl salmon-testes DNA (Sigma) and 50 レM zinc acetate was prepared. The prepared protein mixture was incubated at 25 ‘C for 1 h for stabilization and combined with the microarray at 25 ‘C for 1 h. The microarray was washed with PBS containing 50 レM zinc acetate and 0.5 % Tween-20 for 10 min, PBS-50 レM zinc acetate-0.01 % Triton X-100 for 2 min and PBS-50 レM zinc acetate for 2 min. OsNac6 binding experiments were done in triplicated, and experiments for Cbf1, DREB1B and DsRed only were performed once.
The raw data:	DsRed_010709, CBF1_101108, DREB1B_101108, OsNac6_121208, OsNac6_011609, OsNac6_011809 
The processed data:	DsRed_010709_processed, CBF1_101108_ processed, DREB1B_101108_ processed, OsNac6_121208_ processed, OsNac6_011609_ processed, OsNac6_011809_ processed
Data extraction and processing protocols:	Each microarray was scanned three to five times at full laser power intensity and pixel resolution 5. In order to minimize the number of saturated spots, different photomultiplier tube (PMT) gain settings were applied ranging from 550 to 780 for Cy3 and from 550 to 600 for Cy5. The fluorescence was quantified and bad spots were excluded automatically using GenePix Pro version 5.1 software (Molecular Devices). The background-subtracted median intensities were obtained and typically 0.01 - 0.05 %  (20 ? 100) of spots showed saturated Cy3 intensity. The microarray was rescanned whenever the number of saturated spots was not in this range. The rank-ordered signal distribution demonstrated a deep leftward slope followed by a heavy right tail. As the probes in the deep slope region differed by only one base, we assumed that the signal distribution was due to a specific interaction between the protein and features on the microarray. Two independent linear models, y=ax+b, were applied in the deep and the heavy right tail region using the R statistical language. In one of the examples with OsNAC6, the slope and y-axis intercept of the deep slope are -68.3 and 53074.8, respectively; those of the heavy tail are 0.0207 and 2,283, respectively. These values provide an extrapolated rank of 745 for OsNAC6. To extract preferred elements, it was performed in 2 steps. First, feature sequences from ranks 1 to 745 were grouped using Perl script language. Any sequence that possessed at least five bases was matched; among these, at least three bases were contiguously matched to the highest one. These groups were ranked according to the highest sequence. Second, the frequency of 6-10mer was counted for each group. Frequency differences were compared and ＾TTACGTAA￣ was enriched as the highest among them. Also, our results were equivalent to the sequence defined in previous reports (e.g., ｀CACGTG¨ for Cbf1). Additionally, the CBF1/DREB1B transcription factor bound to ｀CCGAC¨, a C-repeat/cold- and dehydration-responsive DNA regulatory element (CRT/DRE) in Arabidopsis.
Array Design:	The microarray was manufactured by Agilent technology (Santa Clara, CA, USA). The quadruple 9-mer protein binding microarray (Q9-PBM) consisted of 232,145 quadrupled probe features was designed which includes 131,072 features from all possible 9-mers and 101,073 replicated features out of them. Each 9-mer was concatenated four times, followed by a complementary sequence to a primer (5'-CGGAGTCACCTAGTGCAG-3') and a 5-nt thymidine linker to the slide. A microarray slide has totally 243,504 spot addresses formatted with 267 column and 912 rows. Beside 232,145 quadrupled probes, 1,474 random sequences from yeast genome, 8,081 blank features and 1,804 features from manufacturer¨s concern were included.