Bacterial strains, plasmids and growth conditions
Bacterial strains used in this study are shown in . Tables S2
list the plasmids and the oligonucleotides, respectively. A phoB
knock-out derivative from the BW25113 strain was requested from Keio collection 
. This phoB
disruption was then transferred into MG1655 strain by P1 transduction 
and named MG1655_PhoB_KO. The MG1655_PhoB_FLAG which carries a 3xFLAG tag at the 3′ end of phoB
gene was constructed from the BW25113 strain using an epitope tagging approach 
For PhoB ChIP-chip experiments, strains MG1655 and MG1655_PhoB_FLAG were grown in Morpholinepropanesulfonic acid (MOPS) minimal medium with 200 μM K2HPO4 and 0.4% glucose. shows the time point for cell harvesting and the cultivation of MG1655, MG1655_PhoB_FLAG, and MG1655_PhoB_KO under phosphate-limiting and phosphate-sufficient conditions. The time point at OD600nm of 1.0 was selected since phosphate was used up and PhoB had a higher activity for ChIP-chip assay. For the gene expression microarray experiments, MG1655 and MG1655_PhoB_KO were compared under the same conditions as the ChIP-chip assay. To compare the promoter activity of the upstream regions, promoter::luciferase gene fusion plasmids were constructed, and luminescence was measured for both MG1655 and MG1655_PhoB_KO strains at the same time point as the two above experiments (). The ChIP-chip experiments and the reporter gene assays were carried out in at least biological triplicates, while the gene expression microarray experiments were performed in two biological replicates.
Determination of phosphate concentration
To determine the concentration of orthophosphate, an ascorbic acid method described previously was applied in biological triplicates with slight modifications 
. After overnight culturing of the MG1655, MG1655_PhoB_KO, and MG1655_PhoB_FLAG strains in MOPS minimal medium containing 1000 μM K2
and 0.4% glucose, cultures were diluted in 1
100 ratio in MOPS minimal medium containing 200/1000 μM K2
and 0.4% glucose and grew at 37°C. At each time point, cultures were collected, centrifuged at 12,000 g for 5 min, and then 1 ml supernatants were added to 160 μl reaction solution (1 N sulphuric acid, 0.1 mM potassium antimonyl tartrate, 4.8 mM ammonium molybdate and 30 mM ascorbic acid (added lastly)). After 10 min incubation of supernatants with the reaction solution, the light absorbance at 880 nm was measured. By interpolation of the standard curve, the phosphate concentration was determined.
Chromatin immunoprecipitation (ChIP) experiment
To identify the genome-wide DNA-binding profile of PhoB, ChIP assays were performed on MG1655_PhoB_FLAG and MG1655. The ChIP assay protocol was modified from Byung-Kwan Cho et
. The MG1655_PhoB_FLAG strain expresses the PhoB-FLAG fusion protein where the FLAG tag can be recognized by anti-FLAG antibody and used for ChIP assaying 
. The MG1655 strain, which expresses no FLAG tag, was used as a control group. Cultures were grown to an OD600
value of 1.0 and treated with 1% formaldehyde for 10 min. To quench the reaction, glycine was added at the final concentration of 0.125 M for 5 min. Cells were centrifuged at 12,000 g at 4°C for 20 min and washed two times with the washing buffer (10 mM Tris-HCl (pH 7.4), 0.1 M NaCl, 1 mM EDTA and 0.5% Tween-20). The washed cells were then lysed with the lysis buffer (10 mM Tris-HCl (pH 7.4), 0.1 M NaCl, 1 mM EDTA and 0.5% Tween-20, 8 kU/ml lysozyme, 1 mM PMSF, and protease inhibitor cocktail (Sigma)) for 30 min at 4°C. The lysates were sonicated (Bioruptor) to result in DNA fragments ranging from 100 bps to 1000 bps with the average size of 500 bps. After sonication, the lysates were centrifuged at 12,000 g for 20 min at 4°C and the resulting supernatants were used for immunoprecipitation.
To eliminate the non-specific bindings between the magnetic beads coated with Protein G (Invitrogen) and the anti-FLAG antibody, the magnetic beads were pre-incubated with 0.05 mg/ml anti-FLAG antibody (Sigma). Similarly, for the purpose of eliminating the non-specific bindings between our lysates and the beads, lysates were also pre-cleared by incubating them with the beads without the anti-FLAG antibody. To immunoprecipitate the PhoB-FLAG-DNA complex, beads pre-incubated with the antibody were added in both lysates from MG1655_PhoB_FLAG and MG1655 strains at 4°C overnight. The beads were washed once with IP buffer (10 mM Tris-HCl (pH 7.4), 0.1 M NaCl, 1 mM EDTA, and 0.05% [v/v] Tween-20 and 1 mM fresh PMSF), twice with ChIP wash buffer I (10 mM Tris HCl (pH 7.4), 300 mM NaCl, 1 mM EDTA, 0.1% Tween-20 and 1 mM fresh PMSF), three times with ChIP wash buffer II (10 mM Tris-HCl (pH 7.4), 500 mM NaCl, 1 mM EDTA, 0.1% [v/v] Tween-20 and 1 mM fresh PMSF), once with ChIP wash buffer III (10 mM Tris-HCl (pH 7.4), 250 mM LiCl, 1 mM EDTA, 0.1% [v/v] Tween-20 and 1 mM fresh PMSF) and once with TE buffer (10 mM Tris-HCl (pH 7.4) and 1 mM EDTA). After removing the TE buffer, beads were incubated twice with elution buffer (50 mM Tris-HCl (pH 7.4), 10 mM EDTA and 1% SDS) at 65°C for 15 min and the two resulting eluted solutions were combined.
After incubating the combined eluted samples with proteinase K (Sigma) to the final concentration of 10.5 U/ml at 42°C for 2 hours, the reverse cross-link procedure was performed by incubating at 65°C overnight to unlink the covalent bonds formed by formaldehyde between peptides and DNA. Samples were then treated with RNase A (Sigma) to the final concentration of 26 μg/ml, followed by purifying DNA from the RNase A-treated samples using the PCR purification kit (Qiagen).
Whole genome tiling array analysis for ChIP-chip experiments
The NimbleGen 385 K high density tiling array for E. coli K12 MG1655 (Cat. No. 05542901001) was used for our ChIP-chip assay. The instructions of the NimbleGen's protocol (version 2.0) were followed for all procedures. Immunoprecipitated samples were amplified by whole genome amplification kit (Sigma) twice and pooled together. The amplified samples from MG1655_PhoB_FLAG were labelled with Cy5 dye while the control samples from MG1655 were labelled with Cy3 dye. After the hybridization step, the arrays were washed and then scanned with an Axon scanner (GenePix 4000B).
The scanned TIF image files were then processed by NimbleScan software to generate the intensity pair files. The R package Ringo 
was used to read the pair files, and the limma package 
was used for within- and between-array normalization 
. The averaged values of normalized Cy5 and Cy3 intensities from triplicate samples were used to calculate the log2
-ratios (Cy5/Cy3). The enriched regions were then identified by the CMARRT package 
with a controlled error rate set at 0.05. Our ChIP-chip data had been submitted to NCBI GEO database and the GSE Series record is GSE21857.
To find the position weight matrix (PWM) of PhoB binding sites, E. coli
K12 MG1655 sequences of all the enriched regions were extracted from NCBI RefSeq (accession no. NC_000913). We used the MEME program 
to search for the most significant conserved pattern with pattern length ranging from 18 to 22 bps (accomplished by using the –minw 18–maxw 22 options of MEME). The range was selected because the previously reported PhoB binding pattern is 18 bps in length 
, while the structure information indicated that the site is 22 bps 
. A seven-order background model was built from the whole E. coli
K12 MG1655 reference sequence (accomplished by using the –bfile <background model file> option in MEME). In addition, sites on both strands were allowed (accomplished by using the –revcomp option). The sequence logo 
was then used to present the PWM graphically.
Gene expression microarray and analysis
The Affymetrix E. coli
Genome 2.0 array was used to investigate gene expression status in the presence and the absence of PhoB activity. The E. coli
K12 MG1655 and MG1655_PhoB_KO strains grew in MOPS minimal medium containing 200 μM K2
and 0.4% glucose. At an OD600
of 1.0, cultures were treated with 10 mg/ml lysozyme and 10% SDS at 4°C for 5 min to lyse bacterial cells. Then, the protocol for total RNA purification using TRIZOL reagent (Sigma) was followed. The Affymetrix standard protocol was then applied for cDNA synthesis, fragmentation, biotin labelling and hybridization. The raw CEL files were normalized by a robust multi-array average approach 
. The microarray data have also been included in GSE21857 of NCBI GEO database. To assess statistically significant differential expression, we applied linear models and empirical Bayes methods 
through the limma package, and the Benjamini and Hochberg's q-value threshold was set at 0.05. The filtered results were considered as the differentially expressed genes. To investigate the functions of the differentially expressed gene, the functional categories of the clusters of orthologous group (COG) were used 
Construction and assay of promoter::luciferase fusions
The promoter regions of the PhoB targets identified in ChIP-chip experiments were amplified by PCR from the MG1655 strain using the primers listed in Table S3
. After treatment with NheI and NcoI restriction enzymes, the digested linear products were then ligated into a NheI-NcoI digested pGL3-basic vector (Promega). The pGL3 plasmid contains a promoterless luciferase gene. The cultivation condition was in MOPS minimal medium supplemented with 200 μM K2
and 0.4% glucose at the same condition as the experiments for ChIP-chip assay and gene expression microarray (). The luciferase activities were measured using a luciferase assay system (Promega).
PhoB-His fusion protein purification
In order to overexpress the PhoB-His fusion protein, the PhoB coding region was cloned into a pET21d (+) plasmid. This constructed vector expressing the PhoB-His(6x)
fusion protein was transformed into BL21. Overnight cultures were diluted 1
500 into 250 mL LB cultures containing 100 μg/μl ampicillin. The cultures were grown at 37°C until an OD600
of 0.4~0.6, then treated with 1 mM IPTG to induce PhoB-His(6x)
expression and then grown at 37°C for another 2 hours. After centrifugation, the pellets were resuspended in 10 ml lysis buffer (20 mM NaH2
, 500 mM NaCl, 20 mM imidazole, and 1 mg/ml lysozyme). The cells were lysed for 30 min at 4°C and then the lysates were cleared by centrifugation at 14000 g for 30 min at 4°C. After applying the lysate to the Ni-sepharose column (GE Healthcare), the column was washed two times by 4 ml wash buffer (20 mM NaH2
, 500 mM NaCl, and 30 mM imidazole). The elution was performed by applying 1 ml elution buffer (20 mM NaH2
, 500 mM NaCl, 500 mM imidazole) to the column four times. The eluted samples were dialyzed in the storage buffer (25 mM Tris-HCl, 50 mM NaCl, 0.1 mM EDTA, and 0.1 mM DTT (pH 7.4)). The concentration of PhoB-His fusion protein was determined by the Bradford assay (Bio-Rad) using the bovine serum albumin (BSA) as the standard.
Gel mobility shift experiments
The synthetic single-stranded 60 bp DNA fragments centered at PhoB putative binding sites were used in these experiments (Table S3
). DNA fragments were first 3′-end labeled with biotin using a DNA 3′ End Biotinylation Kit (Pierce) and then annealed before use. Before the binding assay, the PhoB-His fusion protein was phosphorylated in the reaction buffer (50 mM Tris-HCl, 10 mM MgCl2
, 0.1 mM DTT, and 20 mM acetylphosphate) at 37°C for 75 min 
. The phosphorylated PhoB-His fusion protein was then used in the mobility shift assays. Each binding reaction contained 20 fmol 3′-end biotin labeled dsDNA, 20 mM Tris-HCl (pH 7.0), 50 mM NaCl, 1 mM DTT, 10 mM MgCl2
, 100 μg/ml BSA, and 0.5 μg/ml poly dI-dC with various amounts of PhoB-His(6x)
fusion protein (see and S4
). Reactions were incubated for 15 min at 37°C, and then loaded onto a 6% native polyacrylamide gel running at 100 V in 0.5X TBE buffer. After separation, samples were blotted to Amersham Hybond-N membranes using a Hoefer TE 70 device. The labeled biotin signals were transferred and detected using a LightShift Chemiluminescent EMSA Kit (Pierce) according to the manufacturer's instructions. For each tested target, at least two to three biological replicates were performed and the best figure was picked and shown in and S4