Culture conditions and treatment with FeSO4
Isolated colonies of Chromobacterium violaceum, ATCC 12472 strain, were inoculated in liquid Luria Bertani (LB) medium for 1618 hours at 28C and shaken at 200rpm.
The treatment was performed in an Erlenmeyer flask diluting the pre-cultivated bacteria with liquid LB medium (1:10) in a final volume of 100mL. The FeSO4 solution was prepared at an initial concentration of 500mM and was further filtered with 0.22 μm GHP membrane disc filters [Acrodisc] and added at a final concentration of 9mM. This concentration was chosen after a screening of treatment concentrations in which at only 9mM of iron we observed different patterns of proteins synthesis at 1D-SDS-PAGE, in contrast to the negative control as well as previous results of our group that have demonstrated the high resistance of C. violaceum to iron (data not shown). The negative control consisted of C. violaceum grown in a LB broth. The treatment exposed the culture for four hours under the same conditions as described above. A measurement of C. violaceum growth was performed between 04 hours and the optical density was read at 600nm in a 96-well spectrophotometer. The negative control and the experimental conditions were performed in biological triplicate.
Total intracellular iron measurement
Total intracellular iron concentration was estimated as described at Barbehenn et al. [18]. First, C. violaceum cultures (100mL) were grown until reaching 0.40.5 OD when the treatment began. After the four-hour treatment, the samples were centrifuged at 4C, 2700g for 15min. Then, samples were washed with 50mM EDTA, followed by another wash with ultrapure water. Finally, the phenanthroline assay was performed as described by Barbehenn et al. [18]. The negative control and the experimental condition were performed in biological triplicate.
Protein extraction
After four hours of treatment, the samples were centrifuged at 2880g and 4C for 20minutes. The supernatant of all samples (including negative control) were discarded and then the samples were washed in 50mM EDTA pH8.0 and centrifuged under the same conditions previously mentioned. Then the samples were washed once more in 10mM TrisHCl, pH8.5 solution and centrifuged at 2880g, 4C for 25minutes. Cells were lysed in a 300 μL 400 μL extraction buffer containing 7M urea, thiourea 1M, 50mM DTT, 0.5% CHAPS, and 30mM Tris pH8.5. Proteins were precipitated by adding 1mL of acetone, vortexing the samples and then centrifuging at 10,000g, 4C for 3minutes. This process was performed a second time, and the sample was centrifuged again for 5minutes. Finally, the proteins were solubilized in the same extraction buffer (300 μL 900 μL).
Antioxidant activity evaluation of Chromobacterium violaceum
In order to evaluate if the iron treatment promotes oxidative stress in Chromobacterium violaceum, antioxidant enzymes catalase (CAT) and superoxide dismutase (SOD) activities were measured on total protein extracts. Catalase Assay Kit 707002 and Superoxide Dismutase Assay Kit 706002 (Cayman Chemical, Ann Arbor, MI) were used according to manufacturers recommendations to quantify the catalase and superoxide dismutase activity level, respectively. The total antioxidant activity was evaluated using the Antioxidant Assay kit from Sigma-Aldrich (CS0790) according to manufacturers instructions. The protein concentration values were used to normalize the enzyme activity. The experiment was performed in biological triplicate.
SDS-PAGE
The total protein extracts were quantified by the Bradford method [19] and 20 μg of protein from each sample was resolved on a polyacrylamide gel under denaturing conditions (SDS-PAGE) at 12%. The marker used was the Precision Plus Proteins WesternC Standard from Bio-Rad. The gel was stained using the Coomassie Colloidal from Sigma-Aldrich.
In gel digestion and peptides extraction
Following electrophoresis, each lane was excised in nine fragments according to the proteins density. All the proteins from each lane were excised and each group of proteins was analyzed independently by mass spectrometry (see below). Proteins were extracted and digested from these fragments as in accordance to the revised Shevchenko et al. protocol [20]. The dye and the SDS were quickly removed by washing the fragments three times in 50% acetonitrile solution (ACN) and 10mM ammonium bicarbonate. Then the bands were dehydrated in ACN at 100%, reduced with 10mM dithiothreitol (DTT) at room temperature and alkylated with 50mM iodacetamide (IAA) in a dark environment. Then, the bands were washed again with 100mM ammonium bicarbonate, dehydrated with 100% ACN, and rehydrated with 100mM ammonium bicarbonate; this was done twice. The bands were then hydrated in a trypsin solution (Trypsin Gold, Mass Spectrometry Grade from Promega [v5280]) prepared according to manufacturers instructions. 3550 μL of trypsin at 20 μg/mL was added to the samples kept on ice. Then we added 50mM of ammonium bicarbonate, sufficient to cover the bands during incubation at 37C for 1618 hours.
For peptide extraction, 1030 μL of 5% formic acid were added to the bands. After 10minutes of incubation at room temperature, the supernatant containing peptides was transferred to another tube. Then a second extraction solution (5% formic acid and 50% ACN) was added in enough volume to cover the bands. The supernatant was transferred to a previously prepared tube containing the already extracted peptides. The extraction of the peptides performed with the second solution was repeated once again. Finally, the solution containing the digested peptides was concentrated in Eppendorfs Concentrator Plus.
Mass spectrometry data acquisition
After the in gel digestion, the samples were loaded onto the liquid chromatography NanoAcquity UPLC system (Waters) connected with an ESI-Q-Tof premier (Waters) mass spectrometer. The tryptic peptides from each sample (4.5 μL) were separated on a BEH130-C18 column (100 μm × 100 mm) at a 600 nL/min flow rate. The gradient varied from 2 to 98% ACN in 0.1% formic acid for 45minutes. The instruments data acquisition mode was set to a data dependent top three and controlled using MassLynx v.4.1.
Protein identification
The ProLuCID search engine v 1.3 [21] was used to compare experimental spectra against those theoretically generated from C. violaceum ATCC 12472 sequence downloaded from Uniprot in January 2013, plus those from 127 common contaminants to proteomic experiments (e.g., Keratin, BSA, etc.). The search was limited to tryptic and semi-tryptic peptide candidates; carbamidomethylation and oxidation of methionine were imposed as fixed and variable modifications, respectively. The search engine accepted peptide candidates within a 50-ppm tolerance from the measured precursor m/z and used the XCorr and Z-Score as the primary and secondary search engine scores, respectively.
The validity of the Peptide Sequence Matches (PSMs) was assessed using the Search Engine Processor (SEPro) v.2.2.0.1 [22]. Briefly, identifications were grouped by charge state (+2 and ≥ +3) and then by tryptic status (tryptic or semi-tryptic), resulting in four distinct subgroups. For each group, the XCorr, ZScore, DeltaCN, and DeltaMass values were used to generate a Bayesian discriminant function. The identifications were sorted in a non-decreasing order according to the discriminator score. A cutoff score was established to accept a false-discovery rate (FDR) of 1% based on the number of decoys. This procedure was independently performed on each data subset, resulting in a false-positive rate that was independent of tryptic status or charge state. Additionally, a minimum sequence length of six amino acid residues was required. Results were post processed to only accept PSMs with less than 10ppm and proteins supported by two or more independent evidences (e.g., identification of a peptide with different charge states, a modified and a non-modified version of the same peptide, or two different peptides). This last filter led to a 0% FDR in all search results for all our sample analyses.
Differential proteomics and functional analysis
The PatternLabs updated ACFold module was employed to pinpoint differentially expressed proteins between the control and iron exposed condition [23],[24]. The revised ACFold module presents increased sensitivity under the Benjamini-Hochberg q-value [25] bound by applying a variable fold-change that varies with the AC-test p-value as a power law [24].
Proteins uniquely identified in one condition (control or iron) were pinpointed according to PatternLabs Approximate Area Proportional Venn Diagram module. To better cope with the limitations from undersampling, we only considered proteins identified in two replicates of each condition, and not found in any replicates of the other condition.
The functional categorization of the proteins was assessed using PatternLabs Gene Ontology Explorer (GOEx) module [26]. Our data analysis used the gene ontology database (OBO v1.2 - http://www.geneontology.org/GO.format.obo-1_2.shtml), downloaded February 16th, 2013 and the C. violaceum gene ontology annotation in the Uniprot text file format of its protein sequences.
Functional domain analysis
The protein sequences were compared [27] against the protein database from NCBI for automatic annotation and functional domains were predicted using the Conserved Domain Database (CDD) [28].
The structural prediction of hypothetical ORFs was performed using the server for protein homology detection HHpred (data not shown) [29]. FASTA sequences from each hypothetical protein were submitted to the server and the PDB database from February 23rd of 2013 was used. No specific organism proteome was selected, the method for multiple sequence alignment (MSA) generation was HHblits [30], up to 3 MSA generation iterations was used, the secondary structure score was applied, and the alignment mode was set to local.
Protein-protein interaction network analysis
The metasearch tool STITCH 3.1 (http://stitch.embl.de) was used to estimate protein-protein interaction networks related to iron response. STITCH is a tool used to explore known and predicted interactions between proteins, and chemical or physical agents. These agents interconnected by evidences derived from experiments, databases, and literature. One network was generated composed of proteins that were identified exclusively in the iron treatment or having an increased expression after being exposed to the metal. As the interaction targets were derived from experimental data, a high confidence index (0.700) was used to generate the networks. All prediction methods were activated: neighborhood, gene fusion, co-occurrence, co-expression, experiments, and databases text mining.
The assembled network was exported to be subsequently analyzed in Cytoscape 2.8.2 and Cytoscape 3.0 [31]. The former was used to calculate the centrality indexes and the later for the remaining analysis. The most relevant proteins sub-networks were selected using the Cytoscape MCODE v. 1.4.0 plugin (Molecular Complex Detection) [32]. The MCODE analysis parameters were: loops inclusion; degree cutoff of 2; haircut option enabled (which leads to deletion of nodes cluster with a single connection); fluff option enabled, node score cutoff at 0.2; K-core at 2, and maximum depth of 100. The only clusters used were those in which the MCODE index score was greater or equal to 2.5.
The network centrality calculations were computed from local networks and topologies. The networks bottlenecks were identified through a Betweenness vs. Node Degree chart with the values generated by the CestiScaPe 1.21 plugin installed at Cytoscape 2.8.2. Betweenness indicates the extent to which a particular node is among all other nodes in a network and usually shows the influence of this node on information propagation within the network [33],[34]. Node Degree corresponds to the number of connections that a particular node has with other adjacent nodes. High node degree levels are called hubs [35]. Thus, a particular node with high node degree and betweenness values represents a bottleneck, or a protein that interconnects many biological processes [34].
Total RNA extraction from C. violaceum and cDNA synthesis
Isolated colonies of C. violaceum were cultured in absence and the presence (9mM) of iron in the previously mentioned conditions. Further, 2mL of the culture was used to extract and purify total RNA using the RNAspin Mini Isolation kit according to manufacturers instructions (GE, catalog number 25-0500-72).
Once extracted, the total RNA was used to synthesize cDNA with random primers with the High capacity cDNA reverse transcriptase kit according to manufacturers instructions (Applied Biosystems, catalog number 4368814).
RT-PCR analysis
The Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) was used to validate the expression of the hypothesized operon as a unique transcript. As such, RT-PCR was performed to provide the transcript evidence of regions encompassing the ORFs CV_0869 (the first) and CV_0867 (the last). Primers for the above mentioned genes were designed using Primer 3 (v. 0.4.0) software.
RT-PCR was performed using AmpliTaq Gold 360 Master Mix (Catalog Number 4398881, Applied Biosystems). One microlitrer of cDNA was used and a final volume of 25 μL was used according to the manufacturers instructions. Then, the PCR steps were the following: initial denaturation at 98C for 5min, 40cycles of 98C for 30sec, 59.8C for 30sec, 72C for 90sec and a final extension at 72C for 7min were followed. Reactions using water and RNA instead of cDNA were carried out as negative controls. To verify the size of the amplicons, 5 μL of the PCR reaction was loaded on 1% agarose gel and submitted to electrophoresis. The DNA ladder of 1 Kb (Promega) was used as reference. An additional file (Additional file 1) shows the sequence of the primers used in this work.
Validation of expression by Real-Time quantitative PCR
The quantitative real-time PCR was used to validate the proteome analysis by verifying the expression of the genes CV_0868 and CV_0867, both comprising what we hypothesized as a newly described operon. Seven nanograms of cDNA (produced as mentioned above, from the iron-culture and control) and a final concentration of 0.25 nM of each primer were applied in a 10 μL reaction using Power SYBR Green PCR Master Mix (Applied Biosystems, 4367659) in an One-Step cycling with the following conditions: 95C for 3min, 40cycles of denaturation at 95C for 3sec, annealing/extension at 60C for 15sec. The 16S rRNA was used as endogenous control and to normalize the expression of the other two genes. The relative quantification was assessed by ΔCt comparative analysis. The primers were designed using Primer 3 (v. 0.4.0). Primers sequences are described in Additional file 1. Statistical analysis was performed according to the t-test. Results were considered significant for p <0.05. Two biological replicates were used.
