Transcriptional regulator ArcA mediates expression of oligopeptide transport systems both directly and indirectly in Shewanella oneidensis

In γ-proteobacterial species, such as Escherichia coli, the Arc (anoxic redox control) two-component system plays a major role in mediating the metabolic transition from aerobiosis to anaerobiosis, and thus is crucial for anaerobic growth but dispensable for aerobic growth. In Shewanella oneidensis, a bacterium renowned for respiratory versatility, Arc (SoArc) primarily affects aerobic growth. To date, how this occurs has remained largely unknown although the growth defect resulting from the loss of DNA-binding response regulator SoArcA is tryptone-dependent. In this study, we demonstrated that the growth defect is in part linked to utilization of oligopeptides and di-tripeptides, and peptide uptake but not peptide degradation is significantly affected by the SoArcA loss. A systematic characterization of major small peptide uptake systems manifests that ABC peptide transporter Sap and four proton-dependent oligopeptide transporters (POTs) are responsible for transport of oligopeptides and di-tripeptides respectively. Among them, Sap and DtpA (one of POTs) are responsive to the SoarcA mutation but only dtpA is under the direct control of SoArcA. We further showed that both Sap and DtpA, when overproduced, improve growth of the SoarcA mutant. While the data firmly establish a link between transport of oligopeptides and di-tripeptides and the SoarcA mutation, other yet-unidentified factors are implicated in the growth defect resulting from the SoArcA loss.


Results
the arcA mutation compromises peptide utilization in S. oneidensis. Our preliminary analysis had demonstrated that the growth defect of the S. oneidensis arcA mutant in complex medium LB is dependent on tryptone 18 . As the SoarcA mutant cultivated in mineral medium MS with lactate as the carbon source (dubbed as MS-L throughout this study) is not defective with respect to aerobic growth (Figs 1A, S1A), this study was initiated by systematically assessing effects of tryptone on growth of the wild-type and ∆SoarcA strains in MS-L with tryptone addition. Clearly, growth of the wild-type increased with tryptone when its concentrations were no more than 0.2%, and became less sensitive to further augment to 0.5% in tryptone amounts (Fig. 1B). By generation times for 2 h during the exponential phase (Table S1), there was a 1.5-fold difference in growth rates between cells grown with 0.25% tryptone and without. In contrast, the ∆SoarcA strain was not responsive to the treatment with tryptone at 0.2% or lower, and grew slightly faster with 0.5% tryptone (Fig. 1C). These data verify that the SoarcA mutation compromises the growth-promoting effect of tryptone. Therefore, we reasoned that utilization of oligopeptides and/or amino acids is impaired by the mutation because tryptone is abundant in such nutrients, a result of incomplete enzymatic hydrolysis of casein.
To test this, impacts of tryptone, casamino acids, and an amino acid mixture on growth of the wild-type and ∆SoarcA strains were compared. In contrast to tryptone, neither casamino acids nor the amino acid mixture could elicit a significant difference in growth rates of both strains (Fig. S1B,C). This observation rules out the possibility that amino acids are an important factor responsible for the growth defect of the ∆SoarcA strain. Subsequently, we tested effects of oligopeptides (O-P, peptides of 4 and more residues, Table 1) of different lengths on growth of the wild-type and ∆SoarcA strains. Additions of a variety of oligopeptides only modestly improved growth of the wild-type (compared to that in MS-L) but showed no detectable effects on growth of the ∆SoarcA strain (Fig. 1D,E). Similar results were obtained with a mixture of di-tripeptides (DT-P) ( Table 1). When both O-P and DT-P (O/DT-P) were present, growth of the wild-type was further accelerated whereas the ∆SoarcA strain remained unaffected. These data therefore suggest that the SoarcA mutant is defective in utilization of short peptides.
the ∆SoarcA strain is heavily defective in growing on o/Dt-p. Growth monitored above was in fact mainly supported by lactate, the carbon source in MS-L, which interferes with the contribution of the peptide mixtures under test. As an attempt to show the growth defect of the SoarcA mutant more clearly, growth on O/ DT-P, casamino acids, and an amino acid mixture as carbon sources was assessed ( Fig. 2A, 2B, 2C). Both the wild-type and ∆SoarcA strains were able to grow in MS with each of these carbon sources at concentrations up to 0.5% (Fig. S2). The amino acid mixture at all testing concentrations supported these two strains to grow comparably (Fig. 2C), manifesting that the SoarcA mutation does not have a significant impact on utilization of amino acids. In contrast, O/DT-P at all testing concentrations were not only more effective than the amino acid mixture in supporting growth, but more importantly, distinguished from the wild-type from the SoarcA mutant with respect to growth rate ( Fig. 2A). Clearly, the difference in growth rates between them was substantial (Table S1). In the case of casamino acids, the growth difference between the two strains was negligible, noticeable, and significant with 0.1%, 0.2%, and 0.5% respectively (Fig. 2B). This further supports the impact of the SoarcA mutation in utilization of O/DT-P because small peptides make up ∼6% of casamino acids (250-500 (molecular weight), ∼6%, BD Bacto TM ).
Peptide digestion is not significantly affected by the SoarcA mutation. Utilization of peptides by the cell depends on two distinct steps, import from the outside and digestion into amino acids. To evaluate the contribution of peptide digestion to the growth difference between the wild-type and ∆SoarcA strains, we measured activity of metalloaminopeptidases, which are predominant in short peptide hydrolysis 20 . Both the wild-type and ∆SoarcA strains were grown in MS-L with tryptone to the mid-log phase and cells were collected by centrifugation. The pelleted cells were resuspended and disrupted by sonication with external cooling on ice, and after removal of cellular debris by centrifugation, cell extracts were assayed for aminopeptidase activity using fluorescent peptide-AMC (7-amido-4-methylcoumarin) substrates 21 . As shown in Fig. 3A, the amount of hydrolyzed peptide of the cell extracts from both the wild-type and ∆SoarcA strains linearly increased with time. The difference in peptidase activity between two strains, by normalized total protein amounts of each sample, was statistically insignificant. To verify that the hydrolysis is mainly attributed to metalloaminopeptidases, activities in the wild-type and ∆SoarcA cell extracts were assayed in the presence of 1,10-phenanthroline (OPT), a metal chelator functioning as unspecific and potent metalloprotease inhibitor 21 . Clearly, OPT inhibited the hydrolysis of peptide-AMC strongly; at 1 mM the maximum inhibition was achieved. In addition, we examined the effect of serineproteases on peptide hydrolysis. In the presence of 1 mM AEBSF (4-(2-aminoethlyl)benzenesulfony fluride), a broad spectrum and irreversible inhibitor for serineproteases 22 , the amount of fluorescence in the wild-type cell extracts was also affected (Fig. S3), suggesting a role of serineproteases in peptide digestion. Growth on LB plates. Cultures of indicated strains prepared to contain approximately 10 9 cfu/ml were regarded as the undiluted (dilution factor, 0), which were subjected to 10-fold series dilution. Five microliters of each dilution was dropped on indicated agar plates. Results were recorded after incubation of 18 h and 30 h for LB and MS-L respectively. Successful complementation by expressing a copy of the arcA gene in trans (ArcA) has been performed and reported repeatedly before. Data for cultures grown in liquid LB were shown in  However, given that the impact of AEBSF was substantially weaker than that of OPT, it is apparent that metalloaminopeptidases are a major contributing factor for peptide digestion. Importantly, the data obtained from the wild-type and ∆SoarcA strains were similar. Additionally, we performed the droplet assay to assess susceptibilities of the wild-type and ∆SoarcA strains to OPT and found that they were similar on MS-L plates containing tryptone (Fig. 3B). These data thus rule out the possibility that peptide hydrolysis is critically responsible for the growth defect of the ∆SoarcA strain.
Selectivity of Sap and pots in peptide transport. In bacteria, oligopeptide transport is predominantly mediated by transporters of three families: the proton-dependent oligopeptide transporter (POT), the ATP-binding cassette (ABC) peptide transporter, and the electrochemical potential-driven oligopeptide transporter (OPT) 23,24 . In the S. oneidensis genome, genes encoding POT and ABC peptide transporter system but not OPT are present 25 . For confirmation, the S. oneidensis proteome was screened for homologues of transporters in these three families whose oligopeptide transport activity has been established, including ABC peptide transporter systems of E. coli (OppABCDF) 26 and Pseudomonas aeruginosa (DppABCDF) 27 , as well as a variety of POTs and OPTs typified by E. coli DtpA (formerly YdgR or TppB) and DtpB (formerly YhiP) 28,29 , and Shewanella pealeana Spea_3867 (Spe1) 23 . BLASTp screening revealed an ABC peptide transporter system, SapABCDF, and four POTs, SO_0002, SO_1277, SO_1505, and SO_3195 (Table 2). Among them, SO_0002 and SO_1277 have been proven to be di-tripeptide transporters biochemically 30,31 , and therefore are named as DtpA and DtpB respectively.
To determine selectivity of Sap and POT peptide transport, we constructed deletion strains for each of them. While all four POT genes are transcribed from single-gene operons, a single operon, sapABCDF, encodes the ABC peptide transporter system. In MS-L containing tryptone, ∆sap (a strain lacking the entire sap operon) mutants showed a significant defect in growth, implicating that Sap plays a major role in transporting short peptides contained in tryptone (Fig. 4A). For complementation, we manipulated expression of the sap operon by  www.nature.com/scientificreports www.nature.com/scientificreports/ using IPTG-inducible promoter Ptac 32 . Defects of ∆sap in growth in tryptone-containing MS-L were corrected by expressing the missing sap genes in trans with IPTG ranging from 0.05 to 0.5 mM (Fig. 4A). According to previous calibration of the Ptac promoter within pHGE, 0.5 mM IPTG confers the activity of the Ptac promoter approximately 2000 Miller units [33][34][35] , more than 6 times of that of the sap promoter in the wild-type (refer to Fig. 6B). Therefore, these data, in addition to validate the correlation between the mutation and phenotypes, also indicate that the Sap peptide transporter produced in the wild-type is sufficient for importing short peptides contained in tryptone to support full growth on lactate as the carbon source. The selectivity of Sap was then investigated with O-P and DT-P as the carbon source. While the ∆sap strain grew only slightly slower than the wild-type on DT-P, it could barely do so on O-P ( Fig. 4B,C). Expectedly, growth of the ∆sap strain on O-P was restored by expressing the sap operon in the presence of 0.05 mM IPTG. Moreover, we found that growth on O-P, unlike that in the tryptone-containing MS-L, further increased with IPTG concentrations (up to 0.5 mM), manifesting enhanced O-P uptake by overproduced Sap (Fig. 4B). On the contrary, Sap in overabundance had little effect on growth of the ∆sap strain on DT-P (Fig. 4C). These data indicate that the Sap system is primarily for uptake of peptides with 4 amino acid residues or larger.
In the case of POTs, depletion of each POT alone did not significantly affect growth of the wild-type in tryptone-containing MS-L (Fig. S4). Although this observation implies that POTs are not critical in growth improvement by tryptone when lactate is used as the carbon source, it may be a result of functional overlapping among the four putative POTs. To test this, a strain (∆pot) devoid of all four POT genes was constructed and characterized with peptides as the carbon source. Compared to the wild-type, this strain displayed normal and substantially impaired growth on O-P and DT-P as the carbon source, respectively (Fig. 5A,B). The dependence of DT-P import on POTs was further verified by assessing the susceptibility of the ∆pot strain to alafosfolin, a di-peptide analog antibiotic that is a good substrate for POTs 28 : the ∆pot strain exhibited substantially elevated resistance whereas the ∆sap strain was almost unaffected (Fig. 5C). For complementation, we manipulated expression of the dtpA gene by using IPTG-inducible promoter Ptac. Effects of dtpA expression on growth of the POT-deficient mutant on di-tripeptides were evident: significantly improved with 0.01 mM IPTG and fully restored to the wild-type level with 0.05 mM (Fig. 5B). DtpA in overproduction (0.5 mM IPTG) was able to further enhance growth on di-tripeptides and increased susceptibility of the ∆pot strain to alafosfolin (Fig. 5B,C). These data clearly manifest that POTs dictate transport of di-tripeptides, and that these POTs, at least some of them, can functionally complement one another. peptide transport has a role in explaining the ∆SoarcA growth defect. To determine whether these peptide transporters are implicated in the growth defect of the ∆SoarcA strain, we compared expression of their coding genes in the wild-type and the mutant with qRT-PCR. Cells of both strains grown with the short-peptide mixture as the carbon source to the mid-log phase were used. Analysis of transcripts of these genes revealed that the sap and dtpA genes displayed significantly reduced transcription in the ∆SoarcA strain, approximately by 30% and 3-fold respectively, while the remaining three genes were transcribed comparably in these two strains (Fig. 6A). We then utilized an integrative lacZ reporter system for confirmation 36 . DNA fragments of ∼400 bp upstream of the coding sequences for these operons, which are sufficiently long to cover entire promoters for most of operons, were amplified and cloned into the reporter vector pHGEI01. The resulting vectors, verified by sequencing, were introduced into relevant S. oneidensis strains for integration and subsequent removal of the antibiotic marker 37 . From similarly prepared cells, we surprisingly found that the differences in expression of the sap and dtpA genes were enhanced, and dtpB and SO_1505 were also modestly down-regulated in the mutant (Fig. 6B). These data indicate that at least sap and dtpA are subject to positive regulation of SoArcA, although it is clear that more investigations are needed to address the difference in data revealed by these two methods.
To assess whether the down-regulated expression of sap or dtpA is responsible for the growth difference between the wild-type and ∆SoarcA strains on O-P and DT-P, we monitored growth of the ∆SoarcA strain with sap and dtpA expressed at different levels. Consistent with compromised production of Sap and POTs, the SoarcA mutant exhibited significantly impaired growth on O-P and DT-P (Fig. 6C,D). Importantly, the growth difference between the wild-type and ∆SoarcA strains on O-P and DT-P decreased with forced expression of sap and dtpA respectively. In the presence of 0.05 mM IPTG, these transporters significantly improved growth of the SoarcA mutant on the respective carbon source (Fig. 6C,D), contrasting undetectable effects on the growth of the wild-type (Fig. S5). These data support that the SoarcA mutant suffers from impairment in peptide import.  www.nature.com/scientificreports www.nature.com/scientificreports/ In addition, we found that the SoarcA mutant with overproduced Sap and DtpA by 0.5 mM IPTG grew at further accelerated rates, but it still was inferior to the wild-type (Fig. 6C,D). For comparison, the same treatments enable ∆sap and ∆pot strains to grow faster than the wild-type (Figs 4B, 5B). Thus, while these data indicate that peptide transport has a role in determining growth of the SoarcA mutant, other factors are involved.
dtpA is under the direct control of SoArcA in S. oneidensis. The promoter region of dtpA (encoding one of POTs) but not the sap operon has been predicted to have an SoArcA-binding site (GGTAAATAGATGTAA; E. coli and S. oneidensis consensus, GTTAATTAAATGTTA) 13,14 . However, the confidence for this site is rather low, based on its weight value (6.5) obtained from genome screening. To test whether these genes are under the direct control of SoArcA, electrophoretic motility shift assay (EMSA) was used with the recombinant S. oneidensis His 6 -tagged SoArcA protein produced in E. coli and purified as described before 13,14 . As phosphorylation of SoArcA is required for its specific binding, only SoArcA (SoArcA-P) phosphorylated by carbamoyl phosphate was used. The DNA fragments, approximately 300 bp in length centered by the predicted SoArcA-binding motif (if available) of the genes to be tested, were amplified with 33 P end-labeled primers. Among the tested PCR sequences in EMSA with phosphorylated SoArcA, dtpA showed SoArcA(-P) binding activity (Fig. 7A). In contrast, retardation of the sequences of sap and genes for other two POTs was not observed.
Interactions in vivo between SoArcA and dtpA and sap promoter region sequences were then tested with bacterial one-hybrid (B1H) analysis, which has proven to be a robust technique applicable to a wide variety of different transcriptional factor families and has been frequently used in S. oneidensis 33,38,39 . This is enabled by the fact that E. coli arcB expression restores the wild-type phenotype of an arcS/hptA deletion in S. oneidensis and that, vice versa, upon ectopic production ArcS/HptA can functionally complement the loss of EcarcB in E. coli [10][11][12] . In the B1H system, vectors containing 'bait' (DNA) and 'target' (DNA-binding regulator) are co-transformed into  Alafosfolin susceptibility determined by disk diffusion assay. Ten µl of 2 mg/ml alafosfolin was added to filter paper discs (8 mm) on cultures of indicated strains. Plates were incubated at 30 °C for 24 h. Asterisks not associated with a bracket represent statistically significant difference compared to the wild-type value (*P < 0.05; **P < 0.01; ***P < 0.001). Asterisks associated with the bracket represent the difference between the two linked strains. Experiments were performed independently at least 5 times, and data were presented as the average and error bars representing standard errors. (2019) 9:13839 | https://doi.org/10.1038/s41598-019-50201-4 www.nature.com/scientificreports www.nature.com/scientificreports/ reporter cells and positive interaction between them allows growth in the presence of 3-amino-1,2,4-triazole (3-AT). When cultured under aerobic conditions, growth was detected only from the positive control pair SoArcA/P SO_1661 (Promoter region of SO_1661) 13,14 with 20 colonies on average (Fig. 6B). Given that similar studies with other regulators/cognate DNAs usually generate ∼300 colonies 33,39 , we reasoned that this low efficacy is likely due to that i) EcArcA proteins may compete for the target DNA and ii) most of SoArcA proteins are not phosphorylated under test conditions. To test the first possibility, the EcarcA gene was knocked out from the reporter strain and the resulting mutant was verified by genetic complementation with respect to increased sensitivity to many dyes such as toluidine blue and crystal violet 40,41 (Fig. S6A). However, with this strain the B1H assay gave out similar results (Fig. S6B), indicating that the interference from EcArcA is negligible. To improve phosphorylation, we cultivated cells under fumarate-respiring conditions. In the absence of oxygen, more than 200 colonies were obtained from the reporter strains with P SO_1661 /SoArcA, either with or without the EcarcA gene (Figs 6B, S6B). Importantly, the reporter strains containing P dtpA /SoArcA was able to form 50 or so colonies on 3-AT plates whereas P sap /SoArcA failed to support growth on 3-AT. These data, collectively, indicate that SoArcA directly regulates transcription of dtpA but not genes for other peptide transporters in S. oneidensis.

Discussion
The Arc system is among the most intensely investigated regulatory systems in γ-proteobacteria and its physiological function(s) known to date is largely derived from studies on the E. coli paradigm. However, as a TCS, the Arc system is rather unusual. First, the Arc system is only known to exist in four γ-proteobacteria families: Alteromonaceae, Enterobacteraceae, Vibrionaceae, and Pasterurellaceae, although its components, OmpR-family regulators and hybrid sensor histidine kinases, are common in bacteria 1 . Second, unlike most of TCSs, the components of all Arc systems, including both the typical (ArcB-ArcA) and the atypical (without a full length ArcB), are never encoded by genes in a single operon, not even in proximity in most cases 42 . Third, the atypical Arc system, as in S. oneidensis, is in fact quite common among the Alteromonads 1 .
Although regarded to be a global regulator with pleiotropic impacts in physiology, the canonical Arc system functions primarily as a global repressor of aerobic metabolic pathways 2,43 . However, this may be one side of the regulatory effects about Arc systems. There are only several overlaps in the ArcA-controlled regulons of E. coli and S. oneidensis, indicating that the physiological function of S. oneidensis ArcA is substantially different from that of the E. coli counterpart 13 . The best example is growth under aerobic conditions. In contrast to the E. coli paradigm, www.nature.com/scientificreports www.nature.com/scientificreports/ which appears to have a rather limited role during aerobic growth, the SoArc system is critical 6,13,44 . However, the growth phenotype resulting from the SoArcA loss is conditional, unlike the defect in cell envelope, another major defect which is independent of growth parameters tested.
The dependence of the growth phenotype on tryptone indicates that it is associated with nutrients, oligopeptides in particular 18 . Following this lead, in this study we examined roles of peptide transporters and peptidases in S. oneidensis. Peptidases, including both metalloaminopeptidases and serineproteases, are dispensable for the growth defect resulting from the SoArcA loss, ruling out the possibility that peptide hydrolysis is critically involved. On the contrary, oligopeptide transport is implicated. ABC peptide transporter consisting of SapABCDF and four POTs are primarily responsible for uptake of oligopeptides and di-tripeptides respectively. By using both qRT-PCR and lacZ-reporters, we concluded that expression of dtpA (encoding one of POTs) and sap is significantly down-regulated in the absence of SoArcA. Importantly, both DtpA and Sap, when manipulatedly produced, are able to improve growth of the SoarcA mutant, supporting a link between the growth defect resulting from the SoArcA loss and peptide transport. Despite this, it is clear that there are other factors implicated in the growth defect of the SoarcA mutant.
We also presented evidence from both in vitro and in vivo experiments to show that transcription of dtpA but not the sap operon is directly mediated by SoArcA. The interaction between SoArcA and the dtpA promoter region appears weak, likely due to low conservation of its SoArcA-binding site as suggested before 13,14 . In S. oneidensis and E. coli, it has been demonstrated that phosphorylation is essential for biological activity of ArcA and ArcA can be phosphorylated under anaerobic as well as aerobic conditions 8,[10][11][12]15,19,45 . However, studies from us and others have argued that SoArcA proteins exist mostly in the phosphorylated form when cells grow aerobically 11,13,15 , contrasting the scenario observed in E. coli because the histidine kinase domain of its EcArcB is activated under anaerobic conditions 45,46 . Results of the B1H analysis performed in E. coli certainly support this proposal. In E. coli cells grown aerobically, positive interactions were detected only between SoArcA and DNA fragments imbedded with a highly conserved binding motif (SO_1661). Under oxygen-free conditions, more EcArcB became activated, which in turn phosphorylated more SoArcA. As a result, SoArcA exhibited stronger binding to the SO_1661 promoter, and more importantly, it was able to interact with the dtpA promoter region, validating that SoArcA directly regulates transcription of the dtpA gene in vivo.
During the investigation, we noted a significant difference in expression levels of genes for peptide transporters obtained from qRT-PCR and lacZ-reporters. With the lacZ-reporters, all five operons were found to be down-regulated in the mutant whereas qRT-PCR only differentiated dtpA and sap. We speculate that this may be due to compromised translation efficacy in the absence of the SoArc system. Translation is carried out by ribosome and there is a linear relation between the ribosome mass fraction and the growth rate 47 . This understanding resonates with our previous findings that the SoarcA mutation concertedly down-regulates the majority of components in protein synthetic machinery 13,18 . In addition, a remarkable number of translation-associated proteins, such as translation initiation and elongation factors and tRNA synthases, are present in less amounts in www.nature.com/scientificreports www.nature.com/scientificreports/ the ∆SoarcA strain than the wild-type with respect to transcripts and proteins. On the contrary, the SoArcA loss has no effect on expression of most of the transcription machinery members. We are working to test this notion.
The Arc system of E. coli was initially recognized to confer resistance to dyes such as toluidine blue O and methylene blue, photosensitizers that facilitate generation of reactive oxygen species in the presence of light 40,48 . However, the underlying mechanism for this phenotype remains largely elusive, implying that it may be unexpectedly complex 41,49 . In parallel, we previously showed that the cell envelope of S. oneidensis is impaired in the absence of the Arc system. A whole-genome screening with dual-effect transposon constructs, which are able to both inactivate genes by interruption and overexpress genes after the insertion site, fails to identify suppressors for the phenotype 19,50 . The failure suggests that the observed envelope defect, the same as the growth defect, unlikely due to altered expression of single gene/operon, is a result of multi-fold complex impacts in physiology. We continue our effort to identify the important proteins for both the growth and envelope defects as they are the key to better understand the physiological significance of this atypical Arc system.

Materials and Methods
Bacterial strains, plasmids, and culture conditions. A list of all bacterial strains and plasmids used in this study was given in Table 3. Information for primers used in this study was given in Table S2. Chemicals are from Sigma-Aldrich Co. unless otherwise noted. E. coli and S. oneidensis strains under aerobic conditions were grown in Lysogeny Broth (LB, Difco, Detroit, MI) medium, which was modified to contain tryptone (10 g/L), yeast extract (5 g/L), and NaCl (5 g/L), at 37 °C and 30 °C for genetic manipulation. When needed, the growth medium was supplemented with chemicals at the following concentrations: 2,6-diaminopimelic acid (DAP), 0.3 mM; ampicillin sodium, 50 µg/ml; kanamycin sulfate, 50 µg/ml; and gentamycin sulfate; 15 µg/ml. The amino acid mixture was composed of 500 µg/ml of each alanine, cysteine, glycine, histidine, aspartic acid, glutamic acid, phenylalanine, asparagine, glutamine, methionine, leucine, isoleucine, proline, serine, threonine, lysine, and valine, 50 µg/ml of tryptophan and tyrosine. The MS basal medium was sterilized by in an autoclave, and required supplements sterilized by filtration were added before usage. Fresh medium was inoculated with overnight cultures grown from a single colony by 1:100 dilution, and growth was determined by recording the optical density of cultures at 600 nm (OD 600 ). For anaerobic growth, mid-log phase aerobic cultures were pelletted by centrifugation, purged with nitrogen, suspended in fresh medium containing 20 mM fumarate as the electron acceptor prepared anaerobically to an OD 600 of ∼0.02.

Mutant construction and complementation.
In-frame deletion strains derived from E. coli MG1655 and S. oneidensis MR-1 were constructed by the Red recombination deletion method and the att-based Fusion PCR method, respectively 52,53 . In the case of S. oneidensis, two fragments flanking the target gene were amplified by PCR with outside primer containing attB and the gene specific sequence and inside primer containing the linker and the gene specific sequence, which were joined by the second round of PCR with two outside primers. The fusion fragments were introduced into plasmid pHGM01 by using Gateway BP clonase II enzyme mix (Invitrogen) according to the manufacturer's instruction, resulting in mutagenesis vectors, which were maintained in E. coli DAP auxotroph WM3064. The vectors were subsequently transferred into relevant S. oneidensis strains via conjugation. Integration of the mutagenesis constructs into the chromosome was selected by resistance to gentamycin and confirmed by PCR. Verified transconjugants were grown in LB broth in the absence of NaCl and plated on LB containing 10% sucrose. Gentamycin-sensitive and sucrose-resistant colonies were screened by PCR for deletion of the target gene. Mutants were verified by sequencing the mutated regions.
Growth of S. oneidensis mutation strains generated in this study was measured by recording the optical density at 600 nm (OD 600 ) values in triplicate with the wild-type as the control in MS-based media as mentioned above 51 . For genetic complementation of the mutants and inducible gene expression, genes of interest generated by PCR were placed under the control of Isopropyl β-D-1-thiogalactoside (IPTG)-inducible promoter Ptac within pHGE-Ptac 32 . After verification by sequencing, the resultant vectors in E. coli DAP auxotroph WM3064 were transferred into the relevant strains via conjugation. peptidase activity assays. Fluorescence peptidase assays were performed using L-Leu-AMC substrate 21 .
Reaction mixtures contained 20 mM Tris-Cl (pH 7.5), 10 μM L-Leu-AMC, and 150 μl of extracts of cells grown in MS-L-T to the mid-log phase. The peptidase activity was monitored at 460 nm with a Synergy 2 Pro200 Multi-Detection (Microplate Reader (Tecan) by detection of 7-amino-4-methylcoumarin (MCA) released from the carboxyl terminus of N-terminally blocked peptides. Negative controls (no enzyme) were run in triplicate to account for thermal degradation of substrates as the background. One unit of protease activity was defined as the amount of enzyme required to release 1 μmol of MCA per min per mg protein. Protein concentration was determined with a bicinchoninic acid assay kit with bovine serum albumin as a standard according to the manufacturer's instructions (Pierce Chemical).
Droplet assays. Droplet assays were employed to evaluate growth inhibition on plates 54 . Cells of the mid-log phase (were collected by centrifugation and adjusted to 10 9 cfu/ml (colony forming unit), which was set as the undiluted (dilution factor 0). Ten-fold series dilutions were prepared with fresh medium indicated. Five microliters of each dilution was dropped onto LB or MS-L-T plates containing required agents as indicated in the figure. The plates were incubated for 24 h or longer before being read. All experiments were conducted at least three times. www.nature.com/scientificreports www.nature.com/scientificreports/ 100 mM Tris/HCl (pH 7.0), 10 mM MgCl 2 , 125 mM KCl, 50 mM dilithium carbamoyl phosphate for 60 minutes at room temperature. The probes used for EMSA were prepared by PCR with 33 P end-labeled primers 13 . The binding reaction was performed with ∼25-50 fmol (∼2-5 nM) labeled probes and various amounts of protein in 12 µl binding buffer containing 100 mM Tris/HCl (pH 7.4), 20 mM KCl, 10 mM MgCl 2 , 2 mM DTT, 0.2 μg/μl poly(dI·dC), and 10% glycerol at 15 °C for 60 minutes and resolved on pre-run 4.8% polyacrylamide native gels. Band shifts were visualized by autoradiography.
Bacterial one-hybrid (B1H) assay. B1H system was used to investigate DNA-protein interaction in vivo in E. coli cells 38 . Briefly, plasmid constructs were created by cloning the bait 'DNA' and target 'the SoarcA gene' in the pBX-cmT and pTRG vectors and verified by sequencing. The resultant plasmids were used to co-transform BacterioMatch II Validation Reporter Competent Cells on M9 salt agar plates containing 25 mg/ml chloramphenicol and 12.5 mg/ ml tetracycline with or without 3-amino-1,2,4-triazole (3-AT). DNA fragments of ∼300 bp for SO_1661 and the 16s rRNA gene promoters (P SO1661 and P 16s respectively) were used as positive and negative controls. The plates were incubated for 24 h and then moved to room temperature for an additional 16 h (the colonies indicating positive interaction usually appeared between 18 and 24 h). For the assay under anaerobic conditions, plates were incubated in an anaerobic glovebox (Coy Manufacturing Co.) until the colonies were fully developed. The positive interactions were confirmed by streaking colonies on plates containing both 3-AT and streptomycin (12.5 mg/ml). other analyses. Homologues of proteins of interest were identified via a BLASTp search of the NCBI's nonredundant protein database, using the amino acid sequence as the query. Experimental values were subjected to statistical analyses and presented as means ± standard error of the mean (SEM). Student's t-test was performed for pairwise comparisons of groups.