Research Article
15 August 2016

Regulation of Gene Expression in Shewanella oneidensis MR-1 during Electron Acceptor Limitation and Bacterial Nanowire Formation


In limiting oxygen as an electron acceptor, the dissimilatory metal-reducing bacterium Shewanella oneidensis MR-1 rapidly forms nanowires, extensions of its outer membrane containing the cytochromes MtrC and OmcA needed for extracellular electron transfer. RNA sequencing (RNA-Seq) analysis was employed to determine differential gene expression over time from triplicate chemostat cultures that were limited for oxygen. We identified 465 genes with decreased expression and 677 genes with increased expression. The coordinated increased expression of heme biosynthesis, cytochrome maturation, and transport pathways indicates that S. oneidensis MR-1 increases cytochrome production, including the transcription of genes encoding MtrA, MtrC, and OmcA, and transports these decaheme cytochromes across the cytoplasmic membrane during electron acceptor limitation and nanowire formation. In contrast, the expression of the mtrA and mtrC homologs mtrF and mtrD either remains unaffected or decreases under these conditions. The ompW gene, encoding a small outer membrane porin, has 40-fold higher expression during oxygen limitation, and it is proposed that OmpW plays a role in cation transport to maintain electrical neutrality during electron transfer. The genes encoding the anaerobic respiration regulator cyclic AMP receptor protein (CRP) and the extracytoplasmic function sigma factor RpoE are among the transcription factor genes with increased expression. RpoE might function by signaling the initial response to oxygen limitation. Our results show that RpoE activates transcription from promoters upstream of mtrC and omcA. The transcriptome and mutant analyses of S. oneidensis MR-1 nanowire production are consistent with independent regulatory mechanisms for extending the outer membrane into tubular structures and for ensuring the electron transfer function of the nanowires.
IMPORTANCE Shewanella oneidensis MR-1 has the capacity to transfer electrons to its external surface using extensions of the outer membrane called bacterial nanowires. These bacterial nanowires link the cell's respiratory chain to external surfaces, including oxidized metals important in bioremediation, and explain why S. oneidensis can be utilized as a component of microbial fuel cells, a form of renewable energy. In this work, we use differential gene expression analysis to focus on which genes function to produce the nanowires and promote extracellular electron transfer during oxygen limitation. Among the genes that are expressed at high levels are those encoding cytochrome proteins necessary for electron transfer. Shewanella coordinates the increased expression of regulators, metabolic pathways, and transport pathways to ensure that cytochromes efficiently transfer electrons along the nanowires.


Shewanella oneidensis encodes an array of enzymes that allow it to use a diverse set of electron donors and acceptors that range from oxygen, dimethyl sulfoxide (DMSO), and nitrate to insoluble acceptors, such as Fe(III) oxide or Mn(IV) oxide. Reduction of insoluble acceptors occurs through a series of electron transfer proteins and molecules that span the inner membrane, periplasm, and outer membrane and transfer electrons from the quinone pool to the cell exterior. Multiple mechanisms for extracellular electron transfer (EET) have been studied in S. oneidensis, including direct contact, secretion of soluble electron shuttles, such as flavins, and the formation of structures termed bacterial nanowires that transfer electrons micrometers away from the cell body (14). Each of these EET mechanisms requires outer membrane cytochromes, such as MtrC and OmcA (4, 5).
Until recently, it was assumed that the type IV pili were important for S. oneidensis nanowire formation and function, similar to the nanowires described in Geobacter sulfurreducens (6, 7). We previously demonstrated that pili are not required for the formation of S. oneidensis MR-1 nanowires. Instead, these structures appear to be extensions of the outer membrane that contain the decaheme cytochromes MtrC and OmcA (8). Atomic force microscopy and fluorescence microscopy images suggest that S. oneidensis nanowires begin as outer membrane vesicles that fuse together to form filamentous structures (8). Extending the outer membrane provides the cell with a greater surface area poised for electron transfer once an appropriate electron acceptor is encountered.
Membrane tubes, similar in appearance to S. oneidensis MR-1 nanowires, are being discovered in many bacterial species and have diverse functions. For example, chains of vesicles in Myxococcus xanthus are important for cell-cell signaling; outer membrane exchange between cells facilitated by these structures can help manage stress at the population level (9, 10). Cryo-electron microscopy (cryo-EM) images of the M. xanthus vesicle chains show characteristics similar to those we observed for S. oneidensis nanowires using atomic force microscopy and fluorescence microscopy (8, 11, 12). Recently, tube-like membrane connections have been identified between Desulfovibrio vulgaris and Clostridium acetobutylicum, as well as between Escherichia coli and Acinetobacter baylyi, for the purpose of exchanging cellular materials and cross-feeding between each pair of species (13, 14). As membrane extensions and chains of vesicles are observed in increasing numbers of bacterial species, it is becoming clear that these structures are a common mechanism employed by bacteria to interact with each other and their environment.
Such immense rearrangement of cell structure likely requires coordinated responses at both the transcriptional and posttranscriptional levels. In this work, we investigated the changes in gene expression that occur during electron acceptor (O2) limitation and nanowire formation. These two events are temporally linked; previous work has shown that when S. oneidensis is limited for electron acceptors, nanowire structures form (1, 8). By the time oxygen levels become undetectable in the chemostat, S. oneidensis MR-1 has already mounted a significant transcriptional response, increasing the transcript abundance of genes important for heme production and cytochrome c maturation and localization. Many genes that are part of central metabolism had increased expression, suggesting that altering energy metabolism is an essential part of the S. oneidensis MR-1 response during the time of oxygen limitation and nanowire formation. We identified regulatory factors that contribute to changes in gene expression, such as the cyclic AMP receptor protein (CRP) and the extracytoplasmic function sigma factor RpoE. The rapid transcriptional response to alter energy metabolism and produce nanowires suggests that the cells have regulatory cascades poised to respond when electron acceptor-limiting conditions are encountered. Our transcriptome results and mutant analyses are consistent with independent pathways for extending the outer membrane to form filamentous structures and altering energy metabolism in the cell to ensure the extracellular electron transfer capability of the nanowires.


Bacterial growth.

A complete list of strains used in this study can be found in Table 1. Shewanella oneidensis MR-1 and its derivatives were grown in Luria-Bertani (LB) broth with the appropriate antibiotics. Escherichia coli strains were grown in LB broth at 30°C or 37°C with the appropriate antibiotics as shaking cultures. The antibiotic concentrations used were kanamycin at 50 μg/ml, spectinomycin at 50 μg/ml, chloramphenicol at 20 μg/ml, tetracycline at 10 μg/ml, and gentamicin at 10 μg/ml when required; the inducer arabinose was used at 1 mM, when required.
TABLE 1 Strains and plasmids used in this work
Strain or plasmidGenotype or descriptionResistanceaSource or reference
Strains (Shewanella oneidensis)   
    MR-1Wild type 65
    ompW::pDS3.1 mutantMR-1 ΔompW 34
    SR694MR-1 Δcrp 50
    SR672MR-1 ΔcyaC 50
    JG0001E. coli JM109 Thermo Fisher Scientific
    SEA001E. coli MG1655 λrpoHP3::lacZ ΔlacX74 66
    JG0467MR-1 pRpoESpecrThis work
    pBAD43Arabinose-inducible promoterSpecr67
    pRpoErpoESO in pBAD43SpecrThis work
    pMP220Promoterless lacZTetr68
    pRpoHpromrpoH promoter region in pMP220TetrThis work
    pOmcApromomcA promoter region in pMP220TetrThis work
    pMtrCprommtrC promoter region in pMP220TetrThis work
    pScyApromscyA promoter region in pMP220TetrThis work
    pDmsEpromdmsE promoter region in pMP220TetrThis work
    pCcmFpromccmF promoter region in pMP220TetrThis work
Specr, spectinomycin resistant; Tetr, tetracycline resistant.

Strain construction.

All plasmids used in this study are listed in Table 1, and oligonucleotide sequences are listed in Table S5 in the supplemental material. The E. coli two-plasmid system was designed to test promoter-lacZ fusions for RpoESO regulation. The promoter regions were first amplified from S. oneidensis MR-1 genomic DNA by PCR using the indicated primer pairs in Table S5. After cleavage by the appropriate restriction endonucleases, the PCR fragments were ligated upstream of the promoterless lacZ gene (β-galactosidase) in the plasmid pMP220. To construct the rpoH promoter-lacZ fusion, two complementary single-stranded oligonucleotides (SORpoHPF and SORpoHPR) were annealed and directly ligated into the EcoRI and PstI sites of pMP220, creating pRpoHprom. GenScript (Piscataway, NJ) synthesized the S. oneidensis MR-1 mtrC promoter region (from bp −1119 to bp −1 relative to the translation start site), and this was used as a template for PCR amplification using the primers MtrCPF and MtrCPR. The resulting fragment was digested with KpnI, ligated into pMP220, and the orientation confirmed by sequencing. Each of these plasmids was transformed into JM109 harboring either pRpoE or pBAD43. To construct pRpoE, rpoE was PCR amplified from MR-1 genomic DNA with the primers RpoEF and RpoER. After cleavage with the appropriate restriction endonucleases, the resulting fragment was ligated into plasmid pBAD43 and transformed into JM109 or SEA001 for overexpression.

Chemostat growth conditions.

S. oneidensis MR-1 was grown in biological triplicate cultures in continuous flow reactors (BioFlo 110; New Brunswick Scientific) in 1 liter of a chemically defined medium, as previously described (8). The medium contained 19.8 mM sodium lactate, 9.9 mM ammonium chloride, 3 mM PIPES [piperazine-N,N′-bis(ethanesulfonic acid)] buffer (pH 7.2), 1.34 mM KCl, 4.4 mM KH2PO4, 7.5 mM NaOH, 1.34 mM Na2SO4, 1 mM MgCl2, 1 mM CaCl2, 0.05 mM ferric nitrilotriacetic acid, and supplements of trace minerals (Wolfe's mineral solution), vitamins (Wolfe's vitamin solution), and 2 mg/ml each of l-glutamic acid, l-arginine, and dl-serine (1, 2). Throughout the experiment, agitation was maintained at 200 rpm to minimize mechanical shear forces, pH was continuously monitored and maintained at 7.0 by using 1 M HCl, and the temperature was kept constant at 30°C.
For each of the three independent cultures, growth was initiated in batch mode by injecting 5 ml of an overnight LB culture of S. oneidensis MR-1 into the bioreactor containing 1 liter of the medium described above. A 20% dissolved oxygen tension (DOT) level was maintained in automatic DOT control mode for 20 h until the culture reached an optical density at 600 nm (OD600) of 0.75. At this point, the cells had exhausted the supply of nutrients (electron donors), and the input rate of air that maintained DOT at 20% had reached a minimum, indicating the culture had reached stationary phase. The bioreactor was then switched to continuous mode, with a dilution rate of 0.05 h−1, while the DOT was still maintained at 20%. After 48 h in continuous mode, a reference sample of cells was harvested for RNA isolation (control). The automatic DOT control was then turned off, and the DOT was reduced from 20% to effectively 0% (within 3 to 5 min) by manually controlling the air input rate. Once the DOT was no longer measurable by a polarographic O2 electrode, the first sample was taken (t = 0). However, it should be noted that this was not an anaerobic state; air was still being provided to the reactor but with the DOT just below detection of the polarographic O2 electrode, so as to ensure electron acceptor limitation while still supporting a cell density of 7.7 × 108 cells/ml (1, 2). The detection limit of the oxygen sensor (InPro 6800; Mettler Toledo) is 0.08% DOT or 6 ppb. Samples were subsequently taken from each of the three independent cultures at 15-min intervals for an hour.

RNA isolation and sequencing.

A total of 10 ml of cells was removed from the bioreactor at each time interval and mixed with ice-cold 5% citrate-saturated phenol in ethanol to prevent further transcription. Samples were taken from three independent biological replicates. Total RNA was prepared using a hot phenol extraction, as previously described (15). The samples were treated with DNase I (Life Technologies) for an hour at 37°C. The resulting samples were quantified, and quality was checked using a formaldehyde-agarose gel. RNA samples were treated with Ribo-Zero (Epicentre) to remove rRNA and then used for library construction and sequencing at the University of Southern California Epigenome Center, Los Angeles, CA. The libraries were constructed and barcoded using the Illumina library preparation kit, according to the manufacturer's instructions, and 50-bp single-end reads were obtained for each of the 18 samples.

MR-1 pRpoE perfusion tests.

S. oneidensis pRpoE (JG0467) or wild-type MR-1 was grown as preculture in LB broth. During early exponential-growth phase, 50 μg/ml streptomycin was added to the cell culture. Once the culture reached an OD600 of 2.4 to 2.8, the cells were pelleted by centrifugation, washed twice, suspended in a defined minimal medium, and examined using the perfusion imaging platform, as previously described (8). S. oneidensis cultures reach a higher optical density in LB-rich medium than in the chemically defined medium described for the chemostat cultures.

Fluorescence microscopy.

The perfusion flow imaging system was set up as previously described (8), limiting oxygen in order to induce nanowire formation. To visualize the nanowires/vesicles by fluorescence microscopy, the membrane stain FM 4-64FX (Life Technologies) (Fig. 2A) was used. For each experimental run, we added 25 μg of FM 4-64FX (dissolved in 200 μl of deionized water) to the autoclaved serum bottle containing the 100 ml of perfusion medium.

Quantitative RT-PCR.

For quantitative reverse transcription-PCRs (RT-PCRs), up to 5 μg of each of the same RNA samples used for RNA sequencing (RNA-Seq) analysis was reverse transcribed using SuperScript III (Life Technologies). Dilutions of the subsequent cDNA were added to SYBR Select master mix (Life Technologies), with each primer added to a final concentration of 2 μM. Reactions were performed using an ABI 7300 at the Penn State Genomics Core Facility, University Park, PA, and the resulting data were analyzed via the 2−ΔΔCT method, using recA for normalization. rpoB was also used for normalization, with similar results. The sequences of all primers used for this analysis can be found in Table S5 in the supplemental material.

Bioinformatic analysis of RNA-Seq data.

Raw Illumina reads (mean, 34,391,001 reads) obtained from each of the three independent sets of samples were trimmed and filtered with Trimmomatic (16) to remove adapters with up to 2 mismatches, with a palindrome read alignment accuracy of 40 and a sequence match accuracy of 15. This program trimmed reads further by removing leading low-quality or N bases (below quality 3), removing trailing low-quality or N bases (below quality 3), and scanning each read with a 4-base-wide sliding window. The parameters were set for cutting when the average quality per base drops below 15 and then to drop any reads below 36 bases long. On average, 99.71% of the reads (mean, 34,291,326 reads) were retained.
Alignment to the S. oneidensis MR-1 chromosome (GenBank accession no. NC_004347) and megaplasmid (GenBank accession no. NC_004349) was performed using Bowtie (maximum seed mismatch of 2 and seed length of 28) (17, 18). On average, 97.7% (range, 96.47 to 98.7%) of the reads aligned to the genome/megaplasmid reference. The number of trimmed reads assigned to genes was estimated with the Python script HTseq-count (19) (mean, 15,800,292 reads per sample), and differentially expressed genes across the three independent experiments were determined with the Bioconductor package DESeq2 (20) using the likelihood ratio test (LRT). P values were adjusted using the Benjamini-Hochberg (BH) correction (21), with a false-discovery rate (FDR) cutoff of 10%. Genes were considered differentially expressed if they had a BH-FDR-adjusted P value of ≤0.05.
Differentially expressed gene profiles were clustered using the Mfuzz Bioconductor package (2224), which soft clusters gene expression time-series data using fuzzy c-means. The cluster number (c = 4) was estimated using minimum centroid cluster distance and the fuzzifier (m = 1.77) parameter, using methods proposed by Schwämmle and Jensen (25). Differentially expressed genes were processed using BiNGO, an add-in for Cytoscape version 3.2.1 (26), and analyzed for Gene Ontology (GO) categories or pathways that are statistically overrepresented in our data set. The frequency of identifying each category in our data set was compared with that of the whole S. oneidensis MR-1 genome, using a significance level of 0.05 as a threshold, and using a Benjamini-Hochberg false-discovery rate correction.

Position weight matrices for CRP-binding sites used in RSAT analysis.

Putative CRP-binding sites were based on a known transcription factor position weight matrix (PWM) formulated from an alignment of 236 known CRP-binding sites in E. coli, with a consensus sequence length of 22 (

Predicting RpoESO-dependent promoters in S. oneidensis.

A total of 49 known E. coli σE-dependent promoters, as determined by Rhodius et al. (27), were aligned to generate a PWM. This matrix was entered into PromoterHunter, which is part of phiSITE (28), and used to scan the 1,100-bp upstream regions of rpoH, omcA, mtrC, scyA, dmsE, and ccmF for similar sequences. This matrix was also used to identify putative RpoE-regulated promoters upstream of the differentially expressed genes in this transcriptome analysis using Regulatory Sequence Analysis Tools (RSAT) (29).

β-Galactosidase assays.

The two-plasmid system constructs in E. coli used to test putative RpoESO-regulated promoters were grown on LB plates with the appropriate antibiotics. A total of 50 μl of a suspension from a single colony was added to tubes containing 5 ml of LB, with the appropriate antibiotics, and 1 mM arabinose was included for the overexpression of S. oneidensis rpoE. All cultures were grown in LB broth for 14 to 16 h at 30°C, with shaking. A total of 100 μl of cell suspension was added to 900 μl of Z-buffer with 2-mercaptoethanol, and β-galactosidase activity was measured (30). Because E. coli rpoE was still present in these strains, the background activity was determined by measuring β-galactosidase activity from control strains containing an empty vector without S. oneidensis rpoE and a second plasmid containing the appropriate promoter-lacZ fusion. The β-galactosidase activity from this strain was used to normalize results from strains that overexpress S. oneidensis rpoE, thereby generating relative Miller units.

Accession number(s).

Newly determined sequence data have been deposited in the Gene Expression Omnibus database under accession number GSE79964.


Differentially expressed genes during electron acceptor limitation and nanowire formation.

To examine the transcriptional changes that occur when S. oneidensis forms bacterial nanowires, we used RNA-Seq to identify differential gene expression when cells are grown in a chemostat (1, 8). The growth conditions for robust nanowire formation required lactate as the electron donor and undetectable levels of dissolved oxygen as the limited electron acceptor (see Materials and Methods). At the time points specified, samples were obtained for isolating RNA. The RNA isolated from three independently grown cultures was sequenced, the transcript abundance at each time point was compared to the reference sample (DOT, 20%), and the significantly differentially expressed genes were identified, as determined by the adjusted P values assigned to each from the likelihood ratio test (see Materials and Methods).
We have previously shown that S. oneidensis produces chains of outer membrane vesicles that fuse to form elongated structures termed bacterial nanowires (8). Vesiculation and extracellular filaments morphologically consistent with vesicular extensions become visible using scanning electron microscopy at the t = 0 point (data not shown). This correlates the time of nanowire formation with the time of electron acceptor limitation, consistent with previous studies that verified the conductance of individual nanowires (1, 2). These structures can still be observed after 30 min of electron acceptor limitation. The simultaneous occurrence of electron acceptor limitation and nanowire formation suggests that genes differentially expressed under these conditions could be important for one or both events.
From 4,555 genes contained on both the chromosome and the megaplasmid, we identified 465 genes with decreased expression and 677 genes with increased expression in one or more time points. Most of the changes in gene expression occurred within 15 min or 30 min after the DOT in the chemostat reached undetectable limits, although the expression of certain genes decreased or increased almost immediately, with significant changes detected at t = 0 (Fig. 1A). A complete table of the differentially expressed genes and their relative fold changes at each time point can be found in Table S1 in the supplemental material.
FIG 1 (A) Hierarchical clustering of differentially expressed genes (adjusted P value [Padj] ≤ 0.05). The scale is the average of the normalized count values transformed so that the mean is 0 and the standard deviation is 1. The top row of bars indicates the assigned cluster of each gene, where cluster 1 is blue, cluster 2 is green, cluster 3 is red, and cluster 4 is yellow. Control indicates the sample taken when the DOT was 20%. (B) A comparison of the fold changes determined by quantitative PCR (gray bars) for ompW, cusB, hemN, frdB, ubiE, and sirE with the fold changes determined by the RNA-Seq results (white bars). Numbers on the x axis indicate the minutes after electron acceptor limitation. Error bars denote standard error.
Quantitative RT-PCR was performed using primers for several differentially expressed genes, including cusB, ompW, hemN, frdB, ubiE, and sirE (Fig. 1B), and the resulting fold changes were compared with the RNA-Seq results. Overall, the fold changes determined by quantitative RT-PCR (qRT-PCR) agreed with the calculated fold changes in gene expression from the RNA-Seq analysis, thereby validating our approach.
Several operons exhibit similar levels of gene expression changes. For example, the genes ccmF, ccmG, ccmH, and so_0269, found in an operon for cytochrome maturation, have a similar pattern of approximately 2-fold to 4-fold increased expression throughout the experiment. In addition, extracellular electron transfer genes mtrA and mtrB, located together in an operon, have similar changes in transcript abundance. The fold changes observed for mtrA and mtrB are also consistent with those observed for the closely linked mtrC and omcA genes, encoding outer membrane cytochromes.
To determine if specific metabolic pathways were differentially expressed during the time of S. oneidensis MR-1 nanowire formation, genes were grouped according to their expression patterns using the Bioconductor package MFuzz (23, 24). Four distinct patterns emerged: cluster 1, genes with an immediate decrease in expression; cluster 2, genes with a gradual increase in expression; cluster 3, genes with a gradual decrease in expression; and cluster 4, genes with an immediate increase in expression (Fig. 2A). Because this is a soft clustering method, a number of genes are found in both clusters 2 and 4 or in both clusters 1 and 3 (Fig. 2B).
FIG 2 (A) Expression patterns for each of the four clusters identified by MFuzz (see Materials and Methods). (B) Venn diagram depicting the number of genes in common between cluster 1 (blue) and cluster 3 (red), and between cluster 2 (green) and cluster 4 (yellow).
Overrepresented Gene Ontology (GO) terms in each cluster were determined using BiNGO, a tool from the Cytoscape software platform (26, 31). BiNGO analyzes the frequency of each GO term in a data set and statistically compares it to the frequency of that GO term in the entire genome. The GO term analysis of clusters 1 and 3 (decreased differential gene expression) showed that no statistically significant overrepresented pathways were present in cluster 1. All overrepresented pathways in this group are from cluster 3 (see Table S2 in the supplemental material). Most of the pathways that are decreased in expression are related to amino acid biosynthesis and histidine metabolism (Fig. 3A). The total number of pathways overrepresented in clusters 2 and 4 (increased differential gene expression) is significantly higher than those in clusters 1 and 3 (Fig. 3A and B). In cluster 2 (genes with a gradual increase in gene expression), anaerobic respiration, copper transport, and electron carrier activity are particularly overrepresented compared to their frequency in the whole genome. Cytochrome complex assembly, heme and tetrapyrrole biosynthesis, peptidoglycan biosynthesis, quinone biosynthesis, ferrous iron transport, lactate transport, and control of gene expression are particularly overrepresented in cluster 4 (genes with an immediate increase in gene expression). Overrepresentation of genes associated with energy derivation from the oxidation of organic compounds, formate dehydrogenase, and oxidoreductase activity is common to both clusters 2 and 4 (Fig. 3B). A complete list of the GO terms overrepresented in each cluster is in Table S2.
FIG 3 Selected pathways overrepresented in the genes with decreased expression (A) or increased expression (B), shown as the ratio of the frequency that a particular category is identified in our data set to the frequency of that category in the genome as a whole.

Pathways and genes with decreased expression.

Genes for amino acid biosynthesis pathways have decreased expression during electron acceptor limitation (Fig. 3A; see also Table S1 in the supplemental material). The cultivation conditions during the time of nanowire formation must require conservation of the metabolic energy normally used for amino acid biosynthesis. In S. oneidensis, attenuation mechanisms are predicted to regulate operons for histidine, branched-chain amino acid, threonine, tryptophan, and phenylalanine biosynthesis (32). mRNA levels of the transcriptional repressor ArgR, which inhibits the expression of the arginine biosynthetic pathway genes (32), decreased 2-fold, suggesting that ArgR alone cannot be the reason for the observed decrease in expression of the arginine biosynthesis genes (Table 2). Transcript attenuation and other posttranscriptional regulatory mechanisms may contribute to some of the decreased expression of genes for amino acid biosynthesis and other pathways represented in cluster 3.
TABLE 2 Selected genes discussed in this work
ProcessGeneGene nameCluster(s)log2 FC by time (min)a
Cytochrome maturationSO_0259ccmE21.171.451.711.651.56
 SO_0269 2, 40.771.
Heme biosynthesisSO_0027hemG43.243.973.653.653.66
Quinone biosynthesisSO_1183ubiF40.781.161.010.970.93
Flavin biosynthesisSO_0142ribB21.521.621.841.972.10
 SO_3466ribE2, 40.190.410.400.370.42
Protein translocationSO_4202tatA41.572.121.871.801.86
 SO_0939 40.951.
 SO_1413 41.171.381.121.321.11
 SO_1782mtrD −0.31−0.44−0.60−0.51−0.45
Energy metabolismSO_0101fdnG40.620.510.470.470.29
 SO_0102fdnH2, 40.750.610.590.600.50
 SO_0103fdnI2, 40.590.510.620.630.55
 SO_1483aceB2, 41.901.231.951.681.30
Molybdenum cofactor biosynthesisSO_0065mogA41.021.180.930.890.80
 SO_4449moaE2, 41.071.651.651.601.60
 SO_4451moaC2, 41.001.601.591.451.55
Transcription factorsSO_0096hutC3−0.80−0.80−0.73−0.64−0.52
 SO_0393fis −0.32−0.22−0.40−0.43−0.55
 SO_1342rpoE 0.410.660.770.780.58
 SO_4780hisL1, 3−0.85−0.34−1.21−1.38−1.16
 SO_A0153 45.106.375.995.485.14
 SO_A0154 45.927.016.425.875.60
OtherSO_1673ompW2, 45.505.495.595.635.65
FC, fold change.

Pathways and genes with increased expression. (i) ompW has significantly increased expression.

One of the genes with the greatest fold change in expression was ompW, which is predicted to encode a small porin protein similar to the E. coli OmpW. The expression of S. oneidensis ompW increased approximately 40-fold during electron acceptor limitation (clusters 2 and 4; Table 2; see also Table S1 in the supplemental material). This is in agreement with a previous study showing that OmpW protein levels increase under low-oxygen conditions (33).
Although the S. oneidensis ompW mutant produces only about 60% of the wild-type current in a microbial fuel cell (34), it still produced tubular membrane extensions with the same frequency and temporal expression as the wild-type strain (Fig. 4). Additionally, the ratio of the number of produced outer membrane vesicles to filamentous nanowires and the average nanowire length in the ΔompW mutant remained similar to those in the wild type (data not shown). This suggests that the function of OmpW under electron acceptor-limited conditions is independent from the formation of filamentous membrane extensions.
FIG 4 Fluorescence microscopy visualizing nanowire formation in wild-type MR-1 and mutants lacking ompW, crp, or cyaC gene products. Arrows indicate examples of nanowire extending from individual cells. Scale bars = 5 μm.

(ii) Expression of genes involved in peptidoglycan synthesis.

Multiple bacterial species have been identified to form membrane extensions or chains of vesicles (11, 12, 14, 35). It is not known why vesicles sometimes remain attached to the cell body and become extended tubular structures. Most mechanisms described for outer membrane vesicle production include severing the connections between the peptidoglycan layer and the outer membrane (36). The S. oneidensis peptidoglycan synthesis pathway is overrepresented in the genes with increased expression during the time of nanowire formation (Fig. 3B; see also Table S2 in the supplemental material). Many of the genes involved in the early steps of peptidoglycan synthesis have around 1.5- to 2-fold increased expression. The gene encoding Lpp, the lipoprotein that connects the peptidoglycan layer to the outer membrane, however, has nearly 2-fold decreased expression (see Table S1 in the supplemental material). A recent study on E. coli proposed that an increase in peptidoglycan synthesis may lead to vesicle production because the peptidoglycan is being turned over faster than the connections to the outer membrane can be formed (37). This might result in a loose and unanchored outer membrane that more easily buds into a vesicle.

(iii) Electron transfer protein gene expression.

Many genes important for anaerobic metabolism showed increased transcript abundance as the culture was shifted from 20% to 0% DOT. The expression of genes encoding the complexes of [NiFe]-hydrogenase, formate dehydrogenase, as well as several terminal reductases, such as fumarate reductase, sulfite reductase, DMSO reductase, and nitrate reductase, is increased during the time of nanowire formation. The cytochrome-containing subunits of these reductase complexes, in particular, show the greatest increase in gene expression, such as the nearly 30-fold increase in expression of the cytochrome subunit of DMSO reductase. The expression of mtrA, mtrC, and omcA, the genes encoding the decaheme cytochromes central to extracellular electron transfer in S. oneidensis, increased, similar to our previous qRT-PCR results (8). Many other cytochromes are also differentially expressed (Table 2; see also Table S3 in the supplemental material). Interestingly, the expression of the mtrC homolog mtrF does not change during the time of nanowire production, and the expression of the mtrA homolog mtrD actually decreases. This indicates that while mtrABC expression increases, increased mtrDEF expression is not required in response to electron acceptor limitation or for forming nanowires, consistent with other reports that the expression of these genes is much lower than mtrABC when grown on Fe(III) citrate or Fe(III) oxide, and that the expression of mtrDEF instead increases during aerobic aggregation (38, 39).
Our results indicated that in addition to increased expression of mtrA, mtrC, and omcA during nanowire formation, the expression of genes important for the maturation and localization of these proteins also increased (Table 2; see also Table S3 in the supplemental material). The heme b biosynthesis pathway increased from 2- to 16-fold after electron acceptor limitation. However, the expression of ctaA and ctaB, which encode heme A and heme O synthases, respectively, decreased nearly 3-fold, suggesting that most new heme biosynthesis is being channeled into heme b production for incorporation into these newly made cytochromes. The entire cytochrome maturation (ccm) pathway, including dsbD, has between 1.5- and 4-fold increased expression; without this pathway, S. oneidensis is unable to respire (40). The twin-arginine translocation (Tat) system, encoded by tatABC, translocates mature cytochromes across the cytoplasmic membrane into the periplasm. tatABC exhibits 3- to 4-fold increased expression compared to the reference sample at every time point in our experiment (Table 2). The coordinated increased expression of these pathways clearly indicates that during electron acceptor limitation, S. oneidensis increases cytochrome production along with its capacity to properly localize these multiheme proteins near the outer membrane during the time of nanowire formation.
The final insertion of cytochromes, such as MtrC and OmcA, into the outer membrane of S. oneidensis requires the type II general secretion pathway (GSP) (41). During the time of nanowire formation, the expression level of individual type II secretion system (T2SS) genes varies; for example, gspC has less than 2-fold increased expression in our experiment, while gspA and gspB have slightly decreased expression (around 1.2-fold). S. oneidensis strains lacking GspG or GspD, essential components of the T2SS, have been shown previously to produce nanowires similarly to the wild type but are severely limited in current production and ferrihydrite reduction (1, 34, 42). Although the T2SS is important for proper insertion of cytochromes into the outer membrane, it is not likely required for forming tubular membrane extensions. This is consistent with a model in which extension of the outer membrane and the localization of electron transport proteins to these membrane extensions are regulated independently of one another.
Several of the genes encoding proteins predicted to be important for heavy metal efflux, including cusABF, copA, so_A0153, and so_A0154, had 30- to 60-fold increased expression over the course of our experiment (Table 2). The role efflux plays in nanowire formation or survival during electron acceptor limitation is not clear. However, there is some evidence that excess copper may negatively impact heme biosynthesis by targeting the coproporphyrin oxidase HemN, which has an iron-sulfur cluster sensitive to copper even under anaerobic conditions (43). Therefore, one possible reason for the higher expression of copper/silver efflux proteins could be to protect the heme biosynthesis pathway, thereby increasing the capacity for cytochrome biosynthesis and maturation.

(iv) Expression of other genes central to energy metabolism.

Under oxygen limitation with lactate as an electron donor, S. oneidensis increases the expression of genes important for lactate transport around 2-fold to 8-fold (Fig. 3; see also Table S2 in the supplemental material). The cells use a combination of fermentation and respiration for energy, and formate becomes an important redox intermediate in the cell (44). Under these conditions, pyruvate formate-lyase encoded by pflB is a key metabolic enzyme, and the glyoxylate shunt composed of isocitrate lyase (AceA) and malate synthase (AceB) is thought to provide the NADPH required for biosynthetic reactions (41). When grown with lactate and limited oxygen, S. oneidensis MR-1 fermentation and central carbohydrate metabolism are predicted to be under the global control of two regulators, the repressor PdhR and the activator HexR. The PdhR repressor and the CRP activator are thought to regulate the expression of pflB, depending on the growth conditions (32). The glyoxylate shunt encoded by the aceBA operon is predicted to be positively regulated by HexR and negatively regulated by four repressors, PdhR, TyrR, LiuR, and PsrA (32). Our transcriptome data confirm these predictions. pflB has between 2-fold and 5-fold increased expression, which reflects the elevated levels of the genes encoding its regulators; pdhR has no significant change, and crp has a 3-fold increase (Table 2; see also Table S1 in the supplemental material). Both crp and pflB are located in cluster 4 and have similar expression patterns. The expression of the aceBA operon is increased 2- to 4-fold, similar to the 2- to 3-fold increased expression of its activator's gene, hexR. The genes encoding the repressors TyrR and LiuR have a 2-fold decreased expression, while the expression of genes pdhR and psrA shows no significant change (Table 2; see also Table S1). hexR and aceBA are located in cluster 4 and demonstrate a rapid increase in expression during the time of nanowire formation. Additionally, the formate dehydrogenase complexes, hydrogenase complexes, and lactate dehydrogenase exhibit anywhere from 2-fold up to 32-fold increased expression (Table 2; see also Table S2 in the supplemental material), highlighting their importance in metabolism during electron acceptor limitation.
Molybdenum cofactors (MOCOs) are important components of various enzyme complexes, including the previously mentioned formate dehydrogenase complexes, as well as the terminal nitrate and DMSO reductases. To accommodate the increased expression of these complexes, molybdate transport (modABCD) and molybdenum cofactor biosynthesis (moaABCDE) operons also exhibit between 1.5-fold and 4-fold increased expression. A MOCO-sensitive riboswitch is predicted to help regulate the expression of these genes (32).
Genes encoding flavin and quinone biosynthesis enzymes also have 1.5-fold to 4-fold increased expression (see Table S3 in the supplemental material). Quinones are required components of the electron transport chain during respiration. During extracellular electron transfer, the periplasmic cytochrome CymA is directly reduced by the quinone pool of the electron transport chain and is then able to transfer electrons to other cytochromes, such as MtrA. Flavins are important molecules for extracellular electron transfer in S. oneidensis, either as soluble secreted electron shuttles and/or as a cofactor bound to the outer membrane cytochromes to enhance electron transfer rates (3, 5, 45). These results suggest that under the specified experimental conditions, S. oneidensis reprograms transcription to maintain high and efficient flux of electron transport components to the cell envelope.

Major regulators contribute to differential gene expression during electron acceptor limitation and nanowire formation.

How does the cell sense the changing conditions of electron acceptor limitation and transmit that information to signal the production of bacterial nanowires? It is likely that the physiological response is an energetically costly process, requiring complex regulatory changes. For example, the response leading to S. oneidensis biofilm formation depends on the coordinated expression of multiple multisubunit structures, such as pili and adhesins, as well as large shifts in metabolism (46). Similar complex levels of regulation are probably necessary to control the extension of the outer membrane in the form of nanowires.
A cis-regulatory map of the S. oneidensis genome was previously generated based on similarity to alignments of transcription factor binding motifs in E. coli (47). The transcription factors with the highest number of conserved target binding sites in S. oneidensis include the cAMP receptor protein (CRP), the nucleoid-binding protein FIS, the oxygen-responsive transcriptional regulator Fnr (EtrA), and the aerobic respiration control protein ArcA. crp and arcA have 1.5- and 3-fold increased expression, respectively, during the time of nanowire formation, while fis expression does not change, and fnr has a slight decrease in expression (Table 2; see also Table S1 in the supplemental material). A comparison of our data to transcriptome data from S. oneidensis fnr, arcA, or crp mutants (4850) shows significant overlap between the genes putatively regulated directly or indirectly by each of these factors (Fig. 5). For example, nearly half of the genes predicted to be positively regulated by Fnr (48) are also positively regulated during electron acceptor limitation in a chemostat (Fig. 5), suggesting that these factors might contribute to the regulation of nanowire formation.
FIG 5 Venn diagrams showing the genes positively (left) or negatively (right) regulated by CRP, CyaC, and/or during nanowire production and ArcA, Fnr, and/or during nanowire production.

(i) Roles of CRP and CyaC transcription regulators.

CRP is well characterized as a major regulator of changes in gene expression when S. oneidensis MR-1 cells switch to anaerobic metabolism (5, 50). In fact, a strain lacking CRP (Δcrp mutant) is not able to grow anaerobically and has lower heme content than wild-type cells (50). The differentially expressed genes in our data set were directly compared with the genes differentially expressed in S. oneidensis MR-1 Δcrp relative to the wild type (Fig. 5). Genes showing decreased expression in a Δcrp mutant require CRP for activation. Of the 254 genes with decreased expression in a Δcrp mutant, 64% have increased expression during nanowire production. This suggests that the increase in CRP production might account for a significant proportion of the altered gene expression in our analysis.
S. oneidensis MR-1 encodes three adenylate cyclases that generate cAMP, namely, CyaA, CyaB, and CyaC. All are produced to different degrees during electron acceptor limitation and the time of nanowire formation (cyaA has ∼2.5-fold decreased expression, cyaB has ∼1.5-fold increased expression, and cyaC ∼2-fold increased expression). CyaC, a membrane-bound adenylate cyclase, has been shown to play the largest role in CRP-dependent anaerobic gene expression (50). Of the 938 genes identified with decreased expression in a ΔcyaC mutant, 159 genes also have increased expression during the time of nanowire formation, including ompW, genes important for flagellar biosynthesis, the ccm locus, and many genes encoding cytochromes.
A majority of the differentially expressed genes in the Δcrp mutant strain and in our data set are regulated by CRP and during nanowire formation. Many of the genes differentially expressed in both the ΔcyaC mutant strain and during nanowire formation, however, are positively regulated in one data set but negatively regulated in the other. Although crp and cyaC exhibit a 2- to 3-fold increase in gene expression during the time of nanowire formation, only 81 genes are positively regulated by both (Fig. 5). The regulation of adenylate cyclase activity might be in direct response to oxygen limitation, and cAMP may have functions in addition to the activation of CRP (5).
Because of the importance of CRP and cAMP in anaerobic metabolism and the substantial number of genes that each might influence during electron acceptor limitation, we tested whether either factor was required for the production of nanowires. Strains lacking either crp or cyaC produce nanowires with similar abundance and temporal expression as wild-type S. oneidensis MR-1 in the perfusion flow imaging platform, indicating that these factors are not required for forming tubular membrane extensions (Fig. 4). The Δcrp and ΔcyaC mutant strains are known to have decreased heme content, likely due to the importance of these factors in regulating the expression of cytochromes, such as mtrC and omcA. S. oneidensis lacking both mtrC and omcA still produces nanowires, but they are nonconductive due to a lack of the electron transfer components localized along these membrane extensions (1, 2). Because mtrC and omcA expression is decreased more than 10-fold in a crp mutant and 3- to 4-fold in a cyaC mutant (50), strains lacking crp or cyaC still extend their membranes but may not be capable of extracellular electron transport.

(ii) Identification of putative CRP-binding sites.

The contribution of CRP to gene expression during electron acceptor limitation and nanowire formation was investigated in silico. Because the genes in each of the clusters share similar expression patterns (Fig. 2A), it is possible that some of these genes have common regulators. The upstream regions of genes from each cluster were examined for specific transcription factor binding motifs. A position weight matrix was designed using an alignment of known E. coli CRP-binding sites (Fig. 6). This matrix was used to scan the upstream regions of the genes that are differentially expressed during the time of nanowire formation. Using the pattern matching program from Regulatory Sequence Analysis Tools (RSAT []) (29), we identified putative binding sites for CRP in the 250-bp-upstream regions of the genes in each cluster. A complete list of the genes with identified CRP-binding motifs can be found in Table S4 in the supplemental material.
FIG 6 Representation of the alignment of known E. coli CRP-binding motifs (A) and E. coli σE-regulated promoters (B) from which the position weight matrices were generated. These PWM were used to scan for similar motifs upstream of the differentially expressed genes in this experiment. Sequence logos were generated at
A larger proportion of the identified putative CRP-binding motifs were located upstream of genes in clusters 2 and 4 (genes with increased expression) than in clusters 1 and 3 (genes with decreased expression). Among the genes identified to have potential CRP-binding motifs are genes with known CRP-binding sites, such as mtrC and omcA (51). Also included are genes encoding other cytochromes (e.g., so_0939, cctA, and mtrA), genes important for cytochrome maturation (e.g., sirE), genes involved in heme biosynthesis (e.g., hemH), and other electron transfer-related genes. Some transcription factors were also predicted to have CRP-binding motifs in this analysis, including the oxidative stress-responsive OxyR, the trimethylamine n-oxide (TMAO)-responsive TorR, and the zinc- and cadmium-responsive ZntR (Table 2; see also Table S4 in the supplemental material). The relatively large number of transcription factors differentially regulated during the time of nanowire formation and electron acceptor limitation suggests that the networks regulating these phenomena are complex.

(iii) Role of the extracytoplasmic function sigma factor RpoE.

We previously showed that S. oneidensis nanowires are extensions of the outer membrane of the cell and appear to form from the fusion of membrane vesicles (8). Regulatory pathways traditionally known for their response to cell envelope stress have also long been associated with the regulation of vesicle production. These pathways, such as the extracytoplasmic function sigma factor RpoE (σE) and the two-component system CpxAR, monitor for interruptions in protein flux to the outer membrane by scanning for protein accumulation in the periplasm (52, 53). In E. coli and other bacteria, members of the σE regulon, such as chaperones and proteases, ensure the efficient transport of properly folded proteins to the outer membrane. During extracytoplasmic stress, the levels of vesicle formation in E. coli and Pseudomonas aeruginosa change in response to the increased or decreased activity of either σE or specific members of its regulon, such as the protease DegP (5456). During oxygen limitation and the time of nanowire production, S. oneidensis rpoE has roughly 2-fold increased expression (Fig. 7A).
FIG 7 (A) Quantitative PCR confirms that expression of rpoE is slightly increased during nanowire formation. Error bars indicate standard error. (B) Cells overexpressing rpoE produce more vesicles than wild type after long-term electron acceptor limitation. Fluorescence microscopy images showing wild-type and MR-1 overexpressing rpoE after 5 h of incubation in the perfusion flow imaging system. Cells are visualized with the membrane stain FM 4-64FX. Arrows indicate nanowires extending from individual cells, and triangles indicate vesicles or blebs attached to individual cells. Scale bars = 5 μm.
RpoE shares 72% amino acid sequence identity in E. coli and S. oneidensis (Fig. 8A). In fact, RpoESO can activate transcription from an E. coli σE-dependent promoter. Overexpression of RpoESO increased the activity of a chromosomal E. coli σE-dependent LacZ reporter 6-fold, indicating that RpoESO recognizes and directs transcription from E. coli σE-dependent promoters (Fig. 8B). Because the two sigma factors recognize similar promoter sequences, we used a bioinformatics approach to identify candidate RpoESO-regulated promoters in S. oneidensis. Known E. coli σE-dependent promoters (27) were aligned to generate a position weight matrix. Through the PromoterHunter tool in phiSITE (28), we used this position weight matrix to scan the upstream regions of specific S. oneidensis genes for putative RpoESO-dependent promoters (Fig. 8C). To confirm that RpoESO directs transcription from the promoter region of these genes, a two-plasmid in vivo reporter assay was constructed with E. coli. Transcriptional fusions were generated between the lacZ reporter gene and the upstream DNA regions of the genes encoding the cytochromes OmcA, MtrC, DmsE, and ScyA, the cytochrome maturation protein CcmF, and the heat shock sigma factor RpoH. S. oneidensis RpoE was expressed from an arabinose-inducible promoter on a separate plasmid. Overexpression of the gene encoding RpoESO resulted in increased β-galactosidase activity from the fusions of the promoters for rpoH, omcA, mtrC, ccmF, and, to a lesser extent, scyA (Fig. 8D).
FIG 8 (A) Amino acid sequence alignment of S. oneidensis RpoE (SoRpoE) and E. coli RpoE (EcRpoE). Asterisks denote identity between the two sequences; two dots denote strong conservation, and one dot denotes weak conservation. (B) RpoESO can transcribe the E. coli σE-dependent promoter rpoHP3. Overexpression of RpoESO (gray bars) has 6-fold higher β-galactosidase activity from an rpoHP3-lacZ promoter fusion than an empty vector strain (white bars). (C) Alignment of putative RpoESO-regulated promoters. The E. coli consensus sequence for σE-dependent promoters is provided at the bottom of the alignment for reference. Underlined bases indicate consensus sequences, and dashes indicate placeholders inserted due to the variable lengths of the spacer regions. (D) Overexpression of rpoESO (gray bars) increases β-galactosidase activity from the indicated promoter-lacZ fusion compared to a strain not expressing rpoESO (white bars). For all β-galactosidase assays, the values are an average of the results from at least 3 independent cultures, and the error bars indicate standard error.
Additional putative RpoESO-regulated promoters were identified upstream of the differentially expressed genes in our analysis using the pattern matching RSAT program (29), as previously used to identify possible CRP-regulated genes. The same E. coli position weight matrix employed in the targeted approach was used for this expanded RSAT search for possible RpoESO-regulated promoters (27) (Fig. 6B). Similarly to CRP, more RpoESO-regulated promoters were predicted upstream of genes from the clusters with increased expression (clusters 2 and 4) than from the clusters with decreased expression (clusters 1 and 3). RSAT analysis confirmed the putative RpoESO promoters upstream of mtrC and ccmF, which we determined experimentally to be transcribed by RpoESO. Our analysis identified promoters upstream of genes previously described to have RpoESO-regulated promoters: surA, fkpA, and so_0516 (57). In addition, RpoESO-regulated promoters were predicted upstream of genes important for cytochrome maturation (ccmDEG and dsbD), heme biosynthesis (hemE), copper efflux (cusB), molybdenum cofactor biosynthesis (mogA), and gene regulation (arcA, cyaC, tyrR, and pspC). This suggests that RpoESO plays a role in altering gene expression during the time of nanowire formation. A complete list of the RpoESO-regulated promoters predicted by RSAT can be found in Table S4 in the supplemental material.
S. oneidensis MR-1 overexpressing rpoE (pRpoE) produces nanowires with a frequency similar to that of the wild type. After 5 hours of electron acceptor limitation in the perfusion flow imaging platform, MR-1 pRpoE produces an extra burst of vesicles that does not occur in the wild type (Fig. 7B). E. coli with altered RpoE activity has been reported to produce membrane blebs (58). An S. oneidensis rpoE mutant has been grown and maintained in LB medium but does not grow in minimal medium (57). Therefore, MR-1 lacking rpoE cannot be grown under the chemostat conditions required to specifically trigger nanowire formation by limiting available electron acceptors. As a result, we cannot elucidate the total contributions of RpoE regulation to membrane vesicle and nanowire formation in S. oneidensis MR-1.


In this work, we have demonstrated that Shewanella oneidensis MR-1 reprograms gene expression during electron acceptor limitation and nanowire formation to promote efficient flux of the extracellular electron transport machinery to the cell envelope. Because nanowire formation occurs as electron acceptors become limited, the genes we identified as differentially expressed under these conditions represent transcriptional regulatory changes that either respond to oxygen limitation or promote nanowire formation, or a combination of both. At least two major events must occur to build a functional nanowire: (i) extension of the outer membrane into a tubular structure and (ii) localization of electron transfer proteins along the membrane extensions. Our work demonstrates a clear transcriptional response that simultaneously increases the expression of genes encoding cytochromes, heme biosynthesis enzymes, cytochrome maturation proteins, translocation machinery, copper efflux, molybdenum cofactor biosynthesis, and quinone and flavin biosynthesis. This alone emphasizes the importance of transcriptional regulation of electron transfer components during electron acceptor limitation and nanowire formation.
The ompW gene displays a 40-fold increase in expression as an immediate response to electron acceptor limitation. Since the level of OmpW protein is increased under oxygen electron acceptor limitation (33), one hypothesis is that it may facilitate the extracellular transport of cations that might occur in concert with extracellular electron transport as a means of maintaining electrical neutrality. The recently characterized OmpW of Caulobacter crescentus has been shown to transport cations (59). This might explain the decreased current output in microbial fuel cells in the S. oneidensis mutant that lacks OmpW (34).
S. oneidensis mutants lacking mtrC and omcA produce nanowire structures with the same abundance and length as the wild type but are nonconductive (2). These observations suggest that the production of tubular membrane extensions is regulated independently from functionalizing the nanowires with electron transport machinery. Separate regulatory mechanisms for the formation of membrane extensions and their function exist in pathogenic bacteria that produce membrane sheaths around their flagella to reduce their antigenicity. Mutants of Brucella melitensis that no longer produce flagella still generate the membrane sheaths, even though they are hollow (60). In a similar fashion, S. oneidensis still extends its membrane as a response to electron acceptor limitation, even if it lacks cytochromes and cannot efficiently transfer electrons to the outside of the cell.
Other groups have compared the expression of S. oneidensis cytochromes in the presence or absence of oxygen (61). The expression patterns of scyA, fccA, ccpA, mtrA, mtrC, and omcA in our study are similar to those in this other work. Different expression patterns, however, were observed in our results for other cytochrome-encoding genes, including sirA, napB, nrfA, cymA, so_4047, and so_4048. These differences in expression may be due to different growth conditions used in these studies. Gene expression from cultures grown in rich LB medium using shake flask batch cultures cannot be easily compared with expression from cultures grown in minimal medium using a chemostat (33), in which cultivation conditions are tightly regulated. Because of these highly controlled conditions, the culture in this study is not strictly anaerobic, and the limited availability of soluble oxygen as an electron acceptor is the only trigger for nanowire formation.
CRP and CyaC are not required for forming vesicles and extending the outer membrane during nanowire formation (Fig. 4). Due to the loss of hemes in cells lacking CRP, it is unlikely that these membrane extensions transfer electrons. CRP was previously shown to be required for the transcriptional activation of mtrC and omcA, although the promoters did not share identical regulation in the presence and absence of oxygen (62). This indicates that regulatory factors other than CRP may contribute to the regulation of mtrC and omcA under these specific growth conditions. We identified RpoESO-regulated promoters upstream of both mtrC and omcA, indicating that at least one additional factor can contribute to the regulation of these cytochromes. Because many uncharacterized transcription factors have increased expression levels, it is likely that additional regulation occurs at both the transcriptional and posttranscriptional levels.
S. oneidensis RpoE (RpoESO) is important for growth at high and low temperatures, responding to high salt and oxidative stress, and for resistance to the β-lactam antibiotic ampicillin during batch culture growth in rich medium. As in E. coli, RpoESO-regulated promoters have been predicted upstream of many genes important for cell envelope maintenance and stress response in S. oneidensis (57). This suggests that due to its conserved role at the cell envelope, the RpoESO regulon might be an important part of the response that leads to bacterial nanowire formation. Most regulation of σE activity is known to be posttranscriptional in E. coli, including activation via regulated proteolysis of its anti-sigma factor RseA, or both directly and indirectly via guanosine pentaphosphate [(p)ppGpp] (63, 64), neither of which would be observable in our analysis of transcript abundance. Posttranscriptional activation of RpoE would allow for a rapid initial response to conditions, such as electron acceptor limitation; even a slight increase in rpoE gene expression may be significant for regulation during the time of nanowire formation.
All Gram-negative bacteria examined to date produce outer membrane vesicles. σE, encoded by rpoE, is associated with outer membrane vesicle formation in many bacterial species, including E. coli, Salmonella enterica, and P. aeruginosa (36, 56). rpoE expression increases 1.5- to 2-fold during electron acceptor limitation in S. oneidensis MR-1. Overexpression of this sigma factor results in a late burst of vesicle formation or membrane blebbing, but initial nanowire formation remains unaffected (Fig. 7B). The identification of RpoESO-dependent transcription of mtrC, omcA, ccmF, and scyA suggests that RpoESO plays a role in regulating the expression of extracellular electron transfer genes under certain conditions.
Our transcriptome analysis shows that the central carbon metabolism regulatory genes (hexR, pdhR, tyrR, liuR, and psrA) are differentially expressed as predicted when oxygen is limiting. We have also shown that several known regulators, including CRP and RpoE, must play key roles. The transcription regulatory network of S. oneidensis MR-1 nanowire formation during electron acceptor limitation is complex and must be poised to quickly sense and respond to changing conditions. According to the comparative analysis of 13 Shewanella species genomes, S. oneidensis MR-1 has a transcription regulatory network that includes 60 transcription factors and approximately 400 transcription factor binding sites (5). Further analysis of the S. oneidensis MR-1 genome has 211 predicted one-component systems and 47 predicted two-component systems involved in environmental sensing and signal transduction (5). The task of identifying the primary components responsible for sensing oxygen limitation and regulating nanowire formation will likely require the development of a method for high-throughput evaluation of many S. oneidensis MR-1 mutants. Our study establishes that as soon as oxygen limitation is reached in a chemostat, a rapid and significant transcriptional response occurs in S. oneidensis, resulting in the alteration of energy metabolism and formation of electrically conductive nanowires. The transcription regulatory network involved in this response and the physiological impact of the production of nanowires will require further experimental work.


We thank Margie Romine for providing the ompW mutant. We also thank Mike Gorka, Karla N. Piedl, and especially Allen T. Phillips and Daâd Saffarini for their critical reading of the manuscript.

Supplemental Material

File (zam999117387sd1.xlsx)
File (zam999117387sd2.xlsx)
File (zam999117387sd3.xlsx)
File (zam999117387sd4.xlsx)
File (zam999117387sd5.xlsx)
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.


Gorby YA, Yanina S, McLean JS, Rosso KM, Moyles D, Dohnalkova A, Beveridge TJ, Chang IS, Kim BH, Kim KS, Culley DE, Reed SB, Romine MF, Saffarini DA, Hill EA, Shi L, Elias DA, Kennedy DW, Pinchuk G, Watanabe K, Ishii S, Logan B, Nealson KH, Fredrickson JK. 2006. Electrically conductive bacterial nanowires produced by Shewanella oneidensis strain MR-1 and other microorganisms. Proc Natl Acad Sci U S A 103:11358–11363.
El-Naggar MY, Wanger G, Leung KM, Yuzvinsky TD, Southam G, Yang J, Lau WM, Nealson KH, Gorby YA. 2010. Electrical transport along bacterial nanowires from Shewanella oneidensis MR-1. Proc Natl Acad Sci U S A 107:18127–18131.
Okamoto A, Kalathil S, Deng X, Hashimoto K, Nakamura R, Nealson KH. 2014. Cell-secreted flavins bound to membrane cytochromes dictate electron transfer reactions to surfaces with diverse charge and pH. Sci Rep 4:5628.
Gralnick JA, Newman DK. 2007. Extracellular respiration. Mol Microbiol 65:1–11.
Fredrickson JK, Romine MF, Beliaev AS, Auchtung JM, Driscoll ME, Gardner TS, Nealson KH, Osterman AL, Pinchuk G, Reed JL, Rodionov DA, Rodrigues JL, Saffarini DA, Serres MH, Spormann AM, Zhulin IB, Tiedje JM. 2008. Towards environmental systems biology of Shewanella. Nat Rev Microbiol 6:592–603.
Reguera G, McCarthy KD, Mehta T, Nicoll JS, Tuominen MT, Lovley DR. 2005. Extracellular electron transfer via microbial nanowires. Nature 435:1098–1101.
Vargas M, Malvankar NS, Tremblay PL, Leang C, Smith JA, Patel P, Snoeyenbos-West O, Nevin KP, Lovley DR. 2013. Aromatic amino acids required for pili conductivity and long-range extracellular electron transport in Geobacter sulfurreducens. mBio 4(2):e00105-13.
Pirbadian S, Barchinger SE, Leung KM, Byun HS, Jangir Y, Bouhenni RA, Reed SB, Romine MF, Saffarini DA, Shi L, Gorby YA, Golbeck JH, El-Naggar MY. 2014. Shewanella oneidensis MR-1 nanowires are outer membrane and periplasmic extensions of the extracellular electron transport components. Proc Natl Acad Sci U S A 111:12883–12888.
Pathak DT, Wall D. 2012. Identification of the cglC, cglD, cglE, and cglF genes and their role in cell contact-dependent gliding motility in Myxococcus xanthus. J Bacteriol 194:1940–1949.
Vassallo C, Pathak DT, Cao P, Zuckerman DM, Hoiczyk E, Wall D. 2015. Cell rejuvenation and social behaviors promoted by LPS exchange in myxobacteria. Proc Natl Acad Sci U S A 112:E2939-2946.
Remis JP, Wei D, Gorur A, Zemla M, Haraga J, Allen S, Witkowska HE, Costerton JW, Berleman JE, Auer M. 2014. Bacterial social networks: structure and composition of Myxococcus xanthus outer membrane vesicle chains. Environ Microbiol 16:598–610.
Ducret A, Fleuchot B, Bergam P, Mignot T. 2013. Direct live imaging of cell-cell protein transfer by transient outer membrane fusion in Myxococcus xanthus. eLife 2:e00868.
Benomar S, Ranava D, Cardenas ML, Trably E, Rafrafi Y, Ducret A, Hamelin J, Lojou E, Steyer JP, Giudici-Orticoni MT. 2015. Nutritional stress induces exchange of cell material and energetic coupling between bacterial species. Nat Commun 6:6283.
Pande S, Shitut S, Freund L, Westermann M, Bertels F, Colesie C, Bischofs IB, Kost C. 2015. Metabolic cross-feeding via intercellular nanotubes among bacteria. Nat Commun 6:6238.
Rhodius VA, Mutalik VK, Gross CA. 2012. Predicting the strength of UP-elements and full-length E. coli σE promoters. Nucleic Acids Res 40:2907–2924.
Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30:2114–2120.
Langmead B. 2010. Aligning short sequencing reads with Bowtie. Curr Protoc Bioinformatics Chapter 11:Unit 11.7.
Langmead B, Trapnell C, Pop M, Salzberg SL. 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 10:R25.
Anders S, Pyl PT, Huber W. 2014. HTSeq—a Python framework to work with high-throughput sequencing data. Bioinformatics 31:166–169.
Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15:550.
Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Series B Stat Methodol 57:289–300.
Futschik ME, Carlisle B. 2005. Noise-robust soft clustering of gene expression time-course data. J Bioinformatics Comput Biol 3:965–988.
Futschik ME. 2015. Mfuzz: soft clustering of time series gene expression data. R package, version 2.30.0.
Kumar L, Futschik ME. 2007. Mfuzz: a software package for soft clustering of microarray data. Bioinformation 2:5–7.
Schwämmle V, Jensen ON. 2010. A simple and fast method to determine the parameters for fuzzy c-means cluster analysis. Bioinformatics 26:2841–2848.
Maere S, Heymans K, Kuiper M. 2005. BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448–3449.
Rhodius VA, Suh WC, Nonaka G, West J, Gross CA. 2006. Conserved and variable functions of the σE stress response in related genomes. PLoS Biol 4:e2.
Klucar L, Stano M, Hajduk M. 2010. phiSITE: database of gene regulation in bacteriophages. Nucleic Acids Res 38:D366–D370.
Medina-Rivera A, Defrance M, Sand O, Herrmann C, Castro-Mondragon JA, Delerce J, Jaeger S, Blanchet C, Vincens P, Caron C, Staines DM, Contreras-Moreira B, Artufel M, Charbonnier-Khamvongsa L, Hernandez C, Thieffry D, Thomas-Chollier M, van Helden J. 2015. RSAT 2015: Regulatory Sequence Analysis Tools. Nucleic Acids Res 43:W50–W56.
Miller JH. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T. 2003. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13:2498–2504.
Rodionov DA, Novichkov PS, Stavrovskaya ED, Rodionova IA, Li X, Kazanov MD, Ravcheev DA, Gerasimova AV, Kazakov AE, Kovaleva GY, Permina EA, Laikova ON, Overbeek R, Romine MF, Fredrickson JK, Arkin AP, Dubchak I, Osterman AL, Gelfand MS. 2011. Comparative genomic reconstruction of transcriptional networks controlling central metabolism in the Shewanella genus. BMC Genomics 12(Suppl 1):S3.
Elias DA, Tollaksen SL, Kennedy DW, Mottaz HM, Giometti CS, McLean JS, Hill EA, Pinchuk GE, Lipton MS, Fredrickson JK, Gorby YA. 2008. The influence of cultivation methods on Shewanella oneidensis physiology and proteome expression. Arch Microbiol 189:313–324.
Bretschger O, Obraztsova A, Sturm CA, Chang IS, Gorby YA, Reed SB, Culley DE, Reardon CL, Barua S, Romine MF, Zhou J, Beliaev AS, Bouhenni R, Saffarini D, Mansfeld F, Kim BH, Fredrickson JK, Nealson KH. 2007. Current production and metal oxide reduction by Shewanella oneidensis MR-1 wild type and mutants. Appl Environ Microbiol 73:7003–7012.
McCaig WD, Koller A, Thanassi DG. 2013. Production of outer membrane vesicles and outer membrane tubes by Francisella novicida. J Bacteriol 195:1120–1132.
Kulp A, Kuehn MJ. 2010. Biological functions and biogenesis of secreted bacterial outer membrane vesicles. Annu Rev Microbiol 64:163–184.
Schwechheimer C, Rodriguez DL, Kuehn MJ. 2015. NlpI-mediated modulation of outer membrane vesicle production through peptidoglycan dynamics in Escherichia coli. Microbiologyopen 4:375–389.
Reyes C, Murphy JN, Saltikov CW. 2010. Mutational and gene expression analysis of mtrDEF, omcA and mtrCAB during arsenate and iron reduction in Shewanella sp. ANA-3. Environ Microbiol 12:1878–1888.
McLean JS, Pinchuk GE, Geydebrekht OV, Bilskis CL, Zakrajsek BA, Hill EA, Saffarini DA, Romine MF, Gorby YA, Fredrickson JK, Beliaev AS. 2008. Oxygen-dependent autoaggregation in Shewanella oneidensis MR-1. Environ Microbiol 10:1861–1876.
Bouhenni R, Gehrke A, Saffarini D. 2005. Identification of genes involved in cytochrome c biogenesis in Shewanella oneidensis, using a modified mariner transposon. Appl Environ Microbiol 71:4935–4937.
Shi L, Deng S, Marshall MJ, Wang Z, Kennedy DW, Dohnalkova AC, Mottaz HM, Hill EA, Gorby YA, Beliaev AS, Richardson DJ, Zachara JM, Fredrickson JK. 2008. Direct involvement of type II secretion system in extracellular translocation of Shewanella oneidensis outer membrane cytochromes MtrC and OmcA. J Bacteriol 190:5512–5516.
Bouhenni RA, Vora GJ, Biffinger JC, Shirodkar S, Brockman K, Ray R, Wu P, Johnson BJ, Biddle EM, Marshall MJ, Fitzgerald LA, Little BJ, Fredrickson JK, Beliaev AS, Ringeisen BR, Saffarini DA. 2010. The role of Shewanella oneidensis MR-1 outer surface structures in extracellular electron transfer. Electroanalysis 22:856–864.
Azzouzi A, Steunou AS, Durand A, Khalfaoui-Hassani B, Bourbon ML, Astier C, Bollivar DW, Ouchane S. 2013. Coproporphyrin III excretion identifies the anaerobic coproporphyrinogen III oxidase HemN as a copper target in the Cu+-ATPase mutant copA of Rubrivivax gelatinosus. Mol Microbiol 88:339–351.
Pinchuk GE, Geydebrekht OV, Hill EA, Reed JL, Konopka AE, Beliaev AS, Fredrickson JK. 2011. Pyruvate and lactate metabolism by Shewanella oneidensis MR-1 under fermentation, oxygen limitation, and fumarate respiration conditions. Appl Environ Microbiol 77:8234–8240.
Kotloski NJ, Gralnick JA. 2013. Flavin electron shuttles dominate extracellular electron transfer by Shewanella oneidensis. mBio 4(1):e00553-12.
Gödeke J, Binnenkade L, Thormann KM. 2012. Transcriptome analysis of early surface-associated growth of Shewanella oneidensis MR-1. PLoS One 7:e42160.
Liu J, Xu X, Stormo GD. 2008. The cis-regulatory map of Shewanella genomes. Nucleic Acids Res 36:5376–5390.
Cruz-Garcia C, Murray AE, Rodrigues JL, Gralnick JA, McCue LA, Romine MF, Loffler FE, Tiedje JM. 2011. Fnr (EtrA) acts as a fine-tuning regulator of anaerobic metabolism in Shewanella oneidensis MR-1. BMC Microbiol 11:64.
Gao H, Wang X, Yang ZK, Palzkill T, Zhou J. 2008. Probing regulon of ArcA in Shewanella oneidensis MR-1 by integrated genomic analyses. BMC Genomics 9:42.
Charania MA, Brockman KL, Zhang Y, Banerjee A, Pinchuk GE, Fredrickson JK, Beliaev AS, Saffarini DA. 2009. Involvement of a membrane-bound class III adenylate cyclase in regulation of anaerobic respiration in Shewanella oneidensis MR-1. J Bacteriol 191:4298–4306.
Gao H, Wang X, Yang ZK, Chen J, Liang Y, Chen H, Palzkill T, Zhou J. 2010. Physiological roles of ArcA, Crp, and EtrA and their interactive control on aerobic and anaerobic respiration in Shewanella oneidensis. PLoS One 5:e15295.
Raivio TL, Silhavy TJ. 1999. The σE and Cpx regulatory pathways: overlapping but distinct envelope stress responses. Curr Opin Microbiol 2:159–165.
Connolly L, De Las Penas A, Alba BM, Gross CA. 1997. The response to extracytoplasmic stress in Escherichia coli is controlled by partially overlapping pathways. Genes Dev 11:2012–2021.
Schwechheimer C, Sullivan CJ, Kuehn MJ. 2013. Envelope control of outer membrane vesicle production in Gram-negative bacteria. Biochemistry 52:3031–3040.
Altindis E, Fu Y, Mekalanos JJ. 2014. Proteomic analysis of Vibrio cholerae outer membrane vesicles. Proc Natl Acad Sci U S A 111:E1548-E1556.
Macdonald IA, Kuehn MJ. 2013. Stress-induced outer membrane vesicle production by Pseudomonas aeruginosa. J Bacteriol 195:2971–2981.
Dai J, Wei H, Tian C, Damron FH, Zhou J, Qiu D. 2015. An extracytoplasmic function sigma factor-dependent periplasmic glutathione peroxidase is involved in oxidative stress response of Shewanella oneidensis. BMC Microbiol 15:34.
Hayden JD, Ades SE. 2008. The extracytoplasmic stress factor, σE, is required to maintain cell envelope integrity in Escherichia coli. PLoS One 3:e1573.
Benz R, Jones MD, Younas F, Maier E, Modi N, Mentele R, Lottspeich F, Kleinekathofer U, Smit J. 2015. OmpW of Caulobacter crescentus functions as an outer membrane channel for cations. PLoS One 10:e0143557.
Ferooz J, Letesson JJ. 2010. Morphological analysis of the sheathed flagellum of Brucella melitensis. BMC Res Notes 3:333.
Jin M, Jiang Y, Sun L, Yin J, Fu H, Wu G, Gao H. 2013. Unique organizational and functional features of the cytochrome c maturation system in Shewanella oneidensis. PLoS One 8:e75610.
Kasai T, Kouzuma A, Nojiri H, Watanabe K. 2015. Transcriptional mechanisms for differential expression of outer membrane cytochrome genes omcA and mtrC in Shewanella oneidensis MR-1. BMC Microbiol 15:68.
Grigorova IL, Chaba R, Zhong HJ, Alba BM, Rhodius V, Herman C, Gross CA. 2004. Fine-tuning of the Escherichia coli σE envelope stress response relies on multiple mechanisms to inhibit signal-independent proteolysis of the transmembrane anti-sigma factor, RseA. Genes Dev 18:2686–2697.
Costanzo A, Nicoloff H, Barchinger SE, Banta AB, Gourse RL, Ades SE. 2008. ppGpp and DksA likely regulate the activity of the extracytoplasmic stress factor σE in Escherichia coli by both direct and indirect mechanisms. Mol Microbiol 67:619–632.
Myers CR, Nealson KH. 1988. Bacterial manganese reduction and growth with manganese oxide as the sole electron acceptor. Science 240:1319–1321.
Costanzo A, Ades SE. 2006. Growth phase-dependent regulation of the extracytoplasmic stress factor, σE, by guanosine 3′,5′-bispyrophosphate (ppGpp). J Bacteriol 188:4627–4634.
Guzman LM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177:4121–4130.
Spaink HP, Okker RJ, Wijffelman CA, Pees E, Lugtenberg BJ. 1987. Promoters in the nodulation region of the Rhizobium leguminosarum Sym plasmid pRL1JI. Plant Mol Biol 9:27–39.

Information & Contributors


Published In

cover image Applied and Environmental Microbiology
Applied and Environmental Microbiology
Volume 82Number 171 September 2016
Pages: 5428 - 5443
Editor: A. M. Spormann, Stanford University
PubMed: 27342561


Received: 1 June 2016
Accepted: 22 June 2016
Published online: 15 August 2016


Request permissions for this article.



Sarah E. Barchinger
Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania, USA
Sahand Pirbadian
Department of Physics and Astronomy, University of Southern California, Los Angeles, California, USA
School of Biosciences, University of Exeter, Exeter, United Kingdom
Carol S. Baker
Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania, USA
Kar Man Leung
Department of Physics and Astronomy, University of Southern California, Los Angeles, California, USA
Nigel J. Burroughs
Warwick Systems Biology Centre and Mathematics Institute, University of Warwick, Coventry, United Kingdom
Mohamed Y. El-Naggar
Department of Physics and Astronomy, University of Southern California, Los Angeles, California, USA
Molecular and Computational Biology Section, Department of Biological Sciences and Department of Chemistry, University of Southern California, Los Angeles, California, USA
John H. Golbeck
Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania, USA
Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania, USA


A. M. Spormann
Stanford University


Address correspondence to John H. Golbeck, [email protected].

Metrics & Citations



  • For recently published articles, the TOTAL download count will appear as zero until a new month starts.
  • There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
  • Citation counts come from the Crossref Cited by service.


If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.

View Options

Figures and Media






Share the article link

Share with email

Email a colleague

Share on social media

American Society for Microbiology ("ASM") is committed to maintaining your confidence and trust with respect to the information we collect from you on websites owned and operated by ASM ("ASM Web Sites") and other sources. This Privacy Policy sets forth the information we collect about you, how we use this information and the choices you have about how we use such information.
FIND OUT MORE about the privacy policy