Genomic Plasticity Enables a Secondary Electron Transport Pathway in Shewanella oneidensis

Abstract

Microbial dissimilatory iron reduction is an important biogeochemical process. It is physiologically challenging because iron occurs in soils and sediments in the form of insoluble minerals such as hematite or ferrihydrite. Shewanella oneidensis MR-1 evolved an extended respiratory chain to the cell surface to reduce iron minerals. Interestingly, the organism evolved a similar strategy for reduction of dimethyl sulfoxide (DMSO), which is reduced at the cell surface as well. It has already been established that electron transfer through the outer membrane is accomplished via a complex in which β-barrel proteins enable interprotein electron transfer between periplasmic oxidoreductases and cell surface-localized terminal reductases. MtrB is the β-barrel protein that is necessary for dissimilatory iron reduction. It forms a complex together with the periplasmic decaheme c-type cytochrome MtrA and the outer membrane decaheme c-type cytochrome MtrC. Consequently, mtrB deletion mutants are unable to reduce ferric iron. The data presented here show that this inability can be overcome by a mobile genomic element with the ability to activate the expression of downstream genes and which is inserted within the SO4362 gene of the SO4362-to-SO4357 gene cluster. This cluster carries genes similar to mtrA and mtrB and encoding a putative cell surface DMSO reductase. Expression of SO4359 and SO4360 alone was sufficient to complement not only an mtrB mutant under ferric citrate-reducing conditions but also a mutant that furthermore lacks any outer membrane cytochromes. Hence, the putative complex formed by the SO4359 and SO4360 gene products is capable not only of membrane-spanning electron transfer but also of reducing extracellular electron acceptors.

INTRODUCTION

Dissimilatory iron reduction is a respiratory process in which proton gradient-dependent energy generation at the cytoplasmic membrane is coupled to the reduction of ferric iron (1–4). Microbial catabolic iron reduction has been studied intensively since its discovery as a respiratory process in the 1980s (2, 3, 5–9). For dissimilatory iron-reducing bacteria, the physiological challenge of this form of respiration is the existence of ferric iron at neutral pH primarily as crystalline iron minerals (10). Hence, an electron acceptor that cannot diffuse through the membranes of Gram-negative cells has to be reduced (11). As an answer to this physiological challenge, microbes have evolved an extended respiratory chain from the cytoplasmic membrane through the periplasm and across the outer membrane to transfer respiratory electrons to the iron mineral (for recent reviews, see references 4 and 12 to 16). The development of a protein complex enabling outer membrane-spanning electron transfer was most probably a key event in the evolutionary process resulting in modern mineral-respiring organisms. In Shewanella oneidensis MR-1, this complex is formed by the periplasmic decaheme c-type cytochrome MtrA, the outer membrane decaheme c-type cytochrome MtrC, and the β-barrel protein MtrB (17–21). MtrB seems to bring the two cytochromes in sufficient proximity to enable interprotein electron transfer (17, 18). The importance of this complex for respiration on ferric iron under anoxic conditions is displayed by the severe phenotypes of strains with mutations in the corresponding genes (19, 22–28). These mutants are severely impaired or unable to use ferric iron as an electron acceptor. Thus, the phenotype is independent of the species of the non-membrane-permeating iron form (24, 28, 29). Even the colloid ferric citrate cannot be used (23, 29). Interestingly, the complex has an intrinsic assembly control. In the absence of MtrA, MtrB is not detectable and complex formation is omitted (18, 30). Moreover, the presence of MtrA seems to be crucial for periplasmic stability of MtrB, since the requirement of MtrA can be mitigated by turning off expression of the gene encoding the periplasmic protease DegP (30). Hence, under native promoter conditions, an MtrA null mutant is always an MtrB null mutant as well (18, 30). In contrast, MtrB production and formation of the MtrAB subcomplex could be observed even in the absence of MtrC (18, 30).

The chromosome of S. oneidensis MR-1 contains three additional homologs of the gene encoding MtrB (31). mtrE is part of a gene cluster that is similar to the mtrABC cluster, while dmsF and SO4359 are parts of operons that furthermore contain the genetic information for dimethyl sulfoxide (DMSO) reductases, which—in contrast to the case for other bacteria—are localized not to the periplasm but to the surface of the outer membrane (31). Nevertheless, only the cluster containing dmsF is upregulated and used under DMSO-reducing conditions, while the other shows no differential expression when aerobic and DMSO-reducing conditions are compared (31). Adjacent to all genes homologous to mtrB is always a gene similar to mtrA, which demonstrates the necessary interplay of both corresponding proteins for membrane-spanning electron transfer (see Fig. S1 in the supplemental material).

Coursolle and Gralnick studied the functionality of MtrB paralogs in a ΔmtrB strain. Only expression of mtrB itself or mtrE could rescue the mutant phenotype. These findings are consistent with the high similarity of MtrDEF to MtrABC (26). Additional complementation experiments with a strain deficient in all periplasmic MtrA homologs revealed that only mtrD and dmsE expression could partly compensate for the loss of mtrA under ferric iron-reducing conditions, whereas replacement with SO4360 did not increase reduction rates compared to those of the mutant (25). Moreover, overexpression of the SO4360 gene product could not compensate for the loss of DmsE under DMSO-reducing conditions (25).

The goal of the present study was to identify alternative or less dominant pathways for iron reduction encoded within the S. oneidensis MR-1 chromosome. Therefore, we screened for gain-of-function mutants in a ΔmtrB strain. As an answer to the above question, we observed that transcriptional activation of SO4360 and SO4359, caused by upstream insertion of a mobile genetic element, ISSod1, led to a strain with a regained ability to respire on ferric citrate. Subsequent experiments indicated that the mtrAB homologs SO4359 and SO4360 were required for the restored growth. Notably, coexpression of these genes in trans in an mtrB deletion mutant could complement for iron reduction. More importantly, the expression of SO4360 and SO4359 in trans in a strain depleted of all outer membrane c-type cytochromes and mtrA (referred to here as the ΔOMCA strain) was also sufficient to complement the mutant for growth under iron-reducing conditions. Hence, it was discovered not only that the SO4359 and SO4360 gene products are redundant to MtrB and MtrA but that these two proteins most likely also function as a terminal reductase localized at the cell surface.

MATERIALS AND METHODS

Aerobic growth of S. oneidensis MR-1 and Escherichia coli.

E. coli and S. oneidensis MR-1 strains (Tables 1 and 2) were grown aerobically as batch cultures in LB medium consisting of 1% (wt/vol) Bacto tryptone, 1% (wt/vol) NaCl, and 0.5% (wt/vol) yeast extract. All flasks were shaken continuously at 180 rpm and incubated at 37°C and 30°C, respectively. If necessary, 2,6-diaminopimelic acid (DAP) (100 μg ml⁻¹), kanamycin (Kan) (50 μg ml⁻¹), or tetracycline (Tet) (15 μg ml⁻¹) was added to the medium. Growth was determined by measuring the optical density at 600 nm (OD₆₀₀).

Table 1.

S. oneidensis MR-1 strains used in this study

Strain	Relevant genotype	Reference or source
JG 7	Wild-type strain	6
JG 55	ΔmtrB	27
JG 69	ΔmtrB pBADmtrBSTREP	27
JG 171 (ΔOMCA)	ΔmtrD-F ΔomcA ΔmtrC ΔSO_1659 ΔSO_2931 ΔmtrA Para mtrBSTREP	30
JG 240 (ΔmtrBS)	ΔmtrB suppressor strain	This study
JG 330	ΔmtrBS ΔSO4358 ΔSO4357	This study
JG 363	ΔmtrBS ΔSO4359	This study
JG 395	ΔmtrBS ΔmtrC	This study
JG 536	ΔmtrBS ΔSO4359 pBAD_SO4359STREP	This study
JG 592	ΔmtrBS ΔSO4359 ΔSO4360 pBAD_SO4359STREP	This study
JG 593	ΔmtrBS ΔSO4359 ΔmtrA pBAD_SO4359STREP	This study
JG 560	ΔmtrB pBAD_SO4360/59STREP	This study
JG 594	ΔOMCA pBAD_SO4360/59STREP	This study

Table 2.

E. coli strains used in this study

Strain	Relevant genotype	Reference or source
JG 22 (DH5αZ1)	aciq PN25-tetR Spr deoR supE44 Δ(lacZYA-argFV169) ϕ80dlacZΔM15	32
JG 98 (WM3064)	thrB1004 pro thi rpsL hsdS lacZΔM15RP4–1360 Δ(araBAD)567 ΔdapA1341::[erm pir(wt)]	W. Metcalf, University of Illinois
JG 304	JG98 pMQ_ΔSO4359	This study
JG 305	JG98 pMQ_ΔSO4357 ΔSO4358	This study
JG 306	JG98 pMQ_ΔmtrC	This study
JG 526	JG98 pSB_ΔmtrA	This study
JG 557	JG98 pSB_ΔSO4360	This study
JG 558	JG22 pBAD_SO4359STREP	This study
JG 559	JG22 pBAD_SO4360 SO4359STREP	This study
JG 555	JG22 pBAD_wt::mCherry	This study
JG 556	JG22 pBAD_ISSod1::mCherry	This study

Anaerobic growth of S. oneidensis MR-1.

All S. oneidensis MR-1 strains used in this work are listed in Table 1. These strains were grown aerobically at 30°C in LB medium or anaerobically in minimal medium (4 M) [1.27 mM K₂HPO₄, 0.73 mM KH₂PO₄, 5 mM HEPES, 150 mM NaCl, 485 μM CaCl₂, 9 mM (NH₄)₂SO₄, 5 μM CoCl₂, 0.2 μM CuSO₄, 57 μM H₃BO₃, 5.4 μM FeCl₂, 1.0 mM MgSO₄, 1.3 μM MnSO₄, 67.2 μM Na₂EDTA, 3.9 μM Na₂MoO₄, 1.5 μM Na₂SeO₄, 2 mM NaHCO₃, 5 μM NiCl₂, and 1 μM ZnSO₄, pH 7.4] supplemented with 50 mM lactate as a carbon and electron source and 50 mM ferric citrate, 50 mM fumarate, or 20 mM DMSO as an electron acceptor. Anaerobic medium bottles were sealed with rubber stoppers. Oxygen was removed from the medium by repeatedly flushing the headspace of each bottle for 2 min with nitrogen, followed by a 2-min application of vacuum. Nitrogen gas and vacuum cycles were repeated 25 times before bottles were autoclaved. Anaerobic 150-ml serum bottles were inoculated with an S. oneidensis MR-1 culture to an initial OD₆₀₀ of 0.01. The optical density during anaerobic growth on ferric citrate was measured at a wavelength of 655 nm to avoid scattering caused by ferric iron.

Ferrous iron measurements.

Ferrous iron concentrations were measured using the ferrozine assay described by Stookey (33), with minor modifications. Briefly, 100 μl of sample was first acidified and diluted by addition of 900 μl 2 M HCl. Subsequently, 20-μl aliquots of acidified samples and ammonium iron(II) sulfate standard solutions were supplemented with 180 μl of a 1-mg ml⁻¹ ferrozine [3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine] and 50% ammonium acetate solution. If necessary, samples were diluted further. Absorption was determined at 562 nm.

Fractionation of cells.

The periplasmic fraction of anaerobically grown S. oneidensis MR-1 cells was isolated by use of polymyxin B as described by Pitts et al. (34). After polymyxin B treatment, the cells were harvested at 15,000 relative centrifugal force (RCF) and 4°C for 10 min, resuspended in 100 mM HEPES (pH 7.4) containing 10% glycerol and 0.1 mg/ml DNase I, and passed through a French pressure cell at 137 MPa. Unbroken cells were removed by a centrifugation step at 3,000 RCF and 4°C for 10 min. The supernatant was centrifuged at 208,000 RCF and 4°C for 60 min. The pellet contained the cell membranes. Separation of outer and cytoplasmic membranes was conducted according to the method of Leisman et al. (35).

Promoter prediction.

The BPROM bacterial promoter predictor was used to identify entire (−35/−10) putative promoter regions (SoftBerry, Mt. Kisco, NY). E. coli-based predictions were deemed suitable due to a recent study on single-molecule characterization of the σ⁷⁰ transcription factor of S. oneidensis MR-1 indicating that it recognizes −35/−10 regions with a motif similar to that of E. coli (36).

Construction of pBAD_SO4359_STREP and pBAD_SO4360/59_STREP.

The vector pBAD_mtrB_STREP (Table 3) (27) was digested using the NcoI and PmeI enzymes. The SO4359 and SO4360/59 genes were amplified using primers 1 plus 2 and 3 plus 2, respectively. In both cases, the SO4359 gene product was thereby modified to contain the sequence for a C-terminal streptavidin (Strep) tag (Table 4). PCR fragments were cleaved with BspHI and cloned into the linearized pBAD backbone. The resulting plasmids, pMS17 and pMS18, were transformed into competent E. coli DH5α cells (Table 3). For heterologous protein expression of the SO4359_STREP or SO4360/59_STREP gene product in E. coli, cells were grown in LB medium to an optical density at 600 nm of 0.6 and then induced by addition of 0.15 mM arabinose to the medium. Cells were harvested 4 h after induction.

Table 3.

Plasmids used in this study

Plasmid	Relevant genotype	Reference or source
pBAD202	Para neo	Invitrogen (Karlsruhe, Germany)
pMS3	pBAD_mtrBSTREP	27
pMS17	pBAD_SO4359STREP	This study
pMS18	pBAD_SO4360 SO4359STREP	This study
pMS19	pBAD_wt::mCherry	This study
pMS20	pBAD_ISSod::mCherry	This study
pMQ150	cen6 r6k ura3 neo bla	37
pMS21	pMQ150_ΔSO4359	This study
pMS22	pMQ150_ΔmtrC	This study
pMS23	pMQ150_ΔSO4357 ΔSO4358	This study
pSB377	r6k Tetr	38
pMS15	pSB_ΔSO4360	This study
pMS16	pSB_ΔmtrA	This study

Table 4.

Primers used in this study

Primer no.	Name	Sequence	Purpose
1	BspHI_SO4359_for	CATGTCATGAAGTTAAGTAAAACGACAATTGC	pBAD_SO4359STREP
2	SO4359STREP_rev	TTATTTTTCGAACTGCGGGTGGCTCCAGGCGCCAAAGCTTTTCTTATATAAGAAACTG	pBAD_SO4359STREP
3	BspHI_SO4360_for	CATGTCATGAAAAAAATACTTTTATTTAAGCTAATATTTATTAGCG	pBAD_SO4360/59STREP
3	SO_4362_seq_rev	GCTTGTGCAAACTCATCTGC	ΔmtrBS determination
4	recG_seq_rev	GTGCGGTCTTCGTAGCGCAG	ΔmtrBS determination
5	pBAD_mtrBS_prom_for	GATCAATTCGCGCGCGAAGGCGAAGCGGCATTGCTGCAAGTCGTTAGCAC	Promoter fusion to mCherry
6	mCherry_pBAD_rev	CCGCCAAAACAGCCAAGCTGGAGACCGTTTTTATTTGTATAACTCATCCATAC	Promoter fusion to mCherry
7	mCherry_natprom_for	CAATATTGGGATTGTATTTTAATATGGTTTCCAAAGGGGAAG	Promoter fusion to mCherry
8	natprom_mCherry rev	CTTCCCCTTTGGAAACCATGGACCCAGCACCTGCATATAG	Promoter fusion to mCherry
9	mCherry_ISSod1_for	GTTTGGGTTGACCAGTACACATGGTTTCCAAAGGGGAAG	Promoter fusion to mCherry
10	ISSod1_rev mCherry	CTTCCCCTTTGGAAACCATGTGTACTGGTCAACCCAAAC	Promoter fusion to mCherry
11	pSB_mtrA_for	CTTAACGGCTGACATGGGAATTCCTGCAGCCCGGGCCATCACAATGGCAATGTCTG	ΔmtrA in ΔmtrBS strain
12	pSB_mtrA_rev	CAAGCTCAATAAAAAGCCCCACCGCGGTGGCGGCCGG	ΔmtrA in ΔmtrBS strain
13	pSB_SO4360_for	CAAGCTCAATAAAAAGCCCCACCGCGGTGGCGGCCACCGTTATGGCATTGCTGAC	ΔmtrA in ΔmtrBS strain
14	pSB_SO4360_rev	CTTAACGGCTGACATGGGAATTCCTGCAGCCCGGGCCTGTCATCGCTTACCATTG	ΔmtrA in ΔmtrBS strain
15	pSB377_for	CTGACATGGGAATTCCTGCAGC	pSB377 test primer
16	pSB377_rev	CTGCTATCGATGACCTTCATGTTAAC	pSB377 test primer
17	mtrC_up_for	ATGATTACGAATTCGAGCTCGGTACCCGGGGCTTATCGTCTTGGTGACAGC	ΔomcB in ΔmtrBS strain
18	mtrC_up_rev	TTTGCCCAAGCAGGGGGAGC	ΔomcB in ΔmtrBS strain
19	mtrC_down_for	GCTCCCCCTGCTTGGGCAAATTTTTTCCCTGCATAGGTTTGGC	ΔomcB in ΔmtrBS strain
20	mtrC_down_rev	CGGCCAGTGCCAAGCTTGCATGCCTGCAGGGCATGCTTAAGTTGCCACCAG	ΔomcB in ΔmtrBS strain
21	SO4357/58_up_for	ATGATTACGAATTCGAGCTCGGTACCCGGGGGCCTTTGTAGGGTGCAAATTC	ΔSO4357/58 in ΔmtrBS strain
22	SO4357/58_up_rev	TAGTTGTAATAAATATGGATAGCGC	ΔSO4357/58 in ΔmtrBS strain
23	SO4357/58_down_for	GCGCTATCCATATTTATTACAACTATAGAATCATCCCCTAAATTTAAAAGC	ΔSO4357/58 in ΔmtrBS strain
24	SO4357/58_down_rev	CGGCCAGTGCCAAGCTTGCATGCCTGCAGGATCGGGGTGGATGTATATTC	ΔSO4357/58 in ΔmtrBS strain
25	SO4359_up_for	ATGATTACGAATTCGAGCTCGGTACCCGGGGGTTGCTACGGGTAAATAAATG	ΔSO4359 in ΔmtrBS strain
26	SO4359_up_rev	ATTTAGGGGATGATTCTAATG	ΔSO4359 in ΔmtrBS strain
27	SO4359_down_for	CATTAGAATCATCCCCTAAATTACTCACTCCATTACTTCAG	ΔSO4359 in ΔmtrBS strain
28	SO4359_down_rev	CGGCCAGTGCCAAGCTTGCATGCCTGCAGGCATAATATTCACGCAGAGGTG	ΔSO4359 in ΔmtrBS strain

Promoter fusion to mCherry reporter gene.

In vitro recombination of DNA fragments was done as described by Gibson et al. (39). The pBAD_mtrB_STREP vector (30) was digested using the enzymes PmeI and NsiI. The mCherry reporter gene was amplified using primers 6 plus 7 or 6 plus 9, depending on the promoter used for fusion (Table 4) (40). The ISSod1 sequence was amplified using primers 5 and 10, whereas the fragment of the native promoter region was generated using primers 5 and 8 (Table 4). The vector backbone and the DNA fragments sharing terminal sequence overlaps were used in equimolar concentrations in a one-step isothermal reaction mix resulting in plasmids pMS19 and pMS20, respectively (Table 3). The reaction was performed in a final volume of 20 μl for 1.5 h at 50°C. After incubation, samples were dialyzed and subsequently transformed into E. coli DH5αZ1 competent cells.

Construction of markerless gene deletion mutants.

Gene deletion mutants were constructed according to the protocol of Schuetz et al. (27). All primers used in this work are listed in Table 4. Regions of 500 bp flanking mtrC (primers 17 to 20), SO4357 and SO4358 (primers 21 to 24), or SO4359 (primers 25 to 28) were amplified (Table 4). The resulting fragments contained regions overlapping the vector pMQ150 sequence and each other (37, 41). The suicide plasmid pMQ150 was cleaved with BamHI and SalI. The two fragments and the vector were then combined using the Gibson method as described above, resulting in a pMQ150 suicide vector for each respective deletion mutant (pMS21 to pMS23) (Table 3). E. coli strain WM3064 was used as the conjugal donor strain for mating with S. oneidensis MR-1 strains.

Gene disruption of mtrA and SO4360 by use of pSB377.

Expression of mtrA or SO4360 was omitted using the suicide vector pSB377 (38). A DNA fragment homologous to mtrA or SO4360 was amplified using the primer pair 11 plus 12 or 13 plus 14, respectively (Table 4). The plasmid was digested using NotI. The linearized vector and the respective amplified fragment were combined by the Gibson method as described above, resulting in plasmids pMS15 and pMS16 (Table 3). WM3064 was used as the conjugal donor strain for mating with S. oneidensis MR-1 strains. A second crossover event was not performed, which resulted in gene disruption of mtrA or SO4360. The vectors pMS15 and pMS16 were sequenced prior to mating, using the primer pair 15 plus 16 (Table 4). Gene disruption was used to construct strains JG592 and JG593 (Table 1).

Immunodetection of Strep-tagged proteins.

For visualization of Strep-tagged recombinant proteins through immunodetection, fractions were run in 10% SDS-polyacrylamide gels and blotted onto nitrocellulose membranes (Roth, Karlsruhe, Germany) by semidry transfer blotting (Bio-Rad, Munich, Germany). Western blotting was performed following a standard procedure (42). Immunodetection was conducted using a primary antibody specific for the Strep epitope (Qiagen, Hilden, Germany) and a secondary anti-mouse antibody conjugated to alkaline phosphatase (AP) (Sigma-Aldrich, Munich, Germany) according to the manufacturer's instructions. For signal development, an AP-conjugate substrate kit (Bio-Rad) was used according to the manufacturer's instructions. Imaging of the blot was performed using a ChemiDoc XRS+ imaging system (Bio-Rad, Munich, Germany).

Solexa genome sequencing.

Isolation of chromosomal DNA was conducted according to the method of Marmur (43). Sequencing was performed on an Illumina HiSeq 2000 instrument with a read length of 46 bp by GATC Biotech. The reads were assembled using BWA software (44). Identification of single nucleotide polymorphisms (SNPs) as well as insertions and deletions was done with SAMtools (44) and the assembly viewer Tablet (45).

Fluorescence microscopy.

Prior to microscopy, cells were fixed in 4% formaldehyde for 1 h at 4°C and washed twice in phosphate-buffered saline (PBS). Subsequently, the samples were immobilized on slides coated with poly-l-lysine at 46°C and then covered with 0.1% agarose. Samples were dehydrated with 100% ethanol and stained with DAPI (4′,6-diamidino-2-phenylindole), using a 1-μg/ml solution. Microscopy was performed with a Leica DM 5500 B upright research microscope (Wetzlar, Germany) equipped with a Leica DFC360FX camera (Wetzlar, Germany) and an HCX PL Fluotar 100.0 × 1.30 oil-immersion objective. The exposure time for detection of the mCherry reporter was 391 ms, with a gain of 2.9, and that for DAPI was 159 ms, with a gain of 1.8. Image processing was carried out using the Leica Application Suite (LAS AF Lite), version 2.6.0.

LC-MS/MS.

SDS-PAGE was conducted according to published procedures (46). For in-gel digestion, the excised gel bands were destained with 30% acetonitrile (ACN), shrunk with 100% ACN, and dried in a vacuum concentrator (model 5301; Eppendorf, Hamburg, Germany). Digestion with trypsin was performed overnight at 37°C in 0.05 M NH₄HCO₃ (pH 8). About 0.1 μg of protease was used for one gel band. Peptides were extracted from the gel slices with 5% formic acid. All liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses were performed on an ion-trap mass spectrometer (Agilent 6340; Agilent Technologies) coupled to a model 1200 Agilent nanoflow system via an HPLC-Chip Cube electrospray ionization (ESI) interface. Peptides were separated on an HPLC-Chip with an analytical column of 75 μm in diameter and 150 mm in length and a 40-nl trap column, both packed with Zorbax 300SB C₁₈ (5-μm particle size). Peptides were eluted using a linear acetonitrile gradient at 1%/min at a flow rate of 300 nl/min (starting with 3% acetonitrile). MS/MS analyses were performed using data-dependent acquisition mode. After an MS scan (standard enhanced mode), a maximum of three peptides was selected for MS/MS (collision-induced dissociation [CID], standard enhanced mode). Singly charged precursor ions were excluded from selection. The automated gain control was set to 350,000. The maximum accumulation time was set to 300 ms. Mascot Distiller 2.3 was used for raw data processing and for generating peak lists, essentially with standard settings for the Agilent ion-trap instrument. Mascot Server 2.3 was used for database searching with the following parameters: peptide mass tolerance, 1.1 Da; MS/MS mass tolerance, 0.3 Da; enzyme, “trypsin,” with 2 uncleaved sites allowed for trypsin; and variable modifications, carbamidomethyl (C), Gln → pyroGlu (N-terminal Q), propionamide (C), and oxidation (M). For protein and peptide identification, the Shewanella and NCBI databases were used.

RESULTS

Isolation of an mtrB suppressor mutant strain.

To investigate whether alternative routes for dissimilatory iron reduction exist in S. oneidensis MR-1, we screened for an mtrB suppressor strain which regained the ability to respire on iron. Therefore, an mtrB null mutant was inoculated into anoxic 4 M minimal medium supplemented with lactate as an electron donor and ferric citrate as an electron acceptor. After prolonged incubation (30 days), ferric iron reduction was observed by a color change of the medium. At this point, the cells were harvested and an aliquot was used for inoculation of a further serum bottle with anoxic medium containing ferric citrate as an electron acceptor. This step was repeated twice until the incubation time for complete ferric iron reduction was decreased to 48 h (data not shown). Aliquots were streaked onto agar plates supplemented with lactate and ferric citrate and incubated under anoxic conditions. An isolated single colony, referred to as the ΔmtrB^S strain, was chosen for molecular and biochemical characterization. The ability of the ΔmtrB^S strain to respire on ferric citrate was determined by monitoring ferrous iron formation and optical density over time. As Fig. 1 indicates, the ΔmtrB^S strain grew at about 30% of the wild-type rate, while the original ΔmtrB mutant was not able to thrive under these conditions. The suppressor strain grew at wild-type rates, aerobically in LB medium and anaerobically in 4 M minimal medium with fumarate or DMSO as a terminal electron acceptor (data not shown).

Fig 1.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase of the optical density (right y axis, dashed lines). The S. oneidensis MR-1 wild type, a ΔmtrB strain, and the isolated ΔmtrB suppressor strain were grown in anaerobic 4 M minimal medium supplemented with 50 mM lactate and 50 mM ferric citrate. Error bars represent standard deviations for three individual replicates.

Involvement of SO4359 in ferric iron reduction.

Further experiments aimed at identifying possible key players responsible for the suppression of the mtrB mutation. Hence, ΔmtrB^S and wild-type cells growing independently on ferric citrate were harvested at exponential phase and fractionated. Membrane and periplasmic fractions of both strains were separated using SDS-PAGE. Due to previously described homologs of mtrB and mtrA carried in the S. oneidensis MR-1 chromosome, we sought for proteins with similar molecular weights to those of MtrA and MtrB that were detectable in the mutant but not the wild-type samples. A comparison of the mass spectrometry data obtained from the membrane fractions of the wild type and the ΔmtrB^S strain revealed masses in the ΔmtrB^S samples that most probably derived from the SO4359 gene product. These peptide signals were absent in the wild-type samples. The MtrA homolog encoded by SO4360 could be detected in both the membrane and soluble fractions of the suppressor strain and could not be observed in the wild type. Solexa sequencing of the mutant strain corroborated a possible involvement of SO4359 in complementing the ΔmtrB mutant. A suspicious gap within the gene cluster carrying SO4362 to SO4357 was spotted when reads were aligned to the genome sequence of the ancestral strain. Subsequently, a markerless deletion mutant of the mtrB homolog SO4359 in the ΔmtrB^S strain was constructed. The resulting mutant was unable to respire on ferric citrate and could be complemented by SO4359 expression in trans via a pBAD expression system (Fig. 2). Furthermore, the expressed gene had the information for a C-terminal Strep tag that was used for immunodetection (SO4359_STREP). Cell viability was achieved only when expression of SO4359_STREP was induced with 0.15 mM arabinose. Higher concentrations of arabinose in 4 M minimal medium supplemented with lactate as an electron donor and ferric citrate, fumarate, or DMSO as an electron acceptor had a lethal impact on the cells, indicating cell lysis due to toxic effects of SO4359 overexpression (data not shown). However, the ability to respire on ferric citrate of the trans-complemented ΔmtrB^S ΔSO4359 strain with SO4359_STREP supports the observation of a direct involvement of the SO4359 gene product in respiratory iron reduction of the suppressor strain (Fig. 2).

Fig 2.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase of the optical density (right y axis, dashed lines) for similarly grown ΔmtrB^S, ΔmtrB^S ΔSO4359, and ΔmtrB^S ΔSO4359 pBAD_SO4359_STREP strains on anaerobic 4 M minimal medium supplemented with 50 mM lactate and 50 mM ferric citrate. SO4359_STREP expression was induced with 0.15 mM arabinose in the growth medium. Error bars represent standard deviations for three individual replicates.

Investigation of the underlying mechanism for gene upregulation.

All experiments conducted so far pointed toward expression of SO4359 as a physiological answer to the selective pressure of iron reduction in the absence of MtrB. Comparison of the ΔmtrB^S genome to the ΔmtrB reference genome by use of Tablet software (45) revealed a gap within the SO4362-to-SO4357 gene cluster. Recently, a cymA suppressor mutant strain was isolated, and the suppression was found to be due to transcriptional activation of SirC and SirD by insertion of an insertion sequence (IS) element forming a hybrid constitutive promoter (47). Sequence analysis of the region containing the identified gap within SO4362 revealed an insertion sequence belonging to the same type of family (ISSod1). Analysis of the hybrid region by use of bioinformatic tools (BPROM; SoftBerry, Mt. Kisco, NY) displayed a putative σ⁷⁰-dependent promoter whereby the predicted −35 region is derived from the IS element and the predicted −10 region is part of SO4362 (Fig. 3).

Fig 3.

Incorporation of the ISSod1 insertion sequence into SO4362. The SO4360 sequence from nucleotides 48 to 79 is shown. Black arrows, coding sequence; gray arrows, newly formed σ⁷⁰-dependent promoter with −35 and −10 promoter regions. The poly-N sequence indicates that the inserted ISSod1 sequence is longer than the sequence shown (58).

Formation of a new hybrid promoter was tested experimentally using a reporter gene assay based on the detection of mCherry fluorescence. The entire region upstream of SO4360, including the native promoter upstream of SO4362 (positions 411,554 to 413,284 of S. oneidensis MR-1) from the suppressor strain (containing ISSod1) or the wild type, was cloned upstream of the reporter gene. Both designed vectors were transformed into E. coli, and expression of the mCherry reporter was monitored using fluorescence microscopy. As Fig. 4 shows, only the construct containing the genomic region from the suppressor strain including the ISSod1 sequence led to a detectable production of mCherry (Fig. 4).

Fig 4.

Probe for functionality of the newly formed constitutive hybrid promoter of the suppressor strain. The genomic region upstream of SO4359/60 of the suppressor or wild-type strain was cloned upstream of the mCherry reporter gene. Cells were grown aerobically on LB medium and monitored with an HCX PL Fluotar 100.0 × 1.30 oil objective with an exposure time of 957 ms (upper panels). Viability of the cells was illustrated by DAPI staining (middle panels). Overlays are shown in the lower panels. The construct including the genomic region from the suppressor strain, as well as the ISSod1 sequence, showed detectable mCherry expression (upper right panel), whereas expression of the reporter gene was mostly omitted when regulated by the wild-type promoter without the ISSod sequence.

Taken together, the data show that the ISSod1 sequence likely forms a constitutive hybrid promoter together with a part of SO4362, which consequently results in upregulation of the SO4362-to-SO4357 gene cluster. This supports the hypothesis of an involvement of SO4359 in respiratory iron reduction of the suppressor strain.

MtrA or gene product of SO4360 as a periplasmic electron carrier?

As previously observed in S. oneidensis MR-1, MtrA seems to have a fundamental function, beyond its function as a periplasmic electron carrier, in the establishment of an outer membrane-spanning complex (18, 30). Due to the involvement of MtrA in complex formation, several possibilities for the reacquired ability to transfer electrons to ferric citrate by the suppressor mutant arise. (i) MtrA interacts with the protein encoded by the mtrB homolog SO4359. MtrA is involved in both the stability of the β-barrel protein and periplasmic electron transfer in a putative MtrA-SO4359 protein-MtrC complex. (ii) MtrA serves solely as an electron carrier in the putative complex, but the protein encoded by the mtrA homolog SO4360 is crucial for periplasmic stability of the SO4359 protein. (iii) Alternatively, the gene product of SO4360 replaces MtrA as the periplasmic electron carrier, but MtrA is necessary to prevent DegP-dependent degradation of the SO4359 protein. (iv) The protein encoded by SO4360 serves as an electron carrier and assists in establishment of the outer membrane complex.

To investigate whether MtrA or the MtrA homolog encoded by SO4360 is essential for the suppression of MtrB, we constructed mutants of the ΔmtrB^S strain lacking either mtrA and SO4359 or SO4360 and SO4359 (strains JG592 and JG593) (Table 1). SO4359_STREP was expressed in trans in the individual mutant strains by the addition of 0.15 mM arabinose. Subsequently, the strains were tested for the ability to respire on ferric citrate by monitoring Fe(II) formation as well as changes in optical density over time. The resulting growth curves revealed that a deletion of mtrA in the ΔmtrB^S strain did not affect the strain's ability to reduce ferric citrate at all, whereas a deletion of SO4360 in the ΔmtrB^S strain severely impaired the ability to grow on ferric citrate (Fig. 5).

Fig 5.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase of the optical density (right y axis, dashed lines) for similarly grown ΔmtrB^S, ΔmtrB^S ΔSO4359 ΔmtrA pBAD_SO4359_STREP, and ΔmtrB^S ΔSO4359 ΔSO4360 pBAD_SO4359_STREP strains on anaerobic 4 M minimal medium supplemented with 50 mM lactate and 50 mM ferric citrate. Error bars represent standard deviations for three individual replicates.

In addition, sole expression of SO4359_STREP in trans in an mtrB mutant strain also was not sufficient to complement for iron respiration (data not shown), which furthermore suggests that coexpression of SO4360 is indispensable for growth of the ΔmtrB^S strain under iron-reducing conditions. As a result of the above experiments, it is likely that MtrA is involved neither as a periplasmic electron carrier nor as an essential part for complex formation.

To answer the question concerning an involvement of the gene product of SO4360 in the stability of the SO4359-encoded protein, SO4359_STREP was expressed in trans in a ΔmtrB^S ΔSO4359 ΔSO4360 mutant strain. As a positive control for immunodetection, a similar experiment was conducted with a ΔmtrB^S ΔSO4359 ΔmtrA mutant strain. Cells were grown anaerobically on 4 M minimal medium supplemented with 50 mM lactate and 50 mM fumarate. Subsequently, the cells were harvested at exponential phase and fractionated. Membrane fractions were separated into outer membrane and cytoplasmic membrane fractions by use of N-lauroylsarcosine (35). Equal amounts of protein (50 μg) were separated using SDS-PAGE. Subsequent immunodetection revealed a sharp band for the ΔmtrB^S ΔSO4359 ΔmtrA mutant strain, whereas the signal for the SO4359_STREP protein was scattered over the entire blot for the ΔmtrB^S ΔSO4359 ΔSO4360 sample (Fig. 6A). These results seem to indicate an involvement of the SO4360 MtrA homolog in stability of the SO4359 protein, although a complete loss of the signal, as shown recently for MtrB in the absence of MtrA, was not observed (30). Similar experiments were also conducted in E. coli, using a vector that contained either SO4359_STREP or both SO4359_STREP and SO4360 (referred to here as SO4360/59_STREP). Surprisingly, and in contrast to the previously observed MtrA-dependent stability of MtrB in S. oneidensis MR-1 and E. coli, expression of SO4359_STREP was not affected, or even was negatively affected, by coexpression of SO4360 in E. coli (Fig. 6B).

Fig 6.

(A) Immunodetection of SO4359_STREP protein in total membrane (TM), outer membrane (OM), inner membrane (IM), and soluble fractions of anaerobically grown cells of either the ΔmtrB ΔmtrA ΔSO4359 pBAD_SO4359_STREP or ΔmtrB ΔSO4360 ΔSO4359 pBAD_SO4359_STREP strain. (B) OM, IM, and soluble fractions with heterologous expression of either the SO4359_STREP or SO4360-SO4359_STREP protein in E. coli. Fifty-microgram protein samples were loaded onto a 10% SDS-PAGE gel.

Thus, the results suggest an involvement of the SO4360 protein in electron transfer, but the periplasmic electron carrier is not involved or has a minor role in the stability of the SO4359 protein in the suppressor strain.

The SO4359-SO4360 module.

Having shown that MtrA is not involved in electron transfer and that expression of the mtrB homolog SO4359_STREP alone does not complement either a ΔmtrB^S ΔSO4359 ΔSO4360 strain or a ΔmtrB strain, the next question was if the SO4359-SO4360 module could complement a ΔmtrB strain under dissimilatory iron-reducing conditions. Therefore, the vector containing SO4360 and SO4359_STREP under the control of an arabinose-inducible promoter was transformed into an S. oneidensis MR-1 ΔmtrB strain. Subsequently, this strain was inoculated into 4 M minimal medium with lactate as an electron donor and ferric citrate as an electron acceptor. Gene expression was achieved by supplementing the medium with 0.15 mM arabinose. Induction of SO4360/59_STREP resulted in the ability of the mtrB mutant to respire on ferric citrate to the same extent as that of the ΔmtrB^S strain (Fig. 7). Without the addition of arabinose, the ΔmtrB strain with SO4360/59_STREP was severely impaired in iron respiration. Therefore, it seems reasonable to conclude that the lack of MtrB is compensated in the suppressor strain by production of both the SO4359 and SO4360 proteins.

Fig 7.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase in optical density (right y axis, dashed lines). Growth of the ΔmtrB^S strain was compared to that of a ΔmtrB pBAD_SO4360/59_STREP strain in which expression was either induced with 0.15 mM arabinose or not induced. Error bars represent standard deviations for three individual replicates.

Deciphering the final reductase in the ΔmtrB^S strain.

In S. oneidensis MR-1, wild-type electron transfer over the outer membrane under anaerobic iron-reducing conditions occurs only when the MtrCAB proteins form an integral membrane-spanning complex. Having shown that the mtrB homolog SO4359 and the mtrA homolog SO4360 are essential for electron transfer to extracellular electron acceptors, the remaining aim was to elucidate the terminal ferric citrate reductase of the suppressor strain. Our hypotheses considered two possible enzymes which could be responsible for electron transfer to iron under anoxic conditions. The first was the c-type cytochrome MtrC, which today is well established as highly important for ferric iron reduction of the wild type (23, 24, 26). Specifically, the SO4359 protein could replace MtrB, and due to the high similarity of the β-barrel proteins, MtrC could also interact with the SO4359 protein. In that case, the SO4359 protein would serve as a sheath in which the SO4360 and MtrC proteins exchange electrons. The second hypothesis centered on the SO4358 and SO4357 DmsAB homologs, which could catalyze electron transfer to ferric iron in a rather unspecific reaction. In order to test the two hypotheses, we created in-frame gene deletion strains lacking either MtrC or both the SO4357 and SO4358 proteins in a ΔmtrB^S background and tested them for the ability to grow on ferric citrate. Surprisingly, both strains were not affected in ferric iron reduction and growth under iron-reducing conditions, as indicated in Fig. 8.

Fig 8.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase of the optical density (right y axis, dashed lines) for similarly grown ΔmtrB^S, ΔmtrB^S ΔmtrC, and ΔmtrB^S ΔSO4357/58 strains. Strains were inoculated into anaerobic 4 M minimal medium supplemented with 50 mM lactate and 50 mM ferric citrate. Error bars represent standard deviations for three individual replicates.

The only outer membrane cytochrome that can complement an mtrC deletion mutant is the highly similar cytochrome MtrF (17, 26, 29). Hence, it might be possible that the SO4359 and SO4360 proteins form a complex together with the c-type cytochrome MtrF.

To exclude interactions of the SO4360 and SO4359 proteins with MtrF or any other remaining outer membrane c-type cytochromes, we transformed SO4360 and SO4359_STREP into an S. oneidensis MR-1 strain deficient in all outer membrane c-type cytochromes and mtrA (ΔOMCA strain) (30). Thereafter, we again determined whether growth deficiency of this strain compared to the ΔmtrB^S strain could be observed. Expression of SO4360 and SO4359_STREP in the ΔOMCA strain was achieved by adding 0.15 mM arabinose. The growth curve clearly revealed that the strain lacking any outer membrane c-type cytochromes was able to respire on ferric citrate when SO4360 and SO4359_STREP were expressed. When expression of the periplasmic electron transfer protein and the β-barrel protein was omitted, the ΔOMCA strain lacked the ability to respire on ferric citrate (Fig. 9). To exclude a possible redundancy of the SO4357 and SO4358 final reductases to any known final ferric iron reductases and vice versa, SO4357 and SO4358 were deleted in the ΔOMCA strain. Hence, this strain does not contain any cell surface-localized terminal reductases, which might transfer electrons to insoluble iron minerals. Still, the phenotype of this strain could still be rescued using the plasmid carrying SO4360/59_STREP (see Fig. S2 in the supplemental material). Therefore, it seems as if a putative complex of the SO4359 and SO4360 proteins not only can complement an mtrB deletion but also can function as a terminal reductase for ferric citrate.

Fig 9.

Correlation between ferric citrate reduction (left y axis, solid lines) and an increase of the optical density (right y axis, dashed lines) for similarly grown ΔmtrB^S strain and induced or uninduced ΔOMCA pBAD_SO4360/59_STREP strain on anaerobic 4 M minimal medium supplemented with 50 mM lactate and 50 mM ferric citrate. Error bars represent standard deviations for three individual replicates.

DISCUSSION

Catabolic reduction of extracellular electron acceptors requires extended respiratory chains. This at least is the core understanding we have from previous studies that describe mechanisms for metal respiration in Shewanella and Geobacter strains (18, 26, 27, 48–51). So far, it is not known how Geobacter species transport respiratory electrons through the outer membrane. However, bacteria of the genus Shewanella achieve electron transfer over the outer membrane via a complex consisting of an outer membrane β-barrel protein and a periplasmic decaheme c-type cytochrome (4, 17, 18, 20). Apparently, this kind of module evolved in Shewanella to couple to either outer membrane cytochromes, as in the case of MtrAB and MtrDE, or proteins of the DMSO reductase family, as in the case of DmsEF and SO4359-60 (31). The number of MtrAB-like modules varies considerably within the sequenced Shewanella species. For instance, the genome of Shewanella denitrificans lacks this kind of module, and consequently, this is so far the only known Shewanella species that is incapable of respiring on ferric iron (52). The genome of Shewanella sediminis, on the other hand, encodes nine homologous MtrAB clusters, while the chromosomes of most other species encode two or three. Interestingly, in S. oneidensis MR-1, only two of the four clusters, namely, MtrCAB and DmsEFABGH, have an identified physiological function, being linked to anaerobic iron- and DMSO-reducing conditions, respectively (23).

Besides the role of MtrAB-like modules in iron and DMSO reduction in Shewanella strains, there is also strong evidence that similar clusters are used by ferrous iron-oxidizing organisms such as Gallionella capsiferriformans, Sideroxydans lithotrophicus, and Rhodopseudomonas palustris TIE-1 (53–57). Regarding ferrous iron oxidation, the MtrAB-like modules seem to be necessary for iron oxidation, as shown, for example, by deletion mutant analysis of the phototrophic iron-oxidizing bacterium Rhodopseudomonas palustris TIE-1 by Jiao and Newman (56). Further evidence for an involvement of MtrAB homologs in ferrous iron-oxidizing bacteria is provided by a recent work from Liu et al. showing complementation of an S. oneidensis MR-1 mtrA mutant by heterologous expression of mtoA, encoding a periplasmic decaheme c-type cytochrome of the neutrophilic ferrous iron-oxidizing bacterium Sideroxydans lithotrophicus ES-1 (57). It is so far not established whether ferrous iron oxidation takes place in the periplasm or at the cell surface, but the similarity to MtrAB-like modules and the possible detrimental effect of the formation of ferric iron minerals in the periplasm favor a surface-localized oxidation.

Interestingly, ferrous iron-oxidizing bacteria do not seem to need an outer membrane cytochrome, such as a ferrous iron oxidase. This oxidase might be dispensable due to the solubility of ferrous iron, which would imply that in these organisms the periplasmic MtrA homolog is the ferrous iron oxidase. This seems well possible, since a number of c-type cytochromes can catalyze at least the reverse reaction, i.e., the reduction of soluble or colloidal ferric iron forms such as ferric citrate or ferric-nitrilotriacetic acid (NTA) to ferrous iron (27, 34, 49). Hence, iron oxidation or reduction might not necessarily need specialized enzymes but rather proteins that have a suitable redox potential window and redox centers that are localized at least partly at the protein periphery.

Our data suggest that the SO4359-SO4360 module is a cluster that fulfills similar requirements to those fulfilled by the above-mentioned iron oxidases. In the absence of an outer membrane-localized terminal reductase, the corresponding outer membrane complex encoded by SO4359 and SO4360 seems to be able to interact with iron. These two proteins therefore encompass the final step of an alternate or secondary pathway for iron reduction in S. oneidensis MR-1. In contrast, although it is part of the native ferric iron reductase complex, the MtrAB heterodimer does not form a functional iron reductase in the absence of a cell surface-localized final reductase such as MtrC or MtrF. The structural and kinetic differences between MtrAB and the SO4359-SO4360 complex allowing for their physiological disparities are not known. One hypothesis to explain the physiological differences is that the pore of the SO4359 protein is wider and therefore allows ferric iron forms to reach the periplasmic cytochrome and that the SO4360 MtrA homolog protrudes through the pore and contacts ferric iron at the cell surface. A similar hypothesis was raised in the recent work of Bücking et al., which shows the necessity of SNPs in mtrA and mtrB to complement a deletion mutant lacking all outer membrane cytochromes (54). Bücking et al. speculated that the inability of the wild-type MtrAB subcomplex to transfer electrons could be due to steric hindrance of MtrB. We believe that at least one further hypothesis can be raised. As shown by Hartshorne et al., the redox properties of MtrA and MtrC are modulated upon formation of an MtrCAB complex (18). Since modulation of redox properties might be a result of conformational changes, one could also hypothesize that the interaction of the final reductase MtrC with the MtrAB subcomplex could probably be responsible for both (i) modulation of the redox properties of MtrA and MtrC and (ii) conformational changes of MtrB, thus enabling electron flow from MtrA via MtrC to the terminal electron acceptor.

Functional redundancy and complex formation control.

It seems to be a disadvantage to the cell to insert MtrB-like proteins into the outer membrane without an attached periplasmic c-type cytochrome. This was shown by experiments that displayed an MtrA-dependent stability of MtrB (18, 30). In other words, an MtrA deletion mutant is not able to insert MtrB into the outer membrane, most probably because MtrB is degraded in the periplasmic space (30). This dependency does not seem to be as strict for the SO4359 integral membrane protein. Expression experiments conducted with S. oneidensis MR-1, using plasmids containing a copy of either SO4359 or SO4359 in conjunction with SO4360, revealed that the SO4359 protein is more prone to degradation in the absence of the SO4360 decaheme cytochrome. However, these results are not comparable to those for the complete loss of MtrB in an mtrA null mutant strain. Moreover, the control of complex formation seems to be completely absent in E. coli as a host. Heterologous expression experiments revealed independent production of the SO4359 protein even when the SO4360 periplasmic c-type cytochrome was not coexpressed. However, expression of the SO4359 β-barrel protein was not sufficient to rescue the phenotype of an mtrB null mutant. Hence, although the SO4359 protein seems to be produced and inserted into the outer membrane, the interaction with the adjacently encoded SO4360 decaheme cytochrome cannot be overcome by an interaction with MtrA. A functional substitution of one MtrA-like decaheme cytochrome for another one has so far been shown only in the case of MtrA and MtrD (26). mtrD expression can complement an mtrA mutation in a strain that produces MtrCDB, but this complementation comes with a >2-fold decrease of the iron reduction rate compared to that of an isogenic strain that expresses the native MtrCAB complex constituents (26). So far, it is not clear if the reason for this decrease in the ferric iron reduction rate is the loss of a certain percentage of MtrB due to periplasmic degradation or a diminished interaction of MtrD with MtrC. However, most importantly, as combinatorial expression profiles in our and previous studies have shown, no significant iron reduction occurs without a functional MtrAB-like module (25, 26, 54).

To date, the biological reason for why MtrB-like proteins in S. oneidensis MR-1 seem to be protected from periplasmic degradation in the presence of specific periplasmic cytochromes as interaction partners is unclear. A direct interaction of the two corresponding proteins in the periplasm prior to insertion of the β-barrel protein in the outer membrane has not been shown so far. Many questions still remain open. Could other chaperone-like proteins be involved in the MtrA-dependent stability of MtrB? Where and when does the interaction of MtrA and MtrB occur? Are there any specific domains of the respective proteins which could be responsible for the stability of the β-barrel protein? An opportunity to understand the remaining questions might be given by the MtrAB-like module described here, which reveals a milder structural dependence of the SO4359 protein on the SO4360 protein yet a remaining functional dependency on each other.