One defining characteristic of Geobacter species is the expression of abundant c-type cytochromes. The genome of G. sulfurreducens contains over 100 putative c-type cytochrome-encoding genes (21). Although c-type cytochromes are generally involved in electron transport reactions, some heme-containing proteins have alternative physiological functions. Examples include catalase, peroxidases, bacterioferritin, and redox-sensing transcriptional regulators (5, 8, 10, 11, 29).

Genetic studies have implicated several G. sulfurreducens c-type cytochromes in Fe(III) reduction and electricity production (4, 16, 17, 20). Furthermore, G. sulfurreducens cytochromes can influence each other's expression. Deletion of an outer membrane-associated monoheme c-type cytochrome, OmcF, negatively affected Fe(III) reduction and decreased the levels of mRNAs for two previously characterized outer membrane cytochromes, OmcB and OmcC (14), one of which, OmcB, had been shown to play a critical role in Fe(III) reduction (16). Here we present another potential example of regulatory interactions between cytochromes. Two homologous multiheme cytochromes, OmcG and OmcH, were found to be involved in Fe(III) reduction and required for OmcB expression at the posttranscriptional level.

An OmcA-, OmcH-, and OmcG-deficient triple mutant (strain DLBK03) (Fig. ​(Fig.1A)1A) was constructed by replacing the omcA-omcH-omcG cluster with a kanamycin resistance cassette via homologous recombination as previously described (14, 17, 22), using the primers indicated in Table ​Table1.1. The genotype of the triple mutant was confirmed by Southern blotting genomic DNA digested with KpnI (locations of KpnI sites are indicated in Fig. ​Fig.1A)1A) as previously described (6, 14). Growth of the triple mutant and that of the wild-type strain in medium in which acetate was the electron donor and fumarate was the electron acceptor (6) were indistinguishable (data not shown). However, when log-phase acetate-fumarate-grown cultures (A₆₀₀ = ~0.5) were inoculated (3%) into acetate-Fe(III)-citrate medium (6), the triple mutant was impaired in Fe(III) reduction (Fig. ​(Fig.1B).1B). The wild-type strain completely reduced the Fe(III) in the medium within 2 days and had a doubling time of approximately 7 h, whereas the triple mutant had a doubling time of 18 h and required 6 days to complete Fe(III) reduction (Fig. ​(Fig.1B).1B). Fe(II) concentrations were determined by the ferrozine assay, and cell densities were determined by epifluorescence microscopy using acridine orange staining as previously described (12, 14, 25).

To determine which gene was responsible for this phenotype, three single mutants (strains DLBK09, DLBK10 and DLBK11) (Fig. ​(Fig.1A)1A) were constructed using the methods described above and the primers indicated in Table ​Table1.1. The rates of Fe(III) reduction by the three single mutants were comparable to that of the wild type (Fig. ​(Fig.1B),1B), indicating that deletion of more than one cytochrome was required for impairment of Fe(III) reduction. The high degree of identity between the N termini of OmcG and OmcH suggested that they might have similar physiological functions and thus be able to compensate for each other's absence. An omcH omcG::kan double mutant (strain DLBK15) (Fig. ​(Fig.1A;1A; Table ​Table1)1) was therefore constructed and screened as described above. The phenotype of this double mutant was comparable to that of the triple mutant (Fig. ​(Fig.1B1B).

Complementation studies were performed to investigate the functional redundancy of omcG and omcH. An omcG expression vector, pRG5-omcG, was constructed by amplifying the omcG coding sequence with primers Ex4776F and Ex4776R (Table ​(Table1)1) using previously described amplification conditions (14), inserting the coding sequence into pCR 2.1-TOPO (Invitrogen, Carlsbad, CA), excising it with EcoRI and HindIII, and ligating it into the EcoRI and HindIII sites of the expression vector pRG5 (14). The omcH expression vector, pRG5-omcH, was constructed via a similar strategy. The omcH coding sequence was amplified with primers Ex4779F and Ex4779R (Table ​(Table1),1), and flanking EcoRI and BamHI sites were utilized for insertion of the omcH coding sequence into pRG5. The omcG and omcH coding sequences were subsequently sequenced to screen for PCR artifacts. Following transformation of the triple mutant with the two expression vectors, spectinomycin-resistant colonies were screened for the simultaneous presence of both the plasmid and the omcA omcH omcG::kan mutation by using the primers indicated in Table ​Table1.1. A representative transformant of each type, DLBK03/pRG5-omcG and DLBK03/pRG5-omcH, was selected for phenotypic analysis. Expression of either omcG or omcH in trans in the triple mutant restored the wild-type phenotype (Fig. ​(Fig.1C),1C), confirming that omcG and omcH have overlapping physiological functions and play a critical role in Fe(III) reduction.

Western blot analysis was performed to determine whether the 78-kDa band was in fact OmcB. In order to generate an OmcB-specific antiserum, an 846-bp fragment of the omcB gene (A116 to N397) was amplified using primers Omcbpepnoti and Omcbpephindiii (Table ​(Table1),1), digested with NotI and HindIII, and inserted into the NotI and HindIII sites of the expression vector pET15b (Novagen, WI). Competent Escherichia coli strain Rosetta2 (DE3) (Novagen, WI) was transformed with the resulting plasmid, pET15b-omcBpep. The His-tagged OmcB peptide was successfully overexpressed and purified by Ni-nitrilotriacetic acid affinity chromatography as recommended by Novagen, and its identity was confirmed by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Cross-reacting antibodies were removed by immunoabsorption using an acetone extract prepared from an OmcB-deficient mutant (16) as described by Sambrook et al. (27), except that cells were disrupted by sonication and cell debris and unbroken cells were removed by centrifugation at 10,000 × g for 10 min. The specificity of the antiserum was confirmed by probing immunoblots of wild-type, OmcB-deficient, and OmcC-deficient outer membrane-enriched fractions (Fig. ​(Fig.2B).2B). Outer membrane-enriched fractions were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Bio-Rad, Hercules, CA). Immunoblots were probed with the OmcB-specific antiserum according to established protocols (details are at http://aroianlab.ucsd.edu/protocols/western.htm), and immunoreactive bands were visualized using an alkaline phosphatase-conjugated goat anti-rabbit secondary antibody and SigmaFast 5-bromo-4-chloro-3-indolylphosphatase/nitroblue tetrazolium tablets (Sigma, St. Louis, MO) according to the manufacturer's instructions.

Western blot analysis confirmed that the 78-kDa band missing from the outer membrane-enriched fraction of the triple mutant was OmcB (Fig. ​(Fig.2C).2C). The OmcB protein was also undetectable in the inner membrane-enriched and soluble fractions of the triple mutant (data not shown). In addition, OmcB was detected at wild-type levels in the two complemented strains, DLBK03/pRG5-omcG and DLBK03/pRG5-omcH (Fig. ​(Fig.2C).2C). Thus, expression of either omcG or omcH in trans fully restored both the ability to reduce Fe(III) (Fig. ​(Fig.1C)1C) and wild-type levels of OmcB expression (Fig. 2A and C) in the triple mutant.

One possible explanation for the effect of omcG and omcH deletion on the level of OmcB is that OmcB, OmcG, and OmcH are part of a complex required for Fe(III) reduction and the absence of OmcG and OmcH results in the accumulation and degradation of OmcB in the periplasmic space. The G. sulfurreducens genome contains two homologs (56 to 57% similar) of the periplasmic protease of E. coli, DegP, which degrades misfolded proteins, including outer membrane proteins, that accumulate in the periplasmic space (13, 30). Alternative explanations are also plausible, including direct or indirect interactions between OmcG or OmcH and the intracellular signaling network that affect expression of OmcB.