The c-type cytochrome (OmcB) and the multicopper protein (OmpB) required for Fe(III) oxide reduction by Geobacter sulfurreducens were predicted previously to be outer membrane proteins, but it is not clear whether they are positioned in a manner that permits the interaction with Fe(III). Treatment of whole cells with proteinase K inhibited Fe(III) reduction, but had no impact on the inner membrane-associated fumarate reduction. OmcB was digested by protease, resulting in a smaller peptide. However, immunogold labeling coupled with transmission electron microscopy did not detect OmcB, suggesting that it is only partially exposed on the cell surface. In contrast, OmpB was completely digested with protease. OmpB was loosely associated with the cell surface as a substantial portion of it was recovered in the culture supernatant. Immunogold labeling demonstrated that OmpB associated with the cell was evenly distributed on the cell surface rather than localized to one side of the cell like the conductive pili. Although several proteins required for Fe(III) oxide reduction are shown to be exposed on the outer surface of G. sulfurreducens, the finding that OmcB is also surface exposed is the first report of a protein required for optimal Fe(III) citrate reduction at least partially accessible on the cell surface.
It is important to understand the mechanisms of electron transfer to Fe(III) in Geobacter species because they are the most abundant Fe(III)-reducing microorganisms in a diversity of subsurface environments in which Fe(III) reduction is an important process (Rooney-Varga et al., 1999; Snoeyenbos-West et al., 2000; Röling et al., 2001; Holmes et al., 2002, 2005; Anderson et al., 2003; North et al., 2004; Ortiz-Bernad et al., 2004; Vrionis et al., 2005). The insoluble nature of Fe(III) oxides requires that Geobacter species transfer electrons outside the cell in order to reduce Fe(III) (Lovley et al., 2004). In addition to Fe(III) oxides, Geobacter species are capable of reducing a variety of other extracellular electron acceptors, including Mn(IV) oxides (Lovley & Phillips, 1988), other metals (Lovley et al., 1991, 1993; Caccavo et al., 1994; Ortiz-Bernad et al., 2004), humic substances (Lovley et al., 1996), and electrodes (Bond et al., 2002; Bond & Lovley, 2003; Lovley, 2006).
Most studies of extracellular electron transfer in Geobacter species have focused on Geobacter sulfurreducens because the complete genome sequence (Methé, 2003), a genetic system (Coppi et al., 2001), and an in silico genome-based metabolic model (Mahadevan et al., 2006) are available. Furthermore, G. sulfurreducens can readily be cultured with soluble Fe(III) form, Fe(III) citrate, or fumarate as the electron acceptor.
Previous studies have identified various electron transfer components required for Fe(III) reduction by G. sulfurreducens. Several c-type cytochromes are required for optimal reduction of Fe(III) citrate as well as Fe(III) oxide. They localize in the inner membrane (Butler et al., 2004) or periplasm (Lloyd et al., 2003), as well as in the outer membrane including the c-type cytochrome, OmcB (Leang et al., 2003; Leang & Lovley 2005; Kim et al., 2006). However, other components are exclusively required for Fe(III) oxide reduction, but not the reduction of Fe(III) citrate. These include the c-type cytochromes OmcS and OmcE (Mehta et al., 2005), the multicopper protein, OmpB (Mehta et al., 2006), as well as the electrically conductive pili, known as ‘microbial nanowires’ (Reguera et al., 2005). The pili are clearly displayed outside the cell (Reguera et al., 2005) as are OmcS and OmcE (Mehta et al., 2006). OmpB and OmcB are both considered to be located in the outer membrane, but whether these proteins are exposed to the outer surface has not been determined previously.
The purpose of this study was to further localize OmpB and OmcB in order to better understand their role in Fe(III) reduction. The results suggest that whereas OmpB is highly exposed on the outer surface and is only loosely associated with the outer membrane, OmcB appears to be tightly associated with the outer membrane, with only a portion of the protein exposed to the extracellular environment.
Materials and methods
Bacterial strains and preparation of cell fractions
Wild type (DL1) as well as omcB (DL6, ΔomcB::cam) (Leang et al., 2003) and ompB (ΔompB::spec) (Mehta et al., 2006) mutant strains of G. sulfurreducens are routinely maintained in the authors’ laboratory. These pure cultures were grown under strict anaerobic conditions as described previously (Coppi et al., 2001). Briefly, the growth medium consisted of a carbonate buffered minimal medium with 20 mM acetate as the electron donor and 40 mM fumarate as the electron acceptor. Cell growth was monitored by measuring the OD600 nm (Genesys 2, Spectronic Instruments, Rochester, NY).
Wild type and mutant G. sulfurreducens cells in their late exponential growth phase were harvested by centrifugation (4000 g for 15 min at 4 °C). The supernatants were concentrated 10-fold with a centrifugal filtration system equipped with a 10 kDa-cutoff membrane. Cells from the cultures were disrupted by sonication (Sonic dismembrator F550; Fisher Scientific, PA), and the soluble and insoluble fractions were collected by centrifugation at 257 000 g for 60 min at 4 °C, The soluble fraction of disrupted cells contains cytosolic and periplasmic proteins whereas, the insoluble fraction contains membrane proteins. Membrane proteins were further fractionated into cytoplasmic membrane and outer membrane proteins by solubilizing the first in 1% (w/v) sodium laurylsarcosine, as described elsewhere (Nikaido, 1994; Kaufmann & Lovley, 2001).
Western blot analysis
Proteins were separated by electrophoresis in 10% sodium dodecylsulfate (SDS) polyacrylamide gels. Western blot analysis was performed by transferring the proteins to Immuno blot polyvinylidene difluoride membranes (Bio Rad, CA). The membranes were probed with polyclonal antibodies raised against a peptide of OmcB or OmpB (Kim et al., 2006; Mehta et al., 2006). A polyclonal alkaline phosphatase-conjugated anti-rabbit antibody (Sigma, MO) was used as a secondary antibody. OmcB and OmpB were visualized by staining with a SigmaFast™ 5-bromo4-chloro-3-indolyl phosphatase/nitroblue tetrazolium tablet (Sigma, MO).
Preparation of resting cell suspensions
The preparation of resting cell suspensions was carried out as previously described (Shelobolina et al., 2007). Briefly, cells from late-exponential phase cultures were harvested by centrifugation as described above and washed twice in an osmotically balanced wash buffer (NaHCO3, 2.5 g L−1; NH4Cl, 0.25 g L−1; NaH2PO4·H2O, 0.006 g L−1; KCl, 0.1 g L−1; NaCl, 1.75 g L−1). Cells were resuspended in the wash buffer and a portion was subjected to proteinase K (Sigma, MO) treatment for 30 min, as described below. The other untreated portion served as a control. The proteinase K reaction was stopped by adding a protease inhibitor (Roche, NJ) to the sample. The cells were then washed twice and resuspended in wash buffer. Untreated and proteinase K-treated cell suspensions were incubated in a minimal buffer (NaHCO3, 2.5 g L−1, NH4Cl, 0.25 g L−1, and NaH2PO4·H2O, 0.006 g L−1) with 5 mM acetate as the electron donor and 20 mM Fe(III) citrate or 10 mM fumarate as the electron acceptors to assay Fe(III)- and fumarate-reducing activities, respectively. The rate of fumarate reduction by resting cell suspensions was monitored by measuring the concentration of fumarate over time by HPLC (Shimadzu LC-6A, Kyoto, Japan). Samples were eluted from an Aminex HPX-87 H column (300 mm × 7.8 mm, Bio-Rad Laboratories, CA) at a rate of 1 mL min−1 for 20 min using 8 mM H2SO4 as an eluent. The fumarate concentration of the samples was monitored at a wavelength of 210 nm using matching retention times of fumarate standards of known concentration as controls. Fe(III) reduction was determined as the amount of HCl-extractable Fe(II) using a ferrozine assay, as previously described (Lovley & Phillips, 1988). Protein concentration was determined with the bicinchoninic acid method using bovine serum albumin as a standard (Smith et al., 1985).
Protease susceptibility assay
Cells grown to the mid-exponential phase were harvested by centrifugation for 10 min at 6000 g, at 4 °C, washed twice, and resuspended in 10 mM HEPES (pH 7.5) containing 500 µM MgCl2. The final cell density was 38.7 mg wet cells mL−1. The cells were incubated with or without 1 U mL−1 proteinase K at 37 °C, for different lengths of time (i.e. 10, 20, and 30 min). A protease inhibitor was then added to stop the proteolytic reaction. Cells were recovered by centrifugation at 6000 g for 1 min and washed twice in HEPES buffer containing the protease inhibitor. The washed cells were subjected to protein analyses by SDS-polyacrylamide gel electrophoresis (PAGE) followed by coomassie blue staining. The c-type cytochromes in the protein preparations were separated by Tricine-SDS PAGE, and visualized by heme-specific staining (Thomas et al., 1976; Francis & Becker 1984).
Protein immunodetection by transmission electron microscopy (TEM)
For immunolocalization of the OmcB and OmpB proteins, mid-exponential phase cultures of the wild type and the omcB and ompB mutant strains were adsorbed onto carbon-coated copper grids and fixed with 1% glutaraldehyde. Immunolabeling was performed at room temperature using primary antibodies raised against the OmcB and OmpB proteins [1 : 100 in phosphate buffered saline–bovine serum albumin (PBS–BSA) buffer for 1 h] and NanoGold®-conjugated secondary antibodies (Nanoprobes; 1 : 50 in PBS–BSA buffer for 30 min) following the manufacturer's recommendations. After immunolabeling, cells were treated with the Goldenhance™-EM reagent (Nanoprobes, Yaphank, NY) for 5 min to increase the size of the gold particles and were negatively stained with the vanadium-based, Nanovan™ stain (Nanoprobes, Yaphank, NY), also following the manufacturer's recommendations. Immunolabeling was also performed in cell suspensions following the manufacturer's recommendations. Briefly, cells were harvested by centrifugation (300 g for 5 min) and washed with PBS buffer containing 0.02 M glycine. Cells were incubated for 3 min with the primary antibody in a PBS–BSA buffer with gentle agitation and with the NanoGold®-conjugated secondary antibodies before treatment with the Goldenhance™-EM reagent (Nanoprobes, Yaphank, NY). The immunolabeled cells were washed twice in PBS–BSA and fixed with glutaraldehyde before being immobilized on the copper grid. Immunolabeled samples were examined with a JEOL 100S transmission electron microscope operated at 80 V. Samples containing strains in which the gene for OmcB (Leang et al., 2003) or OmpB (Mehta et al., 2006) was deleted were labeled, respectively, with the OmcB and OmpB anti-sera and used as controls for nonspecific binding of the primary antibody and background noise.
Results and discussion
Protease digestion of whole cells removes capacity for Fe(III) reduction
Information on the localization of proteins required for Fe(III) reduction in Geobacter species can aid in understanding the mechanisms of this process. One strategy for evaluating whether proteins are exposed on the outer surface of the cell is to determine whether these proteins are susceptible to protease digestion in whole-cell preparations.
Microscopic examination revealed that cells of G. sulfurreducens treated with proteinase K (1 U mL−1) for up to 30 min remained intact (Fig. 1). The protease-treated cells reduced fumarate as well as controls, which were not treated with protease (Fig. 2a). Fumarate is reduced at the inner membrane (Butler et al., 2004). Thus, these results demonstrated that the protease treatment did not disrupt electron transfer processes within the cell. However, the rate of Fe(III) citrate reduction in protease-treated cells was only 21% of that in untreated controls (Fig. 2b). These results suggested that protease treatment removed outer surface proteins that were required for Fe(III) citrate reduction.
When the proteins in the proteinase K-treated cells and untreated cells were separated with SDS-PAGE, the total protein profile of the treated cell samples did not differ significantly (Fig. 3a), suggesting that proteinase K did not have access to the majority of the most abundant cell proteins. However, when the proteins were stained for heme there were notable differences in the composition and in the intensity of the bands (Fig. 3b). The intensity of the heme-containing proteins at high molecular mass range decreased with increasing time of proteinase K treatment, with a corresponding increase in the intensity of lower molecular weight bands. The 9.6 kDa periplasmic c-type cytochrome, PpcA (Lloyd et al., 2003), remained intact throughout the proteinase K treatment (Fig. 3b), providing further evidence that the outer membrane was not disrupted during proteolysis.
Exposure of the c-type cytochrome OmcB on the outer cell surface
Although G. sulfurreducens is predicted to have a multitude of outer membrane c-type cytochromes, only one of these cytochromes that are predicted to be in the outer membrane, OmcB, has been definitely shown to be necessary for optimal reduction of soluble, chelated Fe(III) (Leang et al., 2003; Kim et al., 2006). As previously discussed (Leang et al., 2003), topology prediction programs such as signalp (http://www.cbs.dtu.dk/services/SignalP/), hmmtop (http://www.enzim.hu/hmmtop), and sosui (http://www.proteome.bio.tuat.ac.jp/sosuiframe0.html) indicate that OmcB is likely to be associated with the outer membrane of G. sulfurreducens because it contains a signal peptide homologous to those of lipoproteins, which is followed by a cysteine residue after the putative cleavage site that is thought to serve as the specific lipid attachment site of the protein to the membrane.
Associated with the loss of the capacity for Fe(III) citrate reduction in the proteinase K-treated cells was an apparent protease-catalyzed digestion of OmcB (Fig. 4a). When the proteins of cells treated with proteinase K for different periods of time were separated on SDS-PAGE gels and treated with an antibody specific for OmcB, there was a progressive loss of the 85-kDa band associated with OmcB over time. There was no loss in untreated controls. The loss of OmcB from the protease-treated cells was accompanied by the appearance of an additional band with a lower molecular weight (c. 40 kDa) that reacted with the OmcB antibody (Fig. 4a). This suggested that only a portion of the OmcB in intact cells was accessible to proteinase K.
In order to further evaluate the localization of OmcB, whole cells were treated with OmcB antibodies and gold-conjugated secondary antibodies and examined with TEM (Fig. 5). OmcB was not detected even though other, surface-exposed proteins can be detected with this same technique. However, OmcB appeared to be primarily localized in the outer membrane because when the cell was fractionated the OmcB antibodies detected OmcB in the outer membrane fraction with only traces detected in the soluble cell fraction, and in culture supernatant fluids (Fig. 6a). OmcB was absent in the cytoplasmic membrane fraction. These results suggest that although OmcB is localized in the outer membrane, only a portion of OmcB is exposed outside the cell. The OmcB antibodies were developed with a peptide of 281 amino acids from the central section of the OmcB sequence (Kim et al., 2006), which apparently is not accessible to the antibody when OmcB is localized in the intact outer membrane in vivo.
Localization of the putative multicopper protein, OmpB
In order to further evaluate the approaches used to localize OmcB, the putative multicopper protein, OmpB, was studied with similar techniques. A study in which the gene for OmpB was deleted demonstrated that OmpB is required for the reduction of Fe(III) oxide, but not Fe(III) citrate (Mehta et al., 2006). OmpB is predicted to be localized on the outer surface of G. sulfurreducens because when the gene necessary for the proper functioning of a type-II secretion system is deleted, OmpB accumulates in the periplasm (Mehta et al., 2006).
Localization of OmpB with polyclonal antibodies raised against OmpB suggested that OmpB is loosely associated with the outer surface of G. sulfurreducens. OmpB was primarily detected in culture supernatant fluids (Fig. 6b). OmpB was also detected in the outer membrane but only in small amounts even after loading four times more protein than that used to detect OmcB in the outer membrane fractions (Fig. 6b). OmpB was only weakly present in the soluble cell fraction, and was absent in the cytoplasmic membrane fraction.
TEM-immunogold analysis revealed that the OmpB antibodies could access OmpB on the cell surface and that OmpB was uniformly distributed on the cell (Fig. 5c). In these studies, culture samples were directly applied and fixed to the TEM copper grids; thus, supernatant proteins were also present in the cell surroundings and OmpB was detected in the culture supernatant fluids of wild-type cells, in agreement with cell fractionation studies. When a strain in which the gene for OmpB had been deleted was treated with the antibody, the signal was very weak, demonstrating low levels of antibody cross-reactivity (Fig. 5d). If the wild-type cells were washed before adsorption and fixation to the copper grids, the signal for OmpB was absent (Fig. 5e), demonstrating that OmpB is very loosely associated with the cell surface.
Proteinase K treatment of whole cells resulted in progressive loss of OmpB over time without the production of a secondary band (Fig. 4b). This is consistent with the results from the localization studies, which suggest that OmpB is primarily exposed outside the cell.
Topological and structural analyses with the signalp, hmmtop, and sosui programs predicted that OmpB is a soluble protein with a single transmembrane domain spanning amino acids 6 to 28. This transmembrane motif also includes the predicted cleavage site for OmpB's signal peptide (between the alanine and phenylalanine at positions 26 and 27, respectively), but, unlike OmcB, no cysteine residue that may function as a membrane anchor was found. This is consistent with the experimental data that OmpB is loosely bound to the outer membrane.
The results demonstrate for the first time that G. sulfurreducens requires protein(s) exposed on the outer surface of the cell in order to reduce soluble, Fe(III) citrate. Previous studies have demonstrated that several proteins especially required for the reduction of Fe(III) oxide are localized on the outer cell surface, but none of these have been required for the reduction of Fe(III) citrate. For example, pili are required for Fe(III) oxide reduction but not Fe(III) citrate reduction and, because of their electrical conductivity and specific binding of Fe(III) oxides, may be the final conduit for electron transfer from the cell to Fe(III) oxides (Reguera et al., 2005). Deleting the genes for either of the outer membrane c-type cytochromes, OmcS or OmcE, inhibits Fe(III) oxide reduction, but has no impact on Fe(III) citrate reduction (Mehta et al., 2005). Both of these cytochromes are readily sheared off the outer cell surface, suggesting that they are exposed on the outside of the cell (Mehta et al., 2005). In a similar manner, deleting the gene for OmpB inhibits Fe(III) oxide reduction (Mehta et al., 2006), but not the reduction of Fe(III) citrate, and the studies described here demonstrate that OmpB is on the outside of the cell and so loosely associated with the cell surface that it is often primarily recovered in the culture supernatant. The OmpB that is associated with the cell surface is evenly distributed, unlike the conductive pili that are also required for Fe(III) oxide reduction, but localized to one side of the cell (Reguera et al., 2005).
Like OmcS, OmcE, and OmpB, OmcB can be considered to be an outer surface protein, but with some important differences. OmcB is required for optimal Fe(III) citrate reduction as well as Fe(III) oxide reduction (Leang et al., 2003). Furthermore, OmcB appears to be much more tightly associated with the outer membrane than OmcS, OmcE, or OmpB. The results presented here demonstrate that although a portion of OmcB is exposed on the outer surface of the cell, a significant portion is likely to be embedded within the outer membrane.
These considerations are consistent with the different properties of insoluble Fe(III) oxides and soluble, chelated Fe(III). Access of redox proteins to Fe(III) oxides is expected to be much more sterically hindered than for soluble Fe(III). Therefore, in order for proteins to have the potential to transfer electrons to Fe(III) oxide directly they must be significantly displayed outside the cell. In contrast, as long as a redox-active portion of a protein such as OmcB is exposed on the outer cell surface, it is likely to have the possibility of transferring electrons to soluble Fe(III). Further analysis of the localization of other proteins predicted to be in the outer membrane of G. sulfurreducens is warranted in order to better understand their potential role in extracellular electron transfer.
This research was supported by the Office of Science (BER), US Department of Energy, Grant nos. DE-FG02-02ER63423 and DE-FC02-02ER63446.