Electron transfer from Shewanella algae BrY to hydrous ferric oxide is mediated by cell-associated melanin

Abstract

Shewanella algae BrY uses insoluble mineral oxides as terminal electron acceptors, but the mechanism of electron transfer from cell surface to mineral surface is not well understood. We tested the hypothesis that cell-associated melanin produced by S. algae BrY serves as an electron conduit for bacterial–mineral reduction. Results from Fourier transform infrared spectroscopy and cell surface hydrophobicity assays indicated that extracellular melanin was associated with the cell surface. With H₂ as electron donor, washed cell suspensions of melanin-coated S. algae BrY reduced hydrous ferric oxide (HFO) 10 times faster than cells without melanin. The addition of melanin (20 µg ml⁻¹) to these melanin-free cells increased their HFO reduction rate two-fold. These results suggest that cell-associated melanin acts as an electron conduit for iron mineral reduction by S. algae BrY.

Shewanella, Melanin, Iron oxide reduction

1 Introduction

Although dissimilatory metal-reducing bacteria (DMRB) are known to use insoluble minerals as terminal electron acceptors, the mechanism(s) of electron transfer remains unclear. DMRB have been reported to use soluble electron shuttles for iron mineral reduction. Environmental humic compounds serve as terminal electron acceptors and electron shuttles for iron reduction by DMRB [1,2]. These soluble humic compounds transfer electrons to iron minerals through the redox cycling capabilities of their quinone moieties [3]. Shewanella oneidensis excretes quinones which play a role in extracellular electron transfer and have been proposed to act as electron shuttles for anaerobic respiration of minerals [4]. Geobacter sulfurreducens uses extracellular cysteine as an electron carrier to iron(III) oxides [5].

The DMRB Shewanella algae BrY produces the extracellular heteropolymer melanin [6]. The polyquinoid nature of melanin has the electrochemical properties of an amorphous semiconductor capable of redox cycling [7–10]. Soluble melanin acts as a terminal electron acceptor for S. algae BrY and as an electron shuttle for iron mineral reduction in the absence of bacteria [6]. In the absence of soluble electron shuttles, mineral reduction by DMRB requires close physical contact [11,12]. The mechanism(s) of electron transfer from the cell surface to minerals during physical contact has been postulated to be dependent on a surface-associated cellular component [12–16].

Surface-associated melanin in the hydroquinonic state reduces soluble Fe(III), which allows the yeast Cryptococcus neoformans to accumulate Fe(II) [17]. The redox cycling properties of quinones allow these molecules to be cycled back and forth from the reduced hydroquinone state to the oxidized quinone conformation [18]. Therefore, surface-associated melanin may have potential to serve as an electron conduit from the bacterial surface to a mineral surface during dissimilatory mineral reduction.

With several species of the γ-Proteobacteria, including Shewanella colwelliana, melanin production occurs outside the cell by the autooxidation and self-assembly of melanin precursors. For pyomelanin production, tyrosine is converted to homogentisic acid, which is then excreted from the cell [19–22]. Outside the cells, the homogentisic acid is oxidized abiotically and these oxidized products form polymers with molecular masses ranging from 12 000 to 120 000 Da [6,21,23]. Here we demonstrate that extracellular melanin becomes associated with the cell surface of S. algae BrY and significantly increases the rate of anaerobic iron mineral reduction.

2 Materials and methods

2.1 Culture conditions and cell preparation

S. algae BrY [24] was used throughout this study. Cultures were grown and maintained as described previously [6]. Cells from melanin-producing cultures (tyrosine-supplemented) were termed ‘melanized’, while those grown under conditions that did not produce melanin (e.g. not supplemented with tyrosine) were termed ‘non-melanized’.

2.2 Determination of cell-associated melanin

Melanin production by S. algae BrY and melanin characterization was conducted as previously described [6]. To collect cell-associated melanin, cells were harvested by centrifugation and washed three times with 30 mM sodium bicarbonate buffer, pH 6.8 (SB buffer) to remove any soluble melanin. Melanin was not detected in spent SB buffer after the third wash. The cells were acidified with 10 ml of 6 N HCl to pH 1.5 and subsequently lysed. After centrifugation (6000×g, 10 min, 4°C), the supernatant fluid was discarded and 10 N NaOH was added to the pellet until the pH reached 12. After centrifugation, 6 N HCl was added to the resulting dark solution which was acidified to pH 1.5. The precipitate was collected by centrifugation, washed with deionized water and dried at 60°C.

To determine the degree of cell-associated melanin and soluble melanin production, cultures were grown in lactate basal medium [6] with the following concentrations of tyrosine: 0, 0.25, 0.5, 1.0, 2.0 g l⁻¹. Cultures were sampled 48 h and 144 h after the onset of melanin production. The cells were harvested by centrifugation and the melanin content of the spent, cell-free culture medium was determined spectrophotometrically at a wavelength of 400 nm [6,21]. The cells were washed three times with SB buffer and analyzed for melanin content by determining the quinone content of the washed cells (see Section 2.3). Bacterial density was determined with acridine orange staining and an epifluorescence microscope [25].

2.3 Quinone assay

The presence of cell-associated quinones was determined by the nitroblue tetrazolium/glycinate assay as described previously [26]. The chemical solutions were kept on ice until use. For these assays, 100-µl cell suspensions (OD₆₀₀= 0.5) that were washed with 20 mM phosphate buffer were used. Values for melanized cells were corrected for background activity by subtracting the values obtained for non-melanized controls.

2.4 Fourier transform infrared spectroscopy (FTIR)

Melanized and non-melanized cultures were grown as described above. The cells were harvested by centrifugation, washed three times in sterile dH₂O and lyophilized for 24 h. The lyophilized cells (1–3 mg) were combined with dry KBr powder as described previously [27]. FTIR analyses were performed as described previously [6] using a Nicolet 520 FTIR spectrophotometer (Nicolet Instruments, Madison, WI, USA) with a mercury–cadmium–telluride detector, against a reference of a non-absorbing KBr matrix.

2.5 Hydrophobicity assay

Melanized and non-melanized cells were harvested by centrifugation and washed three times in 20 mM potassium phosphate buffer (pH 6.8). The cells were resuspended to a final OD₆₀₀ of 0.3 (Spec. 21 spectrometer, Unicam, Rochester, NY, USA). Hydrophobicity was determined by the bacterial adhesion to hydrocarbon (BATH) test as described previously [28]. Samples of dilute bacterial suspensions were added (3 ml each) to 10-ml glass cuvettes. After 25 µl of pure hexadecane (Sigma, St. Louis, MO, USA) were added to each suspension, they were mixed by vortexing at low speed for 1 min as previously described [28]. Optical densities at 600 nm of the suspensions were read prior to and 30 min after agitation with hexadecane. Bacteria that adhered to the hexadecane droplets rose to the top of the liquid, resulting in a decrease in absorbance. The degree of hydrophobicity was calculated as the log₁₀ ([initial OD₆₀₀ minus final OD₆₀₀]×100).

2.6 Hydrous ferric oxide (HFO) reduction

Aerobic cultures of melanized and non-melanized S. algae BrY were grown as described above. For one control, cells were grown with lactate and phenylalanine (2 g l⁻¹) instead of tyrosine. Phenylalanine is a tyrosine analogue and melanin was not produced with lactate and phenylalanine (2 g l⁻¹) in minimal medium within 48 h. The cells were harvested at 48 h of growth as described previously [6]. Washed cells (final density 10⁹ ml⁻¹) were added to sterile anaerobic vials containing 50 µl of 6 mM insoluble HFO in 10 ml SB buffer (pH 7.4) with a headspace of N₂ (80%), CO₂ (20%). To one set of duplicate tubes containing non-melanized cells, 100 µl of a 2 g l⁻¹ melanin solution [6] was added (final concentration 20 µg ml⁻¹). Two abiotic controls (without cells) were tested: (1) HFO only and (2) HFO and 20 µg ml⁻¹ melanin. The electron donor, H₂, was added (10 kPa) aseptically to all vials and they were incubated with agitation (150 rpm) in the dark at 28°C. At different time intervals, samples were removed for analysis of Fe(II) as previously described [1,6]. HFO reduction rates were calculated as previously described [6].

2.7 Statistical analysis and reproducibility

Experiments were conducted in duplicate or triplicate and performed at least twice. Statistical analysis was conducted with Student's t-test.

3 Results

3.1 Cell-associated melanin

Washed cells from melanin-producing cultures were darker than cells from non-melanin-producing cultures suggesting that some of the melanin produced associated with the cells. Acid treatment and lysis of the washed melanin-producing cells, followed by adjustment to pH 12, revealed the presence of a dark soluble pigment with the same chemical characteristics as melanin [6,26,29]. The pigment had an absorbance spectrum (300–800 nm) similar to that of commercial humic acid (Aldrich, Milwaukee, WI, USA) (data not shown). In addition, the slope of log absorbance vs. wavelength (400–600 nm) scan of the pigment's spectrum was equal to −0.0035, which is indicative of melanin [6,29]. Following precipitation, the washed, dialyzed, dried pigment was a black powder with the properties indicative of melanin [6,29]. The presence of melanin was not detected in non-melanized cells.

Melanin production in the culture supernatant increased as a function of initial tyrosine concentration, but the relative amount of cell-associated melanin reached saturation with 0.5 g l⁻¹ tyrosine (Table 1). Thus, cell-associated melanin levels did not correlate with extracellular melanin concentrations at 48 or 144 h after pigment production began. The amount of cell-associated melanin increased with time although soluble melanin production ceased. The levels of cell-associated melanin were not significantly different (P<0.05) in the cultures receiving 0.5–2.0 g l⁻¹ tyrosine, but increased two-fold at 144 h. The cell-associated melanin content of cultures supplemented with 0.25 g l⁻¹ tyrosine also increased, but at a significantly lower amount (P<0.05) than with cultures at the higher tyrosine concentrations. At 48 h after pigmentation, cell densities ranged from 3.6 to 3.9×10⁹ ml⁻¹ and did not vary significantly between treatments, indicating that tyrosine concentration had a minimal effect on growth with lactate. Cell densities remained constant until the termination of the study.

1

Effect of tyrosine concentration on the production of soluble and cell-associated melanin

Initial tyrosine concentration (g l−1)	48 h		144 h
	Cell-associated melanin (fg cell−1)	Soluble melanin (mg ml−1)	Cell-associated melanin (fg cell−1)	Soluble melanin (mg ml−1)
0	0	0	0	0
0.25	29±1	0.16±0.01	57±2	0.17±0.05
0.50	53±4	0.46±0.09	113±10	0.63±0.03
1.00	49±3	1.05±0.05	126±16	1.44±0.01
2.00	52±6	1.28±0.05	116±19	1.51±0.01

Time after onset of pigmentation.

3.2 Surface properties of melanized cells

The molecular surface composition of lyophilized bacteria was analyzed by FTIR spectroscopy to determine their physicochemical properties [30] and to detect cell surface-associated melanin. The FTIR spectra of melanized and non-melanized cells differed indicating definite surface differences (Fig. 1). The FTIR bands from melanized cells showed definite signature peaks for the presence of melanin [6,29]. The peaks at the following wavelengths and their respective structures include bands at 1650–1600 cm⁻¹, 1510 cm⁻¹, and 1410–1310 cm⁻¹, which are indicative of aromatic C=C conjugated with C=O and/or COO⁻ groups [31], aromatic C=C bonds [31] and OH groups of phenolics [32], respectively. These results support the hypothesis that the surface differences between melanized and non-melanized cells are due to surface-associated melanin.

1

FTIR scans of lyophilized whole cells of S. algae BrY. Scans of melanized and non-melanized cells are offset for clarity.

Given the hydrophilic nature of melanin [33,34], cell surface hydrophobicity for melanized and non-melanized cells of S. algae BrY was assayed by the BATH test [28]. Hydrophobicity values for melanized and non-melanized cells were 1.95 and 1.92, respectively. These values are significantly different (P<0.01) and demonstrate that melanized cells were less hydrophobic than non-melanized cells. With melanin associated on their cell surface, the cells would be expected to display a decrease in hydrophobicity. This decrease in hydrophobicity supports the hypothesis of surface-associated melanin.

3.3 HFO reduction

The effect of surface-associated melanin on insoluble Fe(III) oxide reduction was tested and the results are shown in Fig. 2. Melanized cells reduced more HFO than non-melanized cells. By the use of a non-linear regression analysis of Fe(II) evolution, the HFO reduction rates were determined and were found to differ significantly (P<0.01). The HFO reduction rate for melanized cells (1.0 µmol h⁻¹) was 10 times faster than the rate for non-melanized cells (0.1 µmol h⁻¹). No HFO reduction occurred in the abiotic controls of HFO or HFO and melanin (data not shown). Cells grown with lactate and supplemented with phenylalanine instead of tyrosine did not produce melanin prior to harvesting and did not reduce HFO at a significantly greater rate (P<0.05) than the non-melanized cells (data not shown). The addition of soluble melanin to non-melanized cell suspensions accelerated HFO reduction to a rate of 0.2 µmol h⁻¹, double that of the non-melanized cells without supplemental melanin. These data support the hypothesis that melanin is surface-associated and has the potential to significantly increase the rate of iron mineral reduction by S. algae BrY. Based on the amount of supplemental melanin used in this experiment with a cell density of 10⁹ cells ml⁻¹, approximately 20 fg of soluble melanin was required per cell to double the HFO reduction rate of non-melanized cells.

2

HFO reduction by melanized and non-melanized S. algae BrY cells. Cell suspensions (approx. 10⁹ cells ml⁻¹) in SB buffer (pH 7.2) were incubated with 50 µl of 6 mM HFO and N₂/CO₂ (80%/20%) headspace and demonstrated HFO reduction with H₂ (10 kPa) as electron donor. Melanized cells (◯), non-melanized cells (•), the addition of 20 mg l⁻¹ of melanin to non-melanized cells (□). HFO reduction was not detected in controls consisting of HFO only (█) or HFO plus 20 mg l⁻¹ of melanin (▵). Only the final time points are shown for the abiotic controls.

4 Discussion

Extracellular humic compounds, including melanin, are known to become cell-associated [20,35]. In this study, we presented several lines of evidence that demonstrate melanin was associated with the cell surface of S. algae BrY including: (1) cell pellets of melanized cells were pigmented; (2) chemical and spectral analysis of the cell-associated pigment was indicative of melanin; (3) the melanin-producing cells had an increased quinone content; (4) melanized cells had signature FTIR absorbance bands and decreased surface hydrophobicity.

Our previous study showed that the addition of soluble melanin increased the rate of bacterial HFO reduction during anaerobic respiration [6]. In this study, we showed that melanized cells had an accelerated rate of HFO reduction compared to non-melanized cells. The addition of melanin to non-melanized cell suspensions also increased the rate of HFO reduction with H₂ as the electron donor, but not to the same level as the melanized cells. These results suggest that the amount of surface-associated melanin was the rate-determining factor.

An increase in iron reduction rates has been reported for the DMRB Shewanella putrefaciens 200 (formerly Pseudomonas sp. 200) when cells are grown in basal medium supplemented with yeast extract and beef peptone compared to those cells grown in unsupplemented basal medium [36]. This increased rate of iron reduction was attributed to increased cytochrome content ascribed to growth in the supplemental carbon sources. An alternative explanation is that the cells produced surface-associated melanin.

In the absence of soluble electron shuttles, increased HFO reduction rates by S. algae BrY are dependent on their ability to adhere to HFO, which is a function of increased hydrophobicity [11,37]. Based on the BATH assays, melanized cells were actually less hydrophobic than the non-melanized cells. Consequently, they should be less adherent to HFO. Thus, the observed increase in HFO reduction by melanized cells was probably not due to an increased contact or adhesion to HFO, but to the redox cycling capacity of the surface-associated melanin.

Because of the redox cycling capacity of hydroquinones, Fe(III) reduction rates by DMRB are increased even at low concentrations [6,38]. In this study, melanized cells containing approximately 100 fg melanin cell⁻¹ had a 10-fold increased rate of HFO reduction compared to non-melanized cells, while the addition of soluble melanin (approximately 20 fg cell⁻¹) doubled the rate of HFO reduction by non-melanized cells. These results also indicate that small concentrations of melanin were responsible for increased HFO reduction rates. The ability of non-melanized cells to reduce HFO, although at a much slower rate than melanized cells, indicates that S. algae BrY may have several mechanisms for HFO reduction.

Melanin production is believed to be a response to changing growth conditions such as anoxia or carbon limitation [39]. Faced with imminent anoxia, nonfermentative, facultative anaerobes such as S. algae could derive a potential bioenergetic benefit from even small amounts of melanin production. Current studies are focused on developing melanin-deficient mutants to further elucidate the role of melanin production on Fe(III) oxide reduction.

Acknowledgements

This research was supported in part by AES Hatch Grant 389, and by the College of Life Science and Agriculture, The University of New Hampshire-Durham. This is Scientific Contribution number 2148 from the NH Agricultural Experimental Station. The authors thank Sterling Tommellini for assistance with FTIR and Robert E. Mooney and William Chesboro for discussions and assistance.

Author notes

1

Present address: Department of Biology, Whitworth College, Spokane, WA 99251, USA.