Microbial manganese(III) reduction fuelled by anaerobic acetate oxidation

Summary

Soluble manganese in the intermediate 1III oxidation state (Mn³⁺) is a newly identified oxidant in anoxic environments, whereas acetate is a naturally abundant substrate that fuels microbial activity. Microbial populations coupling anaerobic acetate oxidation to Mn³⁺ reduction, however, have yet to be identified. We isolated a Shewanella strain capable of oxidizing acetate anaerobically with Mn³¹ as the electron acceptor, and confirmed this phenotype in other strains. This metabolic connection between acetate and soluble Mn³⁺ represents a new biogeochemical link between carbon and manganese cycles. Genomic analyses uncovered four distinct genes that allow for pathway variations in the complete dehydrogenase-driven TCA cycle that could support anaerobic acetate oxidation coupled to metal reduction in Shewanella and other Gammaproteobacteria. An oxygen-tolerant TCA cycle supporting anaerobic manganese reduction is thus a new connection in the manganese-driven carbon cycle, and a new variable for models that use manganese as a proxy to infer oxygenation events on early Earth.

Introduction

Microbial Mn cycling in marine and freshwater environments is generally assumed to consist of Mn²⁺ oxidation to solid Mn(IV) oxides in oxic aquatic environments and Mn(IV) reduction to Mn²⁺ in anoxic ecosystems (Post, 1999; Lovley, 2013; Vandieken and Thamdrup, 2014). Thus, field- and laboratory-based studies of microbial Mn reduction have focused largely on Mn(IV)-driven pathways. This dogma was recently overturned with the finding that soluble Mn³⁺ can dominate the soluble Mn pool at μM concentrations in sediment pore waters (Trouwborst et al., 2006; Madison et al., 2011; 2013) and water columns (Trouwborst et al., 2006; Schnetger and Dellwig, 2012; Luther III et al., 2015; Oldham et al., 2015) in marine environments. Soluble Mn³⁺ may form abiotically by Mn(IV) reductive dissolution of Mn(IV) oxides by siderophore-like ligands (Duckworth and Sposito, 2005; 2007) or during microbial Mn²⁺ oxidation (Tebo et al., 2005; Webb et al., 2005) or Mn(IV) reduction (Hui et al., 2012). The resulting Mn can persist in solution stabilized by organic (Oldham et al., 2017) or inorganic (Yakushev et al., 2009) ligands. Although Mn³¹ is soluble, electron transport and protein secretion pathways involved in extracellular metal reduction are required for electron transfer to Mn³⁺ (Szeinbaum et al., 2014). Therefore, Mn³⁺ has the potential to act as an important soluble extracellular electron acceptor fuelling heterotrophs under anoxic conditions.

Acetate is one of the most abundant volatile fatty acids fuelling manganese reduction in aquatic environments (Vandieken et al., 2012; 2014) and is also one of the primordial organic carbon substrates for microbial life (Russell and Martin, 2004). In anoxic sediments, acetate is one of the most abundant short chain fatty acids, (typically 0.01 to 1 mM) (Blair et al., 1991; de Graaf et al., 1996; Finke et al. 2007). Acetate can be produced via biotic or abiotic processes or imported from thermophilic ecological processes (Balba and Ned-well, 1982; Tor et al., 2001; Heuer et al. 2009; Ferry 2015). Acetate-producing microorganisms include acetogenic autotrophs that produce acetate via the Wood-Ljungdahl Pathway (WLP) (Ragsdale and Pierce, 2008) or acetogenic heterotrophs that produce acetate during incomplete oxidation of fermentation products (Lovley, 2008).

The prevalence of acetate in marine and freshwater sediments has resulted in a wealth of biogeochemical studies on anaerobic acetate metabolism under sulphate-reducing and methanogenic conditions, yet outside two members of the Geobacter family, little is known about anaerobic acetate metabolism under metal-reducing conditions (Lovley et al., 2004; Finke et al., 2007; Lovley, 2008; Elifantz et al., 2010; Vandieken et al., 2012; 2013; 2014). Geobacter and Shewanella are model dissimilatory metal-reducing microbes distinguished not only by differences in the biochemical pathway for extracellular metal-reduction but also by the apparent inability of the Shewanella genus to oxidize acetate coupled to metal reduction, which Geobacter carries out readily (Bird et al., 2011; Lovley, 2013).

Redox potentials of acetate (–290 mV) and Mn(III) (430 mV) (Kostka et al., 1995) at circumneutral pH suggest that anaerobic oxidation of acetate coupled to Mn³⁺ could support microorganisms in suboxic environments. The anaerobic oxidation of acetate coupled to Mn³⁺-pyrophosphate reduction is described by the following overall stoichiometry (Eq. (1)):

$$C{H}_{3}COOH+4M{n}_2^{\left(III\right)}{\left(P_2{O}_7\right)}_3+H_{2}O\to 2HC{O}_3^{2-}+8M{n}^{2+}+10H^{+}+12P_2{O}_7^{2-}$$ (1)

Here, we show that Mn³⁺ can act as an electron acceptor during acetate oxidation in anoxic enrichment cultures from a coastal salt marsh, dominated by Shewanella. We confirm this phenotype in Shewanella isolates from our enrichment and related strains, thus expanding the ecological niche of Shewanella to include anaerobic acetate oxidation with metal reduction, and show that Mn³⁺ can support microbial metabolism.

Results and discussion

Distribution of the main redox species in salt marsh sediments

Sediment depth profiles of the main redox species in creek bank sediments resemble previously obtained data in this environment (Taillefert et al., 2007; Hui and Taillefert, 2014). Dissolved oxygen was consumed within a couple of millimetres from the sediment-water interface, Mn²⁺ increased from immediately suboxic reached a maximum at ~140 μM in the first 20 mm, and decreased below detection limit by 60 mm (Fig. 1A left panel). As expected from thermodynamic considerations (Froelich et al., 1979), the onset of Fe(III) reduction was located below the Mn²⁺ production zone (45 mm), and Fe²⁺ reached concentrations as high as 600 mM at 65 mm then gradually decreasing with depth. Soluble organic-Fe(III) complexes, hypothesized to be produced during microbial Fe(III) reduction (Beckler et al., 2015), mirrored the Fe²⁺ profile, reached current intensities as high as 400 nA around 90 mm, then remained constant throughout the rest of the measured depths (Fig. 1A middle panel). Concentrations of nitrite and nitrate were generally low, decreased slightly with depth, and oscillated between 2 and 6 μM and between 7 and 12 μM, respectively (Fig. 1A right panel). In turn, the concentration of ammonium generally increased with depth below 18 mm as expected from the decomposition of organic matter during anaerobic Mn(IV) and Fe(III) respiration. Dissolved sulphides or FeS_(aq) clusters, indicative of sulphate reduction (Beckler et al., 2016), were not detected in these sediments. Overall, these profiles indicate that metal reduction dominates these sediments and suggest that Mn³⁺, Mn + III or +IV and Fe(III)-reducing microbial communities may be separated vertically.

Fig. 1. Mn³⁺ reduction activity of the enrichment culture (year 2) from SKIO.

A. Electron acceptor depth profiles of sediment core samples used for inoculum. Highlighted in grey dashed boxes is the 20–40 mm zone below the sediment-water interphase selected for inoculation.

B. Mn²⁺ production by sediment-free Mn³⁺-reducing enrichment culture amended with 200 mM acetate and 0.1, 1, or 14 mM SO^2–₄, indicating that sulfate-reducing bacteria were not involved.

C. Extent of Mn³⁺ reduction of the enrichment culture incubated for 60 days under anoxic conditions and amended with sodium molybdate (Mo), 140 mM; sulphate-free (No S) (10 μM for assimilation); carbonyl cyanide m-chlorophenyl hydrazone (CCCP), 200 μM; dinitrophenol (DNP), 360 and 3600 μM; ampicillin (Amp),2.3 mM; chloramphenicol (Chl), 62 and 620 μM; or heat killed (HK). [Colour figure can be viewed at wileyonlinelibrary.com]

Mn³⁺ reduction activity in enrichment cultures and in Shewanella isolates

To identify microbes mediating acetate oxidation coupled to Mn³⁺ reduction, we enriched a microbial community from the Mn²⁺-rich zone between the oxic and Fe(III)-reducing layers of these salt marsh sediments (hatched area in Fig 1A). The enrichment continued to sustain Mn³⁺ reduction from soluble Mn³⁺ (Mn³⁺-pyrophosphate) with acetate as the sole electron donor under anoxic conditions, after seven transfers spanning two years. At this transfer, 16S rRNA gene amplicon analysis revealed that the culture was dominated by Shewanella (Supporting Information Fig. S1). Sulphides were not detected in the cultures that showed Mn³⁺ reduction, and the rate of reduction of Mn³⁺ did not increase with increasing sulphate concentrations in the medium (Fig 1B), suggesting that Mn³⁺ reduction was not mediated by dissolved sulphides. Lack of sulphur resulted in decreased Mn reduction, probably due to lack of S for assimilation, as no sulphides were detected in sulphate-amended cultures and addition of molybdate enhanced Mn³⁺ reduction. Finally, Mn³¹ reduction only occurred in the absence of respiratory and growth inhibitors (Fig 1C), confirming the direct and biological nature of Mn³⁺ reduction by the microbial community.

From this enrichment, we recovered an isolate, strain MN01 that displayed 98% whole genome average nucleotide identity to Shewanella algae BrY (Caccavo et al., 1992) sequenced in this study and 100% to S. algae ACDC (Clark et al., 2013)5JCM 21037^Tand 99% to S. haliotis DW01^T (Kim et al., 2007) (hereafter the ‘S. algae clade’, since 16S rRNA gene (Supporting Information Fig. S2), and whole genome sequence suggest that S. haliotis DW01^T is a S. algae genomic species (Szeinbaum et al, submitted).

Members of the genus Shewanella are described as incomplete oxidizers of organic carbon, producing acetate as a result of the anaerobic oxidation of 3C compounds such as lactate (Hunt et al., 2010; Pinchuk et al., 2011; Lovley, 2013). Recently, S. loihica PV-4 and S. denitrificans were reported to use acetate as electron donor for nitrate reduction, but not metal reduction (Yoon et al., 2013). Strain MN-01, unlike any other Shewanella strain studied thus far, couples oxidation of acetate to reduction of either soluble Mn³⁺ pyrophosphate (Fig. 2A) or Fe³⁺-citrate (Supporting Information Fig. S7A). Like all other metal-reducing Shewanella, strain MN-01 also uses lactate for metal reduction (Supporting Information Fig. S7B). To determine the stoichiometric relationship between the consumption of acetate and Mn³⁺, we carried out incubations with increasingly limiting acetate concentrations and measured the corresponding changes in Mn³⁺ reduction (Fig. 2B). The increase in the extent of Mn³⁺ consumption as a function of acetate concentration suggests direct coupling of Mn³⁺ reduction to respiration. Under our experimental conditions, the relationship between acetate consumed and Mn³⁺ reduced is 1:6 (Fig. 2B and Supporting Information Fig. S10) less than the theoretical maximum if acetate is completely oxidized to CO₂, in which 8 electrons will reduce 8 molecules of Mn³⁺ to Mn²⁺. In anoxic incubations of strain MN-01, dissolved inorganic carbon (¹³DIC) was produced from ¹³C-acetate only in the presence of Mn³⁺, but was absent in Mn³⁺-free, heat-killed, or abiotic controls, confirming that acetate oxidation is biologically mediated and coupled to metal reduction. In these incubations, we observe this 1:8 ratio between 13C-DIC produced and Mn³⁺ reduced. Thus, the fs/fe – or ratio of electrons destined for synthesis (fs) vs. energy generation (fe) (Rittmann and McCarty, 2012) is approximately 1:3, indicating that 75% of the acetate consumed is used for energy generation (a detailed description of our calculations can be found in the supplementary section). The minor proportion (25%) of acetate not used for energy generation may be directed for synthesis. In accordance, under our growth conditions, we measured an increase in 16S rRNA copy numbers in strain MN01 from 1 × 10⁸ to 4 × 10⁸ right before the onset of Mn reduction (Supporting Information Fig. S10), suggesting that acetate consumption with Mn³⁺ reduction is predominantly an energy-generating process.

Fig. 2.

Acetate-dependent Mn³⁺ reduction activity in Shewanella algae strain MN-01 under anoxic conditions. (A) Mn³⁺-reduction activity during anaerobic incubations with excess or lack of acetate in isolate S. algae MN-01. S. oneidensis MR-1, unable to reduce Mn³⁺ with acetate as electron donor, is added here as a negative control; (B) The extent of acetate-dependent Mn³⁺ reduction activity correlates directly with acetate concentrations; (C) Mn³⁺ reduction ability with acetate as the sole electron donor by different Shewanella strains. [Colour figure can be viewed at wileyonlinelibrary.com]

Equal rates and extent of ¹³DIC production (Fig. 3A) and Mn³⁺ consumption (Fig. 3B) by MN-01 with ¹³C1-and ¹³C2-labelled acetate demonstrates complete oxidation of acetate to CO₂. To ensure that the acetate-oxidizing capacity of MN01 is not an artefact of the enrichment conditions, we also tested S. algae strains JCM 21037, BrY and S. haliotis for Mn³⁺ reduction. These strains were isolated under a variety of conditions, on a global scale. Growth on acetate and lactate (Supporting Information Figs. S3–S6) showed that the strains had similar growth rates under 1–2% NaCl, and thus Mn³⁺ reduction rates were comparable under these conditions to avoid inherent growth rate differences. All of the S. algae strains tested, as well as S. haliotis, were capable of Mn³⁺ reduction with acetate (Fig. 2C), thus expanding the ecological niche of the Shewanella genus. Microbial Mn³⁺ reduction fuelled by anaerobic acetate oxidation may thus be an underappreciated pathway for organic matter mineralization (Madison et al., 2013, Vedamati et al., 2015).

Fig. 3. Complete acetate oxidation in isolated strain Shewanella algae strain MN-01.

A. ¹³C labelled dissolved inorganic carbon (¹³DIC) produced during anaerobic incubations of S. algae MN-01 with Mn³¹ as electron acceptor and ¹³C-labelled acetate at C1 or C2.

B. Mn³⁺ reduction rates during incubations with ¹³C-labelled acetate. [Colour figure can be viewed at wileyonlinelibrary.com]

Potential biochemical pathways for acetate oxidation coupled to metal reduction

The metabolic pathway for anaerobic acetate oxidation by metal-reducing bacteria has only been examined in G. sulfurreducens and G. metallireducens (Galushko and Schink, 2000; Tang et al., 2007). In Geobacter, anaerobic acetate oxidation proceeds with the reverse TCA cycle in the oxidative direction. The TCA cycle is typically described as oxidative or reductive, reflecting the roles they played in the organisms where these cycles were discovered, but not their underlying biochemistry. At least three types of TCA cycles work oxidatively but are biochemically distinct. One of the key differences lie in the enzymes that transform 2-oxoglutarate, and are: (i) 2-oxoglutarate dehydrogenase, (ii) 2-oxoglutarate:ferredoxin oxidoreductase and (iii) 2-oxoglutarate carboxylase which leads to a succinyl-coA bypass with succinate semialde-hyde dehydrogenase producing succinate. Although these distinctions are not typically emphasized these differences may reflect underlying evolutionary and ecological forces.

More appropriate is to characterize the reverse (reductive) TCA cycle in Geobacter as ferredoxin-driven, the general feature of this cycle, and specify whether it is operating in the oxidative or reductive direction. The ferredoxin-driven TCA cycle is characterized by three key enzymes:(i) 2-oxoglutarate:ferredoxin oxidoreductase (2-OFOR), which oxidizes 2-oxoglutarate to succinyl CoA, (ii) citrate lyase and (iii) pyruvate:ferredoxin oxidoreductase (Thauer, 1988) (Fig. 4A and B left). These highly O₂-sensitive enzymes, are reversible, enabling the ferredoxin-driven TCA to be used for autotrophy (Thauer, 1988; Hugler et al., 2005). Another unique trait of the ferredoxin-driven TCA pathway in Geobacter is the generation of acetyl-CoA by succinylCoA:acetyl-CoA transferase, bypassing phosphorylation of acetate (Thauer, 1988; Galushko and Schink, 2000). The use of ferredoxin, NADP and succinyl-CoA transferase allows the ferredoxin-driven TCA cycle to transfer electrons to acceptors with lower redox potential such as sulphate and metals. Indeed, sulphate-reducing Deltaproteobacteria oxidize acetate anaerobically via the ferredoxin-driven TCA cycle (Thauer, 1988) (Fig. 4B left). Gammaproteobacteria such as Shewanella, however, are facultative anaerobes that do not encode the ferredoxin-driven TCA cycle. Instead, oxidoreductases are replaced by non-reversible dehydrogenases, rendering them oxygen-tolerant but limiting this cycle to heterotrophy, (Thauer, 1988) (Fig. 4B right). Thus, we will refer to the heterotrophic, oxygen-tolerant cycle as dehydrogenase-driven, in contrast to ferredoxin-driven. A complete dehydrogenase-driven TCA cycle has not yet been described as a pathway for acetyl-CoA oxidation coupled to energy generation under anoxic conditions, except in Cyanobacteria. In Cyanobacteria, however, 2-oxoglutarate decarboxylase (2OGDC) and succinate semialdehyde dehydrogenase (SSDH) convert 2-oxoglutarate to succi-nate and generate NADPH (Zhang and Bryant, 2011), bypassing the generation of succinyl-CoA from either 2-OFOR or 2-oxoglutarate dehydrogenase.

Fig. 4. Genetic comparison among metal-reducing Proteobacteria.

A. Genes linked to energy generation via the TCA cycle in proteobacteria during acetate-drivenmetal reduction. Green and black boxes indicate presence and absence of the gene respectively, in each strain. Shaded boxes highlight major differences in the TCA cycle between acetate oxidizers and non-oxidizers. Phylogenetic relationships between proteobacterial lineages are only for reference and not to scale.

B. Differences in the biochemistry of carbon oxidation between Shewanella and Geobacter during (i) incomplete lactate oxidation to acetate by Shewanella oneidensis MR-1 and acetate oxidation in (ii) Geobacter and (iii) proposed pathway for complete lactate oxidation to CO₂ in Shewanella algae. [Colour figure can be viewed at wileyonlinelibrary.com]

To identify possible pathways mediating complete acetate oxidation, we compared the genomes of the Shewanella strains with and without the capability of acetate oxidation coupled to Mn³⁺ reduction to identify possible pathways mediating complete acetate oxidation. We also compared the central metabolic differences between other facultative and strictly anaerobic metal-reducing proteobacteria. The gammaproteobacterium Pantoea agglomerans, the only other facultative anaerobe reported to reduce Fe(III) with acetate (Francis et al., 2000), was included along with Betaproteobacterium Rhodoferax ferrireducens (Finneran et al., 2003). No Epsilon-and Alphaproteobacterial isolates are available that couple acetate oxidation to metal reduction.

Four distinct enzymes involved in theTCA cycle - aconitate hydratase 2 (AcnA) (Supporting Information Table S1), SSDH (Supporting Information Table S1), and two NADPH:-quinone oxidoreductases (Supporting Information Table S1) (Jordan et al., 1999) - may allow metal-reducing Gammaproteobacteria to use acetate (Fig. 4A, highlighted in grey). These enzymes are absent from Shewanella species unable to couple acetate oxidation to metal reduction. The genes involved in carbon metabolism and energy generation in the S. oneidensis MR-1 and Shewanella MN01 genomes are available in the supplementary section (Supporting Information Tables S1–S7, Fig. 4A). S. algae encodes two aconitate hydratases, AcnA and AcnB, which isomerize citrate to isocitrate in the TCA cycle. AcnA is more stable, displays a higher affinity for citrate, and operates over a wider pH range than AcnB (Jordan et al., 1999), the only aconitate hydratase encoded by all other Shewanella genomes. Among the genomes analysed in this study Cyanobacterial-like SSDH containing the GabD1 domain is only found in Gammaproteobacteria and is absent from non-metal reducing members (Fig. 4B), which only contain the semialdehyde dehydrogenase with a mitochondrial GabH domain. When respiring anaerobically, the NADP⁺/NADPH couple has a redox potential more negative than NAD/NADH (–370 vs. −280 mV), closer to the redox potential of ferredoxin (–400 mV) (Wang et al., 2013). Thus, a NADP-dependent (mena)qui-none oxidoreductase may allow the dehydrognase-driven TCA cycle to reduce electron acceptors other than O₂. The other described route for acetate oxidation is the oxidative acetyl-CoA pathway (Wood-Ljungdahl pathway) (Hori et al., 2011; Ticak et al., 2014), but the absence of genes encoding the acetyl-CoA synthase/carbon monoxide dehydrogenase complex (Supporting Information Table S2) renders this pathway unlikely in Shewanella. Strain MN-01 also encodes the complete mtr pathway necessary for Mn³⁺ reduction in Shewanella (Szeinbaum et al., 2014). Based on our comparative genomic analysis, and considering that both C1 and C2 from acetate are oxidized to CO₂, we propose that in S. algae, acetate oxidation during Mn³⁺ reduction proceeds via the complete dehydrogenase-driven TCA cycle, perhaps using the SSDH bypass that Cyanobacteria employ as an obligate route.

Environmental implications of acetate oxidation with Mn³⁺ reduction supported by distinct TCA cycles

Recent findings indicate that dissimilatory Mn(IV) reduction undergoes two successive one-electron transfer steps and suggest that Mn(IV) is first solubilized by reduction to Mn³⁺, which is then respired as the terminal electron acceptor linked to energy generation (Hui et al., 2012). In accordance, our results show that acetate oxidation can be linked to Mn³⁺ reduction at millimolar levels (pure cultures) as well as with environmentally relevant concentrations (enrichment cultures). Widespread detection of soluble Mn³⁺ suggest that siderophore- or humic-stabilized Mn³⁺ in sediments (Madison et al., 2011; 2013) and water columns (Trouwborst et al., 2006; Dellwig et al., 2012; Schnetger et al., 2012; Oldham et al., 2017) plays an important but overlooked role as an energy-generating electron acceptor to heterotrophs (Hansel, 2017). An oxygen tolerant TCA cycle is a suitable biochemical pathway used by facultative anaerobes for the transfer of electrons from acetate to dissolved Mn³⁺ (Fig. 4), under fluctuating redox conditions. The high abundance of Mn³⁺ at oxic/anoxic interfaces (Trouwborst et al., 2006; Madison et al., 2013) may be able to support microbial activity and sustain active populations, as other substrates more suitable for growth become available (Fig. 5).

Fig. 5.

Working model of anaerobic acetate oxidation coupled to Mn³⁺ reduction by Shewanella in suboxic zones. Mn(III) is produced by Mn(IV) reduction or Mn(II) oxidation below the oxycline, and above ferruginous and sulphidic zones. Gammaproteobacteria such as Shewanella can utilize Mn³⁺ as an extracellular electron acceptor and persist under redox and carbon and electron acceptor fluctuations. [Colour figure can be viewed at wileyonlinelibrary.com]

Evolutionary implications of acetate oxidation with Mn³⁺ reduction supported by distinct TCA cycles

Geological evidence for microbial Mn (III/IV) oxides as early as 2.9 Ga and deep phylogenetic origins of metal-reducing microbes support an origin for Mn respiration prior to the Great Oxidation Event at ~2.4 Ga (Crowe et al., 2013; Lovley, 2013; Lyons et al., 2014; Fischer et al., 2015), a period in which oxygenic photosynthesis by Cyanobacteria made oxygen likely available in surface waters. Differences in cytochrome architectures (Bird et al., 2011) between Delta- and Gammaproteobacteria and differences in central metabolism biochemistry suggests that the enzymatic machinery for acetate oxidation coupled to metal reduction might have evolved independently, perhaps reflecting different oxygen regimes. In this model, strictly anaerobic Deltaproteobacteria (e.g., Geobacter) retained the oxygen-sensitive enzymes of the ferredoxin-driven TCA cycle, present today in other Epsilon- and Deltaproteobacteria typically found in sulphidic environments (Johnston et al., 2009; Berg, 2011), Gammaproteobacteria (e.g., Shewanella) may have evolved under higher oxygen exposure from photosynthetic activity (Johnston et al., 2009; Lyons et al., 2014; Planavsky et al., 2014; Kurzweil et al., 2016) thus requiring the dehydrogenase-driven TCA cycle, with the potential for an SSDH bypass This new link between the carbon and manganese cycles in which acetate oxidation fuels soluble Mn³⁺ reduction is therefore important to the interpretation of the roles of modern and ancient Mn redox, and Mn-driven carbon cycles.

Experimental procedures

Sediment sampling and field measurements

Triplicate sediment cores (7.5 × 50 cm) were collected from a perennial creek bank at the Salt Marsh Ecosystem Research Facility (SERF) of the Skidaway Institute of Oceanography, Georgia as described by (Taillefert et al., 2007). Vertical pore water profiles of the major redox chemical species O₂, Mn²⁺ ,Fe²⁺, and $\sum {\text{H}}_2\text{S}=\left({\text{H}}_2\text{S}+{\text{HS}}^{-}+{\text{S}}^{(0)}+{\text{S}}_{\text{x}}^{2-}\right)$ as well as qualitative signals for soluble organic-Fe³¹ complexes and FeS_(aq) molecular clusters were obtained with gold/mercury amalgam (Au/Hg) voltammetric microelectrodes (Luther et al., 2008) linked to a potentionstat (AIS, Model DLK-60) deployed on a computer-controlled micromanipulator (AIS, MAN-1). Dissolved oxygen was quantified by linear sweep voltammetry, while other species were detected by cathodic square wave voltammetry (Taillefert et al., 2000). All voltammetric measurements were conducted at a scan rate of 200 mV s⁻¹ from −0.1 to −1.8 V and included a conditioning step of 10 s at −0.1 V. When needed, a cleaning step of 10 s at −0.9 V was added before each measurement. Voltammograms were integrated using the semi-automatic integration program VOLTINT in Matlab (Bristow and Taillefert, 2008). Pore water was extracted from ~5 mm sections of the sediment core by centrifugation (3500 rpm) under N₂-atmosphere and analysed for ${\text{NO}}_3^{-}$ and ${\text{NO}}_2^{-}$ by HPLC with UV detection (Beckler et al., 2014) and for ${\text{NH}}_4^{+}$ spectrophotometrically (Riley, 1953).

Enrichment of microbial communities that couple anaerobic acetate oxidation to Mn³⁺ reduction

Enrichment cultures were initiated by transferring 2.5 g of homogenized sediment from the Mn²⁺-rich redox zone (30–50 mm depth interval) to 100 mL of anaerobic growth medium in serum bottles under a N₂ atmosphere. Growth medium consisted of sterile sulphate-free artificial seawater (Schut et al., 1993) diluted two-fold with anoxic deionized water to mimic salt marsh ionic strength. Replicate enrichments were amended with either 100 μM, 1 mM, or 14 mM sodium sulphate. Soluble Mn³⁺ (as Mn³⁺-pyrophosphate) and acetate were both added to final concentrations of 200 μM from anoxic stock solutions using an N₂-flushed syringe. Mn³⁺-pyrophosphate stock solutions were prepared by solubilization of 500 mM Mn₂O₃ in 40 mM sodium pyrophosphate at pH 6.5 (Kostka et al., 1995). The enrichments were incubated at room temperature in the dark and 5% (v/v) aliquots were transferred to fresh media every 45–60 days with an N_2-flushed syringe.

Inhibition of microbial Mn³⁺ reduction activity

A set of control enrichment cultures was incubated in the presence of a suite of metabolic inhibitors to differentiate between microbial and chemical Mn³⁺ reduction activity. The bacterial sulphate reduction inhibitor sodium molybdate (14 and 140 mM) was added to determine if sulphide produced by sulphate-reducing bacteria acted as an electron shuttle to Mn³⁺ (i.e., the enrichment conditions selected for sulphate-and not Mn³⁺-reducing bacteria). The antibiotics ampicillin(1.1 and 2.3 mM) and chloramphenicol (62 and 620 μM) were also added to replicate enrichment cultures to confirm the microbial basis of Mn³⁺ reduction activity. Respiratory inhibitors 2,4 dinitrophenol (360 and 3600 μM) and carbonyl cyanide m-chlorophenylhydrazone (CCCP, 200 μM), were added to assess whether Mn³⁺ reduction activity was electron transport chain-linked.

Cultivation-independent phylogenetic affiliation of the dominant populations of the enrichment

Biomass from the enrichment cultures displaying Mn³⁺ reduction activity with acetate as electron donor was collected by centrifugation (10 000 × g, 25 min at RT). Total community DNA was extracted and purified by standard phenol:chloroform extraction (Sambrook and Russell, 2006). The 16S SSU rRNA gene was amplified using universal primers U1 corresponding to the V1–3 hypervariable region (Supporting Information Table S1). Approximately 30 ng of enriched community DNA was used in 50 μl PCR reaction mixtures consisting of iProof DNA polymerase (BioRad) and an annealing temperature of 40°C. The amplification product was purified by electrophoretic separation on 1.5% agarose gels (Qiagen) prior to library construction. Emulsion PCR was carried out according to the manufacturer’s instructions. Sequencing of the 16S rDNA amplicon libraries was carried out on a 314 chip using the Ion Torrent PGM system. Taxonomic assignment of amplified nucleotide sequence analyses were carried out with Galaxy, an open source, web-based platform for genomic analyses (Blankenberg et al., 2010).

Isolation and phylogenetic assignment of purified bacterial strains that couple anaerobic acetate oxidation to Mn³⁺ reduction

Aliquots from the enrichment cultures displaying Mn³⁺ reduction activity were serially-diluted and spread on agar growth medium and incubated for 5–7 days under anoxic conditions. For phylogenetic identification of the purified strains, total DNA was extracted from single purified bacterial colonies, and subsequently used as template for PCR amplification of full-length 16S rDNA with primers 8F and 1489R. The resulting sequence was queried against the nucleotide collection using nBLAST search (NCBI).

Genomic DNA from isolated strain Shewanella strain MN01 and S. algae strain BrY (ATCC 51181) was purified with a DNA purification kit (Sigma-Aldrich, St. Louis, MO) after growth of an overnight culture in LB at 37°C and sequenced at the Center for Integrative Genomics (Georgia Institute of Technology, Atlanta, GA). DNA was sheared by sonication (Sonicman, Brooks Automation, Spokane, WA) to an average insert size of 650 bp, and DNA libraries were prepared for Illumina paired-end sequencing as described by the manufacturer (Illumina, San Diego, CA). Libraries were sequenced to a read length of 100 bp on an Illumina HiSeq 2500 system. Genome assembly of the resulting 11 649 110 paired-end reads was carried out in CLC Genomics Workbench 8.0.2 (CLC Bio-Qiagen, Aarhus, Denmark; http://www.clcbio.com). Reads were filtered based on quality scores (discarding sequences with at least one N) and length (> 90 nt). Contigs <2000 bp were removed from further analysis. De novo assembly generated a total of 129 contigs with an average length of 47 187 bp and average coverage of 140. The resulting genome sequences have been deposited at DDBJ/EMBL/GenBank under the accession numbers LIRM00000000 (S. algae MN01) and MDKA00000000 (S. algae BrY).

Analytical methods

Mn²⁺ and ∑H₂S concentrations in the enrichment and pure cultures were determined by cathodic square wave voltammetry as described above. Mn³⁺ reduction was monitored spectrophotometrically by absorbance at 480 nm of culture supernatants filtered through 0.2 μm pore filters (Kostka et al., 1995). Fe(III) reduction activity was monitored by following the production of Fe²⁺ via the ferrozine method (Stookey, 1970).

Quantification of acetate was performed with an Agilent 1100 series HPLC and a UV-VIS detector at 210 nm. Acetate was quantified by HPLC (Agilent 1100 series) with UV detection at 210 nm using a Supelcogel™ C-610H ion exclusion column (30 × 7.8 mm) employing an isocratic method with 0.1% phosphoric acid as the eluent at a flow rate of 0.3 ml/min. Samples for total dissolved inorganic carbon isotope (δ¹³C-DIC) analysis were collected into 2 ml crimp-seal autosampler vials fitted with 500 μl glass inserts. Samples were analysed for δ¹³C-DIC and [DIC] using a liquid chromatography-isotope ratio mass spectrometry (LC-IRMS) using an Isolink interface coupled to a ThermoFisher Delta V1 IRMS (Brandes, 2009). Although measured precision is lower in vials with inserts compared to without, the reduced sample volume allowed us to sample incubations more frequently. The system was calibrated with sodium bicarbonate standards in Milli-Q water (0–2500 μM DIC). Isotope and concentration standardization was accomplished using NaHCO₃ solutions, made by directly weighing the solid into 25 ml volumetric flasks, adding deionized water, sealing with Parafilm and rapidly mixing. Solutions were then rapidly transferred to 2 ml crimp-seal vials and sealed with Teflon-faced butyl-rubber septa (Brandes, 2009). The laboratory internal standard NaHCO₃ was calibrated against the NBS 19 calcite standard using a ThermoFisher Gasbench II interfaced to the Delta V+ IRMS (Torres et al., 2005), and all isotope values are reported versus the VPDB (Vienna Pee Dee Belemnite) reference scale (Allison et al., 1995).

Supplementary Material

Supplemental Material

Fig. S1. Mn(III) reduction activity of the 2-year enriched culture. (A)Abundance distribution of major taxa detected via Ion Torrent 16S SSU rRNA amplicon sequencing.

Fig. S2. Phylogenetic relationships between isolate MN-01 and other Shewanella type strains based on full-length 16S rRNA gene sequences. The tree was constructed using the maximum parsimony method; the bootstrap values (100 replicates) and scale bar of 1 substitution per 100 nt are shown.

Fig. S3. Growth of strains (A) S. haliotis, (B) S. algae BrY, (C) S. algae ACDC and (D) strain MN01 with lactate at 37°C under varying NaCl concentrations (%).

Fig. S4. Growth of strains (A) S. haliotis (B) S. algae BrY (C) S. algae ACDC and (D) strain MN01 with acetate at 37°C under varying NaCl concentrations (%).

Fig. S5. Growth of strains (A) S. haliotis (B) S. algae BrY (C) S. algae ACDC and (D) strain MN01 with lactate at 30°C under varying NaCl concentrations (%).

Fig. S6. Growth of strains (A) S. haliotis (B) S. algae BrY (C) S. algae ACDC and (D) strain MN01 with acetate at 30°C under varying NaCl concentrations (%).

Fig. S7. Growth of strains (A) S. haliotis (B) S. algae BrY (C) S. algae ACDC and (D) strain MN01 with lactate at 25°C under varying NaCl concentrations (%).

Fig. S8. Growth of strains (A) S. haliotis (B) S. algae BrY (C) S. algae ACDC and (D) strain MN01 with acetate at 25°C under varying NaCl concentrations (%).

Fig. S9. Mn(III)-reduction activity of the isolated strain Shewanella algae MN-01. (A) with lactate or acetate as electron donor and Fe³⁺ as electron acceptor; and (B) with lactate or acetate as electron donor and Mn³⁺ as electron acceptor

Fig. S10. Incubation of strain MN01 under anaerobic conditions with acetate and Mn³⁺-pyrophosphate. The plot shows acetate consumption (white and, Mn³⁺ reduction (light grey) on the primary axis and 16S rRNA gene copy numbers (dark grey) on the secondary axis.

Table S1. Primers used in this study.

Table S2. Shewanella algae MN-01 genes putatively involved in anaerobic acetate oxidation via the TCA cycle, and similarity to homologs in Shewanella oneidensis MR-1 (if present).

Table S3. Shewanella algae MN-01 genes putatively involved in anaerobic acetate oxidation via the Wood-Ljungdahl pathway, and similarity to homologs in Shewanella oneidensis MR-1 (if present).

Table S3. Shewanella algae MN-01 genes encoding flavinbased electron bifurcation-related proteins, and similarity to homologs in Shewanella oneidensis MR-1 (if present).