Iron (Fe) is abundant in the Earth's crust, occurring mostly in the solid phase in mineral structures. Smectite is a common 2:1 clay mineral 1 that can contain substantial Fe and play important roles in contaminant degradation, [2][3][4] heavy metal immobilization, [5][6][7] and nutrient cycling. 8 Trioctahedral Fe(II)smectite, a major product of basalt alteration, 9,10 is widespread in the ocean crust, [11][12][13] as well as other subsurface terrestrial and marine settings, [14][15][16] forming a large electron pool that may support neutrophilic Fe(II)-oxidizing bacteria (FeOB) growth.
Most FeOB cultivation uses aqueous Fe(II), [17][18][19] so it is unclear if FeOB can grow on solid Fe(II), including smectite. Fe(II) clays can be oxidized by heterotrophic nitrate-reducing bacteria. [20][21][22] However, these studies used nitrate-reducing bacteria that largely rely on organic substrates, producing nitrite that can abiotically oxidize Fe(II), including in clay, [23][24][25] so it is unclear to what extent Fe oxidation supports growth. In contrast, the Fe(II) phyllosilicate biotite has been shown to support chemolithotrophic growth of the nitratereducing, Fe(II)-oxidizing culture KS. 26 The fact that biotite and smectite are both 2:1 sheet silicates suggests that FeOB growth on smectite clays should also be possible.
To grow on Fe(II)-smectite, microbes need mechanisms to be able to uptake electrons from the solid electron source.
Previous studies have proposed two potential pathways for microbial Fe(II) oxidation: Cyc2 and MtoAB/PioAB pathways. [27][28][29][30] Cyc2 is used in oxidizing dissolved Fe(II), as demonstrated in our study on the microaerophilic FeOB Sideroxydans lithotrophicus ES-1. 31 Genes for MtoAB were expressed at a very low level and not responsive to Fe(II)citrate. 31 MtoAB is hypothesized to be used to interact with solid Fe(II), similar to the interaction of the homologous MtrAB(C) with Fe(III) minerals. As a sparingly soluble Fe(II) mineral, smectite can be used to reveal the mechanisms of solid Fe(II) oxidation.
Microbial interactions with smectite depend on the mineral structure and reactivity, so we need to carefully choose the substrate and track Fe(II)/Fe(III). The most commonly used Fe clay in microbial experiments is the nontronite NAu-2, [20][21][22]32,33 a natural Fe(III) clay standard. To test microbial oxidation, previous studies have reduced Fe(III) in NAu-2 to Fe(II) chemically or microbially. [20][21][22] The reduced Fe(II) remains in the NAu-2 dioctahedral structure, with variable distribution. 34,35 In contrast, native Fe(II)-smectites are trioctahedral, and this structural difference could cause variation in reactivity, which depends on the distribution of Fe in the structure. However, trioctahedral smectite is generally unavailable for experimentation because of its sensitivity to oxygen. 36 Studies of dioctahedral smectites show that edge-Fe is more reactive, 35,37 but interior-Fe within the crystal lattice could be accessible via electron hopping between Fe. 35,[38][39][40][41][42][43] To track microbial usage of smectite Fe, it is necessary to monitor different Fe fractions. This will tell us whether Fe(II) in smectite is dissolved and subsequently oxidized to form secondary minerals or if Fe(III) is retained in smectite. This monitoring will also show whether microbes can use interior-Fe, which will control the extent to which microbes can grow on smectite.
In this study, we investigated S. lithotrophicus ES-1, a type strain of the widely distributed Fe-oxidizing genus Sideroxydans. [44][45][46][47][48][49] Importantly, ES-1 has genes that encode both Cyc2 and MtoAB, 27,28 which enables us to differentiate the functions of the two pathways in oxidizing aqueous and solid Fe(II) using the same microbe. In this work, we tested ES-1 growth on chemically synthesized trioctahedral Fe(II)smectite, characterized the products, and demonstrated the distinct roles of Cyc2 and MtoAB in microbial Fe oxidation. These results help us to better understand how clay minerals support life and how microbes drive Fe redox reactions that impact fates of contaminants, heavy metals, and nutrients.
Biotic and Abiotic Smectite Oxidation. The Fe(II)smectite was synthesized using an established sol-gel method 50 with a modification to maintain anaerobic conditions during hydrothermal treatment. 51 The chemical formula of the synthesized trioctahedral smectite with low vacancies is Ca 0.17 (Fe 1.71 II Al 0.35 Mg 0.78 )(Si 3.60 Al 0.40 )O 10 (OH) 2 . The ratio of Fe/Mg/Al in the mineral is 53:24:23, which is a representative of natural trioctahedral Fe(II)-bearing smectites. 14,16,[52][53][54][55] ES-1 cells were cultured in modified Wolfe's minimal medium (MWMM) buffered with 20 mM 2-(N-morpholino)ethanesulfonic acid (MES) and adjusted to pH 6.0. Trioctahedral Fe(II)-rich smectite (1 g/L) was added into the culture as the electron donor. The headspace was filled with 78% N 2 /20% CO 2 /2% O 2 gas mix and refreshed every day with a slight overpressure in the bottles. Triplicate setups of small batch cultures were prepared to monitor cell growth and Fe consumption. The setups included cell and smectite, cell-free and smectite-free incubations. To acquire a sufficient amount of mineral products for characterization, cultures were scaled up to 250 mL. The cell number was recorded by direct cell counting using fluorescence microscopy after staining the cells with Syto13 in a Hausser counting chamber.
Fe Species Analysis from Small Batch Cultures. The consumption of different solid Fe(II) species can be tracked by sequential extraction in three steps: (i) CaCl 2 extraction to remove basal plane-Fe(II), (ii) NaH 2 PO 4 extraction to release Fe(II) bound to the edge sites as well as reactive interior-Fe(II), 42 and (iii) ammonium hydrogen fluoride (NH 4 HF 2 ) dissolution of less-reactive interior-Fe(II). At selected time points, 1.5 mL of culture suspension was taken using a syringe and needle in the anaerobic glove box and filtered through a 13 mm PTFE membrane (0.22 μm). The filtrate was collected to measure the aqueous Fe(II), and the filter membrane material was resuspended in 1.5 mL of 1 M CaCl 2 (pH 7) for 4 h to extract the Fe(II) sorbed on the basal plane. Then, the suspension was centrifuged at 16,000 g for 2 min to collect the supernatant. The remaining solids were washed once with deoxygenated DI water and extracted with 1.5 mL of 1 M NaH 2 PO 4 (pH 5) for 18 h to isolate the Fe(II) associated with the edge OH-groups. 42 After the phosphate extraction, the remaining solids were dissolved in 1.5 mL of 20 g/L NH 4 HF 2 . All centrifugation and washing steps were performed under anoxic conditions. Fe(II) was measured by a 1,10-phenanthroline assay. The total Fe concentration was determined by adding 10% hydroxylamine-HCl to reduce Fe(III) in the samples, followed by the 1,10-phenanthroline assay. 56 TEM. Transmission electron microscopy (TEM) was carried out on the late-log phase smectite culture using a Zeiss LIBRA 120 transmission electron microscope operating at 200 kV. The samples were deposited onto a lacey carbon-coated copper grid and washed twice with one drop of nanopure water after sitting for 30 min. The sample grid was air-dried before imaging.
Mineral Collection from Scaled-Up Cultures. Biotic and abiotic mineral samples were harvested from 250 mL cultures to perform Fe K-edge X-ray absorption spectroscopy and Mossbauer spectroscopy on day 2 and day 7 to represent the mid-log and late-log growth phases, respectively. The mineral products were harvested by centrifuging at 4000 g for 10 min and washed three times using deoxygenated water. The samples were dried in a desiccator in the anaerobic chamber. second shells were considered. For each scattering path, the amplitude reduction factor (S 0
2 ) was set as 0.9, while the interatomic distance (R); the coordination numbers (N) for Fe-Fe, Fe-Mg, and Fe-Al; and a Debye-Waller factor representing disorder (σ 2 ) were determined using nonlinear least-squares fitting. The coordination numbers of Fe-O and Fe-Si were set as 6 and 4, respectively, and relative values of N for Fe-Fe, Fe-Mg, and Fe-Al were held in proportion based on the stoichiometry in the smectite formula.
Experimental Setup for RNA Extraction from Scaled-Up Cultures. To compare the different mechanisms ES-1 uses to oxidize solid and aqueous Fe(II), smectite and Fe(II)-citrate were provided as the electron donors. Smectite was added once (at the time of inoculation) to reach a concentration of 1 g/L, while Fe(II)-citrate was amended daily to reach a target concentration of 500 μM. Cells were harvested at the mid-log phase (day 2) and late-log phase (day 5 for Fe(II)-citrate culture and day 7 for smectite culture) by filtering through a 0.22-micron membrane (Millipore GTTP). In addition, an Fe(II)-source switch experiment was performed to further investigate if the substrate change induces changes in gene expression in response to different Fe(II) sources. When the cells reached the late-log phase, the Fe(II) source was switched by inoculating smectite-grown cells into fresh media supplied with Fe(II)-citrate and vice versa (details are provided in the SI). Cells were harvested when they reached the mid-log phase after the Fe(II) source switch. RNA was extracted from the cells collected on the filter membrane using Qiagen RNeasy Micro kit as previously described. 31 Reverse-Transcription Quantitative PCR (RT-qPCR). Total RNA (0.5 ng) was input as a template to synthesize cDNA using the Maxima First Strand cDNA Synthesis Kit (Thermo Scientific). Gene expression of three cyc2 genes and mtoA was quantified with gyrB as the reference gene. Primers designed for the target genes were adopted from our previous study. 31 Quantabio SYBR Green supermix and the Bio-Rad CFX96 Real-Time PCR system were used to perform the quantification assays using the program from our previous study. 31 Each biological replicate was run in two technical replicates.
Proteomics. Protein Extraction. To obtain a sufficient amount of biomass for proteomics, we combined four 250 mL ES-1 cultures grown on either smectite or Fe(II)-citrate and collected at the late-log phase by filtration through a 0.22 μm membrane (Millipore GPWP). Each growth condition was sampled in triplicate. Then, the membranes were cut into small pieces and vortexed with lysing matrix E and 2 mL of lysis buffer, which is composed of 100 mM Tris/HCl, 10 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0, 0.05% Tween-20, and 1% w/v sodium dodecyl sulfate (SDS). After vortexing, the cell lysate was centrifuged at 14,000 g for 10 min. The supernatant was collected and concentrated using a 10 kDa spin concentrator (Millipore). Then, the protein concentration was determined by a fluorescence assay using the Qubit protein BR assay kit. The lysates were then digested, desalted, and analyzed in the Mass Spectrometry Core at the University of Delaware. Detailed information on the digestion, desalting, and mass spec analysis processes is provided in the SI.
Data Processing and Protein Quantification. The protein database ofS. lithotrophicusES-1 was downloaded from Uniprot (www.uniprot.org). MaxQuant v1.6.3.4 was used for protein quantification, and the false discovery rate (FDR) of less than 1% was required. 62 Filtering, bioinformatics, and statistical analyses of the MaxQuant output were performed in Perseus v1.6.15.0. 63 The label-free quantitation (LFQ) intensities were log (base 2) transformed, and the missing values were replaced by applying a normal distribution downshift using the Perseus default setting. To identify the differentially expressed proteins, a two-sample t-test was performed with the cutoff p-value less than or equal to 0.05. The raw proteome data file was deposited to MassIVE with an accession number of MSV000089149.
Consumption of Different Fe(II) Species. Different concentrations of smectite were added to 80 mL ES-1 culture to investigate whether ES-1 can grow on Fe(II)-smectite and if smectite concentration influences the biomass yield. No cell growth was observed in the smectite-free culture. The ES-1 cell number increased 4.7 fold as the smectite concentration increased from 0.5 to 1 g/L, suggesting ES-1 used smectite as the substrate to grow (Figure 1A). The maximum cell number was around 2.1 × 10 7 cells per mL, which is comparable to the maximum cell number we have observed in aqueous Fe(II) culture. 31 Increasing the smectite concentration to 2 g/L did not lead to higher cell density; therefore, a smectite concentration of 1 g/L was chosen for the further experiments. TEM imaging from air-dried samples shows that the smectite Environmental Science & Technology pubs.acs.org/est Article https://doi.org/10.1021/acs.est.2c05142 Environ. Sci. Technol. XXXX, XXX, XXX-XXX C mineral particles are nanosized (Figures 1B and S1), which provides a large surface area to interact with microbes. Cells are straight or curved rods, and no outer-membrane extensions were observed in the TEM images.
Many Fe(II)-bearing minerals have been found to release dissolved Fe(II) into solution, [64][65][66] which can support FeOB growth together with the mineral-associated Fe(II). The presence of different Fe(II) sources could confound interpretations of the oxidation pathway since microbes might use different pathways to access the solid and aqueous Fe(II). To test if Fe(II)-smectite releases aqueous Fe(II), the dissolved Fe(II) concentrations were measured in MWMM media with 1 g/L of smectite in the absence of oxygen. Smectite released Fe(II) rapidly initially; then, the release rate slowed down and ceased after 6 days (Figure S2A). The final dissolved Fe(II) concentration was ∼80 μM in the 1 g/L smectite culture (Figure S2A) without constant dissolution, which accounts for ∼2% of the total Fe(II). To test if the dissolved Fe(II) is the main source supporting ES-1 growth, we compared ES-1 cell density in 1 g/L of smectite culture with a daily supply of 100 μM aqueous Fe(II) culture. Fe(II) monitoring showed that aqueous Fe(II) was completely consumed during incubation before another feeding, so there was no Fe(II) accumulation. The final cell density in the smectite culture was one magnitude higher than that in the aqueous Fe(II) culture (Figure S2B), suggesting that ES-1 growth on dissolved Fe(II) could not account for the cell yields in the smectite culture, so we conclude that cells mainly grow on the solid Fe(II) in the smectite structure.
Fe(II)-smectite contains three solid Fe(II) species: basal plane-Fe(II), edge-Fe(II), and interior-Fe(II). 35,37 Our data show that the CaCl 2 -extractable basal plane-Fe(II) only accounts for a small proportion of Fe(II) in smectite (1-5%) (Figure 2), which is consistent with the proportion in reduced nontronite. 42 The NaH 2 PO 4 -extractable reactive Fe(II) and NH 4 HF 2 -dissolved less-reactive interior-Fe(II) are the major Fe(II) fractions in the unaltered smectite (Figure 2). During the incubation, the proportion of reactive NaH 2 PO 4extractable Fe(II) decreased quickly after inoculation, while the less-reactive interior-Fe(II) remained unchanged until day 12 (Figure 2). Therefore, ES-1 preferentially oxidized the more reactive Fe(II), especially the edge-Fe(II) in the smectite structure, which is consistent with previous reports on biotic smectite reduction and oxidation. 22,67 After day 12, the NaH 2 PO 4 -extractable Fe(II) ceased decreasing and the lessreactive interior-Fe(II) began to decrease, suggesting that the interior-Fe(II) was used by the cells, despite the fact that cells cannot directly contact the interior-Fe(II).
Comparison of Biotic and Abiotic Smectite Oxidation. Biotic oxidation of aqueous Fe(II) has been demonstrated to outcompete abiotic oxidation at low oxygen levels, 68,69 but it is unclear if that is also true for Fe(II) mineral oxidation. Therefore, Fe(II)/Fe total changes were monitored and compared between biotic and abiotic oxidation in a small batch (80 mL), initially using a phenanthroline assay. By this method, the mean values of Fe(II)/Fe total in the biotic group were lower than the values in the abiotic group at most of our sample time points, but the difference was not significant due to the data scatter, with the exception of the last data point on day 15 (p < 0.05, t-test) (Figure 3A). We then compared the biotic and abiotic oxidation rates and extents of the two major Fe(II) fractions: reactive NaH 2 PO 4extractable Fe(II) and less-reactive interior-Fe(II) dissolved by NH 4 HF 2 . There was no significant difference between biotic and abiotic consumption of NaH 2 PO 4 -extractable Fe(II) and less-reactive NH 4 HF 2 -extracted Fe(II) (p > 0.05, t-test), though there appears to be more biotic oxidation of NH 4 HF 2 -extracted Fe(II) at 15 days. Overall, based on these phenanthroline assay results, it is inconclusive if microbes catalyze Fe(II)-smectite oxidation at a different rate or extent than strictly abiotic processes.
Fe K-edge XANES and MBS were better able to resolve the difference in oxidation extent between bulk biotic and abiotic products. These products were collected from a scaled-up batch (250 mL) with the same synthesized smectite to obtain sufficient amounts of minerals on days 2 and 7, which correspond to ES-1 mid-log and late-log growth phases. XANES data for samples collected on day 7 showed a higher oxidation extent than the day 2 samples (Figure 4A), given the higher Fe K-edge positions in day 7 samples. On day 7, the XANES LCF showed that the biotic oxidation product had less Fe(II) (62.3 ± 1.0%) than the abiotic oxidation product (80.3 ± 1.4%) (Figure 4B and Table S1). MBS fitting at 295 K agrees with XANES results in that day 7 biotic oxidation product shows less Fe(II) than the abiotic product though the difference may not be significant due to noise in spectra (69.1 ± 3.9 vs 74.0 ± 1.1%) (Figure 5A and Table S2). These results from scaled-up cultures (250 mL) may not fully agree with the phenanthroline data from small batch (80 mL) cultures because the different headspace volume and vessel geometry could affect oxygen diffusion, which could result in different oxidation rates. Altogether, our results suggest that microbes can catalyze faster smectite oxidation than abiotic exposure to oxygen, but further characterization is needed to understand the conditions under which this occurs.
Local Structural Changes from Biotic and Abiotic Oxidation. The oxidation of Fe(II) causes a charge imbalance in the octahedral sheet of smectite, which could cause some Fe(III) to be ejected from the structure and precipitate as Fe(III) (oxyhydr)oxide. 59 To test if there is secondary Fe(III) (oxyhydr)oxide formation, we analyzed the same samples using MBS at lower temperatures, including 13 and 5 K (Figures 5B and S7), at which secondary Fe(III) (oxyhydr)oxides would be expected to form a sextet. The unaltered clay shows a doublet at 13 K, which corresponds to the dominantly ferrous components. The day 7 13 K spectra fit shows a small proportion of a partially collapsed Fe(III) "sextet" (labeled as (b)Ferric; 16% in day7_bio and 10.9% in day7_abio). To thoroughly evaluate the possibility of ferrihydrite (Fh), we tried to force a larger ferrihydrite component during fitting and could accommodate at most 6.6% (day7_bio) and 1.8% (day7_abio) without decreasing the goodness of the fit (Chi∧2). However, the fitting program optimally includes less than 0.03% Fh in both sample fits, suggesting that any secondary Fe(III) oxyhydroxide formation is likely minor and could be explained by oxidation of minor dissolved Fe(II) in solution and that the vast majority of the Fe(III) formed was retained in the smectite structure (detailed fitting description is provided in the SI section 2.3.3).
To further investigate the fate of oxidized iron, the unaltered smectite and its oxidation products were studied using EXAFS spectroscopy. Iron oxidation should produce local structural changes, specifically contractions in interatomic distances caused by the smaller ionic radius of Fe(III) compared to Fe(II). In addition, a mixture of octahedral iron in both ferrous and ferric oxidation states should increase the apparent disorder of neighboring shells of atoms as determined by EXAFS spectroscopy. For the unaltered smectite, the average distance between iron and oxygen that comprises its first coordination shell was 2.10 Å (Table S4), which is consistent with other reports. 51,59 The Fe-O distance contracted to 2.07 Å in the biotic oxidation product on day 7, which is the most oxidized sample. This sample also showed an increase in disorder in the oxygen shell (σ 2 = 0.015 for Fe-O), as would be expected for coexisting Fe(II) and Fe(III). The interatomic distance between Fe and the octahedral sheet cations (Fe, Mg, and Al) does not display a significant contraction (Table S4). The coordination number (N) for the unaltered sample is consistent with Fe fully contained in the octahedral sheet and was modeled as a random distribution of the neighboring cations as determined from the smectite chemical formula. For the oxidized samples, all spectra were well fitted by holding these N values constant (Figure 6). The Debye-Waller factor (σ 2 , a measure of disorder) increased slightly for the most oxidized samples (Table S4). The interatomic distances contracted and Debye-Waller factors slightly increased for the Fe-Si shells for the most oxidized samples. These Environmental Science & Technology S5). Samples were taken from mid-log and late-log growth phases in Fe(II)-citrate and smectite cultures; then, the Fe(II) source was switched and one more set of samples was taken after the switch. The RT-qPCR data show that cyc2_1 and mtoA were upregulated (p < 0.05, t-test), while cyc2_2 and cyc2_3 were downregulated in smectite culture vs Fe(II)-citrate culture in the single Fe source experiment (p < 0.05, t-test). In particular, mtoA expression in mid-log smectite culture was 23× higher than that in Fe(II)citrate culture (p < 0.05, t-test) (Table S5). Further the Fe source switching experiment also showed upregulation of cyc2_1 and mtoA when switched from Fe(II)-citrate culture to smectite culture and upregulation of cyc2_2 and cyc2_3 when switched from smectite culture to Fe(II)-citrate culture. The RT-qPCR data imply that Cyc2 and MtoA play different roles in oxidizing aqueous and solid Fe(II).
Expression of Cyc2 and Mto Pathways in Fe(II)-Smectite Culture vs Fe(II)-Citrate Culture. To test the expression of iron oxidases, proteomics was performed on ES-1 grown on Fe(II)-smectite and Fe(II)-citrate. All Cyc2 proteins (Cyc2_1, Cyc2_2, and Cyc2_3) were detected in both Fe(II)citrate and smectite samples, and the differential protein expression analysis shows that Cyc2_1 and Cyc2_3 are less abundant when ES-1 is grown on smectite than on Fe(II)citrate (two-sample t-test, p ≤ 0.05) (Table 1). None of the proteins involved in the Mto pathway were detected in the Fe(II)-citrate samples, while in the smectite cultures, we identified MtoB and the monoheme cytochrome c (Uniprot entry: D5CMP7) encoded by the gene Slit_2494 in the mto operon (Table 1). Since the porin is unstable by itself and cannot exist without the cytochrome, 71 it is reasonable to propose that MtoA is expressed together with MtoB in the smectite culture. The lack of MtoA detection is likely due to the strong disulfide bonds between cysteines and 10 heme groups, so the protein cannot be efficiently denatured and accurately identified by mass spectrometry. The proteomes of the Fe(II)-citrate culture have much higher number of detected proteins compared to smectite yet show no Mtorelated proteins (Table S6). Thus, we conclude that the Mto pathway is expressed for smectite oxidation but not for Fe(II)citrate oxidation. The RT-qPCR and proteomics patterns correspond well and together provide evidence that the Mto pathway is involved in solid Fe(II) oxidation, while Cyc2 mainly contributes to aqueous Fe(II) oxidation.
The constant process of Earth's crustal weathering produces trioctahedral Fe(II)-smectite, 9,10 which could represent a widely available substrate for microbial growth. Yet, its potential as an electron donor for microbial Fe oxidation has not previously been explored because trioctahedral Fe(II)smectites are not generally available for experimentation. As a result, most knowledge about microbial Fe(II) clay oxidation comes from artificially reduced dioctahedral Fe(III)-smectites, [20][21][22] which has yielded initial promising evidence that microbes can oxidize clays. However, these clays may not behave similarly because Fe distribution in the structure is significantly different. Unlike dioctahedral clays, the octahedral sheet of a trioctahedral smectite is fully occupied, allowing for more adjacent Fe atoms, which could lead to faster electrontransfer rates and better performance as an electron donor. Here, we synthesized a trioctahedral Fe(II)-smectite and demonstrated substantial growth of the FeOB S. lithotrophicus ES-1 using clay-bound Fe(II). Our results showed that S. lithotrophicus ES-1 uses the solid Fe(II) in smectite to produce biomass (Figures 1 and 2) and may accelerate mineral Fe(II) oxidation under certain conditions (Figure 4).
To grow on smectite, FeOB need mechanisms to interact with minerals. Previous studies on FeRB proposed that bacterial cells can access solid Fe(III)-minerals, including S2 and S3).
F smectite, by direct contact and/or electron shuttles. [72][73][74][75][76] Unlike the FeRB Shewanella oneidensis, ES-1 does not have a known organic electron shuttle (riboflavin) transporting system. However, ES-1 does possess the Mto pathway, which includes an outer-membrane multiheme cytochrome 27,77,78 that could enable direct contact to oxidize smectite-bound Fe(II). MtoAB is a decaheme cytochrome-porin complex homologous to the Fe(III)-reducing MtrAB in S. oneidensis, in which multiple hemes conduct electrons across the outer membrane, 29,77 thus enabling interactions with extracellular electron sources/sinks. ES-1 expressed Mto pathway proteins on smectite but not on Fe(II)-citrate, which supports a model in which ES-1 directly accesses the solid Fe(II) in smectite using MtoAB (Figure 8A).
In addition to MtoAB, the smectite-grown cells also express Cyc2 (Table 1), including substantial expression of all three Cyc2 proteins, suggesting that they also play a role in smectite oxidation. Cyc2 has a single heme, implying it is an oxidase of dissolved Fe(II). 77 Dissolved Fe(II) could shuttle electrons from smectite to cells. Previous findings in abiotic systems have demonstrated the interfacial electron transfer between Fe(III) and structural Fe in clays. 79,80 Fe(III) produced from dissolved Fe(II) oxidation could therefore be re-reduced to Fe(II) through interfacial electron transfer, as suggested by a previous study on microbial biotite oxidation. 26 Herein, we propose two possible mechanisms for ES-1 smectite oxidation: (1) direct contact to acquire electrons from mineral-bound Fe(II) via multiheme-porin complex MtoAB and (2) indirect oxidation with the involvement of dissolved Fe(II) by Cyc2 (Figure 8A).
Microbial growth is likely a function of smectite mineralogy. Fe(II) is in the middle octahedral layer, sandwiched by two silicate tetrahedral sheets. Previous studies have demonstrated that Fe in edge sites is the most reactive and is accessible to microbes. 37,67,72 However, it is not well-known whether Science & Technology pubs.acs.org/est Article microbes can oxidize interior-Fe(II) though this would clearly affect the extent to which microbes can oxidize smectite. Our results show that ES-1 can use some interior-Fe(II) (Figure 2). This suggests electron transfer between the interior-and edge-Fe (Figure 8B), which has been intensively characterized in abiotic systems. 35,[38][39][40][41][42][43] Although microbes can use the interior-Fe, we observed only partial oxidation and a twostage process (fast to slow reaction) during incubation, which are consistent with previous observations in abiotic systems. 34,35,37,43,59,81 The rapid decrease of NaH 2 PO 4 -extractable Fe(II) during days 0-3 is primarily due to the consumption of reactive edge-Fe(II) by ES-1 (Figure 2). Then, the oxidation rate could be constrained by the electron hopping rate from less-reactive interior-Fe to the edge sites (Figure 8B). Since the driving force of electron hopping is the redox potential gradient between the interior-and edge-Fe, 35,82,83 with the oxidation process, the redox potential gradient could decrease until reaching an equilibrium, which causes the decrease of reaction rate and incomplete oxidation. Partial oxidation results in mixed valent trioctahedral smectite (Figure 8B), with only minor structural changes and no detectable secondary mineral formation (Figures 5B and 6). This mixed valent smectite may then be available for microbial Fe reduction. In this way, trioctahedral Fe(II)-smectite can act as a geobattery to connect the redox cycles between FeOB and FeRB. Environmental Implications. Iron-bearing smectites are practically ubiquitous in soils and sediments, where they can support microbial life and catalyze environmentally significant processes. Previous studies have established that FeRB can reduce Fe(III) in smectite, 32,33,[84][85][86] and resulting redox changes alter the reactivity of smectites toward metals and organic contaminants. 7,32,33,37,87 Our work showed that trioctahedral Fe(II)-smectite can be used as an electron source to support the growth of a common, widely distributed microaerophilic Fe(II) oxidizer. Biotic Fe(II) oxidation may result in accelerated smectite Fe(II) oxidation under some conditions but not others, and yet, in either case, smectite oxidation fuels growth. This suggests that mineral Fe(II) measurements cannot always distinguish biotic versus abiotic effects, potentially rendering biotic effects invisible, thus emphasizing the need for gene-based markers of microbemineral interactions. As smectite Fe(II) oxidation can support the growth of FeOB, this is another way in which clays can fuel nutrient cycling in subsurface settings. FeOB and FeRB can coexist in sediment redox transition zones, 88,89 with autotrophic FeOB using energy from smectite Fe(II) oxidation to fix carbon and FeRB respiring these organics, coupled to smectite Fe(III) reduction. In this way, smectites can support both FeOB and FeRB growth and drive C and N biogeochemical cycling. In turn, FeOB and FeRB can modulate redox and charge of smectites and their reactivity, including sorption and degradation of contaminants.
Intriguingly, FeOB and FeRB appear to use similar mechanisms to oxidize and reduce smectite clays. Our work and previous studies have demonstrated that the multiheme cytochrome-porin Mto/MtrAB homologues mediate electron transfer between microbes and Fe minerals. Mto/MtrAB homologues are widely distributed among bacteria, 90 so Mto/ Mtr-based extracellular electron transfer may be a common mechanism that bacteria evolved to access solid electron sources/sinks. S. lithotrophicus ES-1 has not only Cyc2 and MtoAB but also other multiheme cytochrome genes that may also be useful in accessing additional solid substrates. Such a diverse toolkit may allow FeOB like ES-1 to adapt to changing Fe dynamics, thus playing an important role in sustaining life on geomaterials across diverse habitats.
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.2c05142.
Additional TEM images; data on dissolved Fe(II) release; MBS data collected at lower temperatures; method details on culturing, MBS fitting, and proteomics (PDF)
https://doi.org/10.1021/acs.est.2c05142 Environ. Sci. Technol. XXXX, XXX, XXX-XXX E
https://doi.org/10.1021/acs.est.2c05142 Environ. Sci. Technol. XXXX, XXX, XXX-XXX
https://doi.org/10.1021/acs.est.2c05142 Environ. Sci. Technol. XXXX, XXX, XXX-XXX G
https://doi.org/10.1021/acs.est.2c05142 Environ. Sci. Technol. XXXX, XXX, XXX-XXX H