Introduction

Rich oil and gas resources are distributed in coastal wetlands, such as Louisiana Oilfield in the USA, Shengli, Panjin, and Dagang Oilfield in China1,2,3. In the process of oil development, petroleum substances inevitably enter the soil and water of wetlands through blowouts and leaks. More than 17,000 chemical components have been identified from petroleum, which have disastrous effects on the soil and water system of coastal wetlands4,5,6,7. Additionally, under anoxic or anaerobic conditions, the petroleum substances discharged are directly or indirectly metabolized by methanogens to produce considerable CH4 emissions8. Therefore, it is urgent to minimize the ecological and climatic impacts of petroleum substances in coastal wetlands.

For the removal of oil from the microhabitats of coastal wetlands, methods such as physical, chemical or biological have been attempted in recent decades9,10. Considering cost control and environmental compatibility, microbial technology is proposed as a potential in situ remediation method11,12. The coastal wetland soil contains abundant hydrocarbon-degrading microorganisms, such as Alcanivorax sp., Marinobacter sp., Nocardia sp., which metabolize hydrocarbons as the only carbon source to obtain energy13,14,15. However, due to the lack of terminal electron acceptors such as O2, the metabolism of hydrocarbons usually stops at small-molecule metabolites, which hinders the further removal of recalcitrant hydrocarbons16. Moreover, accumulated fatty acids, acetates, and other reducing organics are also utilized by methanogens to produce CH417. Strengthening the terminal or sub-terminal degradation of hydrocarbons to CO2, the greenhouse effect of which is 1/28 of the same mass of CH4, may be an effective strategy to drive the removal of hydrocarbon macromolecules and alleviate the greenhouse gas effect18.

Electroactive microorganisms metabolize small organics and transfer the generated electrons to extracellular mineral surfaces (e.g., iron/manganese oxides) by using cytochrome proteins, electron shuttles, or nanowires, thus expanding the range of available electron receptors19,20. The synergistic effect of electroactive and hydrocarbon-degrading microorganisms provides feasibility for the complete removal of hydrocarbons from the wetland habitats. In addition to oxidative degradation pathways, Geobacter species are known to perform reductive degradation of hydrocarbons, breaking down complex aromatic structures into intermediate organic products21,22. This dual functionality could significantly influence the fate of hydrocarbons in anaerobic environments. Fe is the most abundant redox-sensitive metal element in the crust, and Fe oxides have been reported to widely exist in wetland soils23,24. Electroactive microorganisms utilize the extracellular Fe oxides (e.g., ferrihydrite and hematite) to respire, that is, dissimilatory iron reduction (DIR), which is closely related to the mineralization of organics. In previous studies with pure cultures, a series of iron-reducing microorganisms, including Ferroglogus placidu, Geobacter metallireducens, and Geobacter sulfurreducens, were found to be able to use aromatics as electron donors25,26,27. Specifically, Geobacter sulfurreducens with whole genome sequence data is the model electroactive strain and involved in a variety of environmental friendly actions like bioremediation, electricity production28. Newly discovered iron-reducing microorganisms such as Desulfuromonas cholroenthenica and Anaeromyxobacter dehalogens have been shown to reduce and remove chlorined compounds via electron transport chain29. Fe(II)-bearing secondary mineralization products, such as goethite, magnetite, and siderite, were formed through the extracellular electron transport chain of electroactive microorganisms30,31. In particular, highly crystalline magnetite with good electrical conductivity (EC) can enhance the interaction of electroactive microorganisms with other functional microflora as a surrogate for the pili-associated cytochrome OmcS32. Additionally, the DIR process is regarded as the primary source of CO2 produced by heterotrophic respiration under anaerobic conditions, and it helps mitigate the greenhouse effect by effectively reducing CH4 emissions23,33.

In order to clarify the effect of electroactive microorganism-driven DIR on carbon fate34, metagenomics was applied for the microbial ecological investigation and an experiment enhancing DIR was performed to the typical coastal oil-contaminated wetland system (from Shengli Oilfield and Dagang Oilfield) in this study. Driven by soil iron transformation, the removal of organics, including hydrocarbons and soil organic matter, was systematically studied. The emissions of CH4 were also taken into account to identify the shifts in metabolic pathways. Moreover, the association between electroactive microbes and indigenous functional microflora was revealed, which expanded their scope of action by the extracellular electron transfer chain. This research reveals how microbial inoculation and Fe oxide transformation synergize to enhance pollutant degradation, methane suppression, and carbon cycling in wetlands.

Results

Microbial ecological investigation of oil-contaminated wetlands

The total petroleum hydrocarbons (TPHs) of the 44 sample sites in Shengli Oilfield (numbered S1#–S44#) ranged from 2358 to 157256 mg kg−1 (Fig. 1a), with a mean value of 60659 mg kg−1. While the TPHs of the 43 sample sites in Dagang Oilfield (numbered D1#–D43#) ranged from 2912 to 252438 mg kg−1, with a mean value of 39857 mg kg−1. Total iron (TFe) and Fe(II) content were determined to identify the iron transformation processes. The TFe contents of S1#–S44# were 3959–17136 mg kg−1, and the Fe(II) contents were 260–16011 mg kg−1. While the TFe contents of D1#–D43# were 6334–17361 mg kg−1, and the Fe(II) contents were 444–15880 mg kg−1. The fluctuation of Fe(II) content indicated the occurrence of differentiated iron transformation processes, typically, iron reduction and oxidation23. The total dissolved organic carbon (TDC) is an organic carbon component in soil that can be easily used by microorganisms, and its content can reflect the degradation of hydrocarbons. The TDC ranges of S1#–S44# and D1#–D43# were 157–4075 mg kg−1 and 185–1240 mg kg−1, respectively. Additionally, the average pH values of S1#–S44# and D1#–D43# were 8.16 ± 0.36 and 8.70 ± 0.37, respectively, indicating that they belonged to alkaline soils. EC, ranging from 281–12727 μS cm−1 in S1#–S44# and from 397–19483 μS cm−1 in D1#–D43#, was positively correlated with the content of soil salt and hydrocarbon metabolites35.

Fig. 1: Microbial ecological investigation of 87 samples from oil-contaminated wetlands in Shengli (S1#–S44#) and Dagang Oilfield (D1#–D43#).
figure 1

a The situation and location coordinates of the sampling sites. b The relative abundances of genes related to Dissimilatory Iron Reduction (DIR-g), Hydrocarbon Degradation (HD-g), Carbon Terminal Metabolism (CTM-g), Methanogenesis (MG-g), and Energy Supply (ES-g) identified by KEGG annotation of metagenomic information (See Supplementary Table 1); The upper and lower boundaries of the box represent the upper and lower quartiles of the data, respectively; The upper and lower lines on the outside of the box represent 1.5 times the interquartile range, respectively; And the symbols of “□” and “◆” represent the mean and abnormal value of the data, respectively. c Spearman correlation between functional genes (DIR-g, HD-g, CTM-g, MG-g, ES-g) and soil physicochemical indices (pH, EC, TDC, TPH, Fe(II), TFe), and the explanation degree of environmental factors for the taxonomic composition (**p < 0.01, *p < 0.05). d RDA to reveal the response of environmental factors to microbial functional genes. The blue lines represent the response variables, and the red lines represent the explanatory variables. The distribution of samples is marked with green dots. The RDA model fit was satisfactory (pseudo-F = 4.7, P = 0.008). e PCoA of samples according to the differentiation of microflora. f The average proportions of microorganisms preforming Dissimilatory Iron Reduction (DIR-m), Hydrocarbon Degradation (HD-m) and Methanogenesis (MG-m) in the samples from S1#–S44# and D1#–D43# (See Supplementary Table 2). g The co-occurrence networks showing the soil microflora of Shengli Oilfield has a higher aggregation degree while that of Dagang Oilfield has more working modules. Colors of nodes indicate the different major phyla. h Spearman correlation between the relative abundance of the typical DIR bacterium Geobacter sulfurreducens PCA obtained by aligning with BWA (Burrows-Wheeler-Alignment Tool) and environmental factors. The symbol “**” represents a highly significant positive correlation (p < 0.01).

Full size image

Metagenomics was applied to identify the microbial ecological function and community composition. 26611070–49194100 of high-quality reads, obtained from 87 soil samples, were assembled and annotated. Based on KEGG Automatic Annotation Server, the genes related to Dissimilatory Iron Reduction (DIR-g), Hydrocarbon Degradation (HD-g), Carbon Terminal Metabolism (CTM-g), Methanogenesis (MG-g), and Energy Supply (ES-g) were identified (Supplementary Table 1). The relative abundance of DIR-g (e.g., MtrABCD) was 0.6–2.0% in each sample, which was involved in the reduction of Fe minerals and could explain the fluctuation of Fe(II) content (Fig. 1b). The relative abundance of HD-g (e.g., bamA, alkB, oah) was 1.5–11.9%, and the relative abundance of CTM-g (e.g., korABCD, sdhABCD), which mainly performed TCA cycle, was 24.0–30.3%, showing the active hydrocarbon degradation and carbon metabolism in the oil-contaminated wetlands. Noticeably, the relative abundance of MG-g (e.g., mtaABC, cdhCDE, fwdABCD) reached 3.3–15.3%, supporting the report that oil fields and reservoirs were the methane emission sources8. Additionally, ES-g (atpABCD, nuoABCD, ndhABCD) mainly plays a role in oxidative phosphorylation, and their high abundance (53.5–67.3%) relative to other functional genes can provide energy guarantee for microbial metabolism.

Based on the Spearman correlation between functional genes and soil physicochemical indices (Fig. 1c), the significant positive correlation between DIR-g and Fe(II) was observed (p < 0.01), confirming the importance of DIR-g in performing the DIR process. The TPH content was positively correlated with HD-g (p < 0.05), indicating that oil leakage could promote the richness of microbial hydrocarbon degradation genes and enhance the degradation and utilization of hydrocarbon in wetlands. TPH and DIR-g also exhibited a significant positive correlation (p < 0.05), likely due to the utilization of hydrocarbons as carbon and energy sources during the DIR process14,35. Moreover, DIR-g could further stimulate the metabolism of terminal organics, leading to its positive correlation with CTM-g (p < 0.01), and the final products of this process are CO2 and H2O. These processes were evidenced by the significant negative correlation between TDC and HD-g, DIR-g, and CTM-g. Additionally, TDC was also negatively correlated with MG-g, which was due to the utilization of small molecule organic metabolites by methanogens. Therefore, the hydrocarbons produced TDC under the action of HD-g, which can form CO2 under the coupling of DIR-g and CTM-g, or CH4 through MG-g. According to the response of environmental factors to functional genes (Fig. 1d), Fe(II) presented an acute angle with DIR-g and HD-g, indicating the close link between DIR and hydrocarbon degradation. The negative angle of TDC with DIR-g and MG-g proved that dissolved organic matter served as a carbon source for both DIR and methanogenesis processes. With the regulation of functional genes in hydrocarbon metabolism, organic acids were gradually accumulated, thus making pH negatively correlated with all functional genes.

Samples from Shengli and Dagang Oilfield presented differentiated microbial community structures and co-occurrence networks (Fig. 1e–g and Supplementary Fig. 1). Microorganisms with the ability to degrade hydrocarbons (HD-m) dominated the microflora (Fig. 1f), with Marinobacter, Alcanivorax and Dietzia being the dominant HD-m for S1#–S44#, while Marinobacter, Salinimicrobium, Halomonas were dominant HD-m for D1#–D43# (see Supplementary Table 2 for details). According to the relative abundance at the genus level, 14 and 13 potential DIR microorganisms (DIR-m) (e.g., Pseudomonas, Shewanella) were identified from the top 500 in S1#–S44# and D1#–D43#, with their total average abundance reaching 14% and 7%, respectively. Geobacter, a typical DIR-m, had a low relative abundance (about 0.1–0.2%) in each sample, which may be due to the complex competitive environment in the soil. However, the key DIR-g (mtrA), HD-g (oah) and CTM-g (korABCD) were significantly positively correlated with Geobacter, indicating its key ecological function in mediating carbon and iron turnover in oilfield (Supplementary Table 3). Specifically, Geobacter was identified mainly as G. sulfurreducens PCA by BWA genome alignment, of which the relative abundance showed a significant positive correlation with the content of Fe(II) and TPH, while a negative correlation with pH (p < 0.01) (Fig. 1h), indicating that high TPH content stimulated the metabolism of G. sulfurreducens PCA to produce H+ accompanied by the reduction of Fe(III). Additionally, Methanothrix was the most dominant methanogens (MG-m) from top 500 genera in S1#–S44# and D1#–D43#, respectively. From Fig. 1c, soil physicochemical properties, especially TPH, Fe(II), and pH, shaped the microbial community structure. These results confirmed that the hydrocarbon degradation was common in the contaminated site of oil field, and DIR could interfere with the succession of microflora and carbon fate. Therefore, this inspires the artificial enhancement of DIR at contaminated sites in oil fields, by introducing Fe minerals and typical DIR-m, Geobacter (G. sulfurreducens PCA), which may simultaneously enhance hydrocarbon removal and mitigate carbon emissions in situ.

Carbon removal from the enhanced DIR system

Two representative soil samples from Shengli Oilfield (SS) and Dagang Oilfield (DS) were subjected to experiments of enhanced DIR. Changes in total solid organic carbon (TSC) content in soil samples from SS and DS indicate the transformation of organics and pollutants. Adding Fe minerals alone decreased TSC content by 0.26% and 0.14% (p < 0.05) in SS and DS, respectively, by day 7 (Fig. 2a, b). Inoculating electroactive bacteria G. sulfurreducens PCA further reduced TSC, with the treated groups (SP and DP) showing an 11% and 6% decrease, respectively, compared to the corresponding control groups (SC and DC) on day 7. The highest TSC reduction was observed in the treated groups SFP and DFP, which were supplemented with both Fe oxide and electroactive microbes, showing 14% and 9% reductions, respectively. TSC content in all treatment groups declined over time, particularly in the first 21 days. Compared to the control group, the enhanced TSC removal in SFP was 118% and 19% higher than in SF and SP, respectively, on day 21. For DFP, it was 210% and 110% higher than in DF and DP, respectively. By day 42, TSC was depleted to 3.38–3.75% in SS and 6.90–7.74% in DS. SFP and DFP showed the highest TSC removal rates of 19% and 17%, respectively.

Fig. 2: The carbon removal from the enhanced dissimilatory iron reduction (DIR) system.
figure 2

The change of soil total solid organic carbon (TSC) in different treatment groups for soil samples from Shengli Oilfield (a) and Dagang Oilfield (b), and the Spearman correlation analysis between pH, EC, TDC, IC, TDN, and TSC with carbon depletion from soil. The colors and numbers shown in (c) indicate the strength of the correlation (*p < 0.05, **p < 0.01). The contents of n-alkanes of C8-40 for soil samples from Shengli Oilfield (d) and Dagang Oilfield (f), and the contents of 16 preferentially controlled PAHs for soil samples from Shengli Oilfield (e) and Dagang Oilfield (g) in different treatment groups. Error bars represent the standard deviation (SD) of the mean across all data points in the figure.

Full size image

Physicochemical properties of the soil related to the carbon reduction process were analyzed using Spearman correlation (Fig. 2c). There was a significant negative correlation between EC and TSC (correlation coefficient = −0.796, p < 0.01) and a significant positive correlation between TDN (total dissolved nitrogen) and TSC (correlation coefficient = 0.181, p < 0.05). TDN was also positively correlated with TDC and negatively correlated with EC and pH. Additionally, IC had significant positive correlations with TDC, EC, and pH.

At the end of the reactor operation, n-alkanes of C8-40 (Alks) and 16 prioritized PAHs (Aros) were quantified to assess petroleum hydrocarbon removal by exogenous Fe minerals and electroactive microbes. The distribution of Alks and Aros in SS and DS varied due to different oil sources (Fig. 2d–g). In SS, the mass ratio of pristane (C19-1: 5.44 ± 0.13 mg kg−1) to phytane (C20-1:16.82 ± 1.08 mg kg−1) was about 1:3, with higher concentrations of C8, C15–C36, and C40, indicating high oil aging36,37. Exogenous Fe oxide and electroactive microbes reduced n-alkanes to varying degrees. The total n-alkane content of C8-40 decreased by 13.4 mg kg−1 in SF, 24.6 mg kg−1 in SP, and 39.4 mg kg−1 in SFP compared to SC (Fig. 2d, f). SFP achieved the best removal efficiencies for medium- and long-chain Alks (C21, C23–34, C40), with a 33% reduction of C40 compared to SC. In DS, the total Alk content (614–972 mg kg−1) was significantly higher than in SS, with the best removal observed in DFP. Compared to DC, DF, and DP, total Alks in DFP decreased by 357.5, 233.4, and 129.3 mg kg−1, respectively. C8 in SF, SP, and SFP was almost completely removed. The best removal of medium- and long-chain Alks (C16–C33) also occurred in DFP.

Aros, less bioavailable petroleum hydrocarbon components, showed a removal trend of SC > SF > SP > SFP in SS and DC > DF > DP > DFP in DS (Fig. 2e, g). Aros with four or more benzene rings, including BaA (Benzo(a)anthracene), BbF (Benzo(b)fluoranthene), and BghiP (Benzo(g,h,i)perylene), had enhanced degradation rates of 16–27% in SFP and 43–66% in DFP. Interestingly, Nap (Naphthalene) with two benzene rings slightly increased in treatment groups compared to controls, likely due to the oxidative and/or reductive degradation of complex organics.

Fe transformation in the enhanced DIR system

The modified five-step continuous extraction method was employed to analyze the response of iron (Fe) in different forms, including the water-soluble fraction (WaF), exchangeable fraction (ExF), carbonate-bound fraction (CaF), Fe-Mn oxides-bound fraction (FeF), organic-bound fraction (OrF), and residual fraction (ReF), to various treatments. Total Fe content in samples from SS and DS ranged from 45965–58597 mg kg−1 and 48252–60416 mg kg−1, respectively (Fig. 3). WaF, ExF, CaF, FeF, OrF, and ReF constituted 0.12–0.46%, 0.12–1.32%, 0.59–11.37%, 37.67–51.06%, 5.19–13.33%, and 30.00–45.22% of the total Fe content, respectively.

Fig. 3: The Fe transformation in the enhanced dissimilatory iron reduction (DIR) system.
figure 3

The changes of Fe content in forms of water-soluble fraction (WaF), exchangeable fraction (ExF), carbonate-bound fraction (CaF), Fe-Mn oxides-bound fraction (FeF), organic-bound fraction (OrF), residual fraction (ReF) on day 7 (a), 21 (b) and 42 (c) in different treatment groups. And the variation of Fe content in the form of Fe(II) and Fe(III) in different treatment groups on day 7 (d), 21 (e) and 42 (f). Error bars represent the standard deviation (SD) of the mean across all data points in the figure.

Full size image

Exogenous Fe oxide increased total Fe content in SF and SFP by 6760–12312 mg kg−1 compared to the control. This addition resulted in a 63–66% reduction in ExF content in SF, SP, and SPF on day 7 (Fig. 3a). In SF, CaF content decreased from 2916 ± 114 mg kg−1 to 432 ± 151 mg kg−1, while OrF and ReF content increased by 29% and 78%, respectively. Electroactive microbes increased CaF content by 62% in SP, and decreased FeF and OrF content by 10% and 3%, respectively. In SFP, OrF content decreased by 12%, while CaF increased by 626%. On day 21, SFP showed significant reductions in CaF and OrF by 2400 mg kg−1 and 485 mg kg−1, respectively, while FeF content increased by 3149 mg kg−1 (Fig. 3b). Over time, OrF content decreased across all treatments (Fig. 3c), with a 46% depletion in SPF and increases in CaF and FeF by 439% and 16%, respectively.

In DS samples, exogenous Fe oxide increased ReF content by 8475–8765 mg kg−1 in DF and DFP, while OrF content decreased by 32–52% and FeF content increased by 7–17%. Exogenous Fe oxide significantly reduced CaF content to 781–1612 mg kg−1 in DF, while the combined effect of Fe oxide and electroactive microbes increased CaF content to 2557–6865 mg kg−1.

Variations in total Fe(II) and Fe(III) content in soil samples were quantified to reveal the iron transformation process. Exogenous Fe oxide increased Fe(III) content by 7316 and 8687 mg kg−1. Compared with SC, exogenous electroactive microbes increased total Fe(II) by 823 mg kg−1 and 2641 mg kg−1 in SP and SFP, respectively, while reducing total Fe(III) on day 7 (Fig. 3d). On day 21, total Fe(III) content in SP and SFP increased by 3030 mg kg−1 and 837 mg kg−1, respectively, while Fe(II) content decreased slightly (Fig. 3e). A similar trend was observed in DS samples, with total Fe(II) increasing by 1082 mg kg−1 and 1847 mg kg−1 in DP and DFP, respectively, accompanied by the consumption of total Fe(III). On day 42, total Fe(III) in DFP increased by 2338 mg kg−1, while total Fe(II) decreased by 1198 mg kg−1 compared to day 7 (Fig. 3f). Surface morphology analysis of soil particles on day 42 revealed the formation of Fe minerals with a three-dimensional structure from exogenous lepidocrocite (γ-FeOOH) (Supplementary Fig. 2).

Carbon emissions from the enhanced DIR system

The CH4 content in the reactor gas was analyzed to determine the end-conversion path of soil organic carbon (SOC). On the 7th day, CH4 emission flux ranged from 2.92–4.91 mL h−1 in SS (Fig. 4a). Adding Fe oxide increased CH4 emission by 0.15 mL h−1 in SF, while emissions in SP and SFP were reduced by 19% and 38% compared to SC. By day 21, CH4 emissions significantly decreased across all groups, with SFP having the lowest at 0.73 mL h−1, 54–69% lower than other treatments. By day 42, CH4 emissions were nearly nonexistent. In DF, CH4 emission flux decreased by 0.15 mL h−1 on day 7 (Fig. 4b), and inoculating with electroactive microbes reduced it further by 1.90 mL h−1. On day 21, emissions in DP and DFP decreased by 32% and 45% relative to DC. Over time, CH4 emissions continued to decline, reaching 0.44–0.88 mL h−1 in DS.

Fig. 4: The carbon emissions from the enhanced dissimilatory iron reduction (DIR) system.
figure 4

The changes in CH4 emission fluxes for soil samples from Shengli Oilfield (a) and Dagang Oilfield (b) in different treatment groups on day 7, 14, 21, 28, 35, and 42.

Full size image

The microbial role in wetland habitats

Microbial community analysis was conducted to understand the biological mechanisms by which exogenous Fe oxide and Geobacter sulfurreducens PCA aid in restoring coastal oil-contaminated habitats. All microbial samples demonstrated sufficient sequencing depth, ensuring the accuracy of the constructed gene libraries (Supplementary Table 4). Both Fe oxide and G. sulfurreducens PCA treatments increased microbial richness in the early stages, with ACE and Chao indices rising by 18–32% and 22–37%, respectively. Additionally, electroactive microbes elevated the Shannon diversity index from 4.0 in SC to 4.5 in SP. However, SFP exhibited significant reductions in ACE, Chao, and Shannon indices, indicating selective species enrichment. Over time, the richness indexes (ACE, Chao) in SC and SFP increased, while in SF, they decreased.

In DS samples, ACE and Chao indices were reduced by 11% in DFP compared to DC. However, these indices increased over time in DC and DFP, consistent with trends in SC and SFP samples from SS. DF and DP showed decreased species richness on day 21, which later elevated. Notably, all diversity indices in DS were higher than those in SS, suggesting richer and more even species diversity with greater developmental variability.

The relative abundance of the top 10 phyla revealed that the predominant phyla in SS were Proteobacteria (28.8–72.9%), Desulfobacterota (8.4–30.9%), and Firmicutes (7.3–23.0%), accounting for 77.0–90.7% of the total, consistent with previous studies (Supplementary Fig. 3a)38,39. Actinobacteriota (1.9–6.5%), Chloroflexi (0.9–10.9%), and Bacteroidota (1.1–4.0%) also played stable roles. Compared with SC, both Fe oxide and G. sulfurreducens PCA reduced Proteobacteria abundance and increased Firmicutes and Desulfobacterota abundances on day 7. Over time, Proteobacteria abundance decreased in SC but increased in SP and SFP. Desulfobacterota and Firmicutes showed significant increases in SF, SP, and SFP.

In DS, the predominant phyla were Proteobacteria (16.1%–50.2%), Actinobacteriota (16.1%–50.2%), Chloroflexi (16.1%–50.2%), Firmicutes (16.1%–50.2%), and Desulfobacterota (16.1%–50.2%), accounting for over 90% abundance (Supplementary Fig. 3b and Supplementary Fig. 4). Proteobacteria were notably enriched in DFP. The abundances of Gemmatimonadota and Acidobacteriota were also higher in DFP. Over time, Proteobacteria abundance increased in DF and DP but decreased in DFP. Desulfobacterota maintained higher abundances in DP and DFP compared to DF.

At the genus level, key bacterial roles in Fe/C transformation were illustrated. Dominant genera in SC included MSBL7, KCM-B-112, Ketobacter, Guyparkeria, Marispirillum, and Phenylobacterium (Supplementary Fig. 3c). Exogenous Fe oxide significantly increased the abundances of Guyparkeria and Alcanivorax in SF. Electroactive bacteria inoculation increased the abundances of MSBL7, Guyparkeria, Halomonas, Desulfitibacter, and Truepera in SP and SFP. On day 21, many genera decreased in abundance in SFP, while Pseudomonas, Ketobacter, Desulfitibacter, and Allorhizobium increased. By day 42, Guyparkeria regained dominance in SFP.

In DS, the microbial community shifted compared to SS. Dominant genera in DC included Nocardioides, KCM-B-112, Mycobacterium, Defluviicoccus, and Streptomyces (Supplementary Fig. 3d). Fe oxide addition enriched Nocardioides, Mycobacterium, Defluviicoccus, and Dietzia on day 7. Highest abundances in DFP included norank_f_norank_o_SBR1031, Mycobacterium, Defluviicoccus, Ketobacter, and Ramlibacter. By day 21, Mycobacterium, Defluviicoccus, and Ketobacter decreased in DFP, while Cavicella, Desulfovirga, and Allorhizobium increased. On day 42, norank_f_norank_o_SBR1031 and Nocardioides returned to dominance in DFP.

The growth conditions of the inoculated Geobacter (G. sulfurreducens PCA) were also investigated (Supplementary Fig. 5). Co-addition of G. sulfurreducens PCA and Fe oxide benefitted Geobacter activity and growth. The highest OTU (2667) of G. sulfurreducens PCA was in SFP on day 7. OTUs decreased over time in SFP but showed an increasing trend on day 21 in SP before decreasing again. In DS, OTUs of G. sulfurreducens PCA were generally lower than in SS, following similar trends.

Microbial co-occurrence networks evaluated potential interspecific interactions among functional microflora (Fig. 5). Functional bacteria, including TPH degradation, nitrogen-transforming, sulfur-cycling, and iron-reducing/electroactive bacteria, were screened from the top 30 genera in SS and DS (Supplementary Fig. 3 and Supplementary Table 5). Cooperation of Fe oxide and G. sulfurreducens PCA increased TPH degrading, sulfur-cycling, and iron reducing bacteria proportions in SFP networks (Supplementary Table 6). TPH degrading and nitrogen transforming bacteria proportions increased significantly in DFP. Functional bacteria showing positive correlations with iron-reducing/electroactive bacteria increased by 4, 14, and 7 in SF, SP, and SFP compared to SC. Positive correlations between functional microorganisms in DFP reached 125, higher by 39, 50, and 42 compared to DC, DF, and DP, respectively (Supplementary Table 7). It should be noted that while DIR and Geobacter sulfurreducens inoculation show promising short-term effects, their long-term ecological impacts remain uncertain and require further study.

Fig. 5: The co-occurrence networks of functional microflora from the top 30 genera in samples of SS and DS.
figure 5

Key nodes include the potential TPH degrading bacteria (C1-32) with red color, nitrogen-transforming bacteria (N1−14) with purple color, sulfur-cycling bacteria (S1−10) with orange color, iron-reducing bacteria (Fe1-5) with blue color and function bacteria that was yet unspecified (S1−10) with gray color. The size of nodes represents the number of connections (i.e., degree) based on the significant correlation of the number of OTU changes (p < 0.05). Blue edge indicates positive correlation and orange edge represents the opposite.

Full size image

Discussion

Fe, the second most abundant metal in coastal oil-polluted wetlands, has active redox properties, driving soil organic conversion through the iron cycle23,25. Metagenomic surveys of oil-contaminated sites highlight the potential of DIR to degrade recalcitrant hydrocarbons and alter terminal electron transfer pathways. In laboratory experiments, adding electroactive microbes and Fe to soil accelerated the depletion of total carbon (TSC), with a 14% and 9% reduction in SFP and DFP treatments, respectively, on day 7. In anoxic coastal wetlands, Fe(II) oxidation is inhibited, while Fe(III) reduction is driven by iron-reducing/electroactive bacteria encoding genes like MtrABCD and using substances such as humic acid and sulfide40,41,42. Humic acid may act as an electron shuttle or be oxidized to provide electrons, while sulfide directly reduces Fe3+ or forms iron-sulfur complexes20,43. The addition of exogenous iron oxides and electroactive bacteria led to significant SOC consumption for Fe(III) reduction, with H2 and NH4+ also serving as electron donors44. As organic carbon decreased, EC increased, showing a significant negative correlation with TSC (p < 0.01). The positive correlation between total dissolved nitrogen (TDN) and TSC (p < 0.05) indicated nitrogen utilization by the microbes. Electroactive microbes enhanced hydrocarbon removal, including C8, and accelerated the degradation of complex pollutants. Extracellular electron transfer by these bacteria relieved electron accumulation in anaerobic soils, stimulating continuous pollutant degradation16,45. Additionally, reduced Fe(II) facilitated phenol oxidase activity, aiding the removal of soluble aromatic compounds46. While this study primarily focuses on oxidative pathways mediated by DIR, the potential contribution of reductive degradation, particularly by Geobacter sulfurreducens, should not be overlooked21,22.

Iron (Fe) in soil exists in various forms: WaF, ExF, CaF, FeF, OrF, and ReF, each with different mobility and bioavailability. Soil samples from SS and DS showed similar total iron content (45965–60416 mg kg−1), with higher proportions of FeF, OrF, and ReF. Inoculating soils with electroactive microbes significantly reduced OrF content in SP and SFP by 3% and 12%, respectively, and increased CaF content by 62% and 626%. The integration of minerals and organic carbon is crucial for carbon sequestration and energy conversion in soil systems47. Active sites on iron oxides can adsorb dissolved organics, forming stable combinations through ligand exchange, ion exchange, and van der Waals forces48,49,50. Moreover, the active Fe(III) from low crystalline iron oxides was demonstrated to form co-precipitation with organic carbon51,52. The DIR process accelerated OrF consumption and CaF formation due to exogenous iron minerals and electroactivity. This is attributed to: (1) the consumption of organic matter bound to iron minerals and the production of carbonate bound to iron minerals, and (2) the reduction of Fe(III) to Fe(II), weakening the bond between organic carbon and iron oxides. OrF content decreased over time, with final depletion rates in SFP and DFP at 46% and 36%, respectively. Notably, the stability of CaFe under varying conditions (e.g., pH and redox condition) is key to remediation, requiring studies on Fe mineral dissolution kinetics in wetlands53. Additionally, ExF content decreased in reactors with iron minerals and electroactive microbes. Total Fe(II) content increased by 2641 mg kg−1 in SFP and 1847 mg kg−1 in DFP due to DIR, consuming Fe(III). SEM analysis showed iron minerals with stereoscopic structures formed from the added ferric oxyhydroxides by electroactive bacteria. Fe(II)-containing secondary minerals (e.g., magnetite) formed from Fe(III) reduction can enhance extracellular electron transfer and direct interspecies electron transfer (DIET), promoting hydrocarbon removal54,55. While magnetite-facilitated extracellular electron transfer suggests potential for DIET, this mechanism was not directly assessed in our study and remains speculative, requiring targeted experiments for confirmation. Reoxidation of Fe(II) to Fe(III) in later stages indicated iron ions acted as electron shuttles, accelerating hydrocarbon consumption.

The long turnover time of SOC results in soil becoming the largest carbon pool on land, 4.4 and 3.2 times than that of biological and atmospheric carbon pools, respectively56,57. The discovery of a novel archaea Ca. Methanoliparum, with complete metabolic pathways for hydrocarbon degradation and methane production, validates the great potential of coastal oil-contaminated soils for greenhouse gas (GHG) emissions8. In this study, the production of CH4 from the reactor was focused because the greenhouse gas contribution of an equal mass of CH4 over 100 years is 28 times than that of CO2 reported by the IPCC18. For reactors applied with SS and DS, inoculation of electroactive microbes resulted in a decrease in CH4 emission flux in the short term. Moreover, under the coupling effect of exogenous iron oxide and electroactive microbes, the CH4 emission flux was reduced by 38% and 40% in SFP and DFP, respectively. Phylogenetically diverse methanogens living in oil-contaminated soils produce CH4 mainly through hydrogenotrophic (ΔG0 = −33 kJ mol−1 hydrogen) or aceticlastic (ΔG0 = −36 kJ mol−1 acetate) methanogenesis17,58. Anaerobic respiration utilizing alternative terminal electron acceptors, such as nitrates and iron minerals (ΔG0 = −40 to −234 kJ mol−1 hydrogen and −69 to −841 kJ mol−1 acetate), have been shown to prevail in competition with the methanogenesis pathway59. So the electroactive microbes may inhibit CH4 emission by transferring extracellular electrons to solid-phase iron oxides to stimulate alternative respiration. Inevitably, more CO2 will be produced as the dissimilated iron reduction process changes the terminal electron transfer path. However, given the significantly higher GHG contribution of CH4 than CO2, stimulating the conversion of dissolved organics to CO2 is an effective way to suppress the GHG effect. In addition, the synthesis of conductive iron minerals can transfer extracellular electrons to a larger electron receiving pool (e.g., quinone groups) to further inhibit CH4 emission20. Therefore, DIR facilitates a biogeochemical coupling mechanism wherein carbon flow is directed toward CO2 rather than CH4 emissions, driven by Fe(III) reduction and electron transfer pathways (Fig. 6). Moreover, under anaerobic conditions, CH4 oxidation coupled with Fe3+ reduction can generate bicarbonate (HCO3−) instead of CO2, thereby mitigating CH4 emissions while limiting CO2 release60. Although not directly assessed in this study, this mechanism could significantly impact greenhouse gas dynamics in oil-polluted wetlands. With the consumption of available organics, the CH4 content from the reactor gradually decreased, and the highest organic carbon removal with the lowest CH4 emission was finally achieved in SFP and DFP.

Fig. 6: The proposed green way to achieve hydrocarbon removal and carbon emission mitigation through in situ reinforcement of dissimilatory iron reduction (DIR) in this work.
figure 6

The red and blue lines represent processes enhanced and inhibited by DIR, respectively. TPHs and OM denote the total petroleum hydrocarbons introduced into the soil from petrochemical activities and the native soil organic matter, respectively, while TDC (total dissolved organic carbon) represents the easily metabolizable organic carbon derived from them.

Full size image

To sum up, the metagenomic survey of 87 samples from the Shengli Oilfield and the Dagang Oilfield confirmed the role of DIR in microflora succession and carbon fate. The coastal oil-contaminated wetland habitat was simulated at the laboratory scale, and the enhanced system of electroactive bacteria-mediated DIR was constructed. Compared with natural conditions, the simultaneous addition of geochemical elements and electroactive microbes enhanced the electron transfer between microorganisms and accelerated the hydrocarbon removal. The anaerobic microbes in coastal wetlands constructed an interspecies interaction system with electroactive microbes through the iron-mediated electron transfer process, which alleviated the accumulation of small organic acids. The goal of accelerating the conversion of petroleum hydrocarbons while greatly reducing CH4 emissions has been achieved for different types of coastal wetland contaminated soils. Unlike previous works that focused on DIR as a standalone mechanism, we explored the potential coupled transformation of Fe oxides into stable forms like conductive mineral, which supports long-term pollutant management. Therefore, a method to regulate the conversion of petroleum hydrocarbons and slow down carbon emissions has been proposed, which could also shed light on the green ecological restoration of oil-contaminated soil/wetland. This study was conducted in a controlled lab environment, challenges such as varying redox conditions, microbial community dynamics, plant-mediated oxygen supply, and competition for electron acceptors from sulfate reduction in wetlands require future field trials to validate these findings. For instance, in natural wetlands, plant-root oxygenation creates aerobic zones, affecting Fe3+ reduction and microbial communities, challenging lab-to-field replication.

Methods

Collection and pretreatment of soil samples

Typical coastal wetland oil-contaminated soils were collected from the 0–20 cm soil surface around the beam-pumping unit in Shengli Oilfield (Shandong, China) and Dagang Oilfield (Tianjin, China). Forty four samples (numbered S1#–S44#) from Shengli Oilfield and 43 samples (numbered D1#–D43#) from Dagang Oilfield with differentiated pollution intensity were investigated for microbial ecology (see Fig. 1a for coordinates). Experiments on enhanced DIR were conducted using two representative soil samples collected from Shengli Oilfield (SS) and Dagang Oilfield (DS). After natural air-drying in a ventilated and cool environment, soil samples were stripped of grit and plant debris before being grounded through a 2 mm sieve.

Culture of electroactive microorganism Geobacter sulfurreducens PCA

To obtain a pure culture of the typical electroactive microorganism G. sulfurreducens PCA (ATCC-51573), the laboratory frozen strain (firstly isolated from surface sediments of a hydrocarbon-contaminated ditch in Norman, Okla) was activated and cultured in NBAF medium at 30 °C under strictly anaerobic conditions57,61. Before inoculating the bacterial source, the medium was bubbled with an 80% N2–20% CO2 gas mixture to remove dissolved O2. The cultured medium was centrifuged to collect electroactive cells, which were resuspended in a sterile medium until a final inoculum of optical density of 0.5 (OD600) was obtained.

Operation of the simulated wetland systems

The coastal wetland system was simulated in a plexiglass cylindrical reactor (10 cm in diameter and 10 cm in height) with a valve for gas collection. For soil samples from Shengli Oilfield, four treatments were performed as follows: treatment 1 contained 1.5 kg of oil-contaminated soil (labeled as SC); treatment 2 contained 1.5 kg of oil-contaminated soil and 1% Fe oxide (labeled as SF); treatment group 3 contained 1.5 kg of oil-contaminated soil and 150 mL of electroactive inoculum (labeled as SP); treatment group 4 contained 1.5 kg of oil-contaminated soil, 1% Fe oxide and 150 mL of electroactive inoculum (labeled as SFP). The Fe oxide used in this study was lepidocrocite (γ-FeOOH) with an average particle size of ~60 nm. Similarly, soil samples from Dagang Oilfield were processed accordingly, including DC, DF, DP, and DFP. After the soil of each group was fully mixed with the additives, the reactor was supplemented with seawater to form a 2 cm water seal to simulate the actual coastal wetland environment. Soil samples from three parallel reactors in different treatment groups were collected weekly (on days 7, 14, 21, 28, 35, and 42 d) for analysis of soil physicochemical properties and microbial community. Additionally, the exhaust gas was collected by the gas bag periodically to investigate the production of CH4.

Determination of soil properties and carbon emissions

The collected soil samples were freeze-dried at −60 °C and passed through a 60-mesh sieve for the determination of soil physicochemical properties. The pH and EC of soil samples was determined by a pH meter and a conductivity meter, respectively, under a soil to water ratio of 1:5 (w/v). The Fe in different forms, including water-soluble fraction (WaF), exchangeable fraction (ExF), carbonate-bound fraction (CaF), organic-bound fraction (OrF), Fe-Mn oxides-bound fraction (FeF), Residual fraction (ReF) were quantified by the modified five-step continuous extraction method62. Briefly: WaF: Extracted with deionized water to isolate Fe in the most labile form. ExF: Isolated using 1 M MgCl2 at pH 7.0 to target exchangeable Fe. CaF: Dissolved with 1 M NaAc at pH 5.0 to quantify Fe associated with carbonate minerals. FeF: Released using hydroxylamines hydrochloride in 25% acetic acid to detect Fe bound to Fe-Mn oxides. OrF and ReF: Extracted with hydrogen peroxide and nitric acid to measure Fe bound to organic matter and silicate minerals, respectively. Each fraction was quantified via UV–Vis spectrophotometry using the o-phenanthroline method. The TDC, total dissolved nitrogen (TDN), and inorganic carbon (IC) were detected by a soluble carbon and nitrogen analyzer (Multi N/C 3100, Analytik Jena Ltd., Germany). The soil total solid organic carbon (TSC) was determined by the combustion method by a TOC instrument with a non-dispersive infrared detector. The soil Fe2+ and Fe3+ were quantitative by the o-Phenanthroline spectrophotometric method using a 96-microplate reader (SPARK 10 M, TECAN Ltd., Switzerland)63. TPH content was quantified gravimetrically by extracting hydrocarbons with dichloromethane and evaporating the solvent at 42 °C. To reduce interference from volatile organic compounds, samples were pre-heated at 40 °C under vacuum to remove non-petroleum volatiles prior to extraction. The gravimetric measurements were calibrated using known TPH standards to ensure accuracy. The Alkyl and Aromatic compounds were separated by the differential adsorption of chromatographic columns, and the contents of C8-C40 n-alkanes (Alks) and 16 priority-controlled PAHs (Aros) were determined by GC-MS as previous method 64,65,66. Gas CH4 content was determined using a gas chromatograph with a thermal conductivity detector.

Identification of microbial communities

Metagenomics was applied to identify the microbial ecological function and community composition of 87 samples from oil-contaminated sites (See Supplementary Method 1 for details). Specifically, raw sequencing reads were quality-checked and trimmed to remove low-quality bases and adapters. The clean reads were assembled de novo with MEGAHIT. The identified function genes of the samples related to Dissimilatory Iron Reduction (DIR-g), Hydrocarbon Degradation (HD-g), Carbon Terminal Metabolism (CTM-g), Methanogenesis (MG-g), and Energy Supply (ES-g) were annotated based on KEGG database. Specifically, the Burrows-Wheeler-Alignment Tool was used to align clean data reads from soil samples to the genome of Geobacter sulfurreducens PCA to determine their relative abundance. The 16S rRNA sequencing was performed for the soil samples from DIR reinforcement system (See Supplementary Method 2 for details).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.