Arsenic-methylating microbial community in sediment along the water flow is correlated with the distance to a low-temperature hot spring

Arsenic (As), a well-known human carcinogen, is ubiquitously distributed in natural environments. Prolonged interactions between geothermal fluids and reservoir host rocks, or inputs from deep magmatic components, can result in the accumulation of dissolved arsenic in geothermal systems (Wang et al. 2018). Arsenic concentrations in hot springs worldwide currently range from 0.01 to tens of milligrams per liter (mg/L) (Smedley & Kinniburgh 2002). The inflow of geothermal waters can result in elevated As concentrations in surface and nonthermal groundwaters. This accumulation has led to serious concerns, especially regarding the waters utilized for drinking and irrigation (Morales-Simfors & Bundschuh 2022). Microorganisms play a significant role in the biogeochemical cycling of arsenic. To date, studies involving microbiology in arsenic speciation in geothermal waters have focused primarily on arsenite [As(III)] and arsenate [As(V)]; some research has been conducted on inorganic thioarsenates and methylated thioarsenates (Hamamura et al. 2009; Yan et al. 2022; Yin et al. 2022; Wang et al. 2023). However, previous studies have shown notable concentrations of methylarsenate (MMA)/dimethylarsenate (DMA) in geothermal springs at various locations, including the Champagne Pool (New Zealand), TengChong (China), and Yellowstone National Park (USA) (Planer-Friedrich et al. 2007; Zhang et al. 2008; Hug et al. 2014). These organic forms substantially contribute to the overall arsenic content in geothermal waters worldwide (Wang et al. 2018). The microbiological and geological mechanism underlying the production of MMA/DMA in geothermal springs has yet to be fully elucidated.

The process of arsenic methylation is catalyzed by arsenite S-adenosylmethionine methyltransferase (ArsM) enzymes, the arsM gene of which serves as a molecular marker for exploring the diversity of arsenic-methylating microbes (Chen & Rosen 2023). Intracellularly, ArsM catalyzes the conversion of As(III) into MMA and DMA, further resulting in the generation of volatile trimethylarsine (TMAs). In this biochemical reaction, As(III) acts as a substrate, and S-adenosylmethionine acts as a methyl donor (Qin et al. 2009; Zhu et al. 2014; Roy et al. 2020). Converting inorganic arsenic to methylarsenic is commonly associated with a decrease in toxicity and is thus considered a detoxification mechanism (Chen & Rosen 2020). In addition, these organic forms of arsenic (MMA and DMA) can undergo demethylation processes, leading to the conversion of inorganic arsenic compounds, specifically As(III) (Zhu et al. 2014). Arsenic methylation is a phenomenon observed in various organisms (Ali et al. 2021). Several prokaryotic strains capable of arsenic methylation, such as Streptomyces sp. GSRB54 (Kuramata et al. 2015), Pseudomonas alcaligenes NBRC14159 (Zhang et al. 2015a), Arcticibacter tournemirensis R1 (Zhou et al. 2022), Clostridium sp. BXM (Wang et al. 2015), Paraclostridium sp. EML (Viacava et al. 2022), Bacillus sp. CX-1 (Huang et al. 2018), Arsenicibacter rosenii SM-1 (Huang et al. 2016), and Streptomyces vietnamensis (Viacava et al. 2020), have been previously isolated and thoroughly studied. These strains are present in both anaerobic and aerobic environments, primarily in paddy soil and other environments, such as freshwater, mines, sludge, wetlands, and composting factories (Zeng et al. 2018; Qiao et al. 2023). Extensive research has been conducted on nongeothermal environments. However, investigations on arsenic methylation in geothermal environments are scarce. Only one report concerning Yellowstone National Park, USA, has documented the involvement of an extremophilic eukaryotic alga of the order Cyanidiales in this process (Qin et al. 2009). Recently, methanogens were proposed to drive arsenic methylation in sulfide-rich hot springs in Tengchong (Wang et al. 2023). However, whether the mechanism of methyl arsenic formation is similar to that of other geothermal environments remains unclear. Here, we collected sediment samples around Yanchi hot spring, which maintains an average temperature of 39 °C throughout the year. The primary aims of the present study were (i) to identify the activity and diversity of the microorganisms involved in arsenite methylation in subgeothermal environments and (ii) to evaluate the effects of distance from the hot spring on the activity and diversity of arsenite-methylating microorganisms.

The temperature was measured using a portable thermometer (LaMotte, Chestertown, MD, USA). The central part of the hot spring maintained a steady temperature of 42 °C, whereas the temperature of the water discharged from the spring was 39 °C. The aforementioned temperatures are year-round. Nine samples were obtained from three distances, and three replicate samples were collected within a 0.2-m range within each of these distances. The sampling locations were 2 m (referred to as YC1), 6 m (referred to as YC2), and 12 m (referred to as YC5) from the hot spring (Figure 1(c)). All the samples were collected at a depth of 15–25 cm below the ground surface. These samples were then promptly placed into sterilized tubes and kept chilled with ice during transport. All the samples reached the laboratory for further analysis within a 12-h timeframe.

Concentrations of arsenic, including total arsenic, soluble arsenic, MMA, and DMA, were assessed utilizing high-performance liquid chromatography coupled with atomic fluorescence spectrometry (HPLCAFS9600). Specifically, an Agilent 1220 HPLC system equipped with a Hamilton PRP-X100 ion exclusion column was utilized following the methods described in previous studies (Zeng et al. 2018). The mobile phase consisted of 15.0 mM (NH₄)₂HPO₄, and the flow rate was set at 1.0 mL/min. The retention times for As(III), DMA, MMA, and As(V) were determined to be 2.58, 3.59, 4.62, and 9.62 min, respectively (Georgiadis et al. 2006). An ion chromatography system (DX-120, Dionex, USA) was used to measure the anion concentrations (Broderick et al. 2005). For the quantification of metal ion concentrations, ICP-AES (inductively coupled plasma-atomic emission spectrometry) was performed using an IRIS Intrepid II XSP instrument manufactured by Thermo Fisher, USA (Chen et al. 2017). The total organic carbon (TOC) and total nitrogen (TN) contents were determined using a carbon/nitrogen analyzer (PRIMACS™ SNC-100, The Netherlands).

Microcosm incubation analyses were also conducted to assess the arsenic methylation activity of the microbial community in the sediment samples. These experiments were performed in triplicate to ensure reliability. In the experimental group, each active microcosm consisted of 50.0 g of sediment and 50.0 mL of mineral salt medium (Bahar et al. 2012). The experimental group was divided into two subgroups. One subgroup was left untreated for the detection of microbial methylation activity under natural conditions. The other subgroup was treated with 0.1 mM As(III) and 25.0 mM lactate in a 100-mL container. As(III) served as the substrate, and lactate acted as the carbon source for the methylation reaction. The control mixtures were subjected to autoclaving. All microcosms were incubated at 39 °C without shaking. Approximately 1.5 mL of the cultures were collected from each vial on days 7 and 21 of the incubation period and subjected to quantitative analysis of MMA and DMA concentrations using HPLCAFS9600.

Total DNA extraction was performed from soil samples (0.5 g) that had undergone 7- and 21-day incubations using the Soil DNA Extraction CZ Kit (DC306, FINDROP, China). The DNA concentrations were detected with a NanoDrop One spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The quantification of the arsM gene abundance was achieved with the CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) (Zhao et al. 2013). The 20-μL qPCR mixture consisted of 10 μL of SYBR™ Green Master Mix (Thermo Fisher Scientific, USA), 0.4 μL of passive reference dye, 0.4 μL of each primer (10 mM), 1.5 μL of template DNA, and 7.3 μL of sterile water without DNA. The qPCR program included an initial denaturation step at 94 °C for 5 min, followed by 40 cycles of amplification (94 °C for 15 s, 60 °C for 30 s, and 72 °C for 30 s). A linear standard curve with R² > 0.99 was obtained in the quantitative assay. The amplification efficiency was within the range of 90–110%. A single peak was observed in the melt curve analysis. Each sample was analyzed in triplicate to ensure reproducibility. The primers used for arsM gene amplification are presented in Table 1.

The V3–V4 hypervariable regions of the arsM genes were amplified with the arsMF and arsMR primers (Table 1) (Jia et al. 2013). The amplicons were purified and quantified. The V3–V4 regions were subsequently sequenced using the Illumina NovaSeq PE250 platform (Magigene Biotechnology Co., Ltd, China). The raw data obtained were subjected to processing, which included demultiplexing based on barcode sequences and subsequent quality filtering (Wang et al. 2023). Quantitative Insights Into Microbial Ecology (QIIME, V1.9.0) software was utilized to filter and assemble the raw sequences. Operational taxonomic units (OTUs) were defined at a 97% similarity threshold using UPARSE (version 7.1) and checked for chimeric sequences with UCHIME (Wang et al. 2023).

Using the BLAST server, a search of the GenBank database was launched to identify homologous protein sequences (http://blast.ncbi.nlm.nih.gov/). We further conducted multiple sequence alignments with ClustalW2 software (https://www.ebi.ac.uk/Tools/msa/clustalw2/). A phylogenetic tree was subsequently constructed through neighbor-joining analysis with MEGA 11.0.13 (https://megasoftware.net/dload_win_beta) following the methodology outlined in a prior study (Zhou et al. 2022). The reliability of the phylogenetic tree was determined by performing 1000 bootstrap replications.

Table 2 presents a summary of the geochemical properties of the sediment samples. The total arsenic concentration in the sediments ranged from 0.090 to 0.128 mg/kg, which is significantly lower than the concentrations previously found in sediments from other hot springs, such as Tengchong (1.6–6406 mg/kg) (Yan et al. 2022), the Tibetan Plateau (101.9–263.7 mg/kg) (Zhang et al. 2017), and Latin America (0.7–111 mg/kg) (Morales-Simfors et al. 2020). Additionally, we established that the sediments contained TOC ranging from 5.56 to 18.70 g/kg and TN ranging from 0.88 to 1.53 g/kg. These components play pivotal roles as essential energy sources, providing both carbon and nitrogen for microbial growth. The sediment also contained sulfate (110.00–136.00 mg/kg), ammonium (2.44–3.66 mg/kg), and nitrate (7.55–45.80 mg/kg).

Table 2

Characteristics of the sediment samples from the Yanchi hot spring

Parameters .	Sediment samples .
	YC1 .	YC2 .	YC5 .
pH	8.33	8.35	8.49
Total As (mg/kg)	0.112	0.128	0.090
Soluble As (μg/kg)	0.042	0.057	0.028
TOC (g/kg)	5.56	18.70	6.46
TN (g/kg)	0.88	1.53	0.52
(mg/kg)	26.80	45.80	7.55
(mg/kg)	3.53	3.66	2.44
(mg/kg)	136.00	151.00	110.00
Cl− (g/kg)	0.666	0.732	0.355
K (g/kg)	17.3	21.7	12.5
Ca (g/kg)	96.4	94.6	173.0
Na (g/kg)	2.52	3.00	2.21
Mg (g/kg)	8.82	8.95	8.79
Fe (g/kg)	21.6	25.2	17.0
Mn (g/kg)	335	349	259

Parameters .	Sediment samples .
	YC1 .	YC2 .	YC5 .
pH	8.33	8.35	8.49
Total As (mg/kg)	0.112	0.128	0.090
Soluble As (μg/kg)	0.042	0.057	0.028
TOC (g/kg)	5.56	18.70	6.46
TN (g/kg)	0.88	1.53	0.52
(mg/kg)	26.80	45.80	7.55
(mg/kg)	3.53	3.66	2.44
(mg/kg)	136.00	151.00	110.00
Cl− (g/kg)	0.666	0.732	0.355
K (g/kg)	17.3	21.7	12.5
Ca (g/kg)	96.4	94.6	173.0
Na (g/kg)	2.52	3.00	2.21
Mg (g/kg)	8.82	8.95	8.79
Fe (g/kg)	21.6	25.2	17.0
Mn (g/kg)	335	349	259

View Large

Given the limited TOC content in the sediment, we investigated microbial arsenic methylation activity by introducing additional organic carbon. As depicted in Figure 2, methylarsenic was detected in samples YC1 and YC2 on both days 7 and 21, at which time the MMA levels were significantly greater than the DMA levels. In YC1, the MMA concentrations were 63.5 and 169.0 μg/L on days 7 and 21, respectively, whereas the corresponding DMA concentrations were only 36.5 and 38.1 μg/L, respectively. Similarly, in YC2, the MMA content was 184.0 and 106.0 μg/L on days 7 and 21, respectively, with DMA concentrations of only 36.0 and 41.9 μg/L on the corresponding days. However, in contrast to YC1 and YC2, YC5 had higher DMA levels than did MMA, which were detected only on day 7, with concentrations of 26.1 μg/L for MMA and 204.0 μg/L for DMA. Overall, all three samples showed significantly stronger methylarsenic activity when supplemented with exogenous carbon and arsenic than under natural conditions, indicating the activation of arsenic methylation microorganisms. Notably, the methylarsenic content in YC1 increased over time, whereas that in YC2 and YC5 decreased.

To gain a comprehensive understanding of the microbial basis contributing to our results, we examined the community of arsM-carrying prokaryotic microorganisms found in both the natural sediment and the treatment sediments collected on days 7 and 21. We identified a total of 2,091 OTUs corresponding to arsM-carrying bacteria. These OTUs were classified into 22 phyla, 183 families, and 586 species. Additionally, we discovered 30 OTUs of arsM-carrying archaea, which were assigned to two phyla, five families, and nine species (Table 3). A rarefaction analysis conducted using QIIME demonstrated that these sequences covered more than 99.84% of the arsM-carrying microorganisms in the sediments (Table 3; Figure S1).

At the genus level, there were slight differences in the community of arsM-carrying microorganisms between YC1 and YC2 (Figure 4(b)). In YC1, the genera with the highest abundance on day 7 were Bacillus, Bilophila, and Geobacillus, whereas on day 21, the genera with the highest abundance were Parageobacillus, Desulfosporosinus, Sorangium, and Rhizobium. In YC2, the genera with the highest abundance on day 7 were Geobacillus, Caenispirillum, Caldilinea, Sulfuricella, Methanosarcina, Thioalkalivibrio, and Pseudomonas, whereas on day 21, the genera with the highest abundance were Rhodopseudomonas, Alkaliphilus, Ktedonobacter, Thiobacillus, Cellulomonas, Mycobacterium, and Pseudomonas.

In contrast, the arsM gene community in YC5 was completely different from that in YC1 and YC2. Acidobacteriota dominated at the phylum level in YC5 on day 7, accounting for 75.0% of the community, followed by Actinobacteriota (7.3%) and Firmicutes (6.6%). Subsequently, on day 21, the abundances of Actinobacteriota and Proteobacteriota increased significantly, composing 60.5 and 14.0%, respectively, of the community (Figure 4(a)). YC5 exhibited a pattern dissimilar from that of YC1 and YC2 at the genus level (Figure 4(b)). On day 7, reads from the unassigned genus emerged, and further sequence analysis identified these reads as members of the Acidobacteriaceae family within the Acidobacteriota phylum. Notably, on day 21, the proportion of Mycobacterium dramatically increased to 60.1%, a tenfold increase compared to its 5.9% abundance on day 7 (Figure 4(b)). Several other genera with relatively smaller proportions included Geobacillus (1.7%), Rhodopseudomonas (2.4%), Desulfotomaculum (3.1%), Desulfococcus (6.0%), and Blastopirellula (3.1%).

Additionally, we discovered five best-matched arsM-carrying methanogens, namely, Methanosphaerula palustris, Methanosarcina barkeri, Methanoculleus marisnigri, Methanosarcina mazei, and Methanosarcina acetivorans, in the three samples (Figure S3).

Arsenic biogeochemical processes include oxidation, reduction, arsenic sulfide reactions, and methylation. However, to date, research on geothermal environments has been predominantly concentrated on the former three processes. Studies on arsenic methylation have largely centered on its role as a detoxification mechanism in prokaryotic and eukaryotic microbes (Chen et al. 2019; Viacava et al. 2020). Nevertheless, limited research has been performed on arsM gene activity and diversity. Furthermore, these studies have been predominantly conducted in nongeothermal environments, such as rice paddies. With an average temperature of 39 °C throughout the year, the Yanchi hot spring provides ideal conditions for studying arsenic-methylating microorganisms in a geothermal environment. In the present study, we found that the sediment along the outflow channel of the hot spring exhibited arsenic methylation capacity. Additionally, arsenic-demethylating species were detected, primarily favoring a niche farther away from the hot spring hole. These observations contribute to advancing our understanding of microorganism-mediated arsenic methylation in geothermal settings.

YC1 and YC2, the two sampling sites closest to the hot spring hole, had significantly greater MMA levels than did DMA on both days 7 and 21 when lactate and arsenite were supplemented. These results were likely due to the similar compositions of the arsM microbial communities (Zhao et al. 2013; Zhai et al. 2017). Actinobacteriota, Firmicutes, and Proteobacteriota were dominant in the bacterial community harboring arsM in YC1 and YC2. Moreover, Methanosarcina exhibited a relatively greater abundance on day 7 in the YC1 and YC2 samples. Methanosarcina sp. can methylate As(III), predominantly generating MMA and DMA to a lesser extent, which cannot be demethylated. Specifically, the arsenic methylation abilities of Methanosarcina mazei Gö and Methanosarcina acetivorans C2A have been extensively studied (Thomas et al. 2011; Wang et al. 2014), and their relatives were also detected in this study (Figure 4, Figure S3).

The difference in methylarsenic production between the YC1 and YC2 microcosms on day 21 compared to day 7 can be attributed to specific microbial enrichment. In YC1, the relative enrichment of Bacillus on day 21 led to greater levels of methylarsenic bacteria than those on day 7. A previous study showed that the purified S-adenosylmethionine methyltransferase of Bacillus sp. CX-1 efficiently methylated As(III), with DMA produced as the final product (Figure S2) (Huang et al. 2018), indicating the accumulation of methylarsenic rather than its volatilization. In contrast, the decrease in methylated arsenic observed in YC2 on day 21 compared to day 7 may be due to the greater relative enrichment of Rhodopseudomonas. Rhodopseudomonas species have been reported to promote arsenic methylation and volatilization, as they have been identified as active arsenite-methylating organisms in paddy soils (Figure S2) (Qin et al. 2006; Chen et al. 2014; Zhao et al. 2015). Due to continuous arsenic volatilization, it was reasonable to see that the MMA and DMA levels decreased on day 21 in YC2. Interestingly, some arsM-carrying bacteria in YC2 were also found to be sulfate reducers or oxidizers, such as Sulfuricella, Desulfitobacterium, and Thioalkalivibrio (Sorokin et al. 2012; Li et al. 2018; Kojima et al. 2021). This result suggested that the cycling of arsenic in this hot spring may also couple with S cycling. This finding is further supported by the highest recorded sulfate level of 152 mg/kg in the geochemical parameters of YC2 (Table 2).

On day 7, the DMA level in the microcosm from YC5 was initially greater than that in the microcosm from MMA. However, both the DMA and MMA levels decreased to undetectable levels by day 21. The community of microorganisms harboring the arsM gene in YC5 was distinct from that in YC1 and YC2, with a notable abundance (75.0%) of Acidobacteriota observed on day 7. Acidobacteriota have been widely associated with arsenic methylation processes (Figure S2) (Yan et al. 2022). Subsequently, on day 21, the dominance of Actinobacteriota (60.5%) and Proteobacteriota (14.0%) increased. The Proteobacteriota, which carry the arsM gene, are key players in arsenic methylation and volatilization (Figure S2) (Bennett et al. 2012), as shown at the genus level by the observed increases in the abundances of Pseudomonas, Dehalobacter, and Rhodobacter (Figure 4(b)). Previously, Pseudomonas alcaligenes NBRC14159 was found to rapidly methylate As(III) to DMA and trimethylarsine oxide (TMAO) (Figure S2) (Zhang et al. 2015a). More importantly, an approximately tenfold increase was observed in the proportion of Mycobacterium, affiliated with Actinobacteriota, from 5.9 to 60.1%. Mycobacterium species are known for their strong ability to demethylate methylarsenic species (Lehr et al. 2003; Zhu et al. 2017). Consequently, on day 21, almost no detectable levels of methylarsenic were found in our study.

Furthermore, the phylum-level abundance of arsM-carrying bacteria within these three microbial community samples differed from what was observed in the Tengchong Hot Springs (Wang et al. 2023). Additionally, the five identified methanogen species in our samples differed from the nine best-matched arsM-carrying methanogens previously reported in the Tengchong hot spring (Wang et al. 2023); these included Methanosarcina horonobensis, Methanosarcina thermophila, Methanolacinia petrolearia, Methanolacinia paynteri, Methanoculleus taiwanensis, Methanoculleus thermophilus, Methanocella paludicola, Methanolobus halotolerans, and Methanothrix soehngenii.

Our study exclusively investigated microbial communities harboring the arsM gene in geothermal environment samples stimulated by lactate. However, it remains unclear exactly how many types of microorganisms, and under what conditions, participate in the actual arsenic methylation process in situ. Metagenomic and transcriptomic data should be integrated into future research on the composition and ecophysiology of active microbial functional groups involved in arsenic methylation and demethylation under various geothermal conditions. This approach could aid in elucidating the mechanisms of microbial-driven arsenic biogeochemical processes in actual geothermal environments. Additionally, understanding the microbial mediation of arsenic biogeochemical cycles in geothermal ecosystems could enhance our ability to manage or manipulate specific microbial functional groups in certain environments. Biological indicators or mechanisms identified through this research could be instrumental in signaling or remediating arsenic-contaminated environments.

In this study, we discovered arsenic-methylating species in the sediments of a low-temperature thermal spring. Arsenic demethylating species were also detected in a niche located farther from the hot spring hole. The microbes in the two sediments near the hot spring hole continuously methylated arsenite after stimulation with lactate and arsenite supplementation, sharing similar arsM-carrying communities. Sediments slightly farther from the hot spring showed both methylating and demethylating activities. The latter were enriched in As-demethylating species such as Mycobacterium. These findings enhance the existing understanding of microorganism-mediated arsenic methylation in low-temperature, low-arsenic geothermal environments.

Y.Y. and C.X.M. conceived and designed the experiments and conducted the experiments. Y.Y. and Y.L.Y. analyzed the data. The manuscript was written by Y.Y. Y.Y. and Y.L.Y. contributed to the provision, use, and description of the reagents/ materials/analysis tools.

This work was financially supported by the Chenguang Program of Shanghai Education Development Foundation and Shanghai Municipal Education Commission (No. 19CGB03).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.