Effects of nickel and cobalt on methane production and methanogen abundance and diversity in paddy soil

Rice field sites and soil sampling

Non-contaminated paddy soil was collected from a paddy field located in the Experimental Station of Jiaxing Academy of Agricultural Sciences in Zhejiang Province (30°77′N, 120°76′E). This site is in the center of the Hangjiahu Plain in the Yangtze River Delta. Soil samples were collected from the 20-cm surface layer in three plots without pollution. Five soil cores (diameter: 8 cm, height 20 cm) were collected from each plot, and the 15 cores were mixed together. Part of the soil was stored at 4 °C before the incubation experiments, and another part was naturally air-dried for testing the soil properties. The dried soil was ground through a 2-mm stainless-steel sieve, and the physical and chemical properties of the tested soil were as follows: pH (H₂O), 6.10; soil organic carbon, 49.7 g kg⁻¹; total nitrogen, 2.72 g kg⁻¹; ammonia, 1.50 mg kg⁻¹; nitrate, 0.34 mg kg⁻¹; available Fe, 2.97 g kg⁻¹; and sulfate, 0.41 g kg⁻¹.

Incubation experiments

Fresh soil samples (10 g dry soil) were incubated in 120-mL serum bottles, which were sealed with screw-caps with rubber stoppers. Based on China’s soil environmental quality limits, 5 mL of the solution containing different amounts of Ni²⁺ (NiCl₂, 0, 50, 100, 200 and 500 mg kg⁻¹ dry soil) or Co²⁺ (CoCl₂, 0, 25, 100, 200 and 500 mg kg ⁻¹ dry soil) were added. Each treatment has three replicates. The treatments with the Ni additions were referred to as Ni50, Ni100, Ni200, and Ni500; additionally, ck1 referred to the control without Ni nor acetate addition, and ck2 referred to the control without Ni addition but with acetate added. The treatments with Co additions were referred to as Co0, Co25, Co100, Co200, and Co500. Flasks were preincubated at 25 °C for 12 d, and acetate (0.3 mmol g⁻¹ dry soil) was subsequently added. The batch experiments of Ni and Co addition were incubated 16 and 21 days after acetate addition, respectively. Methane production potential was determined using gas chromatography (GC-900, Shanghai Kechuang Chromatography Instrument Co. Shanghai, China) with a hydrogen flame ionization detector (FID). Methane concentrations were calculated per kilogram dry weight (g_dw) of the soil.

The methane production (P) was calculated as follows (Wassmann et al., 1998), (1)${\displaystyle P=\frac{dc}{dt}\times \frac{V_H}{W_S}\times \frac{MW\times T_{st}}{MV\times \left(T_{st}+T\right)}}$ where dc∕dt is the measured change in the mixing ratio of methane in the headspace over time (µmol mol⁻¹ d⁻¹); V_H the volume of headspace (L), Ws the dry weight of soil (g), MW the molecular weight of methane (g), MV the molecular volume (L), T the temperature (K), and Tst the standard temperature (K). Methane production was expressed as mg CH₄ kg⁻¹ dry soil d⁻¹.

Ni and Co fractionation

To assess the variance in the liability and bioavailability of Ni and Co in the soil after incubation, a sequential extraction procedure was used, based on the methods in Tessier, Campbell & Bisson (1979). In brief, Ni and Co were sequentially extracted using MgCl₂ (pH 7.0, 1.0 mol L⁻¹), NaAc (pH 5.0, 1.0 mol L⁻¹), NH₂OH-HCl (0.04 mol L⁻¹), HNO₃ (0.02 mol L⁻¹), HNO₃-HClO₄-HF digests, and the extracts were detected using inductively coupled plasma - mass spectrometry (ICP-MS 7500cs, Agilent). These metal fractions were primarily associated with exchangeable, carbonation-bound, iron-bound and manganese oxide-bound, and organic matter-bound to organic matter and phases, respectively.

DNA extraction and fluorescence quantitative PCR

Soil DNA from each bottle of the different treatments was extracted according to the procedure provided for the Soil DNA Isolation Kit (MoBio UltraClean, San Diego, CA, USA). DNA were extracted in duplicate to avoid the quantification bias due to incomplete and random extraction of soil microbial DNA. DNA quality and quantity were analyzed by a Quawell^® Q5000 spectrophotometer (Beckman Coulter, Brea, CA, USA). Fluorescence quantitative PCR was based on the methanogen gene mcrA. The functional gene sequences were amplified using the specific primer pair ME1 (5′-GCMATGCARATHGGWATGTC-3′) and ME2 (5′-TCATKGCRTAGTTDGGRTAGT-3′) (Hales et al., 1996). Each reaction (25 µL) was carried out with 12.5 µL of SYBR Premix Ex Taq™ (Takara, Dalian, China), 0.2 µM of each primer and 1-10 ng of DNA. After amplification, a melting curve was generated (65–98 °C, 0.2 °C per read, 6-s hold) to test product specificity. The standard curve was made using a known copy number of plasmids with ten-fold serial dilution, which included the mrcA fragments from Methanosarcina that were ligated to the vector (p-GEM T-easy, Promega, Madison, WI, USA).

High-throughput sequencing

The archaeal 16S rDNA was amplified based on the custom degenerate primer pair 1,106 (5′-TTWAGTCAGGCAACGAGC-3′) and barcode-1378 (5′-TGTGCAAGGAGCAGGGAC-3′) (Watanabe, Kimura & Asakawa, 2006) to construct a library. The PCR (25 µL) was performed with a 10  ×   buffer (2.5 µL), dNTP (2.5 mM of each), primers (3.0 pmol of each), Taq DNA polymerase (0.5 U rTaq, TaKaRa, Dalian, China) and DNA template (∼50 ng). The PCR cycling steps were as follows: hot start at 95 °C for 3 min, followed by 28 cycles of 94 °C for 30 s, 54 °C for 30 s, 72 °C for 30 s, and ending with a 72 °C extension for 10 min. To minimize PCR bias, each sample was amplified in triplicate, and the three samples were subsequently mixed together for sequencing.

Sequencing of the archaeal 16S r DNA was conducted using the MiSeq Illumina platform at Majorbio Biotechnology Company (Shanghai, China). Sequences were screened (<200 bp, no chimeras) from the resulting data with routines by Usearch (vsesion 7.0) for quality control. Sequences were aligned with the Silva database (http://www.arb-silva.de/) and subsequently analyzed using QIIME (version 1.7.0, Caporaso et al., 2010). The cutoff of the operational taxonomic units (OTUs) was at the 97% similarity level, and this resulted in 7153 OTUs that were clustered using Uclust (Edgar, 2010). Each OTU was annotated in a taxonomic way using the Ribosomal Database Project (RDP) classifier (Cole et al., 2007) based on a representative sequence and a training set that was extracted from the Silva108 database (Quast et al., 2012). All Illumina sequence data reported in this paper have been deposited in the NCBI Sequence Read Archive (SRA) database (accession no. SRP145544).

Statistical analysis

The differences between values of the mean metal concentration, methane production, and gene abundance in the different treatments of this study were statistically analyzed by one-way analysis of variance (one-way ANOVA) and considered significant at the level of p < 0.05, and the Duncan test was used for post hoc analysis; all analyses were conducted using SPSS for Windows.

Fractionation analysis of nickel and cobalt in paddy soil

According to the Tessier method, five forms of Ni and Co in the paddy soil were measured (Fig. 1). The metal-exchangeable and carbonate-bound Ni accounted for 33.9–35.1% and 43.4–61.4% of the total Ni for control and contaminated soils, respectively. The total Ni generally equaled to that of the total Ni added (i.e., 43.0–422.4 mg kg⁻¹ dry soil), while the available Ni content was distributed at 24.5–229.2 mg kg⁻¹ dry soil (Fig. 1B). Similarly, the total Co concentration ranged from 15.8 to 468.4 mg kg⁻¹ dry soil, and the available Co ranged from 1.78 to 267.3 mg kg⁻¹ dry soil (Fig. 1D). The amounts of available cobalt, including the exchangeable and carbonate-bound cobalt, increased greatly, even with the low concentration cobalt additions (58.7–229.1 mg kg⁻¹ dry soil).

Effects of nickel and cobalt on kinetics of methane production in paddy soil

The effect of the addition of nickel on the methane dynamics in paddy is shown in Figs. 2A and 2B. Methane production occurred at very low levels (<500 mg CH₄ kg⁻¹dry soil) in the group without added sodium acetate and heavy metals (i.e., the control group, ck1) (Fig. 2A). The methane production rate (12 mg CH₄ kg⁻¹ dry soil d⁻¹) and the maximum methane accumulation (266 mg CH₄ kg ⁻¹ dry soil) of the control group (ck1) were much lower than those of the other groups that had added sodium acetate (280–561 mg CH₄ kg⁻¹ dry soil d⁻¹; 3,641–4,232 mg CH₄ kg⁻¹ dry soil) (Fig. 2B). Under anaerobic conditions, no distinct difference in methane production was observed between the treatments with different nickel concentrations during the initial preincubation stage (i.e., 0–12 days) (Fig. 2A). The methane production rate (561 mg CH₄ kg⁻¹ dry soil d⁻¹) in the 50 mg kg⁻¹ nickel addition was higher than that of ck2 (546 mg CH₄ kg⁻¹ dry soil d⁻¹). However, the maximum methane accumulation was not significantly different. With the further increase in the nickel concentration, the methane production rate and maximum methane accumulation decreased to 284 mg CH₄ kg⁻¹ dry soil d⁻¹. The methane production rates were significantly negatively correlated with the total nickel and the available nickel (r² = 0.958, p < 0.05 and r² = 0.899, p < 0.05, respectively), and the maximum methane accumulation was negatively correlated with the nickel content (r² = 0.910, p < 0.01).

The effect of the addition of cobalt on the methane dynamics in paddy soil is shown in Figs. 2C and 2D. At the low concentration of 25 mg kg⁻¹ of cobalt, the methane production rate (684 mg CH₄ kg⁻¹ dry soil d⁻¹) and the lag-time of the methanogenic process showed no differences from the control group (643 mg CH₄ kg⁻¹ dry soil d⁻¹). With the increase of the cobalt concentration, the generation of methane decreased by 45.4%–76.3%, and the lag-time also became longer. The methane production rates were significantly negatively correlated with the total cobalt and the available cobalt (r² = 0.983, p < 0.05 and r² = 0.981, p < 0.05, respectively).

The mcrA gene abundance in contaminated paddy soil

The effect of nickel on the mcr A gene abundance of the methanogenic functional gene is shown in Fig. 3. The mcr A gene abundance (8.8  ×  10⁴ ± 0.8 × 10⁴ copies g⁻¹ dry soil) of the control group (ck1) was two magnitudes lower than that in the other treatments that were supplemented with acetate substrates. The gene abundance (3.1  ×  10⁷ ± 0.5  ×  10⁷ copies g⁻¹ dry soil) at 50 mg kg⁻¹ was slightly larger than that of ck2 (2.2  ×  10⁷ ± 0.2  ×  10⁷ copies g⁻¹ dry soil), but this difference was not significant. As the nickel addition increased to 500 mg kg⁻¹, the abundance decreased significantly by one order of magnitude, to 9.2  ×  10⁶ ± 0.4  ×  10⁶ copies g⁻¹ soil. Similarly, the abundance of the mcr A gene under the highest concentration of 500 mg kg⁻¹ Co was one order of magnitude lower than that of the control.

Archaeal communities in paddy soil contaminated by nickel and cobalt

The archaeal compositions differed significantly at the genus level in the soils with different Ni and Co concentrations (Fig. 4). The most abundant Methanosarcina showed a relative abundance that varied from 14.5% at Ni500 to 51.3% at Ni50. In the Ni-amendment treatments, the relative abundance of Methanosarcina decreased with the increase of the nickel concentration above 50 mg kg⁻¹. In the ck2 group (37.3%), which had added sodium acetate but no Ni additions, the Methanosarcina increased relative to the ck1 group (31.6%). Additionally, the relative abundance of Methanosarcina increased with the increase of the Co concentrations; at Co500, the increase was up to 39.8%. In treatments of all concentrations of Ni and Co, the other dominant species of archaea (1%–10%) included Methanobacterium, Methanocella, Methanomassiliicoccus and Rice Cluster I. The abundance of Methanomassiliicoccus was the highest under the treatment of Ni100 (15.0%), while that was the lowest under the treatment of Ni500 (4.26%). Low concentration of Co addition promoted the abundance of Methanomassiliicoccus (23.6% for Co25), which decreased with increasing Co concentration (9.52%–18.8%). The abundance of Methanocella ranged from 7.27% to 17.5% with Ni addition, except a low relative abundance (2.01%) observed at Ni200. With the addition of Co, the relative abundance of Methanocella was higher (8.52%–15.3%) compared with the control (Co0, 5.01%). When the Ni concentration was higher than 100 mg kg⁻¹, the relative abundance of Rice Cluster I was lower (3.26%–5.76%) than that in the control and low concentration of Ni (7.02%–8.77%). With Co addition, the relative abundance of Rice Cluster I was much higher (9.27%–14.3%) except for the treatment Co500 (5.26%). The relative abundance of Methanobacterium ranged from 3.76% (Ni500) to 17.8% (Ni100) after Ni addition. Compared with the treatment Co0 (6.52%), the relative abundance of Methanobacterium decreased at low Co concentration (2.25%–3.01%), and increased with increasing Co concentration (5.76%–6.02%). Methanocorpusculum was observed in the control without the addition of metals (1.25% at ck2, ck), but not after the addition of nickel.

The relative abundance of the dominant genus Bathyarchaeota varied in the range 4.32%–12.31% for the Ni treatments and in the range 6.99%–17.8% for the Co treatments (Fig. 4A). Moreover, the lowest abundance was observed in the treatment with the highest metal additions (i.e., Ni500 and Co500), and the highest abundance was detected in Ni100 and Co25, respectively. The relative abundance of the soil Crenarchaeotic group (SCG) was 4.66% in ck1, 8.15% in ck2, and 20.80% in Ni500; however, SCG were not abundant in the Ni50, Ni100 and Ni200 treatments (<2%). Similarly, the SCG was abundant in all Co treatments (7.16%-15.6%) except Co100 (0.34%). In addition, we found that Candidatus Nitrosopumilus showed high relative abundances in the experimental groups with high amounts of metal, including Ni500 (8.32%) and Co200 (4.33%). Parvarchaeota, FHM a11 terrestrial group, South African Gold Mine Gp 1, Candidatus Nitrososphaera, Candidatus Nitrosopelagicus, Group C3, MKCSTA3, Candidatus Methanoperedens, and Rice Cluster II were observed at percentages less than 1% in the treatments with relatively high concentrations of added metals.

The significant impacts of Ni and Co contamination on the archaeal community compositions were further demonstrated by the clustering of the dominant archaea genus corresponding to different Ni or Co levels in the heatmaps (Fig. 5). Similar patterns in shifts of the major archaeal genera were observed in the Ni and Co contamination. Generally, methanogens like Mehtanosarcina, Methanocela, RC-I, Methanobacterium, Methanomassillicoccus and Batyarchaeota, were dominant in the paddy soils and tended to have lower relative abundances at Ni500. Nitrosopumilus and Nitrososphaera were decreased with Ni addition. Soil Crenarchaeotic Group SCG, Methanoperedens and Methanocellales increased at low Ni concentration, and decreased at Ni500. Marine Group II and Methanomicrobiaceae increased at low Co concentration (Co25, Co100).