Introduction

Atmospheric carbon dioxide (CO2) concentration has increased from 280 to 380 ppm since the industrial revolution and is projected to reach 420 to 936 ppm at the end of 21th century depending on scenarios1. The rapid increase of CO2 has drawn much attention because it closely relates to global climate change. Most of the attention has focused on the responses of aboveground and belowground processes as elevated CO2 (eCO2) impacts numerous biogeochemical processes in both natural and agricultural ecosystems1,2.

Soil organic carbon (SOC) pool is three times greater than C stored in the atmosphere or vegetation biomass2. It is therefore imperative to understand how eCO2 impacts soil C dynamics and the associated underlying mechanisms. Studies have demonstrated that eCO2 could affect soil microbial communities, C and nitrogen (N) cycling processes3,4,5. However, how eCO2 could affect C and N dynamics is of much uncertainty as the responses of soil microbial communities to eCO2 could be positive, negative or neutral6,7,8,9,10.

Soil microorganisms play vital roles in soil C and N transformations and nutrient cycles and thus affect soil biological, chemical and physical properties. For instance, the activity of soil microorganisms and microbial community composition determine not only soil C sequestrations and emissions and the decomposition and accumulation of soil organic matter11,12, but also N2-fixation, nitrification and denitrification and the accumulation of plant available NH4+ and NO3−13. However, limited information is available to the effect of eCO2 on soil bacterial and fungal abundance, community structure and in turn C and N dynamics.

Furthermore, global climate change like elevated CO2 is happening along with other environmental issues, including heavy metal pollution, which has become a seriously global problem due to anthropogenic activities, such as mining, smelting and sewage sludge disposal14. Unlike some organic contaminants, heavy metals are often accumulated in living organisms and not biodegradable in the environment. Many heavy metals, including cadmium (Cd), are toxic to soil microbes, which affect soil biogeochemical processes, such as soil organic matter (SOM) decomposition, through altering microbial biomass and activity14,15. Under climate change, the effects of Cd on soil microbial activity may become more complex as potential interactions between climate factors and excessive Cd in soil. However, the interactive consequence of CO2 and Cd on soil microbial community has seldom been investigated.

In this study, loessial soils were treated with four levels of Cd, i.e., 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and two levels of CO2, i.e., 360 (aCO2) and 480 (eCO2) ppm. Soil bacterial and fungal abundance, community structure, CO2 emission, C and N were measured to determine how soil Cd pollution and elevated CO2 separately or in combination could impact soil microbial community and C and N accumulation.

Results

Microbial community

The abundance of soil fungi was affected not only by Cd but also by CO2 concentrations (Fig. 1). Fungal abundance generally linearly increased with incubation time over the eight weeks of Cd exposure under all treatments. Fungal abundance was much greater under eCO2 than under aCO2, regardless of Cd levels. Compared to the non-Cd control, fungal abundance was significantly increased under Cd1.5, but decreased under Cd3.0 and Cd6.0, regardless of CO2 levels.

Figure 1
figure 1

Effects of cadmium and CO2 on soil fungal abundance after 2 weeks (A), 4 weeks (B), 6 weeks (C) and 8 weeks (D) under 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and ambient CO2 (360 ppm, aCO2) or elevated CO2 (480 ppm, eCO2).

Data (means ± SE, n = 3) followed by different letters between Cd treatments for the same CO2 treatment (a, b, c, d) or between aCO2 and eCO2 for the same Cd treatment (x, y) are significantly different at P < 0.05 according to Mann-Whitney multiple comparisons.

Full size image

Soil bacterial abundance was gradually decreased with increase of Cd concentrations. Its abundance was lower under aCO2 than under eCO2 at the same Cd concentration. Bacterial abundance was decreased after 2 to 4 weeks and was lowest at 4 week over the eight weeks of Cd exposure under both aCO2 and eCO2 (Fig. 2). But soil bacterial abundance began to increase after 4 weeks and was highest at 8 weeks of incubation.

Figure 2
figure 2

Effects of cadmium and CO2 on soil bacterial abundance after 2 weeks (A), 4 weeks (B), 6 weeks (C) and 8 weeks (D) under 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and ambient CO2 (360 ppm, aCO2) or elevated CO2 (480 ppm, eCO2).

Data (means ± SE, n = 3) followed by different letters between Cd treatments for the same CO2 treatment (a, b, c, d) or between aCO2 and eCO2 for the same Cd treatment (x, y) are significantly different at P < 0.05 according to Mann-Whitney multiple comparisons.

Full size image

The triplicate running of DGGE profiles was highly reproducible so only one DGGE profile was displayed (Fig. 3). After 4 weeks of incubation, 10 bands under eCO2 while 8 bands under aCO2 were obviously observed, irrespective of Cd concentrations (left panel in Fig. 3). Among these bands, 4 distinctive bands were generated (the left-towards red arrows in the left panel) and 2 bands were disappeared (the right-towards red arrows in the left panel) under eCO2 than under aCO2. Furthermore, after 8 weeks of incubation, 8 bands under eCO2 while 6 bands under aCO2 were obviously observed, irrespective of Cd concentrations (right panel in Fig. 3). Among these bands, 3 distinctive bands were generated (the left-towards red arrows in the right panel) and 1 band was disappeared (the right-towards red arrows in the right panel) under eCO2 than under aCO2. In addition, there were visual differences in DGGE band width between two incubation durations (4 and 8 weeks), while there seemed no distinctively visual differences between Cd0 and other Cd concentrations (no measurements were done to determine the width). Therefore, the DGGE profiles, e. g., the soil microbial community structure, were altered under eCO2 and cadmium pollution over the 8 weeks of soil incubation.

Figure 3
figure 3

Effects of cadmium and CO2 on soil bacterial community structure (DGGE profiles of PCR products from 16S rDNA extractions) after 4 weeks and 8 weeks under 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and ambient CO2 (360 ppm, aCO2) or elevated CO2 (480 ppm, eCO2).

Red arrow showed differences in DGGE bands between aCO2 and eCO2 treatments.

Full size image

Soil C and N dynamics

The CO2 emission showed a decrease trend with the increase of Cd concentrations under either eCO2 or aCO2 (Fig. 4). The cumulative soil CO2 effluxes were generally significantly greater under eCO2 than under aCO2 over the 8 weeks incubation, regardless of Cd treatment levels.

Figure 4
figure 4

Effects of cadmium and CO2 on soil cumulative carbon emission after 2 weeks (A), 4 weeks (B), 6 weeks (C) and 8 weeks (D) under 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and ambient CO2 (360 ppm, aCO2) or elevated CO2 (480 ppm, eCO2).

Data (means ± SE, n = 3) followed by different letters between Cd treatments for the same CO2 treatment (a, b, c, d) or between aCO2 and eCO2 for the same Cd treatment (x, y) are significantly different at P < 0.05 according to Mann-Whitney multiple comparisons.

Full size image

Concentrations of soil C and N were greater in soils treated with higher concentrations of Cd at both CO2 levels. Significantly higher concentrations of soil C and N were observed under aCO2 than under eCO2, regardless of Cd treatments (Fig. 5a and 5b).

Figure 5
figure 5

Effects of 8-weeks cadmium and CO2 treatments on soil carbon (A) and nitrogen (B) concentrations under 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) and 6.0 (Cd6.0) mg Cd kg−1 soil and ambient CO2 (360 ppm, aCO2) or elevated CO2 (480 ppm, eCO2).

Data (means ± SE, n = 3) followed by different letters between Cd treatments for the same CO2 treatment (a, b, c, d) or between aCO2 and eCO2 for the same Cd treatment (x, y) are significantly different at P < 0.05 according to Mann-Whitney multiple comparisons.

Full size image

Discussion

Due to excessive emission of industrial waste, wastewater irrigation and serious abuse of chemical fertilizers and pesticides, numerous heavy metals have been released into the environment16. One fifth of total cultivated land in China was polluted by heavy metals, such as arsenic (As), cadmium (Cd), chromium (Cr) and lead (Pb)17. Heavy metal pollution has been a threat to the earth system. Previous studies have showed that either short-term or long-term exposure to heavy metals could result in changing of soil microbial communities, or decrease in diversity and activities12,13,14,15. Moreover, fungi are more tolerant to heavy metals than bacteria18. Our results in loessial soil were consistent with those previous studies18. Bacteria were more sensitive to Cd stress than fungi because the low (1.5 mg Cd kg−1 soil) treatment restrained the bacteria reproduction but improved fungal reproduction (Fig. 1 and 2). With the increase of Cd concentrations, the fungal and bacterial abundance were gradually decreased under normal CO2 (Fig. 1 and 2). Moreover, the alteration of DGGE profiles might provide direct evidence that the microbial communities might have been influenced by elevated CO2, but not by cadmium pollution at the tested Cd concentrations (Fig. 3). Consequently, no matter whether the soil was incubated under the ambient CO2 or the elevated CO2 concentration, soil respiration was restrained while concentrations of soil C and N were increased (Fig. 4 and 5), indicating that Cd pollution could alter soil biogeochemical cycling.

Elevated CO2 have various impacts on biogeochemical processes in terrestrial ecosystems. There were some different viewpoints on the effect of elevated CO2 on soil microbial abundance. For instance, in free-air CO2 enrichment experiment no significant changes were observed in the relative abundance or composition of fungi19. On the contrary, the fungal biomass in a chaparral ecosystem was significantly increased under 550 ppm elevated CO220 and the relative abundance of fungi in a scrub-oak ecosystem was greater under elevated CO2 than under ambient CO23. Elevated CO2 could stimulate microbial growth, especially bacteria and lead to substantial changes in the activity and structure of soil microbial communities21. Recently, analyses of 16S rRNA genes showed that soil microbial community structures were significantly changed by elevated CO222. Our experiments also demonstrated the bacterial and fungal abundance were increased by eCO2 regardless of Cd levels. Compared with the same Cd concentration, elevated CO2 significantly altered the bacterial and fungal abundance, the bacterial DNA brand and decreased the concentrations of soil C and N, resulting in changes in soil microbial community and soil biogeochemical element cycling. These results support that elevated CO2 could result in a loss of soil carbon4.

Elevated CO2 and Cd exhibited opposite effects on soil microbial communities. Cd pollution restrained microorganism's growth and high concentration of Cd could decrease microbial biomass and activities. As a result, less SOM could be utilized and thus less CO2 was released from soil, which in turn had more C and N retained in the soil. Moreover, with an increase of Cd concentration, the negative effect was more obvious. However, soil microbial relative abundances and activities of soil carbon-degrading enzymes in soils exposed to an elevated CO2 might have been increased, leading to an enhanced degradation of soil organic matter than soils exposed to ambient CO2, which in turn consumed more C and N in soil. Therefore, a combined effect of Cd and eCO2 could display antagonisms on terrestrial C pool by altering soil microbial community.

In conclusion, higher Cd exposure to soil could lead to a decrease of both bacterial and fungal abundance and hence an increasing soil C and N concentrations. Meanwhile, elevated CO2 could alter microbial communities and lead to an increase of soil C emission, resulting in a decrease in soil C and N concentrations. These results suggest that eCO2 could stimulate, while Cd pollution could restrain, microbial reproduction and C decomposition with the restraint effect alleviated by eCO2. Nevertheless, more research is warranted on the combined effects of Cd pollution and elevated CO2 on the dynamics of soil C.

Methods

Soil characteristics

Loessial soils at 0–10 cm depth were collected from a field after maize harvest in a suburban of Xi'an, Shaanxi, China (34° 44′ N, 109°49′ E). According to Pang et al.23, the soil pH was 8.3, moisture was 7.8%, soil SOC was 8.9 g C kg−1 and CaCO3 was 30.11 g kg−1.

Experiment design

The experiment consisted of four Cd and two CO2 treatments in a completely randomized arrangement. The four Cd treatments were 0 (Cd0), 1.5 (Cd1.5), 3.0 (Cd3.0) or 6.0 mg Cd kg−1 soil (Cd6.0) and the two CO2 treatments were ambient (360 ppm, aCO2) or elevated (480 ppm, eCO2). Each treatment had three replicates, for a total of 24 replicates. The soil was put in 24 growth pots. Each pot (15 cm diameter and 10 cm height) contained 1.0 kg soil. The CO2 concentrations were monitored and controlled by a meter (output flux: 25 L min−2, 15 MPa pressure; Shanghai Yichuan Meter Co., Shanghai, China). Cd was applied as CdCl2·2.5H2O (99%, Tianjin Fuchen Chemical Co., Tianjin, China), which was firstly dissolved in distilled water and then thoroughly mixed with soil. The Cd-containing soils were continuously incubated under 25 ± 1°C and 70% of relative humidity over 8 weeks in a plant growth chamber.

Soil microbial community structure

The number of bacteria and fungi was counted using the plate counting technique24. Briefly, approximately 10 g soils from each pot were putted into a 300 ml Erlenmeyer flask containing 90 ml sterilized potassium phosphate buffer (0.35 g KH2PO4, 0.65 g K2HPO4 and 0.10 g MgSO4 in 1.0 liter water), centrifuged at 250 rpm for 15 min and the soil extraction solution was then diluted into 100-fold. Control plates minus soil suspensions were used to check potential contamination. The total number of bacteria was cultured with an agar medium (including 5 g beef, 10 g peptone, 5 g NaCl, 20 g agar and 0.1 g cycloheximide per 1,000 ml distilled water) for 48 h at 37°C. The total number of fungi was cultured with an agar medium (including 10 g glucose, 5 g peptone, 1 g KH2PO, 0.5 g MgSO4·7H2O, 20 g agar and 0.1 g chloramphenicol per 1,000 ml distilled water) for 72 h at 25°C. The colony concentrations were expressed as colony-forming units per gram soil (CFU/g)24.

DNA extraction and PCR-DGGE analysis

Analyses of bacterial community structure were performed with incubated soils at week 4 and week 8, respectively. Using a Fast DNA Spin Kit (Bio 101, USA), the extraction of soil bacterial DNA was tripled (1.0 g soil each) from each treatment according to the manufacturer's bead-beating procedure. The yield and quality of extracted DNA was validated by the 1.0% (w/v) agarose gel electrophoresis with ethidium bromide staining. The bacterial 16S rDNA genes were amplified with the primer EUB341F (5′-CCT ACG GGA GGC AGC AG-3′) with a GC clamp (CGC CCG CCG CGC CCC GCG CCC GTC CCG CCG CCC CCG CCC G) at the 5′ end and the premier EU500R (5′-GTA TTA CCG CGG CTG CTG G-3′)25. Using 50 μl reaction volumes the PCR amplification was carried out with an iCycler Thermal Cycler (Bio-Rad, USA). A final 50 μl reaction mixture contained 1–10 ng DNA extracts as template, 5 μl of 10 × PCR buffer (minus Mg2+), 15 pmol of each primer, 1.25 μl unit Taq polymerase, 200 μmol deoxyribonucleoside triphosphate and 3 μl MgCl2 (25 mmol). A touchdown temperature profile was applied as the annealing temperature decreased from 65°C to 55°C after 20 cycles with a 0.5°C decline from each cycle and then the mixture was kept at 55°C for further 10 cycles. An aliquot of 5 ml PCR products was then run by 1.0% agarose gel electrophoresis with ethidium bromide staining prior to the denaturing gradient gel electrophoresis (DGGE), which was conducted with a Bio-Rad Dcode™ Universal Mutation Detection System (Bio-Rad, USA). The generated PCR product was loaded on 0.8 mm thick polyacrylamide gel (10% acrylamide/bisacrylamide, w/v = 37.5/1) under a 30% to 50% of urea and formamide (100% corresponds to 7 mol urea and 40% formamide, w/v) denaturing gradient that was increased in the electrophoresis running direction. The electrophoresis was run for 5 h under 60°C and 150 V. After the run, the gels were stained with 1 × TAE (containing 0.5 mg ml−1 ethidium bromide) for 20 min, rinsed with distilled water thoroughly and then photographed.

Measurements of soil total carbon and nitrogen

Soils from all 8-week incubation pots were air-dried to a constant weight and sieved through 1.0 mm mesh. Total carbon and nitrogen were determined according to the dichromate oxidation26 and the H2SO4 titrate method26, respectively.

Soil CO2 efflux measurements

After incubation for 2, 4, 6 and 8 weeks, the pot was placed into a PVC (polyvinyl chloride) collar (14.5 cm height and 20.3 cm inside diameter) and the bottom was completely sealed. Soil CO2 efflux rates (μ mol CO2 pot−1 min−1) between 12:00 pm and 15:00 pm were quantified by a LI-8100 soil CO2 flux system (LI-COR Inc., Lincon, NE, USA), which was equipped with a portable chamber (Model 8100-103) being placed into the PVC collar. The CO2 efflux rate was then automatically calculated from exponential regression of increasing CO2 concentrations over 3-min duration. Each measurement for a pot was repeated 3 times and 3 repetition experiments were performed in one day.

Data analysis

Data (means ± SE) were subjected to one-way analysis of variance (ANOVA) and significant differences between treatments were compared using the Mann-Whitney multiple comparison at P < 0.05. Statistical analyses were performed using the SPSS16.0 (SPSS Inc., Chicago, IL, USA).