Environmental characteristics in different conditions
The basic physicochemical properties of overlying water under different conditions, including EC, pH, and DO, are shown in Fig. 1. The electrical conductivities in all conditions were nearly 170 µs/cm at the beginning of the experiment; they increased to 401.5, 303, and 445 µs/cm under oxic, original, and anoxic conditions, respectively. The increased electrical conductivities indicated an increase in the concentrations of conductive substances in the solutions and that these substances were released in relatively larger quantities under anoxic conditions. We observed substantial differences in the DO concentrations under different redox conditions, which was consistent with the different incubation device treatments. The pH values remarkably increased under all conditions after 2 days of incubation and then slightly declined. The pH values were closely related to DO concentrations in the overlying water. A lack of oxygen can promote an acid process in sediment and then affect the pH values in overlying water due to a combination of strengthened bacterial respiration and intensive organic matter degradation (Song et al. 2020). Differences in DOC concentrations were also observed among the oxic, original, and anoxic treatments. The highest increase in DOC was found in anoxic conditions, from 5.3 to a maximum value of 8.7 mg/L. With increases in oxygen, the maxima concentrations of DOC in the original and oxic conditions were 6.3 and 6.8 mg/L, respectively, which were lower than those in anoxic conditions.
Dynamics of nitrogen and phosphorus in the overlying water
Under oxic, original, and anoxic conditions, the concentrations of dissolved TN increased (Fig. 2). We found the largest increase in TN under anoxic conditions; it grew from 10.10 to 19.73 mg/L after 25 days of incubation. No obvious differences in the TN concentrations were found between oxic and original conditions. After 25 days of incubation, the concentrations of NH4+ in the oxic and original conditions decreased to nearly 0 mg/L, while the concentrations of NH4+ increased from 5.48 to 8.90 mg/L under the anoxic condition. Compared with the oxic and original conditions, where it decreased, the NH4+ concentration considerably increased in anoxic conditions. In anoxic conditions, the lack of oxygen also produced low levels of NO3− of less than 0.5 mg/L in most treatments. The NO3− concentrations were similar in oxic and original conditions, increasing to 2.8 mg/L at the end of the experiment. We found the opposite for NO2−, which is a reduction product of NO3−. All conditions showed a decline in NO2− concentrations, and anoxic conditions mostly showed higher NO2− concentrations than oxic and original conditions.
The initial concentrations of phosphorus were low in this experiment; they were 0.13, 0.14, and 0.09 mg/L in oxic, anoxic, and original conditions, respectively. A decrease in phosphorus concentration was found in both oxic and original conditions; it declined to a concentration of 0.05 mg/L at the end of incubation. However, the phosphorus concentration remained constant under anoxic conditions. No significant differences were found between the dissolved TP and PO4− concentrations, indicating a negligible presence of organic phosphorus in the overlying water, and inorganic phosphorus was the main component of the released phosphorus.
Dynamics of DOM in the overlying water
Six fluorescent components were identified from the fluorescence spectra of the overlying water samples with PARAFAC, including five humic-like components (C1, C2, C3, C5, and C6) and one protein-like component (C4). The fingerprint maps and spectral loadings of the six components are shown in Fig. 3. We compared the spectra of the five components to those in the published PARAFAC models using the OpenFluor online spectral library (Murphy et al. 2014), which simultaneously identifies components with a Tucker congruence exceeding 0.95 on the excitation (Ex) and emission (Em) spectra. C1 had two excitation maxima at < 250 and 310 nm and one emission maximum at 410 nm; which could be identified as a terrestrial-delivered, humic-like component (Shutova et al. 2014, Stedmon et al. 2007). C2 had two excitation maxima, at < 250 and 375 nm, and one emission maxima, at 470 nm, and was identified as a terrestrial, humic-like component (DeFrancesco &Guéguen 2021, Yan et al. 2020). C3 was a humic-like component with excitation/emission maxima at 355/425 nm. It is associated with molecules characterized by the lowest aromaticity and is produced during the photodegradation of terrestrial DOM (Lambert et al. 2016, Murphy et al. 2011). C4 had an excitation peak at 275 nm and an emission peak at 325 nm, exhibiting the typical spectral characteristics of protein-like components (Dainard et al. 2019, Zhou et al. 2019). For C5 and C6, the excitation/emission peaks were found at 300/378 nm and 275/463 nm, respectively; they represent humic-like-component characteristics (Gonçalves-Araujo et al. 2015, Kida et al. 2019, Wauthy et al. 2018).
In the original condition, increases in fluorescence intensity were mainly attributed to C1, C2, and C3, which increased from 49.15, 24.76, and 13.53 QSE to 70.66, 40.08, and 25.15 QSE (Fig. 4), respectively. The release of fluorescent components under the original conditions mostly occurred within 3 days of incubation. Under anoxic conditions, all fluorescent components showed a continuous increase in fluorescence intensity. The fluorescence intensities of C1–C5 increased from 55.06, 30.59, 15.22, 17.57, and 6.22 QSE to 143,77, 77.9, 41.43, 35.33, and 22.8 QSE, respectively, which were the highest increases under all conditions. Compared with the original and anoxic conditions, oxic conditions inhibited the increase in the fluorescence intensity. We only found increases in fluorescence intensity under oxic conditions for C1–C4. The distribution of components was similar for all conditions, with C1 accounting for the highest proportion, with a range of 36–41%. The order of intensities of the components was also similar under the original and anoxic conditions, where C1 > C2 > C3 > C4 > C5 > C6. However, a stronger intensity of C5 was observed under oxic conditions, mostly accounting for the second or third largest proportion of the total intensity.
Diversity of microbial community
The alpha diversity of the microbial community was evaluated using the Chao1 richness and Shannon diversity indices, which were calculated using QIIME (Fig. 5). The Chao1 richness index under the three conditions ranged from 3811 to 6411. The Chao1 richness index ranged from 4320 to 6411 under oxic conditions, which was higher than those under the original and anoxic conditions. Anoxic conditions exhibited the lowest Chao1 richness, with an average value of 4765, whereas the averages of the Chao1 richness index in oxic and original conditions were 5385 and 5125, respectively. Increases in the Chao1 richness index in all conditions were similar, and major increases in the Chao1 richness index occurred after 8 days of incubation. The Shannon diversity index ranged from 7.8 to 10.5 for all conditions; higher values were found under the oxic and original conditions.
The structures of the microbial community in the sediment using the taxonomic distributions of the 16S rRNA sequences. The taxonomic percentage of the sequences is illustrated in Fig. 6, and nine dominant phyla were examined, including Proteobacteria, Acidobacteria, Actinobacteria, Firmicutes, Chloroflexi, and Bacteroidetes. The major phylum in the sediment samples was Proteobacteria, with percentages ranging from 33.71–37.12%. The second most abundant phylum in this experiment was Firmicutes, which accounted for 17.84% – 22.52% of the total phyla. Among the different conditions, anoxic conditions had higher abundances of Proteobacteria, Firmicutes, Actinobacteria, and Chloroflexi, while the original and oxic conditions had higher abundances of Acidobacteria and Bacteroidetes.
Expression levels of nitrogen cycle genes under different conditions
Nitrogen-cycle genes, including nifH, amoA, hzo, nirK, and nirS, were characterized in this study (Fig. 7). These nitrogen cycling genes correspond to the fixation, ammonia oxidation, anammox, ammonification, and denitrification processes in the sediment, respectively. At the beginning of the experiment, the levels of nifH gene expression in oxic, original, and anoxic conditions were 7.90×105, 5.86×105, and 3.69×105 copies/g, respectively. The nifH expression increased under oxic conditions and decreased under original and oxic conditions. After 25 days of incubation, the nifH expressions were 3.45 ×105, 6.98 ×105, and 1.34 ×106 copies/g for oxic, original, and anoxic conditions, respectively. Increases in the level of amoA expression were found in all conditions; under oxic, anoxic, and original conditions, they increased from 3.0×104, 1.64×104, and 1.23×104, to 1.59×105, 5.92×104, and 1.72×105 copies/g, respectively. The expression levels of amoA in oxic conditions were generally higher than those in anoxic conditions and lower than those in the original condition. The highest expression of amoA genes was found under original conditions after 25 days of incubation. The level of hzo gene expressions continuously increased under anoxic conditions, from 3.20×105 copies/g to 7.44×105 copies/g. Under original conditions, the level of hzo gene expressions decreased from 4.60×105 copies/g to 2.59×105 copies/g, and then slightly recovered to 4.08×105 copies/g at the end of the experiment. No obvious trend was found for the level of hzo gene expressions under oxic conditions owing to the fluctuations. The expressions of the nirK and nirS genes under the original condition first increased and then decreased to a low level at the end of the experiment. Under anoxic conditions, the expression levels of nirK and nirS slightly increased but were much lower than those under oxic and original conditions. However, under oxic conditions, we observed dramatic increases in nirK and nirS expressions, although the denitrification reaction usually occurs in an anoxic state.