As an important component of global aquatic ecosystems, lakes only account for a small fraction (2.8%) of the land surface (Downing et al., 2006), however play a unique and essential role in biogeochemical cycles due to their possession of diverse microbes, and capability to hasten the nutrients cycling (Hakulinen et al., 2005; Liu et al., 2009; Song et al., 2012; Small et al., 2014; Verpoorter et al., 2014; Li et al., 2015; Huang and Jiang, 2016; Frade et al., 2020). Microorganisms are monitors to variation in the external environment in the lake and may be one of the most sensitive indicators (Kuang et al., 2022). Metabolism of organic and inorganic elements can be changed by microbial communities which impact biogeochemical environments of lake sediments (Song et al., 2012). Moreover, it is reported that the carbon cycle is dominated by the balance between carbon-fixing and carbon-consuming occurred in microorganisms; microbial nitrogen-transforming networks both attenuate and exacerbate human-induced global change; microbial P mineralization is a side-effect of microbial C acquisition and can be driven by it; the remineralization of the organic matter of seafloor is facilitated by sulfate-reducing microorganisms (Spohn and Kuzyakov, 2013; Bowles et al., 2014; Gougoulias et al., 2014; Kuypers et al., 2018). Such natural diverse conditions harbor hotspots of microbes.

Increasing attention have been paid to the environmental regulatory factors leading to eutrophication and changing the constitute of sediment bacterial community, such as pH, N, P, N:P ratio (N/P) and soil organic carbon (SOC; Howarth et al., 2011; Wen et al., 2012; Yi et al., 2021; Kuang et al., 2022). And overdosing or excessive enrichment of nutrients lead to eutrophication. In the past half century, human activities have greatly accelerated the flow of nutrients to lake ecosystem, resulting in extensive eutrophication (Rabalais, 2002). In many industrialized countries, although the input of P drops sharply due to the improvement of wastewater treatment plants, the N pollution is still high (Howarth et al., 2011). In sediments, the decay of algal biomass generates organic carbon and consumes oxygen, which is conducive to N loss through denitrification. Under natural conditions and control experiments, when algae gather, N loss will be reduced (Zhu et al., 2020). Eutrophication may produce positive feedback to increase phosphorus supply. For example, P storage in many sediments decreases and P flux into water increases (Bol et al., 2018). In addition, in sediment, with the progress of eutrophication, the increase of C input and the decrease of oxygen input, sulfide will accumulate faster (Corredor et al., 1999). Even when the organic carbon (OC) content is low, sulfate reduction also accounts for a lot. When the OC concentration is high, the sulfate reduction contribution is higher (Howarth, 1984). Relationship between N, P and OC are also reported that the loss of N and the increase of P leads to lower N/P but higher trophic level index (calculated by Chl-a, Secchi disk transparency and TP). Specifically, in eutrophic lakes, the enrichment of total organic carbon (TOC) and all forms of P in sediment can promote potential denitrification rate and eventually leads to N loss by regulating community composition and increasing the abundance of nirS denitrifiers (Zhang et al., 2018). The eutrophication of surface water is an endemic global problem, and the nutrient load from agriculture is an underlying and persistent driving factor (Withers et al., 2014). Pesticides containing organophosphorus (OPPs) which were widely used in agriculture for their insecticidal properties are one of the most commonly used chemical pesticides in China (Wu et al., 2009). It is common for OPPs to leave residues in ecosystems due to their indiscriminate application (Montuori et al., 2016; Liu et al., 2019). To the best of our knowledge, there are various levels of toxicity associated with the acute and chronic exposure of humans, animals, plants, and insects to OPPs (Sidhu et al., 2019). It is demonstrated that pesticide pollution exists in agricultural ecosystems. For example, in sediments of Sundays and Swartkops Estuary, the highest concentration of OPPs is, respectively, up to 8.07 μg/kg and 13.6 μg/kg (Olisah et al., 2021a,b). The concentration of parathion and paraoxon-methyl (the degradation product of parathion-methyl) which are the most common OPP pollutants in Tai Lake varies from 5.88 to 506 ng/l (Wei et al., 2022).

In shallow lakes, eutrophication is widespread which leads to phytoplankton and algal blooms and then hypoxic or anoxic episodes when the organisms submerge and decompose, and is generally caused by increased P, N and organic matter (Gao et al., 2021; Li et al., 2016). As the largest shallow freshwater lake in the North China Plain, Baiyangdian lake is essential for maintaining ecosystem functioning in this highly populated region (Han et al., 2020). However, increased anthropogenic activities in recent decades have gradually increased the amounts of nutrients discharged into the lake in domestic sewage and industrial wastewater from the Baoding City in the upper reaches via the Fu River. These discharges and nutrients transported from farmland and residential areas have caused nutrient enrichment, resulting in eutrophication of Baiyangdian lake (Zhu et al., 2019). And then, the activities of C, N, P and S cycling in Baiyangdian lake have changed. A vast and uncultured diversity of microorganisms is exposed by molecular genetic techniques, so that there is a huge breakthrough of environmental microbiology (Handelsman, 2004). This is the main advantage of metagenomics, because cultivating new microorganisms is a difficult technology. Massive sequence data generated by metagenome technology allows linking microbial communities with their biogeochemical functions and provides insights about abiotic and biological interactions of microbial communities (Grossart et al., 2020). However, there are a few studies on the influence of environmental factors on the biogeochemistry cycle of shallow lakes under the condition of eutrophication. In this study, we (1) characterize the microbial community structure in the sediments of Baiyangdian, and (2) reveal the effect of environmental factors in controlling microbiome composition of sediment, and (3) show the relationship between relative abundance of important taxa and functional genes for regulating C, N, P and S cycling, as well as their metabolism pathway for further quantifying bacterial functional potential in biogeochemical cycling processes in inland freshwater lake floors under eutrophic threats.

Environmental factors are important to determine the composition of microbial communities. Many studies had shown that TN and TP played important roles in composition of bacterial communities (Song et al., 2012; Bergkemper et al., 2016; Wei et al., 2018). In our study, TN and TP greatly affected the bacterial community composition. TN likely showed a relatively greater impact on bacterial communities than TP. We divided the ten points into group 1 (S2, S3, S10) and group 2 (the rest of the samples sites; Figure 7). The values of TN, EC and N/P of group 1 was significantly higher than group 2 (p < 0.05). And, there is no obvious difference in total phosphorus content. Compared with the highly contaminated situation (value of TN and TP contents were approximately 4,000 and 800 mg/kg) of Baiyangdian in 2018 reported in January this year (Wang et al., 2022), TN and TP have dropped to around 2,400 and 515 mg/kg, respectively now because of somewhat significant results that the government has achieved in pollution control in recent years. But it is still under eutrophication threat. In addition, in site S3, in the condition of the highest N/P up to 9.04, Proteobacteria had the lowest relative abundance while Chloroflexi had the highest. Also, the function of assimilatory nitrate reduction of N fixation was also the strongest but the functions of C fixation (e.g., reductive citrate cycle, dicarboxylate-hydroxybutyrate cycle, reductive acetyl-CoA pathway) were at an average level. The high N/P might be caused by strong capacity of N removal of Proteobacteria (Nielsen et al., 1999; Ren et al., 2018) and the effects of Chloroflexi in N cycling (Rao et al., 2022), because of the lowest abundance of Proteobacteria and the highest abundance of Chloroflexi. TN and TP were important macronutrients for all biota on earth and were two essential components of the energy metabolism and the genetic backup and stable cell structures. For example, recent researches had provided evidence for that TN was the dominant environmental factor to drive the formation of bacterial community structure and that TP concentration was significantly correlated with the distribution of bacterial communities (Song et al., 2012; Fan et al., 2019). In addition, previous researches showed that microbial community composition could varied with addition of N (Wang et al., 2018). A study showed that the effect of TN on the community structure was mainly caused by pushing the variation of the dominated phyla and influencing the functional microbes of C and N metabolism (Liao et al., 2021). For TN, many studies had shown that the relative abundance of Actinobacteria might be restrained by TN (Pascault et al., 2014; Mao et al., 2019; Ouyang et al., 2020). It was consistent with the community distribution of sample S4, but not the communities of other samples. In addition, at site S4, 3-Hydroxypropionate cycling of C fixation was the strongest while the weakest function of denitrification and complete nitrification of N cycling (might be caused by low abundance of Actinobacteria) which might lead to the high level of TN. TN fixation of sediment might be weaker in areas with high water velocity (Duan et al., 2016). Sample S6, S7, S9 were basically sampled from open water; Sample S1 was sampled from Zhaowangxin river estuary. The relatively low TN content and N/P might be caused by high mobility of water. In addition, sample S10 was sampled from entrance of lake and S2, S3 were, respectively, sampled from the area where was more likely facing pressure of extensive water pollution and eutrophication near to Beihezhuang and Caiputai villages, all of them were relatively severely affected by human activities. The N enrichment in these samples was likely due to the geographic location of the sampling sites. Sample S5 was located near the village but the TN content was relatively low, which suggested that the situation was complicated under field conditions, and there might also be an area with relatively light pollution with less human activities. Sediment pH greatly impacted the TN contents by influencing microbial activities. Microorganisms were most active in neutral conditions while were inhibited in the alkali condition (Bai et al., 2005). Besides, from a global perspective, soil salinity was one of the major environmental factors limiting plant growth and productivity (Huang L. et al., 2017). We found that sediment TN was significantly negatively correlated with pH and positively correlated with EC. It was consistent with the findings of Mei et al. (2014). In lakes, reduction of the external P loading was a prerequisite for water quality improvement. However, lakes showed hardly any signs of recovery after reducing of the external P and it was mainly caused by recycling of P from the P-rich sediments (Lürling and van Oosterhout, 2013). The P cycle genes in Baiyangdian area were relatively rich (Figure 4) indicating that the P cycle in Baiyangdian sediments was relatively strong. This might be explained by the statement that the predominance of organic matter remineralization as the predominant pathway of P cycling in temporally varying bottom water redox conditions (Joshi et al., 2015).

Bacterial community played a key role in ecological functions of sediments. Proteobacteria was reported to contribute to N fixation, photosynthesis, S metabolism, methane metabolism and could degrade proteins, polysaccharides and other organic matter and removed N and P (Liao et al., 2019; Ouyang et al., 2020; Rathour et al., 2020; Duyar et al., 2021). As the dominant colonies in Baiyangdian sediment, the relative abundance of Proteobacteria in bacterial community was significant negatively correlated with Actinobacteria indicating their opposite distribution pattern. The dominance of Proteobacteria in bacterial community and opposite distribution pattern between Proteobacteria and Actinobacteria were also found in other plain lakes with different trophic status (Huang W. et al., 2017), suggesting that a universal mechanism was existed to shape the sediment bacterial community. Following Proteobacteria, Actinobacteria, Chloroflexi and Acidobacteria (Figure 2) were the large phyla within the bacterial community. Actinobacteria produced 70% of the world’s natural antibiotics and had the ability to degrade complex organic compounds (Barka et al., 2016). Recent studies on antibiotics in Baiyangdian showed that there was a real risk of antibiotic contamination in the Baiyangdian area (Zhang et al., 2022). Actinobacteria widely existed in ocean, soil and freshwater ecosystem, and many new species had been found (Wu et al., 2021). Actinobacteria had been identified to be potentially important DDT degraders and had been proved to degrade HCH by producing dechlorinating enzymes (Fudala-Ksiazek et al., 2018; Xu et al., 2021). In Baiyangdian Lake, Hu et al. (2010) had proved that organochlorine pesticides residue existed and HCHs were the predominant contaminants. This could explain why Actinobacteria occupied the second most abundant phylum of bacterial community in Baiyangdian, a typical shallow lake, which was also consistent with other studies (Liao et al., 2019; Ouyang et al., 2020). Then, noticeably, the third most abundance phylum of bacterial community in sediment was Chloroflexi. It was involved in the metabolism of organic matter and the transformation of contaminants (Liao et al., 2021). And, researcher had proved that Chloroflexi were rich in the sediments from eutrophic region (Huang W. et al., 2017).

Except for C metabolism, the biogeochemical processes of N and S also affected the biodegradation in sediments. Microbial groups related to N metabolism, including N₂ fixation groups and denitrification groups, could improve N availability, because bioavailable N was a major limiting factor in nutrient poor environments (Li et al., 2022). In our study, we found that the relative abundance of each module of the C-N-S cycle was consistent in each sample site, but its absolute abundance was different. It might indicate that in Baiyangdian lake, due to environmental properties, absolute abundance of each module of each sample site was different, but it showed consistent relative strength in the whole area. For examples, the relative abundance of reductive citrate cycle (M00173), denitrification (M00529) and assimilatory sulfate reduction (M00176) were, respectively, highest in C, N, S in all sediment samples. However, assimilatory nitrate reduction, denitrification and dissimilatory nitrate reduction of nitrogen cycling of group 1 were significantly high than group 2 (Figure 7). This indicated that higher N might accelerate the processes of assimilatory nitrate reduction, denitrification and dissimilatory nitrate reduction of nitrogen cycling. In addition, there were also significant differences in genus of Nocardioides involved in these pathways. This was consistent with the previous reports that the important effect of Nocardioides on nitrogen transformation (Zhang et al., 2021). Microbial processes associated with the S cycle had been extensively studied. Studies had affirmed that the microbial S cycle was among the most active in lakes (Sorokin et al., 2011; Vavourakis et al., 2019). Our research also found that a complete while weaker S cycle between SO₄²⁻ and HS⁻ might occur in Baiyangdian lake sediments compared to N cycle and C fixation (Figure 4). For N cycle, DNRA had been determined to co-occur with denitrification in estuary sediments (Baker et al., 2015). Also in our study, DNRA co-occurred with denitrification in shallow lake sediment and both of them were relatively active. The most abundant mode of C fixation of Baiyangdian lake was the reductive citrate cycle. It was noted that the reductive acetyl-CoA pathway was found to be most common mode of C fixation and regarded as the response to energy limitation (Hügler and Sievert, 2011). However, in our study, this C fixation pathway was not very strong. As we known, C metabolism pathways were one of the core metabolisms in soil bacterial community (Cai et al., 2020). Carbohydrate metabolism degraded complex organic matter into readily biodegradable substances and provided the necessary energy and matter for growth of bacteria. As a consequence, C fixation and methane metabolism and in Baiyangdian sediment were analyzed to reveal the pathways of organic matter conversion. For C cycle, organic matter was degraded by methanogenic archaea which used a battery of specific enzymes and coenzymes to cut CO₂ or methyl compounds down and produced CH₄ by means of biochemical reactions (Cheng et al., 2021). Moreover, in an aerobic state, organic matter could be broken down to produce CO₂ by microorganisms. Methanogenesis (M00357) in methane metabolism and the reductive citrate cycle (M00173) module in carbon fixation were further analyzed (Figure 6). The highest proportion of M00357 module was the main pathway of methane metabolism. In the methane metabolism M00357 module, acetate was converted to acetyl phosphate via ATP: acetate phosphotransferase (EC 2.7.2.1) again into acetyl-CoA via acetyl-CoA: phosphate acetyltransferase, or to acetyl-CoA via acetate: CoA ligase (EC 6.2.1.1). In addition, acetyl-CoA, coenzyme M-HTP heterodisulfide, 2-(Methylthio) ethanesul-fonate and 5-Methyl-5,6,7,8-tetrahydromethanopterin were interconverted through a series of enzymes and coenzymes. The highest proportion of M00173 module was the main pathway of C fixation. The same was, acetyl-CoA, Citrate and (3S)-Citryl-CoA were interconverted through C fixation process. M00357 module and M00173 module were interconnected by acetyl-CoA. Moreover, many key genes in C, N, P, S cycling were closely related with C fixation. This showed that there was a complex interaction between the metabolic pathways of C fixation and key genes of C, N, P, S biogeochemical cycling. For the P cycle, genes involved in regulation of P starvation (such as phoB, phoR) were prevalent in soil ecosystem. K01078 (acid phosphatase) and K03788 (acid phosphatase (class B)) belonging to phosphoesterase were not detected in Baiyangdian sediment metagenome, possibly meaning the absence of phosphoesterase metabolism in microbiota. Gene ppk was reported to be a phylogenetic marker of phosphorus removal communities and was widely used to investigate P removal organisms. In addition, it was closely linked the patterns underlying endogenous P release from sediments (Lv et al., 2015). So, the dominance of S8 of K00937 (ppk) in Baiyangdian sediment metagenome might be associated with the release of underlying endogenous P. The same change trend in the ten sediment samples was also shown in some genes with relatively high abundance, such as, pstABCS, phnCDE and ugpABCE. It might suggest that the strength of P cycling in Baiyangdian lake. And, high microbial potential for efficient phosphate uptake systems (for example, pstSCAB) was also detected in P depleted soil (Bergkemper et al., 2016). These genes with high abundance also were screened in Baiyangdian sediments. Hence, the high abundance of these genes in lake sediments might be a common phenomenon. Phosphonoacetate hydrolase (phnA) might increase TP content by breaking C-P bonds. This interesting conclusion was also found in our previous studies (Campos et al., 2022). Acid phosphatase phoN was involved in the hydrolysis of P and occupied an important position in global P cycling (Rana et al., 2020). Therefore, it might increase the biologically available P by increasing the activity of P-related enzymes, resulting in the decrease of soil P content.

Baiyangdian lake sediment was under weak alkaline conditions. Compared with the highly contaminated situation of Baiyangdian in 2018, TN content have dropped almost 40% and TP content have dropped about 35%, however it was still under eutrophication threat. TN showed a relatively greater impact on bacterial communities than TP. The status of OPPs was at an acceptable level of risk and was not the main source of TP. Proteobacteria was the most main phylum in sediment, and was proved to indispensable contribute to C, N, P, S cycling. At site S4, 3-Hydroxypropionate cycling of C fixation was the strongest while the weakest function of denitrification and complete nitrification of N cycling. Many genera involved in C, N, S, P cycling processes, such as, Thiobacillus, Nitrosomonas, Rhodoplanes of Proteobacteria; Nocardioides of Actinobacteria. Nocardioides had played an important role in many metabolic pathways of C, N, P, S cycling. Many key genes in C, N, P, S cycling were closely related to the module reductive citrate cycle. The relative abundance of each module of C-N-P-S cycling was consistent in each sample site, but its absolute abundance was different. However, the abundance of pathways of assimilatory nitrate reduction, denitrification and dissimilatory nitrate reduction of nitrogen cycling in sediments with higher TN content was higher than those with lower TN content. Besides, Nocardioides with plural genes encoding key enzymes involved in nasAB and nirBD gene were involved in these pathways. Carbon metabolism pathways was the core metabolisms in sediment bacterial community. In Baiyangdian lake, the methanogenesis process (involving genes such as acs, hdrA2…) was the main pathway of methane metabolism and the reductive citrate cycle module (involving genes such as nifJ, oorA…) accounted for the highest proportion of C fixation processes.

The data presented in the study are deposited in the NCBI repository, accession number PRJNA908081.

BK was responsible for writing manuscript. RX was responsible for writing and correcting manuscript. YH, YW, LZ, and ZW were responsible for measurement of environmental factor. JB, KZ, JA, MJ, and WP were responsible for determining research direction. All authors contributed to the article and approved the submitted version.

This study was financially supported by National Natural Science Foundation of China (NSFC) and National Research and Development Agency of Chile(ANID) for the China-Chile Joint Projects on Water Resources Management (grant no. 51961125201 in China and NSFC190012 in Chile), and Fuzhou University Testing Fund (project number: 2022T011).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.