3.1 Diversity of bacteria communities in riverbed sediment
In all studied 17 soil samples, 779,643 high-quality sequences were totally obtained from the V3-V4 region of 16S rRNA gene sequences clustered in 7,214 bacterial OTUs. The rarefaction curve of bacteria in the samples showed that when the number of reads reached about 3000, the Shannon indexes of all samples tended to be flat (Fig. 1a), and the curves tended to be flat, indicating that our sequencing depth was sufficient and can truly reflect the bacterial community in the sediment sample.
Understanding biological diversity is very important as it associate the function and stability of ecosystems (Yu et al. 2020). The bacterial diversity of the riverbed sediment associated with untreated and treated sediments (TC, g-C3N4 and TC/g-C3N4) analyzed to identify whether different treatments shaped the environmental microbiome (Fig. 1b). The treatment way had a little influence on the bacterial richness. The richness (Chao 1) of bacterial communities of H was 3901.70 lower than that of CK (4227.06), and lower than those of samples exposed to TC (4326.21–5068.42), g-C3N4 (4563.89–4939.65) and TC/g-C3N4 (4641.95–4962.17), respectively. However, the exposure of TC, g-C3N4 and TC/g-C3N4 almost not changed the diversity (Shannon index) of bacterial communities. Meanwhile, the Shannon in the samples containing TC/g-C3N4 remained comparatively stable in the range 6.81–6.92.
Based on Bray-Curtis distance, PCoA was applied to examine the beta diversity of the samples to study the differences of bacterial community structure among them (Fig. 1c). The first two axes (PCoA1 and PCoA2) explained 19.07% and 16.37% of the total variance in the sediment bacterial communities, respectively. A clear separation was observed among three sample groups exposed to TC, g-C3N4 and TC/g-C3N4. The difference between H and the samples treated by TC/g-C3N4 was littler than that between H and the sediments treated by TC alone, but larger than that between H and the samples handled with g-C3N4 alone. It might conclude that major changes in bacterial diversities across the samples were attributed to the difference treatments, g-C3N4 alone had little effect on microbial structure, while TC/g-C3N4 had medium influence and TC had great impact on it.
Five-Way Venn diagrams were constructed to understand the unique and shared OTUs among different treatments in each riverbed sediment (Fig. 1d). There was a high specific OTUs number in each sample group and the shared OTUs (1,898) accounted for 26.31% of the total OTUs (7,214). The number of specific OTUs in TC/g-C3N4 treatment group (713) was highest, followed g-C3N4 treatment group (228) and TC treatment group (181), suggesting that the specific richness of bacterial communities in samples exposed to TC/g-C3N4 was the higher than those exposed to TC and g-C3N4. These indicated that TC/g-C3N4 might be capable of increasing the existence of unique species which precisely affected the bacterial communities in sediment, increasing the Chao 1 index as depicted in Fig. 1b.
The TC in soil posed a high risk for bacterial communities (Pan et al. 2016), according to other researchers studies, TC could be used as C sources by surviving bacteria, produce TC-resistant bacteria, increasing the bacterial diversity of soil (Grenni et al., 2018; Ma et al., 2020; Ullah et al. 2019). The TC in environment could significantly restrain the microbe in its resistance spectrum, thus the bacteria adapting to TC gradually replaced those not acclimated the environment with TC (Zhang et al. 2013). Photocatalyst could produce some active groups, for example •O2− and •OH, which could act on microorganism and substances in environment, thus changed the diversity of bacteria. Hou et al. (2020) observed that g-C3N4 had good biocompatibility in the environment and had a positive effect on the sediment environment. Therefore, g-C3N4 could increase the richness of bacterial communities in the sediment, which was similar to the effect caused by other photocatalyst, such as TiO2 (Li et al. 2020). Up to now, a large number of studies on g-C3N4 composite photocatalyst inactivation of pathogenic bacteria, such as Staphylococcus aureus, Salmonella typhimurium, Escherichia coli (Heo et al. 2019; Tang et al. 2019).
Once external substances were added into sediment, there would be some corresponding changes in the environment, in order to cope with the varieties, the genetic characteristics, physiological and biochemical processes of some soil microorganisms will be changed. Thereby, some microbe was inhibited in sediment, while others in the environment may be stimulate to growth, which enhanced the generation and evolution of the microbes that had adapted to the conditions, corresponding promoted the number of them, and increasing the microbial diversity. At the same time, nutrient composition of sediment was changed by adding external material to the sediment, the structure of the bacterial community would be changed accordingly, thus varying the beta diversity. However, the changes in diversity induced by g-C3N4 was basically stable and similar to those induced by other photocatalysts, including biochar (Wu et al. 2019a), TiO2 (Li et al. 2020) implying that it had little influence on species diversity and might be safer for environment.
3.2 Composition of bacterial communities determined by Illumina MiSeq
According to the annotation and abundance of OTU, a relative abundance table for phyla was obtained (Table S2). The number of bacteria phyla hosted in H and CK was 36 and 40, respectively. In TL, TM, TH, there was 41, 43, 43 bacterial phyla, respectively; In PL, PM, PH, there was 45, 43 and 44 bacterial phyla, respectively; And in TLPL, TLPM, TLPH, TMPL, TMPM, TMPH, THPL, THPM, THPH hosted 42, 44, 43, 42, 44, 43, 42, 44, 43 bacterial phyla, respectively. The number of bacteria phyla in different treatment groups was similar, and in the same treatment group, the number was basically unchanged, implying the composition of bacteria phyla might not be effected obviously in the experiments.
The distribution of bacteria phyla in sediments handled by different ways were demonstrated in Table S3. The dominant phyla in all samples were Acidobacteriota, Proteobacteria, Actinobacteriota, Chloroflexi, and their share was about 70%. Uddin et al. (2021) found that the above four phyla were account for about 60% of soil bacteria by analyzing the effect of four different antibiotics on paddy soils. Though the influence of g-C3N4 on sediment had rarely examined till now, the four dominant phyla in our research were agree with their studies. According to other relevant reports, these four bacteria phyla accounted for a large proportion in water and soil (Chopra et al. 2001). and they were also the dominant phylum in sediment, although the abundance of each phyla was various in different sediment (Welz et al. 2018). Due to the high proportion of the advantage bacterium group in the sediment, they had an important effect on maintaining the stability of community structure in different treatments.
The effect of TC on bacterial community structure in sediment was illustrated in Fig. 2a and Table S3. Compared to CK, the abundance of some microorganism were differ from the samples treated by TC, especially the dominant bacteria, such as Actinobacteriota, Acidobacteriota and Firmicutes, that is TC had a greater impact (p < 0.05) on them. However, between different concentrations of TC, the changes of relative bacteria abundance at phyla level was insignificant (p > 0.05) between different concentrations of TC. The average abundance of Actinobacteriota, Acidobacteriota and Firmicutes was significantly decreased in TC treatment (p < 0.001), and its abundance was 17.33%, 14.14% and 6.04%, respectively, however, in CK, the relative abundance was 25.05%, 16.07% and 8.68%, respectively, so these three phyla were regarded to be susceptible to TC application. The results were consistent with other reports, Alexandrino et al. (2017) also found that Actinobacteriota and Firmicutes were sensitive to TC.
The influence of g-C3N4 on bacterial community structure in sediment was demonstrated in Fig. 2b and Table S3. Compared with the CK, the presence of g-C3N4 significantly inhibited (p < 0.001) the growth of Firmicutes and Actinobacteriota in riverbed sediment. But the growth of Acidobacteriota (20.89%) was significantly improved (p < 0.001) by g-C3N4, and the average abundance was increased about 4.82% in comparison of CK. In the PL treatment group, the biomass of Cyanobacteria (0.88%) was significantly lower (p < 0.001) than that of CK (3.50%), this was indicated that g-C3N4 might inhibited the growth of Cyanobacteria. Based on the above results, it could be concluded that Firmicutes, Actinobacteriota, Acidobacteriota and Cyanobacteria were susceptible to g-C3N4 exposure. Cyanobacteria was photoautotrophic organisms in aquatic ecosystems and ubiquitous in riverbed sediment, as well as a typical microorganism in the identification and detection of environmental pollution (Teta et al., 2019). Acidobacteria had the potential to degrade polymeric carbonaceous complexes and actively participate in the cycling of organic matter, and it played an important role in maintaining the structural stability of microorganism in the environment (Kalam et al. 2020). Cyanobacteria proliferate in large quantities, which could affect the microbial structure in the environment (Chen et al. 2020). Our results showed that g-C3N4 could promoted Acidobacteriota, inhibited Cyanobacteria, and was beneficial to environmental pollution remediation. However, The growth of Cyanobacteria was effected by g-C3N4 concentration, its abundance of PH treatment (6.34%) was higher than that of PL treatment (0.88%), it might due to the automatically agglomerate of g-C3N4 at high concentration, inactivating its function.
The impact of TC/g-C3N4 on bacterial community structure in sediment was illustrated in Fig. 2c and Table S3. Compared with Actinobacteriota (25.05%) and Firmicutes (8.68%) in the CK treatment group, the average abundance of Actinobacteriota (16.48%) and Firmicutes (4.97%) was significantly decreased (p < 0.001) in TC/g-C3N4 treatment. But, compared with the CK treatment groups (16.07%), the average abundance of Acidobacteriota increased significantly (p < 0.05) in TC/g-C3N4 treatment groups (20.71%). When the same does of g-C3N4 was added, the biomass of Cyanobacteria treated with low concentration of TC (TLPL = 4.96%, TLPM = 4.14%, TLPH = 3.81%) was higher than that treated with high concentration of TC (THPL = 3.71%, THPM = 2.76%, THPH = 2.16%). Meanwhile, the biomass of Chloroflexi (11.79% − 13.73%) and Gemmatimonadota (3.23% − 5.32%), which were the dominant microorganism in the sample, could be maintained in a stable range with the addition of both TC and g-C3N4, and the biomass of them were similar to that of H group, respectively. Acidobacteria was very important for the environment to keep the structural stability of microorganism as it potentially involved in the degradation of polymeric carbonaceous complexes and the cycling of organic matte (Kalam et al. 2020). Therefore, it was regarded as a kind of potential microorganism for environmental bioremediation and biotechnological applications (Kielak et al. 2017). Our study indicated that g-C3N4 could significantly slow down the growth of Cyanobacteria and promote the growth of Acidobacteriota in the environment under TC pressure, thus it might beneficial for sediment to remediate TC pollution and haven’t adverse effect on environment.
3.4 Taxonomy-based comparisons of microbiota groups
To identify the biomarkers in sediment samples, the microbial communities in different treatments was compared based on taxonomy, and the results were showed in (Fig. 3), in which LEfSe was applied to determine each group that was revealed in cladograms and histogram of LDA scores of 2.5 or more. The larger the score of LDA was, the more remarkable the difference caused by species abundance was. As depicted in the figure, there were obvious changes in dominant bacteria in sediment treated with TC, g-C3N4 and TC/g-C3N4 at the level of phylum, class, order, family and genus level. In the sediment treated with TC, more than 20 microbes including Actinobacteriota, Desulfuromonadia, Myxococcia were significantly enriched (p < 0.05). Similarly, 8 groups of bacteria including Brocadiae, Entotheonellia, Dependentiae were obviously enriched (p < 0.05) in riverbed sediment treated with g-C3N4. However, in the sediment treated with TC/g-C3N4, only 4 microbes were significantly enriched (p < 0.05), such as Babeliae, Subgroup_25 and c_unclassified_p__Actinobacteriota. These indicated that the selected biomakers could clearly distinguish (p < 0.05) these three treatments and the difference caused by TC was the most, orderly followed by g-C3N4 and TC/g-C3N4. TC mainly affected the bacteria with high abundance, while g-C3N4, TC/g-C3N4 mainly influenced the low abundance ones. The results suggested that g-C3N4 might mitigate the side effect caused by TC on sediment bacteria by reducing the difference of microbes in the environment.
3.5 TC residue and Changes in species at genus level caused by TC and g-C3N4
In order to study the effect of TC residue on bacteria community structure, we analyzed the TC concentration in sediment with different treatment, as well as the relative dominant bacterial genus (Fig. 4). As illustrated by Fig. 4a, there were no TC residues were detected out in CK, PL and PH, suggesting that the original TC residue (0.22 mg/kg) came from H were degraded in these samples. In TH, THPL and THPH, the concentration of TC s was 22.68, 14.42 and 7.83 mg/kg respectively. Most of the TC added into the sediment samples were degraded by riverbed sediment, since there were only 22.68 mg/kg were existed in TH after 30 days. The TC concentration in THPL was 8.26 mg/kg lower than that in TH, and higher than that in THPH, indicating that in samples, g-C3N4 was also able to degrade TC and the degradation ability was improved by its concentration. g-C3N4 could effectively degrade TC in the water environment by some active groups, including •O2− and •OH. (Guo et al. 2017; Zhu et al. 2019b). According reports, under the irradiation of visible light, the g-C3N4 could degrade about 40–90% TC in 30–120 min in water (Guo et al. 2019; Panneri et al. 2017), which was much higher than that in our experiment. This phenomenon might attribute to two facts: Firstly, the light intensity under our experimental condition was weak, so only a small number of photon were available for the degradation. Secondly, the composition of the sediment system was complex, which effected the light transmittance and further effected the photocatalytic activity of g-C3N4.
Figure 4b and Table S4 demonstrated the dominant bacterial genus in different treated sediment. Compared with CK group, under the pressure of TC, the abundance of Bacillus (Firmicutes) decreased by 3.27%, while the biomass of RB41 (Acidobacteriota) only increased by 0.57%. The biomass of Bacillus (Firmicutes) in PL (2.21%) and PH (2.47%) was lower than that of CK (5.06%), however, the biomass of RB41 (Acidobacteriota) in PL (2.91%) and PH (2.00%) was higher than that of CK (0.89%). Compared with the TH group, the biomass of RB41 in THPL and THPH group was significantly increased (p < 0.001) to 3.22% and 2.77%, respectively, and the abundance of Bacillus in THPL and THPH group was increased. Under TC exposure, g-C3N4 was beneficial to enrichment of norank_f__Vicinamibacteraceae and norank_f__Gemmatimonadaceae, belonging to the bacteria phyla of Acidobacteriota and Gemmatimonadota, respectively
Previous research had showed that RB41 was actively participated in the carbon cycle in sediment (Ito et al. 2019), RB41 was the dominant and sensitive microorganism in contaminated soil and played a positive role in environmental ecosystem (Ai et al. 2018; Shen et al. 2018). Based on the fact that its abundance increased in the treatment of TC and g-C3N4, it should be sensitive to them, and more to g-C3N4. Bacillus could fully degrade organic matter and eutrophic substances in sewage (Shen et al. 2020), at the same time, Bacillus was also a kind of beneficial bacteria that was beneficial to soil microecological stability. It could inhibit or kill pathogenic bacteria in the environment (Nicholson et al. 2002). It had been reported that antibiotics induced oxidative stress and inhibited the growth of Bacillus cells (Sannasimuthu et al. 2020). Based on the fact that compared to CK,the degree of decline in the content of Bacillus in TH was greater than that of samples contained g-C3N4, it was suggested that TC had obvious toxic effect on Bacillus, while g-C3N4 could reduce the toxic effect of TC on Bacillus. Hence, one inference was that the g-C3N4 might reduce the effect of TC on Bacillus in sediment.
In our experiment, we founded that the addition of g-C3N4 would reduce TC residue in sediment, and this trend was strengthened with the increase of the concentration of g-C3N4 (Fig. 4a). These results indicated that 1) g-C3N4 could degrade some organic substances, including TC, and changed the nutrient composition of sediment, thus affecting the structure of microorganisms; 2) g-C3N4 could act on some microorganisms, change the growth of sensitive bacteria, thus affecting their ecological niche in the sediment. Based on the effect of g-C3N4 on dominant bacteria RB41 and Bacillus, it was indicated that g-C3N4 could improve the ability of sediment remediation and reduce the toxic effect of TC on beneficial bacteria, which was beneficial to the ecological health of sediment.
3.6 Redundancy analysis (RDA) of bacterial community at the phylum level and class level
The relationships between TC, g-C3N4 and bacterial community was analyzed based on RDA score plot and depicted in Fig. 5. The impacts of each factor (TC or g-C3N4) on bacterial community were represented by the length of arrows, and the cosine angle between arrows illustrated their relationship (smaller angle indicated more significant correlation). As shown in Fig. 5a, at the phylum level, axis 1 and 2 of the RDA plots explained up to 19.37% and 3.45%, respectively. Specifically, TC positively affected (p < 0.05) Chloroflexi, Proteobacteria, Myxococcota, Gemmatimonadetes, Actinobacteria and Firmicutes in riverbed sediment. Cyanobacteria had no significant correlation (P < 0.05) with TC. Previous research had shown that Firmicutes and Bacteroidetes were significantly positively correlated to total antibiotics, while Acidobacteria was significantly negatively correlated to them (Gao et al. 2020). Studies had indicated that Proteobacteria and Actinobacteria were dominated community of the TC resistant members (Pala-Ozkok et al. 2019), and Proteobacteria taken a great part in pollutant degradations and environmental remediation (Xu et al. 2020). The concentration of g-C3N4 in riverbed sediment was negatively correlated (P < 0.05) with Actinobacteria, Myxococcota, Firmicutes and Gemmatimonadetes and positively correlated (p < 0.01) with Acidobacteriota, Proteobacteria, Cyanobacteria and Chloroflexi.
The RDA analysis for the relationship between TC or g-C3N4 and bacteria at genus level was illustrated in Fig. 5b, the first axis explained 17.73% of the total variance, whereas the second axis accounted for 4.48% of the total variance. It is noteworthy that TC has significant correlation (p < 0.05) with Gaiella, Nocardioides and Bacillus. Previous studies had also shown that Gaiella at genus level were significantly positively correlated to the total antibiotics (Gao et al. 2020). Nocardioides were crucial for bioremediation, notably, dehalogenation and denitration (Ito et al. 2019). However, g-C3N4 exhibited negatively correlations (p < 0.05) with Gaiella, Nocardioides and Bacillus. Gaiella was the sensitive genera that was negatively correlated to antibiotics perturbation (Uddin et al. 2019), g-C3N4 had significant positive correlation (p < 0.05) with RB41, which was positively correlated with soil carbon content and actively participated in the carbon cycle in soil environment (Ito et al. 2019). g-C3N4 had significant positive correlation (p < 0.05) with RB41, which was positively correlated with soil carbon content and actively participated in the carbon cycle in soil environment.
Using Spearman correlation analysis, the relationship between the top 30 bacterial genus and TC or g-C3N4 were examined to further investigated the influence of TC or g-C3N4 on bacteria community (Fig. 2s). According to the heatmap, TC and g-C3N4 significantly correlated with the major phyla and genus. Acidobacteriota was the main phyla positively affected by g-C3N4 (p < 0.01), Chloroflexi was the main phyla affected by TC (p < 0.05). RB41 (p < 0.05) and Ellin6067 (p < 0.001) were the main genus significantly associated with g-C3N4. These results were consistent with the results obtained with the RDA.