3.1 Cadmium accumulation kinetics in constructed wetlands
Several recalcitrant pollutants, including heavy metals, are found in aquatic ecosystems. Though, our understanding of what happens to Cd in a wetland ecosystem that supports submerged macrophytes is limited. The concentration and removal rates of Cd were measured following a 30-day exposure of water, sediment, and submersed macrophyte leaves (V. natans and P. maackianus) to the metal. There was a decrease in Cd concentration in water (93.8–98.3%) (Fig. S1 and Table S2-S3) and submersed plant leaves (P. maackianus = 8.09–21.9%, V. natans = 16.03–26.17%, and Artificial plant = 0%), while sediment samples showed Cd accumulation (+ 55.5–83.3%) (Fig. S1 and Table S2-S3). Notably, the highest decrease for Cd in the leaves was for V. natans, whereas artificial plants revealed the highest Cd concentration in sediments. This study suggests that submerged plants and their epiphytic bacteria reduced Cd from the overlying water after 30 days of treatment, whereas sediments are considered as sinks or repositories of various pollutants(Guo and Yang, 2016). The high Cd uptake by V. natans in this study may be due to its potential to remediate recalcitrant pollutants in constructed wetlands, while artificial plants lack innate features facilitating them to filter pollutants into the sediments, consistent with previous reports (Chen et al., 2023; Ohore et al., 2021a). Conversely, Qiu et al. (2021) revealed that artificial plants had better performance in nutrients removal (e.g., TN, TP, NH4-N, and COD) than Myriophyllum Spicatum. Thus, the difference in Cd removal rate may be attributed to Cd concentrations and aquatic plant types.
Figure 1
3.2 Cadmium impeded the growth of epiphytic biofilm
Cd had an impact on the development and colonization of epiphytic bacteria. Figure 1 displays the findings of the qualitative analysis of confocal laser scanning microscopy (CLSM) data. Different outcomes for epiphytic bacteria exposed to Cd and control groups were seen when SEM micrographs were visualized: different microbial groups, including algal cells (diatoms), cocci/cocci-like and bacilli/bacilli-like microorganisms, were displayed in SEM micrographs from artificial plants, V. natans, and P. maackianus. Cadmium considerably decreased the microbial aggregates (Fig. 1). It is worth mentioning that epiphytic biofilm microbial assembly on the submersed plant leaf surfaces declined as Cd concentration spiked. The decline in microbial aggregates on submerged plants with increased Cd concentration may be attributed to the Cd toxicity and allelopathic effect (natural plants) on bacteria growth and colonization (He et al., 2021; Huang et al., 2020). Additionally, bacterial and algal cells were mostly encapsulated in an EPS structure that has been densely produced on artificial plants (AP) and P. maackianus (P) with 0.5mg/L Cd concentration, consistent with prior reports(Chen et al., 2023; Ohore et al., 2021b). The presence of more epiphytic microbes on artificial plants than on V. natans may be due to the artificial plants’ inability to discharge allelochemicals, which impede microbes on the plastic surface.
Surprisingly, the bacterial cells (e.g., bacilli/bacilli-like) in VN0.5, AP1, and AP2 become elongated, while AP0.5 and P0.5 exhibited a high EPS production. These findings suggest that cell elongation and EPS production could be attributed to a bacterial defense mechanism against Cd toxicity and Cd-resistance genes dissemination, in agreement with previous studies (Chen et al., 2023; Pu et al., 2021). A reversal mechanism or resistance of the microbes against Cd stress could be the reason for this observation. Notwithstanding the hardness of the diatom silica cell wall, there was a rupture of the algal (diatoms) cell wall on V. natans (VN1 and VN2) and P. maackianus (P1&P2).
Figure 1
3.3 Cadmium application lowered the capacity of wetlands to remove nutrients.
Generally, this study demonstrated a significant decrease in nutrient removal capacity in the Cd treatment groups compared to the Control groups (p < 0.05) (Fig. 2). In week 1, the average removal rate recorded after 30 days increased from 6.9 to 25.5% for TN, but it fell from 5.7 to 50% in the subsequent three weeks (Fig. 2B). Besides, the average TP removal rate increased from 16 to 30%, while in the following three weeks, the removal rate decreased from 12.5 to 97.1% (Fig. 2A). Moreover, the average COD removal rate increased from 4.2 to 14.28% in weeks 1 and 4, whereas it dropped from 9.45 to 21.8% in the second and third weeks (Fig. 2E). Notably, the TN, TP, and COD removal rates also decreased with increasing Cd concentrations. The decrease in TP removal efficacy can be attributed to the limitation in phosphorus removal, which is a result of decreased enzyme activities caused by Cd (Chen et al., 2014). The declined removal capacities of TN and COD are likely due to the effects of Cd stress on the epiphytic biofilm and the microbial communities in the wetland's mesocosm, which are responsible for nutrient metabolism and degradation (Li et al., 2018; Yang et al., 2019), and this finding aligns with those of Ohore et al. (2021b). Whereas the increased COD removal rate with exposure time is possibly due to Cd inactivation in the mesocosmic wetland.
Cd application significantly increased the DO and EC with exposure time, while decreasing the pH in the water column (p < 0.05) (Fig. 2C-D &F). The increase of DO may possibly due to the oxygen release by submersed plant photosynthesis (Pedersen et al., 2013). The spike in EC values may be due to nutrient retention in the water column (Ohore et al., 2021b). The high Cd concentration can exert a deleterious impact on the submerged plant- epiphytic biofilm systems, which are involved in the sequestration of ions from the surrounding water column. Whereas the increase in pH may be ascribed to CO2 emission from the rhizosphere of submerged macrophytes to water, as a result of reduced biofilm and increased chlorophyll contents(Ohore et al., 2021b). The findings of this study collectively indicate that Cd significantly affected the water quality and hindered the nutrient removal capacity of wetlands.
Figure 2
3.4 Cd exposure effect on epiphytic bacteria diversity
To investigate the impact of Cd exposure on bacterial diversity in the leaves of V. natan’s, Potamogeton maackianus, and artificial plants, Illumina sequencing was utilized, and a uniform number of 1988 clean 16S raw sequences were obtained in all samples. Table S1 shows the Shannon, Simpson, Ace, Chao, and Coverage values for all bacteria samples. These alpha diversity indices (Fig. 3A) revealed no significant difference (p > 0.05) between the treatment groups (VN0.5, VN1, VN2, P0.5, P1, P2, AP0.5, AP1, and AP2) regardless of Cd exposure, consistent with a previous study (Mu et al., 2018). However, most α-diversity indices were significantly different (p < 0.05) in epiphytic biofilms and intestinal microbiota under Cd exposure (Huang et al., 2020; Mu et al., 2018). The findings indicate that the epiphytic biofilm may have sustained its microbial diversity profile to counteract the impact of Cd stress and/or the present metal stress may not have been substantial enough to cause notable alterations in the alpha diversity. This could potentially be attributed to factors such as Cd concentration, exposure duration, and host species.
The principal coordinate analysis (PCoA) for the bacterial community explained approximately 50.63% of the overall variation. The analysis demonstrated that there were significant statistical differences in the dissimilarity of bacterial communities across the samples (p < 0.05) (Fig. 3B). These results suggest that Cd pressure drives important roles in the structure of bacterial communities.
Figure 3
3.5 The impact of Cd exposure on the composition of epiphytic bacteria
The impact of Cd exposure on bacterial composition was examined. (Fig. 4). The relative abundance of epiphytic bacterial species (OTUs) in the control group (SM_Cd_0mg/L) declined as Cd concentration rises (Fig. 5A). The results showed that the control group had a greater number of bacterial OTUs compared to the Cd treatment groups (SM_Cd_0.5mg/L, SM_Cd_1mg/L, and SM_Cd_2mg/L) (Fig. 4A). This suggests that the high Cd concentrations had a detrimental impact on the epiphytic bacterial communities of V. natans, P. maackianus, and artificial plants, which is consistent with recent studies highlighting the impact of different heavy metal concentrations on microbial composition in tadpole gut (Mu et al., 2018). Throughout the experiment, the 10 most abundant epiphytic bacterial phyla, on average, were selected for relative abundance analysis (Fig. 4B). The four dominant phyla in all samples were Bacteroidota (51.8%), Gemmatimonadota (49.9%), Acidobacteria (49.01%), and Actinobacteria (48.05%) (Fig. 4B), corroborating previous investigations (Huang et al., 2020; Mu et al., 2018).
Notably, Proteobacteria (14.4-50.52%), Firmicutes (4.5–68.3%), Cyanobacteria (16.4–46.1%), and Actinobacteria (71.5–93.6%) decreased under Cd exposure in P. maackianus treatment groups compared to the control group. In contrast, Gemmatimonadota, Acidobacteria, Myxococcota, and Chloroflexi showed a reverse trend (Fig. 4B). Furthermore, Gemmatimonadota (28.8-96.16%), Acidobacteria (62.5–75%), Verrucomicrobia (58.3–77.3%), and Chloroflexi (50%) decreased under Cd exposure in groups treated with artificial plants in comparison to the control group, whereas Bacteroidota, Proteobacteria, and Firmicutes spiked under Cd exposure in the same substrate. On the other hand, Proteobacteria (11.3–51.8%), Cyanobacteria (6.08–94.35%), Actinobacteria (55.9–86.6%), and Verrucomicrobia (37.8-87.01%) decreased under Cd application in V. natans treatment groups relative to the control, while Bacteroidota, Gemmatimonadota, and Firmicutes increased under Cd exposure in the same submersed plant. Huang et al.(2020) reported that Cd exposure instigated marked changes in the bacterial composition of V. natans and Nile tilapia. The increase/decrease or discrepancy in bacterial community composition under Cd loading noted in this study may be due to the heavy metal concentration and substrate type.
The Firmicutes and Bacteroidota relative abundances were significantly higher in V. natans (p < 0.05) compared to P. maackianus, however, Actinobacteria and Chloroflexi in P. maackianus revealed a reverse trend (Fig. S2A). Additionally, Gemmatimonadota, Acidobacteria, and Myxococcota were significantly higher in P. maackianus (p < 0.05) relative to artificial plants, while Firmicutes, Verrucomicrobia, and Cyanobacteria were significantly (p < 0.05) higher in artificial plants than P. maackianus (Fig. S2B). Conversely, Verrucomicrobiota, Actinobacteria, Chloroflexi, and Proteobacteria, were of higher significant (p < 0.05) occurring phyla in artificial plants than V. natans, though Gemmatimonadota, Acidobacteria, Myxococcota, and Bacteroidota exhibited significant dominance (p < 0.05) in V. natans relative to artificial plants (Fig. S2C). The variation in bacterial genera under Cd application is depicted in Fig. 4C. The dominant genera, namely g__Aquabacterium, g__Flavobacterium, g__Rhizobium, g__Rhodobacter, g__Comamonas, g__Gemmatimonas, g__Pseudomonas, g__Ideonella, and g__Porphyrobacter, accounted for 23.6%, 15.2%, 10.4%, 9.6%, 8.1%, 6.7%, 5.1%, 4.3%, and 3.8% of the bacterial genera in all samples, respectively (Fig. 4C). Notably, there is an increase in relative abundance of g__Nostoc (31.9–33.1%), g__Gemmatimonas (31.7%), g__ Rhizobium (27.1%), and g__Delftia (33.02–33.1%) in VN, AP, and P, whereas g__ Rhizobium (23.4%), g__Comamonas (21.4-22.89%), g__Leptothrix (24.35–24.6%), and g__Gemmatimonas (18.1%) decreased in AP and P. Bacterial communities inhabiting harsh ecological niches may survive only by acquiring resistance genes (e.g., heavy metal resistance genes), ensuring their survival and leading to an increase in their relative abundance. Conversely, the decrease in the relative abundance of bacterial phyla and genera might be attributed to the detrimental impact of high Cd concentrations on epiphytic bacterial communities (Mu et al., 2018). This could explain the observed variations in bacterial abundance. Firmicutes comprise pathogen-causing diseases (Fan et al., 2018), while Cyanobacteria sustain bacterial growth (Jiang et al., 2019). Chloroflexi is a slow-growing oligotrophy (Eo and Park, 2016; Will et al., 2010), whereas Actinobacteria and Acidobacteria show high metabolic versatility (Alvarez et al., 2017) and actively thrive in stressed conditions (Fierer et al., 2007). As a result, these bacterial phyla were linked to the ability to withstand heavy metal exposure and thrived.
Figure 4
3.6 Cd alters the bacterial network structure
To reveal the structure and interaction of the underlying bacterial communities in both control groups (AP0, PM0, and VN0) and Cd treatment groups (AP_Cd, PM_Cd, and VN_Cd), co-occurrence networks were constructed using the top 48 dominating epiphytic bacterial OTUs. (Fig. 5). A total of 391 edges were identified by 48 nodes (OTUs), 412 edges by 48 nodes, 469 edges by 48 nodes, and 423 edges by 48 nodes in CT, AP_Cd, PM_Cd, and VN_Cd, respectively (Fig. 5A-D).
The impact of Cd exposure on bacterial interactions was investigated by computing topological properties. The modularity and average degree were 2.77 and 17.16 in the AP_Cd group, 1.11 and 19.54 in the PM_Cd group, and 4.3 and 17.62 in the VN_Cd group. According to Scheffer et al., 2012, modularity represents the system’s capacity to resist environmental influences, whereas the average degree indicates the strength of the connection between a node and its neighboring nodes. Epiphytic bacterial interactions were more intense and susceptible in P. maackianus than in V. natans and artificial macrophyte biofilms, as shown by a lower average degree and higher modularity in the AP_Cd and VN_Cd groups (Fig. 5B-C) than the PM_Cd group (Fig. 5D)(Montoya et al., 2006). In addition, a high modularity values in VN_Cd and AP_Cd groups than PM_Cd group revealed that AP_Cd and VN_Cd networks have higher modular class (> 0.4) (Zhang et al., 2019). This study demonstrates that the networks of V. natans and artificial plants were more stable and complex than the network of P. maackianus. Notably, there were more edges in VN_Cd than AP_Cd, suggesting that bacterial communities in artificial plants inhibited the bacterial community stability than in natural submersed plants. These findings denote the improvement of bacterial community structure on V. natans than artificial plants under Cd loading, consistent with a recent study by Mu et al. (2021). Despite the existing research on the effects of heavy metals on natural macrophytes, studies on their impact on epiphytes in the plastisphere remain insufficient, necessitating further research. All networks’ nodes (OTUs) were connected with four bacterial phyla. Among these, three (Proteobacteria, Bacteroidota, and Cyanobacteria) were prevalent, making up a significant proportion of 95.84% of nodes (Fig. 5A-D).
Based on the average degree, Flavobacterium and Sphingorhabdus (Bacteroidota), Acinetobacter, Aeromonas, Hydrogenophaga (Proteobacteria), and Exiguobacterium (Firmicutes) were the keystone taxa in V. natans and P. maaackianus, while Novosphingobium, Sphingomonas, Comamonadaceae, and Phenylobacterium were keystone taxa in artificial plants. The bacterial keystone genera in submersed macrophytes and artificial plants are important for plastic degradation and colonization (e.g., Exiguobacterium, Novosphingobium, Sphingobacteria, and Flavobacterium) (Dussud et al., 2018; Oberbeckmann and Labrenz, 2020; Yang et al., 2015), heavy metal biosorption (e.g., Aeromonas)(Qurbani and Rafiqi, 2022), bioremediation of xenobiotic compounds (e.g., Acinetobacter and Phenylobacterium)(Shelly et al., 2020). These findings indicate that the epiphytic bacteria can survive in environments contaminated with xenobiotics, such as heavy metals, hydrocarbons, and herbicides, and may have developed heavy metal resistance mechanisms to survive. Overall, the co-occurrence networks analysis displayed more stable and complex interactions among microorganisms on V. natans and artificial plants than P. maackianus, and those interactions include cooperation, competition, symbiosis, and parasitism among organisms in biofilms.
Figure 5
3.7 Cd altered bacterial metabolic pathways in epiphytic biofilm
KEGG annotation was used to investigate how Cd influences the metabolic pathways of the bacterial community in the epiphytic biofilms (level 2) (Fig. 6). The findings of this study suggest that the bacterial community’s primary functional trait was metabolism, with prevalent metabolic pathways linked to energy metabolism, carbohydrate metabolism, amino acid metabolism, cofactors, and vitamins metabolism, signal transduction, membrane transport, translation, nucleotide metabolism, lipid metabolism, and xenobiotic biodegradation in both artificial and natural submersed macrophytes (Fig. 6). It should be noted that control samples showed high percentage values of metabolic pathways compared with treatment samples, consistent with our recent study (Chen et al., 2023).
According to the findings, Cd concentration decreased along several metabolic pathways in epiphytic bacterial populations, presumably as a result of Cd’s toxic effects or selective pressure. However, Cd loading in VN1 and P2 triggered a spike in 16 metabolic pathways, which may be due to Cd diffusion impediment by EPS and resistant bacterial strains formation (e.g., Cd-resistant bacteria)(Abbas et al., 2018; Gebreyohannes et al., 2019; Hui et al., 2022).
From an evolutionary point of view, bacteria exposed to heavy metals spontaneously evolved numerous toxic metal homeostasis systems such as uptake, efflux, metallochaperones, detoxification, and sequestration, allowing bacteria to resist or remediate toxic heavy metals, especially Cd (Hui et al., 2022; Kim et al., 2018). These homeostatic mechanisms will eventually safeguard their survival, which explains the observed spike in the metabolic pathways.
Comparing the top 18 metabolic pathways, 11 in P. maackianus, 7 in V. natans, and 7 in artificial plants exhibited a significant decline (p < 0.05, Fig. S3A-C) when Cd concentration increased from 0-2mg/L. Noticeably, the metabolism (e.g., carbohydrate, lipid, amino acid, energy, amino acids, nucleotides, cofactor and vitamins, terpenoids and polyketides), xenobiotic biodegradation, membrane transport, signal transduction, translation, and cell motility (Fig. S3A-C) were significantly lower in heavy metal treatment groups than the control. The predicted functional analysis revealed that the epiphytic biofilm ecosystem’s bacterial metabolic function was reduced, indicating the negative effect of Cd stress. For instance, the deterioration in cell powerhouse ATP generation under Cd stress may disturb the bacterial communities in the epiphytic biofilm of P. maackianus, V. natans, and artificial plants. Additionally, the decline in the translation process inhibits protein synthesis induced by Cd stress. These findings shed new light on the adverse effects of heavy metals on aquatic ecology.
Figure 6