Introduction

Wastewater pollution is a threat to both people and aquatic life and contributes to the largest percentage of water-based pollution worldwide. Globally, an estimated 80% of wastewater, which includes domestic and industrial effluents, is discharged into the environment without treatment, releasing an array of harmful contaminants and causing direct harm to aquatic ecosystems1,2,3. Most treatment measures are often limited by economic constraints and difficulty in obtaining professional expertise, making them infeasible in developing countries4. Constructed wetlands (CW) are organic wastewater treatment systems that mimic the ecological functions of natural wetlands and provide an economical, efficient and sustainable way of treating wastewater5,6. Moreover, due to the advantages of low energy requirements, easy operation and maintenance, CW technology has been adopted worldwide as a green technology for environmental wastewater pollution treatment7.

In CW systems, microorganisms play key roles in pollutant removal, such as the degradation of organic pollutants and the removal of excess nutrients8. Moreover, the ability to regulate chemical influx within CWs is found to be strongly linked with microbial diversity and community composition9. For instance, Lee et al.10 reported that bacterial community structure and the specific bacterial consortia within wetlands determine the denitrification potential in CWs. Interestingly, the soil matrix of CWs contains different microbial fauna, viz., aerobic zones are rich in microbial diversity which assists in metal oxidation (bacteria fixes inorganic compounds by reacting only with its organic content) while the latter anaerobic zones are rich in sulfate-reducing bacteria11,12. Apart from decomposition and biodegradation, microorganisms in CWs can also remove organic compounds from wastewater through biosorption, bioaccumulation and speciation transformation13. In addition, microorganisms can improve the remedial capabilities of CWs to pollutants by enhancing phytoremediation14,15. As such, to understand how to further optimise CWs, it is necessary to investigate the microorganisms associated with remedial properties.

To address some of these knowledge gaps, we undertook a study in Sardar Bherry, situated in the East Kolkata floodplain wetland, India, as the model ecosystem to investigate the functionality of a CW receiving wastewater. The East Kolkata wetlands are classified as a site of community importance and included in the list of wetlands of international importance (“Ramsar List”), under the Ramsar Convention, 200216,17. These wetlands are one of the most important engineered ecological systems in India that serve as natural wastewater treatment plants to reuse/treat urban wastewater flowing out daily from Kolkata, India. Hence, our goals were: (1) to determine the spatial succession of the microbiome across wetland from the input site (sewage) to the output site (natural waterways) through the isolation of bacterial specimens; (2) explore the environmental factors that drive variations in bacterial biochemical characteristics; (3) investigate the bacterial contributions to the health of the waterways and aquatic life forms within; (4) evaluate the presence of pathogens, and how the constructed wetland non-pathogenic microbes exhibit inhibitory activity against them; and (5) investigation on the presence of multiple antibiotic resistance (MAR) bacteria in the constructed wetland.

We employed a straightforward cultivation procedure along with unique morphology selection to gain further insights into diversity-driven biology. This ‘deep-cultivation’ approach focused on selectively obtaining pure cultures from all observed unique morphology present within the bacterial isolates. Our approach also investigated the relationship of bacterial isolates with host wellbeing, as well as their involvement within the environmental processes. We unearthed unexpected bacterial traits and demonstrated that there is far more ‘out there’ in terms of biogeochemical cycling, aquatic animal health and immunity, antimicrobial activity and potential bacterial pathogens.

Results

Cultivation and characterisation of 32 bacterial isolates previously undescribed from constructed wetland system

We based our sampling strategy on the hypothesis that uniqueness in morphology, colour and size of bacterial colony are likely to be representative of a divergent bacterial group. A carefully developed culture strategy was employed to target beneficial bacteria, viz., Bacillaceae, Enterobacteriaceae and Aeromonadaceae inhabiting the sampled sediments of the constructed wetland. Using this method developed for selection, we were able to obtain a total of 32 strains in axenic cultures from the constructed Sardar Bherry wetland (Supplementary Figs. 1 and 2). The phylogenetic diversity of the recovered isolates based on 16S rRNA amplicon sequencing data showed that the isolates were from Bacillaceae, Burkholderiaceae, Enterobacteriaceae and Aeromonadaceae. The bacterial isolates include 26 Bacillus species, and one each of Ralstonia sp., Citrobacter freundii, Aeromonas veronii, Enterobacter cloacae, Burkholderia cepacia and Priestia flexa. Most isolates cluster within Bacillaceae, a clade known to confer positive impact on environment and have health benefits. Interestingly, we also found several members of Burkholderiaceae (Ralstonia sp. and Burkholderia cepacia), Enterobacteriaceae (Citrobacter freundii and Enterobacter cloacae) and Aeromonadaceae (Aeromonas veronii) from our sediment samples of wetland (Fig. 1, Supplementary Fig. 3 and Supplementary Table 1). The 32 bacterial isolates were grouped into five categories based on the origin of sediment samples, viz., sewage canal (SC) - 7 isolates, site A - 9 isolates, site B - 6 isolates and site C - 6 isolates and outlet canal (OC)- 4 isolates (Table 1).

Fig. 1: Isolation and phylogenetic comparison of bacterial isolates.
figure 1

Phylogenetic tree of 32 bacterial isolates based on 16S rRNA nucleotide sequences following maximum composite likelihood method by the MEGA11 software. The numbers next to the branches indicate percentage values for 1000 bootstrap replicates. Bootstrap values above 50% are shown at the nodes. The isolates were categorised into 5 clusters indicated by colour shading.

Full size image
Table 1 Sources of isolates with MAR indices of Sardar bherry wetland.
Full size table

Environmental parameters define the distribution of bacterial species in sediments across constructed wetland

The PCA biplot with an ~60% variability in bacterial abundance was observed from the first two components (PC1, 33.5% and PC2, 27.9 %; Fig. 2A). The study also found that samples from wastewater inlet site (SC and site A) were observed to be strongly influenced by bacterial species viz. Bacillus sp., Priestia flexa, Ralstonia sp., Citrobacter freundii and Burkholderia cepacia belonging to Bacillaceae, Enterobacteriaceae and Burkholderiaceae taxon. Moreover, in sampling site B, C and OC Bacillus sp., was found to be more prominent along with Enterobacter cloacae and Aeromonas veronii. The Scatterplot matrix revealed that there was a positive correlation between site A and B (r = 0.51) with respect to bacterial abundance. We noted that bacterial abundance and diversity pattern of inlet site (SC) is positively correlated with site A but inversely related with B, C and outlet sampling site (OC), which indicates that the growth of certain groups of bacteria are gradually reduced spatially from the SC to OC sampling sites. The study also investigated possible correlations among sampling sites based on water and sediment quality parameters. The PCA plot revealed that the water and sediment quality of inlet (SC) and outlet (OC) were quite different and largely distanced from each other (Fig. 2B), whereas sampling sites A, B and C form overlapping clusters indicating some similarity in nature. The associated ANOVA results, however, show that physicochemical parameters are significantly different between the sampling sites, except temperature, total alkalinity, Mg, chlorophyll and NO3- -N.

Fig. 2: Principal component analysis (PCA) for bacterial abundance.
figure 2

A PCA biplot and scatterplot matrix of the bacterial communities found at different sampling sites in Sardar Bherry. B PCA plot showing the differences among sampling sites based on water and sediment quality parameters. One-way ANOVA analysis between the sampling sites. [NS denotes non-significant, *<0.05 and **<0.01].

Full size image

Canonical Correspondence Analysis (CCA) is generally preferred to relate the abundance of bacteria to environmental variables, with a CCA plot prepared comparing bacteria isolates with water and sediment quality present in the sampling wetland (Fig. 3A, B). The results revealed that most bacteria found in inlet SC site favours water quality high in TDS, conductivity, NH4, Ca, phosphorus, sulfate and BOD, but low in pH and chlorophyll. Bacterial isolates mainly found in Site A appear to prefer high DO and low nitrogen levels in water, while bacteria in site B showed better growth in high concentrations of nitrogen with low DO levels. In comparison, OC site bacteria appear not to be strongly influenced by any water quality parameters. In the case of sediment parameters, both SC and OC site bacterial isolates were observed not to be influenced by the sediment quality parameters. In contrast, bacterial isolates from site B and C, appear to have preferences for high of pH in sediment, with site A bacteria favoring sediment rich in N and P.

Fig. 3: Canonical correlation analysis (CCA).
figure 3

A Bacterial isolates, water quality parameters and sampling sites; and B bacterial isolates, sediment quality parameters and sampling sites.

Full size image

Constructed wetland treatment system modulates the phenotype of microbiome

The bacterial strains isolated from the sediment samples were mostly identified as Gram-positive bacteria. Additionally, the strains were observed to contain unique irregular, filamentous and rhizoid morphology on TSA plates. The biochemical test showed that bacterial isolates from sampling site A-C and OC were mostly positive for ONPG, lysine, ornithine, urease, citrate, Malonate utilisation, Esculin hydrolysis, trehalose, glucose and oxidase activity, whereas the bacterial isolates from SC sampling sites were mostly positive for urease, citrate, esculin hydrolysis, trehalose and glucose activity. Moreover, the 32 isolated bacterial strains displayed negative results for Phenylalanine Deamination, H2S, Voges Proskauer’s, Methyl red, indole, Adonitol, Rhamnose, Melibiose, Raffinose and lactose activity (Supplementary Table 2). These results show high similarities with previously reported biochemical characteristics of Bacillus spp., Ralstonia spp., Citrobacter freundii, Aeromonas veronii, Enterobacter cloacae, Burkholderia cepacia and Priestia flexa.

Our findings of the antibiotic susceptibility of bacterial strains isolated from SC to OC site showed that 83.33% of the bacterial isolates from Inlet of wetland (SC) had MAR index of ≥0.2 followed by site A (55.56%), site B (50%), site C (42.86%), with none in the outflow (OC) site ≥0.2 (Supplementary Table 3 and Supplementary Fig. 4). The isolates from OC site of the Sardar Bherry have the lowest range of MAR index and ranged from 0.11 to 0.19 (Table 2). The MAR index results suggest that multidrug resistant bacteria were able to propagate far from the wastewater inlet, with some OC site isolates showing borderline multidrug resistance.

Table 2 Isolation and phylogenetic comparison of bacterial isolates.
Full size table

Spatial succession regulates the pathogenicity of bacterial isolates

The study discovered that bacterial isolates including Bacillus cereus (2 strains from A and 4 from SC site), Ralstonia sp. (1 strain from site A) and Burkholderia cepacia (1 strain from site SC) induced significantly high mortality in fish fingerlings (Labeo rohita) (Fig. 4 and Table 3). The bacterial strains isolated from SC sampling site have the highest number of pathogenic isolates followed by sampling site A. In contrast, no mortality was observed across 24–120 h time point in L. rohita treatment group challenged with bacterial strains isolated from sampling sites B, C and OC. In parallel with the survival assay, results showed that isolates from SC and sampling site A exhibited significantly higher haemolytic activity (Fig. 5A), whereas the sampling site B, C and OC isolates had no effect on in blood agar (Fig. 5B).

Fig. 4: Survival assay of Labeo rohita challenged with bacterial strains isolated from sediment samples of wetland.
figure 4

The experimental fish were injected with 0.2 ml @ 106 CFU/ml of bacterial suspension. The control group of fish injected with 0.2 ml of sterile saline solution displayed no clinical signs or mortality throughout the experiment. The strains inducing significantly high mortality post injection exhibited clinical signs like haemorrhagic patches on the body surfaces.

Full size image
Table 3 Survival assay of Labeo rohita challenged with bacterial strains isolated from sediment samples of wetland.
Full size table
Fig. 5: Haemolysin assay of bacterial species isolated from sediment samples.
figure 5

Among the 32 isolated bacterial species, 8 isolates exhibited haemolytic activity on Tryptone soya broth (TSB) supplemented with 5% defibrinated sheep blood. A Ratio of clear zone and colony diameter was calculated from bacterial strains displayed haemolytic activity (results are expressed as mean ± SE). B 8 bacterial isolates (A3, A6, A7, SC1-3, SC5, and SC6) exhibiting haemolytic activity; Non-pathogenic strains (NPS): example of remaining strains which had no activity on TSB plates supplemented with sheep blood.

Full size image

Non-pathogenic bacterial isolates exhibit antimicrobial activity

We then sought to determine the antimicrobial activity of the recovered non-pathogenic isolates including 20 Bacillus species, Citrobacter freundii A9, Aeromonas veronii B4, Enterobacter cloacea B6 and Pristia flexa SC4 against virulent bacteria. Non-pathogenic Bacillus species isolated from sampling site A-C and OC mostly exhibited antimicrobial activity against virulent Ralstonia sp. A3, Bacillus cereus A6, B. cereus A7, B. cereus SC1, B. cereus SC2, Burkholderia cepacian SC3, B. cereus SC5, B. cereus SC6 and against two model fish pathogens, viz., Vibrio parhaemolyticus and Pesudomonas aeruginosa. Moreover, Bacillus subtilis C4, B. safensis OC1 and B. subtilis OC2 displayed highest inhibition diameter zones against pathogenic tested bacteria. We also discovered that non-Bacillaceae group isolates recovered in this study including A. veronii B4, E. cloacea B6 and P. flexa SC4 also exhibited antimicrobial activity (Fig. 6A, Supplementary Table 4 and Supplementary Fig. 5), with C. freundii A9 displaying a comparatively better inhibitory effect, similar to our observation of Bacillus species against pathogenic strains.

Fig. 6: Characterisation of antibacterial assay of non-pathogenic bacterial isolates.
figure 6

A Antibacterial assay of non-pathogenic bacteria isolates and extracellular proteins (ECPs) against pathogenic bacteria. The non-pathogenic bacteria suspension/ ECP were cultured on TSA plates. After 24–48 h incubation at 37 °C, the diameters of inhibition halos surrounding the non-pathogenic bacteria suspension/ ECP were measured and expressed in millimetres. B SDS-PAGE analysis of extracellular proteins (ECPs) concentrated from bacterial cell free supernatant filtered with a 3 kDa cut-off size amicon filter. Molecular mass standards (M) in kilodaltons (Protein ladder) are shown on the left. Lane 1 - Citrobacter freundii A9 ECP3, Lane 2 - Bacillus subtilis C4 ECP3, Lane 3 - B. subtilis OC2 ECP3 and Lane 4 - B. safensis OC1 ECP3.

Full size image

Bacterial ECPs were tested against the growth of pathogenic strains to determine their potential antibacterial activity. ECPs produced from four non-pathogenic bacterial isolates that showed inhibitory activity against 8 or more pathogenic strains were selected for antibacterial assay. The results demonstrated that among the four tested isolates, ECPs extracted from Citrobacter freundii A9 showed the best inhibitory effect, followed by Bacillus subtilis OC2, B. subtilis C4 and B. safensis OC1. The ECPs obtained from the isolates also exhibited strong inhibitory effect against multi-drug resistant fish pathogens Pseudomonas aeruginosa and Vibrio parahaemolyticus (Supplementary Table 5). SDS-Page analysis showed that C. freundii cells secrete proteins in the 20–100 kDa range, while B. subtilis OC2, B. subtilis C4 and B. safensis OC1 cells secrete prominent protein from ~50–100 kDa, along with other proteins that are differentially secreted by both cell types (Fig. 6B).

Capability of bacterial isolates to remediate high ammonia levels

Among 24 non-pathogenic isolates, 1 strain from SC (Priestia flexa SC4), 4 strains from site A (Bacillus safensis A4, B. safensis A5, B. aerius A8, Citrobacter freundii A9), 3 strains from site B (B. safensis B3, Aeromonas veronii B4, Enterobacter cloacae B6), 4 strains from site C (B. aerophilus C3, B. subtilis C4, B. licheniformis C6, B. pumilus C7) and 4 strains from OC site (B. safensis OC1, B. subtilis OC2, B. pumilus OC3, B. pumilus OC4) exhibited ammonia-oxidising activity (Fig. 7). In most cases, the NH4+ -N concentration was reduced by >90%, from 26 mg/L to ~1 mg/L, within 24 h of growth. The Nitrification products, NO2− -N and NO3− -N were detected during the removal process within the 72 h of incubation. In contrast, 8 bacterial isolates including B. pumilus A1, B. pumilus A2 (site A), B. safensis B1, B. pumilus B2, B. altitudinis B5 (site B), B. licheniformis C1, B. altitudinis C2 and B. licheniformis C5 (site C) displayed ammonia producing activity. The ammonia concentration increased ~2 folds in the bacteria supplemented group as compared to the control.

Fig. 7: Ammonia removal assay of bacterial isolates recovered from Sardar bherry wetland.
figure 7

Cultivation of non-pathogenic isolates at C/N mass ratio 10 and initial concentration of NH4+ -N of 26 ppm. The suspension was incubated at 37 °C under constant agitation (120 rpm) and samples were collected at 24, 48, and 72 h post bacterial addition for ammonia, nitrate, and nitrite estimation. The production of each intermediate nitrification product, NO2− -N and NO3− -N were shown for each non-pathogenic strain isolated from 5 sampling sites A (A), B (B), C (C), SC (D), and OC (E). All the assays were performed with freshly prepared media in three replicates.

Full size image

Discussion

Constructed wetlands offer an economic, self-maintained and cost-effective alternative for the conventional treatment of different types of wastewaters. Therefore, an in-depth analysis of the community structure and diversity of microorganisms within CWs, especially for beneficial candidates, is important to understand its performance patterns and explore optimised strategies. The Sardar Bherry, a constructed wetland that receives large volumes of wastewater, presented a model ecosystem to analyse bacterial biodiversity, wastewater nutrient supply, and potential ecological interactions. Based on a straightforward deep-cultivation procedure coupled with unique morphology selection methods and 16S rRNA gene amplification and phylogenetic tree analysis, we obtained 32 pure cultures of bacterial isolates from sediment samples of wetland, viz., sewage canal (SC) - 7 isolates, site A - 9 isolates, site B - 6 isolates, site C - 6 isolates and outlet canal (OC)- 4 isolates. Our results from the recovered isolates indicate that wastewater nutrient supply, physicochemical parameters and ecological interactions are potential drivers of greater bacterial community structure in the sediments of the Sardar Bherry wetland ecosystem.

Nutrient-rich environments favours bacterial growth that can potentially cause an intensive modification of the environment through the release of harmful metabolites that inhibit the growth of competitors in the community18. The “selfishness” of the winning species impedes species co-existence and thus reduces the levels of biodiversity. Active growth may also lead to competition for essential nutrients, counterbalancing the biotic harshness produced by strong competitors19,20. Additionally, the water and sediment quality parameters have a strong influence on the microbial community structures and their efficiency to remove contaminants21; hence, assessing the relationship of microbial ecology and environmental parameters would possibly lead to identify biomarkers responsible for succession of bacterial communities in CWs. Here we observed that water and sediments in a nutrient rich environment (sewage canal) due to wastewater with high TDS, conductivity, NH4, Ca, phosphorus, sulfate and BOD, but low in pH and chlorophyll, contains lower diversity with higher abundances of Bacillus cereus, Priestia flexa and Burkholderia cepacia.

The effect of nutrient selection appeared to be reduced with the effect dilution in the constructed wetland. The higher values of organic carbon, pH, DO and chlorophyll support the growth of wider bacterial populations likely leading to the increase in community diversity observed in site A, B, C and outlet canal of the wetland22. Similar observations were found by Dai et al.19 who proposed a hunger game hypothesis and reported that along with nutrient availability, biotic and abiotic factors affect the natural microbial communities and their co-existence in aquatic environment. Lew et al.23 and Zheng et al.24 demonstrated that physicochemical parameters of wetlands, in general, chemical oxygen demand, organic carbon, nitrogen and phosphorus levels have distinct effects on bacterial community structure and interactions. As such, the nutrient supply from wastewater likely pushes the wheel of community succession through the course of the CW in this study. The effect of nutrient availability and physicochemical parameters on community stability is uncertain, however, as lower biodiversity often reduces stability, but the higher negative network association by nutrient amendment observed in microbial communities might instead enhance stability25,26.

Biochemical reactions can reveal the vital information necessary for accurately identifying the various bacteria genera within a sample. By their nature, bacteria produce large volumes of enzymes, and it is these enzymes that allow for their identification via biochemical methods27,28. Hence, the type of enzymes produced by a bacterium can usually be used to classify its species given that bacteria have distinct enzymatic profiles. The recovered isolates from sites A-C and OC were mostly positive for ONPG, lysine, ornithine, urease, citrate, Malonate utilisation, Esculin hydrolysis, trehalose, glucose and oxidase activity; whereas, the bacterial isolates from SC sampling sites were mostly positive for urease, citrate, esculin hydrolysis, trehalose and glucose activity. Zhao et al. highlighted the importance of biochemical characterisation in bacterial identification. In the study, a total of 36 bacterial isolates recovered from sewerage samples collected at Nanchang City, Jiangxi Province, China, were identified through specific enzymatic activity29. Syed et al.30 demonstrated that a newly isolated strain can be identified as Enterobacter sp. based on variable biochemical reactions. Our results suggest that the bacteria present in Sardar Bherry are efficient at utilising various substrates from wastewater for energy and growth, however, the substrate of utilisation varies from inlet to outlet canal of CW and is reflected in the differentiation of species recovered from successive locations.

Aquatic environments are key niches for the emergence, evolution and dissemination of antimicrobial resistance31. There is evidence that suggests that urban wastewater can contain various chemicals with antimicrobial properties, and these incidental antimicrobial substances might alter the phenotype of the microbes, potentially causing them to develop resistance to chemotherapy32,33,34. Growing resistance to antibiotics in bacteria has been documented for several decades35,36, with wastewater sites recently being considered as one of the main hotspots for the spread of antibiotic resistance. The main reasons suggested for this occurrence are that wastewaters have been observed to contain high prevalence of subclinical levels of antibiotics, heavy metal ions, and other bactericidal factors present in low concentrations that potentially increase the selection of resistant strains in this environment37,38.

The exposure of bacteria to environmental chemical and/or pollutants may promote diversity of resistance genes in various forms (resistant bacteria able to conjugation, free plasmids/DNA, and phage particles), enabling a high probability of gene transfer within the bacterial community39,40. The antibiotic resistance assay has shown the role of wastewater systems in the proliferation of antibiotic resistance in the constructed wetland that receives urban wastewater. Our study found high levels of environmental antibiotic resistance indicators within the bacteria that thrive in the wastewater system, with multi-resistance patterns to antibiotics being common among the isolates. The significantly high MAR indices indicate that the sediments of Sardar Bherry, especially at SC site, potentially harbours multidrug resistant bacteria. There are several reports which indicate that wastewater effluents has significant effect on development of antimicrobial-resistant bacteria41,42,43. For instance, Gessew et al.44 reported that river site receiving wastewaters have high abundance of multiple drug resistant bacteria (28%), mainly Providencia alcalifaciens, Kluyvera cryocrescens and Citrobacter freundii. The possibility of antibiotic resistance spread from wastewater to CWs is also a major threat that impacts the food chain, with potential downstream ramifications during human consumption45. These observations suggest that CW systems are not able to effectively filter out multiple drug resistant bacteria, and the treatment of wastewater for chemical pollutants prior to discharge into CWs may be required to prevent the propagation of resistant strains carrying into the natural waterways.

The abundance profiles are subject to various interspecies interactions that affect individual organism growth and loss rates46,47. Bacterial communities constitute “social networks” in which the members interact with each other in various ways, including competition for nutrients, cooperation by cross-feeding, communication via secretion, and detection of extracellular substances48. In addition, organisms may also indirectly affect other community members by modifying their environment, termed “niche construction theory”49. For example, excretion of secondary metabolites by actively growing species could change environmental conditions, influencing the growth of other organisms and shifting relative abundance levels50,51. Theoretically, biotic factors (e.g., bacterial strain and ecological interactions) and abiotic factors (e.g., nutrient supplies) intertwine to affect individual species abundance, which determines community composition.

To better understand how ecological factors affect interspecies interaction, we investigated how proximity to wastewater effluence correlated with bacterial pathogenicity in fish species. Haemolysin activity is a target virulence factor that is used to determine the pathogenicity of bacterial specimens and was the criteria used to classify potentially pathogenic strains52,53. We observed that CW areas receiving nutrient rich wastewater, SC and Site A, both contained a higher abundance of pathogenic microbes, resulting in 100% mortality in freshwater model fish species (Labeo rohita), as well as higher haemolytic activity. Urban wastewater in other regions has also been reported to carry potentially pathogenic microbes, which would be detrimental to aquatic animal health and the ecology of the environment if the wastewater effluence were directed into natural waterways without treatment54,55. Moreover, the two sites closest to the wastewater effluence (SC and Site A) had a lower abundance of recovered bacterial isolates that were non-pathogenic with antimicrobial activity against tested pathogenic bacterial strains. This suggests that the ecological conditions at the sites that receive the most wastewater tend to favour the proliferation of pathogenic bacteria, as well as suppress the growth of beneficial/remedial bacteria.

The spatial succession of Sardar Bherry demonstrated the potential remedial properties of a CW ecosystem toward the influx of pathogenic strains. There appeared to be a higher abundance of non-pathogenic bacteria, with comparatively better inhibitory effect against pathogenic microbes in areas further (Sites B, C and OC) from the location of the wastewater effluence. The wetland sediments with low effluent load contained more bacterial isolates (e.g., Bacillus sp.) showing inhibitory activity against pathogenic microbes (Corynebacterium fimi and Listeria sp.)56,57. Hence, the disparate interaction of bacterial strains in either nutrient rich or scarce sediment samples give rise to predominant bacterial communities, which are likely due to secretion of secondary metabolites, forming an equilibrium of stable co-existence and niche partitioning of dominant species arising from competitive exclusion. By conducting this spatial succession study, the study highlights those biotic factors and nutrient availability associated with proximity to wastewater effluence is what likely shapes the bacterial species composition and abundance, whilst demonstrating the filtering process performed by the endemic microbial communities on effluent-introduced pathogenic strains across the CW ecosystem.

Recovery of non-pathogenic bacterial isolates with inhibitory effects on the growth of pathogen strains from Sardar Bherry demonstrates the potential of CW ecosystems as a target for the discovery of new anti-bacterial agents. The antibacterial properties of the non-pathogenic isolates and their ECPs provide a potential source of probiotics or bacteriocins that can be investigated for use in aquaculture (Pereira et al.69). The ECPs tested in this study were able to inhibit the growth of multidrug resistant pathogenic bacteria, making them prime research targets as potential bacteriocins for further development as anti-bacterial agents (Benítez-Chao et al.70; Simons et al.71). Ecosystem structure and microbiome diversity have been reported to be key drivers in the production of ECPs (Garcia-Garcera & Rocha72). The observation of inhibitory ECPs at Sardar Bherry suggests that CWs or other similar sites that receive regular influxes of multidrug resistant bacteria from wastewater (Gessew et al.44) could foster the potential development of ECPs effective against resistant strains due to inter-species interactions with the endemic microbial population. Further investigation is needed to ascertain the therapeutical potential of CW microorganisms for applications in healthcare and aquaculture.

The discovery of isolates that can reduce ammonia levels substantially in solution and the consistent pattern of ammonia oxidising bacteria across the Sardar Bherry CW provides an indicator of the ecosystem’s capability to effectively remediate ammonia. Wastewater has been shown to accelerate the build-up of ammonia and other nitrogenous compounds in the aquatic ecosystem, which causes serious ecological problems such as toxicity in fish, as well as in other aquatic animals58,59. Ammonia in particular can pose issues for the health of fish, increasing physiological stress, resulting in tissue damage and even death (Soler et al.73). Along with the pathogenicity study, this indicates that a better understanding of non-pathogenic CW microorganisms is important for elucidating the key taxa involved in preserving the health of aquatic lifeforms.

The recovery of non-pathogenic bacterial isolates with high ammonia removal capabilities provides a potential source of bioremediating candidates for use in water treatment studies. Ammonia is a common pollutant in wastewater and biological approaches using living organisms have been explored as an inexpensive and efficient form of water treatment (Karri et al.74; Royan et al.75). Aside from aquatic health, the ammonia removal capabilities of CW isolates could be utilised in green biotechnology solutions, such as combining the CW with a microbial fuel cell system for the generation of bioelectricity (Liu et al.76). The Sardar Bherry demonstrates that CW microbiomes are not just efficient treatment systems, but also present a potential bioprospecting trove for microorganisms for use in sustainable solutions.

This study provided insights into the ecological roles of bacteria within the Sardar Bherry CW sediment microbiome via the diversity-driven cultivation and functional characterisation of 32 bacterial strains. By summarising the main functional microorganisms in the CW, we found that the family Bacillaceae is the dominant one, containing microorganisms with a wide range of functions. In addition, the family Enterobacteriaceae and Aeromonadaceae are also frequently detected in CW. Regarding the effects of different pollutants on microbial diversity, we found that different microorganisms respond in different ways. The CW inlets containing high nitrogen and phosphorus levels, microorganisms with high pathogenic potential become dominant in the system. However, once the wastewater enters the CW, there was high abundance of genera of functional microorganisms, which can remove pollutants from CW and have health benefits to aquatic life. While nutrient availability appears to have fundamental coupling with bacterial community diversity, interaction between bacterial isolates potentially contributes, in part, to the assembly of distinct microbial communities in the constructed wetlands. We have shown that the culturing of highly diverse bacteria is indeed possible when their ecological niche is sufficiently well mimicked, and time is allowed for slow-growing strains to propagate.

The application of similar cultivation strategies to other understudied bacterial species will increase the number and diversity of axenic cultures from constructed wetland systems in the future. This discovery-driven culture and identification based on 16S rRNA amplicon sequencing and biochemical method provide valuable preliminary information which can be used to guide further study. Questions remain, however, on whether the available axenic cultures of bacterial isolates reflect bacterial diversity sufficiently, as ~99% of environmental bacteria are uncultivable with countless more hitherto undiscovered microorganisms remaining in the microbial ‘dark matter’. Hence, metagenomics or single-cell techniques are needed to gain additional insights required for expanding the understanding of the bacterial communities present within the constructed wetland treatment system. Nevertheless, the study highlighted that the CW microbiome plays a critical role by providing the microbial-based bioremediation processes for the removal of pollutants and pathogenic bacteria. Further work in this area can aid researchers in finding links between remedial microorganisms and the physicochemical parameters of CWs, and can help develop optimisation strategies by facilitating the discovery of suitable and beneficial microbial candidates.

Methods

Study area and sample collection

The East Kolkata floodplain wetland is situated in the lower Bhagirathi basin bounded by latitude 22˚25′N to 22˚40′N and longitudes 88˚20′E to 88˚35′E, and lies between two rivers, Hooghly River in the northwest and Bidyadhari in the east. These wetlands act as a natural wastewater treatment plant and reuse/treat about 800 million liters of wastewater flowing out daily from Kolkata (Fig. 8). Among the East Kolkata floodplain wetlands, Sardar Bherry, consisting of an area of roughly 140 hectares, receives significant amounts of wastewater and is considered one of the polluted wetlands in the region (based on our previous water quality analysis). Due to these properties, Sardar Bherry was selected as the model CW ecosystem for study and sample collections.

Fig. 8: Pictorial representation of wastewater treatment by constructed wetlands in East Kolkata, West Bengal, India.
figure 8

The raw sewage water from different sewage canals (stage 1) enters wetlands (stage 2) utilised for aquaculture activities and then goes into natural streams/rivers (stage 3). The wetlands in stage 2 are purposed as natural treatment systems of urban wastewater. In the study, Sardar Bherry was selected as the model system to investigate the role of microbiome-based bio-treatment and reuse systems of urban wastewater in constructed wetland treatment systems.

Full size image

To analyse the impact of anthropogenic pressure generated from urban wastewater dilution on the bacterial community, the sediment samples from 17 sampling sites were collected from sewage canals (SC), wetland sites A, B, C and outlet canals (OC) located in Sardar Bherry (Supplementary methods). Sites SC1-SC3, were located in sewage canal inlet sites (numbered from north to south), receiving wastewater from various drainages of the city. Site A1-A4, B1-B4 and C1-C4 were in the constructed wetlands (numbered from north to south). Site C1-C2 is located in the outlet canal site, discharging water to natural Kestopur canal/river. The direction of the water flow is from sites SC to OC (Fig. 9).

Fig. 9: GIS based map of study area and sampling locations.
figure 9

The selected location is Sardar Bherry, a constructed wetland receiving urban wastewater located in the Eastern part of Kolkata city, India, spanning a ~ 140 ha area. Sampling was done from sewage canals (SC1-SC3), wetland sites A (A1–A4), B (B1–B4), C (C1–C4) and outlet canals (OC1-OC2) located within the wetland.

Full size image

Measurement of environmental parameters

The water and sediment samples were quantified for physiochemical parameters. To prepare for the sediment physiochemical quality assessment, the sediment samples were first dried in the shade at room temperature and then ground with a wooden hammer, followed by sieving through a No. 10 mesh sieve (2 mm) and stored in plastic packets. Details of the physicochemical testing procedures for the water and sediment samples can be found in the supplementary methods.

Isolation of bacterial isolates from sediment samples

The sediment samples were used to determine the abundance of total cultivable bacteria following the protocol developed by Guan et al.77 with slight modifications (Supplementary methods). Following the testing for cultivable bacteria, single colonies from Tryptone soya agar (TSA) plates used in the previous step were selected based on uniqueness in morphology, size, and colour, and cultured overnight in Tryptone soya broth (TSB) at 37 °C with shaking at 120 rpm. The TSB culture was streaked into TSA media plates to check the purity of the isolates (based on the growth of colonies). Furthermore, a single colony was inoculated into TSB media, incubated overnight at 37 °C under constant agitation (120 rpm), with stock culture prepared containing 40% glycerol for storage at −80 °C until further use. A total of 32 stock cultures of bacterial isolates were produced from this process.

Identification and phylogenetic comparison of bacterial isolates

The bacterial isolates were identified based on 16S rRNA PCR amplicon sequencing methods. The methods for preparing the isolate cultures for PCR amplification, such as DNA extraction, PCR cycling protocols, as well as the primers used, can be found in the Supplementary methods and Supplementary Table 6, respectively. The amplified gene products were sequenced in forward and reverse direction using the ABI 373xl capillary sequencer (Applied Biosystem, Foster City, CA). Assembly of the 16 S rRNA sequences was performed by aligning forward and reverse sequences using DNA baser 7.0.0. The assembled bacterial 16 S rRNA sequences were aligned and compared with available sequences in the NCBI GenBank using BLAST (http://blast.ncbi.nlm.nih.gov). Afterward, the sequences were submitted to GenBank for the preparation of a phylogenetic tree.

Evolutionary analyses were conducted in MEGA1160 using the Neighbor-Joining method, to create the evolutionary history of recovered bacterial strains with the optimal tree61. The tree is drawn to scale, with branch lengths in the same units as the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Maximum Composite Likelihood method62,63,64,65 and the units represent the number of base substitutions per site, with the analysis performed on 32 nucleotide sequences. Codon positions included were 1st + 2nd + 3rd + non-coding, with ambiguous positions removed for each sequence pair using the pairwise deletion option, resulting in a total of 1444 positions in the final dataset.

Biochemical characterisation and antibiogram assay of bacterial isolates

The 32 bacterial isolates were distinguished using the Gram-staining method with biochemical characterisation performed following standard procedures, viz.; ONPG (β-galactosidase), lysine utilisation, ornithine utilisation, urease, phenylalanine utilisation, nitrate reduction, H2S production, citrate utilisation, Voges Proskauer’s (VP), methyl red, indole, malonate utilisation, esculin hydrolysis, arabinose, xylose, adonitol, rhamnose, cellobiose, melibiose, saccharose, raffinose, trehalose, glucose, lactose, and oxidase (KB003, Hi-media)63,64. Antibiotic resistance testing of the 32 bacterial isolates was done according to the methods given by Batoni et al. and Zidour et al., with details in the supplementary methods. Bacterial sensitivity towards antibiotics was interpreted as sensitive, intermediate or resistant according to the guidelines of the Clinical and Laboratory Standards Institute33,34, with the antibiotic discs used in the antibiogram assay listed in Supplementary Table 7. The antibiotic resistance assay was performed in quintuplicate and is representative of two independent experiments. Multiple antibiotic resistance (MAR) index of each isolate were estimated according to the method developed by Krumperman65.

Acclimatisation of experimental fish and challenge assay

Organization for Economic Cooperation and Development (OECD) guidelines were followed for the handling and care of experimental animals. The animal utilisation protocol was approved by Institutional Animal Ethics Committee, ICAR-Central Inland Fisheries Research Institute, Kolkata, India, (IAEC/2021/04) for the experimental setup.

Healthy Labeo rohita (Length = 132.8 ± 6.1 mm, Weight = 28.87 ± 2.4 g) were procured from a local fish hatchery. The fish were inspected to ensure that there were no external clinical symptoms like haemorrhage, ulcer, discoloration, descaling, or redness in the body of the fish. Additionally, before the challenge assay, fish were randomly selected and screened for the presence of infectious microbes following a standard protocol (Nickum et al.78; Johansen et al.79). The fish were acclimatised in 200L FRP tanks for 2 weeks, supplied with proper aeration, and fed a diet of commercial fish pellets (containing 35% crude protein and 10% crude fat) weighing ~2% of total body weight twice a day. The bacterial challenge assay was prepared from the 32 bacterial isolates recovered from the sediment samples. The 32 bacterial isolates were sub-cultured in 20 ml of sterile TSB in 50 ml Erlenmeyer flasks (Himedia, India) at 37 °C for 24 h. Bacterial cells were collected by centrifuging at 5000 rpm for 5 min and washed thrice with a sterile saline solution. Afterward, the pellets were resuspended in normal saline and the number of cells/ml was estimated through the spread plate method. Intraperitoneal injection of the experimental fish (20 numbers) was done with 0.2 ml @ 106 CFU/ml of bacterial suspension. The control fish were injected with 0.2 ml of saline solution. Afterward, the fish were kept in an FRP tank and observed every 24 h for a total of 120 h. To confirm Koch’s postulate, the bacteria were reisolated and identified from the liver, kidney and blood of the moribund fish66.

Haemolytic activity

The haemolytic activities of the 32 bacterial isolates were conducted according to a standard protocol with slight modifications67. Briefly, the TSA plates were prepared by supplementing 5% defibrinated sheep blood. The pure stock cultures of the 32 bacterial isolates were grown overnight in TSB at 37 °C under constant agitation. Later, the overnight culture was diluted to an OD600 of 0.5, and 2 µl of the diluted culture was spotted in the middle of the haemolysin test plates. The plates were incubated at 37 °C for 48 h, after which the diameters of the clearing zones were measured. The assays were done using freshly prepared media in four replicates.

Antibacterial assay of non-pathogenic bacterial isolates and extracellular proteins

In total, two separate tests were performed to determine the potential antibacterial role of non-pathogenic bacterial isolates (classified based on fish survival and haemolysin assay results) against isolated pathogenic strains (Ralstonia sp. A3, Bacillus cereus A6, B. cereus A7, B. cereus SC1, B. cereus SC2, Burkholderia cepacian SC3, B. cereus SC5 and B. cereus SC6) and two model fish pathogens (Vibrio parhaemolyticus and Pseudomonas aeruginosa). The pathogenic bacteria suspensions (100 μl @107 cells/ml) were spread onto TSA plates using a sterile spreader, and five different non-pathogenic bacteria suspensions (10 µl @105 cells/ml) were placed on agar plates with a micropipette. Later, the agar plates were covered with parafilm tape and were incubated for 24–48 h at 37 °C. Diameters of inhibition halos surrounding the non-pathogenic bacteria colony were measured and expressed in millimetres. Subsequently, the non-pathogenic bacterial isolates showing antibacterial activity were grown separately from pathogenic strains to reconfirm the activity. Testing and quantification of bacterial extracellular proteins (ECPs) were performed on non-pathogenic bacterial isolates that displayed inhibitory effects against 8 or more pathogenic strains to determine their antibacterial activity (Supplementary methods).

Ammonia removal assay

Synthetic wastewater was prepared for use as the basic medium for storage, cultivation and for subsequent studies on nitrogen removal (at C/N mass ratio 10)68. The synthetic wastewater mixture was prepared using (g/L): Na2HPO4·12H2O 21.5, KH2PO4 0.91, NaCl 4, glucose (anhydrous) 2, NH4Cl 0.08 g and 30 mL of each trace solution (pH maintained between 7–7.5). The trace solution was prepared using the following components (g/L): MgSO4 6.42, MnSO4 0.63, and H3BO3 3.36 g. Each solution was distributed evenly using 250 mL shaking flasks and inoculated for experimental analysis. For the assay, non-pathogenic bacterial isolates (based on fish survival, haemolysin and antibacterial assay results) were used to determine the potential nitrogen remediating properties (supplementary methods).

Statistical analysis

The data were arcsine transformed to satisfy normality and homoscedasticity requirements, then subjected to one-way analysis of variances (ANOVA) followed by Duncan’s multiple range test using statistical software statistical package for the social sciences version 24.0 (P value ≤ 0.001). Statistical analyses were also performed to investigate the correlation between water/sediment quality with the sampling sites and bacterial isolates recovered from the wetland, with details of the multivariate data analyses performed in the supplementary methods.