Introduction

Antimicrobial resistance (AMR) is a significant global health threat, with environmental factors playing a critical role in its development and dissemination. Environmental AMR refers to the occurrence and spread of antibiotic-resistant bacteria and resistance genes in various environmental settings, including soil, water, air, and wildlife1,2,3,4. Another factor is Wastewater and Sewage, where the effluents from wastewater treatment plants, hospitals, and pharmaceutical manufacturing can contain high levels of antibiotics and resistant bacteria, which can enter natural water bodies5. The third one is Aquaculture where the reluctant use of antibiotics in fish farming can lead to the contamination of water bodies with resistant bacteria6 and the fourth one is Wildlife where animals in the wild can act as reservoirs and vectors for resistant bacteria, which can be spread through their movement and interactions with human and domestic animal populations7. This antibiotic resistance is commonly spread through four different ways. The first one is through waterways (rivers, lakes, and oceans can transport resistant bacteria and genes over long distances) followed by soil (antibiotic residues and resistant bacteria can persist in soil, affecting microbial communities and potentially entering the food chain), air (aerosols from manure spreading, wastewater treatment, and other sources can disseminate resistant bacteria and genes through the air) and animals (birds, insects, and other wildlife can act as carriers, spreading resistant bacteria across different environments)3,8,9,10.

Apart from general AMR dissemination in the environment, the presence of antimicrobial resistance (AMR) in gutter samples is an emerging concern, particularly in urban and peri-urban areas where gutters often collect runoff from various sources, including households, healthcare facilities, and streets11. Gutters can serve as reservoirs and confluent for resistant bacteria and resistance genes, which can then spread to broader environments. These contaminated gutter water can enter larger water bodies, spreading resistant bacteria and genes to rivers, lakes, and groundwater. This can disrupt natural microbial ecosystems and promote the proliferation of resistant strains through either horizontal gene transfer (conjugation, transformation, and transduction) or biofilm formation12. Humans can also be exposed to these resistant bacteria through direct contact with contaminated water, use of contaminated water for irrigation, and consumption of contaminated food13. These increases the risk of infecting antibiotic-resistant bacteria which leads to difficult-to-treat infections, longer hospital stays, higher medical costs, and finally increased mortality. In this context, it is very much necessary to assess the bacterial population present in the targeted gutter ecosystem and diagnose loads of antibiotic-resistant bacteria in the gutter samples which are one of the water sources of nearby riverine systems14. So, AMR in gutter samples represents a critical aspect of the broader issue of environmental antimicrobial resistance. Addressing this problem requires a holistic approach involving improved waste management, stringent regulations, public education, and ongoing surveillance to monitor and mitigate the spread of resistance in the environment15.

Although these antibiotic resistance scenarios, there are no such reports that can assess the microbiome community and the level of antibiotic resistance present in the street gutter samples which are the secondary water resource of the nearby rivers except Hervé et al.16 reported the diverse eukaryotic microbial community present on the street gutters of Paris, previously but there was no insight of antimicrobial resistance. To consider the present emergency, we need to explore the level of antibiotic resistance in the gutter samples of a semi-urban town in India like Roorkee. Therefore, the present study aims to evaluate the antibiotic-resistant bacterial population in the gutter samples of Roorkee, Uttarakhand, India, and the presence of beta-lactamase in the bacterial population. This study also aims to explore the total bacterial communities present in different gutter samples with their population variations within the six different ecosystems and predict the metabolic pathways present in those microbiomes. This study will signify the public health, environmental impacts, and scientific knowledge to encounter the emerging emergency of the development of environmental AMR.

Results

Physical characterization of the water samples

The physical properties of gutter water showed that the total electrolytes content was highest in HL_NS-5 (1227 µS/cm) and the lowest in HL_NS-6 (646 µS/cm) whereas in all the six samples, the pH were lied between slight acidic to slight basic (6.27–8.1) conditions. Each sample had variations in the quantities of heavy metals such as lead, cadmium, zinc, iron, and arsenic. The iron content was highest in HL_NS-3 (556.03 ppb) followed by HL_NS-2, HL_NS-4, HL_NS-5 whereas the lowest one in HL_NS-6 (39.107 ppb). The zinc concentration was highest in HL_NS-1 (94.24 ppb) whereas HL_NS-3 contained lowest zinc concentration (8.054 ppb). The arsenic content was highest in HL_NS-6 (5.416 ppb) and lowest was in HL_NS-3 (2.722 ppb). The water from HL_NS-1 possessed maximum amount of cadmium (0.096 ppb) and lead (3.046 ppb) content whereas HL_NS-4 possessed lowest cadmium (0.015 ppb) and HL_NS-3 possessed lowest lead content (0.143 ppb) among all the six sample sites (Table 1).

Table 1 Total electrolytes, pH, and heavy metal concentration of six gutter samples.
Full size table

All the experiments were performed in triplicate manner. Data were presented as mean ± SD.

Quantification of bacterial load and antibiotic resistance profiling

The bacterial load from six samples (HL_NS-1 to HL_NS-6) were quantified in terms of colony-forming units per milliliter (CFU/ml). The CFU counts range from 26,000/ml in sample HL_NS-5 to 252,000/ml in sample HL_NS-4, indicating variation in bacterial concentration across the samples (Fig. 1a). Corresponding logarithmic values (log CFU/ml) are provided to normalize the data and facilitate comparison, with values ranging from approximately 4.41 to 5.40. The highest bacterial load was observed in sample HL_NS-4 (252,000 CFU/ml; log CFU/ml 5.40), followed by HL_NS-2 (224,000 CFU/ml; log CFU/ml 5.35), HL_NS-3 (182,000 CFU/ml; log CFU/ml 5.26), HL_NS-1(178,000 CFU/ml; log CFU/ml 5.25), HL_NS-6(28,000 CFU/ml; log CFU/ml 4.45), and the lowest was in HL_NS-5 (26,000 CFU/ml; log CFU/ml 4.41) (Fig. 1a).

The six gutter samples exhibited varying levels of antibiotic-resistant bacterial populations. Among them, sample HL_NS-6 harboured the most resistant bacteria, which were resistant to multiple antibiotic classes including Penicillins (ampicillin, oxacillin), Cephalosporins (cephalothin, cefotaxime), Aminoglycosides (kanamycin), Fluoroquinolones (ciprofloxacin), and Antifolates (trimethoprim), but remained sensitive to the Carbapenem antibiotic meropenem. Similarly, HL_NS-2 contained a highly resistant population that was resistant to all tested antibiotics except the third-generation Cephalosporin, cefotaxime. Both HL_NS-6 and HL_NS-2 showed extreme resistance (90–100%) to the Penicillin antibiotics. In contrast, HL_NS-5 showed the highest antibiotic sensitivity, being susceptible to seven of the tested antibiotics, showing resistance only to kanamycin (25%). Overall, the Penicillin class (ampicillin and oxacillin) exhibited the highest levels of resistance across all samples. The lowest resistance was observed in the Cephalosporin group particularly cefotaxime and cephalothin except in HL_NS-6. Resistance to the Carbapenem (meropenem) and Fluoroquinolone (ciprofloxacin) classes was detected in half of the samples. Five samples (HL_NS-1, HL_NS-2, HL_NS-3, HL_NS-5, HL_NS-6) showed resistance to kanamycin, while another five (HL_NS-1, HL_NS-2, HL_NS-3, HL_NS-4, HL_NS-6) exhibited resistance to trimethoprim (Fig. 1b).

Fig. 1
figure 1

Determination of bacterial load (a) and antibiotic resistance profile (b) of six gutter microbiome. Here, the bars represent the bacterial load in log CFU/ml and dotted line represents the bacterial load in CFU/ml units.

Full size image

Determination of beta lactamase mediated antibiotic resistance

The presence of beta-lactamase genes was assessed to detect the development of primary antibiotic resistance mechanisms acquired by the bacterial populations. Wild type E. coli BL21 (DE3) cells without the plasmid vector (pET28a) did not hydrolyzed the nitrocefin (Fig. 2a) whereas the E. coli BL21 (DE3) cells transformed with the Beta lactamase KPC recombinant plasmid vector hydrolyzed the nitrocefin and converted the color yellow to red (Fig. 2b). The gutter microbiome also shown hydrolyzing efficiency of nitrocefin except HL_NS-5. This observation confirms the presence of beta-lactamase in the gutters’ microbiome (Fig. 2c-h).

Fig. 2
figure 2

Presence of beta lactamase in gutter microbiome. Here, a and b represent the negative and positive control whereas c-h represents HL_NS-1 to HL_NS-6 respectively. The black arrows indicate the red color due to the hydrolysis of nitrocefin.

Full size image

General characteristics of the amplicons and sequencing data

The amplicons of the six gutter samples were obtained from Illumina Miseq sequencing analysis. The V3 region of the samples produced 0.2–0.3 million chimera filtered paired-end sequences and the average number of processed sequences were 0.2 M with the sequence length ranged from 150 to 300 bp and the average sequence length was 160 bp in all the six samples. Sequences were clustered into rarefaction curves of operational taxonomic units (OTUs) of all these six samples based on ≥ 97% similarity. The %GC content of HL_NS-1, HL_NS-3 were 54%, HL_NS-2, HL_NS-4 were 54.5% and HL_NS-5, HL_NS-6 were 53%, respectively and the fred values were > 38 with Phred quality score distribution (Q%) were > 90% in all the six samples.

Taxonomic composition analysis

The microbial community of six different gutter samples was characterized and classified from phylum to genus according to the default QIIME v1.9.1 program. More than 0.2 million reads with GC% <50% were considered in each gutter site. There were 18 different phyla, 139 families and 240 genera keeping 0.05 as the relative abundance cut off values. There was a large percentage of relative abundance from unidentifed bacterial populations. The common phyla present in all the gutter samples were proteobacteria followed by firmicutes, bacteroids, actinobacteria, planctomycetes, chloroflexi, cyanobacteria, tenericutes, and deinococcus. Among them, proteobacteria, firmicutes, bacteroids, and actinobacteria possess 97.5% niche distribution. Apart from the common phyla distribution, there were some phyla were present in the specific micro-environment such as Fusobacteria and Thermotogae in HL_NS-1, Acidobacteria in HL_NS-2, Chlamydiae in HL_NS-3, Candidatus saccharibacteria and Lentisphaerae in HL_NS-4 whereas HL_NS-5, and HL_NS-6 did not possess any specific phyla in that micro-environment (Fig. 3a). The major families present in all the samples were Enterobacteriaceae (15.4%) and Pseudomonadaceae (8.29%) and they are covering 23.7% relative abundance of the total population. Apart from these two, Lachnospiraceae (HL_NS-1 and HL_NS-2), Porphyromonadaceae (HL_NS-2, HL_NS-5, and HL_NS-6), Moraxellaceae, Flavobacteriaceae (HL_NS-2, HL_NS-3, HL_NS-4, and HL_NS-6), Pasteurellaceae (HL_NS-2, and HL_NS-4,), Prevotellaceae (HL_NS-2, HL_NS-3, and HL_NS-5), Comamonadaceae (HL_NS-3, HL_NS-4, and HL_NS-6), Rhodospirillaceae (HL_NS-4, and HL_NS-6), Desulfovibrionaceae, Campylobacteraceae (HL_NS-5, and HL_NS-6) were present in more than one sample sites (Fig. 3b). Beside these dominant families, there was some families were exclusively present in the specific sample site. Clostridiaceae, Staphylococcaceae, Corynebacteriaceae, Chitinophagaceae, Erysipelotrichaceae, Coriobacteriaceae were only present in HL_NS-1 whereas Caulobacteraceae, Acidiferrobacteraceae, Atopobiaceae (HL_NS-2), Blattabacteriaceae, Ruminococcaceae, Alteromonadaceae (HL_NS-3), Zoogloeaceae, Sphingobacteriaceae, Prolixibacteraceae (HL_NS-4), Azonexaceae, Bacteroidaceae, Bacillaceae, Rhodocyclaceae (HL_NS-5), and Eubacteriaceae (HL_NS-6) were exclusively present in the specific gutter samples (Fig. 3b). Among the six microenvironments, the highest number of taxa was present in HL_NS-2 and HL_NS-4 (106 families) whereas the lowest number of taxa present in HL_NS-5 (62 families, Table S1). The major genera present in all the samples were Pseudomonas (8.3%) Escherichia (8.2%), Hydrogenophaga (6.85%), and Candidatus Moranella (5.4%) and they are covering 28.74% relative abundance of the total population. Among the six microenvironments, the highest number of taxa was present in HL_NS-2 (151 genera) whereas the lowest number of taxa present in HL_NS-5 (81 genera, Fig. 3c, Fig. S1, Table S2). Although these diverse taxonomic distribution among the six gutters, the one-way ANOVA of both the family (df = 833 p > 0.05) and genus (df = 1439 p > 0.05) levels showed statistically insignificant variations presence of bacterial population. Apart from the total taxonomic distribution in the gutter samples, it is very important to assess the relative abundance of the total fecal coliform bacteria that includes Citrobacter, Enterobacter, Escherichia, and Klebsiella. The relative abundance (%) of total fecal coliform bacteria (Citrobacter, Escherichia, and Klebsiella) were 14.10661, 4.402987, 7.820675, 9.511516, 5.500269, 11.8824 in HL_NS-01 to HL_NS-06, respectively where the highest fecal coliforms resided in HL_NS-01 (14.10661%) and the lowest was in HL_NS-02 (4.402987%) (Table S2). However, there was no representation of Enterobacter in all the six gutter microbiome. 

Fig. 3
figure 3

Taxonomic composition of microbial communities in six gutter samples. Here, ‘a’ represents the circular representation of correlation of phyla presented in six samples; ‘b’ and ‘c’ represent the heatmap of top ten families and genera of six gutter samples, respectively.

Full size image

Ecological diversity of the microbial populations

The ecological diversity from the represented bacterial abundance was assessed by ecological dominance, evenness, alpha and beta diversity indices (Fig. S2, Fig. 4). The dominance and evenness indicate the minimum and maximum variability of the taxa, respectively resulting to assess competition for niche acquisition. The maximum dominance of families (0.1094) and genera (0.0794) were shown in the HL_NS-4 and HL_NS-5, respectively and lowest in HL_NS-2 in both families (0.03162) and genera (0.02313) (Fig. S2a, 4a). Whereas the ecological evenness was inversely proportional to the dominance of the taxon (Fig. S2b, 4b). The Simpson index (D) among the family’s relative abundance was highest in the HL_NS-2 (0.9684) followed by HL_NS-3 (0.9333), HL_NS-6 (0.9117), HL_NS-1 (0.9083), HL_NS-5 (0.8959), HL_NS-4 (0.8906), respectively whereas the highest Simpson index (D) of genus level were found in HL_NS-2 (0.9769) followed by HL_NS-3 (0.9538), HL_NS-6 (0.9485), HL_NS-4 (0.9245), HL_NS-1 (0.9243), HL_NS-5 (0.9206), respectively. The Shannon index (H) of both family and genus level were also showed similar pattern of indices (Fig. S2c, 4c). The beta diversity of family level such as Whittaker and Cody index of all six sample sites were 0.51361 and 98.5, respectively whereas the Whittaker and Cody index at genus level were 0.87744 and 204, respectively. The Principal Coordinate Analysis (PCoA) of both family and genus showed that the bacterial composition of the sample site HL_NS-1 (positive x-axis and negative y-axis) was highly distant from HL_NS-2 (on the x-axis and negative y-axis) whereas HL_NS-3 and HL_NS-4 were close to each other because they were lied in the same coordinate (negative x-axis and negative y-axis) with a short distance matrix in both family and genera point of view similar to the HL_NS-5 and HL_NS-6 with different coordinates (negative x-axis and positive y-axis) and a large distance matrix in both family and genera level analysis (Fig. S2d, 4d).

Fig. 4
figure 4

Ecological diversity of genera of endophytic bacteria. Here, 'a', 'b', 'c', and 'd' represent ecological dominance, ecological even ness, alpha-diversities and Bray–Curtis beta-diversities of genera, respectively. Here, in 'c', the black and hollow bars represent Simpson and Shannon indices, respectively; 'd' represent the beta diversity of six gutter samples. Here all the data were represented as considering the relative abundance ≥ 0.05 OTUs.

Full size image

Microbial community distribution

The Venn diagram analysis identified a distinct set of bacterial communities consistently present across all six sampling sites during the summer season, comprising 36.7% of bacterial families and 21.25% of genera (Fig. 5a, c). This shared microbial assemblage constitutes the core microbiome, which likely plays a pivotal role in maintaining the structural and functional stability of the gutter ecosystem. In contrast, site-specific taxa, referred to as the peripheral microbiome, were uniquely distributed among the individual samples, underscoring the microbial diversity and ecological heterogeneity characteristic of each location. From the family point of view, 9 taxa were only present in HL_NS-1, 2 in HL_NS-2, 6 in HL_NS-3, 5 in HL_NS-4, 1 in HL_NS-5, 3 in HL_NS-6. Others taxa were shared in between these six samples sites in different permutation and combinations (Fig. 5a, Table S3). From the genera point of view, 18 taxa were only present in HL_NS-1, 11 in HL_NS-2, 21 in HL_NS-3, 13 in HL_NS-4, 2 in HL_NS-5, 6 in HL_NS-6. Others taxa were shared in between these six samples sites in different permutation and combinations (Fig. 5b, Table S4). The family and genera with the six sampling sites were positively correlated and the correlation coefficient were lied between 1 and 0.33 (Fig. 5b, d). The data corresponds the high diversity of microbial population present in the gutter samples during summer season. The principal component analysis (PCA) of the six sampling sites were performed to assess the relative divergence of microbial community presented in the semi-urban gutter ecosystem. The analysis showed that HL_NS-1 and HL_NS-5 were distantly placed (positive x-axis and negative y-axis for HL_NS-1 and negative x-axis and positive y-axis for HL_NS-5) from rest of the four, whereas among the rest four, HL_NS-2 and HL_NS-5, and HL_NS-3 and HL_NS-4 were closely related with each other (Fig. 6a). This observation also correlated with the sampling site’s information (Table S5) where HL_NS-1 and HL_NS-5 were collected from cattle farm and a location adjacent to the highway, respectively. In contrast, the remaining samples collected from areas with high levels of human and other anthropogenic activities. The UPGMA analysis of the microbial communities present in the sample sites showed the similar pattern of clustering with 1000 bootstrap values where HL_NS-1 appeared as outgroup membered followed by HL_NS-5. The HL_NS-3 and HL_NS-4 were closely related. Whereas HL_NS-2 and HL_NS-6 were distantly placed in the UPGMA tree (Fig. 6b).

Fig. 5
figure 5

Venn diagram and correlation coefficient of total families and genera present in the six gutter samples. Here, ‘a’, ‘b’ represents the family level and ‘c’, ‘d’ represents the genus level, respectively. Here all the data were represented as considering the relative abundance ≥ 0.05 OTUs.

Full size image
Fig. 6
figure 6

Relationship assessment between six sample sites on the basis of their microbial taxonomic profiles. Here ‘a’ represents Principal Component Analysis of sample sites; and ‘b’ represents UPGMA similarity matrix (Bray-Curtis) with bootstrap values 1000 of bacterial genera presented in the six gutter samples. Here all the data were represented as considering the relative abundance ≥ 0.05 OTUs.

Full size image

Prediction of metabolic pathways

The metabolic pathways of the microbial communities of the six gutter samples were predicted and the major metabolic pathways were common to all the six gutter samples except Biotin metabolism (present in HL_NS-1, HL_NS-2, HL_NS-3, HL_NS-4, HL_NS-6), Cysteine and methionine metabolism (present in HL_NS-1, HL_NS-2, HL_NS-3, HL_NS-6), beta-Alanine metabolism (present in HL_NS-1, HL_NS-4, HL_NS-5, HL_NS-6), Glutathione metabolism (present in HL_NS-1, HL_NS-4, HL_NS-5, HL_NS-6), Streptomycin biosynthesis (present in HL_NS-1, HL_NS-4, HL_NS-5, HL_NS-6), Fatty acid biosynthesis (present in HL_NS-1, HL_NS-3, HL_NS-4, HL_NS-6), Glycine, serine and threonine metabolism (present in HL_NS-1, HL_NS-2, HL_NS-4, HL_NS-5, HL_NS-6), Glycosphingolipid biosynthesis-globo and isoglobo series (present in HL_NS-2, HL_NS-3, HL_NS-4, HL_NS-6), Lipopolysaccharide biosynthesis (present in HL_NS-2, HL_NS-3, HL_NS-4), and Cyanoamino acid metabolism (present in HL_NS-5) (Fig. 7).

Fig. 7
figure 7

Heatmap of predicted metabolic profiles presented in the bacterial genera of six gutter samples.

Full size image

Discussions

The gutter from semi-urban towns are the major source of pollution in water-bodies especially catchments, rivers etc. Besides conventional water pollution, antibiotic pollution leading to antimicrobial resistance is the emerging threat of the ecosystem15. The present study opened the lid of the Pandora’s box indicating the antibiotic resistance in household adjacent gutter samples. The pH of all the gutter waters remains slightly acidic to slightly basic with high amount of electrolytes content and also possessed certain amount of heavy metals (iron, zinc, arsenic, cadmium, and lead) below the permissible limits17,18 which commonly leads to develop favorable condition for bacterial growth. The gutter samples showed antibiotic resistance towards penicillin, cefalosporins, carbapenems, aminoglycosides, fluoroquinolone, diaminopyrimidines classes of antibiotic. Among these semi urban gutter samples only HL_NS-5 was sensitive to all the five classes of antibiotics except aminoglycosides. Previously, Sayah et al.19 reported the antibiotic resistance E. coli in Domestic- and Wild-Animal Fecal Samples, Human Septage, and Surface Water against Neomycin, Streptomycin, Sulfisoxazole, Tetracycline, Cephalothin antibiotics. After that Amato et al.20 reported the antibiotic resistance E. coli in ground waters especially agricultural run offs and irrigation waters from Valencia, Spain against sulfamethoxazole, trimethroprim, ciprofloxacin, tetracycline, azithromycin, nalidixic acid, cefotaxime, chloramphenicol, ceftazidime, ampicillin, gentanicin antibiotics. From these observations it can be said that due to time intervals, the antibiotic resistance was elevated sharply and gradually it developed pan resistance towards broad spectrum antibiotics. In this study, apart from the determination of specific antibiotic resistance against specific antibiotics, the general beta-lactam resistance of the bacterial community was assessed and as similar to the conventional antibiotic resistance profile, except HL_NS-5, all the five samples showed beta lactam resistance through beta-lactamase activity. This is the first report to assess beta-lactamase based resistance rapidly in an environmental sample.

Pharmaceutical products, including antibiotics, are partially metabolized by humans, leading to their excretion in the environment. Urban sewage systems and gutters, which are the first step of the city wastewater cycle, are hotspots for microbial diversity and activity, potentially causing the primary emergence of antibiotic-resistant bacteria, which can impact population health21. Metagenomic analysis aids in understanding microbial populations in water and sediment, providing functional and genomic information beyond traditional fecal coliform counts for monitoring water quality22. Hu et al.,23 also reported the prevalence of Mycobacterium and Nakamurella in the dormitory drain pipes which may cause indoor contamination and increase disease transmission risk. In this study, the bacterial population resided in the gutter samples was proteobacteria, firmicutes, bacteroids, actinobacteria, planctomycetes predominantly whereas the predominant genera were Pseudomonas, Escherichia, Hydrogenophaga. Previously it was reported that in waste water treatment plants (WWTPs), Actinobacteria, and Proteobacteria were predominantly present in phylum level whereas Mycobacterium, Acidovorax, and Polaromonas were the prevalent genera24. Among all the families and genera, majority were common in all the six sample sites and the uniqueness of the taxa were very less in the specific sample sites. The positive correlation coefficient of all the six gutter samples indicates significant variation in microbial population between the gutter samples. The bacterial population of the six gutter samples possessed common metabolic pathways like citric acid metabolism, carbon metabolism, pentose phosphate metabolism, amino acid metabolism, pantothenate and CoA biosynthesis, porphyrin, butanoate, propanoate, nitrogen, fatty acid metabolism etc. One of the interesting observations was the presence of drug metabolism which signifies the presence of drugs in the gutter samples and these might be cause of the development of antibiotic resistance25,26. The presence of drugs and streptomycin biosynthesis, globo and isoglobo series of glycosphingolipid biosynthesis, lipopolysaccharide biosynthesis pathways in the bacterial community might also be responsible for development of antibiotic resistance in the bacterial population26. Previous study also enlightens the presence of different carbon metabolism pathways (purple sulfur bacteria, purple non-sulfur bacteria, and oxygenic photosynthetic bacteria), methylotrophic, methane oxidizing pathways, nitrogen and iron metabolic pathways in the street gutter samples similar to the present study27. However, there was no such information to enlighten the possible metabolic pathways to develop antibiotic resistance. This study holistically uncovers the gutter water quality, their bacterial population, and antibiotic resistance profiles. The potential metabolism was predicted to understand the root of antibiotic resistance also familiarize in this study.

In a nutshell, microbial population in the semi urban gutter samples and development of its antibiotic resistance is one of the key concerns in environmental microbiology and social health. These cattle-side, roadside, and household gutters harbor and disperse the microbial communities that may develop pathogenesis to adjacent plants, animals, and humans that leads to community diseases. The development of broad-spectrum beta lactam resistance due to beta lactamases production is the key concern towards community health. Although there was no heavy metal contamination risk in the studied environment. The presence of proteobacteria, firmicutes, and bacteroids predominantly in the gutter microbial population leads to this antibiotic resistance. The presence of drug, antibiotic (streptomycin), glycosphingolipid, lipopolysaccharide metabolism pathways in the microbial communities are the important biosynthetic cause of development of antibiotic resistance. These findings suggest that urban gutter water is a reservoir of antibiotic-resistant bacteria with antibiotic resistant metabolic pathways and beta-lactamase genes, posing a potential risk to public health and environmental ecosystems. Therefore, targeted action is necessary on proper cleaning, sanitization and reduction of microbial antibiotic resistance in the semi urban cities of India.

Materials and methods

Sampling and site description

The gutter samples were collected from six locations of the semi-urban town, Roorkee, Uttarakhand, India (Fig. 8, Table S5) in a triplicate manner. Samples were collected during the summer (April 2024) when the gutters are in low flow conditions with higher deposition of solid garbage. The permission for collecting the samples was taken from the Roorkee Nagar Nigam on April 2, 2024. The samples were collected in a sterile container and stored at 4 °C for further studies. All the samples were collected in triplicate manner and the mean data of three samples were reported accordingly.

Fig. 8
figure 8

Gutter sample sites represented in Google Map in semi urban town Roorkee, India.

Full size image

Physical characterization of the samples

The physical properties like pH, and total electrolytes content were determined according to standard protocol28. Briefly, pH and total electrolytes content was determined using a pH meter equipped with conductivity meter from 100 ml of gutter water and the data were recorded. The five heavy metals, iron (Fe), zinc (Zn), arsenic (As), lead (Pb), and cadmium (Cd), are commonly correlated to human activities or ARGs occurrence in the environment29. Ten milliliters of water samples were filtered by a syringe filter 0.45 μm and stored with additional nitric acid HNO₃ 2%30. The heavy metals were assessed using an inductively coupled plasma-mass spectrometer (ICP MS, 8900 ICP-MS Triple Quad). The calibration plots and heavy metal concentrations in the water samples were created using gradients of standard solutions.

Isolation and quantification of bacteria

Bacterial isolation and quantification were performed using the colony-forming unit (CFU) method31. Aseptically collected samples were subjected to tenfold serial dilutions using sterile phosphate-buffered saline (PBS), and 100 µL aliquots from each dilution were plated on nutrient agar using the spread plate technique. The plates were incubated aerobically at 37 °C for 24 h, after which colonies were counted manually. Plates containing 30–300 colonies were selected for accurate enumeration, and the CFU per milliliter was calculated using the formula: CFU/mL = (Number of colonies × Dilution factor)/Volume plated.

Antibiotic resistance profiling

The antibiotic resistance profiles of the total bacterial population of gutter samples were assessed against eight antibiotics (ampicillin, kanamycin, cefotaxime, meropenem, oxacillin, ciprofloxacin, cephalothin, and trimethoprim) according to Clinical Laboratory and Standards Institute (CLSI) guidelines, 202332 by replica plating method in Mueller–Hinton agar (MHA) and incubated for 14 h at 37 ℃. After incubation, the number of resistant colonies forming units (CFU) and the percentage of resistance was calculated33.

Determination of beta lactamase mediated antibiotic resistance

The presence of beta lactamases in the gutter samples were assessed qualitatively by the application of beta-lactamase substrate nitrocefin (1mM)34. Briefly, 50 ml gutter samples were filtered using filter paper with the pore size 0.22µM. Then 200 µl nitrocefin was added on the filter paper and waited for 10 min. The change of colour from yellow to red indicated the presence of beta lactamase in the environmental samples32. Wild type E. coli BL21 (DE3) cells without the plasmid vector (pET28a) was used as negative control whereas the same cells transformed with the beta-lactamase KPC recombinant plasmid vector served as the positive control.

DNA isolation and 16 S rRNA gene Amplicon-based illumina library Preparation

Metagenomic DNA preparation, 16S rRNA gene library preparation and Illumina MiSeq sequencing were done by Biokart India Pvt. Ltd., India. Using the Xploregen kit, the metagenomic DNA was extracted in accordance with the manufacturer’s instructions. DNA concentration and purity were assessed using 1% agarose gel electrophoresis and Nanodrop 2000c spectrophotometer. The universal 16S rRNA gene forward primers 5’-AGAGTTTGATGMTGGCTCAG-3’ and reverse primer 5’-TTACCGCGGCMGCSGGCAC-3’ were used for the amplification of V3-V4 region of 16S rRNA gene. Following the construction of libraries using the PCR amplicons, paired-end sequencing was performed using an Illumina MiSeq sequencing platform35.

Metagenomic data analysis: The base quality, base composition, and GC content were used to determine the quality of the raw metagenomic sequences. Following the appropriate trimming of the adapters, Biokart Pipeline eliminated the chimera of the raw sequences, and QIIME pipeline v1.9.1 was used to double-check the results36. Following the identification and filtering of the V3 areas from the paired-end data, the FLASH or clustelO algorithms were used to generate the consensus sequences from the paired-end data. The V3 region’s trimmed consensus sequence was processed using Biokart Pipeline. Uclust (similarity cutoff = 0.97) and QIIMEv1.9.1 were used to filter (with < 5 reads), and search and analyze the similarity of representative sequences. The SILVA database was used for the taxonomic classifications. According to usual procedure, the OTUs and abundance of each taxon were ordered taxonomically from phylum to genus level37,38. Using the Circos software v0.69-9 (https://circos.ca/intro/tabular_visualization/), the bacterial community of the six gutter samples was correlated at phylum levels39. The top ten taxa of families and genera were then shown using the Heatmapper2 software (http://www.heatmapper2.ca/). Using Bioinformatics & Evolutionary Genomics website (https://bioinformatics.psb.ugent.be/webtools/Venn/), the Venn diagram was used to view the OTU sharing percentages of six gutter samples. Each sample’s species richness, evenness, Shannon, and Simpson diversity indices were calculated, and the Kruskal-Wallis test was used to compare the results. The Bray–Curtis dissimilarity and the Cody index were used to assess the beta diversity using Paleontological Statistics package version 4.01.

Prediction of metabolic pathways: The metabolic KEGG pathways of the communities were predicted and built using Microbiome analyst40 (https://www.microbiomeanalyst.ca/). The heatmap was built using the Heatmapper2 software (http://www.heatmapper2.ca/) with the Spearman Rank Correlation distance measurement method and the clustering was done by average linkage between the hits of the metabolic pathways.

Statistical analysis

GraphPad Prism 8.0 and the Paleontological Statistics package version 4.01 (PAST software) were used to conduct statistical studies of the bacterial community. In accordance with usual procedure, the PCoA was utilized to assess the differences in the bacterial community structure among the six samples using the Bray–Curtis dissimilarity coefficient using Past v4.0135. Using PAST v4.01 software, the unweighted pair group method with arithmetic mean (UPGMA) tree was first built using the bacterial genera found in the gutter samples41. GraphPad Prism 8.0 software was then used to undertake analysis of variance (ANOVA) investigations from the OTUs, maintaining p < 0.05 as a significant level. Mean ± SD (standard deviation) was used to indicate the mean values of the triplicate data.

Deposition of metagenome data: Raw reads of six samples were deposited in fastq format to the National Center for Biotechnology Information. Sequences were deposited under Bioproject PRJNA1212300. The accession number of HL_NS-1, HL_NS-2, HL_NS-3, HL_NS-4, HL_NS-5, HL_NS-6 were SAMN46296458, SAMN46296459, SAMN46296460, SAMN46296461, SAMN46296462, SAMN46296463, respectively (https://www.ncbi.nlm.nih.gov/biosample?LinkName=bioproject_biosample_all&from_uid=1212300).