Introduction

Approximately 356 million smokeless tobacco (SLT) users have been estimated in 140 countries that accounts for one-fourth of the total tobacco consumers [1]. Smokeless tobacco is one of the deadliest reasons for 1.5 million disability-adjusted life years (DALYs) among 113 countries [2]. The prevalence of SLT users is more in lower and lower-middle-income countries like India [1]. The factsheet of the World Health Organization (WHO) states that SLT consumption accounts for 48% of cardiovascular diseases, 23% of chronic respiratory diseases, 14% CMNNDs (communicable, maternal, neonatal, and nutritional diseases), and 10% cancer [3]. The existence of approximately 4000 chemical constituents in smokeless tobacco and derived products makes it a risky chewing/smoking product [4]. Several of these chemicals are derivatives of toxicants and carcinogens that include tobacco-specific nitrosamines (TSNAs), N–nitrosamino acids, volatile aldehydes, polyaromatic hydrocarbons (PAHs), metals, and metalloids. Besides, smokeless tobacco also exhibits a huge microbial load of bacteria, fungi, molds, and other microorganisms [5]. Han et al (2015) report that bacterial load on various SLT ranges from 8.33 ×101 to 2.54 ×105 CFU/g SLT [6], where numerous opportunistic pathogens (Eubacterium, Porphyromonas, and Prevotella) have also been documented. Several bacterial residents of smokeless tobacco assist in the formation of tobacco-specific nitrosamines (TSNAs) [7] as well as acetaldehyde production, a carcinogen [8]. The International Agency for Research on Cancer (IARC) has identified five types of TSNAs and categorized them into group I carcinogens [9].

Oral cavity encounters first with SLT products along with their inhabitant microorganisms and thus alters the overall bacterial composition [10]. Dysbiotic microbiota of the oral cavity has been reported for the onset of several diseases including the cancers of the oropharynx [11, 12], where SLT products may further influence the overall bacterial communities.

SLT users are prevalent in India and its consumption leads to one million deaths every year in India [3]. We have limited information on oral microbiota from the Indian community [13], where we have no concrete data on the oral microbiome of SLT consumers. Therefore, the present investigation has been carried out to determine the oral bacterial composition of SLT users from a set of populations of north India and to assess the alteration of bacterial communities due to the SLT consumption as compared to the control group. Our findings revealed significant alterations in the oral bacterial communities among SLT users at various taxon levels (2 phyla and sixteen genera; p-value <0.05) as compared to the non-tobacco chewers. Besides, the phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt) analysis also showed an apparent shift in the functions, especially the genes involved in the metabolism of nitrogen, amino acids, and nicotinate/nicotinamide. The investigation will pave the way to determine the core oral microbiome of SLT users and to understand the influence of smokeless tobacco on a healthy oral microbiome especially from a developing country like India, where the SLT consumers account for 66.6% (approx. 237.4 million) of the total SLT consumers (356 million) among 140 countries [1].

Materials and Methods

Sample Collection, DNA Extraction, and Sequencing

Oral wash samples were collected from SLT users (n=20; code: HTC) and SLT non-users (n =20; code: HI) after their prior consents. The SLT users were in tobacco consumption habits for more than 5 years having 15–25 g of SLT product intake in a week. The protocol of the study was approved by the Institutional Ethical Committee, Netaji Subhas University of Technology, New Delhi (NSUT), as well as the All India Institute of Medical Sciences (AIIMS), New Delhi. None of the participants (HI and HTC) was alcoholic and was not on any medication/antibiotic for the last 3 months at the time of sample collection (Table 1). The health status of the participants was detected as healthy in terms of their diabetic status, systolic BP, and BMI. Besides, none of the participants have any lesions in their oral cavities. Metagenomic DNA extraction, V3 hypervariable region amplification, sequencing, and processing of reads have been carried out as mentioned in our recent publication [12]. Briefly, the metagenomic DNA was extracted using prior treatment with proteinase-K (10 mg/ml) and lysozyme (10 mg/ml) followed by the Qiagen DNeasy Blood and Tissue Kit (Qiagen, USA). The amplicon library of the V3 hypervariable region was prepared and sequenced at the Illumina MiSeq 2500 platform (Illumina, San Diego, USA) (C:\Users\pc\Desktop\HTC NGS sequences). The sequences having a Phred score value >30 were processed using different filters such as conserved region, spacer, mismatch, and consensus read to obtain the quality reads. ClustalO program [14] and UCHIME [15] of USEARCH [16] were other programs used for downstream processing of the sequences.

Table 1 Various parameters of the healthy smokeless tobacco users during this investigation
Full size table

OTU Generation, Taxonomy Assignment, and OTU Clustering

The QIIME program was used for generating operational taxonomic units (OTUs) at the similarity cutoff of 0.97 from the pre-processed sequences (QIIME version 1.9.1). The taxonomy assignment was carried out using the reference of the Greengenes database with the assistance of the PyNAST program.

Analysis of Diversity Indices, Data Interpretation, and the PICRUSt Analysis

Various diversity indices (Chao1, Shannon, Simpson, PD whole_tree, and Good’s estimator Shannon and Simpson indices) were calculated to analyze the diversity of the bacteria in each sample [17] (Table 2). The data have been interpreted for different outputs such as heatmap and correlation graphs using CAMEO analysis tool, (https://github.com/avishekdutta14/CAMEO), SPSS V26, NCSS V20.0.2, and STAMP V2.1.3 [18]. The functional prediction of the 16S rDNA genes was performed using the PICRUSt algorithm version 1.1.0 [19].

Table 2 Diversity indices and estimators obtained among SLT users
Full size table

Interpretation of the Data and Statistical Analysis

The generated OTUs were analyzed for the overall bacterial composition in the healthy oral microbiome of SLT users and non-users at the various taxon level. An attempt was made to determine the influence of tobacco on overall bacterial composition as well as their predicted functions among SLT users as compared to the SLT non-users. The data were evaluated for the assumption of normality. In the case of departure from normality, the equality of means between the groups was compared using Mann Whitney U test at the level of phylum and genera. A General Linear Model procedure was applied to assess the impact of tobacco on microbial diversity of subjects adjusting for categorical predictors such as sex and diet including interaction terms. Principal coordinate analysis (PCoA) plots were generated using the first two principal coordinates and labeled according to smoking status using the Euclidean distance matrix. The statistically significant differences at the level of tier I, tier II, and tier III were performed using a two-sided G-test (w/Yates)+Fisher’s test at a 95% confidence interval to analyze the PICRUSt output using STAMP.

Results

FastQC Analysis and Pre-processing of the Reads

The paired-end amplicon sequencing from twenty different samples revealed 8,080,889 high-quality reads (~680 Mb) which were converted into 86,270 OTUs on passing through four different filters along with the removal of consensus reads and chimeric sequences, and singletons (Supplementary table 1, 2, 3, and 4). It could be due to the use of QIIME 1.9.1-UCLUST that usually generates a large number of counterfeit OTUs [20]. However, UCLUST offers the fastest runtime with reasonably good performance; therefore, it can be used for microbiome analysis [21].

Bacteriome of SLT Users

Overall Phyla

On taxonomic assignment, the generated OTUs were assigned into 15 phyla, 23 classes, 46 orders, 99 families, and 180 genera. Overall analysis revealed that Proteobacteria (25.836%) dominates over Firmicutes (23.844%) among SLT users followed by Bacteroidetes (19.664%), Fusobacteria (3.222%), Actinobacteria (5.392 %), Candidate division _GN02 (0.081 %), Spirochaetes (0.363 %), Candidate division _SR1 (0.3059 %), and Candidate division _TM7 (0.197 %). Other small contributors of the SLT user’s oral cavity were Tennericutes (0.0789%), Synergistetes (0.033%), Cyanobacteria (0.262 %), Chloroflexi (0.0016%), Elusimicrobia (0.001%), and Acidobacteria (Fig. 1a).

Fig. 1
figure 1

The relative abundance of phyla (a) and top twenty genera, (b) present in the oral cavity of each SLT user.

Full size image

Overall Genera

The taxonomic assignment at the genus level revealed a total of 180 assigned genera. Overall analysis among twenty participants revealed that Neisseria (14.30%) was among the dominant genera followed by Streptococcus (12.20%) and Prevotella (9.493%). It was further extended by Hemophilus (4.403%), Porphyromonas (4.251%), Rothia (4.091%), Granulicatella (2.978%), Actinobacillus (2.831%), Veillonella (1.963%), and Fusobacterium (1.863%). The remaining genera exhibited their portion below 1% (Fig. 1b).

Diversity Indices

Subject HTC-8 and HTC-12 showed the maximum richness or diversity having fourteen phyla followed by HTC-1, HTC-2, HTC-3, HTC-4, HTC-6, HTC-7, HTC-10, HTC-13exhibited thirteen phyla, while subjec, HTC-15, HTC-16, HTC-17, and HTC-20 having thirteen phyla in each subject. A total of twelve phyla were observed in the subjects HTC-5, HTC-9, HTC-11, HTC-14, HTC-18, and HTC-19 (Supplementary file 1). The calculated diversity indices were also observed highest in HTC-18 and HTC-12 having the Shannon index of 6.93 and 6.25 respectively. The values of Simpson indices (SI) for the subject HTC-18 also showed highest in all subjects having a value of 0.9761331 followed by HTC-12 (0.965). It was followed by HTC-17 (SI: 0.960) and HTC-3 (SI: 0.958). While the subject HTC-2 showed the least richness as compared to the indices of other subjects, all the subjects showed Good estimator’s values > 99. Comparative analysis between SLT and non-SLT users were further analysed for various diversity indices (Fig. 2)

Fig. 2
figure 2

Box plots illustrating alpha diversity indices (Chao1, Simpson index, and Shannon index) in bacterial microbiomes of SLT users (HTC) and non-users (HI) samples. Median values and inter-quartile ranges have been indicated in the plots

Full size image

Inter-individual Differences at the Level of Phylum and Genera

Comparative analysis of the top ten phyla among twenty SLT users revealed that Proteobacteria were dominant in half of the participants (HTC-1, HTC-4, HTC-5, HTC-7, HTC-9, HTC-10, HTC-13, HTC-14, HTC-17, and HTC-20), where HTC-20 exhibited the maximum OTUs (48.950 %) of Proteobacteria, while HTC-8 showed a minimum of 7.49% OTUs. Similarly, Firmicutes were observed as dominant phyla among eight participants (HTC-2, HTC-3, HTC-6, HTC-8, HTC-11, HTC-12, HTC-18, and HTC-19). Bacteroidetes were dominant only in one SLT user, i.e., HTC-15. However, it was of second most abundant phylum in more than half of the participants (HTC-2, HTC-5, HTC-6, HTC-7, HTC-9, HTC-10, HTC-11, HTC-12, HTC-17, HTC-18, HTC-19, and HTC-20) (Supplementary file 1).

Inter-individual differences at the genus level resulted in high variation. Six subjects (HTC-4, HTC-7, HTC-13, HTC-14, HTC-17, HTC-20) were populated by Neisseria, while Streptococcus showed its dominance in five participants (HTC-1, HTC-3, HTC-6, HTC-11, HTC-18). Two oral cavities were dominated by the genera Prevotella (HTC-2 and HTC-15) and Rothia (HTC-8 and HTC-16); however, Prevotella was found to be the second contributor in seven subjects (HTC-4, HTC-10, HTC-13, HTC-17, HTC-18, HTC-19, and HTC-20) (Supplementary file 2).

Shared Genera

Of the one hundred eighty assigned genera, thirty-eight genera showed their occurrence among all the twenty SLT users (Supplementary file 3). The top twenty genera include Neisseria, Streptococcus, Prevotella, Haemophilus, Porphyromonas, Rothia, Granulicatella, Actinobacillus, Veillonella, Fusobacterium, Actinomyces, Leptotrichia, Aggregatibacter, Capnocytophaga, Treponema, Campylobacter, Corynebacterium, Parvimonas, Selenomonas, and Mogibacterium (Fig. 1b). Peptostreptococcus, Bacteroides, Candidate Division_TG5, Cardiobacterium, Catonella, Microbacterium, Streptomyces, Paenibacillus, and Bifidobacterium were the other highly shared genera among 85 to 95% of the SLT users. It was followed by Lactobacillus, Peptococcus, and Candidatus division SHD-231 that were shared among 80% of the total SLT oral cavities. Approximately seventy-five genera were common in less than 80% of the oral cavities. Approximately twenty-five genera were least shared and observed among two participants only. Halomonas, Candidatus_Tammella, Roseomonas, Marinilactibacillus, Roseburia, Arcobacter, Sneathia, and Xanthobacter are the name of few such genera. Saccharopolyspora, Acetobacter, Bilophila, Limnobacter, Brevibacillus, and Tessaracoccus are among the thirty genera that were not shared and identified at the individual level. On the Spearman correlation analysis, heatmaps of shared genera showed a comparatively weak correlation among SLT users (Fig. 3a) as compared to the control group (Fig. 3b).

Fig. 3
figure 3figure 3

Spearman correlation heatmaps for (a) HTC and (b) HI were generated using un-weighted pair group (average group) clustering method, including only those genera that were shared in both the groups and/or identified as significantly different in between the groups using NCSS software version 20.0.2. HI groups exhibit comparatively strong positive correlation with each other as compared to the HTC group

Full size image

Comparative Analysis of SLT Users and Non-users

Phyla Level

Among SLT users, the total number of assigned phyla was reduced to 15 as compared to the 20 bacterial phyla in SLT non-users [12]. These depleted phyla were Chlorobi, Thermi, Gemmatimonadetes, Nitrospirae, and Candidate division_OD1 as compared to the phyla detected in the oral microbiome of SLT non-users.

Genera Level

Here, the total assigned genera were increased among SLT users as compared to the control group. Among SLT users, the total assigned genera were 180 than that of 162 among the SLT non-users (Supplementary table 5). Moreover, 122 genera were common in between the groups, where 58 genera were unique in the SLT group. Similarly, forty genera of SLT non-users were missing among SLT users.

Statistical Analysis

Of the fifteen common phyla, Fusobacteria and Tenericutes were detected as significantly different (p-value < 0.05) among both groups. The p-values and FDR-adjusted p-values are shown in Supplementary table 5. Genera level analysis revealed a total of sixteen genera, viz., Fusobacterium, Porphyromonas, Parvimonas, RFN20, Oribacterium, Lactobacillus, Gordonia, Coprococcus, Moryella, Catonella, Microbacterium, Streptomyces, Desulfobulbus, Enterococcus, SHD.231, and, Heamophilus that were significantly different in between the two groups (Fig. 4). SLT users showed more prevalence of Fusobcaterium, Porphyromonas, RFN20, Parvimonas, Gordonia, Moryella, Catonella, Microbacterium, Desulfobulbus, Enterococcus, and SHD.231, while other remaining genera were relatively higher in SLT non-users (Supplementary table 5). The p-values and FDR-adjusted p-alues are shown in Supplementary table 5.

Fig. 4
figure 4

The OTU (%) distribution of statistically significant genera between SLT users (HTC) and non-users (HI). Data was normalized by taking center log ratio (clr)

Full size image

Effect of Tobacco Consumption at the Level of Gender and Diet

Only one genus Atopobium (p-value <0.001) was identified as significantly different at the level of gender, while Mycoplasma (p-value 0.027) was significantly different at the level of diet.

PICRUSt Analysis

The relative abundance of predicted genes revealed that the majority of the OTUs belong to the genes involved in metabolic pathways, i.e., 47.14% followed by the genes of genetic information/processing (21.63%), environmental information processing (12.74%), and human diseases (0.98%). Tier I showed a significant increase in the count of genes involved in cellular processes, environmental information processing, human diseases, and metabolism, among HTC users as compared to the control group (Benjamini–Hochberg FDR-adjusted value <0.05, q value −1e−15) (Supplementary fig 1a). Tier II represents the extension of tier I and its analysis revealed a significant abundance for the genes that belong to amino acid metabolism, cellular processes, and signaling, metabolism of cofactors and vitamins, xenobiotic biodegradation, and metabolism in HTC users, while SLT non-users showed an abundance of genes for carbohydrate metabolism, replication, and repair, lipid metabolism, transcription, translation, and nucleotide metabolism in tier II analysis (q value: −1e−15) (Supplementary fig. 1b). Tier III further expands the tier I– and tier II–level genes. Overall, tier III analysis showed an enhancement in the number of genes involved in nitrogen metabolism, amino acid metabolism, and nicotinate/nicotinamide metabolism, as well as toluene degradation in SLT users (Fig. 5). Besides, SLT users also showed more abundance for the genes involved in bacterial secretion systems, bacterial motility proteins, and other transporters, while the genes of starch and sugar metabolism, glycolysis/gluconeogenesis, DNA repair and recombination proteins, ABC transporter, and phosphotransferase system (PTS) were more abundant in SLT non-users (Benjamini–Hochberg FDR-adjusted q value= 1e−15).

Fig. 5
figure 5

The PICRUSt analysis at tier III reveals significant differences among the genes involved in toluene/nitrotoluene degradation as well as nicotinate/nicotinamide degradation

Full size image

Discussion

India bears a high burden of smokeless tobacco consumers due to the occurrence of a wide variety of SLT products that are easy to access to every age group [1]. Tobacco consumption is one of the major reasons for preventable deaths globally, where India accounts for approximately one million deaths every year due to tobacco consumption in any form [3].

Our study encompasses the first extensive characterization of the oral microbiome of SLT users from the Indian population. Here, we also report the effect of smokeless tobacco on a healthy oral microbiome and observed that alterations occur at various taxon levels and so forth their predicted functions also deviate accordingly. Various forms of tobacco products have been reported to influence the microbiome of healthy oral cavities [10, 22] but such information is lacking from the Indian community. We observed that the count of total genera was 180 in SLT users as compared to 162 in the control group. Of these 180 genera, 122 genera were shared in both the groups, where 58 were unique to SLT users only. It may be due to the introduction of inhabitant tobacco bacteria that compete with the autochthonous bacteria of the oral cavity for nutrients and other limiting substrates and consequently enhance the overall bacterial composition [23, 24]. The depletion of 40 genera in SLT users as compared to the control group is also noteworthy and envisages that several toxicants, heavy metals, alkaloids, and other xenobiotic compounds of SLT may be the possible reasons after it [5]. In a traditional cultivation approach, SLT has been reported to alter (promote/inhibit) the growth of eleven isolated oral bacteria, where SLT promoted the growth of seven strains and inhibited 4 strains that exhibit a role in biofilm formation [22]. The Middle Eastern smokeless tobacco product Dokha has also shown a significant association with oral microbiome dysbiosis among the UAE population [25].

The statistical analysis uncovered that the beta diversity between the groups was consistent (p-value 0.7) (Supplementary fig. 2). However, differently significant candidates were identified at the level of the various taxa. Two phyla (Fusobacteria, and Tenericutes) were identified as significantly different between the two groups (p-value <0.05) (Supplementary table 5). Elevated count of Fusobacteria among SLT users indicates poor oral health. Fusobacteria belongs to the highly common anaerobic Gram-negative bacteria of the oral cavity that act as a connecting link between early and late colonizers of dental plaque [26]. Due to several pathogenic characteristics like production of adhesins, co-aggregation with other bacteria, susceptibility towards several potent antibiotics, incorporation of outer membrane protein (OMPs), and their abundance during poor dental hygiene makes it a potential candidate for the onset of several periodontal diseases [27]. It further suggests that tobacco consumers are more prone to Fuosobacterial diseases. A marginal decrease in the count of Tenericutes (p-value: 0.049) in SLT users is the subject of further research, especially in oral health. Mycoplasma, a highly studied genus of the phylum Tenericute, has been associated with the pathogenesis of severe respiratory and urogenital tract infections [28]. The count of Tenericutes was comparatively higher in the patients with periodontitis as compared to the healthy controls [29].

Statistical analysis revealed only sixteen significantly distinct genera in between these two groups (p-values <0.05). Besides, Fusobacterium as discussed above, other genera such as Porphyromonas, Parvimonas, Gordonia, Moryella, Catonella, Microbacterium, Desulfobulbus, Enterococcus, and SHD.231 were also significantly elevated in SLT users. Fusobacterium, Porphyromonas, Haemophilus, Enterococcus, and Lactobacillus have been identified as highly altered genera during tobacco-associated oral dysbiosis also [30]. The association of Fusobacterium and Porphyromonas in dental biofilm formation is well established and paves the way for several oral diseases [26, 27]. Parvimonas (formerly known as the member of Micromonas) belong to Gram-positive anaerobic cocci (GPAC) and shares properties with Peptostreptococcus [31] which further affirm our findings that these antecedent bacteria may act as an early colonizer in biofilm formation and allow the adhesion of Fusobacterium- and Porphyromonas-like Gram-negative bacteria [26, 27]. Parvimonas micra is the only reported genus of this taxon to date [32, 33] and also has been reported as a prominent oral pathogen [31]. Similarly, the genus Catonella and Moryella also exhibit only one species Catonella morbi [34] and Moryella indoligenes [35] respectively which cause endodontic infections and periodontitis. Moryella is a weak saccharolytic bacteria [35] that produce acid during carbohydrate fermentation and may create an acidic environment favorable for adherence of other strong saccharolytes into the oral cavity and elevates their count during oral illness [36, 37]. In our case, the count of Moryella was significantly higher in SLT users (p-value 0.02) which suggests further exploring its role in oral health. Elevated count of Enterococcus (a lactic acid bacteria) may also suggest a similar function in SLT user’s oral cavity. Besides, members of Enterococcus have been recognized as the transient constituents of the oral microbiome advocate its appearance in the oral cavity due to external sources [38]. Tsuzukibashi et al (2015) suggest that Microbacterim does not constitute the human normal oral microflora and exhibit in denture wearing oral cavities [39]. Gordonia and Microbacterium count were another significantly abundant methylotrophic actinobacteria in SLT users that further suggest them as transient genera in the healthy human oral cavity [39, 40]. The members of these genera have been recognized as environmental strains having a role in nitrogen metabolism [41, 42]. Therefore, these genera may play a key role in the transformation of nicotine to tobacco-specific nitrosamines [43, 44]. Such findings encourage us to identify the inhabitant bacteria of smokeless tobacco products for enhancing our information on transient bacteria. Increased Desulfobulbus count in SLT users corroborates our findings with the previous reports, where enhanced Desulfobulbus count has been identified in the patients of periodontal diseases [45, 46]. Anaerobic and chemoorganothrophic Desulfobulbus potentially transform sulfur into sulfate/sulfide and plays a crucial role in dental caries [47]. The prevalence of methylotrophic bacteria and Desulfobulbus in SLT user’s oral cavity can also be linked for halitosis.

We obtained a lower count of Lactobacillus and Haemophilus in SLT users. It may be due to the inhibitory effect of SLT products on several other oral isolates [7, 48]; therefore, it needs to be examined carefully using standard strains by traditional cultivation approaches. In an interesting study, Lactobacillus count was reduced in presence of a higher concentration of nitrate in the medium [49]; therefore, we may hypothesize that the oral cavity of tobacco consumers exhibits an abundance of nitrate/nitrite that can be estimated in terms of TSNAs (2.3–27 μg/g SLT powder in India; 50). Lactobacillus and Haemophilus have been reported as health compatible commensals and thus their reduction signals for poor oral health [24]. Other significantly depleted genera in SLT users were Oribacterium, Caprococcus, and Streptomyces. The elevated count of Oribacterium has been reported in oral cancers [51, 52]. Streptomyces is chiefly a soil dweller while Coprococcus is more prominent in fecal samples; therefore, their occurrence in the oral cavity may be considered as conjectural [53, 54].

Analysis at the level of gender and food habits further validates our findings that food habits (vegetarians and non-vegetarians) and gender do not affect the overall bacterial composition if the population belongs to the same demographic regions [13]. Only two genera Atopobium and Mycoplasma were statistically different at the gender and diet level respectively. Atopobium, a Gram-positive H2S-producing rod is used as a biomarker for the diagnosis of OSCC (oral squamous cell carcinoma) and plays a crucial role in bacterial vaginosis and colorectal cancer [55]. However, its significant abundance at the gender and diet level needs to be further examined.

The statistical analysis of predicted genes in between the groups revealed significant findings at tiers I, II, and III using STAMP version 1.2.3. These levels depict the hierarchical functional pathways of the Kyoto Encyclopedia of Gene and Genome (KEGG) that were obtained by using the output of the PICRUSt analysis. Differences between mean proportion showed a significant increase in the genes involved in metabolism, cellular processes, environmental information processing, and human diseases among HTC users. It indicates that tobacco and its inhabitant bacteria showing its influence in deciding the overall oral bacterial composition and therefore associated pathways are also being affected [56]. Tier II analysis uncovered extended analysis, where mean abundance proportion was relatively higher for the genes involved in the amino acid metabolism, xenobiotic biodegradation, and metabolism activity in HTC users. Oral bacteria encounters first with tobacco flakes that contain several recalcitrant compounds and toxicants [12, 50], and therefore bacteria may shift their metabolic activities for degrading such xenobiotic compounds [57]. A higher proportion of genes of toluene/nitrotoluene degradation, nicotinate/nicotinamide degradation, and nitrogen metabolism among SLT users at the tier III level further affirms the influence of tobacco on the overall bacterial metabolic activity as compared to the healthy participants. We hypothesize here that an increase in genes of nitrogen metabolisms indicates the conversion of several nitrogen derivatives and their transformation into tobacco-specific nitrosamines on reaction with nicotine into the oral cavity of SLT users.

Conclusion

The present investigation is the first information on the oral microbiome of SLT users from the Indian population that enhances our understanding of the oral bacterial composition of SLT users. Overall, we conclude that smokeless tobacco significantly alters (enrich/deplete) the oral microbiome. Enrichment of the SLT’s oral cavity with several pathogenic and anaerobic bacteria suggests the onset of various oral diseases. The information of significantly altered genera could be explored further for making biomarkers for an early diagnosis of tobacco-induced oral disease. However, further research is required in this direction with a large sample size by considering varied socio-economic factors to reveal a better picture. An attempt should also be made to explore the inhabitant microorganisms of various commercial tobacco products to establish a correlation with the oral bacteria. Shotgun metagenomics could be a better tool to uncover the overall dynamics of the oral bacteria in terms of their functional characteristics, where we can directly retrieve the information of the genes involved in nitrogen metabolism (especially the nitrogen reductase gene). This study has significance as India is among the largest consumers of SLT users, where we have no concrete information on the oral bacterial composition of SLT users/SLT products.