The oral cavity comprises different habitats, such as the teeth, tongue, gingival sulcus, that provide a complex ecosystem for differential microbial growth [37, 38]. The sheltered, cavitated-trenches of dentin-caries lesions, particularly in S-ECC, appear to be a unique habitat in this context [10, 39], as reinforced by the complex, polymicrobial consortia noted in the occlusal and proximal cavitated caries milieus in the present study. As far as we are aware, this is the first report to illustrate the richness and the diversity of site-specific microbes of occlusal and proximal caries lesions in S-ECC and compare the compositional differences and diversity of taxa in these two sampled sites.
Proximal caries lesions harbor a richer and diverse microbiota than occlusal caries lesions
Using established metrics, Bray Curtis dissimilarity analysis and Shannon diversity indices, the microbiota of occlusal and proximal deep dentinal caries lesions in S-ECC were found to be distinct in terms of the richness of high- or low- abundant taxa and species (Fig. 6b). Furthermore, MDS assessment indicated clearly that caries-microbial consortia from the proximal and occlusal samples of the same patient did not typically cluster together.
This phenomenon may be due to the anatomical and structural differences of the occlusal and proximal surfaces of deciduous teeth, and/or the intrinsic ecological differences in these two localities. For instance, the occlusal niches are constantly exposed to the ebb and flow of saliva with its arsenal of immune challenges, and the masticatory forces due to the intermittent food intake accompanied by the incessant tongue movements. Further, it has a much more dynamic environment than the proximal niches in between teeth, which are more sedate and well protected from such extrinsic stresses [30, 40]. In addition, the sheltered proximal cavitated locales are almost unreachable to routine oral hygiene measures [30, 40]. Thus, it is tempting to speculate that the inherent features overarching the proximal and distal caries ecosystems may be the keys reasons for the significant diversity in the microbiota of these two sites.
S. mutans and Veillonella alcalescens are the two most prevalent species recovered from all caries sites: We noted that S. mutans and Veillonella alcalescens were the most prevalent species, amongst all evaluated caries sites. There is an ample narrative on the link between S. mutans and dental caries due to its superior acidogenic and aciduric potential [12, 41, 42]. Our data confirms a report by Aas et al. (2008), who also noted that the most prevalent species in either the occlusal or proximal caries of children with S-ECC are S. mutans . Another observation that substantiates the work of the latter group (2008)  is the profusion of Veillonella alcalescens, which we noted in both the occlusal and proximal deep-dentine locales (p≥0.05). Classically, Veillonella spp. and S. mutans are known to be co-located and intimately associated with the caries process. The former is thought to nutritionally metabolize the carboxylic acids produced by streptococci, and thereby suppress the cariogenicity of the mutans-group of streptococci . On the contrary, others have noted that Veillonella alcalescens and S. mutans in tandem produce more acids than each of the species separately . Veillonella species also easily coaggregate with various oral microbes, including Streptococcus spp.  thus suggesting a high degree of synergism and mutualism between them [44, 46], as was noted here.
Atopobium parvulum is the most prevalent species in the occlusal caries lesions
We noted a highly significant (p=0.01) prevalence of a lactate-producing species, Atopobium parvulum, predominantly in the occlusal cavities. Previous studies have also mentioned isolation of the Genus Atopobium from the carious dentin of children [47, 48]. Atopobium is not only acidogenic but is also known to be aciduric in nature . The reason why this species predominates in occlusal rather than proximal lesions is unclear but could be due to the aforementioned ecological factors, as also described by Kleinberg and Jenkins (1964) . They measured the salivary flow and pH in different parts of teeth and observed contrasting pH values even on different surfaces of subjacent proximal teeth, which they surmised would impact the preferential microbial colonization and the eventual propensity for caries .
An abundance of acidogenic and aciduric flora in the proximal cariogenic locale: Detailed analysis of our data indicated the rich and diverse presence (p≤ 0.05) of acidogenic and aciduric taxons, particularly in proximal cavitated niches, in the following order: Propionibacterium acidificiens, Leptotrichia sp., Bifidobacterium dentium, and species of genera Lactobacillus. This is not surprising as these attributes are essential prerequisites for survival in a very low pH cariogenic niche. Others, too, have reported similar findings. Gross et al. (2012) observed P. acidificiens in dentinal caries lesions of children in ECC , while Obata et al (2014)  noted its avidity to dentine collagen, which may be another reason for their preponderance in deep dentinal lesions, in comparison to early enamel caries. Furthermore, Downes and Wade (2009) have described the saccharolytic attributes of P. acidificiens with the production of acetic, propionic, and succinic acids as end products of dietary carbohydrate metabolism . Other too have confirmed the aciduric potentials of P. acidificiens [53, 54], a criterion essential for survival in a low pH milieu of deep caries lesions.
Akin to P. acidificiens, the Genus Leptotrichia, a recognized putative cariogen, was noted in our cohorts in significant numbers (p≤0.05). Aas et al. (2008) have also described the high prevalence of Leptrotrichia in deep-dentine caries of deciduous teeth  which are known to ferment many mono- and disaccharides to lactic acid .
The acidogenic and aciduric Bifidobacteriaceae, which play a contributory role in caries progression , was also highly prevalent in proximal cavities (p <0.05). In line with our observation, Becker et al. (2002) identified Bifidobacterium species as the most prevalent cariogen and even outnumbered S. mutans in dentinal caries of children . Furthermore, in a very early in vitro study, Van Houte et al. (1996) have shown that Bifidobacteria had the potential to reduce the pH of glucose-supplemented media, upto to < 4.2 , adequate for demineralization of both enamel and dentine .
Historically, the two major cariogens were considered to be mutans-group streptococci and lactobacilli, and the latter is particularly known to be found in the advancing front of the dentinal caries lesions . Indeed, lactobacilli are a critical secondary pathogen in dental caries [20, 59]. Hence it is not surprising that we noted a spectrum of 22 lactobacillus species in the dentine caries samples of both the proximal and occlusal lesions. Of these, three species, L. salivarius, L. gastricus, and L. ghanensis stood out as significantly more prevalent in the proximal cavities (p≤0.05) (Fig. 2a).
Apart from the foregoing predominant genera, less well-known others were noted in significant proportions in the proximal cavitated niches (p=0.01). These were Actinobacteria belonging to Bifidobacteriaceae, namely Parascardovia denticolens and Scardovia inopinata. Supportive of these findings, Mantzourani et al. have also identified Bifidobacterium dentium, P. denticolens, S. inopinata from the caries samples of primary teeth . An intriguing characteristic shared by these species is their ability to degrade complex carbohydrates, including dextran  which potentiates the production of demineralizing acids within the cariogenic biofilm, even in the absence of fermentable carbohydrates . Furthermore, another recent study has reported the synergism in acidogenicity in dual-species biofilms of P. denticolens, S. inopinata, and B. dentium, with S. mutans . Taken together, it is clear that such microbial consortia play a decisive role in the pathobiology of deep caries lesions, in particular.
Several other attributes of the isolates mediate cariogenicity in S-ECC
There are several major pathogenic attributes of putative cariogenic flora that make them fit for their role and thus stand out as key cariogens, especially in deep dentinal lesions. These include, apart from their acidogenic and aciduric potential, potency to adhere and colonize both enamel and dentinal surfaces, collagenolytic and proteolytic potential to degrade dentinal collagen, and ureolytic properties that assist degradation of metabolic urea of biofilm microbiota. Some or most of these attributes are noted in the predominant species in both occlusal and proximal caries in this study.
Dentin tissue is essentially a hydroxyapatite mineral crystallite collagen matrix . Indeed, type I collagen comprises up to 90% of the organic material of the extracellular dentinal matrix . Cariogens S. mutans and B. dentium can preferentially colonize dentin as they possess adhesins (SpaA adhesins) that mediate their attachment to collagen [65, 66], while the former has additional collagen-binding proteins Cnm and Cbm [67, 68]. Hence, the duality of key attributes, their adhesive and acidogenic nature, makes S. mutans eminently suitable to be a key cariogen, particularly dentinal caries.
As opposed to the richness of acidogenic cariogens such as S. mutans, we also noted a significant proportion of health-associated ureolytic species C. matruchotii, Haemophilus parainfluenzae, and Actinomyces naeslundii in varying proportions in both the proximal and occlusal samples. C. matruchotii had been suggested as a nucleating species of the biofilm bacterial community in several studies. Welch et al. (2016)  noted a multi-genus conglomerate of nine taxa structured around filamentous corynebacterial cells. C. matruchotii, in particular, raises the biofilm pH by using acetate and lactate produced by the acidogenic plaque microbes . Furthermore, significant numbers of other ureolytic species such as H. parainfluenzae and A. naeslundii were also noted in our samples, as was earlier reported by Ma et al. (2015) in S-ECC plaque . The profusion of these urease-producing taxons amongst the abundant acidogenic, aciduric microbial consortia, particularly in the proximal deep-dentine milieu, is intriguing. It may be an attempt at keeping the pH to moderate survivable levels of the biofilm community. However, further studies are required to explore their role in such consortia.
It appears that microbe-mediated acidification of the plaque biofilm is not just the prime mover of caries progression . The acidic milieu, in turn, can activate endogenous dentin-embedded and salivary matrix metalloproteinases (MMPs) and cysteine cathepsins, which are thought to play a significant role in caries development [63, 72] as proposed by Takahashi and Nyvad .
For instance, the trio, Prevotella nigrescens, Fusobacterium nucleatum, and Aggregatibacter actinomycetemcomitans produce both intra- and extracellular gelatinolytic proteinases that may activate latent pro-MMP-9 . Furthermore, in our study, F. nucleatum was detected in both proximal and occlusal lesions. The latter is a core constituent of dental biofilms and plays a pivotal role in bridging microbes of early and late colonizing species . Thus, despite their relatively low numbers, the synergistic impact of the proteases of the trio, F. nucleatum, A. actinomycetemcomitans, and P. nigrescens, may significantly contribute to the structural disintegration of the collagenous scaffold of the dentine, in tandem with collagenase producing other species of Prevotella.
Several studies have reported the high prevalence of several Prevotella spp., particularly in the dentine caries [21, 48, 71, 75]. Indeed, this led Teng et al. to suggest the relative abundance of a panel of seven keystone Prevotella spp. from caries lesions of ECC  . Our findings, with a total of 27 Genera belonging to Prevotella species, with 23 species present at both the occlusal and proximal sites, and four species (P. intermedius, P. nanceiensis, P. marshii, and P. fusca) only at the proximal sites, confirm the assertions of Teng et al . Others too have echoed these sentiments and surmised that the overexpression of collagenases by Prevotella species during proteolytic metabolism might significantly contribute to the progression of dental caries especially at the advancing dentinal front . The rich aggregates of Prevotella Genus with such collagenolytic attributes that were recovered from our samples included P. denticola, P. histicola, P. melaninogenica, P. multisaccharivorax, and low prevalent, P. nigrescens, and P. intermedia. Indeed, all of these species have been previously recorded by others as isolates from cariogenic lesions [78–82] (Fig. 5). Our findings, therefore, further substantiate the view that Prevotella species play a leading role in proteolytic digestion and the progression of dentinal caries. However, further work is needed to ascertain the specific mechanisms by which they mediate such changes.
Collagenolytic microbes are highly prevalent in S-ECC (Fig. 5)
Although the acidogenic and aciduric attributes of S. mutans are well known, their ability to degrade human collagen (acid-soluble, type I) - the major constituent of dentine, by their extracellular proteases is poorly recognized. Some studies including, new metabolomic research, indicate the overexpression of collagenase gene expression in S. mutans in dental caries [79, 83]. As mentioned above, we had a rich harvest of S. mutans across all caries samples, albeit with a significant preponderance in the proximal niches. Their heavy presence in deep dentinal caries where collagen is plentiful appears to be a likely reflection of the tenacity and avidity of S mutans for collagen and fibronectin, as well as the abundance of peptidases and collagenases they possess . Finally, in this context, a range of several other cohabitant microbes with known collagenolytic attributes [66, 79, 80, 83, 84] was also isolated from dentin eco-niches. They were, B. dentium, S. inopinata of the phylum Actinobacteria, Selenomonas noxia and Veillonella parvulum . These consortia involved in initiation and degradation of the demineralized organic matrix of the dentinal tissues, and the activation of host-derived proteases ratifies and add credence to the ecological hypothesis of dentine proposed by Takahashi and Nyvad , as well as the tissue-dependent caries propagation hypothesis of Simon et al.  which states that while acid-producing bacteria are the prime movers of enamel penetration, the dentin degrading collagenolytic organisms which destroy deeper dentinal tissues are co-contributors in caries propagation.
Microbiota in caries lesions may act as reservoirs for other local and systemic infections
S-ECC, if not intervened and appropriately treated, is likely to have far-reaching effects, even extending to adulthood, and impact the general health of these children. In the concluding section, we briefly discuss the possible impact of our findings on the development of local or systemic disease.
Inquimbert et al.  and few others [86–88] have shown that early colonization of ECC lesions by periodontopathic species may be construed as a marker of periodontal disease risk later in life. In the investigation, they identified a number of periodontopathic organisms such as Campylobacter gracilis, S. noxia, and P. intermedia known to be associated with the initiation and progression of periodontal infection from ECC lesions [85–88].
We also noted several pathogens such as Campylobacter concisus, Capnocytophaga granulosa, Neisseria bacilliformis, and Granulicatella adiacens, implicated in the oral-systemic disease axis, amongst caries microbiota. The two former organisms (C. granulosa, N. bacilliformis) are implicated in abscess development and bacteremia secondary to focal infections [89, 90]. In addition, the oral C. concisus strains have been associated with human irritable bowel syndrome [91, 92]. Furthermore, members of the "HACEK" group bacteria, i.e., Haemophilus sp., Aggregatibacter sp., Cardiobacterium hominis, Eikenella corrodens, Kingella sp., and G. adiacens, a nutritionally variant streptococcus, all known to cause bacterial endocarditis [93, 94], were also prevalent in the dentinal caries samples. Thus, it is tempting to speculate that reservoirs of these microbes within cavitated lesions of S-ECC may act as potential reservoirs that may contribute to the foregoing systemic diseases in these children.
Limitations of the study
Our study has few limitations. The current data were derived from a relatively small sample of children and needed to be confirmed in a larger cohort, ideally from another geographic locale. Further, our report encompasses species-composition of the deep caries lesions in general but does not appertain or relate to a specific depth of the lesion. Therefore, future studies are required to evaluate the microbiome composition in dentinal cavities of varying depths to clearly understand the natural history of S-ECC and the compositional variations of the microbiota during lesion progression towards the pulpal axis.