*3.3. Di*ff*erent Genera and Species between BS and Control Group*

Heatmap analysis was employed to see overall patterns of microbial abundance at different levels of taxonomy. At all levels from species to order, the control group was isolated from the BS group (Figure 2), suggesting that the microbial composition was quite different between the two groups. The only exception was the sample "C12", which formed its own lineage within the BS group. Since the diversity pattern of C12 was much different from the other samples in the control group, we excluded the sample from further statistical analysis. To test the hypothesis that changes in microbial diversity affected BS formation, the abundance (read number) of each OTU was compared between the groups using the Mann–Whitney U test. At the genus level, six genera—*Abiotrophia, Eikenella, Granulicatella, Neisseria, Porphyromonas* and *Streptococcus*—were significantly different between the two groups (*p* < 0.05). *Neisseria*, *Porphyromonas* and *Streptococcus* contained 11, 14 and 31 species within them, respectively (Figure 3). All six genera were more abundant in the BS group than the control group. In addition, the genus *Selenomonas* exhibited the opposite trend and was more abundant in the control group (*p* = 0.053). At the species level, 19 species were identified as significantly different between the two groups (*p* < 0.05, Table 1). The number of each sample represented the identified OTUs. Except for *Selenomonas noxia, Corynebacterium\_uc* and *JQ454562\_s* (*Prevotella*)*,* all the species were higher in the BS group than the control group. Twelve out of 19 species were included in the genus showing significance.

**Figure 2.** Heatmap of the oral microbiome relative abundance. The species having more than 5% of the total reads are shown. The Bray–Curtis method was used for calculating cluster dissimilarity in the clustering analysis (*p* = 0.05, significantly different). B: black stain group, C: control group.

**Figure 3.** Abundance of the predominant bacterial groups between the black staining and control groups. Genus *Neisseria* (**A**), genus *Porphyromonas* (**B**), *Streptococcus* (**C**). Asterisks indicate species significantly different between the two groups (*p* < 0.05).



## *3.4. Trend Analysis among the Three BS Scale Groups*

In the heatmap at the species level, clusters based on microbial abundance were correlated to BS scale severity (Figure 2). For example, the samples "B1", "B2" and "B9", which had been determined as having "BS scale 2" severity according to Shori's method, were clustered together in similarity (Figure 2). A similar correlation was observed at all taxonomical levels, suggesting that BS scale severity may be correlated with the bacterial composition (Figure 2). Therefore, a dose-dependent relationship between the BS scale and abundance was examined using a nonparametric Jonckheere's trend test in the "clinfun" R package. A total of 47 species showed a trend among the BS scale clusters: 39 species exhibited an "increasing" trend when the BS scale increased (*p* < 0.05), whereas eight species were found to exhibit a "decreasing" tendency (*p* < 0.05). Principal component analysis (PCA) for these selected species revealed that species showing an "increasing" trend contributed to an increase in BS scale severity (Figure 4). This microbial community may play an essential role in the formation of BS. The most abundant genera were *Neisseria*, *Porphyromonas*, *Streptococcus* and *Capnocytophaga*. In contrast, the "decreasing" population, which showed abundance in the control group, consisted of *Corynebacterium*, *Prevotella*, *Selenomonas*, *Capnocytophaga* and *Leptotrichia*. Four genera were found in both groups: *Capnocytophaga*, *Corynebacterium*, *Leptotrichia* and *Prevotella*.

**Figure 4.** Principle component analysis of the selected species showing an increasing or decreasing trend according to the severity of the black stain (BS) scale. Approximately 71% of the total variance among the individual samples was explained by the first two components.

#### **4. Discussion**

Conventional BS therapy has been focused on the mechanical removal of black pigments themselves because the exact mechanism of BS prevalence had yet to be identified [1]. In this study, we used pyrosequencing analysis of supragingival dental plaque of children with and without BS to understand the microbiological aspects of BS, which may lead to better prevention and treatment for BS. The genera *Porphyromonas*, *Streptococcus* and *Neisseria,* were abundant in the microbiome of children with BS, while the genera *Corynebacterium*, *Prevotella* and *Selenomonas\_g1* were abundant in children without BS in this study. This study suggests that dynamics in the oral microbiome may cause the formation of BS in children, and the microbial population in the above genera would play important roles in changing microbial dynamics.

The genus *Porphyromonas* was significantly abundant in the BS group (Figure 2). All the species belonging to this genus exhibited a similar pattern in abundance (Figure 3). Four *Porphylromonas* sp. showed a trend to increase as the BS scales increased (Figure 4). *Porphyromonas* is an obligate anaerobic bacterium that produces a black pigment called porphyrin [11]. This pigment can protect the bacterial cell from oxidative stress by reducing oxygen. *P. gingivalis* is one of the most well-known species that produces black pigments [12] and is implicated in the progression of periodontal disease [4]. However, this species was not one of the pathogens observed in this study. *Porphyromonas catoniae* is the most common species in this genus (Figure 3A). It is frequently found in the oral microbiota of healthy children [9] and is known to be a pioneer bacterium that is established in the mouths of babies [13]. In addition, it is a non-pigmented species [14]. However, it is possible that this species may play an opportunistic role in the formation of BS because this species exhibits activities of diverse cellular enzymes like trypsin, which is important in bacterial colonization [13]. The other *Porphyromonas* spp. may contribute to the accumulation of black pigments.

The genus *Streptococcus* was also observed abundantly in the BS group (Figures 2 and 3B). *S. cristatus*, *S. sanguinis*, *S. oralis*, *S. mitis* and *S. gordonii* are known as oral biofilm bacteria that colonize the tooth surface [15,16]. Since these species are normal tooth microbiome and do not produce black pigments, they may contribute to the formation of microbial beds on teeth and interact with other species producing black pigments. Although the genus *Streptococcus* makes up 10% of the total microbiome, *S. mutans* was detected in only one subject (BS2, five reads). This rare detection of *S. mutans* is consistent with a recent report of BS patients where *S. mutans* was not found. This may be due to the antagonistic effect of the predisposed *Streptococcus* spp. *S. sanguinis* and *S. gordonii* that are known to inhibit the growth of *S. mutans* by producing hydrogen peroxide [17]. A recent study using pyrosequencing suggests that the genus *Streptococcus* was most abundant but showed no difference between the groups [18]. Another pyrosequencing study reported that this genus was observed frequently in subjects having both caries and pigments, but not in subjects having only pigments [19]. Interestingly, a recent microbiome study showed that the population of *P. catoniae* was decreased in children with caries and increased in children with healthy oral status [9]. Since the *P. catoniae* population was found to be increased in our BS group, it may have a negative influence on caries development (Figure 3A).

*Veillonella dispar* and *V. rodentium* were identified as positive candidates for BS formation (Figure 4). *V. dispar* produces hydrogen sulfide (H2S) from L-cysteine [20]. The genus *Actinomyces* can produce H2S and has been known to be a strong candidate for BS formation. However, *Actinomyces gerencseriae* was detected in this study, and there is no report about the H2S production of this species.

Microbial composition is continuously affected by the environment of the oral cavity, and each population interacts with others in the same community. Therefore, a decrease in population density may result in the breakdown of one community, presenting the opportunity for a new population or community to thrive. We hypothesized that different microbial compositions would appear in different BS scales. The BS scale indicates different environmental factors affecting microbial composition. In that hypothesis, our trend analysis exhibited positive or negative candidates in BS formation. In Figure 4, some species decreased as BS scales increased, and the others were increased in proportion to the BS

scale. These species may play core roles in BS formation. However, their exact roles must be identified in further studies.

Recently, two pyrosequencing studies were performed to analyze the microbiota of BS [18,19]. One was performed for 25 subjects [18], while the other study consisted of 111 subjects [19]. Li et al. [18] reported that the genera of *Actinomyces*, *Tannerella*, *Treponema*, *Corynebacterium*, *Cardiobacterium* and *Hemophilus* were more abundant in children with BS. In another study, the genera *Leptotrichia* and *Fusobacterium* contributed to BS formation, while the genera *Streptococcus* and *Mogibacterium* were important in the formation of both caries and pigments. Interestingly, three studies, including ours, did not reveal a strong consistency between them. Different ethnic backgrounds and dietary styles may provide confounding effects, suggesting that a more careful sampling design is required for this kind of study.

The limitation of this study is the small sample size. Larger sample size is needed for safe experiments and the potential confounding factor that the pyrosequencing may present. Further research will be designed by reflecting this limitation. In addition, it is meaningful to compare the study with existing healthy subjects in place of the small control group. In the study of healthy children, Proteobacteria was abundant in the group without dental caries, and Firmicutes and Bacteroides were abundant in the group with dental caries [21]. A study of healthy elderly people found that Firmicutes and Veillonella were abundant at the phyla level [22,23]. This was markedly different from the BS group in this study, in which Neisseria, Porphyromonas and Streptococcus were significant.
