Shotgun metagenomic sequencing reveals skin microbial variability from different facial sites

Abstract

Biogeography (body site) is known to be one of the main factors influencing the composition of the skin microbial community. However, site-associated microbial variability at a fine-scale level was not well-characterized since there was a lack of high-resolution recognition of facial microbiota across kingdoms by shotgun metagenomic sequencing. To investigate the explicit microbial variance in the human face, 822 shotgun metagenomic sequencing data from Han Chinese recently published by our group, in combination with 97 North American samples from NIH Human Microbiome Project (HMP), were reassessed. Metagenomic profiling of bacteria, fungi, and bacteriophages, as well as enriched function modules from three facial sites (forehead, cheek, and the back of the nose), was analyzed. The results revealed that skin microbial features were more alike in the forehead and cheek while varied from the back of the nose in terms of taxonomy and functionality. Analysis based on biogeographic theories suggested that neutral drift with niche selection from the host could possibly give rise to the variations. Of note, the abundance of porphyrin-producing species, i.e., Cutibacterium acnes, Cutibacterium avidum, Cutibacterium granulosum, and Cutibacterium namnetense, was all the highest in the back of the nose compared with the forehead/cheek, which was consistent with the highest porphyrin level on the nose in our population. Sequentially, the site-associated microbiome variance was confirmed in American populations; however, it was not entirely consistent. Furthermore, our data revealed correlation patterns between Propionibacterium acnes bacteriophages with genus Cutibacterium at different facial sites in both populations; however, C. acnes exhibited a distinct correlation with P. acnes bacteriophages in Americans/Chinese. Taken together, in this study, we explored the fine-scale facial site-associated changes in the skin microbiome and provided insight into the ecological processes underlying facial microbial variations.

Keywords: Chinese; Cutibacterium acnes (C. acnes); Propionibacterium acnes bacteriophage; biogeography; facial skin microbiome; fine-scale; shotgun metagenomic sequencing.

FIGURE 1

Comparative characterization of the skin microbiome present in three anatomical sites in Chinese samples. Boxplots comparing Shannon index of (A) overall skin microbiome and (B) bacteria, fungi, and virus microbiome of three anatomical sites in Chinese samples. (C) Principal coordinate analysis (PCoA) plot illustrating the comparison of the overall composition of skin microbiome between three sites in the Chinese population. The PERMANOVA test is used to determine significance. Boxplots indicate the distribution of samples along the PC1 and PC2. (D) Stack plot of the 17 differential species ranked by relative abundance. Linear discriminant analysis (LDA) effect size (LEfSe) histograms on the right showed the microbial comparisons of three anatomical sites, with an LDA threshold of 3.0. (E) Heat map of the Spearman’s correlation between differential species. (F) The distance distribution lines on the left showed the Bray-Curtis distance from the center of three facial sites to the forehead, to the cheek, and to the back of the nose, respectively. The boxplot on the right quantified and compared the difference. FH, the forehead; CK, the cheek; NS, the back of the nose. P-values were adjusted using the false discovery rate (FDR) correction. The significance levels are: ns, not significant, P > 0.05; *P < 0.05; ^**P < 0.01; ^***P < 0.001.

FIGURE 2

Comparative characterization of the skin microbiome present in three anatomical sites in American samples. (A) PCoA plot illustrating the comparison of the overall composition of skin microbiome between three sites in the Chinese population. The PERMANOVA test is used to determine significance. (B) Functional differences of the skin microbiome in three sites of Chinese samples. The heat map represents the relative abundance of function in three sites. (C) The heat map represents the relative abundance of 554 differential genes in three sites in Chinese samples. (D) Different categories that 554 differential genes belong to. (E) Boxplots comparing seven differential genes related to the porphyrin metabolism pathway in three anatomical sites (K02230: cobN, cobaltochelatase CobN; K00231: PPOX, protoporphyrinogen/coproporphyrinogen III oxidase; K02259: COX15, heme a synthase; K02224: cobB-cbiA; cobyrinic acid a,c-diamide synthase; K01698: hemB, porphobilinogen synthase; K13542: cobA-hemD, uroporphyrinogen III methyltransferase/synthase; K01885: EARS, glutamyl-tRNA synthetase). The relative abundance of facial porphyrin (F) and Cutibacterium acnes, Cutibacterium avidum, Cutibacterium granulosum, and Cutibacterium namnetense (G) in three sites. FH, the forehead; CK, the cheek; NS, the back of the nose. The significance levels are: ns, not significant, P > 0.05; *P < 0.05; ^**P < 0.01; ^***P < 0.001.

FIGURE 3

Comparative characterization of the skin microbiome present in three anatomical sites in American samples. (A) PCoA plot illustrating the comparison of the overall composition of skin microbiome between three sites in the American population. The PERMANOVA test is used to determine significance. Boxplots indicate the distribution of samples along the PC1 and PC2. (B) LEfSe histograms for the microbial comparisons of three anatomical sites, with an LDA threshold of 3.0. (C) PCoA plot illustrating the gene composition of skin microbiome between three sites in the American population. The PERMANOVA test is used to determine significance. (D) Functional differences of the skin microbiome in three sites of American samples. The heat map represents the relative abundance of function in three sites. (E) The heat map represents the relative abundance of 131 differential genes in three sites. FH, the forehead; CK, the cheek; NS, the back of the nose. P-values were adjusted using the FDR correction. The significance levels are: ns, not significant, P > 0.05; *P < 0.05; ^**P < 0.01; ^***P < 0.001.

FIGURE 4

The correlation between Cutibacterium and Propionibacterium acnes bacteriophages in Chinese and American samples. Heat map of the results of the Spearman correlation between four species in genus Cutibacterium and P. acnes bacteriophages in three sites and in Chinese (A) and American (B) samples. FH, the forehead; CK, the cheek; NS, the back of the nose. P-values were adjusted using the FDR correction. The significance levels in the Spearman correlation are: *P < 0.05; ^**P < 0.01; ^****P < 0.001.