Genomic characterization of theC. tuberculostearicumspecies complex, a ubiquitous member of the human skin microbiome

Abstract

ABSTRACTCorynebacteriumis a predominant genus in the skin microbiome, yet its genetic diversity on skin is incompletely characterized and lacks a comprehensive set of reference genomes. Our work aims to investigate the distribution ofCorynebacteriumspecies on the skin, as well as to expand the existing genome reference catalog to enable more complete characterization of skin metagenomes. We used V1-V3 16S rRNA gene sequencing data from 14 body sites of 23 healthy volunteers to characterizeCorynebacteriumdiversity and distribution across healthy human skin.Corynebacterium tuberculostearicumis the predominant species found on human skin and we identified two distinctC. tuberculostearicumribotypes (A & B) that can be distinguished by variation in the 16S rRNA V1-V3 sequence. One is distributed across all body sites and the other found primarily on the feet. We performed whole genome sequencing of 40C. tuberculostearicumisolates cultured from the skin of five healthy individuals across seven skin sites. We generated five closed genomes of diverseC. tuberculostearicumwhich revealed thatC. tuberculostearicumisolates are largely syntenic and carry a diversity of methylation patterns, plasmids and CRISPR/Cas systems. The pangenome ofC. tuberculostearicumis open with a core genome size of 1806 genes and a pangenome size of 5451 total genes. This expanded pangenome enabled the mapping of 24% moreC. tuberculostearicumreads from shotgun metagenomic datasets derived from skin body sites. Finally, while the genomes from this study all fall within aC. tuberculostearicumspecies complex, the ribotype B isolates may constitute a new species.IMPORTANCEAmplicon sequencing data combined with isolate whole genome sequencing has expanded our understanding ofCorynebacteriumon the skin. Human skin is characterized by a diverse collection ofCorynebacteriumspecies butC. tuberculostearicumpredominates many sites. Our work supports the emerging idea thatC. tuberculostearicumis a species complex encompassing several distinct species. We produced a collection of genomes that help define this complex including a potentially new species which we are callingC. halluxbased on a preference for sites on the feet, whole-genome average nucleotide identity, pangenomics and growth in skin-like media. This isolate collection and high-quality genome resource sets the stage for developing engineered strains for both basic and translational clinical studies.Microbiomes are shaped by taxa that are both characteristic to those sites and functionally important to that community. The genusCorynebacteriumis one such taxa for the human skin and nares. Foundational studies using 16S rRNA gene sequencing and shotgun metagenomics by our lab (1, 2) and others (3) have establishedCorynebacteriumas common members of the skin microbiome. WhileCorynebacteriumhave been positively correlated with the resolution of dysbiosis associated with eczema flares (4), the importance of the Corynebacterium spp. is less defined for skin disease severity in primary immune deficient patients (5, 6).Corynebacteriumspp. are predominant members of the human aerodigestive tract microbiome (nares, oral cavity and respiratory tract) (3) and participate in microbe-microbe interactions with members of nasal microbiome (7, 8).Corynebacteriumhave been shown to engage with the host immune system, specificallyC. accolens-promoted IL23-dependent inflammation in mice on a high-fat diet (9).C. bovisandC. mastiditishave been shown to predominate the microbiome of a ADAM10-deficient mouse model (10) as well as an ADAM17-deficient mouse model of eczema (11). Finally,C. tuberculostearicumhas been shown to induce inflammation in human epidermal keratinocyte cell cultures (12). These studies establishCorynebacteriumspp. as key members of the skin microbiome capable of both microbe-microbe and microbe-host interactions.A critical resource for understanding the biology ofCorynebacteriumon the skin is a robust collection of complete reference genomes, including isolates collected from a variety of individuals and body sites. Previously published genome collections from skin- or nares-resident species includeStaphylococcus epidermidis(13),Cutibacterium acnes(14) and the recent comparative analysis ofDolosigranulum pigrum(15). Of note, while emerging bioinformatic methods and pipelines are now being employed to extract nearly-complete genomes (MAGs) from metagenomic assemblies of skin samples (16), MAGs are not yet a substitute for genomes from cultured isolates to understand strain level or pangenomic diversity. In addition to functional prediction, comparative genomics is increasingly being used to augment conventional microbiological methods to define or redefine taxonomic boundaries (17, 18), as well as describe the full extent of diversity within these boundaries (19). A pangenome, which encompasses the complete set of genes present within a set of genome sequences, enables the characterization of gene-level heterogeneity within a taxonomic group. The pangenome is commonly subdivided into the ‘core’ genome, referring to genes present in all strains, and the ‘accessory’ or ‘dispensable’ genome, referring to those present in only one or some isolates. (The accessory pangenome can be further subdivided to reflect a wider range of gene uniqueness,e.g.singletons.) Thorough characterization of taxa is limited by the availability of representative and high-quality genome assemblies. Unfortunately, with the exceptions of clinically relevantCorynebacteriumspp. (e.g.,C. diphtheriae,C. striatumandC. pseudotuberculosis), the genus is inadequately sequenced, with 75% of species having fewer than six genomes. This includes common skin-associated species likeC. tuberculostearicumwith just five unique isolate genomes, only two of which are from skin.This work seeks first to characterize the distribution ofCorynebacteriumacross 14 skin sites from 23 healthy volunteers. The second goal of this work focuses on what we identify as the predominant skinCorynebacteriumspecies,C. tuberculostearicum. We have sequenced 23 distinctC. tuberculostearicumstrains (n=40 genomes before dereplication), a five-fold increase in the number of publicly available, unique genomes (n=5). In addition to short-read assemblies, we generated five complete genomes which, along with the type strain (DSM44922), demonstrate thatC. tuberculostearicumgenomes are largely syntenic and carry a number of methylation systems as well as a CRISPR/Cas system. Genes from theC. tuberculostearicumgenomes in our collection fall into 5451 gene clusters comprising the species pangenome. This expanded pangenome, as compared to existing public references, improved the mapping ofC. tuberculostearicummetagenomic reads from unrelated healthy volunteers. In addition, we have identified a distinctC. tuberculostearicumclade that is highly enriched on the feet that may represent a new species, tentatively designatedCorynebacterium hallux.