Human skin microbiota: high diversity of DNA viruses identified on the human skin by high throughput sequencing

Abstract

The human skin is a complex ecosystem that hosts a heterogeneous flora. Until recently, the diversity of the cutaneous microbiota was mainly investigated for bacteria through culture based assays subsequently confirmed by molecular techniques. There are now many evidences that viruses represent a significant part of the cutaneous flora as demonstrated by the asymptomatic carriage of beta and gamma-human papillomaviruses on the healthy skin. Furthermore, it has been recently suggested that some representatives of the Polyomavirus genus might share a similar feature. In the present study, the cutaneous virome of the surface of the normal-appearing skin from five healthy individuals and one patient with Merkel cell carcinoma was investigated through a high throughput metagenomic sequencing approach in an attempt to provide a thorough description of the cutaneous flora, with a particular focus on its viral component. The results emphasize the high diversity of the viral cutaneous flora with multiple polyomaviruses, papillomaviruses and circoviruses being detected on normal-appearing skin. Moreover, this approach resulted in the identification of new Papillomavirus and Circovirus genomes and confirmed a very low level of genetic diversity within human polyomavirus species. Although viruses are generally considered as pathogen agents, our findings support the existence of a complex viral flora present at the surface of healthy-appearing human skin in various individuals. The dynamics and anatomical variations of this skin virome and its variations according to pathological conditions remain to be further studied. The potential involvement of these viruses, alone or in combination, in skin proliferative disorders and oncogenesis is another crucial issue to be elucidated.

Figure 1. Cutaneous virome.

Relative abundance of reads from beta, gamma and unclassified papillomavirus, from circoviruses and other single strand DNA viruses and from the respective polyomaviruses (MCPyV, HPyV6, 7 and 9) within the whole eucaryotic DNA viral reads detected on the skin surface of each sample.

Figure 2. MCPyV genome sequence coverage versus MCPyV viral load in the six samples.

Genome coverage is expressed as number of reads per nucleotide and viral load is in genome copies per ng of whole DNA. Coefficient of correlation is: R² = 0.894; and best fit regression equation is: [MCPyV viral load]  = 4.1 [MCPyV coverage]- 28,850.

Figure 3. Phylogenic tree of L1 nucleotide sequences.

Bayesian phylogeny (using the Tamura-Nei model TN93 with gamma distribution and invariant sites) of the main alpha, beta, gamma, mu and nu papillomaviruses infecting human inferred from their L1 nucleotide sequences (the list of the taxa included all sequences available from the PAVE database and are available at: http://pave.niaid.nih.gov/#prototypes?type=human). The new sequences (red branches) and species reported here belongs to the Gammapapillomavirus genus only, therefore, Alphapapillomavirus and Betapapillomavirus subtrees have been collapsed for clarity reasons. The tree is rooted by a bird Etapapillomavirus: the Fringilla coelebs Papillomavirus (FCPV). The added animal papillomaviruses to the phylogeny are the Roussetus aegyptiacus papillomavirus type 1 (RAP), the Bovine papillomavirus 8 (BPV8) and the Ovine papillomavirus (OPV) type 3. Posterior probabilities are reported for each node and sequences obtained during this study are depicted by grey/red branches. Each Human papillomavirus is noted HPV followed by the number of the species and by its Genbank accession number.