DNA sequencing of samples collected by swabbing the skin surface is the most common method used to evaluate the skin microbiome because it is simple, noninvasive, and has been shown to result in higher consistency than other sampling methods (Bjerre et al., 2019). However, traditional metagenomic and 16S rRNA gene sequencing do not differentiate DNA from viable and inviable bacteria. We thus implemented a method that allowed us to quantitatively determine the fraction of DNA in a sample that is specifically from intact bacteria. The ability to molecularly differentiate between viable and inviable cells would enable us to both validate our FISH results using an orthogonal method on untreated, living skin, and address the possibility that the low numbers of intact surface bacteria observed by FISH were due to biopsy sterilization. To this end, we utilized the cell-impermeable small molecule propidium monoazide (PMA), which binds irreversibly to double-stranded DNA upon photoactivation to inhibit PCR amplification (Figure 2A; Nocker et al., 2006). When PMA photoactivation is performed before the cell lysis step of DNA isolation, the genomic DNA inside viable bacteria is protected from PMA binding because PMA is cell-impermeable, while cell-free DNA or DNA within permeabilized bacteria becomes PMA-bound. To quantify bacterial DNA, we combined the use of PMA with droplet digital PCR (PMA-ddPCR). Calculating the ratio of ddPCR counts between samples without PMA and samples with PMA allowed us to generate a viability score for any given bacterial population. A similar approach has been used to assess the viability of bacteria from environmental waste water samples (Yang et al., 2017).

Figure 2 with 2 supplements with 2 supplements see all Download asset Open asset Propidium monoazide-droplet digital PCR (PMA-ddPCR) and viability scores for human skin and non-skin microbiomes. (A) Schematic of the PMA-ddPCR workflow. (B) Sampling scheme showing each skin site that was sampled. Colors indicate site type (sebaceous in blue, moist in green, dry in red). (C) PMA-ddPCR on skin and non-skin microbiome sites shows that the viability score of the skin microbiome is significantly lower than other microbiome sites. ****p≤0.0001 for Student’s t-test on pooled skin and non-skin samples. Four volunteers contributed skin and non-skin microbiome samples. Additional samples were collected from some individuals and represent biological replicates. N = 8 for glabella, N = 6 for retroarticular crease, N = 5 for lower back, hair shaft, nares, and dorsal forearm, N = 3 for antecubital fossa, tongue, saliva, and plaque, N = 2 for popliteal fossa, and N = 1 for human feces. Each human skin sample site consists of samples from four different individuals. Some volunteers were sampled multiple times on different days (at least 2 wk apart). For glabella, one volunteer was sampled four times, one volunteer was sampled two times, and two volunteers were sampled one time. For retroauricular crease, two volunteers were sampled two times, and two volunteers were sampled one time. For lower back, one volunteer was sampled two times and three volunteers were sampled one time. For hair shaft, all samples came from one volunteer. For antecubital fossa, three volunteers were sampled one time. For popliteal fossa, two volunteers were sampled one time. For nares, one volunteer was sampled two times and three volunteers were sampled one time. For dorsal forearm, one volunteer was sampled two times and three volunteers were sampled one time. Tongue, saliva, and plaque all represent one sample from three different individuals. For raw ddPCR counts, see Figure 2—figure supplement 2A and B. (D) PMA-ddPCR on follicle contents and forehead swabs from five individuals. Mean viability score for follicle contents is 0.15 and for forehead is 0.013. All error bars indicate standard deviation.

To assure that PMA-ddPCR would allow us to reliably gauge the fraction of viable cells in a population, we first validated that it generated the expected results using known ratios of either exponentially-growing or stationary-phase Escherichia coli cultures combined with heat-killed E. coli cells (Figure 2—figure supplement 1A). To determine whether PMA-ddPCR also works for skin-resident bacterial species, we also applied PMA-ddPCR to four of the most common skin microbiome bacteria, S. epidermidis, C. acnes, M. luteus, and C. striatum. We performed serial dilutions of each bacterial culture and in parallel performed both colony-forming units (CFU) plating and ddPCR with or without PMA. In every case, the amount of DNA in a sample (assessed by PMA-ddPCR) correlated well with the number of culturable bacteria in high and low bacterial abundance scenarios (Figure 2—figure supplement 1B). Together, these controls confirm that PMA-ddPCR is a reliable method for assessing the amount of DNA in a sample present within intact bacteria.

We next applied PMA-ddPCR to human skin microbiome samples by swabbing the skin of four healthy human volunteers at eight sites (glabella, retroauricular crease, lower back, hair shaft, antecubital fossa, popliteal fossa, nares, and dorsal forearm) (Figure 2B, Figure 2—figure supplement 2A and B). PMA-ddPCR revealed that the viability scores for these sites ranged between 0.02 and 0.12 (0 represents a fully nonviable population, 1.0 represents a fully viable population), indicating that the majority of bacterial DNA found on the skin surface are not associated with viable cells (Figure 2C). To investigate whether this was a skin-specific phenomenon, we tested several non-skin microbiome sites (tongue, saliva, plaque, and feces). We found that in all non-skin microbiome sites, the viability score was significantly higher than for the skin, ranging from 0.4 (saliva) to 0.87 (feces) (Figure 2C).

We next sought to address whether our PMA-ddPCR viability scores accurately represent of the number of viable bacteria on the skin surface. The low number of detectable bacteria on the skin surface (either by FISH or PMA-ddPCR) does not imply that the skin surface is sterile but rather that the majority of the bacterial DNA on the skin surface are not from these viable cells. Indeed, culturing bacteria directly from the skin is common (Byrd et al., 2018). To quantify the number of viable bacteria directly, we plated a small amount of each sample using the standard conditions for culturing skin microbes (5% sheep blood in tryptic soy agarose plates incubated both aerobically and anaerobically). To determine whether PMA-ddPCR or traditional ddPCR better represented the number of viable skin microbiome bacteria, we compared our results to a standard curve generated with known numbers of S. epidermidis. For each sample, the PMA-ddPCR quantification closely matched this standard curve while the samples lacking PMA showed no overlap (Figure 2—figure supplement 1C and D). The highest abundance skin microbiome species are readily culturable, such that quantifying the DNA from viable skin surface bacteria should be able to accurately predict CFUs upon plating. Quantifying the bacterial DNA in skin microbiome samples without the use of PMA resulted in DNA quantities that were, on average, 82 times higher than predicted by the standard curve, while the use of PMA brought this value down to just 1.3 (Figure 2—figure supplement 1E). Using ddPCR counts to predict CFU showed similar results, as ddPCR in the absence of PMA yielded values that predicted CFU counts 58.5 times greater than those measured, while PMA-ddPCR yielded values that predicted CFU counts that were on average only 1.28 times greater than the actual cultured CFU (Figure 2—figure supplement 1F). Since we expect a ratio near 1, these findings lend further support for our conclusion that PMA-ddPCR reflects the viable microbiome better than the approaches lacking PMA.

Our FISH results from biopsied skin suggested that viable bacteria may be protected below the skin surface. To test this hypothesis in untreated skin, we collected facial follicle contents and used PMA-ddPCR to compare the viability of facial follicle contents and skin surface samples (Figure 2D). The average viability for follicle contents was greater than tenfold higher for follicle contents than for the skin surface, confirming that cells in hair follicles are more viable than those on the skin surface. Together, our results from FISH on biopsies and PMA-ddPCR on skin swabs independently support the conclusion that the skin surface is populated by few viable bacterial cells, indicating that the surface of healthy, non-sterilized human skin is sparsely colonized. While these data indicate that bacterial DNA on the skin surface is predominantly not associated with viable bacterial cells, we note that our results do not suggest that there are no viable cells on the skin surface. Rather, our data indicate that the majority of the bacterial DNA on the skin surface are not within bacteria such that using PMA provides a much more accurate estimation of the viable skin microbiome.