Molecular analysis of the prevalent microbiota of human male and female forehead skin compared to forearm skin and the influence of make-up

Authors


  • The sequence data from this study have been submitted to GenBank under accession no. HQ16212HQ16344.

Bernhard Redl, Division of Molecular Biology, Biocenter, Innsbruck Medical University, A- Fritz Pregl Str. 3, 6020 Innsbruck, Austria. E-mail: bernhard.redl@i-med.ac.at

Abstract

Aims:  To compare the bacterial diversity of two different ecological regions including human forehead, human forearm and to estimate the influence of make-up.

Methods and Results:  Twenty-two swab-scraped skin samples were analysed by profiling bacterial 16S rRNA genes using PCR-based sequencing of randomly selected clones. Of the 1056 clones analysed, 67 genera and 133 species-level operational taxonomic units (SLOTUs) belonging to eight phyla were identified. A core set of bacterial taxa was found in all samples, including Actinobacteria, Firmicutes, and Proteobacteria, but pronounced intra- and interpersonal variation in bacterial community composition was observed. Only 4·48% of the genera and 1·50% of the SLOTUs were found in all 11 subjects. In contrast to the highly diverse microbiota of the forearm skin, the forehead skin microbiota represented a small-scale ecosystem with a few genera found in all individuals. The use of make-up, including foundation and powder, significantly enlarged the community diversity on the forehead skin.

Conclusions:  Our study confirmed the presence of a highly diverse microbiota of the human skin as described recently. In contrast to forearm skin, gender does not seem to have much influence on the microbial community of the forehead skin. However, the use of make-up was associated with a remarkable increase in the bacterial diversity.

Significance and Impact of the Study:  This study enhances our knowledge about the highly complex microbiota of the human skin and demonstrates for the first time the significant effect of make-up on the bacterial diversity of the forehead skin.

Introduction

The human skin has a number of essential protective functions in ensuring homeostasis of the whole body as it forms a barrier against harmful effects of the environment (Nemes and Steinert 1999; Madison 2003; Elias and Choi 2005; Proksch et al. 2008). It is colonized by a large number of micro-organisms representing a complex ecosystem. Originally, Price (Price 1938) classified the microbiota of human skin into two groups. Those organisms that grow on the surface and in the stratum corneum within the outmost layers of the epidermis and in the follicles of the sebaceous glands were called resident flora. They form a relatively stable population (Roth and James 1988; Wilson 2005), able to proliferate on the skin surface and in the outer sweat glands (Montes and Wilborn 1969). A second group of micro-organisms was called transient flora. They originate from external sources, through the contact of humans with the environment. These micro-organisms do not attach firmly to the skin and their composition varies permanently. The resident bacterial flora plays an important physiological role in protecting the skin from being infected by external pathogenic micro-organisms (Wilson 2005). Any change in the skin microbial balance can lead to harmful effects (Meloni and Schito 1991; Brook 2000; Wilson 2005). Thus, research on the microbiota has a long tradition. The use of molecular approaches instead of culture-dependent assays has helped to uncover the great diversity of skin microbiota (Dekio et al. 2005; Gao et al. 2007; Fierer et al. 2008; Grice et al. 2008, 2009). These studies revealed a low level of interpersonal consensus with respect to bacterial community membership and structure, which is approximately equal to intrapersonal variation. Also, a highly dynamic microbiota composition was observed to fluctuate greatly over time and gender variations have been found (Gao et al. 2007; Fierer et al. 2008; Grice et al. 2008, 2009).

In this work, we have investigated the diversity of skin bacteria present on two different sites of the body, namely the forehead and the forearm. An important feature of the forehead skin is the presence of both sebaceous and eccrine glands in high densities (Wilson 2005). This results in enriched nutrient supplementation and in elevated levels of antimicrobial peptides. The forehead is described as a relatively oily and moist region and is one of the most acidic regions of the skin. Because of its exposed nature, the forehead is subject to large temperature variations. Furthermore, the use of cosmetics might influence the number and species of bacteria present on the forehead (Holland and Bojar 2002). Forearm skin is described as a relatively dry environment, with a few eccrine and apocrine glands, and only a low density of sebaceous glands (Wilson 2005). However, the skin of the forearm is in certain climate zones often covered with clothes, which might affect the local environment, and thus altering the growth conditions of the microbiota.

Materials and methods

Sample collection

Specimens from superficial skin were obtained from the central forehead and the volar left forearm of 11 healthy young subjects, five men and six women, with no history of dermatological disorders or other chronic medical disorders and with no current skin infections. Samples were collected between late winter and early spring from subject living in the area of Innsbruck, which has a classical alpine climate. The mean age of the subjects was 25·8 years (range 22–29 years). The skin characteristics of all subjects are presented in Table S1. None of the subjects had received any antibiotics for at least 1 month. Female subjects F2, F4 and F5 used make-up (foundation and powder) every day, whereas female subjects F1, F3 and F6 never used make-up but used skin cleansers and skin cosmetics like moisturizers. None of the male subjects used make-up. Subjects were instructed not to wash with anything for an 8-h interval prior to sampling. Samples were taken from 2 × 2 cm of skin, from the central area of the forehead and volar left (flexor) forearm, midway between wrist and elbow, by swabbing the skin for 1 min with a sterile cotton swab that had been soaked in sterile 0·15 mol l−1 NaCl with 0·1% Tween 20. Samples from forehead and forearm were taken at the same time. To minimize sample cross-contamination, a fresh pair of sterile gloves was used for each individual sampling process. Reagent controls consisted of sterile cotton swabs moistened with a solution of sterile 0·15 mol l−1 NaCl and 0·1% Tween 20 and placed directly in 1·5-ml microcentrifuge tubes. The study was approved by the Ethics Committee of the Medical University Innsbruck (AN3616/4·13).

DNA extraction from swabs

DNA was extracted using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). As Gram-positive bacteria are more resistant to lysis than Gram-negative bacteria, the manufacturer’s protocol for genomic DNA isolation from Gram-positive bacteria was followed with modifications. The cotton tip of each swab was broken off directly into a 1·5-ml tube to which 180 μl of lysis buffer has been added. The tubes were capped and shaken by hand horizontally for 30 s. The remaining steps were performed as suggested by the manufacturer.

16S rRNA gene amplification, cloning procedure and sequencing

PCR amplification was performed with primers specific for conserved bacterial 16S rRNA sequences. PCR primers F 8: 5′-AGAGTTTGATYMTGGCTCAG-3′ and R 1401: 5′-CGGTGTGTACAAGGCCCG-3′ amplified a bacterial 16S rRNA gene fragment from nucleotide positions 8–1401 (Escherichia coli numbering, GenBank J01859). These primers span over 8 (V1–V8) of the nine hypervariable regions of the bacterial 16S rRNA and were chosen because most sequences in databases are available for regions V1–V8 (Chakravorty et al. 2007). PCR amplification was performed using the following conditions (final concentrations): 1× GoTaq reaction buffer (Promega, Madison, WI, USA), 0·2 mmol l−1 dNTPs, 50 pmol of each primer and 2·5 U GoTaq Polymerase (Promega) and 10 ng of DNA in a total volume of 50 μl. Twenty-seven cycles of PCR amplification were performed. Each of them entailed denaturation at 95°C for 60 s, annealing at 48°C for 60 s and primer extension at 72°C for 60 s. PCR products were analysed on 1·2% agarose gels, stained with ethidium bromide. The PCR fragments were gel eluted using Wizard® SV Gel and PCR Clean-Up System (Promega). Ten nanograms of DNA was ligated with the pGEM-T vector (Promega) and transformed into Ecoli DH5α competent cells. The complete insert was cycle sequenced using M13 primers for amplification and T7 and SP6 primers for sequencing on an ABI PRISM 3100 Genetic Analyser (Applied Biosystems, Carlsbad, CA).

Database and phylogenetic analysis

Analysis of closest relatives was performed by comparison of the entire amplified rDNA sequence (without primer sequences) with sequences available in the Ribosomal Database Project (RDP) II (RDP Release 10·23) and GenBank databases, by using the standard nucleotide-nucleotide Blast program (Blastn-megablast; National Center for Biotechnology Information, Bethesda, MD). Forty-eight clones were analysed from each skin sample. Sequences were examined for chimerism using Bellerophon at Greengenes (DeSantis et al. 2006).

Statistical analysis

The total number of species-level operational taxonomic units (SLOTUs) (SLOTU cut-off was 98%) that may be present in the sampled human skin and its associated confidence intervals were calculated by using the nonparametric richness estimator Chao1 (EstimateS ver. 5.0.1.) (Hughes et al. 2001).

Principal coordinate analysis (PCoA) was performed within the UniFrac online suite of tools (http://bmf2.colorado.edu/unifrac/index.psp). For this purpose, a phylogenetic tree generated with mega ver. 4.0 (Tamura et al. 2007), along with environmental labels, was imported into UniFrac (Lozupone et al. 2006). To facilitate the visualization of sample dissimilarity and diversity, the first two orthogonal principal axes were obtained based on the sample dissimilarity and were plotted to show the distribution of samples in a two-dimensional space. The P test, also available on the UniFrac suite of tools, assessed trees for nonrandom distributions of lineages according to environment. All P values reported were also corrected for multiple comparisons (Martin 2002; Lozupone et al. 2006).

Results

Clone libraries, classification and distribution of clones

Samples from superficial skin of 11 healthy subjects were collected, and a total of 1056 clones were analysed. Initially, 25 clones containing putative chimeric sequences were removed from the phylogenetic analyses and were replaced by new ones from the same master plate and the same subject to obtain 48 clones for each skin sample.

These experiments revealed the presence of eight phyla, 67 genera and 133 SLOTUs, when using a SLOTU cut-off of 98% (Table S2). Of the 1056 clones investigated, 1051 clones (99·53%) shared a nt sequence homology of ≥98% with cultivated type strains and five clones (0·47%) showed only significant homology to noncultured bacteria. These include uncultured members of Acidobacteria and members of the TM7 phylum. No clones representing previously uncharacterized phylotypes were found.

Actinobacteria, Firmicutes, and Proteobacteria were the predominant phyla, representing 1027 (97·25%) clones and accounting for 36 (59·75% of the clones), 44 (25·09%) and 37 (12·41%) SLOTUs. These three phyla were observed in all 11 subjects (Fig. 1a,b). The phylum Bacteroidetes was observed in samples from eight subjects, and the phylum Fusobacteria in samples from three subjects. Acidobacteria were observed in samples from two subjects, and each of the phyla Deinococcus-Thermus and TM7 were observed in one subject (Fig. 1a,b).

Figure 1.

 Relative abundance of the most prevalent bacterial groups associated with each microenvironment depicted for each of the subjects investigated. (a) Forehead skin, (b) forearm skin. Superscripts indicate phylum: 1, Actinobacteria; 2, Firmicutes; 3, Proteobacteria; 4, Bacteroidetes; 5, Acidobacteria; 6, Deinococcus-Thermus; 7, Fusobacteria. F, female; M, male. (inline image) Propionibacteria spp.1; (inline image) Corynebacteria spp.1; (inline image) Other Actinobacteria1; (inline image) Bacillales2; (inline image) Lactobacillales2; (inline image) Clostridiales2; (inline image) γ-Proteobacteria3; (inline image) α-Proteobacteria3; (inline image) β-Proteobacteria3; (inline image) Bacteriodales4; (inline image) Flavobacteriales4; (inline image) Acidobacteriales5; (inline image) Deinococcales6; (inline image) Fusobacteriales7; (inline image) TM7 Phylum.

Three genera including Propionibacterium (46·78% of all clones), Staphylococcus (14·39%) and Corynebacterium (5·40%) were observed in all 11 subjects analysed. However, of the 67 genera, 30 genera (44·78%) were identified in only one subject, 11 (16·42%) in two subjects and 6 (8·96%) in three subjects; only 17 (25·37%) genera were observed in ≥4 subjects, indicating substantial interpersonal variation at the genus level.

A total of 133 SLOTUs were found, reflecting the microbiota of the forehead and of the volar left forearm skin (Table S2). The number of SLOTUs in each subject ranged from 16 to 42 (mean: 26·55 ± 7·57) (Fig. 2a,b). The most numerous species identified belonged to the genera Corynebacterium, Streptococcus and Staphylococcus. Propionibacterium acnes was the most prevalent species detected in all 11 subjects, accounting for 45·74% of all clones analysed.

Figure 2.

 Distribution of species-level operational taxonomic unit richness of bacterial communities present on the forehead (a) and forearm skin (b). Subjects F2, F4 and F5 were make-up users.

To estimate the SLOTU richness, the expected number of SLOTUs was calculated using the Chao1 nonparametric estimator. The results were visualized by collector’s curves of observed and estimated SLOTU richness (Fig. S1a,b), and the 95% confidential intervals were calculated to estimate the precision of the richness estimation (Hughes et al. 2001). We estimated that the microbiota of the forehead samples contains c. 65 SLOTUs (95% CI, range 50–107). Consequently, the 42 SLOTUs observed in this study represented 65·6% (95% CI, 46·7–68·0%) of the estimated species or phylotypes. On the forearm skin, we observed 128 SLOTUs in our study. However, the Chao1 score estimated that the microbiota of the forearm samples contains c. 170 SLOTUs (95% CI, range 149–211). Based on this prediction, the present study identified 75·3% (95% CI, 65·9–78·5%) of the SLOTUs in this bacterial ecosystem. This is in agreement with Gao et al. (Gao et al. 2007), who calculated c. 246 SLOTUs (95% CI, range 217–301) for this area. The Chao1 estimates for both sampling sites stabilized to the end of the collector’s curves indicating that the estimated richness might not change with further sampling (Fig. S1a,b).

Forehead skin microbiota and the effect of gender

For the forehead samples, we observed four phyla (Fig. 1a), with Actinobacteria being the most frequently isolated microbial phylum in both men and women, representing 75·38% of all clones of the forehead skin (Table S2). Subjects F4 (make-up using) and M3 contained an increased proportion of Firmicutes sequences compared with the other male and female subjects (Fig. 1a). Furthermore, the phyla Firmicutes and Proteobacteria were detectable in both sexes. The phylum Bacteroidetes could only be detected in female subjects (Fig. 1a). From the 30 genera observed on the forehead skin, Propionibacterium (73·11% of all clones) and Staphylococcus (15·91% of all clones) were most abundant. These two genera were detected in all 11 subjects and together they accounted for nearly 90% of all clones. These findings indicate a small ecosystem on the forehead skin (Fig. 1a), which was also evident on the species level (Fig. 2a). Propionibacterium acnes was the most predominant species on the forehead, accounting for 72·54% of all clones, but Staphylococcus epidermidis (12·31%) was also frequently isolated from the forehead (Table S3).

When comparing the samples from all subjects, the forehead skin of women seems to harbour larger bacterial diversity on the genus level (Fig. 1a). On this site, women harboured on average 7·5 genera per subject and men three genera per subject (Table S2). Also, on the species level women showed an increased bacterial diversity compared with male subjects (P ≤ 0·01) (Fig. 2a). However, when excluding subjects using make-up (F2, F4 and F5) in the comparison of the forehead microbiota of women and men, no significant differences (P = 0·26) in the bacterial diversity on the species level were detected, and also the numbers of detected SLOTUs were similar (Fig. 2a). Thus, gender does not seem to be a factor relevant for microbial diversity on the forehead of the subjects investigated.

Forearm skin microbiota A and the effect of gender

On the forearm skin, eight phyla were identified (Table S2 and Fig. 1b). Actinobacteria, Firmicutes, Proteobacteria, Bacteroidetes and Fusobacteria were found on both men and women. The phylum Deinococcus-Thermus was only detected in men, and an uncultured member of the phylum Acidobacteria was only present in women.

The most frequently isolated genera on the forearm were Propionibacterium (20·45%), Staphylococcus (12·88%) and Corynebacterium (10·04%). In total, only 14 genera of the 62 genera detected on the forearm were observed in ≥4 subjects, indicating substantial interpersonal diversity at the genus level. A comparison of the microbiota diversity of women and men showed clear differences on the genus level (mean 18·5 genera per subject vs 14·4 genera per subject). A similar result is also obtained when comparing the number of SLOTUs found (P ≤ 0·01) (Fig. 2b). On the female forearm skin, an average of 25·2 SLOTUs per subject were found whereas an average of 18·4 SLOTUs were found on male forearm skin. The most abundant species on the forearm was P. acnes, followed by Staph. epidermidis, Rhizobium tropici and Micrococcus luteus (Table S3).

Noteworthy, certain genera were exclusively found in make-up-using women (Table S2). Examples include Aerococcus (exclusively found in F5, both on forehead and forearm), Lautropia (forehead of F5; forearm of F4) and Gordonia (forehead of F2 and F5; forearm of F2).

Effect of make-up on the forehead skin microbiota

Female subjects F2, F4 and F5, who used make-up every day, showed significantly higher microbial diversity (P = 0·04) on their forehead skin than other female or male subjects (P = 0·27) (Fig. 1a). These subjects harboured an average of 10·33 genera/subject, compared to 5·66 genera per subject in the female subjects without make-up and to 3·2 genera per subject in male subjects (Table S2). Higher microbial diversity of the make-up users is also evident from the number of SLOTUs found with an average of 13·33 SLOTUs per subject present on the forehead skin of make-up users compared to six SLOTUs per subject in the female subjects without make-up and to 4·2 SLOTUs per subject in male subjects (Fig. 2a). Furthermore, the genera Selenomonas (in F5), Aggregatibacter (in F5) and Aquicella (in F5) were exclusively found in make-up-using women.

Thus, the use of make-up appears to be the most important factor for causing differences in the bacterial diversity on the forehead skin of the subjects investigated. Skin nature does not appear to be a highly relevant factor, because only female F4 had a more oily skin, but all other women had normal or dry forehead skin (Table S1).

Comparison of forehead and forearm microbiota

The comparison of the microbiota of forehead and forearm skin clearly showed that the bacterial diversity of the forearm was considerably greater than that of the forehead (Fig. 1a,b). Even when including the data from the make-up users, only 30 genera were observed on the forehead compared to 62 genera on the forearm (Table S2). Also on the species level the microbiota of the forearm showed a much higher biodiversity than that of the forehead. This is manifested by a mean of 23 SLOTUs per subject compared to a mean of 7·18 SLOTUs per subject on forehead skin (Fig. 2a,b). However, the bacterial communities from both sites showed some similarity in qualitative terms. Of the 42 SLOTUs found on the forehead skin only five SLOTUs, including Acinetobacter haemolyticus, Aggregatibacter segnis, Aquicella siphonis, Paracoccus sp., Selenomonas noxia were not present in the forearm samples (Table S2). Nevertheless, the overall microbiota differed substantially amongst the 11 subjects analysed (Fig. 1a,b). Propionibacterium acnes was the only species present in all subjects on the forehead, and only four of the 42 SLOTUs were observed in more than ≥4 subjects (Table S2). Conversely, 66·67% of all of the individual SLOTUs were identified only in a single individual. The microbiota of the forearm revealed similar results. Again P. acnes was the only SLOTU present in all 11 subjects (Table S2), 14 SLOTUs of the 128 SLOTUs were observed in more than ≥4 subjects and 54·69% of all of the individual SLOTUs were identified in a single subject (Table S2).

This study also revealed that Gram-positive bacteria were much more abundant on the skin areas investigated (Table S5). About 95% of the bacteria found on forehead areas of both sexes were Gram-positive species. Gram-positive bacteria were also predominant on the forearm areas (78·75 and 71·88%). The forearm areas of women showed slightly higher numbers of Gram-negative species than male forearms (27·08% compared to 21·05%).

To evaluate sample diversity and the relationship amongst samples PCoA was performed. This analysis suggested that the forehead skin samples from male subjects were more similar to each other than they were to forearm skin samples (Fig. 3). An exception was M3, probable because of the high numbers of Staph. epidermidis in this sample. Additionally, the forehead samples of subject F1, F3 and F6 (women not using make-up) clustered closely to the forehead skin samples of the male subjects. The forehead samples of the make-up-using women (F2, F4 and F5) were more distant. Clustering of the forearm samples of the 11 subjects revealed that they were not as closely related to each other as the forehead samples, clearly indicating greater diversity amongst the forearm samples (Fig. 3).

Figure 3.

 Principal coordinate analysis of species-level operational taxonomic unit relatedness of 22 skin samples from 11 subjects. The axes are labelled with the per cent of the variation explained by each principal component. The circle represents all forehead skin samples, which cluster closely together. F, female; M, male; f, face; a, arm.

To determine whether individuals share a common skin microbial diversity profile, the P test was used to assess intrapersonal and interpersonal variation (Table S4). Analysis of each pair of samples showed that no significance was found in skin samples from the same subject when both sites were considered. Furthermore, this analysis showed that if significance appeared between two samples from different subjects, it did not follow any trend.

Discussion

Analysis of 1056 clones isolated from the superficial forehead and forearm skin revealed eight phyla, 67 genera and 133 SLOTUs. Of these, the forehead microbiota harboured four phyla, 30 genera and 42 SLOTUs, with an estimated species coverage of c. 65·6%, and the forearm microbiota harboured eight phyla, 62 genera and 128 SLOTUs, with estimated species coverage of c. 75·3%.

The most abundant phyla on the forehead as well as on the forearm were Actinobacteria, Firmicutes and Proteobacteria. These three phyla have also been reported to be predominant on other skin areas (Dekio et al. 2005; Gao et al. 2007; Fierer et al. 2008; Grice et al. 2008, 2009). The phylum Bacteroidetes could only be harboured on the forehead of female subjects. We cannot rule out the possibility, however, that increased sampling would show the presence of this phylum also in male subjects. On the forearm skin, the phylum Deinococcus-Thermus was only detected in men, and an uncultured member of the phylum Acidobacteria was only present in women. Again, we cannot rule out the possibility that with increased sampling these two latter phyla might be found in both genders.

Although there appears to be a core set of phylotypes present on superficial human skin, the overall microbiota differed substantially amongst the 11 subjects. Three genera (Propionibacteria, Staphylococcus, and Corynebacteria) were commonly found, but only 1·5 % of the SLOTUs and 4·48% of the genera were found in all 11 subjects. In contrast, 50·38% of all of the individual SLOTUs and 46·27% of the genera were identified only in a single individual. These data indicate that the superficial human skin microbiota is highly diversified with a low level of interpersonal consensus. Similar findings were reported in recent studies (Gao et al. 2007; Fierer et al. 2008; Grice et al. 2008, 2009). A pairwise comparison of all subjects revealed that significance did not follow any trend, suggesting that occurrence of significance was more random than directed amongst different subjects. The low interpersonal consensus is also evident when comparing our results to the study of Gao et al. (2007), because of the 67 genera found in our study only 36 genera were also found in the previous study.

The most frequently isolated organism on the forehead as well as on the forearm was P. acnes, which was the only species present in all skin samples of the ten subjects. Moreover, the forehead microbiota was dominated by P. acnes accounting for 72·54% of all clones. In contrast, P. acnes represented only 18·94% of the clones isolated from the forearm skin. Thus, we conclude that a rich population of Propionibacteria is a stable characteristic of the skin of the forehead. So far, no other exposed microenvironment of the human body has been found to be similarly dominated by a single bacterial species (Evans 1975). Propionibacterium, and P. acnes in particular, is the most dominant organism in normal human skin, as shown in prior culture-based studies and in molecular analyses (Fredricks 2001; Gao et al. 2007, 2008). Although it has been suggested that P. acnes is associated with serious skin disorders, such as acne vulgaris and rosacea (Leyden 2001), it is likely that it might protect the skin against more aggressive microbes (Eady and Ingham 1995; Leyden 2001; Bek-Thomsen et al. 2008). It is not yet clear what factors dictate the level of the Propionibacteria population, but it seems probable that the critical substances will be found in sebum (Evans 1975).

It is interesting to note that subject F4 appears to be a major outlier with respect to community composition. On the forehead skin of subject F4, P. acnes was found to be underrepresented. Instead, an increased proportion of Firmicutes sequences was detected. These sequences appear to be derived from Staphylococcus species, suggesting that subject F4 was colonized with Staphylococcus, as found by cultivation studies in c. 5–10% of healthy adults (Nagase et al. 2002). As subject F4 was using make-up, this different bacterial composition might be attributed to an influence of make-up.

Of the ten most common species on the forearm skin found by Gao et al. (2007), six (including P. acnes, Corynebacterium tuberculostearicum, Corynebacterium mucifaciens, Streptococcus mitis, Staph. epidermidis and Corynebacterium amycolatum) were also found in our study. However, in our study only P. acnes, Staph. epidermidis and Strep. mitis were present in high numbers. Acinetobacter junii and Deinococcus AJ549111 were not found in our study, but we detected other closely related Acinetobacter and Deinococcus species (Table S2). Enhydrobacter aerosaccus, identified in a study by Gao et al. (Gao et al. 2007), was not detected in subjects of the present study.

We also detected some potential pathogens on the forehead as well as on the forearm skin, such as Staphylococcus aureus, Neisseria meningitis, Streptococcus pneumonia and Haemophilus influenzae. These organisms are considered as transient colonizers of the human skin producing various kinds of skin-damaging toxins and are suspected to play a role in serious skin diseases (Roth and James 1988; Wilson 2005). They are ubiquitously present in the environment, e.g. in soil, tap water and crops. Therefore, these species are able to cope with harsh environmental conditions. Consequently, they are well equipped to survive the rapidly changing conditions of the surface of the skin and it is possible that the skin microbiota may be a reservoir of such opportunistic pathogens derived from the environment (Dekio et al. 2005).

It was reported previously that Gram-positive bacteria are more abundant on superficial human skin than Gram-negative bacteria (Roth and James 1988; Wilson 2005; Gao et al. 2007). This was confirmed by our study. On the other hand, from culture-based studies it has also been suggested that the skin of male individuals, being more humid because of enhanced sweat production, is more frequently colonized by Gram-negative bacteria than that of women, as high moisture is necessary for survival of these bacteria (Roth and James 1988). This observation could not be confirmed by our study. We found that the skin of women was more frequently colonized by Gram-negative species than that of men, which is in accordance with previous molecular genetic analysis (Gao et al. 2007). As Gram-negatives were more abundant on the forearm, it might be suggested that occlusion of the skin by dressing could have a profound effect on the micro-environment of the forearm skin area by reducing the evaporate water loss. Thus, increasing the water content of the occluded region may lead to a higher humidity suitable for the growth of Gram-negative bacteria (Wilson 2005).

In contrast, the cutaneous microbiota of the forehead is always exposed to the external environment and therefore subject to UV, temperature and enhanced mechanical stress. Probably Gram-negative bacteria are more sensitive to these conditions because of their cell wall structure (Roth and James 1988; Wilson 2005). This might explain our finding of their less abundant presence on the forehead skin.

Analysis of the distribution of SLOTUs between forehead and forearm revealed that only five of the 42 SLOTUs found in the forehead samples were not shared with the forearm samples, but quantitatively there was more discrepancy between the two regions. The forehead skin harboured only a few species (mean: 7·18 SLOTUs per subject). Whereas the forearm microbiota showed a more complex community structure (mean: 23 SLOTUs per subject), suggesting that the forearm skin is a highly diverse ecosystem. This difference in bacterial diversity between the forehead samples and the forearm samples was statistically significant (= 0·03). Together these results demonstrate that the phylotype richness of the forearm was higher than that of the forehead skin area. These results were surprising, as the forehead was expected to have higher levels of diversity than the forearm because of the more frequent contact with potential invaders from the environment. On the other hand, a higher concentration of antimicrobial substances on the forehead could inhibit the growth of some species and the dominant presence of P. acnes could protect this site from other species by competing for nutrients, niches and receptors (Gao et al. 2008).

Comparing all forehead samples from men and women revealed significant diversity amongst them. However, when samples from make-up-using women (F2, F4 and F5) were excluded, these differences between male and female samples were not longer observed. Thus, the use of make-up appears to strongly interfere with the composition of the bacterial communities. It is important to note, however, that further experiments using higher numbers of samples are needed to substantiate this conclusion. For instance, the observed differences might also be a consequence of a bacterial contamination of the make-up. In fact, especially the presence of Aquicella, which is usually found in thermal water, strongly indicates a contamination of make-up. Use of skin cleansers and skin moisturizers does not appear to influence the skin in a way as make-up does. Female subjects F1, F3 and F6 in our study never used make-up but used skin cleansers and skin cosmetics like moisturizers in their daily life. These subjects did not show increased numbers of phylotypes on the forehead skin.

Acknowledgements

The research was partially funded by a COMET project from the Austrian Research Promotion Agency (FFG). The authors thank Rajam Csordas and Alexandra Lusser for critical reading and correction of the manuscript.

Ancillary