Staphylococcus aureus enters hair follicles using triacylglycerol lipases preserved through the genus Staphylococcus

Letter

To the editor,

Bacteria belonging to the genus Staphylococcus are a group of organisms that readily inhabit human skin and the upper respiratory tract. Staphylococcus aureus (S. aureus) is a leading cause of soft tissue infections and bacteraemia (Miller and Cho, 2011), but is closely related to many other species of Staphylococcus that are skin commensals and rarely cause disease. These species densely populate follicular structures that may serve as a reservoir for long term colonization (Nakatsuji, Chiang et al., 2013). Furthermore, there is growing evidence that various other microorganisms live in hair follicles (HFs) and interact with skin immune cells (Chen, Fischbach et al., 2018). For example, Cutibacterium acnes produces short chain fatty acids by fermenting lipids in sebum, and these substances have influence on host immune function (Sanford, O’Neill et al., 2019). Thus, it is important to further understand the mechanisms by which both pathogenic and commensal microbes establish colonization of the HF.

In this study, we sought to understand how Staphylococcus can enter the HF. Since sebum is secreted from sebaceous glands and fills in the infundibulum of HFs, we hypothesized that sebum functions as a hydrophobic barrier while lipases may enable bacteria to enter the HF independently of penetration of the stratum corneum. We focused on triacylglycerol lipases that degrade triglycerides, which are the main components (60%) of sebum(Picardo, Ottaviani et al., 2009), and studied the response to deletion mutants of the S. aureus known lipases.

S. aureus produces 2 secreted lipases termed gehA (SAL-1) and gehB (SAL-2) and the expression of these genes are under the control of the accessory gene regulator (agr) quorum sensing system (Horswill and Gordon, 2020). We compared the capacity of a methicillin-resistant S. aureus (MRSA) USA300 LAC wild type (WT) parent strain, a double lipase mutant S. aureus strain (SA Δlipases), a S. aureus agr mutant (SA Δagr), and a mutant S. aureus strain of all 10 known secreted proteases (SA Δproteases) to induce skin damage and inflammation. Topically applied S. aureus WT induced inflammation, erythema and crust within 48 hours, Δlipases and Δproteases strains induced less inflammation, and the agr mutant was least capable of promoting inflammation (Fig. 1a). Measurement of interleukin-6 mRNA and measurement of transepidermal water loss (TEWL) of the skin revealed that murine Il6 mRNA in the whole skin was decreased by 56% with the SA Δlipases strain compared to the S. aureus WT strain control (Fig. 1b). TEWL was decreased by 47% in SA Δlipases compared to the S. aureus WT strain (Fig. 1c). Overall these data reveal how S. aureus lipases play a role in skin inflammation and damage, and that these likely cooperate with other previously discovered agr-regulated skin damaging factors such as S. aureus proteases (Kolar, Ibarra et al., 2013).

Figure 1.

S. aureus (SA) USA300 wild-type parent strain (SA WT) and mutants lacking lipases (SA Δlipases), proteases (SA Δproteases), or a functional agr quorum sensing system (SA Δagr) were applied at 1e⁶ CFU/cm² to murine back skin 24h after removal of hair by shaving. (a) Representative images of mice 48 hours after S. aureus application (dotted area). (b,c) qPCR assessment of murine Il6 mRNA normalized to the housekeeping gene Gapdh and measurement of skin transepidermal water loss (TEWL) after 48h colonization by S. aureus bacterial strains (n=4). (d) qPCR for universal 16S rRNA bacterial abundance in serial 20um horizontal sections of mouse skin 4 hours after application. Depth=estimated distance from the surface (n=3). (e) Representative images of immunofluorescent assessment of S. aureus (green) and nuclear DAPI staining (blue) of murine skin sections 2 hours after application of bacteria on the skin. Dotted lines show the relative area of the hair follicle (HF) infundibulum (scale bar = 100μm). (f) Proportion of S. aureus observed in the infundibulum compared to surface S. aureus (% penetration) for S. aureus WT and mutant strains (n= number of HFs observed in 4–5 mice per condition). (g) Similar assessment as in f of individual S. aureus lipases knockouts gehA and gehB. All results are representative of at least two independent experiments and standard error of the mean. (b,c) One-way ANOVA analysis was used for statistical analysis and significance is indicated by: p<0.05 *; p<0.01 **, p<0.001 ***; p<0.0001 ****.

Next to determine if S. aureus lipases allow for penetration of microbes into the HF, the S. aureus WT and mutant strains were applied topically to shaved back skin at 1×10⁶ CFU/cm² for 2 hours. 16S rRNA abundance was quantified in sequential 20 μm horizontal sections of frozen skin. This analysis revealed that the SA Δlipases strain as well as the SA Δagr and SA Δproteases strains all lacked the ability to penetrate past 100μm of the skin surface (Fig. 1d). This depth of penetration corresponds with the depth of the infundibulum that is typically within approximately 100 μm of the surface. Furthermore DNA for the S. aureus WT and SA Δproteases strains appeared greater within the 60–100um depth than the SA Δlipases and SA Δagr strains. This suggested that S. aureus lipases play a role in initial penetration of bacteria into the lipid-rich infundibulum. To confirm this, immunostaining of murine skin sections was done with an antibody specific to S. aureus and revealed that only S. aureus WT and SA Δproteases strains were frequently observed to infiltrate into the infundibulum, but mutants lacking lipases or an active agr system (lacking capacity to secrete lipases and proteases) could not (Fig. 1e). Furthermore, by counting individual staining within multiple follicles, and applying the agr and total protease mutant compared to specific S. aureus mutants to either lipase gehA or gehB, we quantified what fraction of total bacterial staining could be seen within the infundibulum. This analysis revealed that GehB was the primary driving initial HF penetration (Fig. 1f,g). Overall these data suggest that the specific S. aureus lipase, gehB, is essential for penetration into the upper HF of mice.

Lipases are widely expressed throughout the genus Staphylococcus. In case of S. aureus, SAL-1 (encoded by gehA) primarily reacts with short-chain glycerides, while SAL-2 (encoded by gehB) hydrolyzes both short- and long-chain triglycerides (Cadieux, Vijayakumaran et al., 2014). Two equivalent lipases are reported in Staphylococcus epidermidis: SEL-1 (encoded by gehC) and SEL-2 (encoded by gehD), respectively (Longshaw, Farrell et al., 2000). Other Staphylococcal species are known to secrete lipases, and some of their genomes are sequenced. To reveal the evolutionary differences of major lipases among the genus Staphylococcus, we performed database search and analyzed their similarity using NCBI BLAST (www.ncbi.nlm.nih.gov/BLAST) followed by the analysis conducted in MEGA X software (www.megasoftware.net). As expected, our data suggested that lipases are highly preserved among those species, and all the staphylococcal lipases are produced as pre-pro-enzymes with a signal peptide in the pre-region and are secreted as proenzymes needing a specific cleavage for mature configuration. Amino acid positions of the catalytic triad (serine-histidine-aspartate motif) are quite similar among the mature lipases (Fig. 2a). Generally, lipases are members of a large group of enzymes possessing the α/β hydrolase fold with the preserved arrangement of the catalytic residue (His-Ser-Asp or Cys) located in a tight loop after the β5 strand (Ollis, Cheah et al., 1992). In addition, the calculation indicated that based on the similarity they are categorized to two groups: one including SAL-1 and the other including SAL-2 (Fig. 2b), indicating that all Staphylococcal species encode two lipases. This high identity of lipases in the genus Staphylococcus implies important roles in metabolism and adaptation to the environment (Nguyen, Luqman et al., 2018), and suggests our observations in S. aureus may apply to other species of Staphylococcus in the HF.

Figure 2.

Protein sequences of lipases were compared among the genus Staphylococcus. (a) Protein sequences similar to the two major lipases encoded by S. aureus USA300 (SAL-1 and SAL-2). All sequences had three portions (SP, PP, and ML) and a catalytic triad (Ser, Asp, and His) in its ML region. Protein sequence identity of ML was calculated: each column shows the identity base on SAL-1 and SAL-2 of S. aureus USA300. (b) Protein sequences were aligned by ClustalW algorithm and its phylogenetic dendrogram was depicted by neighbor-joining tree method. The bar indicates genetic distance. SP, signal peptide; PP, propeptide; ML, mature lipase

These observations add important new insight into the previously known contributions of S. aureus proteases to skin damage (Nakatsuji, Chen et al., 2016; Williams, Costa et al., 2019), and suggest that expression of lipases by Staphylococcus enables bacteria within this genus to establish residence in the HF.

Data availability

No datasets were generated during the current study.