Electricity-producing Staphylococcus epidermidis counteracts Cutibacterium acnes

Staphylococcus epidermidis (S. epidermidis) ATCC 12228 was incubated with 2% polyethylene glycol (PEG)-8 Laurate to yield electricity which was measured by a voltage difference between electrodes. Production of electron was validated by a Ferrozine assay. The anti-Cutibacterium acnes (C. acnes) activity of electrogenic S. epidermidis was assessed in vitro and in vivo. The voltage change (~ 4.4 mV) reached a peak 60 min after pipetting S. epidermidis plus 2% PEG-8 Laurate onto anodes. The electricity produced by S. epidermidis caused significant growth attenuation and cell lysis of C. acnes. Intradermal injection of C. acnes and S. epidermidis plus PEG-8 Laurate into the mouse ear considerably suppressed the growth of C. acnes. This suppressive effect was noticeably reversed when cyclophilin A of S. epidermidis was inhibited, indicating the essential role of cyclophilin A in electricity production of S. epidermidis against C. acnes. In summary, we demonstrate for the first time that skin S. epidermidis, in the presence of PEG-8 Laurate, can mediate cyclophilin A to elicit an electrical current that has anti-C. acnes effects. Electricity generated by S. epidermidis may confer immediate innate immunity in acne lesions to rein in the overgrowth of C. acnes at the onset of acne vulgaris.

Bacterial culture. S. epidermidis (ATCC 12,228) was cultured in tryptic soy broth (TSB) (Sigma, St. Louis, MO, USA). C. acnes (ATCC 6919) was cultured on Reinforced Clostridium Medium (RCM, Oxford, Hampshire, England) under anaerobic conditions using a Gas-Pak (BD Biosciences, San Jose, CA, USA). Bacteria were cultured at 37 °C until the logarithmic growth phase. Bacterial pellets were harvested by centrifugation at 5000 × g for 10 min, washed in phosphate-buffered saline (PBS), and then suspended in PBS or TSB for further experiments. [10 7 colony-forming unit (CFU)/mL] was incubated in 10 ml rich media (10 g/L yeast extract (Biokar Diagnostics, Beauvais, France), 3 g/L TSB, 2.5 g/L K 2 HPO 4 , and 1.5 g/L KH 2 PO 4 ) with and without 2% (20 g/L) of polyoxyethylene glycol 400 monolaurate designated as PEG-8 Laurate (C 28 H 56 O 10 ) by The International Nomenclature of Cosmetic Ingredients (INCI) (Taiwan NJC Corporation, Ltd, Chiayi, Taiwan). In some experiments, S. epidermidis was pretreated with 1 μM TMN 355 (UNI-ONWARD, New Taipei, Taiwan) 33 , under aerobic conditions at 37 °C with shaking at 200 rpm for 12 h before adding into rich media. The 0.002% (w/v) phenol red (Sigma) in rich media acted as a fermentation indicator. A color change from red-orange to yellow indicated the occurrence of bacterial fermentation which was detected by optical density (OD) 560 nm. To determine if PEG-8 Laurate affects the bacterial growth, bacteria (10 7 CFU) was incubated with 2% PEG-8 Laurate in TSB media for 24 h. Bacteria were diluted 1: 10 0 −1: 10 5 into PBS and 10 μl from each dilution was spotted onto a TSB agar plate to count CFU. Plates were incubated at 37 °C for 12 h for S. epidermidis and 72 h for C. acnes to count the colony numbers.

Fermentation of bacteria. S. epidermidis
Detection of bacterial electricity. The voltage difference between the electrodes (cathode and anode) was used to detect the bacterial electricity in vitro. A carbon felt (2.5 cm × 10 cm) and a carbon cloth (10 cm × 10 cm) (Homy Tech, Taoyuan, Taiwan) were used as an anode and a cathode, respectively. The carbon cloth was wrapped up in a nafion membrane N117 (6 cm × 6 cm) (Homy Tech), a proton exchange membrane (PEM), and placed in a 10 cm diameter petri dish. Anode and cathode were connected by copper wires, which in turn were bridged to external resistance of 200 Ω. Bacteria (S. epidermidis or C. acnes) cultured overnight to 10 7 CFU in rich media with/without 2% PEG-8 Laurate (200 μL) were pipetted onto the surface of the anode. In some experiments, S. epidermidis was pretreated with 1 μM TMN 355 for 12 h before pipetting onto the anode. The voltage difference (mV) against time (min) was monitored by a digital multimeter (Lutron, DM-9962SD, Sydney, Australia). The voltage was recorded every 10 s to plot a graph of voltage against time 34 .
The interference of S. epidermidis with the growth of C. acnes in vitro. C. acnes ATCC 6919 (10 7 CFU) was incubated with 0.22 µm-filtered supernatants obtained from the culture of S. epidermidis (10 7 CFU) in the presence or absence of 2% PEG-8 Laurate for 60 or 300 min. After incubation for 1 h, C. acnes was serially diluted and spotted onto a C. acnes selective agar plate to count CFU. In other experiment, rich media containing C. acnes ATCC 6919 (10 7 CFU) was added into a 10 cm diameter petri dish where a carbon felt, a carbon cloth and PEM were placed. S. epidermidis (10 7 CFU) pretreated with/without 1 µM TMN 355 in rich media supplemented with/without 2% PEG-8 Laurate was pipetted on the surface of the anode. After that, C. acnes was collected from the petri dish, serially diluted, and spotted onto a selective agar plate containing rich media plus furazolidone (10 µg/mL) to count CFU. Results in our previous publication 9  www.nature.com/scientificreports/ that furazolidone (10 µg/mL) in a selective agar plate completely inhibits the growth of S. epidermidis, but not C. acnes. To measure the membrane permeability 35 , C. acnes was suspended in 0.5 mM PBS (pH 7.4) containing 10 µg/mL crystal violet (Sigma) followed by centrifugation at 9300 × g for 10 min. After further incubation at 37 °C for 10 min, the suspension was centrifuged at 13,400 × g for 15 min. The OD 590 of the supernatant was measured using untreated C. acnes as a blank. The OD 590 value of 10 µg/mL crystal violet solution was considered as 100%. The percentage of crystal violet uptake was calculated as follows: (OD 590 value of sample/ OD 590 value of crystal violet solution) × 100 = Percentage uptake of crystal violet.
High-performance liquid chromatography (HPLC) analysis. Cultured media of S. epidermidis (10 7 CFU) with/without 2% PEG-8 Laurate after 60 or 300 min of incubation were centrifuged at 5000 rpm for 10 min. The supernatants were filtered through a 0.22 μm microfiltration membrane and butyric acid in supernatants was detected according to the previous protocol 36 . The concentrations of butyric acid were quantified based on a calibration curve of a butyric acid analytical standard.

Statistical analysis.
Experiments were repeated at least three times to ensure reproducibility. Data are presented as mean values ± standard deviation (SD). Statistical significance was determined using Student's unpaired two-tailed t-test, as indicated in the legend (*P < 0.05, **P < 0.01, ***P < 0.001 and ns = non-significant).

PEG-8 Laurate induces electricity production by S. epidermidis. Bacteria develop biofilms on
microbial fuel cell (MFC) electrodes, enhancing electricity production during the EET process 38 . The biofilm matrix produced by the bacteria is conductive, which allows electrons to move efficiently to the electrodes. Addition of organic acids such as acetic acid generates electrons by oxidation of organic acids by bacteria. Here we investigate the electrogenicity of S. epidermidis ATCC 12228, a non-biofilm forming skin bacterium, in the presence of 2% PEG-8 Laurate, a carbon-rich PEG ester of lauric acid. S. epidermidis 10 7 CFU and 2% PEG-8 Laurate in rich media were added onto the surface of the anode. Addition of the same volume of S. epidermidis (10 7 CFU) alone or 2% PEG-8 Laurate alone acted as controls. As shown in Fig. 1, little or no voltage change was detected in media containing S. epidermidis or PEG-8 Laurate alone throughout the monitoring period of 360 min. However, a marked increase in voltage to 4.4 mV was detected when 2% PEG-8 Laurate was added to media containing S. epidermidis. Voltage production peaked at 60 min, and remained steady for 180 min before declining to a baseline voltage of 0.2 mV at 300 min. This result clearly illustrates the biofilm-independent electrogenic properties of skin S. epidermidis in the presence of PEG-8 Laurate.
Electrogenic S. epidermidis impedes the growth of C. acnes. Since both S. epidermidis and C. acnes are bacterial members in an acne microbiome 39 , we sought to assess if electricity generated by S. epidermidis in the presence of PEG-8 Laurate can alter the growth of C. acnes. The culture media of S. epidermidis plus PEG-8 Laurate which elicited the high and low voltage peaks at 60 and 300 min, respectively ( Fig. 1), were added into a culture of C. acnes for 60 min. The incubation of C. acnes with media of S. epidermidis alone was included as a control. As shown in Fig. 2a,b, the number (7.0 ± 0.5 × 10 7 CFU) of C. acnes incubated with media collected from 60 min after the culture of S. epidermidis plus PEG-8 Laurate was significantly lower than that (11.6 ± 0.8 × 10 7 CFU) incubated with media collected from in the culture of S. epidermidis alone. Higher C. acnes counts were detected when C. acnes incubated with media collected from 300 min after culture of S. epidermidis in the presence (11.0 ± 0.5 × 10 7 CFU) or absence (11.3 ± 0.8 × 10 7 CFU) of PEG-8 Laurate. This result suggests that high electricity in media collected from 60 min after culture of S. epidermidis plus PEG-8 Laurate hindered the growth of C. acnes.
Our previous studies have demonstrated that S. epidermidis can mediate fermentation to produce SCFAs including butyric acid to suppress the growth of C. acnes 40 . We next measured the production of butyric acid in the culture media of S. epidermidis with/without PEG-8 Laurate for 60 and 300 min by HPLC. As shown in Fig. 2c Fig. S2). This result demonstrates that cyclophilin A is an essential mediator of electricity production in S. epidermidis.

Production of electrons by S. epidermidis is detected by ferrozine assays and induces the lysis of C. acnes.
In the presence of electrons, ferric (Fe 3+ ) ammonium citrate is converted to ferrozine-cheltable irons (dark brown) which can be quantified by measurement of OD 562 . As shown in Fig. 3b, S. epidermidis plus PEG-8 Laurate induced a higher amount of ferrozine-cheltable irons than S. epidermidis alone. Inhibition of  www.nature.com/scientificreports/ this effect (Fig. 3c,d), illustrating a cyclophilin A-mediated pathway of electricity production in S. epidermidis against C. acnes.

Electron production mediated by cyclophilin A in S. epidermidis attenuates viability of C. acnes in vivo.
Integrating all in vitro data above supports the model that electrogenic S. epidermidis suppresses the growth of C. acnes, thus we employed a mouse ear model to investigate the counteraction of S. epidermidis to C. acnes in vivo. Ears of ICR mice were intradermally injected with S. epidermidis and C. acnes in the presence or absence PEG-8 Laurate for 24 h. As shown in Fig. 4a,b, the number (5.7 ± 0.3 × 10 7 CFU) of C. acnes from mouse ears injected with S. epidermidis and C. acnes in the presence of PEG-8 Laurate was significantly lower than that (55.7 ± 0.9 × 10 7 CFU) injected with S. epidermidis and C. acnes in the absence of PEG-8 Laurate. This result suggests that PEG-8 Laurate provokes electricity production in S. epidermidis against C. acnes. An extremely low voltage difference (< 0.5 mV) was detected when C. acnes and PEG-8 Laurate were pipetted on an anode ( Fig. 1 and Fig. S3). To verify the cyclophilin A-mediated pathway of electricity production in S. epidermidis against C. acnes in vivo, S. epidermidis was pretreated with or without TMN 355 before intradermally injecting into mouse ear with C. acnes and PEG-8 Laurate. The effect of lowering the number of C. acnes by S. epidermidis plus PEG-8 Laurate was significantly reversed when TMN 355-pretreated S. epidermidis was injected into mouse ear (Fig. 4c,d). These data suggest that PEG-8 Laurate-induced electricity mediated by cyclophilin A in S. epidermidis contributes to bacterial interference in vivo.

Discussion
Antibiotics without bacterial selectivity for the treatment of acne vulgaris carry a risk of developing antibiotic resistant C. acnes and may result in dysbiosis in the skin human microbiome. Transferring electron from the cytosol to the exterior of the cell via the EET process represents an alternative strategy that may selectively target www.nature.com/scientificreports/ pathogenic bacteria. Media containing a high content of electricity (> 4 mV) collected from the culture of S. epidermidis plus PEG-8 Laurate for 60 min exerted a marked anti-C. acnes activity (Fig. 2a,b). Although a lower electricity (~ 1.2 mV) was detected when both C. acnes and S. epidermidis plus PEG-8 Laurate were simultaneously present in media (Fig. 3a,c), the anti-C. acnes activity of electrogenic S. epidermidis still remained. Although low electricity of C. acnes in the presence of PEG-8 Laurate was detectable (Fig. 1), electron acceptors on the cell wall of C. acnes may interrupt the intensity of electricity produced by S. epidermidis. It has been reported that electrogenesis can be influenced by different bacteria when they are co-cultured 41 . Addition of C. acnes in the presence of PEG-8 Laurate onto the surface of anode created an extremely low voltage difference (< 0.5 mV) which did not affect the growth of S. epidermidis (Fig. S3), supporting that S. epidermidis, not C. acnes, exerted the electrogenicity and antibacterial activity when both bacteria were co-cultured in the presence of PEG-8 Laurate. www.nature.com/scientificreports/ It has been reported that membrane-bound electron transport protein entrapped inside an electrochemically active biofilm facilitates electron transfer 42 . For example, electron transfer in Shewanella oneidensis (S. oneidensis) or many lactic acid bacteria in the human gut occurs directly by the formation of biofilm, as shown to occur on the surface of electrodes through extensions in the form of nano-wires 43,44 . Interestingly, an indirect electron transfer was also detected in S. oneidensis that involves excreting redox-active mediators such as flavin molecules that act as small diffusible shuttle molecules to transfer electron between electrodes 44 . However, the formation of biofilm is associated with the production of exopolysaccharides (EPS), quorum sensing signaling molecules and other stress factors such as heavy metal stress 45 , salinity 46 , pH 47 , nutrient starvation 48 , and pathogen invasion. Moreover, the products from microbial fermentation, such as hydrogen, formate, or acetate as redox mediators, prove an advantage over the electron transfer by the biofilm formation process 49,50 . A decrease in electron transport observed in S. oneidensis by inhibiting fermentation supports the model that fermentation is likely to be associated with enhancing electron transport 51 . As shown in Fig. S2, S. epidermidis can utilize PEG-8 Laurate to undergo fermentation which may generate SCFAs as redox mediators for electricity production. We, here, identify S. epidermidis ATCC 12,228 as a non-biofilm producing skin commensal bacteria which may generate electricity through PEG-8 Laurate as an electron donor 31,52 . The electricity produced by S. epidermidis plus PEG-8 Laurate can be enhanced by addition of FMN (Fig. S4), supporting flavin-based EET in Gram-positive bacteria 53 .
The addition of S. epidermidis plus PEG-8 Laurate induced a higher voltage (4.4 mV) within 60 min, which remained elevated for ~ 120 min, before declining to baseline by 300 min (Fig. 1). This indicates high electron availability at 60 min, with little remaining 300 min after addition of PEG-8 Laurate into bacterial media. Incubation of C. acnes with media collected at 60 min induced a significantly greater reduction in C. acnes viability relative to the incubation of C. acnes with media collected at 300 min (Fig. 2). There was no statistically significant difference in the content of butyric acid present in media of S. epidermidis plus PEG-8 Laurate at 60 min versus 300 min after culture (Fig. 2c,d). A butyric acid concentration less than 0.1 mM was produced within a 300 min culture of S. epidermidis plus PEG-8 Laurate (Fig. 2c,d). Data from our previous study demonstrated that the minimum bactericidal concentration (MBC) of butyric acid for C. acnes was 10 Mm 54 , indicating that the amount of butyric acid produced within a 300 min culture of S. epidermidis plus PEG-8 Laurate was not sufficient to kill C. acnes. Results in Fig. 2 indicated that electrons, not butyric acid, in culture media exerted the anti-C. acnes property in vitro. Furthermore, direct delivery of a voltage at 4.4 mV to C. acnes in vitro in the absence of S. epidermidis plus PEG-8 Laurate reduced the growth of C. acnes (Fig. S5). Although the reduction (< one log 10 ) of C. acnes growth by 4.4 mV voltage was less than that (> two log 10 ) by S. epidermidis plus PEG-8 Laureate (Fig. 3c), the results in Fig. S5 clearly demonstrated the anti-a. acnes property of electricity. However, we cannot rule out the possibility that SCFAs and other fermentation metabolites potentiated the effect of the electrons to fully eradicate C. acnes in vivo.
Genome-based antagonism between S. epidermidis and C. acnes highlighted the expression of antimicrobial substances in S. epidermidis against C. acnes 55 . Succinate in the metabolites of glycerol fermentation of S. epidermidis effectively inhibited the growth of C. acnes in vivo 56 . The C. acnes phylotype IA1, a high predominance of phylotype in acne lesions, can trigger inflammatory responses including activation of Toll-like receptors (TLRs), secretion of pro-inflammatory cytokines and infiltration of immune cells 57,58 . Although we do not know whether electricity generated by S. epidermidis can directly activate immune cells to eliminate C. acnes in vivo, it has been reported that electric fields increased the phagocytosis in macrophages 59 . Inhibition of cyclophilin A in S. epidermidis by 1 µM TMN 355 for 12 h completely abolished the electricity production (Fig. 3a). Intradermal co-injection of C. acnes with TMN 355 pretreated S. epidermidis in the presence of PEG-8 Laurate only partially reduced the suppressive effect of S. epidermidis plus PEG-8 Laurate on the growth of C. acnes in vivo (Fig. 4c,d). Several microbes have been identified in subepidermal compartments of normal skin 60 . It is worth investigating how S. epidermidis pretreated with or without TMN 355 works together with other skin microbes to fully eliminate C. acnes in vivo. The PEG-8 Laurate fermentation activity of S. epidermidis still remained after pretreatment of S. epidermidis with TMN 355 (Fig. S2), indicating that SCFAs can be produced by TMN 355-preteated S. epidermidis. Both electrons produced by S. epidermidis and SCFA-activated skin immunity 54 may be required for complete eradication of C. acnes in vivo. Electron production by PEG-8 Laurate fermentation of S. epidermidis may provide an immediate innate immunity to lyse C. acnes at an early stage of C. acnes overgrowth. We further show that cyclophilin A is a key mediator of electron production of S. epidermidis. Cyclophilin is highly conserved across genera in both prokaryotes and eukaryotes. TMN 355, a potent cyclophilin A inhibitor, can down-regulate the gene expression of cyclophilin A and Ndh2 in S. epidermidis (data not shown). This is not due to toxicity since the inhibition of cyclophilin A by TMN 355 does not influence bacterial fermentation or survival (Figs S1 and S2). This is in line with a previous study that identified a flavin-based EET mechanism in Gram-positive bacteria that utilizes membrane-anchored Ndh2 for electricity production 61 . Our results demonstrate for the first time that cytoplasmic cyclophilin A in S. epidermidis is an essential component for electricity production.
Low-intensity direct currents or low-frequency alternating electric fields use conductive electrodes to produce free radicals, modify pH or alter exopolysaccharide matrix in bacterial biofilm. The interaction of the electromagnetic field with charged particles present in that matrix causes electron-mediated bacterial cell lysis 26 . In eukaryotic cells, including cancer cells, cell death through an electric field is associated with plasma membrane permeabilization, cytochrome C release into the cytoplasm, disorientation of the spindle microtubules or internucleosomal DNA fragmentation, resulting in necrotic cell transformation 62,63 . The increase in lysis of C. acnes (Fig. 3d) is likely to be due to electrolysis activity generated by S. epidermidis in the presence of PEG-8 Laurate. This has been reported to result from several processes including cell wall degradation 64 , or inactivation or breakdown of membrane proteins involved in cellular respiration glycolysis or cell division such as glycerol-3-phosphate dehydrogenase (GPDH) 65 or filamenting temperature-sensitive mutant Z (FtsZ) 66,67 . These proteins share the similar dipole moments with tubulin in eukaryotes 68 , which is known to undergo disruption by low electric field. In addition, exposure of bacteria to a high pulsed electric field induces plasma membrane www.nature.com/scientificreports/ permeabilization and disrupts cell wall integrity, leading to loss of viability 64 . In the case of Gram-positive bacteria, a high electric pulse can lead to a change in the zeta potential of the surface charge especially at the phosphoryl groups of teichoic acids, causing cell wall degradation 69 . In summary, deciphering the electrogenic characteristics of each bacterium leads to a better understanding of complex interplay involved in the human microbiome. Here, we present a novel modality to inhibit the growth of C. acnes using cyclophilin A-mediated electricity generated by skin S. epidermidis in the presence of PEG-8 Laurate and illustrate an electric pathway for drug targeting in acne vulgaris. www.nature.com/scientificreports/