Generation of Multiple Stable Dermcidin-Derived Antimicrobial

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector ORIGINAL ARTICLE

Generation of Multiple Stable Dermcidin-Derived Antimicrobial Peptides in Sweat of Different Body Sites Siegbert Rieg1, Silke Seeber2, Heiko Steffen1, Andreas Humeny2, Hubert Kalbacher3, Stefan Stevanovic4, Akihiko Kimura5, Claus Garbe1 and Birgit Schittek1

Antimicrobial peptides (AMPs) are effector molecules of innate immunity. Dermcidin (DCD), a recently discovered AMP with broad-spectrum activity, is produced constitutively by the eccrine sweat glands and secreted into sweat. In this study, we investigated the proteolytic processing, site-specific expression, and stability of DCD peptides in eccrine sweat. Using surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF-MS) and reversed-phase high-pressure liquid chromatography analysis, we identified in eccrine sweat 14 proteolytically processed DCD peptides. Semiquantitative SELDI-TOF-MS analysis indicated that processing of DCD-1L is individually different, but generates a few dominant peptides. At body sites with a high probability for contact with pathogenic microorganisms, a high amount of antimicrobial active DCD peptides was detected in sweat. Furthermore, we show that the secretion rate of DCD is constant during a period of prolonged sweating and that DCD peptides are stable in sweat over several hours. Other known AMPs like the human cathelicidin LL-37 and a-orb-defensins were not detected in significant quantity in eccrine sweat. Owing to the durable and abundant presence, DCD-derived peptides contribute to the first line of defense by building a constant barrier that overlies the epithelial skin. Journal of Investigative Dermatology (2006) 126, 354–365. doi:10.1038/sj.jid.5700041; published online 22 December 2005

INTRODUCTION (Wiedow et al., 1998) are produced constitutively by keratino- Epithelia of multicellular organisms are continuously exposed to cytes, the human cathelicidin LL-37 (Frohm et al., 1997) and the a large number of microorganisms. The remarkable resistance b-defensin-2(HBD-2)and-3(HBD-3)(Harderet al., 1997, 2001) to infection indicates effective defense mechanisms along the are induced in keratinocytes under inflammatory conditions. epithelial barriers. Within this first line of defense, antimicrobial We recently discovered a new AMP named dermcidin peptides (AMPs) are emerging as important effector molecules. (DCD) with a broad spectrum of activity and no homology to In mammalian skin, AMPs are produced by cells within other known AMPs (Schittek et al., 2001). DCD is expressed the epithelial lining, are delivered by circulating cells, or are in the eccrine sweat glands, secreted into sweat, and produced by adjacent cells of the epithelial appendages. In transported via sweat to the epidermal surface. In contrast human skin, keratinocytes are a major source of antimicrobial to other AMPs such as human cathelicidin LL-37 and the active peptides or larger proteins. Whereas the ribonuclease human b-defensins, DCD is not induced in epidermal RNase 7 (Harder and Schroder, 2002), psoriasin (S100A7) keratinocytes under inflammatory skin conditions like atopic (Glaser et al., 2005), and secretory leukoprotease inhibitor dermatitis, psoriasis, or lichen planus, but is constitutively produced by the eccrine sweat glands (Rieg et al., 2004). 1Department of Dermatology, Eberhard-Karls-University Tu¨bingen, Tu¨bingen, Thus, DCD takes part in the constitutive innate defense of Germany; 2Institute of Biochemistry, Friedrich-Alexander-University of human skin. Recently, we showed that patients with atopic Erlangen-Nu¨rnberg, Erlangen, Germany; 3Medical and Natural Sciences dermatitis have a reduced amount of DCD peptides in sweat, Research Center (MNF), Eberhard-Karls-University Tu¨bingen, Tu¨bingen, which correlated with diminished antimicrobial activity of Germany; 4Department of Immunology, Institute for Cell Biology, University of Tu¨bingen, Tu¨bingen, Germany and 5Department of Forensic Medicine, eccrine sweat in vivo (Rieg et al., 2005). Wakayama Medical University, Wakayama, Japan Analogous to other known AMPs, a precursor protein of Correspondence: Dr Birgit Schittek, Department of Dermatology, DCD is produced, which is proteolytically processed to Eberhard-Karls-University Tu¨bingen, Liebermeisterstr. 25, Tu¨bingen generate antimicrobial active DCD peptides. The DCD- D-72076, Germany. E-mail: [email protected] precursor protein consists of 110-amino-acid residues with a Abbreviations: AMP, antimicrobial peptide; DCD, Dermcidin; RP-HPLC, 19-amino-acid N-terminal signal peptide (Schittek et al., reversed-phase high-pressure liquid chromatography; SELDI-TOF-MS, surface-enhanced laser desorption ionization time-of-flight mass spectrometry 2001). Antimicrobial active DCD peptides are derived from Received 12 May 2005; revised 11 August 2005; accepted 19 August 2005; the C-terminal region of the precursor protein, DCD-1L published online 22 December 2005 (consisting of 48 amino acids) and DCD-1 (47 amino acids,

354 Journal of Investigative Dermatology (2006), Volume 126 & 2005 The Society for Investigative Dermatology S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

lacking the last leucine). Both peptides show antimicrobial 2004). To analyze whether LL-37 and other known AMPs or activity against a variety of pathogenic microorganisms proteins like a- and b-defensins and lysozyme are present in including Staphylococcus aureus, E. coli, Entercoccus faeca- sweat, we performed SELDI-TOF-MS from sweat samples lis, and Candida albicans under in vitro conditions resem- from 18 healthy volunteers and in parallel separated eccrine bling human sweat. In human sweat, 1–10 mg/ml of DCD-1 is sweat by RP-HPLC and performed antimicrobial assays from found, a concentration that is toxic to most microorganisms the fractions. From antimicrobial active and nonactive tested (Schittek et al., 2001; Flad et al., 2002). Using surface- fractions, we performed mass spectrometry and Edman enhanced laser desorption ionization time-of-flight mass sequencing. Using both methods besides DCD peptides, we spectrometry (SELDI-TOF-MS) technology, we described that could not detect other known AMPs in sweat in significant besides these peptides, shorter peptides with 43–46 amino quantity (Figures 1 and 2). Neither LL-37 (4,493.3 Da) nor the acids could be identified in eccrine sweat (Flad et al., 2002). derivatives KR-20 (2,468.9 Da), RK-31 (3,800.5 Da), and In this study, we investigated (a) the spectrum of KS-30 (3,644.3 Da) as well as the b-defensins HBD-1 proteolytically processed DCD peptides and (b) the presence (3,928.6 Da), HBD-2 (4,328.2 Da), HBD-3 (5,155.2 Da), or of other AMPs like human defensins and the human a-defensins HNP-1 (3,442.1 Da), HNP-2 (3,371.0 Da), and cathelicidin LL-37 in eccrine sweat of healthy volunteers HNP-3 (3,486.1 Da) are present in eccrine sweat. As the using reversed-phase high-pressure liquid chromatography above-mentioned cathelicidins and defensins are cationic (RP-HPLC) and SELDI-TOF-MS. Furthermore, we analyzed peptides, we carried out additional SELDI-TOF-MS analysis whether (a) the amount of DCD peptides in sweat differs with a cation exchange chip (WCX2). In accordance with between several body sites, (b) the rate of DCD secretion measurements with a reversed-phase chip (H4), none of the remains constant over time, (c) whether DCD peptides are above-mentioned cationic AMPs could be detected (Figure stable in sweat, and (d) whether DCD-derived peptides are 3a). Furthermore, to rule out a weak affinity of human present in tears, saliva, and nasal secretions. cathelicidin, a-orb-defensins to the reversed-phase chip, we analyzed synthetic LL-37, HNP-2, and HBD-1 peptides on a RESULTS reversed-phase chip. Synthetic LL-37 could be identified Several DCD-derived peptides are present in human sweat when added to eccrine sweat samples that did not contain Using SELDI-TOF-MS, we previously identified in human LL-37-corresponding masses in previous analyses in concen- sweat from healthy persons five proteolytically processed trations as low as 2 mM (Figure 3b). In SELDI analyses of sweat peptides derived from the full-length DCD protein sequence with low amounts of DCD-derived peaks in the molecular (Schittek et al., 2001; Flad et al., 2002). Among those are the weight range of 4,493 Da, LL-37 can still be detected in a peptides DCD-1L (48mer) and DCD-1 (47mer), which are concentration of 1.1 mM (Figure 3b). Accordingly, HNP-1/-2 antimicrobially active on different microorganisms. and HBD-1 were detected in concentrations of 2.9 and To get the spectrum of all possible processed DCD 2.5 mM, respectively (Figure 3c); HBD-2 was detected in a peptides in sweat, we performed SELDI-TOF-MS analysis of concentration of 2.3 mM (Figure 3d). Furthermore, using physical-exercise-induced sweat of the forehead of 18 healthy Western blot analysis, the a-defensins HNP-1, HNP-2, and volunteers (Figure 1). In addition, we separated sweat of four HNP-3, b-defensin HBD-2 or lysozyme could not be detected healthy volunteers by RP-HPLC and performed from the in eccrine sweat, whereas DCD was readily detectable collected fractions mass spectrometry and Edman sequencing (Figure 3e). Considering the sensitivity of our Western blot as described earlier (Figure 2) (Schittek et al., 2001). In analysis, we conclude that the defensins are either absent Table 1, all identified processed DCD-derived peptides are from eccrine sweat or present at concentrations lower than summarized. We identified 13 different DCD peptides derived 2 mM. These results indicate that DCD-derived peptides are from the C-terminal end and one N-terminal-derived peptide the predominant AMPs in sweat, and that neither defensins named YDP-42 and its dimeric form. DCD-1 (pos. 63–109) nor LL-37 are present in significant quantity, that is, and SSL-46 have lost the last leucine or valine residue, antimicrobial active quantity, in sweat. respectively, compared to the parental peptide DCD-1L and SSL-45 has in addition lost the serine. SSL-29 and SSL-25 are Processing of DCD-1L is individually different, but generates shorter peptides and have the same N-terminus as DCD-1L, a few dominant peptides beginning with SSL. The other peptides lack the first three SELDI analysis of sweat of 18 healthy persons indicates an amino acids SSL from DCD-1L, beginning with the amino individual proteolytic pattern of DCD-derived peptides in acids LEK, and are either 41–45 amino acids long (LEK-41 and sweat (Figure 1). Interestingly, in most individuals, we LEK-45) or only 24 or 26 amino acids long (LEK-24 and LEK- identified the DCD peptides DCD-1L (15 of 18), DCD-1 (13 26). Interestingly, nine of the 13 C-terminal-derived DCD of 18), SSL-46 (14 of 18), LEK-45 (18 of 18), LEK-44 (12 of peptides are anionic, two peptides have a neutral charge and 18), LEK-43 (14 of 18), and the shorter DCD peptides that two shorter peptides – SSL-25 and LEK-24 – are cationic. have either a neutral charge or are cationic: SSL-29 (16 of 18) and SSL-25 (11 of 18), respectively (Table 2). Among those, at LL-37 and defensins are not expressed in significant quantity least one of the most prominent ones, DCD-1L, LEK-45, or in eccrine sweat SSL-29, is always present. It has been described that the AMP LL-37 and truncated forms Despite these dominant peptides in most sweat samples, of this peptide are present in sweat (Murakami et al., 2002, there are individual differences in the amount of each peptide

www.jidonline.org 355 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

2,000 3,000 4,000 5,000 6,000 2,000 3,000 4,000 5,000 6,000 30 40 4,818.4 4,606.2 30 4,818.3 4,530.9 20 4,318.7 20 2,868.2 4,530.5 5,733.6 (O) 5,733.6 (O) 4,705.5 10 4,705.5 10 2,408.0 4,301.8 2,866.6 2,581.3 0 0

4,706.0 4,529.4 15 20 5,733.6 4,819.1 4,818.0 4,301.4 5,733.6 10 (O) (O) 2,867.3 4,418.2 10 2,580.6 4,418.3 2,409.2 2,867.4 5 2,263.8 4,705.7 0 0

30 4,703.9 15 4,818.2 5,733.6 4,817.2 4,530.6 10 20 2,867.5 (O)

(O) 5,733.6 2,580.2 4,318.5 10 5 2,350.3 4,603.7 2,408.4 4,704.7 2,864.0 0 0

60 4,319.4 4,530.4

20 40 4,818.1 4,417.2 5,733.6 (O)

(O) 4,529.6 20 10 2,411.2 4,416.9 5,733.6 4,705.1 2,867.0 2,264.5 2,866.1 0 0 15 4,318.5 4,317.8 5,733.6 40 4,530.2 10 4,605.8 4,606.4 2,867.4 2,867.1 (O) 20 (O) 5,733.6 5 2,580.5 4,817.3 4,405.3 2,263.7 4,705.1 4,818.7 0 0 2,000 3,000 4,000 5,000 6,000 2,000 3,000 4,000 5,000 6,000

Figure 1. Spectra of SELDI-TOF-MS-analysis of human sweat of 10 healthy volunteers using a reversed-phase (H4) protein chip. These spectra are representative of the SELDI-TOF-MS analysis of overall 18 analyzed sweat samples. Shown are masses from 2,000 to 6,000 Da. After baseline subtraction, spectra were normalized to the bovine insulin peak (5,733.6 Da). From the masses, several DCD-derived peptides could be identified (summarized in Table 1). The deviation of the masses measured in the SELDI-TOF-MS analyses to the theoretical masses listed in Table 1 is around 71.5 Da in the mass ranges 3,000–5,000Da and around 73 Da in the mass ranges 2,000–3,000Da.

as well as the proteolytic pattern. Whereas in most peptides in the sweat of healthy subjects using SELDI-TOF- individuals (13 volunteers), seven to nine C-terminal-derived MS. Sweat samples collected from various body sites of five DCD peptides are present in sweat, in a few individuals (five healthy volunteers were applied together with 100 fmol of persons) only four to five DCD-derived peptides could be bovine insulin (MW 5,733.6 Da) on a reversed-phase (H4) identified in sweat. Irrespective of the number of total Protein Chip Array (Ciphergen Biosystems, Freemont, CA). processed peptides, most individuals have two to four major The spectra were analyzed by Protein Chip Software 3.1 DCD-processed peptides in sweat (Table 2). This indicates (Ciphergen Biosystems, Freemont, CA). After baseline sub- that each individual processes the DCD peptides differently, traction, all spectra were normalized to the bovine insulin but that there are prominent peptides such as DCD-1L, DCD- peak for semiquantification of the DCD-derived peptides. As 1, SSL-46, SSL-29, SSL-25 and LEK-45 that are present in the can be seen in Figure 4, the amount of DCD-derived peptides majority of sweat samples analyzed. is different at the various body sites. Whereas at the thorax and legs relatively low levels of DCD peptides are present in The amount of DCD-derived peptides in sweat is dependent eccrine sweat, on the face, arms, and especially in the palms, on the body site the amount of DCD peptides in sweat is high. In the region of To analyze whether there are differences in the amount of the axilla, where in contrast to the other analyzed body sites DCD-derived peptides secreted into sweat at various body sweat is secreted predominantly, but not exclusively by sites, we performed semiquantitative analysis of DCD apocrine sweat glands, the amount of DCD peptides is also

356 Journal of Investigative Dermatology (2006), Volume 126 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

6 2 1: DCD-1L DCD peptides. After incubation of sweat for 24, 48, and up to 2: DCD-1 3.30 72 hours, C-terminal DCD peptides were still detected by 14.78 3: SSL-46, LEK-44 Western blot (Figure 6a). Moreover, SELDI-TOF-MS analysis 4: LEK-43 of eccrine sweat after 4, 16, and 40 hours of incubation at 5: LEK-44, SSL-25 room temperature revealed mass profiles still containing 3 1 6: YDP-42 prominent DCD peaks after 16 and 40 hours of incubation, yet a slow but continuous shift towards the processed DCD peptides LEK-43, LEK-42, LEK-41, and SSL-29 can be 5 29.22 observed (Figure 6b). These results indicate that DCD peptides in sweat are relatively stable. 4 10.40 DCD peptides are not present in nasal secretion, tears, and

27.57 saliva Next, we analyzed whether DCD-derived peptides are 30.61

Absorption (215 nm) present in other body fluids like nasal secretions, tears, and 7.57

5.54 saliva. As can be seen in Figure 7, DCD peptides are not 11.62 19.09 38.18 present in these body fluids, whereas the a-defensins HNP-1, 4.85 6.61 HNP-2, and HNP-3 are present in saliva and to a lesser extent 25.70

24.96 in tears. Other AMPs like b-defensins and LL-37 were not 20.53 17.76

48.85 detected by SELDI-TOF-MS analysis in tears, nasal fluid, and 47.25 43.73

45.97 saliva. As the sensitivity of the SELDI-TOF-MS method for the detection of b-defensins and LL-37 is in the range of 2 mM (see above), these peptides are either absent or present at 5045403530252015105 concentrations lower than 2 mM in these body fluids analyzed Retention time (minutes) ex vivo. Lysozyme was present in all analyzed body fluids, Figure 2. RP-HPLC analysis of eccrine human sweat (spectrum of a single except sweat (data not shown and see Figure 3e). healthy volunteer representative of four). Among other coeluted proteins in some fractions, we identified in fractions 1–6 by mass spectrometry and DISCUSSION Edman microsequencing 1 DCD-1L, 2 DCD-1, 3 SSL-46 and LEK-44, The eccrine sweat gland is one of the major cutaneous 4 LEK-43, 5 LEK-44 and SSL-25, and 6 YDP-42. Fractions 1–3 and 5 are appendages with approximately 2–4 million glands that are antimicrobially active against S. aureus. The antimicrobial activity of these fractions correlates well with the activity of the corresponding synthetic distributed over nearly the entire body surface (Sato et al., peptides to S. aureus (see Schittek et al., 2001; Rieg et al., 2005). 1989). For long, only its role in thermoregulation was appreciated. Recently, DCD, a novel human AMP, was described as being a major constituent of eccrine sweat, low (data not shown). This reflects the fact that DCD is only thereby indicating that the eccrine apparatus is capable of expressed in eccrine sweat glands and not in apocrine sweat functioning as an effector organ in the first line of defense of glands (Schittek et al., 2001; Minami et al., 2004). These human skin. In our present investigation, we analyzed the results imply that at body sites with a high probability for a proteolytic processing of DCD in eccrine sweat. Furthermore, contact with pathogens (palms, face, arms), a pronounced we examined site-specific DCD expression as well as the amount of antimicrobial DCD peptides is secreted into sweat. kinetics of secretion and degradation of DCD peptides in eccrine sweat. Additionally, we addressed the question of DCD is secreted into sweat at a constant rate whether other AMPs are present in significant amounts in Next, we asked whether the secretion rate of DCD into sweat human eccrine sweat. changes over time. Therefore, we collected sweat from five Using SELDI-TOF-MS analysis and RP-HPLC fractionation, volunteers at different time points during a prolonged episode we detected 14 DCD-derived peptides in eccrine sweat, 13 of sweating – immediately at the beginning and 20 and peptides of the C-terminus, and one peptide of the 40 minutes afterwards – and performed a semiquantification N-terminus of the full-length DCD sequence. Of the different using SELDI-TOF-MS. Figure 5 shows that the amount of truncated peptides, DCD-1L, DCD-1, SSL-46, LEK-45, LEK-44, DCD-derived peptides in sweat does not appreciably change LEK-43, and SSL-29 were found in at least two-thirds of all during a sweating period of 40 minutes. This indicates that individuals. Our analysis revealed an average of seven the secretion rate of DCD keeps constant over time. different DCD-derived peptides (ranging from 4 to 10 DCD peptides) in eccrine sweat of healthy volunteers. Thus, by DCD-derived peptides are stable in sweat postsecretory proteolytic processing, the DCD gene product Are DCD peptides rapidly destroyed by proteases in sweat or gives rise to a whole group of truncated DCD peptides with are they stable? To answer this question, we collected native different biochemical properties and yet unknown functional sweat and kept it at room temperature up to 72 hours. Sweat activities. samples were evaluated by Western blot using the mono- DCD-1L and DCD-1 exhibit antimicrobial activity against clonal G-81 antibody, which recognizes all C-terminal-derived Gram-positive organisms including S. aureus, E. faecalis, and

www.jidonline.org 357 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

Table 1. DCD-derived peptides identified in human sweat Name Amino-acid positions/amino-acid sequence Theor. mass (Da) pI Charge

DCD-1L 63–110 4,818.5 5.07 2 SSLLEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDSVL DCD-1 63–109 4,705.3 5.07 2 SSLLEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDSV SSL-46 63–108 4,606.2 5.07 2 SSLLEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDS SSL-45 63–107 4,519.1 5.15 2 SSLLEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLD SSL-29 63–91 2,869.3 5.97 0 SSLLEKGLDGAKKAVGGLGKLGKDAVEDL SSL-25 63–87 2,412.8 9.40 +2 SSLLEKGLDGAKKAVGGLGKLGKDA LEK-45 66–110 4,531.2 5.08 2 LEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDSVL LEK-44 66–109 4,418.0 5.08 2 LEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDSV LEK-43 66–108 4,318.9 5.08 2 LEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLDS LEK-42 66–107 4,231.8 5.15 2 LEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVLD LEK-41 66–106 4,116.3 5.62 1 LEKGLDGAKKAVGGLGKLGKDAVEDLESVGKGAVHDVKDVL LEK-26 66–91 2,582.0 6.24 0 LEKGLDGAKKAVGGLGKLGKDAVEDL LEK-24 66–89 2,353.7 8.38 +1 LEKGLDGAKKAVGGLGKLGKDAVE YDP-42 20–61 4,302.6/8,605.2 5.62 1 YDPEAASAPGSGNPCHEASAAQKENAGEDPGLARQAPKPRKQR The sequence of full-length DCD is: MRFMTLLFLTALAGALVCAYDPEAASAPGSGNPCHEASAAQKENAGEDPGLARQAPKPRKQR SSLLEKGLDGAKKAV GGLGKLGKDAVEDLESVGKGAVHDVKDVLDSVL DCD-1L (amino-acid positions 63–110) is marked in bold, and the signal peptide (amino-acid positions 1–19) is marked in italics.

Gram-negative organisms including E. coli (Schittek et al., Proteolytic processing of DCD peptides and other sweat- 2001). Further investigations revealed an extended anti- derived peptides and proteins may principally be due to host microbial spectrum including S. epidermidis (Vuong et al., proteases, which can be keratinocyte-derived (Komatsu et al., 2004), Pseudomonas putida, methicillin-resistant S. aureus, 2003) or are produced by eccrine sweat glands. In eccrine as well as rifampicin- and isoniazid-resistant Mycobacterium sweat, diverse proteases have been described, among these tuberculosis (Lai et al., 2005). Additionally, shorter DCD kallikrein and kininase II (Fraki et al., 1970; Hibino et al., fragments like SSL-46 (amino-acid positions 63–108) (Rieg 1994), acid proteases resembling cathepsin B1 and cathepsin et al., 2005) and SSL-25 exhibit antimicrobial activity (data D (Fraki, 1976), and matrix metalloproteinase 19 (Sadowski not shown). Thus, analogous to the processing of other et al., 2003). Recent evidence suggests that the sweat-derived human AMPs like cathelicidin LL-37, proteolytically derived aspartate protease cathepsin D and a 1,10-phenanthroline peptides exhibiting antimicrobial activity originate from the sensitive carboxypeptidase (together with a yet unidentified C-terminus of the full-length DCD peptide. Interestingly, endopeptidase) are involved in the postsecretory processing proteolytically processed DCD peptides possess net charges of DCD (Baechle et al., submitted). Furthermore, proteases between 2 and þ 2; thus, initially, anionic DCD-1L and derived from sebaceous glands may contribute to the DCD-1 are the source of the generation of neutral or cationic degradation of AMPs on the skin surface (Frohlich et al., shorter DCD fragments due to subsequent processing. 1995; Komatsu et al., 2003). It is noteworthy that the Whether the alteration from anionic to cationic peptides expression of the keratinocyte-derived matrix metalloprotein- leads to a different antimicrobial spectrum and additional ase 19 is upregulated under inflammatory conditions; there- functional activities or is part of the inactivation of DCD fore, the host is capable of modulating the proteolytic peptides is subject to current investigations. capacity on the outermost surface (Sadowski et al., 2003).

358 Journal of Investigative Dermatology (2006), Volume 126 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

ac 2,000 3,000 4,000 5,000 6,000 3,000 3,500 4,000 4,500 5,000 7.5 2,412.2 3,369.4 4,705.3 4,317.7 15 5 3,440.6 4,301.0 10 HNP-1/2 Sweat 4,416.7 2.5 4,302.4 4,421.0 5,733.6 5 4,530.0 2.9 M 4,605.4 0 0 125 4,316.1 100 b 75 3,000 3,500 4,000 4,500 5,000 HNP-1/2 60 4,228.9 50 4,414.9 3,943.9 4,113.7 4,489.8 0.29 M 25 40 LL-37 0

20 4,114.4 4,317.3 22 M d 3,000 3,500 4,000 4,500 5,000 25 0 4,603.5 15 20 4,315.1 3,945.1 15 4,115.5 4,527.9 HBD-1 10 4,317.5 10 3,927.2 LL-37 4,816.5 2.5 M 5 5 2.2 M 0 4,490.1 100 0 4,604.4 4,316.8 75

3,000 3,500 4,000 4,500 5,000 50 HBD-1 4,490.9 4,528.9 100 25 0.25 M 4,817.5 75 0 50 LL-37 3,280.3 3,000 3,500 4,000 4,500 5,000 25 22 M 3,537.6 4,317.4 4,325.2 0 75

50 7.5 3,957.4 4,229.7 HBD-2 4,114.5 4,416.0 25 2 M 5 4,529.2 LL-37 0 2.5 1.1 M 4,490.3 4,317.4 0 40 3,957.4 4,230.1 kDa 4,416.6 HBD-2 e 1234 20 4,114.5 14 0.2 M 6.5 DCD 0 3.5 14 6.5 HNPs 3.5 14 6.5 HBD-2 3.5 123

14 Lysozyme 6.5

Figure 3. LL-37 and defensins are not present in significant quantity in eccrine sweat. (a) Spectra of SELDI-TOF-MS analysis of eccrine human sweat using a cation exchange (WCX2) protein chip: corresponding masses of the cationic AMPs LL-37, its derivates, or a-orb-defensins are not detected. (b) Spectra of SELDI- TOF-MS analysis of synthetic LL-37 spiked into sweat samples using a reversed-phase (H4) protein chip: LL-37 (4,493.3 Da) is marked with an arrow and is clearly detectable at a concentration of 20 and 2 mM (first two panels). When LL-37 is spiked into sweat with low amounts of DCD-derived peaks in the molecular weight range of 4493 Da, LL-37 can still be detected in a concentration of 1.1 mM (bottom two panels). (c and d) Spectra of SELDI-TOF-MS analysis of (c)amixof a-defensins HNP-1 (3,442.1 Da) and HNP-2 (3,371.0 Da) or (d) b-defensins HBD-1 (3,928.6 Da) and HBD-2 (4,328.2 Da) spiked in sweat using a reversed-phase (H4) protein chip. The defensins are marked with an arrow and are clearly detectable at a concentration of B2 mM.(e) Western blot analysis of 15 ml of human eccrine sweat (lanes 1 and 2 in the first three panels and lanes 2 and 3 in the last panel) with antibodies against dermcidin (G-81), a-defensins (HNPs), b-defensin 2 (HBD-2), or lysozyme. As positive controls, we used the respective synthetic peptides (lane 3: each 100 ng; lane 4: 500 ng total peptide) in the first three panels and in the last panel 15 ml saliva (lane 1) for the detection of lysozyme. From these data, we conclude that the a-andb-defensins and lysozyme are either not present in eccrine human sweat or present at amounts lower than 2 mM.

www.jidonline.org 359 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

Table 2. DCD-derived peptides identified in human sweat of 18 individuals Percentage of individuals Individual/ (numbers with DCD peptide 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 peptide)

DCD-1L X XXXXXXXX XX X XX X83% (15/18) DCD-1 XX X XXXXXXX X X X X 78% (14/18) SSL-46 XXXX XX X X XXX X X X 78% (14/18) SSL-45 X X X X X X 33% (6/18) LEK-45 XXX X X XXX XXX XX XXXX X 100% (18/18) LEK-44 X X XX X XXXX X X X 67% (12/18) LEK-43 X X XXXXXXX XXXXX 78% (14/18) LEK-42 X 6% (1/18) LEK-41 X 6% (1/18) SSL-29 XXX X XXXXXXX X XXX X 89% (16/18) SSL-25 X X XX XX XXX X X 61% (11/18) LEK-26 X X X X X X X 38% (7/18) LEK-24 XXX X 22% (4/18) DCD-29 and LEK-26 have a neutral charge, and SSL-25 and LEK-24 are cationic peptides. The other peptides are anionic. Marked in bold are peptides that are predominantly present in the sweat sample of the individual.

Apart from host proteases, processing of DCD and potentially were performed for the a- and b-defensins, with similar other AMPs in eccrine sweat may be due to microbial- results. As human a- and b-defensins and LL-37 are cationic derived proteases. Recently, degradation of human cathe- peptides, we completed our experiments with the use of a licidin LL-37 by the S. aureus-derived metalloproteinase cation exchange chip. Consistent with the results of the aureolysin and the glutamylendopeptidase V8 protease was reversed-phase chip, none of these AMPs was detectable in described (Sieprawska-Lupa et al., 2004). Accordingly, eccrine sweat. Furthermore, we confirmed our results using P. aeruginosa elastase, E. faecalis gelatinase, or P. mirabilis Western blot analysis of human sweat for the a- and metalloprotease are able to degrade and inactivate LL-37. b-defensins and lysozyme. The data indicated that these Interestingly, besides antimicrobial DCD-derived peptides peptides/proteins are either absent in eccrine sweat or present and proteases, which generate shortened DCD fragments, at amounts lower than 2 mM. In previous studies, expression of eccrine sweat additionally contains protease inhibitors such b-defensins HBD-1 and HBD-2 in eccrine sweat glands was as cystatin A (Zeeuwen et al., 2001). It is tempting to described by in situ hybridization and immunohistochemistry speculate whether these protease inhibitors play a role in the (Fulton et al., 1997; Ali et al., 2001; Rieg et al., 2004); protection of proteolytic degradation of sweat-derived however, these peptides have not been identified in human proteins and peptides by inhibition of either host- or sweat until now. In summary, we could not isolate other microbial-derived proteases. AMPs in several independent experiments and with different Expression of a second AMP in eccrine sweat was methods, and therefore conclude that in human sweat described recently. The human cathelicidin LL-37 was isolated ex vivo DCD-derived peptides are the main AMPs detected by Western blot analysis in eccrine sweat in that are present in significant functional amounts. concentrations of 0.1 mM (Murakami et al., 2002). In our Our results reveal a constant DCD secretion during investigation using SELDI-TOF-MS and RP-HPLC analysis 40 minutes of thermally induced sweating. This indicates a followed by mass spectrometry and Edman sequencing, we durable secretion of DCD in eccrine sweat, which is not could not detect the human cathelicidin LL-37, shortened LL- subject to rapid exhaustion in periods of prolonged sweating. 37-derivates, or the unprocessed precursor protein. Whereas Additionally, we show that – despite further proteolytic Murakami et al. (2002, 2004) analyzed concentrated sweat processing generating shorter DCD derivates – DCD-derived (20:1 or 50:1) by Western blot analysis or RP-HPLC for the peptides are still present 72 hours after incubation at room detection of LL-37 peptides, we used crude sweat samples for temperature. This indicates that DCD-derived peptides are SELDI-TOF-MS analysis. To rule out a weak affinity of LL-37 relatively stable in sweat. Our findings are in accordance to the reversed-phase chip, we analyzed synthetic LL-37 with previous reports of Sagawa et al., which managed to peptide spiked in sweat with the same protocol. Using this detect DCD-derived peptides by ELISA (for medicolegal approach, synthetic LL-37 could clearly be detected in aspects) in aged sweat for up to 11 weeks (Sagawa et al., concentrations as low as 2 mM, the concentration required 2003). Even in the absence of macroscopic sweating, a daily for antimicrobial activity of LL-37. The same experiments amount of B200 ml eccrine sweat is perspirated. Thus, a

360 Journal of Investigative Dermatology (2006), Volume 126 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

a 2,000 3,000 4,000 5,000 6,000 a 2,000 3,000 4,000 5,000 6,000 4,817.7 30 5,733.6 4,530.7 4,705.0 6 2,411.4 2,309.4 4,704.9 20 4,417.8 4,416.9 4

10 5,733.6 2,409.4 4,301.9 4,606.3 2 2,868.0 0 0

b 5,733.6 b 6 4,818.9 6 5,733.6 4,302.0 4 4,417.4 2,866.6 4 2,309.8 2 4,608.1 4,704.5 2 0 0 c 20 4,818.8 c 4,417.4 5,733.6 15 4,705.6 6

10 4,531.3 2,309.3 5,733.6 4 2,411.8 4,705.2 4,302.5 5 2,408.9 2,867.0 2 0 0 d 6 5,733.6 Figure 5. SELDI-TOF-MS-analysis of eccrine sweat collected at different 4,818.4 4 time points during a prolonged period of sweating. Sweat was collected at 4,302.4 the beginning, time (a)0,(b) 20 minutes afterwards, and (c) 40 minutes after 2,866.6 2 4,605.7 the beginning of sweating. Spectra shown are from sweat of a single healthy 2,408.6 volunteer representative of five. 0

e at the palms, soles, and face, our results underline observa- 100 4,818.0 tions of site-specific expression of AMPs with pronounced 75 4,704.6 DCD secretion in body areas, which are most often exposed 50 4,530.0 to pathogens and minor trauma. 4,416.7 25 2,408.2 4,302.0 5,733.6 In conclusion, our results indicate that DCD is the 0 predominant AMP expressed in eccrine sweat. Multiple DCD-derived peptides are generated by proteolytic proces- Figure 4. SELDI-TOF-MS-analysis of eccrine sweat collected from different sing, confirming the concept that human eccrine sweat body sites: (a) forehead, (b) thorax, (c) arms, (d) legs, and (e) palms. Shown represents a suitable and efficient vehicle to deliver innate are masses from 2,000 to 6,000 Da. After baseline subtraction, spectra were defense molecules onto the skin surface. Through the normalized to the bovine insulin peak (5,733.6 Da). Spectra shown are from sweat of a single healthy volunteer representative of five. continuous and abundant presence of DCD peptides in sweat, a constitutive first line of defense is established, which overlies the epithelial barrier of the skin. constitutive antimicrobial barrier overlying the epithelial sheets of the skin is secured by constant baseline DCD MATERIALS AND METHODS sweating, by durable DCD secretion when intense sweating Collection of sweat, tears, saliva, and nasal secretions occurs and by the continuous presence of DCD-derived For SELDI-TOF-MS analysis of the proteolytic pattern of DCD peptides owing to proteolytic processing, which occurs with peptides, eccrine sweat was collected from 18 healthy volunteers moderate to low kinetics. (seven women and 11 men, mean age 28.0 years, without skin or Site-specific expression of AMPs has been described for systemic diseases, who were on no systemic medications, except for HBD-1 and HBD-2, with pronounced expression in plantar two women on oral contraceptives, five volunteers were smokers) skin and in the scalp (Ali et al., 2001). Accordingly, the newly after sweat induction by physical exercise on a bicycle ergometer. described AMP psoriasin (S100A7) can be detected in highest For each participant, 5 ml of the initial portion of sweat of the concentrations on the skin surface of face, scalp, palms, and forehead was collected and frozen immediately at 201C until soles (Glaser et al., 2005). We here demonstrate regional analysis. Participants were asked not to wash or apply ointments differences in the concentration of DCD-derived peptides in 6 hours prior to the investigation. For the analysis of body site- eccrine sweat with the highest amounts of DCD peptides in specific differences and DCD secretion over time, sweating was the sweat of palms, arms, and forehead. Taken together with thermally induced by sauna visits. Of five healthy individuals (two the fact that the density of the eccrine sweat glands is highest women and three men, mean age 29.4 years), sweat samples were

www.jidonline.org 361 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

a kDa 0 24 48 72 (h) Semiquantitative SELDI analysis of sweat, nasal secretions, 6.5 tears, and saliva DCD The sweat samples were analyzed on a reversed-phase (H4) or a 3.5 weak cation exchange (WCX2) Protein Chip Array (Ciphergen Biosystems, Freemont, CA). Sweat samples were centrifuged 45 Albumin 1 minute at 13,000 r.p.m. (16,000 Â g) to remove particles or debris, 30 but were not sterile filtered. For reversed-phase (H4) analysis, 1 mlof the supernatant was added to 4 ml50mM sodium phosphate buffer b 3,000 4,000 5,000 6,000 m 4,816.1 (pH 6.5). Protein Chip Arrays were pretreated with 3 l of 50% (vol/ vol) acetonitrile/water for 1 minute, and after 2 minutes of drying, 40 4,529.0 4,604.1 1 ml of the diluted sweat solution was added. After drying (8 minutes), 2,406.5 20 0 hour spots were washed four times with 3 mlH2O (by flushing each spot 2,866.3 4,316.5 4,703.5 5,733.6 three times) and air-dried. For SELDI analysis with a weak cation 0 exchange chip (WCX2), 1 ml of the supernatant was added to 4 mlof M 4,317.7 50 m sodium acetate, pH 4.0 (binding buffer). Protein Chip Arrays 30 2,866.9 were equilibrated for 5 minutes with 3 mlof50mM sodium acetate (pH 4.0), and then the binding buffer was replaced by 1 ml of the 4,530.5 20 3,958.0 4 hours 4,417.4 2,579.3 4,115.6 diluted sweat sample. After 30 minutes of incubation in a humidity 10 5,733.6 chamber, spots were washed three times with 3 ml of the binding 0 buffer and once with 3 mlH2O, and air-dried.

4,313.4 After drying 1 ml of matrix consisting of a saturated solution of 40 sinapinic acid in 50% acetonitrile/water (vol/vol), 0.5% trifluoracetic 2,864.6 4,226.7 acid with 100 fmol/ml bovine insulin (MW 5,733.6 Da) was added. 16 hours 20 4,111.8 SELDI-TOF-MS analysis of each sweat sample was performed in 2,993.6 4,601.8 triplicate. Chips were read with the following instrument settings: 5,733.6 0 laser intensity 230, detector sensitivity 8, detector voltage 19.7 kV,

4,115.3 and positions 18–78 were read with an increment of 5, resulting in 10 3,957.6 4,317.3 12 different spot positions. Ten laser shots were collected on each 2,866.9 position (total shots collected and averaged: 120/sample), and two 40 hours 5 warming shots were fired at each position, which were not included 5,733.6 in the collection. The acquired mass range for the detection of DCD 0 peptides included 0–12,500Da with a focus mass at 4,800 Da. Furthermore, we performed several SELDI analyses from the Figure 6. Stability of DCD-derived peptides in eccrine sweat. (a) Western respective samples in which we focused on different molecular blot analysis of eccrine human sweat after incubation periods of 0, 24, 48, weight ranges, from 0 to 12,500 Da with a focus mass at 4,800 or and 72 hours at room temperature using the monoclonal antibody G-81, 10,000 Da and from 10–25kDa with a focus mass at 19 kDa. which recognizes C-terminal-derived DCD-peptides. DCD peptides were still detected in sweat after incubation of up to 72 hours at room temperature. Concentrating on molecular masses between 10 and 16 kDa, we Detection of human albumin (MW 67 kDa) was used as a loading control. (b) detected lysozyme, using SELDI-TOF-MS in tears, nasal secretions, SELDI-TOF-MS analysis of native human sweat after incubation periods of 0, and saliva, but not in human sweat. Calibration was carried out with 4, 16, and 40 hours at room temperature using a reversed-phase (H4) protein the bovine insulin peak (MW 5,733.6 Da). The spectra were chip. After baseline subtraction, spectra were normalized to the bovine analyzed by Protein Chip Software 3.1 (Ciphergen Biosystems, insulin peak (5,733.6 Da). DCD peptides are still detectable after 40 hours of Freemont, CA). After baseline subtraction, all spectra were normal- incubation at room temperature, yet a slow but continuous shift towards the ized to the internally obtained bovine insulin peak for semiquanti- processed DCD peptides LEK-43, LEK-42, LEK-41, and SSL-29 can be observed. fication. We did provide evidence that the semiquantification of DCD peptides using SELDI-TOF-MS is adequate since the results are taken at respective body sites (palmar sweat could only be obtained in accordance with the data we achieved with ELISA analysis of from two volunteers) and at three different time points (initial portion human sweat (Rieg et al., 2005). Previous analyses revealed a high of sweat of the forehead, after 20 and 40 minutes). Sweat was sensitivity of the SELDI-TOF-MS analysis with the detection of incubated at room temperature for up to 40 hours to analyze the synthetic DCD-1 peptide by SELDI-TOF-MS in concentrations as low kinetics of proteolytic degradation of DCD by RP-HPLC analysis as 0.27 mM (Flad et al., 2002). Samples of the other body fluids were and SELDI-TOF-MS. Saliva and nasal secretions of five healthy analyzed accordingly, using the reversed-phase chip (H4). individuals were collected and frozen immediately at 201C. Sampling of tears was carried out with the help of sliced onions, HPLC analysis of eccrine sweat and tears were also immediately frozen at 201C. The ethics Crude sweat was centrifuged 5 minutes at 13,000 r.p.m. committee of the medical faculty of Tu¨bingen University approved (16,000 Â g). Eight hundred microliters of the supernatant was all described studies, and participants gave written, informed applied to an RP-HPLC on a Nucleosil C18 column (150 Â 10 mm) consent. The study was conducted according to the Declaration of with 7 mm particles and a 100-A˚ pore size with a flow rate of 2.5 ml/ Helsinki Principles. minute. Solvent A was 0.055% aqueous trifluoroacetic acid and

362 Journal of Investigative Dermatology (2006), Volume 126 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

a 2,000 3,000 4,000 5,000 6,000 2,939.2 3,833.7 15 BS 2,846.6 2,724.5 2,875.1 3,720.2 2,752.8 2,911.2 10 2,382.3 2,410.2 3,512.73,637.2 4,019.7 5,733.6 5 3,442.0 0 2,939.3 40 SR

20 2,410.4 4,048.2 2,875.2 3,371.4 3,442.63,513.7 2,722.0 5,733.6 0 4,048.0 10 AV 2,875.1 2,939.3 5 5,733.6 2,410.2 2,722.5 3,748.7

0 b 5,232.0 15 BS 3,979.7 5,379.1 4,127.3 10 3,498.4 3,824.3 2,615.7 3,645.4 5,152.15,299.9 5 2,689.3 5,071.6 5,733.6 0 2,721.2 5,379.3 3 SR 5,733.6 2,938.4 4,126.9 2 3,970.8 5,298.7 1 0

7.5 5,379.6 TW 2,720.9 5 2,874.3 5,733.6 4,045.9 5,299.2 2.5 3,977.0 4,127.9 5,232.2

0 c 5,733.6 3 BS 2 2,866.9 3,407.5 3,337.1 5,319.1 5,675.8 1 5,009.4 0 3,407.0 3,336.7 SR 5,324.9 4 5,733.6 5,080.3 2 3,278.7 5,179.4 3,450.7 4,878.75,008.9 5,675.8

0 30 TW 3,442.0 3,370.7 20 3,486.5 10 4,434.8 5,060.6 5,214.2 5,370.7 5,733.6 0

Figure 7. Spectra of SELDI-TOF-MS analysis of (a) tears, (b) nasal fluid, and (c) saliva from three healthy volunteers. The a-defensins HNP-1 (3,442.1 Da), HNP-2 (3,371.0 Da), and HNP-3 (3,486.1 Da) are marked in a box. Shown are masses from 2,000 to 6,000 Da. After baseline subtraction, spectra were normalized to the bovine insulin peak (5,733.6 Da). solvent B was 80% acetonitrile in 0.05% aqueous trifluoroacetic zation time of flight (MALDI-TOF) analysis (G2025A, Hewlett-Packard, acid. A linear gradient of 5% B to 80% B over 40 minutes was used. Waldbronn, Germany) and N-terminal Edman microsequencing The resulting elution peaks were collected and lyophilized. DCD- with an Applied Biosystems 494 protein sequencer (Applied derived peptides were identified by matrix-assisted laser desorption ioni- Biosystems, Weiterstadt, Germany).

www.jidonline.org 363 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

Western blot analysis of eccrine sweat Flad T, Bogumil R, Tolson J, Schittek B, Garbe C, Deeg M et al. (2002) After physical exercise, sweat of the forehead was collected and Detection of dermcidin-derived peptides in sweat by ProteinChip technology. J Immunol Methods 270:53–62 analyzed by Western blot. Sweat was centrifuged 3 minutes at Fraki JE, Jansen CT, Hopsu-Havu VK (1970) Human sweat kallikrein. 13,000 r.p.m. (16,000 Â g) to remove particles and frozen either Biochemical demonstration and chromatographic separation from immediately (0 hour) or after incubation at room temperature (24, several other esteropeptidases in the sweat. Acta Dermatol Venereol 48, and 72 hours). Fifteen microliters of the samples were separated 50:321–6 by SDS-PAGE (using 15% acrylamide with 0.05% bisacrylamide), Fraki JE (1976) Human skin proteases. Separation and characterization of two transferred electrophoretically (1,200 mA/h) onto polyvinylidene acid proteases resembling cathepsin B1 and cathepsin D and of an inhibitor of cathepsin B1. Arch Dermatol Res 255:317–30 difluoride membrane sheets (Immobilon-PSQ, Millipore, Germany) Frohlich E, Schaumburg-Lever G, Klessen C (1995) Immunocytochemical and with a tank type blotter, and then blocked in 10% nonfat milk in PBS immunoelectron microscopic demonstration of cathepsin B in human (pH 7.4) for 30 minutes at room temperature. After incubation with malignant melanoma. Br J Dermatol 132:867–75 the primary antibodies overnight at 41C, the membrane sheets were Frohm M, Agerberth B, Ahangari G, Stahle-Backdahl M, Liden S, Wigzell H washed with phosphate buffer and incubated with the respective et al. (1997) The expression of the gene coding for the antibacterial secondary biotin-conjugated polyclonal antibodies (1:1,500 in peptide LL-37 is induced in human keratinocytes during inflammatory disorders. J Biol Chem 272:15258–63 blocking solution). After washing in phosphate buffer, the strepta- Fulton C, Anderson GM, Zasloff M, Bull R, Quinn AG (1997) Expression of vidin–AP conjugate (Roche, Mannheim, Germany) was used for the natural peptide antibiotics in human skin. Lancet 350:1750–1 detection of biotin-labeled secondary antibody. The membrane was Glaser R, Harder J, Lange H, Bartels J, Christophers E, Schroder JM (2005) immersed in CDP-Star solution (Western Lightning Chemilumines- Antimicrobial psoriasin (S100A7) protects human skin from Escherichia cence Reagents for AP; Roche, Mannheim, Germany) for 10 minutes, coli infection. Nat Immunol 6:57–64 and then exposed to X-ray film (Eastman Kodak, Rochester, NY). Harder J, Bartels J, Christophers E, Schroder JM (1997) A peptide antibiotic Western blot analysis was performed using the following primary from human skin. Nature 387:861 antibodies: monoclonal mouse anti-DCD antibody (G-81; 1:2,000 in Harder J, Bartels J, Christophers E, Schroder JM (2001) Isolation and characterization of human beta-defensin-3, a novel human inducible blocking solution; Sagawa et al., 2003), rabbit anti-HNPs polyclonal peptide antibiotic. J Biol Chem 276:5707–13 antiserum (1:1,000 dilution, kindly provided by H Kalbacher, Harder J, Schroder JM (2002) RNase 7, a novel innate immune defense Tu¨bingen, Germany), goat anti-human HBD-2 polyclonal antiserum antimicrobial protein of healthy human skin. J Biol Chem 277:46779–84 (1:1,000 dilution, Cell Concepts, Umkirch, Germany), mouse anti- Hibino T, Takemura T, Sato K (1994) Human eccrine sweat contains tissue human lysozyme monoclonal antibody (1:200 dilution in 4% BSA in kallikrein and kininase II. J Invest Dermatol 102:214–20 phosphate buffer; Acris, Hiddenhausen, Germany), and mouse anti- Komatsu N, Takata M, Otsuki N, Toyama T, Ohka R, Takehara K et al. (2003) human albumin monoclonal antibody (1:1,000 dilution; Sigma). The Expression and localization of tissue kallikrein mRNAs in human epidermis and appendages. J Invest Dermatol 121:542–9 peptides LL-37, DCD-1L, and a-andb-defensins were either purchased Lai YP, Peng YF, Zuo Y, Li J, Huang J, Wang LF et al. (2005) Functional and from Sigma or kindly provided by H Kalbacher (Tubingen, Germany). ¨ structural characterization of recombinant dermcidin-1L, a human antimicrobial peptide. Biochem Biophys Res Commun 328:243–50 Antimicrobial assays Minami Y, Mede K, Furukawa F, Sagawa K, Kimura A, Tsuji T (2004) Antimicrobial assays were performed using the colony-forming units Cutaneous mixed tumors: an immunohistochemical study using two 0 (CFU) assay as described previously (Rieg et al., 2005). S. aureus antibodies, G –81 and C811443B. J Dermatol Sci 36:180–2 (ATCC 25923) cultures were grown to mid-exponential growth Murakami M, Lopez-Garcia B, Braff M, Dorschner RA, Gallo RL (2004) Postsecretory processing generates multiple cathelicidins for enhanced phase and washed twice with 10 mM sodium phosphate buffer (pH topical antimicrobial defense. J Immunol 172:3070–7 7.0). Bacterial concentration was estimated photometrically at 7 Murakami M, Ohtake T, Dorschner RA, Schittek B, Garbe C, Gallo RL (2002) 600 nm. Absorbance of 1.0 corresponded to 1.97 Â 10 cells of Cathelicidin anti-microbial peptide expression in sweat, an innate S. aureus. After dilution to a concentration of 106 CFU/ml, 10 mlof defense system for the skin. J Invest Dermatol 119:1090–5 the dilutions was incubated at 371C for 2 hours with HPLC fractions Rieg S, Garbe C, Sauer B, Kalbacher H, Schittek B (2004) Dermcidin is of human sweat of four healthy volunteers (2 ml sweat of each was constitutively produced by eccrine sweat glands and is not induced in epidermal cells under inflammatory skin conditions. Br J Dermatol fractionated) in a total volume of 30 mlin10mM sodium phosphate 151:534–9 buffer (pH 7.0)/10 mM sodium chloride. After incubation, cells were Rieg S, Steffen H, Seeber S, Humeny A, Kalbacher H, Dietz K et al. (2005) diluted 1:100 in 10 mM sodium phosphate buffer (pH 7.0) and 90 mlof Deficiency of dermcidin-derived antimicrobial peptides in sweat of the diluted bacterial suspension was plated in triplicate on blood agar. patients with atopic dermatitis correlates with an impaired innate defense of human skin in vivo. J. Immunol 174:8003–10 Sadowski T, Dietrich S, Muller M, Havlickova B, Schunck M, Proksch E et al. CONFLICT OF INTEREST (2003) Matrix metalloproteinase-19 expression in normal and diseased The authors state no conflict of interest. skin: dysregulation by epidermal proliferation. J Invest Dermatol 121:989–96 ACKNOWLEDGMENTS Sagawa K, Kimura A, Saito Y, Inoue H, Yasuda S, Nosaka M et al. (2003) This work was supported by a grant from the University of Tu¨bingen (fortu¨ne Production and characterization of a monoclonal antibody for sweat- 1182-0-0) and Deutsche Forschungsgemeinschaft (DFG SCHI 510/3-1 and 3-2). specific protein and its application for sweat identification. Int J Legal Med 117:90–5 Sato K, Kang WH, Saga K, Sato KT (1989) Biology of sweat glands and their REFERENCES disorders. I. Normal sweat gland function. J Am Acad Dermatol 20:537–63 Ali RS, Falconer A, Ikram M, Bissett CE, Cerio R, Quinn AG (2001) Expression Schittek B, Hipfel R, Sauer B, Bauer J, Kalbacher H, Stevanovic S et al. (2001) of the peptide antibiotics human beta defensin-1 and human beta Dermcidin: a novel human antibiotic peptide secreted by sweat glands. defensin-2 in normal human skin. J Invest Dermatol 117:106–11 Nat Immunol 2:1133–7

364 Journal of Investigative Dermatology (2006), Volume 126 S Rieg et al. Dermcidin-Derived Antimicrobial Peptides in Sweat

Sieprawska-Lupa M, Mydel P, Krawczyk K, Wojcik K, Puklo M, Lupa B Wiedow O, Harder J, Bartels J, Streit V, Christophers E (1998) Antileukopro- et al. (2004) Degradation of human antimicrobial peptide LL-37 by tease in human skin: an antibiotic peptide constitutively produced by Staphylococcus aureus-derived proteinases. Antimicrob Agents Chemother keratinocytes. Biochem Biophys Res Commun 248:904–9 48:4673–9 Zeeuwen PL, Van Vlijmen-Willems IM, Jansen BJ, Sotiropoulou G, Curfs JH, Vuong C, Voyich JM, Fischer ER, Braughton KR, Whitney AR, DeLeo FR et al. Meis JF et al. (2001) Cystatin M/E expression is restricted to differentiated (2004) Polysaccharide intercellular adhesin (PIA) protects Staphylococ- epidermal keratinocytes and sweat glands: a new skin-specific protei- cus epidermidis against major components of the human innate immune nase inhibitor that is a target for cross-linking by transglutaminase. system. Cell Microbiol 6:269–75 J Invest Dermatol 116:693–701