Peptides 32 (2011) 845–851

Contents lists available at ScienceDirect

Peptides journal homepage: www.elsevier.com/locate/peptides

Anti-Legionella activity of staphylococcal hemolytic peptides
A. Marchand, J. Verdon, C. Lacombe, S. Crapart, Y. Héchard, J.M. Berjeaud ∗
Université de Poitiers, Laboratoire de Chimie et Microbiologie de l’Eau – UMR 6008 CNRS, IBMIG – UFR Sciences Fondamentales et Appliquées, 1 rue du Georges Bonnet, 86022 Poitiers Cedex, France

a r t i c l e

i n f o

a b s t r a c t
A collection of various Staphylococci was screened for their anti-Legionella activity. Nine of the tested strains were found to secrete anti-Legionella compounds. The culture supernatants of the strains, described in the literature to produce hemolytic peptides, were successfully submitted to a two step purification process. All the purified compounds, except one, corresponded to previously described hemolytic peptides and were not known for their anti-Legionella activity. By comparison of the minimal inhibitory concentrations, minimal permeabilization concentrations, decrease in the number of cultivable bacteria, hemolytic activity and selectivity, the purified peptides could be separated in two groups. First group, with warnericin RK as a leader, corresponds to the more hemolytic and bactericidal peptides. The peptides of the second group, represented by the PSM from Staphylococcus epidermidis, appeared bacteriostatic and poorly hemolytic. © 2011 Elsevier Inc. All rights reserved.

Article history: Received 15 December 2010 Received in revised form 19 January 2011 Accepted 19 January 2011 Available online 1 February 2011 Keywords: Legionella pneumophila Staphylococci Hemolytic peptides

1. Introduction Legionella pneumophila is a waterborne pathogenic bacterium responsible for severe pneumonia called Legionnaire’s disease [6,17]. In the environment, L. pneumophila is ubiquitously found in fresh water and could survive within biofilms and free-living amoeba [2]. L. pneumophila could be found at high level in manmade water systems such as air conditioning and cooling towers, spas. These systems are mainly responsible for outbreaks as they might produce contaminated water droplets, which are inhaled by people. L. pneumophila reaches the lungs and multiplies within macrophages [15]. In 2006, there were more than 6000 reported cases in Europe leading to 400 deaths ca. Concerning L. pneumophila growth in water systems, various methods of control have been used: chlorine, monochloramine, heat. However, these treatments are not fully efficient and, after a lag period following the treatment, L. pneumophila might be able to recolonize the system [19]. Therefore, it is of primary importance to find new treatments to restrain L. pneumophila growth. Recently, it was shown that hemolytic peptides secreted by Staphylococcus warneri RK inhibited the growth of Legionella spp. [8]. Beside an original 22 amino acids long peptide named warnericin RK, two delta-hemolysins were characterized [20]. Hemolysins are among the most important virulence factors of Staphylococci. Staphylococcus aureus, which is a potentially pathogenic coagulase-positive specie of this genus, produces four

∗ Corresponding author. Tel.: +33 5 49 45 40 06; fax: +33 5 49 45 35 03. E-mail address: jean-marc.berjeaud@univ-poitiers.fr (J.M. Berjeaud). 0196-9781/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.peptides.2011.01.025

hemolysins – , , and [26]. -Hemolysin, also named -toxin or -lysin, is a well known peptide, which mechanism of action has been described [4,22]. -Hemolysins were found in most species of coagulase-negative Staphylococci and some sequences (Table 1) were characterized for S. intermedius [9], S. epidermidis [12], S. simulans and S. warneri [18]. Numerous hemolytic peptides (Table 1) with various other activities were found to be secreted by different Staphylococci. Thus S. epidermidis was described to produce a complex of peptides, named phenol-soluble modulin (PSM) [13], combining proinflammatory activity and a role in the detachment of biofilm [23,27]. The PSM complex is composed of three peptides of 22, 44 and 25 amino acids named PSM , PSM and PSM , respectively. PSM corresponds to the S. epidermidis -hemolysin [13]. Genes encoding PSM related peptides are present in all sequenced S. aureus strains. The production of the peptides, PSM , PSM and -toxin, was shown to be higher in the community-associated methicillin-resistant S. aureus (CA-MRSA) strains, than in the hospital associated strains (HA-MRSA), and contributes significantly to their virulence [24]. S. haemolyticus secretes three hemolytic peptides, which inhibit the growth of Neisseria gonorrhoeae. They were named gonococcal growth inhibitor (GGI) I, II and III and consist of 44 amino acids linear peptides with high sequence homologies (65–75%) [7,25]. The SLUSH A, B and C peptides, each composed of 43 amino acids with closely related sequences (>76%), are secreted by S. lugdunensis. Their sequences are clearly distinct from known hemolysins but are similar to the S. haemolyticus GGIs [5]. From the three hemolytic peptides secreted from S. cohnii, H1 and H3 are composed of 43 amino acids and resemble to SLUSH peptides, whereas the shorter H2, containing only 18 amino acids,

846

A. Marchand et al. / Peptides 32 (2011) 845–851

Table 1 Amino acid sequences of the hemolytic peptides produced by Staphylococci described in the literature. Peptide name (producer species) Warnericin RK (S. warneri) Hemolysin-H2 (S. cohnii) PSM- 1 (S. aureus) PSM- 2 (S. aureus) PSM- 3 (S. aureus) PSM- 4 (S. aureus) PSM- (S. epidermidis) PSM- (S. epidermidis) -Hemolysin-I (S. warneri) -Hemolysin-II (S. warneri) -Hemolysin-I (S. aureus) -Hemolysin (S. epidermidis) -Hemolysin (S. intermedius) -Hemolysin (S. simulans) PSM- (S. epidermidis) GGI I (S. haemolyticus) GGI II (S. haemolyticus) GGI III (S. haemolyticus) SLUSH-A (S. lugdunensis) SLUSH-B (S. lugdunensis) SLUSH-C (S. lugdunensis) Hemolysin-H1 (S. cohnii) Hemolysin-H3 (S. cohnii) Sequence MQFITDLIKKAVDFFKGLFGNK MDFIIDIIKKIVGLFTGK MGIIAGIIKVIKSLIEQFTGK MGIIAGIIKFIKGLIEKFTGK MEFVAKLFKFFKDLLGKFLGNN MAIVGTIIKIIKAIIDIFAK MADVIAKIVEIVKGLIDQFTQK MSIVSTIIEVVKTIVDIVKKFKK MAADIISTIGDLVKLIINTVKKFQK MTADIISTIGDFVKWILDTVKKFTK MAQDIISTIGDLVKWIIDTVNKFTKK MMAADIISTIGDLVKWIIDTVNKFKK MAGDIISTIVDFIKLIAETVKKFTK MAGDIVGTIGEFVKLIIETVQKFTQK MSKLAEAIANTVKAAQDQDWTKLGTSIVDIVESGVSVLGKIFGF MQKLAEAIAAAVSAGQDKDWGKMGTSIVGIVENGITVLGKIFGF MEKIANAVKSAIEAGQNQDWTKLGTSILDIVSNGVTELSKIFGF MSKLVQAISDAVQAQQNQDWAKLGTSIVGIVENGVGILGKLFGF MSGIVDAITKAVQAGLDKDWATMATSIADAIAKGVDFIAGFFN MSGIIEAITKAVQAGLDKDWATMGTSIAEALAKGIDAISGLFG MDGIFEAISKAVQAGLDKDWATMGTSIAEALAKGVDFIIGLFH MSGIVEAISNAVKSGLDHDWVNMGTSIADVVAKGADFIAGFFS MSDFVNAISEAVKAGLSADWVTMGTSIADALAKGADFILGFFN Reference Verdon et al. [20] Mak et al. [11] Wang et al. [24]

Mehlin et al. [13] Tegmark et al. [18] Wiseman [26] McKevitt et al. [12] Ji et al. [9] Tegmark et al. [18] Mehlin et al. [13] Watson et al. [25]

Donvito et al. [5]

Mak et al. [11]

appears unique [11]. These peptides, whose sequences are presented in Table 1, could be separated in two groups according to their size, from 18 to 26 amino acids for the small peptides, and from 43 to 44 for the long peptides. Nevertheless, like warnericin RK and S. warneri -hemolysins, all these peptides are hydrophobic, induce lysis of erythrocytes but are generally described to be non-active against bacteria, except for GGI peptides active against gonococci. Because warnericin RK presents a high hemolytic activity, which is a major drawback for its therapeutic potency, we decided to test if other Staphylococci could produce anti-Legionella peptides with the final objective to detect a peptide with no or a low hemolytic activity. In this paper we report the selection of Staphylococci producing peptides inhibiting the growth of Legionella. The antibacterial peptides were purified and identified and their activities towards Legionella as well as erythrocytes were measured. Based on the comparison of the activities of the purified peptides, as well as warnericin RK and -hemolysin of S. warneri, two classes of antiLegionella peptides were proposed. 2. Materials and methods 2.1. Bacterial strains, culture media and growth conditions The Staphylococcus strains were grown in Brain Heart Infusion (BHI, Difco) liquid medium at 37 ◦ C for 18–20 h under shaking (250 rpm). L. pneumophila Lens was grown at 37 ◦ C either on buffered charcoal yeast extract (BCYE) agar plates for 96 h or in buffered yeast extract (BYE) liquid medium for 24–30 h under shaking (150 rpm). 2.2. Purification of anti-Legionella peptides In order to characterize anti-Legionella peptides, a two-step purification procedure was conducted as described below. Cells were removed, from an overnight staphylococcal culture, by centrifugation (9000 × g, 15 min, 4 ◦ C) and the supernatant was heated at 70 ◦ C for 15 min. The resulting sample was named crude extract (CE). Chromatographic steps were conducted on a Dionex 3600 HPLC instrument composed of a Dionex P680 HPLC binary pump, an iso-

cratic Dionex UltiMate 3000 pump used as a loading pump and a Dionex UltiMate 3000 UV detector. Firstly, the CE was loaded onto a hydrophobic interaction column (POROS 20 HP2, PerSeptive Biosystems, 4.6 mm × 80 mm). Elution was monitored at 220 nm and 280 nm and carried out, at a flow rate of 2 mL/min, using a water/acetonitrile/trifluoroacetic acid 0.05% (v/v) gradient. After washing for 5 min with 10% acetonitrile, elution was achieved in 25 min with a 20 min linear gradient from 10 to 100% acetonitrile, followed by a 5 min wash with 100% acetonitrile. The collected active fraction was lyophilized. Secondly, the lyophilized fraction was dissolved in 900 L of 50% aqueous acetonitrile and injected onto a Kromasil C8 reverse-phase ˚ HPLC analytical column (5 M, 100 A, 4.6 mm × 250 mm). Separation was carried out using a water/acetonitrile/trifluoroacetic acid 0.05% (v/v) solvent system. After an initial 5 min wash with 50% or 70% (according to the staphylococcal strain) acetonitrile, elution was achieved in 40 min at a flow rate of 0.8 mL/min with a 30 min linear gradient from 50 or 70 to 100% acetonitrile, followed by a 10 min wash with 100% acetonitrile. Each collected fraction was lyophilized and stored at −20 ◦ C for further studies. Synthetic non-formylated peptides, warnericin RK, and H2U, were purchased from Genscript (Piscataway, USA). PSM was provided by Dr. T. Jouenne from the PBS laboratory (CNRS UMR6270, Rouen, France). 2.3. Mass spectrometry For molecular weight determination, the HPLC fractions were analyzed by mass spectrometry, on a Perkin Elmer Sciex API 165 mass spectrometer equipped with an ion-spray source. Each sample was resuspended in 50% acetonitrile/0.1% formic acid and was analyzed by infusion at a flow rate of 5 L/min. The instrument scale for the mass-to-charge (m/z) ratio was calibrated with the ions of the ammonium adduct of polypropylene glycol. Scan data were obtained with LC2-Tune and mass calculation was done with Biomultiview 1.2 (Software package Sciex). 2.4. Peptide titration Peptide concentration was determined by the bicinchoninic acid assay as described by the supplier (Sigma) with bovine serum albumin as a standard. 25 L of each sample were mixed with 200 L of

A. Marchand et al. / Peptides 32 (2011) 845–851

847

a solution of bicinchoninic acid and copper sulfate 50:1 (v/v). The preparation was incubated at 37 ◦ C for 30 min and the absorbance at 595 nm was measured by a microtiter plate reader OPSYS MR (Thermo Labsystems). 2.5. Anti-Legionella activity assays 2.5.1. Spot on lawn assays Anti-Legionella activity was determined using spot on lawn assays as described below. A 107 colony forming units (CFU)/mL suspension of L. pneumophila was spread onto a BCYE agar plate. Samples to be tested (5 L of staphylococcal culture or 50 L of staphylococcal CE) were spotted onto the surface of the agar plate. The plate was then incubated at 37 ◦ C for 96 h. Anti-Legionella activity was revealed by a zone of inhibition around the test samples. Quantitative analysis was carried out by measuring diameters of the zone of inhibition. 2.5.2. Minimal inhibitory concentration Minimal inhibitory concentration (MIC) of the different peptides was determined using microtiter plate assays as described previously [20]. MIC represents the lowest concentration of peptide which totally inhibits the growth of L. pneumophila. 2.5.3. Membrane permeabilization assays and cultivability evaluation Exponentially growing L. pneumophila (OD600 = 0.4–0.8) were washed and resuspended in BYE at a concentration of 106 CFU/mL. The bacterial suspension was treated with various concentrations, ranging from 0 to 10 M, of peptides during 45 min at 37 ◦ C. The generation time, in these conditions, was evaluated and corresponded to 2.5 h (data not shown) (i) One half of the cell suspensions were then analyzed using a flow cytometric approach after bacterial staining with a couple of fluorochromes (SYTO9 and propidium iodide (PI)), as described previously [21]. Flow cytometric measurements were performed on a FACSCanto II flow cytometer (BD Biosciences, Le Pont de Claix, France) with a 488 nm argon excitation laser. A total of 50,000 events were analyzed in each sample, using BD FACSDiVa 6 software (BD Biosciences) for data acquisition and analysis. Optical filters were set up such that PI fluorescence was measured at 670 LP nm and SYTO 9 fluorescence was measured at 530/30 BP nm. A SYTO 9+/PI-gate was drawn to determine the percentage of non permeabilized cells in the control without peptide. The minimal permeabilization concentration (MPC) of peptides corresponds to the concentration leading to 90% of permeabilization of L. pneumophila. (ii) The other half of the bacteria suspensions were diluted to 1/10, 1/100 or 1/1000 depending on the peptide and then spread on a BCYE agar plate using an automatic spiral plater (Whitley Automatic Spiral Plater, Don Whitley Scientific Ltd). Plates were incubated 96 h at 37 ◦ C before colonies numeration using tables provided by the supplier. The results were expressed by the decrease in the L. pneumophila cultivability estimated at the maximum peptide concentration. 2.6. Hemolytic activity assays Hemolytic activity of the peptides was determined by measuring the released hemoglobin from human erythrocytes as described previously [20]. 100% hemolysis was given by adding 0.1% Triton X 100 to the reaction mixture instead of peptide solutions.

Table 2 Anti-Legionella activity of 15 staphylococcal strains and culture supernatants. Staphylococcal strain Anti-Legionella activitya Culture S. warneri RK S. aureus 2850 S. lugdunensis 967 S. saprophyticus 715 S. hominis 373 S. epidermidis 567 S. xylosus 700404 S. cohnii 898 S. haemolyticus 2259 S. lentus 982 S. equorum 4057C S. simulans 4334 S. chromogenes AM1 S. carnosus 4251 S. caprae 2534D + ++ + +++ +++ + ++ +++ +++ ++ + + − − − Crude extract ++ +++ +++ + ++ +++ + + ++ − − − − − − Hechard et al. [8] Laboratory collection Donvito et al. [5] IMI collection IMI collection IMI collection ATCC IMI collection Laboratory collection IMI collection Laboratory collection Laboratory collection Laboratory collection Laboratory collection Laboratory collection Source/referenceb

a “+++” corresponds to an inhibition zone with a diameter superior to 1.1 cm, “++” to a diameter between 0.8 cm and 1.1 cm and “+” to a diameter between 0.6 cm and 0.8 cm. “−” was affected when no inhibition zone was observed. b IMI: Institut für Molekular Infektionsbiologie, Universität Würzburg, Germany.

3. Results 3.1. Screening Staphylococci for the anti-Legionella activity Staphylococcus strains representative of different species, virulent or not, are listed in Table 2. In a first step, anti-Legionella activity of the selected strains was checked by spotting colonies on an agar medium seeded with L. pneumophila Lens. Inhibitory activity was detected when a clear zone appeared around the colonies after the growth of Legionella. From the 15 selected strains, 12 displayed an inhibition zone (Table 2). In order to verify the relationship of this apparent anti-Legionella activity and the secretion of an antibacterial factor, the cell-free culture supernatants, named crude extracts (CEs), of the strains were assayed against L. pneumophila Lens. Nine from the 15 crude extracts displayed activity halos on the agar test revealing the action of secreted anti-Legionella factors (Table 2). From the 15 tested strains, 12 Staphylococci showed inhibition areas around the colonies but only 9 displayed activity in their culture supernatant. For the non active crude extracts prepared from the 3 strains which showed inhibition halos around colonies, the concentration of the inhibiting factor in the supernatant could be too small to be detected. Indeed, as indicated in Table 2 diameters of the inhibition zones varied from a crude extract to another, showing either that the amount of the antibacterial agents varied depending on the species or that the various species produced different factors that differed in their specific activity. Alternatively, the secreted antimicrobials could be denatured during the incubation at 70 ◦ C of the crude extracts. This step was conducted in order to inhibit proteases. In this case, the antimicrobial agents could be proteins but not peptides. This could be verified by using protease inhibitors instead of heat treatment for example. Alternatively, the antimicrobials could be directly subjected to chromatographic separation without heat treatment in order to obtain antibacterial agents separated from proteases. Almost all the peptides described in the literature (Table 1), secreted by the Staphylococci species selected in our study, were successfully purified by the two steps process developed in this work. However, the SLUSH B from S. lugdunensis, as well as hemolysin I from S. warneri RK were detected by mass spectrometry but in too low amounts to be useful for activity assays. In the same way, GGI III from S. haemolyticus was never detected. Our S. aureus strain did not produce any detectable PSM, as it was demonstrated for hospital associated strains contrarily to the community asso-

848

A. Marchand et al. / Peptides 32 (2011) 845–851

ciated strains [24]. Finally, no hemolytic peptide was recovered, using our methodology, from the culture supernatant of the S. cohnii strain. This could be due to the strain used, which either do not produce these particular peptides or in too weak amounts. Indeed, the anti-Legionella activity of the crude extract prepared from this strain was low (Table 2). S. epidermidis, S. haemolyticus, S. lugdunensis and S. aureus, known to produce hemolytic peptides, were of particular interest because they appeared as active as S. warneri RK. 3.2. Purification and identification of anti-Legionella peptides Because we postulated that the anti-Legionella factors were proteinaceous, as found for S. warneri RK, the process was first optimized for the purification of warnericin RK and S. warneri hemolysins I and II. The purification method of these peptides previously described and based on the affinity chromatography on hydroxyapatite was time consuming [20]. Thus, taking account of the procedure described for the purification of the -toxin of S. epidermidis [16], we developed a two steps chromatographic strategy. Briefly, crude extracts (CEs) from all the tested Staphylococci were directly applied on a hydrophobic interaction phase column. The adsorbed anti-Legionella peptides were then eluted from the column by using a rapid gradient of acetonitrile. The fractions displaying an anti-Legionella activity were then injected on an analytical C8 HPLC column. The chromatograms obtained from the CE of S. warneri RK (Fig. 1A) were quite different than the one previously described [20]. Particularly, the relative intensities of the peptides peaks were different. Moreover, all the different antiLegionella peptides previously described were found but only in their N-formyl forms. The same protocol was applied to other antiLegionella CE and the observed HPLC profiles varied depending on the strains (Fig. 1). However in all cases, at least one major peak was detected. The corresponding fractions were tested for antiLegionella activity and analyzed by electrospray ionization mass spectrometry (ESI-MS). Molecular masses of the analyzed peptides were compared to those of peptides known to be secreted by the corresponding Staphylococci. All the peptides were thus identified except one, which was named Haemo 3. This peptide, produced by S. haemolyticus 2259, has a molecular mass of 2259.78 Da. Interestingly, all the purified peptides were formylated on their N-terminal methionine. We failed to purify any peptide from the CE of S. cohnii 898. However, the S. cohnii peptide H2U sequence displayed similarities with the warnericin RK one, thus we decided to acquire the non-formylated synthetic peptide. In the same way, in order to evaluate the impact of the formylation on the peptides activities, synthetic warnericin RK and S. epidermidis PSM were tested. 3.3. Hemolytic and anti-Legionella activities of the peptides The purified peptides were assayed for anti-Legionella and hemolytic activities. The antibacterial activity was first determined by measuring the minimal inhibitory concentration (MIC) which corresponds to the capacity of a peptide to inhibit the growth of Legionella (Table 3). Among all the tested peptides, the formylated forms of warnericin RK, -hemolysins I and II from S. warneri and PSM from S. epidermidis showed the highest inhibitory activities towards Legionella (MIC < 1 M). On the other hand, the SLUSH A and C from S. lugdunensis, and GGI II from S. haemolyticus displayed the lowest inhibitory activity (MIC > 5 M). Interestingly, all the latter peptides displayed a longer amino-acids sequence as compared to the first peptides group. Besides, analysis of the growth of L. pneumophila as a function of the concentration of added peptide, showed two opposite types of observed curve shape. In the first case (Fig. 2A), with nonformylated warnericin RK, the peptide seemed to act through an “all or nothing” mechanism with no inhibition of L. pneumophila growth above the MIC. In the second case (Fig. 2B), with non-formylated PSM , the inhibition seemed “progressive” and dependent of the peptide dose. In an attempt to explain this result, we hypothesized that the different shapes would be related to different modes of action. So, in order to verify such hypothesis, the minimal permeabilization concentration (MPC) of peptides, which corresponds to the concentration leading to 90% of permeabilization of the target bacteria, was determined (Table 3). Briefly, the percentage of permeabilized cells, corresponding to those stained with propidium iodide (IP), was measured using flow cytometry. As pointed out for the MIC, the formylated forms of the peptides produced by S.

Fig. 1. Reverse-phase elution profiles at 220-nm of active fractions obtained from hydrophobic interaction chromatography from Staphylococcus crude extracts. (A) S. lugdunensis 967, (B) S. aureus 2850, (C) S. epidermidis 567, (D) S. haemolyticus 2259 and (E) S. warneri RK.

A. Marchand et al. / Peptides 32 (2011) 845–851 Table 3 Activities of peptides towards L. pneumophila and red blood cells. Peptidea MIC curve shape M Group 1 f-Warnericin RK f- -hemolysin II f- -hemolysinb -hemolysin II Warnericin RK f-Ggi I f-SLUSH C f-SLUSH A f-PSM Haemo 3 f- -hemolysinc PSM f-PSM H2U f-Ggi II 0.30 0.54 1.05 1.09 1.22 4.15 5.16 11.28 0.3 < MIC < 12 0.63 1.38 1.59 1.90 2.69 3.04 13.23 0.6 < MIC < 14 AoN AoN AoN AoN AoN Pro Pro Pro AoN AoN Pro Pro Pro Pro Pro MIC MPC M 0.6 0.5 0.3 0.4 1.22 1.37 0.7 2.0 11.0 >8.4 >10.1 >5.4 4.5 >5.2 >3 Bacterial cultivability decrease log 1.0 1.7 1.7 1.3 3.1 0.3 1.2 1.0 >0.9 0.3 0.1 0.1 0.7 0.5 0.6 0.0