Biochimica et Biophysica Acta 1758 (2006) 1713 1722 www.elsevier.

com/locate/bbamem

WLIP and tolaasin I, lipodepsipeptides from Pseudomonas reactans and Pseudomonas tolaasii, permeabilise model membranes
M. Coraiola a,, P. Lo Cantore b , S. Lazzaroni c , A. Evidente c , N.S. Iacobellis b , M. Dalla Serra a
b

ITC-CNR Istituto di Biofisica, via Sommarive 18, I-38050 Povo (Trento), Italy Dipartimento di Biologia, Difesa e Biotecnologie Agro Forestali, Universit degli Studi della Basilicata, viale dell'Ateneo Lucano 10, 85100 Potenza, Italy c Dipartimento di Scienze del Suolo, della Pianta e dellAmbiente, Universit di Napoli Federico II, via Universit 100, 85155 Portici (Napoli), Italy Received 17 February 2006; received in revised form 18 May 2006; accepted 30 June 2006 Available online 14 July 2006

Abstract The activity of the White Line Inducing Principle (WLIP) and tolaasin I, produced by virulent strains of Pseudomonas reactans and Pseudomonas tolaasii, respectively, was comparatively evaluated on lipid membranes. Both lipodepsipeptides were able to induce the release of calcein from large unilamellar vesicles. Their activity was dependent on the toxin concentration and liposome composition and in particular it increased with the sphingomyelin content of the membrane. Studies of dynamic light scattering suggested a detergent-like activity for WLIP at high concentration (> 27 M). This effect was not detected for tolaasin I at the concentrations tested (< 28 M). Differences were also observed in lipodepsipeptides secondary structure. In particular, the conformation of the smaller WLIP changed slightly when it passed from the buffer solution to the lipid environment. On the contrary, we observed a valuable increment in the helical content of tolaasin I which was inserted in the membrane core and oriented parallel to the lipid acyl chains. 2006 Elsevier B.V. All rights reserved.
Keywords: Lipodepsipeptide; Pseudomonas reactans; Pseudomonas tolaasii; Liposome; Transmembrane pore; Secondary structure

1. Introduction Recent studies have demonstrated that brown blotch disease of Agaricus bisporus and yellowing of Pleurotus ostreatus are complex diseases [13]. Besides Pseudomonas tolaasii, known by long time as the causal agent of the brown blotch disease on A. bisporus [4], also other Pseudomonas spp., including P. reactans, cause the disease symptoms even though in a different way [57]. Furthermore, P. reactans has been recently demonstrated as the causal agent of the yellowing of Pleurotus eryngii [8]. P. reactans is a bacterial entity not yet classified, known as a saprophytic bacterium associated to the cultivated mushrooms and useful in the white line assay for the identification of P. tolaasii [9]: strains of P. reactans produce in culture an

This paper is dedicated to the memory of Gianfranco Menestrina. Corresponding author. Tel.: +39 0461 314157; fax: +39 0461 810628. E-mail address: manucora@itc.it (M. Coraiola).

extracellular substance, called the White Line Inducing Principle (WLIP), known for its ability to interact with tolaasin I, the main virulence factor of P. tolaasii, and to form a white precipitate in the agar [9]. The potential involvement of both the bacteria in the development of some diseases on fungi and the specific interaction occurring between their main secondary metabolites, prompted us to study these molecules in a comparative way. WLIP is a lipodepsipeptide (LDP) with a molecular weight of 1125 Da composed of a N-terminal hydroxydecanoic acid and a peptide moiety of nine aminoacids, six of which are in the D form (Fig. 1). The molecule contains a lactone ring between D-allo-threonine (D-Thr3) and the C-terminal L-isoleucine (Ile9) [10]. WLIP is structurally similar to another lipodepsipeptide called viscosin [11,12], except for the chirality of leucine in position 5 which has D configuration in WLIP and L configuration in viscosin. Also tolaasin I from P. tolaasii is a lipodepsipeptide with a molecular weight of 1985 Da composed of 18 aminoacids, eleven of which are in the D form. The primary structure (Fig. 1), elucidated by Nutkins et al. [13], bears a -hydroxyoctanoic

0005-2736/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2006.06.023

1714

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

Fig. 1. Primary structure of WLIP and Tolaasin I. Non standard amino acids indicated are as follows: aThr, allo threonine; But, 2,3-dehydro-2-aminobutyric acid; Hse, homoserine; Dab, 2,4-diaminobutyric acid.

acid moiety linked to the N-terminus, a sequence of seven successive D-amino acids in the N-terminal region of the peptide (Pro2Val8), with a SerLeuVal repeat, and then alternate L- and D-amino acids. It also contains a 2,3dehydro-2-aminobutyric acid (But) residue at positions 1 and 13, a D-homoserine (D-Hse16), and a D-2,4-diaminobutyric acid (D-Dab17). Finally, a lactone ring is formed between the hydroxyl of D-Thr14 and the C-terminal L-Lys18. In addition to tolaasin I and II [13] other five analogues, namely tolaasin AE, have been shown to be produced by strains of P. tolaasii. The primary structure and the antimicrobial activity of these new analogues have also been reported [14]. Tolaasin I is considered to be the main virulence factor of the bacterium and appeared to cause the mushroom tissue alteration by disrupting the cell membrane through the formation of transmembrane pores [1519]. In comparison to other lipodepsipeptides produced by phytopathogenic or mycopathogenic

bacteria [20] the biological properties of WLIP and its possible role in the disease process are less defined. The novelty of the pathogenetic capability of P. reactans and particularly the fact that avirulent variants of the pathogen have also lost the ability to produce WLIP [3] suggest that this metabolite may be important in the P. reactansmushrooms interactions. However, even though the haemolytic and the antimicrobial activity of WLIP and its interactions with synthetic membranes have been recently demonstrated [2,3,8], the mechanism involved is not yet defined. Here, the effects of WLIP directly on model membranes of different lipid composition, also in comparison to tolaasin I, are reported. Furthermore, data on the LDPs mode of action and the molecularity of the permeabilising unit are analysed. In addition, the changes in the secondary structure occurring when the above molecules passed from the aqueous buffer to the membrane are reported. In particular, it was possible to analyse the conformation of WLIP and tolaasin I during their

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

1715

interaction with the lipid membrane, i.e. their natural target, and, in the case of tolaasin I, to calculate the molecule orientation with respect to the lipid acyl chains.
2. Materials and methods 2.1. Chemicals
Lipid used were egg phosphatidylcholine (PC), 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC) and sphingomyelin (SM) from Avanti Polar Lipids (Alabaster, AL, USA), ergosterol (Erg) from Sigma, cholesterol (Chol) from Fluka. Calcein, EDTA, and Sephadex G-50 were from Sigma, Triton X-100 from Merck. In FTIR experiments solvents used were trifluoroethanol (TFE) from Sigma, dimethyl sulfoxide (DMSO) and hexafluoroisopropanol (HFIP) from Riedel-de Han, and sodium dodecylsulfate (SDS) from Pharmacia. The isolation and purification of LDPs were performed as described earlier [3,14,21] with minor modification.

cumulants. From the first cumulant the mean intensity-weighted diffusion coefficient (D) was obtained. The D value is used to provide information on the size applying the StokesEinstein relation to obtain the mean diameter, d d kT=3pD where kT is the thermal energy (4.12 10 21 J at 25 C) and is the viscosity of the solvent (0.89 10 3 kg/m . s for water at 25 C).

2.4. Fourier-transform infrared spectroscopy experiments

Attenuated total reflection (ATR) Fourier transform infrared spectroscopy (FTIR) was used to investigate the changes in the LDPs secondary structure when they passed from the buffer solution to an organic solvents (TFE, HFIP) to an SDS solution or to the lipid phase. This was done analysing the amide I band as previously described [28]. Briefly, samples were suspended in the buffer or in an appropriate solvent (WLIP was always reconstituted in 20% DMSO to favour its solubility), deposited on a 10-reflections germanium crystal (45 cut) and gently dried by nitrogen flushing. For the analysis of LDPs in the lipid phase, both substances in 10 mM HEPES pH 7.4 (hereafter buffer B) were incubated for 1 h at room temperature with palmitoyl-oleoyl-phosphatidylcholine (POPC) LUVs prepared in buffer B as described above, at a toxin to lipid molar ratio (T/L) of 1/20 and 1/50 for WLIP and tolaasin I, respectively. In the case of WLIP the incubation was performed twice in order to favour toxin binding. After incubation the mixtures were centrifuged in an Optima TL ultracentrifuge (Beckman, USA), using a fixed-angle rotor (TLA-100.2) at 350,000 g, for 3 h at 4 C. The resultant pellet was resuspended in buffer B (100 l) and directly deposited on the germanium crystal. Spectra of hydrated and deuterated film were collected, in an ATR geometry, using a FTS 185 spectrometer (Bio-Rad, USA), with an MCTdetector. By interposing a rotating KRS5 wire-grid polarizer we obtained polarized spectra to calculate the orientation of a structural element [29]. The dichroic ratio of the tolaasin I band at 1738 cm 1 (which corresponds to the stretching of the phospholipid carbonyl groups), provided a good estimation for Riso (i.e. the dichroic ratio for an isotropic distribution of dipoles) [30], and therefore was used to calculate the dichroic spectrum in the amide I, as described in Results. To estimate the secondary structure content, the amide I band in the region 17001600 cm 1 was analysed using a set of Gaussian components whose frequencies were assigned to different structural elements in the standard way as follows: band at 1680 3 cm 1 and 1621 5 cm 1 -structures; bands in the region 1660 3 cm 1, helical-structure; band at 1641 3 cm 1, random coil; band at 1610 5 cm 1, side chains contribution. In the case of tolaasin I, a single band at 1672 cm 1 appeared indicating the presence of trifluoroacetic acid (TFA) from the purification process. Such minor band was digitally removed and the resulting spectra were compared with those collected from samples without TFA (removed by lyophilizing the sample several times in the presence of 10 mM HCl [30]). No significant differences were found and the results after analysis were averaged.

2.2. Calcein release assay

Large unilamellar vesicles (LUVs) loaded with calcein at a self-quenching concentration (80 mM) have been used to measure membrane permeabilisation as reported by Kayalar and Duzgunes [22]. LUVs were prepared using pure lipids (4 8 mg/ml) by extrusion through two stacked polycarbonate filters with 100 nm pores [23]. The external non encapsulated calcein was removed by washing the vesicle suspension through Sephadex G-50 minicolumns, equilibrated with the experimental buffer [24]. The assay was performed with the microplate fluorimeter (Fluostar, BMG, Offenburg, Germany) as described in Dalla Serra et al. [25]. Unspecific binding of protein and liposomes to plastic microplate walls were strongly reduced by pre-treating the microplate with 0.1 mg/ml of Prionex (Pentapharm, CH). Aliquots of washed LUVs were dispensed into each well to a final lipid concentration of 612 M in 200 l of 20 mM TrisHCl, 100 mM NaCl, 1 mM EDTA, pH 7.0 (hereafter buffer A) containing the desired amount of toxin. Interferometric filter at 485 nm in excitation and at 538 nm in emission were used. Maximum calcein release (Fmax) was determined by adding 1 mM Triton X-100 and used as 100%. The extent of permeabilisation, R%, was then calculated as follows: R% 100Ffin Fin =Fmax Fin where Fin and Ffin represent the initial and final (after 45 min) value of fluorescence before and after toxin addition, respectively. Before each assay WLIP was dissolved in DMSO in order to favour its solubility; the same amount of DMSO tested alone did not induce any permeabilisation of LUVs (data not shown). Data of calcein release were further analysed using a statistical model first introduced by Parente et al. [26] and improved by Rapaport et al. [27]. This empirical model allowed the estimation of the three main parameters governing the permeabilisation process, i.e. the partition coefficient (K1) of toxin monomers from the aqueous solution into the lipid membrane, the aggregation process of membrane-bound monomers (K2), and the number of monomers (M) necessary for the formation of an active pore. The reported values were calculated as follows: a set of curves was generated by freely varying each of the three parameters in a wide range, curves with a 2 (chi square distribution) value comprised between the minimal (2 ) and 2 + (2 /2) were selected and min min min their parameters were used to calculate the average SD.

3. Results 3.1. Calcein release assay WLIP and tolaasin I were both able to permeabilise LUVs of different lipid composition, although in most of the cases tolaasin I displayed higher porating activity (Table 1). In particular, on PC LUVs tolaasin I provoked calcein efflux at five times lower concentrations than WLIP. The activities of both toxins showed similar responses to changes in the membrane lipid composition, even if WLIP activity was only slightly influenced by lipid changes. Thus, tolaasin I activity was enhanced by the SM content in the membrane but it decreased with increasing the sterols content (Table 1 and Fig. 2). A similar tendency has already been described for other LDPs [25,31]. However, differences were observed if Chol and Erg were compared: in fact, tolaasin I was

2.3. Dynamic light scattering

The dimension of vesicles was measured by dynamic light scattering (DLS) at a fixed angle (90) following procedures previously reported in Alvarez et al. [24]. The light scattering apparatus (Zetasizer3, Malvern, UK) was upgraded with a 30 mW laser diode light source (675 nm), 64 channels autocorrelator and a quartz cell with a Peltier temperature control. The quartz cell has been filled with 0.5 ml buffer A and 40 l of PC LUVs. The final lipid concentration was 80 M. The calculated autocorrelation functions were submitted to computer analysis (with the Malvern Application Software) using the method of

1716

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

Table 1 Permeabilising activity of WLIP and tolaasin I on vesicles of different lipid composition and a statistical analysis of their pore-formation mechanism Lipid composition a (mol %) PC 100 50 50 50 50 50 50 50
a b

WLIP 1/C50 b (M 1) 0.59 0.04 1.41 0.05 0.70 0.05 0.59 0.01 0.45 0.01 0.70 0.03 0.59 0.08 0.54 0.01 K1 c, d (M 1) 50 20 100 30 100 30 50 20 50 20 100 30 50 20 50 20 M
d, e

Tolaasin I K2 d, f 533 451 200 340 217 247 533 451 533 451 413 440 533 451 533 451 1/C50 b (M 1) 2.8 0.4 4.9 0.2 2.8 0.2 1.2 0.2 0.4 0.1 2.5 0.2 1.3 0.1 1.0 0.2 K1 c, d (M 1) 233 52 673 179 200 50 100 30 50 20 250 54 100 30 50 20 M
d, e

SM 0 50 33 16.5 0 33 16.5 0

Sterol 0 0 16.5* 33* 50* 16.5 33 50

K2 d, f 5.2 2.6 1.7 1.1 12 12 19.7 17.4 0.8 0.3 2.8 1.7 30 20 30 20

81 61 12 1 81 11 1 12 1 81 10 1

61 61 61 71 71 71 71 51

Lipid mixtures are reported on a molar base. PC: egg phosphatidylcholine; SM: sphyngomyelin; the sterol included in the vesicles was cholesterol* and ergosterol. The permeabilising activity is expressed as the reciprocal of the concentration of toxin causing 50% of calcein release (C50). Data are average S.D. of al least two different experiments. c The partition coefficient of toxin monomers into the liposome. d Data are average S.D. of at least three different experiments. e The number of monomers necessary for the formation of an active pore. f The aggregation process of membrane-inserted monomers.

more active on Erg-containing membranes. This effect was more pronounced when sterols and PC were present in equal molar amount, and decreased with the sterol content in the membrane. Interestingly, the analysis of these data with the Parente Rapaport model (parameters are reported in Table 1) showed a direct correlation between the permeabilising activity and the

empirical parameter K1 which is related to the partition coefficient of tolaasin I between water and lipid phase. The number of monomers constituting the active conducting unit (M) was instead always in the range of 6 1. In comparison to tolaasin I, WLIP showed a higher slope of its dose-dependence curve, similar to a detergent-like behaviour [25].

Fig. 2. Dose dependence activity of tolaasin I on liposomes with different lipid composition: Calcein loaded liposomes were exposed to different peptide concentrations. The percentage of calcein release was determined after 45 min as a function of the toxin/lipid ratio (T/L) and expressed as % of the maximal value obtained with TritonX-100. Experiments were done at constant lipid concentration (about 9 M). The true toxin concentration used was reported in the upper scale of the panel. Lines through the points are best fit according to the statistical model described in the text. Best values of fitting parameters are reported in Table 1. Vesicles were prepared with the following compositions expressed in molar ratio: (down triangle) PC:SM (50:50), (up triangle) PC:SM: Chol (50:33:16.5), (circle) PC:SM:Chol (50:16.5:33), (square) PC:Chol (50:50).

Fig. 3. Effect of toxins on vesicle size as determined by dynamic light scattering: The treatment of phosphatidylcholine vesicle with tolaasin I (open square) and WLIP (close circle) caused changes in vesicle average size ( diameter) which was reported vs. the toxin concentration (T), normalized to the vesicle diameter in the absence of toxin. Experiments were obtained at constant lipid concentration (80 M). The toxin/lipid ratio (T/L) was reported in the upper scale of the same panel. Inset: in this case we analysed the variation of the scattered light intensity ( ILS), normalized to the intensity value in the absence of toxin. Other parameters are as above.

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

1717

3.2. Dynamic light scattering Studies of dynamic light scattering (DLS) showed that the treatment of PC LUVs (80 M) with LDPs at a toxin to lipid molar ratio (T/L) lower than 0.3, caused only a slight increase of vesicle average size (Fig. 3) which was less than 20% for both WLIP and tolaasin I. At these toxin concentrations the release of calcein was lower than 70%. Therefore, at such doses both toxins are able to form pore-like lesions which allow calcein efflux without causing any destruction of the vesicles. A further increase of WLIP concentration (higher than 27 M) induced a decrease of vesicles average size with a concomitant marked decrease of the scattered light intensity (Fig. 3 and Inset), typical of a detergent-like activity through vesicles micellization. This effect was not observed in the case of tolaasin I, at least at concentrations up to 28 M, which induced only a slight decrease in the intensity of the scattered light (less than 35%). 3.3. Fourier-transform infrared spectroscopy experiments FTIR spectroscopy was used to determine the LDPs conformation and to follow its changes when molecules passed from a hydrophilic to a membrane mimetic environment. For this reason, FTIR spectra were collected either in buffer solution, in an organic solvent (TFE, HFIP), in SDS solution or in the presence of lipid vesicles. For our analysis

Fig. 5. FTIR-ATR spectra of WLIP in buffer and in lipid mimetic environments: (A) Analysis of the amide I band of deuterated films of WLIP samples deposited from a buffer solution (buffer, thick dotted line) or after binding to the lipid membrane (POPC, solid line). The original spectrum (POPC) was deconvoluted and curve fitted to resolve the component frequencies. The corresponding Gaussian bands are reported as dotted lines, and their sum as a thick dashed line superimposed to the original spectrum. (B) Differential spectra were obtained by subtracting the deuterated spectrum of the soluble form in buffer from that in TFE (1), in HFIP (3), in SDS (2) or in presence of the lipid phase (4).

Fig. 4. Differential spectra after H/D exchange of the WLIP amide protons: Analysis of differences in the amide I band of films of WLIP samples deposited from a buffer solution (B, dashed line), TFE (T, dotted line), HFIP (H, solid line), SDS (S, thick dashed line) and after binding to the lipid membrane (L, thick dotted-dashed line). Differential spectra were obtained by subtracting the hydrogenated spectra of WLIP in the different environments from the corresponding deuterated spectra. Inset: Area of negative peaks between 1710 and 1660 cm 1. All the differences are between normalized spectra, i.e., with the amide I peak area set at 1.

we considered the absorption of the amide I region after a complete deuteration, i.e. after exchange of amide protons with deuterium atoms (amide I). Firstly, the exchange rate is dependent on both the extent of hydrogen bonds (since these bonds must be broken for the exchange to occur), and on the accessibility of a given proton. As a consequence, hydrogen/ deuterium (H/D) exchange provides information about the local structure of the molecule. Moreover, helical and unordered structure contribute to the amide I vibration at almost identical wavenumbers making therefore difficult the precise determination of helix and random coil structures from the IR spectra. However, after H/D exchange it is possible to discriminate between these structures because of changes in the random coil absorption, which shifts to a higher extent than the helical component absorption. In the case of WLIP, differential spectra obtained after the H/D exchange in different environments (Fig. 4) showed a decrease of absorption at 1670 cm 1 and a corresponding increment at lower frequency (1634 cm 1). As appears from a comparison of areas from the negative peaks centred at 1670 cm 1 (inset of Fig. 4), this effect decreased passing from the buffer solution to a more hydrophobic environment (either the solvent or the lipid membrane).

1718

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

Similar changes are shown in Fig. 5B where differential spectra, obtained by subtracting the deuterated spectrum of the soluble form in buffer from those collected in all other environments, are plotted. An excess of absorbance at about 1670 cm 1 was observed, with the concomitant decrease of absorption at 1634 cm 1. The effect was maximum upon interaction with the lipid membrane. By applying the curve fitting procedure to the amide I peak of WLIP in lipid (Fig. 5A), we calculated changes of absorption in comparison with the soluble form in buffer. The results show an increment from 11 to 26% and a decrease from 44 to 26% in the regions centred at 1672 cm 1 and at 1634 cm 1 , respectively. Furthermore, polarization experiments indicated that WLIP had some disordering effect on the lipid membrane: the average orientation of lipid chains with respect to the normal to the plane changed from 34 to 41 upon interaction (Table 3), a typical change induced by proteins inserted into the membrane [29,32]. In the case of tolaasin I, frequency component analysis of the amide I band (Table 2) indicated very small changes, if any, passing from the buffer solution to less polar solvents. The structure was composed predominantly of helical component (comprising 46% of the structure), random coil (28%) and -structure (27%). Interestingly, upon interaction with the lipid membrane we observed a significant increase of the helical structure (from 46 to 63%) at expenses of other components, i.e. random coil (which decreased from 28 to 15%) and -structure (from 27 to 21%). Finally, the orientation of the amide I band for tolaasin I co-sedimented with LUVs was qualitatively analysed by subtracting the 90 polarized spectrum (normalized by Riso) from that at 0 (Fig. 6B): a clear positive deviation was found at 1658 cm 1, typical of the helical component, which suggested an orientation of this element, that became close to parallel to the membrane normal [30]. From a quantitative analysis of the polarized spectra of the inserted peptide (Fig. 6A), it was possible to calculate its dichroic ratio and orientation around the perpendicular to the plane of the membrane (Table 3) [32]. We obtained an average tilt angle () of 47. When we considered changes in the lipid chains of LUVs upon interaction with tolaasin I, we noticed an

Fig. 6. Analysis of FTIR-ATR spectra of tolaasin I in POPC layers with polarizer: (A) Spectra were taken with either parallel (0) or perpendicular (90) polarization. The amide I region of tolaasin I bound to vesicles was reported after subtraction of the lipid contribution. The best fit curve with Gaussian components (dotted lines) was superimposed as a thick dashed line to the 90 polarized trace (solid line). The absorption bands in the parallel and perpendicular configuration were used to calculate the orientation of the corresponding structural element as reported in Table 3. Bands are: h (helix), (-structure), r (random). (B) Dichroic spectrum obtained by subtracting the 90 polarized spectrum (after multiplication by Riso i.e. 1.54) from that at 0.

evident increment of the tilt angle of the lipid chains from 34 to 45, meaning that the binding of the peptide caused an increment of the lipid disorder. This value was used to recalculate the relative orientation of the structure with respect to the lipid chains [32], providing an angle (L) around 20 2. It suggested that tolaasin I was not only bound to the membrane but also inserted in it and fully immersed into the lipid core. 4. Discussion In this paper the permeabilising activity of WLIP and tolaasin I has been compared. Both LDPs are able to damage biological and artificial membranes through the formation of transmembrane pores. This was assessed also recently by Cho et al. [19] who convincingly reported the ion channel forming ability of tolaasin I, and by Lo Cantore et al. [3], who showed evidences for pore formation of either WLIP or tolaasin I on erythrocytes, using osmotic protectants. Consequently, we investigated the activity of both LDPs on artificial membranes formed by different combination of phospholipids, sphingomyelin and sterols, which are the main components of plasma cell membranes. Tolaasin I was more active on SM-containing membranes than on LUVs of phospholipids alone and less active when sterols were present. This feature was also observed for other

Table 2 FTIR spectroscopy determination of the secondary structure of tolaasin I in aqueous buffer, with and without lipids, and in different solvents System % Secondary structure a buffer TFE HFIP SDS POPC 29 24 24 27 21 h 44 45 48 47 63 r 27 30 28 26 15

a : total -structure (i.e. -sheet, -turn); h: helix; r: random. Errors of FTIR determinations are 5%.

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

1719

peptin-like LDPs which did not show any specific requirement for sterols [25,31] and apparently correlated to their cell specificity, being more phytotoxic and less antifungal than the nonapeptides syringomycines, pseudomycines and cormycin, which showed a clear preference for sterols [20,25,31,3336]. In the case of tolaasin I, the slight preference for ergosterol (a fungal sterol) among sterols is consistent with its antifungal activity. Contrarily to our expectations, also WLIP, which is a nonapeptide, was more active on SM-containing LUVs. It could be explained by observing its structure which is more similar to the peptine-like LDPs with only a portion of the peptide moiety involved in the formation of the lactone ring. In comparison to tolaasin I, the inhibitory effect of sterols was less effective on the WLIP activity. Nevertheless the two toxins showed similar responses to changes of the lipid membrane composition, some interesting differences appeared after analysis of data, all suggesting a different mode of action. Firstly, the parameter M (i.e., the molecularity of the permeabilising unit) estimated from the ParenteRapaport model and reported in Table 1, was larger with WLIP and in the same range of detergents [25]. In the case of tolaasin I, Cho et al. [19] described the existence of subconductance states, which could be related to the presence of pores with slightly different molecularities. This instability of the pore structure well correlates with the variability of the parameter M found by our analysis. For both LDPs the activity seemed mainly due to the ability of monomers to bind to the liposome membrane (K1) rather than to the two-dimensional aggregation rate (K2) of the bound peptide. A second difference comes from the light scattering signal. From our results it was evident that WLIP at concentration higher than 27 M showed again a detergent-like activity; in fact, the intensity of the scattered light felt down probably in consequence of vesicles micellization. On the contrary, both toxins at low concentration (at a toxin to lipid molar ratio up to 0.3) caused an increase of the liposome dimensions and a decrease of the intensity of the diffused light. These effects well correlated with vesicles aggregation, that increased the measured diameters, and with precipitation of vesicle macroaggregates, which caused a decrease of signal intensity through a decrease of the concentration of scattering centres. Interestingly, the concentrations of WLIP and tolaasin I used in our assays were of the same order of magnitude required to cause mushroom tissue alteration in vivo [3,16,18]. Also the different activity between the two LDPs was observed, being WLIP about ten times less active than tolaasin I. Our results, together with those reported in Lo Cantore et al. [3], supported the idea of the cell membrane as the main target for both WLIP and tolaasin I. Therefore it seemed interesting to investigate the conformational changes occurring when these molecules interacted with the lipid environment. The secondary structure of both of them was already reported but only in the soluble form [10,21,37]. In particular the conformation of tolaasin I was studied in SDS [21,37]:

in this membrane-like environment it was showed the presence of a left-handed -helix, suggesting a refolding of the structure from the aqueous buffer where the conformation of the whole peptide was largely extended [37]. Considering these results, the structure of peptides was analysed in the lipid-mimetic conditions provided by TFE, HFIP and SDS, as well as in the model lipid membrane of POPC liposome. We expected changes in the secondary structure of peptides passing from the buffer solution to the more hydrophobic environment, with a maximum effect in the presence of the lipid membrane. In the case of WLIP, the rigidity of the molecule, a nonapeptide cyclized via lactone formation between the third and the C-terminal residues [10], could explain the slight variation observed in the amide I spectrum even after the interaction with the membrane. Moreover, the presence in the same molecule of unusual residues made the assignment of a secondary structure to the curve fitting components more difficult. Therefore we focused on differential spectra obtained after H/D exchange or after the buffer substitution with a membrane-like environment. In fact, variations of spectra after H/D exchange provided information about the local structure of the molecule. The decrease of absorption at 1670 cm 1 (and the corresponding increment at 1634 cm 1), for example, indicated the presence of free protons (i.e. not involved in hydrogen bonds) which were able to exchange with deuterium, probably because of their good accessibility allowed by the local structure of the molecule. It could be attributed to a turn configuration with fast exchanging protons. Interestingly, this behaviour was depended on the solvent used and decreased in a lipid-mimetic environment (effect measured by the small negative and positive peaks at 1670 cm 1 and 1634 cm 1, respectively), suggesting a decrease of the proton accessibility and, consequently, a change in the local structure of the molecule. We deduce that the lipid-mimetic conditions enhanced the formation of hydrogen bonds inside the structure which changed, becoming less susceptible to deuteration. Similar changes were confirmed in differential spectra obtained by subtracting the deuterated spectrum of the soluble form in buffer, from those collected in all the other solvents. Variations in the absorption were at the same wavenumbers described above, showing an increment at 1670 cm 1 and a corresponding decrease at 1634 cm 1 with the maximum in the lipid membrane. These changes were in favour of a more stable local structure which was enhanced from the lipid-mimetic environments and that could be able to interact with the membrane and its hydrophobic moiety. In addition, the disordering effect on the vesicles membrane induced by the binding of WLIP supports the hypothesis of an insertion of the molecule into the membrane lipid core. According to what we expected, we noticed an increment of the helical content in the structure of tolaasin I when it passed from the buffer solution to the membrane-like environment. This could be consistent with results obtained by CD and NMR spectroscopy in aqueous solution and in SDS, respectively [21,37], although some observations need

1720

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722

Table 3 Assignment and dichroic ratio of some IR bands observed in lipid vesicles with and without toxin Wavenumber (cm 1) Lipid alone 2920 2850 LUVs + WLIP 2920 2850 LUVs + tolaasin I 2920 2850 1660
a b

Vibration a

Direction () b

Dichroic ratio 1.32 0.01 1.34 0.01

Angle () c 33 1 34 1

as CH2 stretching 90 s CH2 stretching 90

as CH2 stretching 90 s CH2 stretching 90

1.48 0.01 1.52 0.01

41 1 42 1

as CH2 stretching 90 s CH2 stretching 90 Amide I helix 30

1.61 0.01 1.63 0.01 2.51 0.10

45 1 45 1 47 2

as: asymmetric, s: symmetric. Direction of the variation of the dipole moment associated to the vibration with respect to the direction of the main molecular axis (aliphatic chain or -helix axis). c Average angle between the direction of the molecular axis and the perpendicular to the crystal plane (i.e. membrane plane).

to be added. As mentioned above for WLIP, also the peptide moiety of tolaasin I (18 aminoacids) includes unusual residues and with D-chirality, so the standard assignment of secondary structures to the curve fit components could be not always applied. Interesting studies were recently reported on -helical membrane lytic peptides in which a few Lamino acids were replaced with their D-enantiomers: incremented amide I frequencies (16561670 cm 1) compared with pure -helical structures (16451654 cm 1) were observed and assigned to 310-helix or dynamic/distorted helix [38]. Previous studies had correlated this region with 310-helix [39], -turn [40] and interestingly to a left-handed -helix, as well [41,42]. In our analysis we found indeed two components around the typical region of helix absorption, centred at 1654 cm 1 and 1664 cm 1, respectively (data not shown). We noticed that after deuteration the position of both peaks did not change, suggesting a stable structure maintained by hydrogen bonds, like a helix. Moreover, the sum of percentage absorption of these two components kept at a constant value, independently of the environment (i.e. buffer, solvents and membranes), on the contrary the single absorptions changed: the peak at 1654 cm 1 increased passing from buffer to solvents at expenses of peaks at 1664 cm 1. After interaction with the membrane, the contribution of the former was maximum but analysis of the dichroic spectrum showed only one peak at higher wavenumber (1659 cm 1). For all these reasons, we joined the two contributions in an unique peak and we assigned to it an helix structure which finally resulted the conformation adopted from tolaasin I to bind the membrane and to insert in it. The significant decrease in the orientation order parameter of the aliphatic chain region (Table 3) and the orientation of tolaasin I in the lipid bilayer prompt us to interpret our results in favour of a barrel-stave mechanism of

action versus a carpet like hypothesis [43]. This is indeed in agreement with Jourdan et al. [37] who described for tolaasin I the formation of a stable amphipathic helical structure and suggested its insertion into the lipid matrix for the formation of an ionic channel. On the contrary, a carpet-like mechanism was previously proposed for magainin derived lipopeptides which localized on the membrane surface and did not significantly destabilise the acyl-chain order. In fact, Avrahami et al. [44] observed negative order parameters for both natural magainin and lipomagainin (magainin analogues conjugated with lipophilic acids), typical of helices oriented nearly parallel to the membrane surface. In addition, after incorporation of the peptides into the lipid membrane, the lipid order was only poorly disturbed and to the same extent, suggesting that both peptides did not insert deeply into the lipid core. A different mechanism of action could be proposed for WLIP, in agreement with the evidences here reported for membrane disruption in a detergent-like manner. In fact, it does not have a sufficient length to transverse the entire membrane, similarly to syringomicin for which Malev et al. [45] hypothesized a toroidal pore. In summary, the results described in this paper shed light on the permeabilising effects induced by WLIP and tolaasin I on liposome, and on the secondary structure occurring in these molecules during their interaction with the lipid membrane, which is their biological target.

Acknowledgements This work was financially supported by Grant, Progetto Giovani Ricercatori (Biological characterisation of the tolaasin and WLIP, lipodepsipeptides of Pseudomonas tolaasii and P. reactans) from Universit degli Studi della Basilicata and by Provincia Autonoma di Trento (PAT, projects AGRIBIO and SyrTox). It was also partially supported by Fondazione Cariverona (Bando 2004 Integrazione tra tecnologia e sviluppo di settore Bando per progetti di ricerca a Indirizzo Biomedico). References
[1] J.M. Wells, G.M. Sapers, W.F. Fett, J.E. Butterfield, J.B. Jones, H. Bouzar, F.C. Miller, Postharvest discoloration of the cultivated mushrooms Agaricus bisporus caused by Pseudomonas tolaasii, P. reactans, and P. gingeri, Phytopathology 86 (1996) 10981104. [2] P. Lo Cantore, S. Lazzaroni, A. Evidente, C. Cafarchia, N.S. Iacobellis, Production of tolaasin I and WLIP by Pseudomonas tolaasii and P. reactans, their antimicrobial activity and possible role in the virulence of the pathogens, in: S.N. Iacobellis, et al., (Eds.), Pseudomonas syringae and Related Pathogens. Biology and Genetic, Kluwer Academic Publishers, Dordrecht, NL, 2003, pp. 255262. [3] P. Lo Cantore, S. Lazzaroni, M. Coraiola, M. Dalla Serra, C. Cafarchia, A. Evidente, and N.S. Iacobellis, Biological characterisation of WLIP produced by Pseudomonas reactans strain NCPPB1311, Mol. Plant Microbe. Interact. 4 (in press). [4] A.G. Tolaas, A bacterial disease of cultivated mushrooms, Phytopathology 5 (1915) 51.

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722 [5] S.A. Godfrey, J.W. Marshall, J.D. Klena, Genetic characterization of Pseudomonas NZI7a novel pathogen that results in a brown blotch disease of Agaricus bisporus, J. Appl. Microbiol. 91 (2001) 412420. [6] N.S. Iacobellis, P. Lo Cantore, Pseudomonas reactans a new pathogen of cultivated mushrooms, in: N.S. Iacobellis, et al., (Eds.), Pseudomonas syringae and Related Pathogens. Biology and Genetic, Kluwer Academic Publishers, Dordrecht, NL, 2003, pp. 595605. [7] P. Munsch, T. Alatossava, N. Marttinen, J.M. Meyer, R. Christen, L. Gardan, Pseudomonas costantinii sp. nov., another causal agent of brown blotch disease, isolated from cultivated mushroom sporophores in Finland, Int. J. Syst. Evol. Microbiol. 52 (2002) 19731983. [8] P. Lo Cantore, M. Coraiola, M. Dalla Serra, G. Menestrina, S. Lazzaroni, A. Evidente, N.S. Iacobellis, Interaction of tolaasin I and WLIP, lipodepsipeptides of Pseudomonas tolaasii and P. reactans, with biological and model membranes, in: S.N. Iacobellis, et al., (Eds.), Pseudomonas syringae and Related Pathogens. Biology and Genetic, Kluwer Academic Publishers, Dordrecht, NL, 2003, pp. 263273. [9] W.C. Wong, T.F. Preece, Identification of Pseudomonas tolaasii: the white line in agar and mushroom tissue block rapid pitting tests, J. Appl. Bacteriol. 47 (1979) 401407. [10] R.J. Mortishire-Smith, J.C. Nutkins, L.C. Packman, C.L. Brodey, P.B. Rainey, K. Johnstone, D.H. Williams, Determination of the structure of an extracellular peptide produced by the mushroom saprophytic Pseudomonas reactans, Tetrahedron 47 (1991) 36453654. [11] T.R. Neu, T. Haertner, K. Poralla, Surface active properties of viscosin: a peptidolipid antibiotic, Appl. Microbiol. Biotechnol. 32 (1990) 518520. [12] M.V. Laycock, P.D. Hildebrand, P. Thibault, J.A. Walter, J.L.C. Wright, Viscosin, a potent peptidolipid biosurfactant and phytopathogenic mediator produced by a pectolytic strain of Pseudomonas fluorescens, J. Agric. Food Chem. 39 (1991) 483489. [13] J.C. Nutkins, R.J. Mortishire-Smith, L.C. Packman, C.L. Brodey, P.B. Rainey, K. Johnstone, D.H. Williams, Structure determination of tolaasin, an extracellular lipodepsipeptide produced by the mushroom pathogen Pseudomonas tolaasii Paine, J. Am. Chem. Soc. 113 (1991) 26212627. [14] C. Bassarello, S. Lazzaroni, G. Bifulco, P. Lo Cantore, N.S. Iacobellis, R. Riccio, L. Gomez-Paloma, A. Evidente, Tolaasins AE, five new lipodepsipeptides produced by Pseudomonas tolaasii, J. Nat. Prod. 67 (2004) 811816. [15] P.B. Rainey, C.L. Brodey, K. Johnstone, Biological properties and spectrum of activity of tolaasin, a lipodepsipeptide toxin produced by the mushroom pathogen Pseudomonas tolaasii, Physiol. Mol. Plant Pathol. 39 (1991) 5770. [16] C.L. Brodey, P.B. Rainey, M. Tester, K. Johnstone, Bacterial blotch disease of the cultivated mushroom is caused by an ion channel forming lipodepsipeptide toxin, Mol. Plant-Microb. Interact. 4 (1991) 407411. [17] P.B. Rainey, C.L. Brodey, K. Johnstone, Biology of Pseudomonas tolaasii, cause of brown blotch disease of the cultivated mushroom, Adv. Plant Pathol. 8 (1992) 95117. [18] M.L. Hutchison, K. Johnstone, Evidence for the involvement of the surface active properties of the extracellular toxin tolaasin in the manifestation of the brown blotch disease symptoms by Pseudomonas tolaasii on Agaricus bisporus, Physiol. Mol. Plant Pathol. 42 (1993) 373384. [19] K.H. Cho, Y.K. Kim, Two types of ion channel formation of tolaasin, a Pseudomonas peptide toxin, FEMS Microbiol. Lett. 221 (2003) 221226. [20] C.L. Bender, F. Alarcon Chaidez, D.C. Gross, Pseudomonas syringae phytotoxins: mode of action, regulation, and biosynthesis by peptide and polyketide synthetases, Microbiol. Mol. Biol. Rev. 63 (1999) 266292. [21] R.J. Mortishire-Smith, A.F. Drake, J.C. Nutkins, D.H. Williams, Left handed alpha-helix formation by a bacterial peptide, FEBS Lett. 278 (1991) 244246. [22] C. Kayalar, N. Dzgnes, Membrane action of colicin E1: detection of the release of carboxyfluorescein and calcein from liposomes, Biochim. Biophys. Acta 860 (1986) 5156. [23] R.C. MacDonald, R.I. MacDonald, B.P. Menco, K. Takeshita, N.K. Subbarao, L.R. Hu, Small-volume extrusion apparatus for preparation of large, unilamellar vesicles, Biochim. Biophys. Acta 1061 (1991) 297303.

1721

[24] C. Alvarez, M. Dalla Serra, C. Potrich, I. Bernhart, M. Tejuca, D. Martinez, I.F. Pazos, M.E. Lanio, G. Menestrina, Effects of lipid composition on membrane permeabilization by Sticholysin I and II, two cytolysins of the sea anemone Stichodactyla helianthus, Biophys. J. 80 (2001) 27612774. [25] M. Dalla Serra, G. Fagiuoli, P. Nordera, I. Bernhart, C. Della Volpe, D. Di Giorgio, A. Ballio, G. Menestrina, The interaction of lipodepsipeptide toxins from Pseudomonas syringae pv. syringae with biological and model membranes: a comparison of syringotoxin, syringomycin and syringopeptins, Mol. Plant-Microb. Interact. 12 (1999) 391400. [26] R.A. Parente, S. Nir, F.C. Szoka Jr., Mechanism of leakage of phospholipid vesicle contents induced by the peptide GALA, Biochemistry 29 (1990) 87208728. [27] D. Rapaport, R. Peled, S. Nir, Y. Shai, Reversible surface aggregation in pore formation by pardaxin, Biophys. J. 70 (1996) 25022512. [28] G. Menestrina, V. Cabiaux, M. Tejuca, Secondary structure of sea anemone cytolysins in soluble and membrane bound form by infrared spectroscopy, Biochem. Biophys. Res. Commun. 254 (1999) 174180. [29] A. Menikh, M.T. Saleh, J. Gariepy, J.M. Boggs, Orientation in lipid bilayers of a synthetic peptide representing the C-terminus of the A(1) domain of shiga toxin. A polarized ATR-FTIR study, Biochemistry 36 (1997) 1586515872. [30] E. Goormaghtigh, V. Raussens, J.M. Ruysschaert, Attenuated total reflection infrared spectroscopy of proteins and lipids in biological membranes, Biochim. Biophys. Acta 1422 (1999) 105185. [31] G. Menestrina, M. Coraiola, V. Fogliano, A. Fiore, I. Grgurina, A. Carpaneto, F. Gambale, M. Dalla Serra, Antimicrobial lipodepsipeptides from Pseudomonas spp: a comparison of their activity on model membranes, in: S.N. Iacobellis (Ed.), Pseudomonas syringae and Related Pathogens. Biology and Genetic, Kluwer Academic Publishers, Dordrecht, NL, 2003, pp. 185198. [32] G. Menestrina, Use of fourier-transformed infrared spectroscopy (FTIR) for secondary structure determination of staphylococcal pore-forming toxins, in: O. Holst (Ed.), Bacterial Toxins, Methods and Protocols, Humana Press, Totowa, NJ, 2000, pp. 115132. [33] C. Julmanop, Y. Takano, J.Y. Takemoto, T. Miyakawa, Protection by sterols against the cytotoxicity of syringomicin in the yeast Saccharomyces cerevisiae, J. Gen. Microbiol. 139 (1993) 23232327. [34] N. Taguchi, Y. Takano, C. Julmanop, Y. Wang, S. Stock, J.Y. Takemoto, T. Miyakawa, Identification and analysis of the Saccharomyces cerevisiae SYR1 gene reveals that ergosterol is involved in action of syringomycin, Microbiology (UK) 140 (1994) 353359. [35] A.M. Feigin, L.V. Schagina, J.Y. Takemoto, J.H. Teeter, J.G. Brand, The effect of sterols on the sensitivity of membranes to the channel-forming antifungal antibiotic, syringomycin E, Biochim. Biophys. Acta 1324 (1997) 102110. [36] A. Scaloni, M. Dalla Serra, P. Amodeo, L. Mannina, R.M. Vitale, A.L. Segre, O. Cruciani, F. Lodovichetti, M.L. Greco, A. Fiore, M. Gallo, C. D'Ambrosio, M. Coraiola, G. Menestrina, A. Graniti, V. Fogliano, Structure, conformation and biological activity of a novel lipodepsipeptide from Pseudomonas corrugata: Cormycin A, Biochem. J. 384 (2004) 2536. [37] F. Jourdan, S. Lazzaroni, B.L. Mendez, P. Lo Cantore, M. de Julio, P. Amodeo, N.S. Iacobellis, A. Evidente, A. Motta, A left-handed alpha-helix containing both L- and D-amino acids: the solution structure of the antimicrobial lipodepsipeptide tolaasin, Proteins 52 (2003) 534543. [38] N. Papo, Y. Shai, Effect of drastic sequence alteration and D-amino acid incorporation on the membrane binding behavior of lytic peptides, Biochemistry 43 (2004) 63936403. [39] D.F. Kennedy, M. Crisma, C. Toniolo, D. Chapman, Studies of peptides forming 3(10)- and alpha-helices and beta-bend ribbon structures in organic solution and in model biomembranes by Fourier transform infrared spectroscopy, Biochemistry 30 (1991) 65416548. [40] H. Susi, D.M. Byler, Resolution enhanced Fourier transform infrared spectroscopy of enzymes, Methods Enzymol. 130 (1986) 290311. [41] V. Giancotti, F. Quadrifoglio, V. Crescenzi, Polyelectrolyte behaviour of phosvitin. Spectroscopic, microcalorimetric and acridine-orange-binding data, Eur. J. Biochem. 35 (1973) 7886.

1722

M. Coraiola et al. / Biochimica et Biophysica Acta 1758 (2006) 17131722 [44] D. Avrahami, Y. Shai, Conjugation of a magainin analogue with lipophilic acids controls hydrophobicity, solution assembly, and cell selectivity, Biochemistry 41 (2002) 22542263. [45] V.V. Malev, L.V. Schagina, P.A. Gurnev, J.Y. Takemoto, E.M. Nestorovich, S.M. Bezrukov, Syringomycin E channel: a lipidic pore stabilized by lipopeptide? Biophys. J. 82 (2002) 19851994.

[42] M.J. Citra, P.H. Axelsen, Determination of molecular order in supported lipid membranes by internal reflection Fourier transform infrared spectroscopy, Biophys. J. 71 (1996) 17961805. [43] L. Yang, T.A. Harroun, T.M. Weiss, L. Ding, H.W. Huang, Barrel-stave model or toroidal model? A case study on melittin pores, Biophys. J. 81 (2001) 14751485.