-tical ander crystalsut 300 nm.

Journal of Colloid and Interface Science 271 (2004) 507–510www.elsevier.com/locate/jcis

Letter to the Editor

Cutin and suberin monomers are membrane perturbants

Jean-Paul Douliez

Unité de Recherche sur les Protéines Végétales et leurs Interactions, INRA, rue de la Géraudière, 44316 Nantes, France

Received 11 August 2003; accepted 10 December 2003

Abstract

The interaction between cutin and suberin monomers, i.e.,ω-hydroxylpalmitic acid,α,ω-hexadecanedioic acid,α,ω-hexadecanediol, 12hydroxylstearic acid, and phospholipid vesicles biomimicking the lipid structure of plant cell membranes has been studied by optransmission electron microscopy, quasielastic light scattering, differential scanning calorimetry, and31P solid-state NMR. Monomers wershown to penetrate model membranes until a molar ratio of 30%, modulating their gel to fluid-phase transition, after which monomealso formed in solution. These monomers induced a decrease of the phospholipid vesicle size from several micrometers to aboThe biological implications of these findings are discussed. 2004 Elsevier Inc. All rights reserved.

Keywords: DMPC membranes; Bola amphiphiles; Vesicles; Fatty acids; Cutin; Suberin

ter-8

e ofpor-has

ath-zedstillerandthe

ng

ac-gened,ellsurstur-lls,ne

i-y-

ashey),eticticd tond

esi-

seheirli-

iza-oftheby

ionder

1. Introduction

Cutin and suberin are plant polymers consisting of esified hydroxyl and dicarboxylic fatty acids with 16 and 1atoms of carbon [1–3]. They are localized at the surfacorgans and their surface properties are of particular imtance [4]. Whereas the biosynthesis of the monomersbeen the focus of numerous studies [5], their routing pway in the plant from the cell where they are synthesito the external layers where they are polymerized isunclear. The participation of lipid colloids, i.e., monomcorpuscles, has been formulated [1]. However, the cutinsuberin monomers are poorly water soluble, suggestingparticipation of proteins or other lipids as cosolubiliziagents.

Another important question is related to the potentialtivity of cutin monomers as messengers in plant–pathointeractions [1]. Upon pathogen attack, cutin is hydrolyzreleasing monomers which can be perceived by plant cfor the induction of resistance [1,6]. How this process occis still an enigma but monomers could induce strong perbations in the membrane of the polymer juxtaposing cepotentially modulating its functions and that of membraproteins.

E-mail address: douliez@nantes.inra.fr.

0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2003.12.020

Several cutin and suberin monomers asω-hydroxylhexa-decanoic acid (ω-OHplm) α,ω-hexadecanedioic acid (dCOplm),α,ω-hexadecanediol (diOHplm), and 12-hydroxstearic acid (12OHSt) are commercial and were usedmodels of cutin and suberin monomers in this study. Twere mixed with dimyristoyl phosphatidylcholine (DMPCa phospholipid forming membrane vesicles as a synthsimplified membrane model of bilayers, i.e., a biomimesystem of cell membranes. Optical microscopy was useobserve the texture of the lipid mixtures, light scattering atransmission electron microscopy for determining the vcle size, and DSC and31P solid-state NMR to follow thethermotropism and phase behavior of the mixtures.

2. Materials and methods

2.1. Lipids

All lipids used in this study were from Sigma. Becaucutin and suberin monomers are not water soluble, tinteraction with DMPC cannot be probed on preformedposomes of this phospholipid and previous cosolubiltion in a solvent is required as in the case of mixturesphospholipids and cholesterol [7]. As a consequence,vesicle-size distribution is polydisperse and is driventhe monomer/phospholipid interactions. A stock solutof DMPC was then prepared in methanol, vortexed un

508 J.-P. Douliez / Journal of Colloid and Interface Science 271 (2004) 507–510

atol-Allwithriedis-ofed,

ev-singloadninginter-

onsserea

e-pidoy60

ed.

ikonVC

as

theventver

thercid.notepacro-

ere

edg

bideter-ize.

tudyrsedpec-lityttyedan-

esi-ervedpa-

oni-mefor

asen,mo-ys-

hatrys-[7].wn

atioofof athe

ersen

ed aplit-ndthissug-snes,andhatiticis

eico-

low-power ultrasonic conditions, and further aliquoted10 mg/ml in glass tubes. Monomers were previously subilized in methanol and added with the phospholipid.samples were dried with a stream of nitrogen, hydratedpure water, freeze-thawed three times, and then freeze-dThis allowed removal of any trace of methanol. The lipid dpersions were further realized with 1 ml (full hydration)a 30 mM Tris–HCl, pH 7.5, buffer. Samples were vortexfreeze-thawed three times, and conserved at−20◦C. Priorto use, each sample was heated to 50◦C for 30 min.

2.2. Solid-state NMR

31P solid-state NMR experiments were performed at seral temperatures on a 400 MHz Bruker spectrometer ua static probe, the sample coil of which was adapted toa 7-mm rotor such as those used with magic-angle-spinprobes. Typically, 350 µl of lipid dispersion was pouredthe rotor. A Hahn echo sequence was used with an inpulse delay of 30 µs, and 1k points in 12k accumulatiwere done with a 90◦ pulse and a spectral width of 7 µand 100 kHz, respectively. Free induction decay signals wzero-filled to 4k points prior to Fourier transform afterbroad line exponential multiplication of 100 Hz.

2.3. Differential scanning calorimetry

Experiments were performed on a microDSC III from Staram (Caluire, France). The amount of 0.85 ml of the lisolution (10 mg/ml) was accurately weighed in a HastellC276 vessel. The sample was scanned between 1 and◦Cupon two successive heating and cooling cycles at 1◦C/min.The DSC traces from the second heating steps were us

2.4. Light microscopy

Observations were made at 20× magnification usingan optical microscope in the phase-contrast mode (NEclipse E-400, Tokyo, Japan) equipped with a 3-CCD Jcamera allowing digital images (768× 512 pixels) to becollected. A drop of the lipid dispersion (about 50 µl) wdeposited on the glass-slide surface (76× 26 × 1.1 mm,RS France) and covered with a cover slide (22× 22 mm,Menzel-Glaser, Germany). With such a lipid volume,dispersion was not confined between glass plates, preing the formation of tubules or giant vesicles [8]. The coslides were cleaned with ethanol and acetone.

2.5. Transmission electron microscopy (TEM)

Samples were deposited on copper grids and furstained with either uranyl acetate or phosphotungstic aThis procedure is widely used in the literature and isexpected to affect the results, i.e., the vesicle size. Prrations were examined with a transmission electron miscope JEOL 100S operating at 80 keV. Observations wmade at various magnifications.

.

-

-

2.6. Light scattering

Quasielastic light-scattering experiments were performwith an HPPS instrument from Malvern (UK) allowinparticle-size determination from 0.6 nm to 6 µm in turmedium. The scattered beam at 173◦ was analyzed and thcorrelation function treated by CONTIN software. Scating intensities are reported as a function of particle sSamples were diluted 10 times.

3. Results and discussion

The pure cutin and suberin monomers used in this sare not water soluble at neutral pH and remained dispeas crystals. Phase-contrast microscopy and visual instion (not shown) were used to investigate their solubiin DMPC membranes at four different molar ratios, faacids/DMPC ofRi = 0.01, 0.1, 0.5, and 1. As commentabove, the interaction between monomers with DMPC cnot be probed on well-defined preformed phospholipid vcles. As a consequence, polydisperse vesicles were obsin the case of pure DMPC [8] as expected from the preration method since no sizing down as extrusion or scation was used. At a molar ratio of 0.1, solutions becaless turbid and vesicles exhibited a lower diameter thanRi = 0.01. At Ri = 0.5 small vesicles were still presentsmall dots in addition to crystals of about 10 µm size. Thit can be estimated that DMPC incorporates about 30%lar cutin or suberin monomers prior to yield a two-phase stem in which crystals coexist. This behavior is similar to tencountered in the case of cholesterol which also forms ctals at ratios higher than 50% molar in model membranesInterestingly, monocarboxylic fatty acids have been shoto incorporate in phospholipid membranes at a molar rof 2 [9]. This reveals that the specific chemical structurethe cutin and suberin monomers, that is, the presencesecond carboxylic moiety or hydroxyl group, hamperscosolubilization in model membranes.

DSC was then used to monitor the effect of monomon the thermotropism of DMPC. This phospholipid whpure exhibits a weak pretransition at 15◦C from a gel phaseto a ripple phase and a main transition at 24◦C to a fluidphase [10]. Increasing the amount of monomers inducshift toward higher temperatures, a broadening or a sting into two overlapping peaks of the main transition, athe disappearance of the pretransition (Fig. 1). First,suggests a possible phase separation, which is furthergested from31P NMR results. Next, this clearly showthat monomers are incorporated in the DMPC membramodulating the interaction between the phospholipidstheir phase transition. This behavior is very similar to tobserved in the case of mixtures of DPPC and palmacid [9] or cholesterol for which the main transitionmarkedly broadened [7] or upon incorporation of linolacid in DMPC membranes [11]. Moreover, a similar therm

J.-P. Douliez / Journal of Colloid and Interface Science 271 (2004) 507–510 509

unts

prosi-ade

tran4]

em-sentciallyatioof

ture

is-olarinin-ig. 2edres-MRas

trasev-olarbed-oids

sticsizera-

icles

nd(B)

ra-hera

col-inetheb-

acid

t anout

with

Fig. 1. DSC thermograms obtained upon incorporation of various amoof ω-OHplm in DMPC. The molar ratio is indicated on the left.

gram was observed in the case of sonicated vesicles, acedure which is known to induce small unilamellar vecles [12], in agreement with the present observations mby phase-contrast microscopy. Moreover, vesicle-shapesitions have been widely studied in lipid mixtures [13,1and may originate from lipid-phase segregation in the mbrane as in the case of raft formation [15]. Then, the preresults suggest that cutin and suberin monomers, espebecause of their molecular shape, induce phase segregin DMPC membranes, resulting in a marked modificationthe phase transitions and the spontaneous vesicle curva

Further investigations were done by31P solid-state NMR,which is a powerful technique for monitoring the coextence of various phases in phospholipid mixtures. At a mratio of 0.01,31P NMR produced a bilayer powder patterncoexistence with an isotropic line, the intensity of whichcreased with the temperature and molar ratio as shown Fin the case ofω-OHplm. The presence of superimpostwo components on the NMR spectra indicates the pence of a two-phase system in slow exchange at the Ntime scale. The powder component can be undoubtedlycribed to the large vesicles as observed by phase-conmicroscopy whereas the isotropic line may occur for seral reasons: particular orientation of the phospholipid phead, local motions in the membrane, phospholipids emded in a cubic phase, and more probably, small size collas previously suggested.

These mixtures were then investigated by quasielalight-scattering experiments to determine the particle(not shown). In the case of pure DMPC and at a molartio of 0.01, a scattering broad peak corresponding to ves

-

-

n

.

-t

Fig. 2. Selected31P solid-state NMR spectra of the mixture of DMPC aω-OHplm at various molar ratios and temperatures. (A) 0.01 at 308 K;0.01 at 323 K, and (C) 0.5 at 323 K.

of diameter higher than 2 µm was obtained. At a molartio of 0.1, the large vesicles were still observed togetwith colloids having a diameter of about 300 nm withbroad polydispersity. The presence of these small sizeloids is consistent with DSC data and with the isotropic las observed by solid-state NMR. At higher molar ratios,small colloids were still observed together with large ojects, the latter corresponding to the presence of fattycrystals.

Fig. 3 shows the lipid dispersion as viewed by TEM amolar ratio of 1 in the case ofω-OHplm after sedimentatioof the fatty acid crystals. Vesicles having a diameter of ab300 nm were observed. This is in very good agreement

510 J.-P. Douliez / Journal of Colloid and Interface Science 271 (2004) 507–510

andnta-

idss to

andingly[16or

kyltherintheersver

thebe

ingin

byer-ers

col-art-ersemrs.andcur

ntoonglat-pro-

ne-theper-size,rig-be-ranesentalltial

r-ernri-andes-

50–

y 61

to,

.

. 10

, in

ys.

45.03)

hys.

95)

39.

Fig. 3. Transmission electron microscopy of a mixture of DMPCω-OHplm at a molar ratio of 1, stained with uranyl acetate after sedimetion of the fatty acid crystals. Scale bar corresponds to 1 µm.

the light-scattering data and prove that small-size colloare formed upon addition of cutin or suberin monomermembranes.

Altogether, these results are evidence that cutinsuberin monomers are membrane perturbants. Interestmost of these monomers are known as bola-amphiphilesand the perturbing effect may arise from the hydroxylcarboxyl moieties localized at both extremities of the alchain. Note that the membrane hydrophobic thickness incase of DMPC is about 25 Å [17] while the cutin and subehydrophobic string length is about 14 Å. In other words,perturbation of the DMPC membranes by such monomcan be associated with the hydrophobic mismatch. Howethe positioning and localization of the monomers withinphospholipid membrane in the small colloids remain todetermined.

Beyond these fundamental aspects, the present findare of particular interest to bio-mimic what happensplants: first, for the elucidation of the path followedmonomers when transported out of the cell to the polymization site (it is shown here that these insoluble monomcan be cosolubilized with phospholipids as small-sizeloids which can serve as vehicles through the plant compments); and second, to predict the effect of free monomthat are released upon pathogen attack on the cell mbranes localized in proximity of cutin or suberin polymeThis work shows that even at low concentrations, cutinsuberin monomers induce a change in the membranevature. Although cells are certainly not transformed ismall colloids, monomers can more probably induce strperturbations in the cell membrane, potentially moduing its functions and that of associated membraneteins.

,]

,

s

-

-

4. Conclusions

It is shown here that cutin and suberin monomers petrate model membranes and exhibit a limit of solubility inphospholipid matrix. These monomers are membraneturbants and induce a marked decrease of the vesicleforming small colloids. The perturbation is supposed to oinate from phase segregation in the phospholipid bilayercause of the hydrophobic mismatch between the membthickness and the monomer string length. Finally, the prefindings bring valuable information on the routing as smcolloids of these monomers in plants and on their poteneffect on cell membranes.

Acknowledgments

I thank J. Davy (LPCM, INRA, Nantes) for having peformed the DSC experiments, L. Lachmanski from Malvfor giving me the opportunity to perform the QLS expements, C. Mangavel and B. Bouchet for the realizationhelp with the TEM experiments, and D. Marion for suggtions and helpful discussions.

References

[1] A. Heredia, Biochim. Biophys. Acta 1620 (2003) 1–7.[2] A. Heredia, A. Matas, E. Dominguez, Biochem. Educ. 28 (2000)

51.[3] J. Graca, L. Schreiber, J. Rodrigues, H. Pereira, Photochemistr

(2002) 205–215.[4] N. Cordeiro, P. Aurenty, M.N. Belgacem, A. Gandini, C.P. Ne

J. Colloid Interface Sci. 187 (1997) 498–508.[5] P.E. Kolattukudy, Adv. Biochem. Eng. Biotechnol. 71 (2001) 1–49[6] P. Schweizer, G. Felix, A. Buchala, C. Müller, J.P. Métraux, Plant J

(1996) 331–341.[7] M.R. Vist, J.H. Davis, Biochemistry 29 (1990) 451–464.[8] J.-P. Douliez, D. Lavenant, D. Renard, J. Colloid Interface Sci.

press.[9] T. Inoue, Y. Yanagihara, S.-i. Misono, M. Suzuki, Chem. Ph

Lipids 109 (2001) 117–133.[10] R. Koynova, M. Caffrey, Biochim. Biophys. Acta 1376 (1998) 91–1[11] F. Castelli, S. Caruso, N. Uccella, J. Agric. Food Chem. 51 (20

851–855.[12] T. Heimburg, Biochim. Biophys. Acta 1415 (1998) 147–162.[13] H.-G. Döbereiner, J.D.N. Käs, I. Sprenger, E. Sackmann, Biop

J. 65 (1993) 1396–1403.[14] E. Sackmann, FEBS Lett. 346 (1994) 3–16.[15] J.R. Silvius, Biochim. Biophys. Acta 1610 (2003) 174–183.[16] J.B.F.N. Engberts, D. Hoekstra, Biochim. Biophys. Acta 1241 (19

323–340.[17] J. Douliez, A. Leonard, E. Dufourc, Biophys. J. 68 (1995) 1727–17