Investigation of polyene macrolide antibiotic-induced

Investigation of polyene macrolide antibiotic-inducedpermeability changes in vesicles by 31Pnuclear magnetic resonance

Can. J. Biochem. Downloaded from www.nrcresearchpress.com by Simon Fraser University on 12/28/15 For personal use only.

H. Id. PIERCE, JR., A . M. UNMU, AND A. @. OEHLSCHLAGER Department of Ckernistry, Sirnorz Fraser University, Burnaby, B.C., Canada VSA IS6 Received October 4, 1977 Revised February 20.11978 Rerce, H. D. Br., Unrau, A. M. & Behlschlager, A. C. (1978) Investigation of polyene macrolide antibiotic-induced permeability changes in vesicles by 31Pnuclear magnetic resonance. Can. J , Biochern. 56,801-807 The permeability of egg yolk lecithin (EYL) vesicles to R3' has been measured by 31Pnuclear magnetic resonance (nmr) spectroscopy. Measurable Pr3+leakage into the rnternal aqueous compartment of EYL vesicles at ambient (21°C) temperature required the presence of small (7-10 mol%) amounts of dicetyl phosphate (BCP). The permeability of DCP-containing vesicles is decreased by incorporation of sterol (cholesterol > ergosterol -- 5,6-dihydroergosterol > zymosterol) into the lipid bilayer. Addition of the polyene macrolide antibiotic, nystatin, to DCP-containing EYL vesicles with and without sterol resulted in increased Pr3' permeability at the three temperatures studied (21-37.5"C). Permeability changes observed upon addition of nystatin to sterol-impregnated. DCP-containing vesicles varied with sterol structure: ergosterol = 5,6-dihydroergosterol > cholesterol = zymosterol. These results are compared with other polyene macrolide induced permeability changes on model and natural membrane systems. Permeability changes induced by nystatin in sterol-free EYL vesicles were generally greater than for comparable sterol-containing vesicles. This is attributed to a nonspecific interaction of the antibiotic witk the latter vesicles. Pierce, H. D. Jr., Unrau, A. M. & Oehlschlager, A. C. (1978) Investigation of polyene macrofide antibiotic-induced permeability changes in vesicles by 31P nuclear magnetic resonance. Can. J . Biochenr. 56.801-807 Nous avons mesure la permeabilite des vesicules de lecithine de jaune d'euf (EYL) au Pr"" par spectroscopie de resonance magnetique nucleaire du 31P.Pour mesurer la perte du Pr" dans le compartimcnt aqueux interne des vCsicules de EYL B temperature arnbiante (2I0C), il h u t la presence de petites quantites 47-10 mol%) de dicktyl phosphate (DCP). La permeabilite des vesicules contenant du DCPest diminuee par I'incorporation d'un stkrol (cholestkrol> ergosterol -- 5 ,B-dihydrsergssterol > zymostCrol) dans la double couche lipidique. L'addition d'un antibiotique macrolide polyknique, la nystatine, aux vesicules EY L contenant du DCP, avec et sans sterol, augmente la permkabilite au PrH aux trois tempkratures essayees (21-37.5"C). Les changements dc permeabilitk causks par ['addition de nystatine aux vesicules contenant du DCP et impregntes de sterol varient selon la structure du sterol: ergostkrsl 5,6-dihydroergostCro1 > cholestkrol zymostkrol. Ces rksultats sont cornparis avec les changements de permkabilitk induits par d'autres macrolides polyeniques sur des modeles et des systemes membranaires naturels. Leschangements de permeabilite suscitis par la nystatine dans des vksicules EYL sans sterol sont generalement plus importante que dans des vCsicules comparavec stkrol. Ceci est attribue B une interaction non sp6cifique de I'antibiotique avec a b l e ~mais , ces dernikres vesicules. [Traduit par le journal]

-

antibiotics for the principal sterols in these cell types. The polyene macrolide antibiotics are considered to Indeed, numerous studies have shown that this class of exert their lethal action by an association with the cyto- antibiotics increases permeability to ions and small phsmic membranes of sensitive organisms that results in molecules of sterol-impregnatedmembranes and that the the loss of the selective permeability properties of the permeability changes are greater for membranes impregmembrane (1). Organism sensitivityto these antibiotics is nated witk fungal sterols (ergosterol) than mammalian imparted by the presence of sterols in the cytoplasmic sterols (cholesterol)(1). In parallel, there has emerged from sterol content membrane. The selective toxicity of polyene macrolides toward fungal cells compared with mammalian cells is yeast culture resistance studies (2, 3) evidence that the considered to reside in the differential aflinity of these development of resistance to low levels of polyene macrolides correlates with loss of b5.7-unsaturationin sterols EYL, egg yolk lecithin; nmr, nuclear of surviving cells. As more resistant cells are selected by ABBREVIATIONS: magnetic resonance DCP, dicetyl phosphate; tlc, thin-layer growth om an antibiotic-containingmedium, nuclear rawchromatography; DMF, dimethylformamide. saturation of major sterols in surviving cells shifts from

CAN. J. BHOCKEM.


ergosterol

cholesterol

A3 to A8 (e-g., 5,6-dihydroergostero1 to zymosterol). At higher levels of resistance (e.g., 1008 U raystatin per milMitre) increasing amounts s f A8-4,4-dimethyl and A84,14-dimethyl nuclear-methylated sterols are found in surviving ceiHs (2, 3). However, there are numerous exceptions to this pattern of development of resistance to polyene, and attempts to correlate sterol structure and polyene resistance have met with only moderate degrees of success. Other factors, e.g., the cell wall, appear to influence polyene sensitivity, but, at present, little is known about how these mechanisms influence the action of pol yenes. Model membrane systems have been extensively employed in recent years to examine a broad range of membrane phenomena since many complexities associated with membranes of the living cell are absent. In the present study, we have examined the effects of the pdyene rnacrolide nystatin on the permeability of kilayered vesicles impregnated with sterols found in sensitive and resistant yeast cultures. The method, which a h avoids the complications of v dialysis, is based on the observation by B y s t r ~ ~and co-workers (4-7) that the magnetic resonance spectra of the 31Pnuclei on either side of the phospholipid bilayer of vesicles can be distinguished by addition of a small amount of lanthanide shift reagent to the vesicle exterior phase. The lanthanide ion associating with the exterior phosphatidyl groups causes a significant downfield shift of the resonance due to the exterior 31Pnuclei. The "'P nuclei of the interior phosphatidyl groups shift only a smdi amount upon addition of lanthanides to the vesicles indicating that the bilayer restricts access of the Hantha i d e ion to these nuclei. It was anticipated that if the vesicle bilayer of such preparations were rendered permeable to the lanthanide by addition of polyene macrolides, the rate at which lanthanide ion concentration in the environs of the interior phosphatidyI phosphorus approached that in the exterior phase would be reflected by the rate of merging of the high field 31Psignal (interior) with the low field 31P signal (exterior),

Materiais and Methods DCP, Trizma-HC1, and Trizma-Base were obtained from Sigma Chemical Co. Pr(N03)3.5H,0 was obtained from Alpha Inorganics. Ergosterol (Matheson, Coleman, and Bell) and cholesterol (Fisher Scientific Co.) were recrystallized from methanol - ethyl acetate. 5,6-Dihydrwrgosterol, zymosterol, and ianosterol were obtained from local sources (2). Nystatin was a gift from E. 8.Squib$ and Sons, Ltd. Pur@catiurz uf Lecithin EYL (Sigma or extracted from fresh egg yolks) (8) was piarified by chromatopraphy on acid-washed alumina (Act I) previously sieved to remove particles greater than 180 mesh. The column (200, alumina:l, lecithin) was developed with CPIC1,-methanol(3: 1, vlv) and 10-rnl fractions were examined by tlc (sijica gel 6 , CHC1,-MeCbH-H,O, 65:25:4, v/v/v, with H,SB, charring). Fractions containing phosphatidyl choline eluted at 400 and 5CXBmB and exhibiting only one component by tlc were combined. Subsequent soivent removal and storage operations were performed under nitrogen. Preparation of Vesdc8cs The appropriate amounts of lipid components were dissolved in CH61, and DCP in CWC1,-MeOM solution was added. Solvent was removed in vacua and the lipid film was kept under vacuum overnight. Degassed Q.O5M Tais buffer (pH 7.3) was added and a coarse dispersion was formed by shaking until no lipid was left on the vessel walls. The coarse dispersion was transferred to ajacketed, side-arm test tube and sonieated under N, at 15-28°C. The sonicator was a Bronwill Scientific Biosonik I11 equipped with a 5 x 3132in. (I in. = 25.4rnm) titanium probe and operated at its maximum output (approximately 100 acsustical watts). hasst lecithin dispersions (=18% wlv) became translucent after 2.4-3 h and were freed of undispersed lecithin, large liposomes, and titanium fragments by centrifugation at 40WOg for 2 h a t 4OC. Dispersions were stored at 4°C and used within 2 days of preparation. The dry weight of the centrifugation pellet (including titanium fragments) was 0.5-495 of initial lipid weight. Initial experiments were accompanied by tlc analysis of the lecithin dispersions a s described above for lecithin purification. We did not ohserve the formation of fatty acids or lysolecithins during this procedure for our initial experiments. It was assumed that this degradation was not important under our ssnication conditions. Vesicle Size Disbribkation The procedure of Lau and Chan (9) was employed with only minor modification for- estimation of the size distribution of the vesicle preparations by electron microscopy. Negatively stained (2% phosphotungstic acid, pH 4.4) veside preparations on 20g)-meshcopper grids coated with Formvar and a thin film of carbon were examined on Philips EM300 electron microscope operating at $0 kV and at magnification Level 15 (29000). Final magnification of the electron photomicrographs was $9 900. The following criteria were employed for measurement of vesicle dimensions: ( I ) vesicles had to be isolated and not touching ang neighbors; (2) for slightly nonsphericai vesicles, the mear 'diameter9 is an average d the longest and shortest dimensions and (3) diameters of vesicles were measured to the neares 2 10 nm. The sizes of approximately 500 vesicles were measurec for several vesicle preparations. The data were treated an6 plotted according to Lau and Chan (9). Kinetic Procedwes Vesicle suspensions, 2-2.51~11,in 12-mm nmr tubes wer


PIERCE ET AL.

FIG. 1. (a] 31Pspectrum of O.1M EYE vesicles in water; (b) "P spectrum of l(ab sampie in the presence s f 0.014M h(N03b3;(c.) 'P spectrum of I(h) sample with 1% finai concentration of Triton X-100 after a 5-min boiling; (d)3iPspectrum of O.1M EYL vesicles prepared in 0.084M m N 0 3 ) 3.5H,O then dialyzed against 0.OlM Tris buffer to 10-4M [NO3-].

maintained at constant temperature (iz0.5"C) in a water bath. N ystatin was introduced into the suspension in DMF solution by rapid syringe injection. The concentration of nystatin stuck solution was adjusted so that a 100-pl aliquot woziid provide the desired concentration in the vesicle suspension. An equivalent m o u n t (100~13of DMF was added to control samples. The suspension was incubated with wystatin for 1-2 h, and then a loQ-pI aliquot of h(NO,), was added to the control and test samples to give a final P~(No,), concentration of=§ mdm]. The 31Pspectrum of a control sample was observed to establish the chemical shifts of inside and outside 3 B Psignals. The chemical were measured in shifts s f 3 B Psignals of test vesicle hertz relative to the chemical shift of the 3'P signals in the appropriate control. Most kinetic runs were carried Out using three control samples and three or four test samples. If probe and bath temperatures were identical, each test sample was used for the determination of several points. If probe and bath temperatures were different, each test sample was used for only one point. For each kinetic point, at least 2048 accumulations were (approximately 9- 17 min) required. The errors in determination of chemical shifts were typically 5-BOHz. "'B Pabe, Fourier Tmnsjbrrn Spectroscopy

Unless otherwise stated, 31P spectra were obtained at 15.08MHz on a modified NMR-Specialties spectrometer with a home-built crossed coil, locked on external H,O, and interfaced to a Nicolet 1082 FT system developed by Professor Wells of this department (Higgs, T. B., Brooke, A. & Wells, E. J . , unpublished work). A pulse angle of agproximateBy 25" and pulse

repetition rake of 300-600 ms gave maximum signal to noise in a minimum time (10).

Results and Discussion initial experiments were directed at the determination of the and pr3+ permeability ofEYL vesicles of prepared Y' sonicatiOn 12). The the vesicles in our preparations were found by electron microscopy (91 to Possess a single bilayer and to have diameters of less tkan 35nm.They gave high resolution 31Pspectra (Fig. ICE)over several weeks. In agreement with the observations of Bystrov and co-workers (4-T), we observed that addition of Pr(NO,), to vesicles resulted in a large downfield shift of the resonance of the exterior nuclei but only a slight shift in of interior ' I P (Fig. Ib). the position of the Addition of Triton X-100 and boiling caused immediate (5 min) merging of the high a d low field 1Psignals 4Fig. Ic). That complete rupture of the vesicles occurred under these conditions was indicated by the addition of sufficient EBTA to complex all Pr3' whereupon a single lp as in Fig. l a was observed. The change in the sh8t of the 3'P nuclei was a nearly linear function of Pd-(N03)3added to the vesicle preparation (Fig. 2) indicating rapid reversible coordination of the Pr3+ with external 31B nuclei. That ("9

C A N . I. BIOCHEM. VOL. 56, 1978

TABLE 1. Effect of DCP concentration an leakage of PI." +into EYL vesicles DCP,


Temperature, "C

moIx

Ergosterol, molz

Nystatin

Leak rate, Wz/ha Nil Nil Ni 1 Nil ~ 0 . 3 1 27 > 140

OChanpe in chemical shift o f inside JIPresa *Sixteen hertz after 1 week. rPhospholipid:nystatin ratio.

The observation that EYE vesicles without DCP were not permeable to R3+is in agreement with the observation of Bystrov and co-workers (4-7) that such preparations were not permeable to highly charged ions such as Mn2* and Eu3* over 24 h and were stable up to 80°C in their presence. The low permeability of B3+ in the vesicles not containing BCP when treated with nystatin indicates that, even over extended periods, nystatin was not absorbed by the vesicles to a sufficient extent to induce Pr3+leakage t o the interior. The permeability changes in EYL vesicles treated with nystatin due to the presence of DCP noted in the present work is in contrast with the results of earlier workers. Sessa and Weissmann (13) have reported that addition s f stearylarnine or DCP t o EYL vesicles did not alter their permeability to encapsulated charged ions such as chromate or glucose when they were treated with polyenes. They concluded from these results that there was no specific interaction between the polyene and these charged bilayer components. The apparent ability of DCP to increase Br3+ permeability in EYL vesicles treated with nystatin in the present study may be due to the small size of the vesicles used and the ability of the polar additive, Pr3+, and polyene to engage in tertiary interaction. Under conditions where R3+leakage was FIG.2. Chemical shift of resonance due to exterior 31Pnuclei observed, the initial 32Bspectrum of the vesicles in Br3+ (from original position in water) versus concentration sf solution varied with time as shown in Figs. 3 and 4. The spectra in Fig. 3 represent the expected change for single R(N03),. Vesicle preparation was 8.8% (wlv). bilayered vesicBes with no B signal intensity remaining exchange with internal phosphatidyl 3ZPnuclei was also at the origin of the signal due to inside "P nuclei. In fact, rapid and reversible was demonstrated by observation of this type of spectral change was seldom encountered in the 3BPspectra of EYL vesicles with encapsulated R3+. the present study. More typically, spectra a s shown in These vesicles, prepared by vesicailation of EYL in the Fig. 4 were encountered in which the initial signal atpresence of Pr(NO,), followed by dialysis of the solution tributable to the inside 3 i P nuclei gradually became two against water, gave the spectra shown in Fig. Id. distinguishable signals, one migrating toward that due to the external "P signal at a rate faster than the second. Effects qf DCP Initial Pr3+permeability experiments conducted by the The faster moving 31Psignal generally accounted for addition s f the pslyene macrolide, nystatin, to EYL- greater than 60% of the intensity attributable to the initial ergosterol vesicles in Pr3+ solutions indicated that no inside 31Presonance. It was broad in appearance, apparmeasurable merging of the 3 i Presonance of the interior ently reflecting inside 31Bresonances of single bilayered phosphatidyl group with that due to the exterior phos- vesicles possessing differing concentrations of R3*in ghatidyl g o u p could be observed over a week unless their internal aqueous compartments. The slower migratDCP was added and the vesicles were maintained at ing 31Bsignal (maximum 12% intensity overall) was considered to be due to multilayered vesicles, which alslightly above ambient temperature (Table 1).


PIERCE ET AE.

NMR QI EY L vesicles

1 A\

F ~ G3.. 31Pspectra of EYL vesicles (0.1Aa) containing 7 mol%

DCP and 7 mol% dihydroergosterol in 2 rnl aqueous solution, 0.05M in Tris buffer, 0.012M in Pr(N03),-5H,0,and lW3Min nystatin (added in 100pl DMF). Vesicles preincubated with nystatin prior t o addition of Pr3+.

though permeable to Pr3+,were much less so than single bilayered species. When the spectra of Fig. 4 were obtained at 40 MHz rather than 15 MHz, the signal of the faster moving inside 3'P resonance became extremely broad and was generally lost in the base line noise. This latter observation makes apparent the basis for recent R3+permeability studies which utilize the loss of inside 3 E Psignal intensity as a measure of Br3+permeability of EYL vesicles (14). In this work, the rates of leakage of Pr3+ into the internal vesicle compartment were measured by the rate of movement of the hster moving internal 31Psignal. In cases where this signal was not easily resolved from the slower moving 31Presonance, additional shift reagent was added (Fig. 4, bottom trace) to resolve the signals.

Elc. 4. 31Pspectra of EYL vesicles containing 7 mol% DCP, and 10 mol% dihydroergosterol, in 2 rnl aqueous solution, O.QSM in Tris buffer, 0.012M in R(N03),.5H,O, and BW3M in nystatin (added in 100pl DMF). Vesicles preincubated with nystatin prior to addition of R3+.

dihydroergosterol > zymosterol. This result is in general agreement with cation permeability studies using other analytical methods (1). The data in Table 2 indicate that the permeability of EYL vesicles with and without sterol increases with increasing temperature. The limited data obtained indicate that the permeability of vesicles containing ergosterol is more sensitive to temperature than those without ergosterol in the range of 21 to 323°C but that permeability changes of the latter are more sensitive to temperature above 32.5"C. It has been previously noted in model membrane cation permeability studies (18) that the permeability of sterol-containing membranes is more sensitive to increasing temperature than those devoid of this component due to the decreasing influence of sterol on the molecular packing order of the phospholipid acyl chains with increasing temperature. In the present experiments, we limited our comparisons to the range 21 to 323°C where this trend was in evidence.

Ejyects of Sterol and Temperature It has been reported that for phosphatidylcholine, cation permeability decreases upon the incorporation of ergosterol or cholesterol into the lipid bilayer (15, 16). Efiect of Nystadire This decrease is considered to be due to an increase in Addition of nystatin to EYL vesicles invariably caused molecular packing order attending the addition of sterol an increase in Br3+permeability. The effect is dependent (17). In the present experiments, incorporation of these most strongly on the temperature and phospholipid: sterols, 5.6-dihydroergosterol or zymosterol, decreased antibiotic ratio. At 21°C and at a phospho1igid:nystatin RHpermeability of the vesicle bilayer. The ability of the ratio of 50:1, vesicles without added sterol leak Pr3+ sterols to decrease PrH permeability decreases in the much more rapidly than those with sterol (Table 3, Fig. following order: cholesterol > ergosterol = 5,6- 5). Addition of Pr3+after leakage to equilibrium of vesi-

CAN. 9. BBOCHEM. VOL. 56, 1978

TABLE 2. Effect of temperature on Pr3 permeability in EYL vesicles with +

ergosterol and without ergosterol Temperature, "C

DCB, molz

21

Leak rate,

I0 mol%

&/h"

7 7 7 7

32.5


&glsstt3r08,

86 124

4.3 6.2 27 76

-I-

-I-

-

10

1 1.8

0.05 0.09

-I-

10

37

Relative rate

540

1420

°Initial rate of shift per hour of resonance due to faster moving internal "P.

TABLE 3. Effect of nystatin on relative leak rates of PrJf into sterol-impregnated EYL vesiclesa

Phospholipid :nystatin -

Cholesterol ErgosteroI 5,6-Dihydroergosterol Zyrnosterol No sterol

Temp., "C

Control

100/ 1

5011

32 32 21 32 32.5 32.5 28

1.8

1.7 0.02 2.2 2.2 2-5 0.84

0.99 2.5 3.3 2.4 26.5 -

1.3 -

125

-

-

-

---

Rate enhancement 1.0

1.5 65

1.5 1.1

18.6 3114

"Relative leak rates of Pr3f into EYL vesicles as measured by downfield movement of the 3'P resonance o f interior "P nuclei in hertz per hour. Vesicles (0.100 MIkf)containingBO anof7 or no sterol and 7 m o l x DCP were suspended in 4 ml0.05 M Tris buffer. DMF (200 pi) was added to the controls, &statin added in 200 crl B M F to a final concentraM and allowed to greincubate with vesicbs prior to addition of Pr(N03j3in a 280-pl tion o f 2 Y 10-3 or 1 >( buffer (lo-= M final concentration). The phosph0lipid:nystatin ratio given is based on total available phospholipid and total available antibiotic and does not take into Zpccsunt the possibility that the solubility of nystatin in sterolContaining membranes could vary with the nature of the sterol.

cles without sterol resulted in a 31Pspectrum with only one peak indicating these vesicles had been lysed or made so permeable that Pr3+equilibmtion was complete in 1Qmin. Addition of R3+to vesicles with ergosterol (Fig. 5) after near equilibration resulted in the downfield shift of only the ?'P resonance attributable to the outside phosphatidyl group (as in Fig, 4, bottom trace). Hence, these vesicles were still intact after exposure to the high concentration of nystatin used in this experiment. At 21°C and at a phospho1ipid:antibiotic ratio of 108:1, the Br3+Beak rates of vesicles with and without sterol became exceedingly slow. M e r 2 days, vesicles of both types (7% DCP) treated with nystatin showed no shift of the signal due to the interior 36Pnuclei. At this phospho1ipid:nystatin ratio, a temperature of 32OC was required to obtain measurable Pr3+leakage. At both nystatin concentrations, the vesicles without sterol were made significantly more permeable to Pr3+ than those with sterol (Table 31, and this difference increased with increasing nystatin concentration. In all experiments conducted, nystatin-treated vesicles, devoid s f sterol, became at least an order of magnitude more permeable to Br3+ than those containing sterol and did not maintain integrity of their internal aqueous compartment. We suspect, therefore, that in these cases nystatin is severely disrupting the bilayer system by a mechanism different than that operating in the case of vesicles containing sterol.

The effect of sterol structure on nystatin-induced permeability changes in sterol-impregnated vesicies is shown in Tab%e3. The ability of this polyene to increase vesicle permeability to R3+varied according to sterol structure with ergosterol and 5,6-dihydroergssterol conferring more sensitivity to nystatin than cholesterol or zymostersl. These relative permeability changes are those expected from studies of sterol composition of yeast strains resistant to nystatin. In these studies 12-31, culture sensitivity (and hence increased cation permeability induced by nystatin) is generally found to increase as ring B unsaturation of the major culture sterol is altered: > A7 > A8* It is noteworthy that even at the IIsw antibi0tic:sterol ratios (1:10) used in these experiments enhancement of vesicle permeability due to the presence of antibiotic was measurable. Comparative studies of permeability changes induced by nystatin at lower antibiotic:sterol ratios were thwarted by insolubility s f the antibiotic. Decreasing the antibiotic:sterol ratio to 1:s generally increased the rate enhancement observed by ergosterolcontaining vesicles by twofold but was accompanied by noticeable turbidity and precipitation due to undispersed nysfatin. Permeability changes in phospholipid bilayers have been reported (18-20) to increase as a power of polyene macsolide concentration and vary according to the relationship: A515


PIERCE ET AL.

=O

807

Acknowledgements We wish to thank Dr. M. Fryberg for suggesting (1972) this approach to the study ofthe polyene macrolide sterol interaction and Dr. A. M. Pierce for measurement of vesicle size distributions. We gratefully acknowledge the Simon Fraser University Resident's Research Fund and the National Research Council of Canada for financial support of this work.

-e ~cleswtth ergosterol treated w ~ t h nystat~n

1

0

1

I

4

,

1

r

12

I

-

T

16

T

-

;0';4'2k13;h

TIME IN HOURS

FIG.5. Rates