Membrane lipid response to clathrin coat protein determined by ...

1 downloads 0 Views 523KB Size Report
Clifford J. Steer$, James S. VincentST, and .... and A, E, and C are adjustable parameters. The band ..... H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.. 17.
Communication

THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 259, No. 13. Iseue of July 10, pp. 8 0 5 2 ~ 5 5 , 1 9 & Printed in U.S.A.

Membrane Lipid Responseto Clathrin Coat Protein Determined by Infrared Spectroscopy

the primary role of the clathrin coat is to anchor receptorligand complexes in coated pit regions in order to inhibit lateral diffusion (8), there is little evidence to support this function. A recent proposal attributes an active role to the clathrin coat in the regulation of membrane fusion. Absence of the coat protein, however, promotes fusion between memPOSSIBLEINVOLVEMENT IN COATED VESICLE branes of coated vesicles and those of intracellular organelles FORMATION* (9). Although the clathrin coat has been implicated ina (Received for publication, March 14, 1984) number of processes, a major functionappears to be its Clifford J. Steer$, James S. VincentST, and involvement in the selective pinching offof the associated Ira W. Levin5 membrane underlying the lattice structure. The dynamics of From the $Laboratory of Biochemistry and Metabolism and this process are, however, not well understood. the $Laboratory of Chemical Physics, National Institute of In an attempt to explain the unique properties of the coated Arthritis, Diabetes, and Digestive and Kidney Diseases, pit region, several studies have focused attention on the National Institutes of Health, Bethesda, Maryland 20205 characteristics of the constrained portion of the membrane (10, 11). The original reports by Pearse (12) that coated Clathrin,themajorstructuralprotein associated vesicles arealmost devoid of membrane cholesterol were with both coated pits and coated vesicles, has been recently corroborated by electrons micrographic studies emimplicated inthe dynamics ofvarious endocytotic proc- ploying filipin, a polyene antibiotic, as acholesterol probe (10, esses. I n a n a t t e m ptot define the mechanisms involved 11).In the latterstudies, the absence of filipin binding in the in the transition from uncoated membranes to clathrin- coated pit regions was interpreted as evidence for a lack of coated pits and then to coated vesicles, we investigated detectable cholesterol in the membrane. These results sugby infrared spectroscopy the lipid perturbations arising from the interactions of the clathrin coat with gested the that coated membranes represent specialized domains bilayers of intact membrane assemblies.A comparison which act as molecular filters for removing cholesterol. This sterol displacement was considered to be important in allowof t h e lipid acyl chain symmetric methylene stretching modes at -2850 cm-’ f o r isolated clathrin-coated ves- ing the membrane to adopt the appropriate curvaturerequired icles, uncoated vesicles, and synaptic membranes at for coated vesicle formation. The role of the clathrin coat in 21, 38,a n d 50 “C indicated that clathrin significantly the development of such a specialized membrane is, however, increases the number of gauche chain conformers in not known. In thepresent report, we examine by infrared spectroscopy the bilayer matrix of t h e coated vesicle system. The on the lipid increase in lipid disorder at 21 “C, accompanying the the perturbation of the clathrin coat protein observed 0.44-cm-’ frequency increase for coated ves- matrix of the isolated coated vesicle membrane. Infrared icles compared to uncoated vesicles, is approximately spectroscopy provides a sensitive technique for monitoring equivalent to the intrachaindisorder incurred inheat- the structural and dynamic properties of both the lipid and ing liquid crystalline dimyristoylphosphatidylcholine protein components present in membrane assemblies. In this multilayers by -10 “C. The implications of these re- communication, wewill be concerned specifically with the sults oncoated vesicle formationare discussed. methylene symmetric carbon-hydrogen stretching modes which are characteristic of the intramolecular acyl chain order of the bilayer lipids (13, 14). Specifically, the frequency of these modes, which appears a t approximately 2850 cm”, is Membrane coatedpits andcoated vesicles have been impli- sensitive to thenumber of gauche bond conformers along the cated in a number of fundamental intracellular functions in lipid hydrocarbon chain. We report and assess the detailed eukaryotic cells (1-4). In particular, their involvement in the frequency shifts and temperature dependence of this spectral endocytosis of biological macromolecules by specific receptor band for coated and uncoated vesicles and for synaptic memsystems has attracted considerable attention. The basic unit branes isolated from bovine brain preparations. Our results of the membrane coat, the triskelion, existsas a M, = 650,000 indicate that the presence of the clathrin coat results in a protein trimer of clathrin subunits (M, = 180,000) in combi- marked conformational disorder in the lipid bilayer matrix. nation with three clathrin-associated proteins (M, = 30,000This membrane perturbation may represent a fundamental 36,000) (5). Assembly of the triskelions on the membrane requirement for the structuralrearrangements associated with leads initially to aplanararray of hexagons which upon the formation of coated pits and coated vesicles. subsequent growth forms a complex, closed polyhedral lattice EXPERIMENTAL PROCEDURES structure associated with the coated vesicle (6, 7). Despite the recent focus of attention upon the assembly Materials-Mes’ was obtained from Calbiochem-Behring. Deutercharacteristics of the clathrin coat, our understanding of its ium oxide was purchased from Aldrich, and ultrapure sucrose was function remainslimited. Although it hasbeen suggested that from Bethesda Research Laboratories. All other chemicals were re-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely t o indicate this fact. llOn sabbatical leave from the Department of Chemistry, University of Maryland, Baltimore County, Catonsville, MD 21228.

agent grade or better. Bovine brains were obtained from Treuth Slaughterhouse (Catonsville, MD) and were kept on ice until processed within 2 h of slaughter. The abbreviations used are: Mes, 2-(N-morpho1ino)ethanesulfonic acid EGTA, ethylene glycol bis(p-aminoethyl ether)N,N,N’,N’-tetraacetic acid.

8052

Spectra Infrared

ofCoated Isolated

Isolation of CoatedVesicles-Coated vesicles were isolated from bovine brain according to the method of Nandiet al. (15). The procedure consisted of a combination of low speed (10,000 X g) and high speed (100,000 X g) centrifugations, including a final single-step gradient centrifugation through an 8% sucrose/D20 solution maintained at pH 6.5 with the same buffer salts as included in the homogenization buffer (Buffer A 0.1 M Mes, pH 6.5, 1 mM EGTA, 0.5 mM MgCl,, and 0.02% (w/v) NaN3). Further purification of the coated vesicles was achieved by Sephacryl S-1000 (Pharmacia Fine Chemicals) gel filtration column chromatography. Typically, 35 mg of coated vesicles in a total volume of 7 ml wereapplied to a Sephacryl S-1000 column (2 X 85 cm) equilibrated with Buffer A. The column was eluted at approximately 50 ml/h in the same buffer, and the samples were collected in 2-ml fractions. The column-purified coated vesicles eluted in the included volume as a well defined peak, as determined by its absorbance a t 280 nm (9). The coated vesicles were pooled and concentrated by centrifugation at 100,000 X g for 60 min. The pellet was allowed to resuspend slowly overnight in 10 ml of Buffer A. Uncoated vesicles were prepared by dialyzing approximately 5 ml of the column-purified coated vesicles against 10 mM Tris-HC1, pH 8.2, for 16 h at 4 "C. The clathrin coat protein inthe supernatantwas removed after pelleting the uncoated vesicles at 100,000 X g for 60 min. Approximately 90% of the clathrin coat protein was dissociated as determined byLowry protein assays (16) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The uncoated vesicles were gently resuspended in several drops of 100 mM Tris, pH 8.2, to attain a protein concentration of approximately 50 mg/ml. A similar protein concentration was used for coated vesicles which, after a similar high speed centrifugation, were resuspended in 0.1 M Mes, pH 6.5. Spectroscopic measurements of both coated vesicles and uncoated vesicles at high protein concentrations were determined within 2 h after resuspension. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and thin section electron microscopy were performed as previously reported (17) to assess the purity of the samples. Preparation of Synaptic Plusma Membrane-Synaptic plasma membranes from bovine brain were isolated according to the procedure of Jones and Matus (18).The plasma membrane band a t the 28.5-34% (w/w) sucrose interface was collected, diluted 2-fold with cold distilled water, and centrifuged at 87,000 X g for 2 h in a Beckman 45 Ti rotor. The slightly brownish center of the pellet was gently aspirated, and the remaining white pellet of synaptic plasma membrane was resuspended in 100 mM Tris-HC1, pH 8.2, to a protein concentration similar to both the coated and uncoated vesicles. Ifnot used immediately, the synapticplasma membranes were quickly frozen in liquid nitrogen and stored for no more than 1 month at -70 "C in 5 mM Tris, pH 8.1. No changes occurred during that period of time in eitherthe lipid composition or in theinfrared spectra. Total cholesterol content was determined using the Sigma Assay Kit, number 350, by the enzymatic oxidation of cholesterol to cholest4-en-3-one. Sample turbidity was corrected by measuring the sample absorbance in the presence of ascorbate. The correction was generally 3-5% of the total absorbance. The phospholipid content was determined by the method of Fiske and SubbaRow (19) after extraction of the lipid by the procedure of Bligh and Dyer (20). Infrared Spectroscopic Measurements-Infrared spectra were obtained with a Perkin-Elmer 580B spectrophotometer controlled by the National Institutes of Health LDACS (Laboratory Data Acquisition System)computer. LDACS consists of a local LSI-11 computer interfaced with the spectrophotometer and a Tektronix 4006 graphics terminal. The local computer communicates with the laboratory PDP 11-70 computer for data storage and manipulation. Since the line width at haif-height of the symmetric CH2 stretch transition at 2850 cm"is approximately 12 cm", the spectra were observed under moderate resolution conditions utilizing spectral slit widths on the order of 0.85 cm". Data were collected a t 0.1-cm" intervals over the 2900-2800-cm" range. A jacketed, variable path length cell equipped with CaF, windows wasthermostatted a t 21 "C for the spectra of the synaptic membranes and coated and uncoated vesicles. A small amount of the hydrated, pelleted, sample, together with one or two drops of the buffer solution, was placed in the center of one window of the disassembled cell; the opposing window wasthen lowered until the sample formed a uniform, bubble-free film between the CaF, plates. Spectra were recorded, stored, and averaged for each sample. The instrument was continuously purged of water vapor with dry nitrogen gas. Water vapor lines were used for spectral calibration;

Vesicles

8053

vibrational frequencies are reported to within f O . l cm", a conservative estimate of the precision and reproducibility of the spectral scans. Library spectra of both Mes and Trisbuffer solutions, for use in a solution subtraction step, were also recorded in the same conditions for which the sample spectra were observed. The cell path length was decreased in small steps in order to develop extensive libraries for the solvent spectra over a 10 to 3-pm path length range. The library solution spectra chosen for subtraction from the sample spectra matched the sample absorbance at 2800cm". Although the lipid 2850-cm" feature shows little absorbance at 2800cm", the broad water absorption band exhibits appreciable intensity at this frequency. The resulting solution-subtracted spectra were smoothed with a 25-point cubic least squares convolution algorithm. No spectral distortions were introduced by the smoothing routine. The 25-point smoothing algorithm consists of fitting a cubic polynomial to points spanning 2.5-cm" portions of the spectrum, the main feature of which has a line width on the order of 12 cm-I. In order to assign more precisely a peak frequency to the 2850cm" feature, the unsmootbed data points from 2857 to 2843cm" were fit to a single gaussian line shape. The 141 data points between 2857 and 2843cm"were fit by a nonlinear Levenberg-Marquardt least squares minimization technique to a single three-parameter gaussian function, A exp(-(X, - B)'/2C), in which X, is a data point and A , E , and C are adjustable parameters. The band center E of the gaussian curve was fit to anuncertainty of less than kO.02 cm".

RESULTS ANDDISCUSSION

The infrared spectra at 21 "C of each of three separate preparations of coated vesicles showed significant shifts in the methylene symmetric stretching mode at about 2850 cm" to lower frequencies when clathrin was dissociated from the vesicle membrane. Membrane spectral shifts of 0.3-0.7 cm" between the various coated and uncoated vesicle samples were observed by monitoring the frequency at theabsorption maximum either before or after the solvent absorbance had been subtracted. Fig. 1 illustrates the water-subtracted spectra of one preparation of coated and uncoated vesicles at 21 "C for which the maximum absorbance for the uncoated vesicle sample is shifted by 0.5 k 0.1 cm" toa lower frequency compared to that of the coated vesicle sample. The peak frequency of this vibrational feature for the synaptic membrane sample is within 0.2 cm" of that €or uncoated vesicles. These shiftswere determined from fits of the absorption peaks to a gaussian line shape with the band centers determined from the maxima of the fitted curves. Three separate preparations of coated vesicles at 21 "C display absorbance maxima shifted 0.44 f 0.22 cm" (95% confidence level for three samples) to lower frequencies when clathrin is dissociated from the vesicles. Values of the frequency maxima and their frequency shifts Av are listed in Table I. The temperature dependence of the 2850-cm" methylene symmetric stretching modes for the coated and uncoated vesicles and synaptic membranes were examined using freshly prepared, highly purified samples. These frequencies also appear in TableI for spectra determined at 38 and 50 "C. The infrared spectroscopic methylene symmetric stretching modes at 2850 cm" are sensitive to the relative numbers of gauche and transconformations along the lipid acyl chains of the membrane bilayer. Both normal coordinate calculations (21) and infrared spectroscopic measurements (22) indicate that themethylene stretching modes increase 3-6 cm" as the gauche conformers are introduced into thehydrocarbon chain in passing from an ordered to a disordered state during, for example, the gel to liquid crystalline phase transition of a multilamellar assembly. For multilamellar systems composed of saturated diacyl chains, the all-transchain form is assumed "C for dimyrisfor the gel state at low temperatures ("40 toylphosphatidylcholine)(23). As the chains disorder intramolecularly due to either an increase in temperature or the

Infrared Spectra ofIsolated Coated Vesicles

8054

band maximum of at least 1.3 cm" to higher frequencies as the temperature i s raised from 21 to 50 "C (see Table I); a clear indication that, although the bilayer chains are intrinsically disordered as a consequence of their heterogeneous composition, further disorder in terms of the introduction of gauche conformers arises on increasing the temperatures of the systems. In addition to each membrane system displaying a temperature dependence for the spectroscopic marker reI flecting intrachain disorder, the clathrin-coated vesicle systems manifest increased numbers of gauche chain conformers, that is, a greater lipid chain disorder, compared to the uncoated vesicle and synaptic membrane assemblies at each temperature studied. Thus, as indicated in Table I, at 38 "C 0.08 the coated vesicles showa 1.00-cm" shift tohigher frequency compared to the uncoated vesicles. (Table I also notes that, B. UNCOATED for each type of system, the greatest increase in intramolecular chain disorder, as a consequence of temperature, occurs in the 0.04 21-38 "C interval.) We conclude that the presence of the clathrin coat clearly introduces a bilayer perturbation leading to an increase in the average number of gauche conformers along the lipid chains. Coated pits, the precursors of coated vesicles, are specialized 0.2 membrane domains which differ fromuncoated tracts in that ,rnM.S they contain a higher density of intramembranous particles 0.16 (24) and integral membrane receptors (25),but lack detectable C. SYNAPTIC levels of cholesterol (10, 11).The cholesterol deficiency was 0.12 MEMBRANE believed to be due to the molecular lipid-filtering capacity of the coated pit (26). This lack of cholesterol was also considered to be important in promoting the degree of fluidity required to facilitate the membrane structural changes involved in coated vesicle formation (10). Recent studies in this laboratory and others, however, have shown that the cholesterol contentof coated vesicles is greater than thatpreviously reported and may not be different from uncoated tracts of 2890 2870 2850 2830 2810 plasma membrane (9, 27). Furthermore, we have evidence to WAVENUMBERS (cm") suggest that the negative response of coated pits and coated vesicles to filipin is not due to abnormally low cholesterol FIG. 1. The infraredabsorbance spectra at21 "C of the lipid acyl chain methylene symmetric modes for coated vesicles content, as previously reported, but rather is a result of the (A), uncoated vesicles ( B ) , and synaptic membranes (C). stabilizing influence of the clathrin coat which inhibits the Buffer and water spectra have been subtracted. For the purpose of characteristic structural perturbationof the filipin-cholesterol presentation, the datahave been smoothed with a 25-point (2.5 cm" complexes.2 These various results, together with the present wide) cubic least squares algorithm. The peak frequency determined data, suggest that although the lipid environments of the by a gaussian fit (see text for discussion) of the coated vesicle stretching modes is shifted to higher frequenciesby 0.5 cm" compared coated pit and coated vesicle are not significantly different to that of the uncoated vesicle system, which, in turn, is shifted to from uncoated membrane domains, the bilayer disorder inhigher frequencies by 0.2 cm" compared to the stretching vibrations duced by the clathrin coat may prove to be one of the major for the synaptic membranes. Note the greater relative absorbance of factors responsible for membrane invagination and coated the methyl symmetric stretching modes at 2872 cm" for the coated vesicles compared to theuncoated vesicles and synaptic membranes. vesicle formation. Spectroscopic evidence for model membrane systems indiThe intensity of this feature in the coated vesicle system is derived cates that the acyl chain gauche to trans conformation ratio from the clathrin protein. increases as the lipid bilayer transforms from the nearly flat introduction of a membrane perturbant,the evolution of structures present in multilamellar systems to the highly gauche forms along the chains may bemonitored by a number curved forms reflecting single-shell unilamellar vesicles (13). of vibrational spectroscopic intensity andfrequency markers, Similarly, the presence of the clathrin lattice on the memincluding the the frequency shift observed for the 2850-~m-~brane surface of intact coated vesicles leads to anincrease in feature. For example, in a model membrane system composed gauche conformers compared to uncoated vesicles. Thus, in of dimyristoylphosphatidylcholinemultilayers, the methylene the coated pit regions of the membrane, the clathrin lattice symmetric stretching mode shifts from about 2850.5 cm" in may induce sufficient chain disorder in the underlying memthe ordered gel state at 5 "C to about 2853.4cm" in the brane region through the introduction of additional gauche disordered liquid crystalline state at 40 "C (22). (The gel to conformers to facilitate the formation of a coated vesicle. liquid crystalline phase transition occurs at 23 "C.) In the Although studies with model membranes suggest that some liquid crystalline form, a 10 "C increase in temperature from penetration intothe membrane bilayer by the ostensibly 30 to 40 "C shifts the methylene stretching mode by about 0.5 extrinsic clathrin coat is necessary to introduce additional cm" from 2852.9 to 2853.4 cm" (22). In the present study, the three systems observed, coated vesicles, uncoated vesicles, C. J. Steer, M. Bisher, R. Blumenthal, and A. C. Steven (1984) J. Cell Biol.. in mess. and synaptic membranes, display a shift of the 2850-cm" . " " L

1 "

I

I

I

!~

Infrared Spectra of Isolated Coated Vesicles

8055

TABLEI Summary of the vibrationalfrequencies in wave numbers (centimeter-') of the methylene symmetric stretching mode for coated vesicle. uncoated vesicle. and s v w t i c membrane assemblies Membrane assembly

Coated vesicles

21 "C"

AV*

2851.47 f 0.16

38 "C'

2853.51 k 0.05 0.44 f 0.22

Uncoated vesicles

2851.03 f 0.15d

2852.51 f 0.03'

AV

1.oo

50 "C'

2854.11 f 0.05 1.24 2852.87 f 0.04'

-0.07 f 0.27 0.49 Synaptic membrane 2851.10 f 0.23 2852.02 f 0.02 2852.44 f 0.02 Reported uncertainties are at the95% confidence level for three separate preparations. Av represents the frequency difference (centimeter") between succeeding membrane assemblies. e Reported uncertainties are standard deviations of the gaussian fit to one highly purified sample preparation. No change in frequency is observed for a pH change from 8.2 to 6.5. e The observed frequency increases for uncoated vesicles compared to synaptic membranes possibly reflect the decreased phospholipid/cholesterol mole ratio in the uncoated vesicle assembly. The determined phospholipid/ cholesterol mole ratios for coated vesicles and synaptic membranes are 1:0.3 and 1:0.6, respectively.

gauche conformers into the chain structure,3 a deep penetration seems unlikely intheintact membrane system in view of the with which the clathrin coat may be removed from the vesicle membrane. Nevertheless, the clathrin coat significantb perturbs the lipid bilayer matrix in both model and isolated coated vesicle systems. REFERENCES 1. Friend, D. S., and Farquhar, M. G. (1967) J. Cell Bwl. 3 5 , 357376 2. Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1979) Nature ( L o n d . ) 2 7 9 , 679-685 3. Pearse. B. M. F..andBretscher. M. S. (1981)Annu. Rev. Biochem. . . 60,85-101 4. Salisbury, J. L., Condeelis, J. S., and Satir, P. (1983) Int. Rev. E m . Pathol. 24, 1-62 5. Ungewickell,E., and Branton, D. (1981) Nature (Lond.) 2 8 9 , 420-422 6. Kanaseki, T., and Kadota, K. (1969) J. Cell B i d . 42,202-220 7. Heuser, J. (1980) J. Cell Biol. 84,560-583 8. Roth, T. F., and Woods, J. W. (1982) in Differentiationand Function of Hematopoietic Cell Surfaces (Marchesi, V. T., Majerus, p., and Gallo, R. c.,eds) pp. 163-181, Alan R. Liss, Inc., New York 9. Altstiel, L., and Branton, D. (1983) Cell 3 2 , 921-929 10. Montesano, R., Perrelet, A., Vassalli, R., and Orci, L.(1979) Proc. Natl. Acad. Sci. U. S. A . 7 6 , 6391-6395 11. McGookey, D. J., Fagerberg, K., and Anderson, R. G. W. (1983) J. Cell Bwl. 9 6 , 1273-1278 '

I. W. Levin, C. J. Steer, and R. Blumenthal, manuscript in preparation.

Av

0.43

12. Pearse, B. M. F. (1976) PFOC. Natl. Acad. Sci.U. S. A. 7 3 , 12551259 13. Levin, I. W. (1984) in Advances inInfrared and RamanSpectroscopy (Clark, R. J. H., and Hester, R. E., eds) Vol. 11, pp. 1-48, Heyden & Son, Ltd., London 14. Huang, C.H., Lapides, J. R., and Levin, I. w. (1982) J. Am. Chem. SOC.104,5926-5930 15. Nandi, P. K., Irace, G., van Jaarsveld, P. P.,Lippoldt, R. E., and Edelhoch, H. (1982) Proc. Natl. Acad. Sci. U. S. A. 7 9 , 58815885 16. Lowry, 0.H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 1 9 3 , 265-275 17. Steer, C. J., Wall, D. A., and Ashwell,G. (1983) Hepatology (Baltimore) 3,667-672 18. Jones, D. H., and Matus, A. I. (1974) Biochim.Biophys.Acta 356,276-287 19. Fiske, C. H.,and SubbaRow, Y. (1925) J. Biol. Chem. 66, 375400 20. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochern. Physiol. 37,911-917 21. Snyder, R. G., Strauss, H. L., and Elliger, C. A. (1982) J. Phys. Chem. 86,5145-5150 22. Mendelsohn, R., Dluhy, R., Taraschi, T., Cameron, D.G., and Mantsch, H. H. (1981) Biochemistry 20,6699-6706 23. Yellin, N., and Levin, I. W. (1977) Biochim. Biophys. Acta 489, 177-190 24. Orci, L., Carpentier,J-L.,Perrelet, A., Anderson, R.G. W., Goldstein, J. L., and Brown, M. S. (1978) Exp. Cell Res. 113, 1-13 25. Steer, C. J., and Klausner, R. D. (1983) Hepatology (Baltimore) 3,437-454 26. Bretscher, M. S. (1976) Nature (Lond.) 260,21-22 27. Simion, T., Winek, D., Brandon, B., Fleischer, S., and Fleischer, B. (1982) J. Cell Bwl. 96,249a