Proc. NatL Acad. Sci. USAVol. 78, No. 1, pp. 293-297, January 1981Cell Biology

Heterogeneous distribution of filipin-cholesterol complexes acrossthe cisternae of the Golgi apparatus*

(intracellular membranes/cholesterol/freeze-fracture)

L. ORCIt, R. MONTESANOt, P. MEDAt, F. MALAISSE-LAGAEt, D. BROWNt, A. PERRELETt, AND P. VASSALLIttInstitute of Histology and Embryology, and *Department of Pathology, University of Geneva Medical School, 1211 Geneva 4, Switzerland

Communicated by George E. Palade, August 21, 1980

ABSTRACT The Golgi apparatus is a key element in the or-dered movement of secretory polypeptides from the rough en-doplasmic reticulum to the plasma membrane during secretion.It has been shown that cisternae that receive membranes from thereticulum are morphologically similar to the latter and that cis-ternae liberating secretory granules resemble the plasma mem-brane. By using an ultrastructural probe for membrane choles-terol, filipin, on freeze-fractured and thin-sectioned exocrine andendocrine pancreatic cells, we have shown that an enrichment infilipin-cholesterol complexes takes place across the stacked cis-ternae of the Golgi apparatus; the reticulum-related (forming) cis-ternae are poor in such complexes, but the secretory granule-re-lated (maturing) cisternae contain numerous complexes. Secretorygranule membrane is also richly labeled with fipincholesterolcomplexes. The heterogeneous cholesterol distribution in themembranes of the Golgi apparatus, as shown by filipin, empha-sizes the polarity of this organelle, in agreement with its role inorganizing the traffic ofthe secretory polypeptides from the roughendoplasmic reticulum to the plasma membrane.

When the membranes from different cell organelles involvedin the synthesis, packaging, and export ofsecretory proteins areexamined, there is good biochemical and morphological evi-dence that the secretory granule membranes resemble theplasma membrane (with which they will ultimately fuse) inmany respects and that those of the rough endoplasmic retic-ulum (RER) do not (1-3). Within the Golgi apparatus, the cis-ternal membranes show a gradient from "RER-like" at the form-ing convex face to "plasma-membrane-like" at the maturingconcave face (1, 4-8). Although changes in protein compositioncan, at least in part, be detected morphologically by changesin the density of intramembrane particles in freeze-fracturedmembranes (9-13), there has so far been no morphological evi-dence for the biochemically measured changes in lipid content(14-18).

In this paper, we provide direct morphological evidence forthe heterogeneous distribution ofa major membrane lipid, cho-lesterol, between the maturing and the forming faces of theGolgi apparatus. To do this, we used a polyene antibiotic, fi-lipin, that specifically binds to cholesterol and related 3,B3hy-droxysterols (19-22). This interaction results in the formationof multimolecular filipin-cholesterol (f-c) complexes that areeasily recognized in freeze-fractured tissues as either protu-berances or pits (23-26) in the fracture faces. The complexesare 20-25 nm in diameter and have a characteristic shape thatmakes it easy to distinguish them from the much smaller (""9nm) intramembrane particles. Based on these properties, thetreatment of tissues with filipin during or after glutaraldehydefixation has been used as a cytochemical technique for deter-mining the location ofcholesterol in freeze-fractured cell mem-

branes (27, 28). Glutaraldehyde fixation is required to preventredistribution of the intramembrane particles (28) and the celllysis (19, 28) that would result from the action offilipin on livingcells. Also, because filipin can diffuse across the plasma mem-brane and reach the intracellular membranes, it can be used toassess their degree of labeling and, hence, their probable rel-itive cholesterol content.

MATERIALS AND METHODSIslets ofLangerhans were isolated from rat or spiny mouse pan-creases as described (29). Batches of -"200 islets were fixed for15 min in 2% glutaraldehyde/0. 1 M cacodylate buffer (pH 7.4)and then incubated overnight at room temperature in the samebuffer containing filipin at 0.2 mg/ml (kindly provided by J. E.Grady, Upjohn) dissolved in a drop ofdimethyl sulfoxide (finalconcentration of dimethyl sulfoxide, 1%). Rat and spiny mousepancreases were minced into very small pieces (<1 mm), fixedin the glutaraldehyde solution described above for 60 min, andincubated in the filipin/cacodylate solution overnight.

After filipin treatment, pellets of isolated islets and piecesofpancreas were washed in 0.1 M cacodylate buffer, soaked inthe same buffer containing 30% glycerol (vol/vol) for 2 hr, andfreeze-fractured at -1100C in a Balzers BAF 301 apparatus(Balzers High Vacuum, Balzers, Liechtenstein) according to ref.30. The tissues were then thawed, and replicas were cleanedwith sodium hypochlorite and then with chloroform/methanol(2:1), rinsed in distilled water, and recovered on Parlodion-coated copper grids (150 mesh).

For thin-section electron microscopy, the glutaraldehyde-fixed, filipin-treated material was postfixed in 1% osmium te-troxide in 0.1 M Veronal acetate buffer, stained en bloc with0.5% uranyl acetate in Veronal acetate buffer, dehydrated byusing graded ethanols and embedded in Epon. Thin sectionswere cut with a diamond knife and further stained with uranylacetate and lead citrate. Freeze-fracture replicas and thin sec-tions were examined in a Philips EM 300 (Philips, Eindhoven,The Netherlands) electron microscope.

Quantitative evaluation of the density of f-c complexes wascarried out as follows. Photographs showing the fracture facesof one or more of the three intracellular compartments inves-tigated (forming Golgi cisternae, maturing Golgi cisternae andtheir associated condensing vacuoles, and mature secretorygranules) were selected (36 pictures of acinar cells; 33 picturesof endocrine cells). The surface of each exposed membranecompartment and the numbers of protuberances or pits rep-resenting f-c complexes present on the exposed membrane

Abbreviation: RER, rough endoplasmic reticulum.* Part of this work was presented at the symposium Biological Cycles,in honor of Sir Hans Krebs' 80th birthday, March 17-19, 1980, Dallas,TX.

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

293

Dow

nloa

ded

by g

uest

on

July

5, 2

020

Proc. NatL Acad. Sci. USA 78 (1981)

FIG. 1. Freeze-fracture replica offilipin-treated acinar cells ofthe spiny mouse exocrine pancreas. (A) Several membrane faces in the Golgi areashow different degrees of filipin labeling. Complexes of f-c appear as 20- to 25-nm protuberances or slightly smaller pits on the membrane fractureface and are easily distinguished by their shape and size from the much smaller (-9 nm) intramembrane particles. The most heavily labeled mem-brane is the concave circular structure in the upper right-hand corner that represents part ofa condensing vacuole (CV1); in the curved and stackedGolgi cisternae, the maturing cisternae (MG) show a moderate, heterogeneous labeling, and the forming cisternae (FG) are not labeled. Comparisonofthe two condensing vacuoles visible in this field (CV1 and CV2) shows that the degree offilipin labeling in these structures often varies. (x43,000.)(B-D) These micrographs suggest a sequence ofcondensing vacuoles (CVs) having an increasing content of f-c complexes. The differences in freeze-fracture appearance between condensing vacuoles and zymogen granules (E) are mainly the irregular shape of the former and their higher intra-membrane-particle content. Also, the CVs are characterized by the presence ofmembrane segments that appear to bulge (or invaginate, dependingon which leaflet of the membrane is exposed) and, in filipin-treated material, contain no filipin-sterol complexes (C, dotted line; D, arrows). NoteinE that the f-c complexes take the form ofprotuberances on the zymogen granule E face (convex) [ZG(E)], and ofpits on the granule P face (concave)[ZG(P)I. (B, rat, x32,000; C, spiny mouse, x49,000; D, rat, x37,000; E, spiny mouse, x41,000.) In all figures, the horizontal bar equals 0.2 gm.

were recorded on a graphic tablet (Tektronix type 4953) con- RESULTSnected to a microprocessor system (IMSAI 8080) programmedto compute the number of complexes per unit area of mem- Freeze-Fracture. The freeze-fracture appearance of the cy-brane. Values were compared by using Student's t test. toplasm of acinar cells is shown in Fig. 1 and that of endocrine

294 Cell Biology: Ord et aL

Dow

nloa

ded

by g

uest

on

July

5, 2

020

Proc. Natl. Acad. Sci. USA 78 (1981) 295

'FG -a .'%

I

'4 V~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i4)Il3~~~~~~~3% -4ONG. 71:*ii. tn ;v

~~~~~~~~~~~~\~~~~~~~~~~i, ~' -, XMG_'

FIG. 2. (A) Freeze-fracture replica of a filipin-treated rat pancreatic endocrine cell showing the typical curved and stacked arrangement of thecisternae of a Golgi apparatus: the outer forming Golgi cisternae (FG) have no filipin label and the inner maturing cisternae (MG) are labeled withfilipin-sterol complexes. Convex circular faces representing the fractured membrane of secretory granules (SG) are also labeled with numerouscomplexes. (x61,000.) (B) Freeze-fracture replica of a filipin-treated rat pancreatic endocrine cell showing the cytoplasm including the Golgi ap-paratus. The pattern of filipin labeling across the Golgi apparatus is similar to that shown inA; f-c complexes are absent from the forming cisternae(FG) but present in both the maturing cisterna (MG) and secretory granules (SG). In this case, filipin labeling is heterogeneous from area to areaof the maturing cisterna. (x41,000.) (C) Thin-section electron micrograph of filipin-treated rat pancreatic B cell showing the Golgi region. Thepresence of f-c complexes induces a characteristic corrugated or scalloped appearance of the thin-sectioned membrane. In tangential sections, com-plexes appear as 20- to 25-nm circles (arrows). Scalloped membrane segments are evident in a maturing cisterna containing a condensing granulecore (arrowhead); the forming cisternae and transitional elements (TE) of the rough endoplasmic reticulum lack these complexes. f-c complexesare also present on a mature secretory granule. The empty vesicle with a scalloped limiting membrane (V) is probably a cross section of the dilatedend of a maturing cisterna. (x 48,000.) (D) Thin-section electron micrograph of appearance of the dilated end of a maturing Golgi cisterna containinga condensing secretory granule core. Corrugation of the membrane shows the presence of f-c complexes (some in tangential section are shown byarrows) on most of the periphery of the dilated cisterna. (x49,000.) (E) Thin-section micrograph of expanded end of maturing Golgi cisterna con-taining a condensing granule core (arrowhead). The limiting membrane has a scalloped appearance showing the presence of f-c complexes on partof its periphery. At the region indicated by arrows, the membrane has a distinct coat and lacks f-c complexes. (x 34,000.) (F) Thin-section electronmicrograph of maturing P-secretory granule in the Golgi area. The granule limiting membrane is corrugated by filipin-sterol complexes exceptat the coated region (dotted line), and another neighboring vesicular profile lacks filipin-sterol complexes in a coated region (dotted line). (x 52,000.)In all figures, the horizontal bar equals 0.2 tum.

Cell Biology: Orci et aL

Dow

nloa

ded

by g

uest

on

July

5, 2

020

Proc. Natd Acad. Sci. USA 78 (1981)

pancreatic cells is shown in Fig. 2 A and B. In both acinar andendocrine tissue exposed to filipin, it was apparent that not allcells had been labeled. This was probably due to the unevenpenetration of filipin into the tissue, an assumption that hasbeen confirmed in the case of acinar tissue by exposing disso-ciated cells to filipin, which.led to the labeling of all cells pres-ent. The following description, therefore, applies to well-la-beled cells only. §

The complexes appear as 20- to 25-nm protuberances orslightly smaller pits that may both be present on the same mem-brane face. A heterogeneous distribution of the f-c complexeswas evident not only from one membrane compartment to an-other but also from area to area within the same compartment.In both acinar and endocrine pancreatic cells, the intracellularcompartment showing the highest density off-c complexes wasthe limiting membrane of the secretory granules, which wereidentified by their characteristic size and regular sphericalshape (see Figs. 1E and 2 A and B). Occasionally, however, inacinar cells showing well-labeled Golgi apparatuses, there werea few secretory granules that did not have f-c complexes. In thezymogen granules, the f-c complexes were usually distributedhomogeneously on the limiting membrane, whereas in the se-cretory granules of endocrine cells, the granule membranesoften had patches free ofcomplexes. When a favorable fractureplane allowed a clear distinction between the two faces of theGolgi complex, the convex forming (or cis) cisternae, whichwere close to the RER membranes, appeared virtually devoidof f-c complexes. On the contrary, the concave maturing (ortrans) cisternae adjacent to the secretory granules showed a.variable concentration of f-c complexes, although these wereusually not as densely packed as those on the secretory granulemembranes themselves (see Figs. 1A and 2 A and B). The dis-tribution of complexes in a single maturing cisterna was fre-quently heterogeneous, consisting of clusters of complexes onan otherwise poorly labeled membrane (see Figs. LA and 2B).However, this was not the only pattern of labeling observed;as shown in Fig. 2A, maturing cisternae sometimes showed arandom pattern of labeling. The clustered type of labeling ver-sus the random type did not seem to be tissue or species spe-cific; both forms were observed, not only in endocrine.andacinar pancreatic cells ofdifferent animals, but also in secretorycells of the anterior pituitary and the mammary glands (data notshown). In acinar cells, the condensing vacuoles were identifiedon the bases of their topographic relationship with maturingcisternae and oftheir irregular shape as compared with the zym-ogen granules (the membrane of the condensing vacuoles oftenshowed circular bumps or depressions that probably correspondto the coated segments visible in thin sections). The limitingmembrane of the condensing vacuoles appeared to contain avarying number of f-c complexes: some condensing vacuoleshad only a few complexes (see Fig. 1 A and B), but others hadmembranes as densely and homogeneously labeled as the ma-ture zymogen granules (see Fig. 1 C and D). In these vacuoles,the circular bulges or. depressions of the limiting membranewere always free of complexes (see Fig. 1 B-D).

The results ofthe quantitative evaluation ofthe density ofthef-c complexes in each membrane compartment studied areshown in Table 1.

Thin Sections. The interaction between filipin and choles-terol was detectable on thin sections offilipin-treated material,because the membranes bearing f-c complexes showed a char-acteristic "corrugated" or "scalloped" appearance. Study ofthin

Table 1. Quantitative evaluation of filipin labelingin exocrine and endocrine pancreatic cells

P face E faceAcinar cellsZymogen

granules 195.8 ± 14.6 (n = 29) 183.3 ± 14.4 (n = 16)Maturing Golgiand con-densingvacuoles 168.2 + 19.3* (n = 25) 156.7 ± 18.9* (n = 43)

Forming Golgi 3.8 ± 2.0 (n = 31)tEndocrine cells

Secretorygranules 320.8 ± 29.7 (n = 34) 377.5 ± 43.4 (n = 39)

Maturing Golgi 176.6 ± 21.9 (n = 19)$ 136.3 ± 23.9 (n = 13)VForming Golgi. 35.8 ± 11.3 (n = 44)t

Complexes (protuberances plus pits) are expressed as number persquare micrometer of membrane area ± SEM. n = number of mem-brane faces studied. Due to the highly heterogeneous distributionof f-c complexes in many membrane areas, the values shown may de-rive from the averaging ofa high number and a low number in adjacentareas of the same membrane.* Comparison with zymogen granules, not significant.t Comparison with maturing Golgi (and condensing vacuoles, if pres-ent), P < 0.001. Values given represent pool ofP- and E-face values.

t Comparison with secretory granules, P < 0.005.

sections of pancreatic endocrine and exocrine cells confirmedand extended the observations made on freeze-fracture replicasby offering the advantage of a. more-precise identification ofsome of the labeled organelles. In thin sections including theentire stack of Golgi cisternae (see Fig. 2C), corrugated mem-brane segments could be observed in the inner maturing cis-ternae, some of which showed an expanded end where con-densation of secretory material occurred (see Fig. 2D). Thin-sectioned material also showed other membrane segments hav-ing heterogeneous filipin labeling. These were the coated re-gions present either on expanded Golgi cisternae containingcondensing secretory material (see-Fig. 2E) and the coated seg-ment present on some maturing secretory granules in the en-docrine pancreatic B cell (see Fig. 2F).

DISCUSSIONThe central role of the Golgi apparatus in the processing andpackaging of proteins synthesized in the RER is well known,particularly in polypeptide secreting cells (for review, see ref.31). It has also been shown that, as the secretory proteins movefrom the RER to the Golgi apparatus and the secretory granules,the limiting membranes also change. Thus, the membranes ofthe RER are biochemically and morphologically different fromthose of the mature secretory granules, which themselves inmany respects resemble the plasma membrane (1-3). Whenthin sections of the membranes of the Golgi cisternae are ex-amined, a progressive alteration in membrane thickness isfound from "RER-like" at the forming face to "plasma-mem-brane-like" at the maturing face (1, 4-8, 32, 33). Isolated Golgimembranes show lipid (including cholesterol) (14-18) and en-.zyme (34) contents that are, in many cases, intermediate be-tween those of the RER and of the plasma membrane. In somefreeze-fracture studies of the Golgi apparatus, an intramem-brane particle-density gradient has been observed across thestacked Golgi cisternae (10-13). These data, together with re-sults obtained by using the osmium impregnation technique(35, 36) and enzyme cytochemistry (37-42), suggest that themembranes of the Golgi cisternae are structurally and func-tionally heterogeneous. Heterogeneity ofthe Golgi membranes

§ In the least-labeled acinar cells, the intracellular compartment pref-erentially marked was the maturing cisternae and condensing vacu-oles of the Golgi apparatus. Some cells showed labeling only in thisarea.

296 Cell Biology: Ord et aL

Dow

nloa

ded

by g

uest

on

July

5, 2

020

Proc. Natd Acad. Sci. USA 78 (1981) 297

has also been shown by biochemical studies on Golgi subfrac-tions (43-46). A finding of particular interest is that the mem-branes of different Golgi subfractions do not bind. digitonin (amolecule known-to form insoluble complexes with cholesterol)to the same extent (45), which led to the suggestion that cho-lesterol may be heterogeneously distributed in Golgi mem-branes (1, 46).Our results provide in situ evidence that an increase of mem-

brane cholesterol content may take place at the maturing faceofthe Golgi apparatus and, thus, contribute to the growing bodyof evidence that the Golgi apparatus is a complex, heteroge-neous system, not only from cisterna to cisterna (cis-trans asym-metry) but also from area to. area within the same cisterna. How-ever, it should be borne in mind that all membrane compartmentsmay not be equally accessible to filipin and also that redistri-bution of both. cholesterol and the f-c complexes probably oc-curs even in fixed membranes (28). Nevertheless, when cellshaving a comparable degree of overall labeling were studied,the heterogeneous distribution of the f-c complexes was con-sistent and reproducible. We feel justified, therefore, in ourinterpretation that the observed heterogeneity representsmeaningful.differences in membrane structure.

Another problem of interpretation concerns the significanceof the form (protuberances or.pits) taken by the f-c complexes.The distribution of each type of deformation of the membranefracture face,. due to the orientation of the f-c complex with re-spect to the plane of the membrane (i.e., toward the cytosol orthe cisternal space in the case of an intracellular vacuole,. givinga pit or a protuberance, respectively, on the membrane P face)appears nonrandom in the respective leaflets of the zymogengranule and the condensing vacuole membrane of the acinarcell, which may reflect an asymmetry of cholesterol contentbetween the leaflets (47).A final point concerns the coated membrane segments that

are present on trans-Golgi cisternae, on maturing secretorygranules of endocrine pancreatic cells (36, 48), and on con-densing vacuoles of acinar cells (49). and are always free of f-ccomplexes. An absence of f-c complexes was recently reportedin discrete regions of the fibroblast plasma membrane found tocorrespond precisely to the coated pits involved in endocytoticplasma membrane internalization (26). Whether.the absence off-c complexes on these segments is also associated, as in the caseof the coated pits of the plasma membrane, with their eventualpinching off from their initial compartment, remains to beelucidated.

We thank Dr. M. Amherdt for-help with the quantitative analysisandM. Bernard, P. Fruleux, P. Sors, M. Sidler-Ansermet, G. Perrelet, andI. Bernard for skillful.technical assistance. This study was supported bythe Swiss National Science Foundation, Grants 3.120.77 and 3.668.80,and by a grant-in-aid from Hoechst-Pharmaceuticals, Frankfurt,Hoechst, Federal Republic of Germany.

1. Meldolesi, J., Borgese, N., De Camilli, P. & Ceccarelli, B. (1978)in Membrane Fusion, Cell Surface Reviews, eds. Poste, G. & Nic-olson, G. L. (North Holland, Amsterdam), Vol. 5, pp. 509-627.

2. Morre, D. J., Kartenbeck, J. & Franke, W. W (1979) Biochim.Biophys. Acta 559, 71-152.

3. Geuze, J. J., Kramer, M. F. & De Man, J. C. H. (1977) in Mam-malian Cell Membranes,,eds. Jamieson, C. A. & Robinson, D. M.(Butterworth, London), Vol. 2, pp. 55-107.

4. Morr6, D. J. (1977) in The Synthesis, Assembly and Turnover ofCell Surface Components, Cell Surface Reviews, eds. Poste, G.& Nicolson, G. L. (North Holland, Amsterdam), Vol. 4, pp. 1-83.

5. Morr6, D. J. & Ovtracht, L. (1977) Int. Rev. Cytol Suppl. 5,61-188.

6. Favard, P. (1977) in Mammalian Cell Membranes, eds. Jamieson,C. A. & Robinson, D. M. (Butterworth, London), Vol. 2, pp.108-140.

7. Whaley, W. G. & Dauwalder, M. (1979) Int. Rev. Cytol 58,199-245.

8. Farquhar, M. G. (1978) in Transport of Macromolecules in Cel-lular Systems, ed. Silverstein, S. C. (Dahlem Konferenzen, Ber-lin), pp. 341-362.

9. Deamer, D. W. (1977) in Mammalian Cell Membranes, eds. Ja-mieson, C. A. & Robinson, D. M. (Butterworth, London), Vol.4, pp. 1-31.

10. Staehelin, L. A. & Kiermayer, 0. (1970)1. Cell Sci. 7, 787-792.11. Vian, B. (1974) C. R. Acad. Sci. Paris Ser. D 278, 1483-1486.12. Orci, L. & Perrelet, A. (1977) in The Diabetic Pancreas, eds. Volk,

B. W. & Wellmann, K. F. (Plenum, New York), pp. 171-209.13. Chailley, B. (1979) BioL. CeL 35, 55-70.14. Keenan, T. W. & Morre, D. J. (1970) Biochemistry 9, 19-25.15. Yunghans, W. N., Keenan, T. W. & Morr6, D. J. (1970) Exp. Mol

Pathol, 12, 36-45.16. Meldolesi, J., Jamieson, J. D. & Palade, G. E. (1971)J. Cell Biol

49, 130-149.17. Meldolesi, J. (1971) in Advances in Cytopharmacology, eds. Cle-

menti, F. & Ceccarelli, B. (Raven, New York), Vol. 1, pp.145-157.

18. Morre, D. J., Keenan, T. W. & Mollenhauer, H. H. (1971) inAdvances in Cytopharmacology, eds. Clementi, F. & Ceccarelli,B. (Raven, New York), Vol. 1, pp. 159-182.

19. Kinsky, S. C. (1970) Annu. Rev. Pharmacol 10, 119-142.20. Norman, A. W., Demel, R. A., De Kruiff, B., & Van Deenen,

L. L. M. (1972)J. Biol Chem. 247, 1918-1929.21. De Kruijff, B., Gerritsen, W. J., Oerlemans, A., Demel, R. A.

& Van Deenen, L. L. M. (1974) Biochim. Biophys. Acta 339,30-43.

22. Norman, A. W., Spielvogel, A. M. & Wong, R. G. (1976) Adv.Lipid Res. 14, 127-170.

23. Verkleij, A. J., De Kruijff, B., Gerritsen, W. J., Demel, R. A.,Van Deenen, L. L. M. & Ververgaert, P. H. J. (1973) Biochim.Biophys. Acta 291, 577-581.

24. Tillack, T. W & Kinsky, S. C. (1973) Biochim. Biophys. Acta 323,43-54.

25. Andrews, L. D. & .Cohen, A. I. (1979) J. Cell Biol 81, 215-228.26. Montesano, R., Perrelet, A., Vassalli, P. & Orci, L. (1979) Proc.

Natl Acad. Sci. USA 76, 6391-6395.27. Elias, P. M., Friend, D. S. &Goerke, J. (1979)1. Histochem. Cy-

tochem. 27, 1247-1260.28. Robinson, J. M. & Karnovsky, M. J. (1980) J. Histochem. Cyto-

chem. 28, 161-168.29. Lacy, P. E. & Kostianovsky, M. (1967) Diabetes 16, 35-39.30. Moor, H. & Mdhlethaler, K. (1963)J. Cell Biol 17, 609-628.31. Palade, G. E. (1975) Science 189, 347-358.32. Grove, S. N., Bracker, C. E. & Morr6, D. J. (1968) Science 101,

171-173.33. Morr6, D. J., Keenan, T. W. & Huang, C. M. (1974) in Cyto-

pharmacology of Secretion, eds. Ceccarelli, B., Meldolesi, J. &Clementi, F. (Raven, New York), Vol. 2, pp. 107-125.

34. Meldolesi, J., Jamieson, J. D. & Palade, G. E. (1971)J. Cell Biol49, 150-158.

35. Friend, D. S. & Murray, M. J. (1965) Am. J. Anat. 117, 135-149.36. Orci, L. (1974) Diabetologia 10, 163-18737. Friend,.D. S. (1969)1. Cell BioL 41, 269-279.38. Hand, A. R. (1971) Am. J. Anat. 130, 141-157.39. Novikoff, P. M., Novikoff, A. B., Quintana, N. & Hauw, J.-J.

(1971)1. Cell Biol 50, 859-886.40. Farquhar, M. G., Bergeron, J. J. M. & Palade, G. E. (1974)1. Cell

Biol 60, 8-25.41. Cheng, H. & Farquhar, M. G. (1976)J. Cell BioL 70, 671-684.42. Novikoff, A. B. & Novikoff, P. M. (1977) Histochem. 1. 9, 525-551.43. Bergeron, J. J. M., Ehrenreich, J. H., Siekevitz, P. & Palade, G.

(1973)J. Cell Biol 59, 73-88.44. Ovtracht, L., Morre, D. J., Cheetham, R. D. & Mollenhauer, H.

H. (1973) J. Microsc. (Oxford) 18, 87-102.45. Amar-Costesec, A., Wibo, M., Thines.Sempoux, D., Beaufay, H.

& Berthet, J. (1974) J. Cell Biol 62, 717-745.46. Borgese, N. & Meldolesi, J. (1980) J. Cell BioL 85, 501-515.47.- Orci, L., Miller, R. G., Montesano, R., Perrelet, A., Amherdt,

M. & Vassalli, P. (1980) Science 210, 1019-1021.48. Orci, L. (1977) in Endocrinology, Proceedings of the 5th Inter-

national Congress on Endocrinology, ed. James, V. H. T. (Ex-cerpta Medica, Amsterdam), Vol. 2,.pp. 7-40.

49. Jamieson, J. D. .-& Palade, .G. E. (1967)J. Cell BioL 34, 577-596.

Cell Biology: Ord et aL

Dow

nloa

ded

by g

uest

on

July

5, 2

020