Proc. Nati. Acad. Sci. USAVol. 82, pp. 5385-5389, August 1985Cell Biology

Clathrin-immunoreactive sites in the Golgi apparatus areconcentrated at the trans pole in polypeptidehormone-secreting cells

(coated cisternae/coated granules/immunocytochemistry/intracellular transport/secretion)

L. ORCI*, M. RAVAZZOLA*, M. AMHERDT*, D. LOUVARDt, AND A. PERRELET**Institute of Histology and Embryology, University of Geneva Medical School, 1211 Geneva 4, Switzerland; and tPasteur Institute, Department of MolecularBiology, 75275 Paris, France

Communicated by J. E. Rall, April 24, 1985

ABSTRACT By analogy with receptor-mediated endo-cytosis, clathrin-coated membrane segments at locations otherthan at the plasma membrane have been implicated in intra-cellular membrane transport/sorting. The crucial role of theGolgi apparatus in these processes prompted us to study thedistribution at this level of clathrin-immunoreactive sites invarious polypeptide hormone-secreting cells. With a polyclonalantibody recognizing the polymerized form of clathrin (struc-tural coats or cages), we found that the Golgi area had thehighest degree of specific labeling of various cytoplasmicregions, including the subplasmalemmal web. Moreover, theGolgi labeling was not homogeneously distributed, most of theimmunoreactive sites being associated with membrane seg-ments at the trans pole. The labeled membranes comprisedregions of trans cisternae with and without condensing secre-tory material, newly formed (maturing) secretory granulesfreshly released from trans cisternae, and typical coatedvesicles. The cis Golgi cisternae, as well as the cytoplasmicregion containing transfer (shuttle) vesicles and transitionalcisternae of the rough endoplasmic reticulum, were onlyweakly labeled. Clathrin concentration, degree of assembly oftriskelions into coats, duration of clathrin association withmembranes, or presence of a coat protein different fromclathrin could account for the low immunoreactivity observedat the cis pole as compared to the trans pole.

The migration of secretory polypeptides from their site ofsynthesis, the rough endoplasmic reticulum (RER), to theirsite of exocytotic release, the plasma membrane, involves aseries of processes requiring fusions and fissions of mem-brane as well as specific targeting (1). By analogy withreceptor-mediated endocytosis (2), which is characterized bythe association of clathrin coats (3) with the invaginatingmembrane, a role of membrane carrier/sorter has also beenproposed for clathrin at intracellular locations other than theplasma membrane (4). During the intracellular migration ofsecretory polypeptides, two steps at which membrane inter-actions are crucial are the transfer of polypeptides from theRER to the cis pole of the Golgi apparatus and their releaseat the trans pole to yield secretory granules/vesicles (1). Todetermine the distribution of clathrin at these two sites, wehave analyzed, using an affinity-purified polyclonal antibodyvisualized by the staphylococcal protein A-gold technique,the labeling of the Golgi area of endocrine cells of the islet ofLangerhans and of the anterior pituitary.

MATERIAL AND METHODSIslets of Langerhans were isolated from male SIVZ rats (5),incubated for 3 hr in threshold stimulatory concentration ofglucose (1 mg/ml), fixed in sodium phosphate-buffered 2%(vol/vol) glutaraldehyde, pH 7.4, dehydrated, and embeddedin Epon (for immunofluorescence) or at low temperature inLowicryl K4M resin (6) (for protein A-gold). Rat anteriorpituitaries fixed in situ by aortic perfusion of fixative wereprocessed as described. Clathrin-immunoreactive sites werelocalized by light microscopy with the indirect immunofluo-rescence method (7) after removal of the Epon resin (8) andby electron microscopy with the protein A-gold technique (9)on Lowicryl sections, using =15-nm (9) or =6-nm colloidalgold particles (10). For the protein A-gold technique, thinsections were incubated overnight at 4°C on a drop ofaffinity-purified antibody to clathrin (polyclonal, lot 3-124)(11) diluted 1:100 (5 pXg/ml) with phosphate-buffered saline,pH 7.4. After incubation, thin sections were rinsed withdistilled water and treated with the protein A-gold solutionfor 1 hr at room temperature. Sections were washed againwith distilled water and stained with uranyl acetate and leadcitrate before observation in a Philips EM 300 electronmicroscope. lThe specificity of immunostaining was tested byusing nonimmune rabbit serum; no labeling was obtained inthis condition.

In pancreatic B cells, the quantitative evaluation of thedistribution of immunoreactive sites was carried out onphotographic prints (x38,000), using an electronic pen con-nected to a graphic tablet and a microprocessor programmedto calculate the number of points (gold particles) per ,tm2 ofthe compartment drawn with the pen (see Fig. 2A). Fourcompartments were evaluated separately: (i) subplasmalem-mal rim (0.25 ,um wide); (ii) cis Golgi region, defined as thecis-most cisterna, the two adjoining ones, and a neighboringcytoplasmic area comprising the transfer (shuttle) vesiclesissued from the RER and the transitional cisternae of thelatter; (iii) trans Golgi region, defined as the trans-mostcisterna, the two adjoining ones, and a surrounding cytoplas-mic area including newly formed secretory granules andvesicles neighboring the cisternae; and (iv) nucleus (as cellu-lar background).The Golgi region of endocrine cells of the anterior pitu-

itary, as well as of islet cells other than B cells, was assessedqualitatively on serial thin sections.

RESULTS

At the light microscope level, indirect immunofluorescencerevealed the presence of anti-clathrin binding spots in both

Abbreviation: RER, rough endoplasmic reticulum.

5385

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

Proc. Natl. Acad. Sci. USA 82 (1985)

FIG. 1. Immunofluorescence staining, with the affinity-purifiedpolyclonal anti-clathrin antibody, of semi-thin sections of the islet ofLangerhans (A) and anterior pituitary (B). In both tissues, theimmunofluorescence reaction is visible as bright spots in the vicinityof the unreactive cell nucleus (asterisk). The crescent shape (arrows)as well as the location of the fluorescent spots suggest that theycorrespond to Golgi regions of secretory cells. (Epon embeddingwithout osmium tetroxide postfixation; both x840.)

islet and anterior pituitary cells. The fluorescence wasparticularly abundant in the juxtanuclear area and oftenshowed a crescent shape (Fig. 1). Both the localization andshape of the immunofluorescence reaction suggested thatGolgi regions of islet cells were stained. At the electronmicroscope level, the immunoreactivity revealed by colloidalgold particles appeared to be of low intensity throughout thecytoplasm except in the Golgi region and, to a lesser extent,in the subplasmalemmal area (Fig. 2A). The quantitativevalues of clathrin immunolabeling per Am2 of the nuclear, cisand trans Golgi, and subplasmalemmal compartments ofinsulin-containing cells are shown in Table 1. In the Golgiregion, a marked asymmetry of the immunolabeling wasobserved between the cis and trans poles of the complex. Atthe trans Golgi pole, there were numerous gold particles thatappeared to be preferentially associated with extensivesegments (<6 ,m) of cisternae with or without condensingsecretory material, and with vesicles and newly formedsecretory granules (Fig. 2 A-D). The coated membranes onthese compartments are not optimally demonstrated onLowicryl-embedded material needed for clathrin im-munolabeling, but these compartments were previously rec-ognized as "coated" in conventional nonimmunolabeledEpon thin sections (12, 13). At the cis pole, the low degree ofimmunoreactivity was distributed over cisternae and thecytoplasmic area between these and the RER. In islet celltypes other than B cells-i.e., glucagon, somatostatin, andpancreatic polypeptide cells-as well as in endocrine cells ofthe anterior pituitary, the marked asymmetry of clathrin

Table 1. Clathrin immunolabeling of various compartments ofthe B cell

Compartment No. of gold particles per ,umlSubplasmalemmal rim 3.34 ± 0.32cis Golgi 3.46 ± 0.41trans Golgi 26.52 ± 1.10Nucleus 0.40 ± 0.11

For the delimitation of the various compartments, see Materialand Methods and Fig. 2A. Values are means ± SEM. P values weredetermined with Student's t test: subplasmalemmal rim versus cisGolgi, not significant; cis Golgi versus trans Golgi, P < 0.001; transGolgi versus nucleus, P < 0.001; subplasmalemmal rim and cis Golgiversus nucleus, P < 0.001. Eleven islets were evaluated. In eachislet, 6-10 B cells were examined; a total of 2183 gold particles in 777,4m2 were counted. Gold particles 6 nm (10).

immunolabeling of the Golgi region was also evident (Figs. 2E and F and 3 A and B). Fig. 3 A and B shows an example ofthe extensive coating of maturing secretory granules in theGolgi area of prolactin cells, as evidenced on serial sections.

DISCUSSIONThe current hypothesis concerning the role of intracellularnonplasmalemmal clathrin coats (4), as well as the analysis ofthe transport of a viral protein showing two waves ofclathrin-coated vesicles along the RER-plasma membraneroute (14, 15), suggests that membrane interactions at boththe cis and trans poles ofthe Golgi apparatus involve clathrin.The results of the present study indicate that a marked levelof clathrin immunoreactivity is detected only at the transGolgi region. The previous immunocytochemical character-ization (11) ofour polyclonal antibody to clathrin showed thatit recognizes clathrin polymerized into coats rather than the"soluble" pool of clathrin, presumably in the form oftriskelions (16). Although the Lowicryl embedding needed todemonstrate clathrin immunoreactive sites at the ultrastruc-tural level does not permit optimal visualization of coats orcages, it is assumed in the following discussion that anti-clathrin-rich sites on immunostained sections correspondmost likely to regions underlined with structural coats. At thecis Golgi pole, microvesicles that are thought to mediate theRER-Golgi transfer of secretory and membrane proteins (1)reside in a cytosolic matrix with a microfibrillar component.Typical clathrin coats are not seen in association with suchvesicles. The immunocytochemical data would not exclude,however, the presence in this region of triskelions of clathrin.Alternatively, and this by analogy with the formation ofsmooth endosomes from clathrin-coated pits (17), one couldenvisage that the association of a clathrin coat with transfervesicles is a more rapid phenomenon than at the trans pole.For example, the role of clathrin coats at this level could berestricted to the step of fission from transitional elements ofthe RER followed by the rapid dissociation of the coats. Itshould be pointed out, however, that the identification ofbona fide clathrin coats on transitional elements has re-mained elusive (18). Other possibilities that must be envis-aged to account for the low labeling at the cis pole are thatonly a limited number of clathrin-coated vesicles are suffi-cient to mediate the RER-Golgi transfer, that clathrin coatsdeparting from the classical pattern (16, 19) are involved, orthat a coat protein different from clathrin is concerned. Thesealternatives cannot be resolved at present, nor do we knowwhether individual clathrin triskelions, without forming a

FIG. 2 (on following page). (A) Thin section of B cells immunostained with anti-clathrin antibody revealed by the protein A-gold technique.The field comprises the three areas (delimited here with vertical bars) in which the immunolabeling was quantitated. From top to bottom:subplasmalemmal rim (spr), cis Golgi region, trans Golgi region. The low degree of immunolabeling of the subplasmalemmal rim and cis Golgi

5386 Cell Biology: Orci et al.

Cell Biology: Orci et al. Proc. Natl Acad. Sci. USA 82 (1985) 538

v- ;W- 4

7p~ ~ ~ ~ ~ c5* %

,30.~ ~ ~ SC~

it ~ ~rt~j'

A-6~~~~~~~~~~~~~~~~~~~~

~~~~~~~~~~~~~~~~~F~~~~~~~~~~~~~Oi<4>4.1`41

t.....

4- V~~~~~~~~~~~~~'¶~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

'p C' ~~~~~~~~~~~~~~~~~~~~~K9' 'a ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l

7,4~ ~ ~ ~ ~

region can be compared with the marked immunoreactivity of the trans region. (Lowicryl embedding; gold particles 15 nm; x34,000.) (B)Thin section of a Golgi area showing the asymmetrical distribution of clathrin immunoreactivity across the Golgi apparatus. While the cis regionhas a low degree of immunostaining, numerous gold particles appear on the trans region, where they are associated with the limiting membraneof both cisternae (arrows) and newly formed secretory granules (asterisks). (Lowicryl embedding; gold particles 6 nm; x47,000.) (C) Goldparticles revealing clathrin immunoreactive sites decorate the membrane of extensive segments of trans Golgi cisternae, as well as of numerousvesicles (arrows). (Lowicryl embedding; gold particles 6 nm; x38,000.) (D) Detail of a maturing secretory granule with extensive clathrinimmunolabeling of the limiting membrane (Lowicryl embedding; gold particles 6 nm; x49,000.) (E) Golgi area in a glucagon-containing A cellof the islet of Langerhans. The cis-trans asymmetry of the clathrin immunostaining is evident, with most gold particles concentrated on elementsat the trans pole, including a newly formed (maturing) secretory granule (arrow). (Lowicryl embedding; gold particles 6 nm; X31,000.) (F)Clathrin immunolabeling of several secretory granules in the Golgi area of an A cell. (Lowicryl embedding; gold particles 15 nm; x42,000.)

87

5388 Cell Biology: Orci et al.

, , * 1 !*b;*s-;Ltow *vl'I

.,..o

** ..i'.

a._

Proc. Natl. Acad. Sci. USA 82 (1985)

(A~~~~~~~~~~~~~~~~~~~~~A

t..'N

s;"04P

SF; tt;. : ]ZOle

a45lt4

|W~t~t~AAWK

FusL\'s'v;W w T- * N ;

"l'izsn':k': +

;7V '

A'A _ t

W.

.,pII

i 6..,

P.I

*Aiwlb'%- -j'..

FIG. 3. (A and B) Two consecutive serial sections of the Golgi area of a prolactin-containing cell showing the characteristic polymorphousshapes of maturing secretory granules. The serial sections allow the identification of the extensive coating (indicated by broken lines) of themembrane on the same granule. (Osmium tetroxide postfixation; Epon embedding; x42,000.) (Inset) Clathrin immunolabeling of severalmaturing granules. (Lowicryl embedding; gold particles = 6 nm; x56,000.) (C) Golgi area of a somatotroph cell after clathrin immunostaining.The asymmetrical distribution of clathrin immunoreactive sites across the Golgi apparatus is evident, with the marked labeling of the peripheryof maturing secretory granules (asterisks) at the trans pole. (Lowicryl embedding; gold particles 6 nm; x35,000.) (Inset) Clathrin

immunolabeling of the limiting membrane of a maturing secretory granule in a somatotroph cell. (Lowicryl embedding; gold particles 6 nm;

x41,000.)

a '

4T%~~~~~~~~~~~~'

S +e ~ e^ ' ' 4',-

4"7 4*.£g t < t* 9> .a iwS Asv !tj ,th~~~~~~. !.;~, . k>i>

r-; -W - It' "'...", 11 :.,4,: 41 .. "

.1 ..V

I.., .., 1.

Proc. Natl. Acad. Sci. USA 82 (1985) 5389

recognizable coat, may mediate some sort of intracellularmobilization of membranes. If, however, as for the plasmamembrane step, one takes typical clathrin coats as themorphologic markers of what we propose to call "receptor-mediated, nonplasmalemmal intracellular transport,"t thepresent results would indicate that RER-Golgi transport,possibly including cis-trans migration, is not comparable totrans Golgi-plasma membrane transport.The asymmetrical clathrin immunolabeling of Golgi mem-

branes can be viewed as an additional expression of hetero-geneity across the Golgi complex, together with previouslyevidenced changes of intramembrane particle and fili-pin-sterol complex concentration (for review, see ref. 13),enzymes (for review see ref. 22), or antigens-e.g., the135-kDa protein (23). Besides their postulated role in mem-brane interactions (4), clathrin-coated vesicles were shown toacidify the vesicle content in vitro (24, 25). Isolated Golgimembranes also contain proton pumps (26) and, in the case

of pancreatic B cells, the relationship between the transcoated compartment and the conversion of proinsulin seems

interesting; a clathrin-coated Golgi-related compartment hasrecently been shown to be linked to the conversion ofproinsulin (27, 28), and a proinsulin converting enzyme was

identified as a thiol protease with an acidic pH optimum (29).The coated compartment of B cells (and of other polypeptidehormone-secreting cells where precursor/product conver-

sion occurs) may be a critically acid compartment involved inthe proteolytic processing of prohormones.

tThe best characterized system in this respect is that involvingmannose 6-phosphate receptors for lysosomal enzymes (for reviewsee ref. 20); recent immunocytochemical data also suggest thepresence of (pro)insulin binding sites on membranes of the Golgiapparatus in B cells (21).

We thank A.-M. Lucini, G. Perrelet, and G. Tripet for technicalassistance, G. Negro and P.-A. Ruttimann for photographic work,and I. Bernard for typing the manuscript. This work was supportedby Grant 3.460.83 from the Swiss National Science Foundation.

1. Palade, G. E. (1975) Science 189, 347-358.2. Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. (1979)

Nature (London) 279, 679-685.

3. Pearse, B. M. F. (1976) Proc. Natl. Acad. Sci. USA 73,1255-1259.

4. Bretscher, M. S. & Pearse, B. M. F. (1981) Annu. Rev.Biochem. 50, 85-101.

5. Lacy, P. E. & Kostianovsky, M. (1967) Diabetes 16, 35-39.6. Armbruster, B. L., Carlemalm, E., Chiovetti. R., Garavito,

R. M., Hobot, J. A., Kellenberger, E. & Villiger, W. (1982) J.Microsc. (Oxford) 126, 77-85.

7. Coons, A., Leduc, E. H. & Connolly, J. (1955) J. Exp. Med.102, 49-63.

8. Maxwell, M. H. (1978) J. Microsc. (Oxford) 112, 253-255.9. Roth, J., Bendayan, M. & Orci, L. (1978) J. Histochem.

Cytochem. 26, 1074-1081.10. Slot, J. W. & Geuze, H. J. (1981) J. Cell Biol. 90, 533-536.11. Louvard, D., Morris, C., Warren, G., Stanley, K., Winkler, F.

& Reggio, H. (1983) EMBO J. 2, 1655-1664.12. Pictet, R., Orci, L., Gonet, A. E., Rouiller, C. & Renold,

A. E. (1967) Diabetologia 3, 188-211.13. Orci, L. (1982) Diabetes 31, 538-565.14. Rothman, J. E. & Fine, R. E. (1980) Proc. Natl. Acad. Sci.

USA 77, 780-784.15. Rothman, J. E., Bursztyn-Pettigrew, H. & Fine, R. E. (1980)

J. Cell Biol. 86, 162-171.16. Ungewickell, E. & Branton, D. (1981) Nature (London) 289,

420-422.17. Helenius, A., Mellman, I., Wall, D. & Hubbard. A. (1983)

Trends Biochein. Sci. 8, 245-250.18. Palade, G. E. & Fletcher, M. (1977) J. Cell Biol. 75, 371a

(abstr.).19. Crowther, R. A. & Pearse, B. M. F. (1982) J. Cell Biol. 91,

790-797.20. Sly, W. S. & Fischer, H. D. (1982) J. Cell Biochern. 18, 67-85.21. Orci, L., Ravazzola, M. & Perrelet, A. (1984) Proc. Natl.

Acad. Sci. USA 81, 6743-6746.22. Farquhar, M. G. & Palade, G. E. (1981) J. Cell Biol. 91,

77s-103s.23. Louvard, D., Reggio, H. & Warren, G. (1982) J. Cell Biol. 92,

92-107.24. Stone, D. K., Xie, X.-S. & Racker, E. (1983) J. Biol. Chem.

258, 4059-4062.25. Forgac, M., Cantley, L., Wiedenmann, B., Altstiel, L. &

Branton, D. (1983) Proc. Natl. Acad. Sci. USA 80, 1300-1303.26. Glickman, J., Croen, K., Kelly, S. & Al-Awqati, Q. (1983) J.

Cell Biol. 97, 1303-1308.27. Orci, L., Halban, P., Amherdt, M., Ravazzola, M., Vassalli,

J.-D. & Perrelet, A. (1984) Cell 39, 39-47.28. Orci, L., Halban, P., Amherdt, M., Ravazzola, M., Vassalli,

J.-D. & Perrelet, A. (1985) J. Cell Biol. 99, 2187-2192.29. Docherty, K., Carroll, R. J. & Steiner, D. F. (1982) Proc.

Natl. Acad. Sci. USA 79, 4613-4617.

Cell Biology: Orci et A