Proc. Nat. Acad. Sci. USA Vol. 72, No. 11, pp. 4501-4505, November 1975
Cell Biology
Cytochemical study of secretory process in transplantable insulinoma of Syrian golden hamster (electron microscopy/enzyme localizations/GERL/endoplasmic reticulum/Golgi apparatus)
ALEX B. NOVIKOFF, ANA YAM, AND PHYLLIS M. NOVIKOFF Department of Pathology, Albert Einstein College of Medicine, Yeshiva University, Bronx, New York 10461
Contributed by Alex B. Novikoff, August 28,1975
the trans element of the Golgi apparatus (like its other elements) shows no such activity.
ABSTRACT Electron microscopy, including phosphatase cytochemistry, indicates thatthe secretory granules of an insulinoma producing proinsulin and insulin are packaged by the endoplasmic reticulum (ER) and especially by a specialized region of ER which we call GERL bcause of the spatial relationship of this region to the Golgi apparatus and its apparent role in producing lysosomes. The granules are not derived from the Golgi apparatus. Preliminary evidence suggests this may be true also of pancreatic f-cells. Since 1967 when Steiner and Oyer (1) discovered proinsulin in a human islet cell adenoma, much has been learned about the proteolytic conversion of proinsulin to insulin (2, 3). Concerning the intracellular localization of this conversion, Steiner et al. (2) have suggested that this conversion "probably begins when the polypeptide reaches the Golgi apparatus and continues for several hours after new secretory granules have formed within the lamellae of the Golgi apparatus." It was of interest to study the roles, if any, of GERL* and of the ER outside of the Golgi zone in such secretory granule packaging. Drs. Ora M. Rosen and Stephen G. Baum of the Department of Medicine made available to us Syrian golden hamsters bearing the transplantable pancreatic islet-cell tumor 2309 initiated by Kirkman in 1962 (6). A recent publication from their laboratories (7) may be consulted for earlier studies on this insulinoma. Areas with well preserved ultrastructure are easy to find in the comparatively homogeneous tumor, and the presence of relatively few secretory granules in the tumor cell as compared to the pancreatic fl-cell facilitates morphological study of temporal events in the secretory process. The biochemical conversion of proinsulin to insulin is thought to be similar in insulinomas and normal fl-cells (2). Preliminary observations in our laboratory, as well as two published accounts on the fl-cell (8, 9), suggest that some basic cytological aspects of the secretory process are similar in tumor and normal cells. As in earlier studies from our laboratory (10, 11), two phosphatase markers are used to distinguish the Golgi apparatus from GERL. The trans (12) element of the Golgi apparatus shows thiamine pyrophosphatase (TPPase) activity whereas GERL does not. On the other hand, GERL shows activity for acid phosphatase [orthophosphoric-monoester phosphohydrolase (acid optimum), EC 3.1.3.2] (AcPase) but
MATERIALS AND METHODS The tumors used in these experiments were removed from etherized hamsters 18-24 days after subcutaneous inoculation in the thigh (7). Thin slices of grossly non-necrotic pink tumor just below the capsule were quickly immersed in a cold (40) solution containing 2.5% glutaraldehyde, 2% formaldehyde (prepared from paraformaldehyde), 0.09 M cacodylate buffer at pH 7.4 and 0.25% CaCl2 (13) and fixed for 3 hr. Further processing for morphological study by electron microscopy has been described elsewhere (10). Electron microscopic localizations of TPPase and AcPase activities were determined as previously described (10, 11) except that in the present experiments the nonfrozen "chopper" sections were 20,um thick. Light microscopic studies will be described elsewhere. TPPase and AcPase preparations emphasize the very extensive nature of Golgi apparatus and GERL, respectively, despite the small cytoplasm to nucleus ratio characteristic of tumor cells. Many bodies that at first appear to be lysosomes in the AcPase preparations prove, when carefully studied under oil immersion, to be portions of GERL. Since many of the apparent secretory granules proved to have an electron microscopic appearance resembling some small microperoxisomes (14), tissue sections were incubated in the 3,3'-diaminobenzidine (pH 9.7) medium that reveals these organelles (15). The secretory granules were invariably unreactive. RESULTS
Golgi apparatus In the electron microscope, the Golgi apparatus appears as a stack of smooth-membraned saccules or elements; when we refer to electron micrographs we will usually use the term Golgi stack. In the insulinoma cells the stack consists of three to five elements. The ER cisterna at the cis (12) aspect of the Golgi stack is characterized by an absence of ribosomes on the membrane surface adjacent to the stack (long arrow, Fig. 1). Material resembling secretory material is not encountered either in these areas of ER or in any of the Golgi elements. Secretory granules are not seen attached to Golgi elements. The Golgi stack is interrupted by numerous small gaps, as previously found in neurons (see Figs. 13-15 in ref. 10). Within these gaps a great many microfilaments are seen, together with elements of GERL or with rough ER in apparent transition to GERL. The trans Golgi element is clearly
Abbreviations: AcPase, acid phosphatase; ER, endoplasmic reticulum; TPPase, thiamine pyrophosphatase. * Our laboratory has suggested that many cell types may package lysosomes and other structures, including secretory granules, in a specialized region of endoplasmic reticulum (ER) designated as GERL (4, 5). The acronym is derived from the spatial relation of this region of ER to the Golgi apparatus and its apparent role in producing Lysosomes.
4501
4502
Cell Biology: Novikoff et al.
Proc. Nat. Acad. Sci. USA 72 (1975)
its ,-t
a
f
'+
*)*ww
s
be A
K4 . . *. 'k* . . 4.e
.%;fr..
Abbreviations in Figures: B, residual bodies; C, coated vesicles; CE, centriole; ER, endoplasmic reticulum; F, microfilaments; G, secretory granule; GE, GERL; M, mitochondrion; N, nucleus; PM, plasma membrane; S, secretory material within expanded areas of GERL; T, microtubule; V, coated vesicles continuous with GERL. Figs 1-10 are portions of insulinoma cells and Fig. 11 is a portion of a j3-ceil from rat pancreas. Figs. 1-3 and 8-11 are from sections stained with lead citrate alone and Figs. 4-7 are from sections stained with both uranyl acetate and lead citrate. FIGS. 1-4. The large arrow in Fig. 1 indicates an ER cisterna lacking ribosomes from its membrane surface facing the Golgi stack. In Figs. 1 and 2, the arrowheads indicate GERL and the small arrows indicate ribosomes in the region of transition from rough ER to GERL (see text). In Fig. 3 the inset is an enlarged area with the apparent origin of three secretory granules from ER. The inset in Fig. 4 is an enlarged area showing apparent fusion of a coated vesicle with a secretory granule. Fig. 1, X37,Q00; Fig. 2, X48,000; Fig. 3, X33,000; inset X49,000; Fig. 4, X30,000; inset, X61,000.
marked by TPPase activity whereas GERL elements are unreactive (Fig. 8). Neither secretory granules nor coated vesicles show TPPase activity. If the tissue is not incubated for TPPase activity or AcPase activity it is difficult, particularly without experience, to distinguish the trans element of the Golgi stack from adjacent elements of GERL (Figs. 1 and 2). However, the distinction becomes evident at once if sections are incubated for the phosphatase activities.
GERL In contrast to the trans Golgi element that is without demonstrable AcPase activity, all elements of GERL show this activity (Figs. 9 and 10). GERL occupies the areas of cytoplasm bounded by the Golgi apparatus. These areas are always at the trans face of the Golgi stack irrespective of the twists and turns of the Golgi apparatus (Figs. 1, 2, 8, 9, and 10).
Cell Biology: Novikoff et al.
Proc. Nat. Acad. Sci. USA 72 (1975)
4503
:Rt *
mXhs+ *ar
~~~~~4
f,
I
4~~~~~~!a.-4~ ~ ~ ~ ~ lf
4.
'44~~~~~~~~~~~~~~~~~ .A
I1
4 .4
'.4-
'}act At('t'X~~~~~~~pp-, tt ~~~~~~~~~~~ "4.
.:I me a, r f , j,
X
C ..
- .4E
I
..X
;~~~Fz ax'
S
:
,
j
.~
-
,S,
14C
Ail
Ld
FIGS. 5-11. Rarely is a cell seen like that of Fig. 5 with numerous secretory granules near the plasma membrane (PM). In Fig. 6 the arrowhead indicates an area where membrane of secretory granule and plasma membrane appear continuous; in adjacent areas the membranes of the two overlap. Fig. 7. The coated vesicle (C) to the right is apparently fusing with a secretory granule. Fig. 8 is from tissue incubated for TPPase activity. The inner (trans) element of the Golgi apparatus (GO) shows reaction product. GERL (arrowheads) and secretory granules (G) lack reaction product. Secretory material (S) is seen within dilated portions of GERL; the uppermost element shows the membrane continuity most clearly. Figs. 9-11 are from sections incubated for AcPase activity. Reaction product is present in GERL (GE) but not in the Golgi apparatus (GO). All secretory granules in Fig. 9 show reaction product. The inset in Fig. 9 shows adjacent secretory granules from another cell, one with reaction prodpct and the other without it. In Fig. 11 reaction product is seen in GERL (arrows) and adjacent secretory granules (G). The Golgi apparatus (GO) is unreactive. Fig. 5, X29,000; Figs. -6 and 7, X94,000; Fig. 8, X33,000; Fig. 9, X17,000; inset X37,000; Fig. 10, X36,000; Fig. 11, X34,000.
GERL elements often appear electron opaque (Fig. 2), probably because they are narrow enough to be included in toto within the thickness of the section and they are thus viewed through their moderately electron-opaque membrane. Adjacent to GERL are numerous small secretory granules (G in the figures). These granules appear to arise by accumulation of electron-opaque secretory material within ex-
panded areas of GERL (S in Figs. 1, 2, and 8). The average maximum diameter of the secretory granules is 0.16 Am. Most significantly, both secretory granules and coated vesicles in the Golgi zone demonstrate AcPase activity. As suggested below, the coated vesicles may be involved in transporting lysosomal hydrolases to secretory granules that appear to bud directly from peripheral ER. Numerous coated vesicles also appear to bud from GERL (C in the figures; see
4504
Cell Biology: Novikoff et al.
particularly the two vesicles at V in Fig. 1). The coated vesicles are readily distinguishable from the secretory granules: the vesicles -lack the content of the secretory granules and they possess characteristic peripheral "spikes". The average maximum diameter of the coated vesicles is 0.11 Mm. In the insulinoma cells it is difficult to draw a sharp dividing line between the rough ER tubules and the smooth GERL tubules. Particularly in or near the small gaps in the Golgi stack, the ribosomes are seen in a distinctive pattern that invites speculation. Some are attached to membranes and some are free in the adjacent hyaloplasm, which creates *the impression that rough ER gives rise to smooth elements of GERL by loss of ribosomes (small arrows in Figs. 1 and 2). Some ribosomes are still present in areas of GERL. Secretory granules Most secretory granules appear to arise from GERL (Figs. 1, 2, and 8) and these show AcPase activity (Fig. 9), but there seems to be a second packaging mechanism. Some secretory granules appear to bud from ER not seen to be part of GERL or to be within the Golgi zone (Figs. 3 and 7). (In the absence of serial sections the relationship of this region of ER to GERL or Golgi zone cannot be ascertained.) This ER is peripherally located in the cell (Fig. 3). In contrast to the granules near GERL, some peripheral secretory granules do not show AcPase activity (inset, Fig. 9). Continuities of peripheral secretory granules and coated vesicles, as if the two were fusing, are encountered (Figs. 4 and 7). Secretory granules are often seen adjacent to the plasma membrane (Figs. 4-6). Images suggesting exocytosis are frequent (Figs. 4 and 6), but unequivocal openings bounded by a continuous membrane have not been encountered. Occasionally, a cell is so sectioned that many secretory granules are seen at or near the plasma membrane (Fig. 5). Neither the frequency of such cells nor their functional state is known. The contents of the secretory granules have various electron opacities, particularly if the thin sections are not stained with uranyl acetate prior to lead citrate staining (Fig. 3). In endocrine cells of the pituitary gland Smith and Farquhar (16) and Farquhar (17) have described the degradation of secretory granules (crinophagy) within two types of lysosomes, multivesicular bodies and residual bodies. Granule degradation is not observed in the insulinoma cells we have studied. Very few autophagic vacuoles (1) are observed in these cells; none was found to include a recognizable secretory granule. Preliminary observations on 3-cells and exocrine cells In the rat pancreas both (3-cells (Fig. 11) and exocrine cells (to be reported elsewhere) show AcPase-positive GERL at the trans aspect of the AcPase-negative Golgi stack. AcPasepositive secretory granules are present near GERL in both cell types. In the exocrine cell (of the rabbit as well as of the rat) the contents of the positive granules appear less condensed and resemble that in condensing vacuoles (18, 19). In at least some (3-cells, not only GERL but the ER elsewhere shows AcPase activity. The (3-cell shows many coated vesicles near GERL and these show AcPase activity. DISCUSSION Two observations suggest that a major area of packaging secretory material in cells of the hamster insulinoma is the specialized region of ER known as GERL: (a) secretory granules appear to bud from GERL, and (b) the cytochemically-
Proc. Nat. Acad. Sci. USA 72 (1975)
demonstrable enzymatic activities of the granules are those of GERL because they show AcPase activity (characteristic of GERL) but not TPPase activity (characteristic of the trans element of the Golgi stack). Coated vesicles do not appear to be involved in this packaging. GERL appears also to play a packaging role in pancreatic (3-cells. The report of Orci et al. (9) is consistent with our preliminary observations on rat pancreas; the same is true of the study of Lazarus et al. (8) on rabbit pancreas. AcPase is present in GERL (called inner cisternae of the "Golgi complex" by them) and in the secretory granules near GERL ["immature granules" (20)]. In insulinoma cells the coated vesicles may play a role in processing the secretory granules that appear to bud from peripheral ER. Demonstrable AcPase activity is missing from some of the peripheral granules, and fusions of coated vesicles with peripheral granules are seen. This raises the possibility that lysomal hydrolases are transported to these granules via the coated vesicles. It may be relevant that in (3-cells of rabbit pancreas, Lazarus et al. (8) found some secretory granules inside rough ER cisternae. Biochemical studies are required for: (a) testing our tacit assumption that when AcPase activity is revealed in a cytological entity it is likely that other lysosomal enzymes are present there as well, and (b) resolving the issues considered by Steiner et al. (4) in their discussion of the possible involvement of lysosomes in converting proinsulin to insulin. The hamster insulinoma may be a good material for such studies since sizeable quantities of relatively homogeneous tissue are readily available. It is possible that secretory granules, coated vesicles, and ER vesicles that form during tissue homogenization could be separated by density equilibration centrifugation and tested for proteolytic and other lysosomal enzyme activities. The data could resolve questions such as the fate of GERL during homogenization and the significance of the different electron opacity among secretory granules. Differences between insulinoma cells and (3-cells to be borne in mind are the relative paucity and small size of the secretory granules in the insulinoma (correlated with lower hormone content) and the presence of AcPase activity in almost all the secretory granules. We gratefully acknowledge the preparation of the final photographs by Mr. George Dominguez. This work was supported by U.S. Public Health Service Research Grant CA06576. A.B.N. is the recipient of the U.S. Public Health Service Research Career Award 5K6 CA14923 from the National Cancer Institute. We are indebted to Drs. Baum and Rosen for numerous discussions and for providing us with the tumor-bearing hamsters; the work of their laboratories is supported by U.S. Public Health Service Research Grant 5P01 GM19100 and National Science Foundation Grant GB43581. 1. Steiner, D. F. & Oyer, P. E. (1967) Proc. Nat. Acad. Sci. USA
57,473-480. 2. Steiner, D. F., Kemmler, W., Clark, J. L., Oyer, P. E. & Rubenstein, A. H. (1972) in Handbook of Physiology-Endocrine Pancreas, eds. Steiner, D. F. & Freinkel, N. (Am. Physiol. Soc., Washington, D.C.), Vol. I, pp. 175-198. 3. Steiner, D. F., Kemmler, W., Tager, H. S. & Rubenstein, A. H. (1974) Advances in Cytopharmacology, eds. Ceccarelli, B., Clementi, F. & Meldolesi, J. (Raven Press, New York), Vol. 2, pp. 195-205. 4. Novikoff, A. B. (1973) in Lysosomes and Storage Diseases, eds. Hers, H. G. & Van Hoof, F. (Academic Press, New York), C pp. 1-41. 5. Novikoff, A. B., Novikoff, P. M., Ma, M., Shin, W.-Y. & Quin-
Cell Biology: Novikoff et al.
6. 7. 8. 9. 10. 11. 12.
tana, N. (1974) in Advances in Cytopharmacology, eds. Ceccarelli, B., Clementi, F. & Meldolesi, J. (Raven Press, New York), Vol. 2, pp. 349-368. Kirkman, N. (1962) Stanford Med. Bull. 20,163-166. Shapiro, S., Eto, S., Fleischer, N. & Baum, S. G. (1975) Endocrinology 97, 442-447. Lazarus, S. S., Volk, B. W. & Barden, H. (1966) J. Histochem. Cytochem. 14, 233-246. Orci, L., Stauffacher, W., Rufener, C., Lambert, A. E., Rouiller, C. & Renold, A. E. (1971) Diabetes 20,385-388. Novikoff, P. M., Novikoff, A. B., Quintana, N. & Hauw, J.-J. (1971) J. Cell Biol. 50, 859-886. Boutry, J.-M. & Novikoff, A. B. (1975) Proc. Nat. Acad. Sci. USA 72,508-512. Ehrenreich, J. H., Bergeron, J. J. M., Siekevitz, P. & Palade, G.
Proc. Nat. Acad. Sci. USA 72 (1975)
4505
E. (1973) J. Cell Biol. 59,45-72. 13. Miller, F. & Herzog, V. (1969) Z. Zellforsch. Mikrosk. Anat.
97,84-110. 14. Novikoff, A. B. & Novikoff, P. M. (1973) J. Histochem. Cytochem. 21, 963-966. 15. Novikoff, A. B., Novikoff, P. M., Davis, C. & Quintana, N. (1972) J. Histochem. Cytochem. 20, 1006-1022. 16. Smith, R. E. & Farquhar, M. G. (1966) J. Cell Biol. 31, 319347. 17. Farquhar, M. G. (1969) in Lysosomes in Biology and Pathology, eds. Dingle, J. T. & Fell, H. B. (North-Holland, Amsterdam), Vol. 2, pp. 462-482. 18. Palade, G. (1975) Science 189, 347-358. 19. Van Heyningen, H. E. (1964) Anat. Rec. 148, 485-489. 20. Novikoff, A. B. & Essner, E. (1962) Fed. Proc. 21, 1130-1142.
Cell Biology
Cytochemical study of secretory process in transplantable insulinoma of Syrian golden hamster (electron microscopy/enzyme localizations/GERL/endoplasmic reticulum/Golgi apparatus)
ALEX B. NOVIKOFF, ANA YAM, AND PHYLLIS M. NOVIKOFF Department of Pathology, Albert Einstein College of Medicine, Yeshiva University, Bronx, New York 10461
Contributed by Alex B. Novikoff, August 28,1975
the trans element of the Golgi apparatus (like its other elements) shows no such activity.
ABSTRACT Electron microscopy, including phosphatase cytochemistry, indicates thatthe secretory granules of an insulinoma producing proinsulin and insulin are packaged by the endoplasmic reticulum (ER) and especially by a specialized region of ER which we call GERL bcause of the spatial relationship of this region to the Golgi apparatus and its apparent role in producing lysosomes. The granules are not derived from the Golgi apparatus. Preliminary evidence suggests this may be true also of pancreatic f-cells. Since 1967 when Steiner and Oyer (1) discovered proinsulin in a human islet cell adenoma, much has been learned about the proteolytic conversion of proinsulin to insulin (2, 3). Concerning the intracellular localization of this conversion, Steiner et al. (2) have suggested that this conversion "probably begins when the polypeptide reaches the Golgi apparatus and continues for several hours after new secretory granules have formed within the lamellae of the Golgi apparatus." It was of interest to study the roles, if any, of GERL* and of the ER outside of the Golgi zone in such secretory granule packaging. Drs. Ora M. Rosen and Stephen G. Baum of the Department of Medicine made available to us Syrian golden hamsters bearing the transplantable pancreatic islet-cell tumor 2309 initiated by Kirkman in 1962 (6). A recent publication from their laboratories (7) may be consulted for earlier studies on this insulinoma. Areas with well preserved ultrastructure are easy to find in the comparatively homogeneous tumor, and the presence of relatively few secretory granules in the tumor cell as compared to the pancreatic fl-cell facilitates morphological study of temporal events in the secretory process. The biochemical conversion of proinsulin to insulin is thought to be similar in insulinomas and normal fl-cells (2). Preliminary observations in our laboratory, as well as two published accounts on the fl-cell (8, 9), suggest that some basic cytological aspects of the secretory process are similar in tumor and normal cells. As in earlier studies from our laboratory (10, 11), two phosphatase markers are used to distinguish the Golgi apparatus from GERL. The trans (12) element of the Golgi apparatus shows thiamine pyrophosphatase (TPPase) activity whereas GERL does not. On the other hand, GERL shows activity for acid phosphatase [orthophosphoric-monoester phosphohydrolase (acid optimum), EC 3.1.3.2] (AcPase) but
MATERIALS AND METHODS The tumors used in these experiments were removed from etherized hamsters 18-24 days after subcutaneous inoculation in the thigh (7). Thin slices of grossly non-necrotic pink tumor just below the capsule were quickly immersed in a cold (40) solution containing 2.5% glutaraldehyde, 2% formaldehyde (prepared from paraformaldehyde), 0.09 M cacodylate buffer at pH 7.4 and 0.25% CaCl2 (13) and fixed for 3 hr. Further processing for morphological study by electron microscopy has been described elsewhere (10). Electron microscopic localizations of TPPase and AcPase activities were determined as previously described (10, 11) except that in the present experiments the nonfrozen "chopper" sections were 20,um thick. Light microscopic studies will be described elsewhere. TPPase and AcPase preparations emphasize the very extensive nature of Golgi apparatus and GERL, respectively, despite the small cytoplasm to nucleus ratio characteristic of tumor cells. Many bodies that at first appear to be lysosomes in the AcPase preparations prove, when carefully studied under oil immersion, to be portions of GERL. Since many of the apparent secretory granules proved to have an electron microscopic appearance resembling some small microperoxisomes (14), tissue sections were incubated in the 3,3'-diaminobenzidine (pH 9.7) medium that reveals these organelles (15). The secretory granules were invariably unreactive. RESULTS
Golgi apparatus In the electron microscope, the Golgi apparatus appears as a stack of smooth-membraned saccules or elements; when we refer to electron micrographs we will usually use the term Golgi stack. In the insulinoma cells the stack consists of three to five elements. The ER cisterna at the cis (12) aspect of the Golgi stack is characterized by an absence of ribosomes on the membrane surface adjacent to the stack (long arrow, Fig. 1). Material resembling secretory material is not encountered either in these areas of ER or in any of the Golgi elements. Secretory granules are not seen attached to Golgi elements. The Golgi stack is interrupted by numerous small gaps, as previously found in neurons (see Figs. 13-15 in ref. 10). Within these gaps a great many microfilaments are seen, together with elements of GERL or with rough ER in apparent transition to GERL. The trans Golgi element is clearly
Abbreviations: AcPase, acid phosphatase; ER, endoplasmic reticulum; TPPase, thiamine pyrophosphatase. * Our laboratory has suggested that many cell types may package lysosomes and other structures, including secretory granules, in a specialized region of endoplasmic reticulum (ER) designated as GERL (4, 5). The acronym is derived from the spatial relation of this region of ER to the Golgi apparatus and its apparent role in producing Lysosomes.
4501
4502
Cell Biology: Novikoff et al.
Proc. Nat. Acad. Sci. USA 72 (1975)
its ,-t
a
f
'+
*)*ww
s
be A
K4 . . *. 'k* . . 4.e
.%;fr..
Abbreviations in Figures: B, residual bodies; C, coated vesicles; CE, centriole; ER, endoplasmic reticulum; F, microfilaments; G, secretory granule; GE, GERL; M, mitochondrion; N, nucleus; PM, plasma membrane; S, secretory material within expanded areas of GERL; T, microtubule; V, coated vesicles continuous with GERL. Figs 1-10 are portions of insulinoma cells and Fig. 11 is a portion of a j3-ceil from rat pancreas. Figs. 1-3 and 8-11 are from sections stained with lead citrate alone and Figs. 4-7 are from sections stained with both uranyl acetate and lead citrate. FIGS. 1-4. The large arrow in Fig. 1 indicates an ER cisterna lacking ribosomes from its membrane surface facing the Golgi stack. In Figs. 1 and 2, the arrowheads indicate GERL and the small arrows indicate ribosomes in the region of transition from rough ER to GERL (see text). In Fig. 3 the inset is an enlarged area with the apparent origin of three secretory granules from ER. The inset in Fig. 4 is an enlarged area showing apparent fusion of a coated vesicle with a secretory granule. Fig. 1, X37,Q00; Fig. 2, X48,000; Fig. 3, X33,000; inset X49,000; Fig. 4, X30,000; inset, X61,000.
marked by TPPase activity whereas GERL elements are unreactive (Fig. 8). Neither secretory granules nor coated vesicles show TPPase activity. If the tissue is not incubated for TPPase activity or AcPase activity it is difficult, particularly without experience, to distinguish the trans element of the Golgi stack from adjacent elements of GERL (Figs. 1 and 2). However, the distinction becomes evident at once if sections are incubated for the phosphatase activities.
GERL In contrast to the trans Golgi element that is without demonstrable AcPase activity, all elements of GERL show this activity (Figs. 9 and 10). GERL occupies the areas of cytoplasm bounded by the Golgi apparatus. These areas are always at the trans face of the Golgi stack irrespective of the twists and turns of the Golgi apparatus (Figs. 1, 2, 8, 9, and 10).
Cell Biology: Novikoff et al.
Proc. Nat. Acad. Sci. USA 72 (1975)
4503
:Rt *
mXhs+ *ar
~~~~~4
f,
I
4~~~~~~!a.-4~ ~ ~ ~ ~ lf
4.
'44~~~~~~~~~~~~~~~~~ .A
I1
4 .4
'.4-
'}act At('t'X~~~~~~~pp-, tt ~~~~~~~~~~~ "4.
.:I me a, r f , j,
X
C ..
- .4E
I
..X
;~~~Fz ax'
S
:
,
j
.~
-
,S,
14C
Ail
Ld
FIGS. 5-11. Rarely is a cell seen like that of Fig. 5 with numerous secretory granules near the plasma membrane (PM). In Fig. 6 the arrowhead indicates an area where membrane of secretory granule and plasma membrane appear continuous; in adjacent areas the membranes of the two overlap. Fig. 7. The coated vesicle (C) to the right is apparently fusing with a secretory granule. Fig. 8 is from tissue incubated for TPPase activity. The inner (trans) element of the Golgi apparatus (GO) shows reaction product. GERL (arrowheads) and secretory granules (G) lack reaction product. Secretory material (S) is seen within dilated portions of GERL; the uppermost element shows the membrane continuity most clearly. Figs. 9-11 are from sections incubated for AcPase activity. Reaction product is present in GERL (GE) but not in the Golgi apparatus (GO). All secretory granules in Fig. 9 show reaction product. The inset in Fig. 9 shows adjacent secretory granules from another cell, one with reaction prodpct and the other without it. In Fig. 11 reaction product is seen in GERL (arrows) and adjacent secretory granules (G). The Golgi apparatus (GO) is unreactive. Fig. 5, X29,000; Figs. -6 and 7, X94,000; Fig. 8, X33,000; Fig. 9, X17,000; inset X37,000; Fig. 10, X36,000; Fig. 11, X34,000.
GERL elements often appear electron opaque (Fig. 2), probably because they are narrow enough to be included in toto within the thickness of the section and they are thus viewed through their moderately electron-opaque membrane. Adjacent to GERL are numerous small secretory granules (G in the figures). These granules appear to arise by accumulation of electron-opaque secretory material within ex-
panded areas of GERL (S in Figs. 1, 2, and 8). The average maximum diameter of the secretory granules is 0.16 Am. Most significantly, both secretory granules and coated vesicles in the Golgi zone demonstrate AcPase activity. As suggested below, the coated vesicles may be involved in transporting lysosomal hydrolases to secretory granules that appear to bud directly from peripheral ER. Numerous coated vesicles also appear to bud from GERL (C in the figures; see
4504
Cell Biology: Novikoff et al.
particularly the two vesicles at V in Fig. 1). The coated vesicles are readily distinguishable from the secretory granules: the vesicles -lack the content of the secretory granules and they possess characteristic peripheral "spikes". The average maximum diameter of the coated vesicles is 0.11 Mm. In the insulinoma cells it is difficult to draw a sharp dividing line between the rough ER tubules and the smooth GERL tubules. Particularly in or near the small gaps in the Golgi stack, the ribosomes are seen in a distinctive pattern that invites speculation. Some are attached to membranes and some are free in the adjacent hyaloplasm, which creates *the impression that rough ER gives rise to smooth elements of GERL by loss of ribosomes (small arrows in Figs. 1 and 2). Some ribosomes are still present in areas of GERL. Secretory granules Most secretory granules appear to arise from GERL (Figs. 1, 2, and 8) and these show AcPase activity (Fig. 9), but there seems to be a second packaging mechanism. Some secretory granules appear to bud from ER not seen to be part of GERL or to be within the Golgi zone (Figs. 3 and 7). (In the absence of serial sections the relationship of this region of ER to GERL or Golgi zone cannot be ascertained.) This ER is peripherally located in the cell (Fig. 3). In contrast to the granules near GERL, some peripheral secretory granules do not show AcPase activity (inset, Fig. 9). Continuities of peripheral secretory granules and coated vesicles, as if the two were fusing, are encountered (Figs. 4 and 7). Secretory granules are often seen adjacent to the plasma membrane (Figs. 4-6). Images suggesting exocytosis are frequent (Figs. 4 and 6), but unequivocal openings bounded by a continuous membrane have not been encountered. Occasionally, a cell is so sectioned that many secretory granules are seen at or near the plasma membrane (Fig. 5). Neither the frequency of such cells nor their functional state is known. The contents of the secretory granules have various electron opacities, particularly if the thin sections are not stained with uranyl acetate prior to lead citrate staining (Fig. 3). In endocrine cells of the pituitary gland Smith and Farquhar (16) and Farquhar (17) have described the degradation of secretory granules (crinophagy) within two types of lysosomes, multivesicular bodies and residual bodies. Granule degradation is not observed in the insulinoma cells we have studied. Very few autophagic vacuoles (1) are observed in these cells; none was found to include a recognizable secretory granule. Preliminary observations on 3-cells and exocrine cells In the rat pancreas both (3-cells (Fig. 11) and exocrine cells (to be reported elsewhere) show AcPase-positive GERL at the trans aspect of the AcPase-negative Golgi stack. AcPasepositive secretory granules are present near GERL in both cell types. In the exocrine cell (of the rabbit as well as of the rat) the contents of the positive granules appear less condensed and resemble that in condensing vacuoles (18, 19). In at least some (3-cells, not only GERL but the ER elsewhere shows AcPase activity. The (3-cell shows many coated vesicles near GERL and these show AcPase activity. DISCUSSION Two observations suggest that a major area of packaging secretory material in cells of the hamster insulinoma is the specialized region of ER known as GERL: (a) secretory granules appear to bud from GERL, and (b) the cytochemically-
Proc. Nat. Acad. Sci. USA 72 (1975)
demonstrable enzymatic activities of the granules are those of GERL because they show AcPase activity (characteristic of GERL) but not TPPase activity (characteristic of the trans element of the Golgi stack). Coated vesicles do not appear to be involved in this packaging. GERL appears also to play a packaging role in pancreatic (3-cells. The report of Orci et al. (9) is consistent with our preliminary observations on rat pancreas; the same is true of the study of Lazarus et al. (8) on rabbit pancreas. AcPase is present in GERL (called inner cisternae of the "Golgi complex" by them) and in the secretory granules near GERL ["immature granules" (20)]. In insulinoma cells the coated vesicles may play a role in processing the secretory granules that appear to bud from peripheral ER. Demonstrable AcPase activity is missing from some of the peripheral granules, and fusions of coated vesicles with peripheral granules are seen. This raises the possibility that lysomal hydrolases are transported to these granules via the coated vesicles. It may be relevant that in (3-cells of rabbit pancreas, Lazarus et al. (8) found some secretory granules inside rough ER cisternae. Biochemical studies are required for: (a) testing our tacit assumption that when AcPase activity is revealed in a cytological entity it is likely that other lysosomal enzymes are present there as well, and (b) resolving the issues considered by Steiner et al. (4) in their discussion of the possible involvement of lysosomes in converting proinsulin to insulin. The hamster insulinoma may be a good material for such studies since sizeable quantities of relatively homogeneous tissue are readily available. It is possible that secretory granules, coated vesicles, and ER vesicles that form during tissue homogenization could be separated by density equilibration centrifugation and tested for proteolytic and other lysosomal enzyme activities. The data could resolve questions such as the fate of GERL during homogenization and the significance of the different electron opacity among secretory granules. Differences between insulinoma cells and (3-cells to be borne in mind are the relative paucity and small size of the secretory granules in the insulinoma (correlated with lower hormone content) and the presence of AcPase activity in almost all the secretory granules. We gratefully acknowledge the preparation of the final photographs by Mr. George Dominguez. This work was supported by U.S. Public Health Service Research Grant CA06576. A.B.N. is the recipient of the U.S. Public Health Service Research Career Award 5K6 CA14923 from the National Cancer Institute. We are indebted to Drs. Baum and Rosen for numerous discussions and for providing us with the tumor-bearing hamsters; the work of their laboratories is supported by U.S. Public Health Service Research Grant 5P01 GM19100 and National Science Foundation Grant GB43581. 1. Steiner, D. F. & Oyer, P. E. (1967) Proc. Nat. Acad. Sci. USA
57,473-480. 2. Steiner, D. F., Kemmler, W., Clark, J. L., Oyer, P. E. & Rubenstein, A. H. (1972) in Handbook of Physiology-Endocrine Pancreas, eds. Steiner, D. F. & Freinkel, N. (Am. Physiol. Soc., Washington, D.C.), Vol. I, pp. 175-198. 3. Steiner, D. F., Kemmler, W., Tager, H. S. & Rubenstein, A. H. (1974) Advances in Cytopharmacology, eds. Ceccarelli, B., Clementi, F. & Meldolesi, J. (Raven Press, New York), Vol. 2, pp. 195-205. 4. Novikoff, A. B. (1973) in Lysosomes and Storage Diseases, eds. Hers, H. G. & Van Hoof, F. (Academic Press, New York), C pp. 1-41. 5. Novikoff, A. B., Novikoff, P. M., Ma, M., Shin, W.-Y. & Quin-
Cell Biology: Novikoff et al.
6. 7. 8. 9. 10. 11. 12.
tana, N. (1974) in Advances in Cytopharmacology, eds. Ceccarelli, B., Clementi, F. & Meldolesi, J. (Raven Press, New York), Vol. 2, pp. 349-368. Kirkman, N. (1962) Stanford Med. Bull. 20,163-166. Shapiro, S., Eto, S., Fleischer, N. & Baum, S. G. (1975) Endocrinology 97, 442-447. Lazarus, S. S., Volk, B. W. & Barden, H. (1966) J. Histochem. Cytochem. 14, 233-246. Orci, L., Stauffacher, W., Rufener, C., Lambert, A. E., Rouiller, C. & Renold, A. E. (1971) Diabetes 20,385-388. Novikoff, P. M., Novikoff, A. B., Quintana, N. & Hauw, J.-J. (1971) J. Cell Biol. 50, 859-886. Boutry, J.-M. & Novikoff, A. B. (1975) Proc. Nat. Acad. Sci. USA 72,508-512. Ehrenreich, J. H., Bergeron, J. J. M., Siekevitz, P. & Palade, G.
Proc. Nat. Acad. Sci. USA 72 (1975)
4505
E. (1973) J. Cell Biol. 59,45-72. 13. Miller, F. & Herzog, V. (1969) Z. Zellforsch. Mikrosk. Anat.
97,84-110. 14. Novikoff, A. B. & Novikoff, P. M. (1973) J. Histochem. Cytochem. 21, 963-966. 15. Novikoff, A. B., Novikoff, P. M., Davis, C. & Quintana, N. (1972) J. Histochem. Cytochem. 20, 1006-1022. 16. Smith, R. E. & Farquhar, M. G. (1966) J. Cell Biol. 31, 319347. 17. Farquhar, M. G. (1969) in Lysosomes in Biology and Pathology, eds. Dingle, J. T. & Fell, H. B. (North-Holland, Amsterdam), Vol. 2, pp. 462-482. 18. Palade, G. (1975) Science 189, 347-358. 19. Van Heyningen, H. E. (1964) Anat. Rec. 148, 485-489. 20. Novikoff, A. B. & Essner, E. (1962) Fed. Proc. 21, 1130-1142.
