Proc. Nati. Acad. Sci. USAVol. 73, No. 8, pp.,2781-2787, August 1976Cell Biology

The endoplasmic reticulum: A cytochemist's view (A Review) *(electron microscopy/enzyme localizations/GERL/Golgi apparatus)

ALEX B. NOVIKOFFDepartment of Pathology, Albert Einstein College of Medicine of Yeshiva University, Bronx, New York 10461

Contributed by Alex B. Novikoff, March 25,1976

ABSTRACT Enzyme cytochemistry has been used, at thelight and electron microscope levels, to "mark" cytoplasmicorganelles of mammalian cells. Catalase cytochemistry per-mitted identification of microperoxisomes, apparently ubiq-uitous organelles that are attached by numerous slender cofl-nections to the endoplasmic reticulum. Thiamine pyrophos-phatase and acid phosphatase cytochemistry can be used todistinguish between the Golgi apparatus and a specializedacid-phosphatase-rich region of smooth endoplasmic reticulum(ER) that appears to be involved in: (a) the formation of lyso-somes and melanin granules: (b) the processing and ackagingof secretory materials in endocrine and exocrine cells; and (c)the metabolism of lipid. The acronym GERL has been given tothis region of smooth ER because it is located at the inner or"trans" aspect of the Golgi apparatus and because it appearsto produce various types of Lysosomes.

Cytochemistry, "marker enzymes," and specializedendoplasmic reticulum regionsBy making possible the visualization of organelles in situ, cy-tochemistry has aided interpretation of biochemical analysesof subcellular fractions isolated from homogenates. It has addeda new dimension to understanding electron microscope ob-servations. Cytoplasmic organelles not usually identifiable bylight microscopy can be observed in normal and altered cellstates.

In our laboratory we have chosen enzyme activities thatsurvive the sufficiently long aldehyde fixation required foradequate structural preservation of the organelles and that stain("mark") one or more cytoplasmic organelle (Table 1). Themanner of preparing the tissue sections and incubation mediais described in the references listed in the table.

During 1966-1970 a series of publications drew attention toan effect of lead ions employed to trap the phosphate releasedin the course of the phosphatase procedures: that the lead ionscould themselves induce hydrolysis of ATP or other phosphateesters. These papers (12-15) cast doubt upon the validity of thecytochemical procedures. In reply I pointed to: (a) specificitiesof the localizations obtained with specific substrates; (b) thesimilar localizations obtained under conditions where lead-induced hydrolysis does not occur; and (c) other observationswhich made me confident that the observed localizations didindeed reveal the intracellular sites of the different phosphatases(16-18). Thus, in 1963 we described the NDPase localizationin the endoplasmic reticulum (ER) of rat liver (Fig. 1) and weconfirmed it by assay of subcellular fractions (1). In addition,when purified from isolated microsome fractions the enzymedemonstrates the same substrate specificity and optimal con-ditions for activity as when studied by the cytochemical pro-cedure (2).

Abbreviations: AcPase, acid phosphatase; ER, endoplasmic reticulum;NDPase, nucleoside diphosphatase; TPPase, thiamine pyrophospha-tase.* By invitation. From time to time, reviews on scientific and techno-logical matters of broad interest are published in the PROCEED-INGS.

The contributions of AcPase cytochemistry to elucidatingthe forms and functions of lysosomes have been described byde Duve (19). Light microscopy of lysosomes with the AcPaseprocedure and of the Golgi apparatus with the TPPase proce-dure, in the small neurons of rat dorsal root ganglia, initiateda series of publications which led to appreciation of specializedregions of ER.

GERL and lysosomes

Light microscopic studies of a wide variety of cell types led usto identify a close relationship between the location of lysosomesand the form and distribution of the Golgi apparatus (20, 2, 21).In AcPase preparations of rat dorsal root ganglia not only thestained spherical lysosomes but also some larger stained areaswere encountered. From their distribution, size, and shape Iconsidered that these areas could be fitted within the cres-cent-shaped regions formed by portions of the Golgi apparatus

Table 1. Cytochemical marker enzymes

Refer-Organelle Enzyme ence

Endoplasmic reticulumin some cell types Nucleoside diphospha- 1, 2

tase (NDPase) (EC3.6.1.6)

in a few cell types Peroxidase 3, 4(EC 1.11.1.7)

GERL and Acid phosphatase 5lysosomes (AcPase) [orthophos-

phoric-monoesterphosphohydrolase(acid optimum),EC 3.1.3.2]

Tyrosinase (mono- 6phenol monoxy-genase) (EC 1.14.18.1)(in melanocytes)

Peroxisomes, Catalase 7, 8including (EC 1.11.1.6)microperoxisomes

Golgi apparatus Thiamine pyrophospha- 9, 2tase (TPPase)

Mitochondria Enzymes, coenzymes 7, 10that oxidize 3,3'-di-aminobenzidine

NADH-tetrazolium re- 11ductase (for lightmicroscopy only)

Plasma membrane Nucleoside phospha- 2tases

2781

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

Proc. Natl. Acad. Sci. USA 73 (1976)

4th)*>t<e

w; ot . t-t

t. W ,,4)-''",''k.

N!4

0'0 G ' <t'S5.X~. .

' ->i< ~ t0e;I

At ..

4.* B B

'Ar

E

:r~~~~ ~ ~:

* it

tRNA N~~~~~~~~~~~~~4

4.KE#':.,<+ > c

.4 .

r

_3 <-{ s

N

¼-i-<5 (

.~~~~A..C....

....1.

/- t + .

.3<s.;^.@8ini K?)

I-.0,r..w

-t,

~i

.I.:batK

*

C ' 7

-A

aFIGS. 1-8. (Legend appears at bottom of the following page.)

0

i't-49t r4&.,.- *?.!;, 1

* ..

*

6

S

2782 Cell Biology: Novikoff

.f, W

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

Endoplasmic reticulum (Review) 2783

(22). This was confirmed by incubating sections for AcPaseactivity and then for TPPase activity. With electron microscopyit was determined that the AcPase-rich region was smooth ER.I named it GERL (23) because this region of smooth ER is lo-cated at the inner or "trans" (24) aspect of the Golgi apparatusand because it appears to produce various types of Lysosomes(25).The structural relation of GERL and Golgi apparatus has

been studied best in the neurons of dorsal root ganglia of the rat(5) and mouse (26). In these cells the relationship appears to bemore complex than in other cell types studied in our laboratory(6, 4, 27). Figs. 5-7 illustrate some aspects of this relationship;also see Fig. 40 in ref. 5. Smooth-surfaced tubules, probably partof GERL, extend into each polygonal compartment of the transelement of the Golgi apparatus. The polygonal arrangementprovides a large surface area where molecular interchangesbetween GERL and Golgi apparatus could occur. However, thefunctional significance of the spatial relations of GERL andGolgi apparatus remains to be elucidated. This is particularlyimportant for understanding events in the secretory cells con-sidered in the next section.

Four types of lysosomes seem to arise from GERL: (a) re-sidual bodies (Fig. 8) in which presumably indigestible residuesof intracellular digestion are visible; (b) coated vesicles whichmay carry lysosomal hydrolases to other cell structures (if sothey would be primary lysosomes); (c) autophagic vacuoles,type 1, in which portions of GERL envelop regions of cytoplasmcontaining organelles like mitochondria, ER, or peroxisomes.(When sealed off, lysosomal hydrolases are apparently releasedinto the vacuole and the sequestered structures undergo deg-radation to become a residual body.); and (d) autophagic vac-uoles, type 2, in which a region of GERL enlarges and portionsof its membrane are internalized, bringing in bits of cytoplasmlike cytosol or glycogen. Degradation by lysosomal enzymesleads to residual body formation, as in autophagic vacuoles, type1. For a fuller description see another review (25).An assumption made by cytochemists, apparently valid thus

far, is that when AcPase activity is found in a cell organelle otherlysosomal hydrolases are probably there as well. Unfortunately,only two cytochemical procedures valid at the ultrastructurallevel are currently available for other lysosomal enzymes, inmost cells, and even these are not as convenient as the AcPaseprocedure. These enzymes are arylsulfatase and an esterase thathydrolyzes thiolacetate. Decker (28) has demonstrated bothenzyme activities in GERL of neurons. Bentfeld and Bainton(29) have shown arylsulfatase activity in GERL of mega-karyocytes.

In addition to the demonstrable acid hydrolase activitiescommon to GERL and lysosomes, GERL often shows otherhallmarks of the residual body type of lysosome: a relativelythick delimiting membrane, an electron-lucent "halo" betweenmembrane and residual body contents, and the presence in the

contents of small ferritin-like grains, myelin-like figures, andother structures (25).GERL and secretion processing and packagingThe cytochemical visualization of TPPase and AcPase activitieshas shed new light on the packaging, and perhaps the finalprocessing, steps of secretory materials. With E. Essner, in theearly sixties, we had demonstrated the presence of AcPase ac-tivity in nascent secretory granules of both endocrine and ex-ocrine glands (20). Recently we have returned to studies en-gendered by these observations. In cells where TPPase activityis demonstrable in only one element of the Golgi apparatus itis the trans element, adjacent to GERL, that shows the activity.Thus, by incubating sections for TPPase activity and parallelones for AcPase activity it can readily be determined whethersecretory granules arise from GERL or from the trans elementof the Golgi apparatus. Thus far the findings are different fromwhat is generally assumed (30, 31).

In an insulin-producing transplantable hamster tumor weconclude (32) that most secretory granules arise from GERLand none from the Golgi apparatus, because the granules showAcPase activity and not TPPase activity (Fig. 9).

In the fi cells of the rat pancreas the granules also arise fromGERL and not the Golgi apparatus as widely thought. Againthe nascent granules show AcPase activity (Fig. 11) and noTPPase activity. Unlike the insulinoma, however, the moremature secretory granules do not have demonstrable AcPaseactivity.We (32) have indicated possible advantages of the trans-

plantable insulinoma for biochemical studies required to testthe assumption that when acid phosphatase activity is presentin granules it is likely that other lysosomal enzymes are thereas well, and to resolve the issues considered by Steiner et al. (31)in their discussion of the possible involvement of lysosomes inconverting proinsulin to insulin. Unfortunately there is cur-rently no valid cytochemical procedure generally accepted forvisualizing, at the electron microscope level, the sites of theproteolytic enzymes involved in converting proinsulin to in-sulin.The presence of AcPase only in the nascent secretory granules

is also seen in the exocrine cells of the pancreas. Fig. 10 showsthe absence of TTPase activity and Fig. 11, the presence ofAcPase in the nascent secretory granules in the pancreas of theuntreated guinea pig. The same is true in the fasted and refedguinea pig pancreas and in untreated hamster, rabbit, and rat(33). The nascent granules are the "condensing vacuoles"considered by Palade and coworkers to be sites of processingand packaging of secretory granules; for a review see Palade(30). We view the condensing vacuole portions as expandedcisternal portions of GERL; tubular ER elements attached tothem often take the form of "rigid lamellae" described byClaude (34) (arrows in Figs. 10 and 12).

FIGS. 1-8 (on preceding page). Fig. 1. Section of rat liver incubated for NDPase activity, with inosine diphosphate as substrate, from Novikoffand Heus (1). Reaction product is seen in the endoplasmic reticulum (ER) and nuclear envelope (NM). The mitochondria (M) are unreactive;the nucleolus is also labeled (Nu). X23,000. Fig. 2. Section of retina of C57 black mouse incubated for AcPase activity. The melanolysosomesof the pigment epithelial cell are ringed with reaction product, particularly evident at the arrows. X22,000. Fig. 3. Section ofmouse Harding-Passey melanoma incubated for tyrosinase activity, from Novikoff et al. (6). The Golgi apparatus (G) and rough endoplasmic reticulum (RER)are unreactive. Reaction product is present in GERL (GE) and in premelanosomes and melanosomes (ME). X22,000. Fig. 4. Section of ratthyroid incubated for peroxidase activity, from Novikoff et al. (4). Reaction product is present in the endoplasmic reticulum (ER) and "A granules"(A), but is absent from the "B granules" and residual bodies (L). A portion of the nucleus is seen at N. X16,000. Figs. 5-7. Serial sections ofa rat dorsal root ganglion incubated for TPPase activity, from P. M. Novikoff et al. (5). Reaction product is restricted to the trans element ofthe Golgi apparatus. GERL (GE) is unreactive. Long arrows indicate tubular structures, probably part of GERL. The short arrows indicatea small gap in the Golgi apparatus in Figs. 6 and 7 but this is closed over in Fig. 5, demonstrating that even at the electron microscope level theGolgi apparatus is a continuous organelle throughout the cytoplasm. X32,000. Fig. 8. A section, 0.5 ,um thick, ofa rat dorsal rat ganglion incubatedfor acid phosphatase activity, from P. M. Novikoff et al. (5). Reaction product is seen in GERL which shows a cisternal portion (C), tubularelements (T), and budding residual bodies (arrows). X40,000.

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

2784 Cell Biology: Novikoff

C'.>AWASir-.G~~~~~~~~~~~~~

W A.: ';A!'4, . T

ss.NMS .A;> . . ; ¢ ;

Proc. Natl. Acad. Sci. USA 73 (1976)

T V ,,' ' ,',,,V.4.

-ZVFtC~~~'~

z k VAk

L L

FIGS. 9-13. Fig. 9. Section of a transplantable hamster insulinoma, incubated for TPPase activity, from Novikoff et al. (32). Only the in-nermost element of the Golgi apparatus (G) shows reaction product. Most of the crescent-shaped area enclosed by the Golgi apparatus is occupiedby GERL. It consists of tubular elements (arrowheads), with nascent secretory granules (S) attached to some, and coated vesicles (C), somestill attached to tubular elements and others apparently free in the cytosol. Also labeled are two separated secretory granules (GR); note thatthey are unreactive. X36,000. Fig. 10. Section of guinea pig pancreas, incubated for TPPase activity, showing part of an exocrine cell. Reactionproduct is present only in the inner element of the Golgi apparatus (G). Condensing vacuoles are indicated by V and "rigid lamellae" by arrows;both are unreactive. X26,000. Fig. 11. Section of rat pancreas, incubated for AcPase activity, showing part of a j3 cell, from Novikoff et al.(32). The Golgi apparatus (G) is unreactive. Reaction product is seen in GERL (arrows) and nascent secretory granules (GR). Mature unreactivesecretory granules are not included in this field. X37,000. Fig. 12. Section ofguinea pig pancreas, incubated for AcPase activity, showing partof an exocrine cell. The Golgi apparatus (G) is unreactive. Reaction product is seen in three condensing vacuoles (V), in "rigid lamellae" (arrows),and in a residual body (RB). X30,000. Fig. 13. Section of liver from a male hamster fed a high-cholesterol diet for 2 days, incubated for AcPaseactivity, from Nehemiah and Novikoff (27). The Golgi apparatus (G) is unreactive. Much reaction product is seen in GERL (GE) and lipolysosomes(LL), within which the lipid site is electron-lucent. Portions of GERL containing very low density lipoprotein-like particles show slight depositsof reaction product (arrows). X29,000.

Strikingly similar results have been reported and conclusionsdrawn in a preliminary report by Hand and Oliver (35) for theexorbital lacrimal gland. Earlier AcPase results on epineph-rine-secreting cells of adrenal medulla were reported byHoltzman and Dominitz (36).

Interesting questions raised by these findings are: (a) Do theelements of the Golgi apparatus contribute to the secretorymaterial in GERL? (b) If so, what is the nature of this contri-bution and by what mechanisms does transfer from Golgi toGERL occur? (c) What roles do the coated vesicles derived fromGERL play?

Despite such questions, it seems most likely to us that somesecretion processing and the secretion packaging in a varietyof secretory cells will prove to be a function of GERL, implying

direct access from ER to secretory granules for at least somesecretory components. The functions of the Golgi apparatus inthese particular instances remain to be elucidated. Again weare limited by the exceedingly few cytochemical proceduresavailable Thus, it would be highly desirable to demonstratecytochemically the glycosyl transferases known to be concen-trated in isolated subcellular fractions enriched in portions ofthe Golgi apparatus, but this is presently not possible.GERL, lysosomes, and lipid transformations inhepatocytesWe have studied two model situations which suggest an in-volvement of GERL and lysosomes [identified cytochemically(AcPase) as well as by ultrastructural hallmarks] in lipid

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

Endoplasmic reticulum (Review) 2785

transformations within hepatocytes. However, we will beginby referring to the studies of Essner and Oliver (37) on hepa-tocytes of the beige mouse, regarded as an analogue to thehuman Chediak-Higashi disease. They wrote, "In addition tomarked enlargement and increased complexity, GERL (andlysosomes) of beige hepatocytes was characterized by accu-mulations of lipid-like inclusions, dense grains, and membra-nous material...." Essner and Oliver suggest that the patho-genesis of Chediak-Higashi disease may involve a defect in lipidmetabolism resulting in the accumulation of lipid within thegreatly enlarged GERL and the huge lysosomes (residualbodies) that arise from it.

J. Nehemiah and 1(27) have studied the hepatocytes of theSyrian golden hamster, untreated and fed a high cholesteroldiet. Unlike any other mammalian hepatocytes we know of, inthe hamster hepatocytes lipid accumulates within lysosomes.The lipid is stainable by Oil Red 0 and is birefringent; inelectron micrographs its site is electron-lucent. Such lipid ispresent in the periportal hepatocytes of untreated male ham-sters. On the high cholesterol diet all hepatocytes, in either maleor female hamsters, show enlargement of GERL and accu-mulation of cholesterol-rich lipid in the residual bodies arisingfrom GERL (also in adjacent areas of ER); cholesterol levels inthe liver rise dramatically. Fig. 13 shows part of the Golgi zonein a hepatocyte of a male hamster fed the cholesterol-rich dietfor 2 days. AcPase reaction product can be seen in GERL,portions of which contain lipid-like (very low density lipopro-tein?) particles, and in lysosomes (residual bodies) filled withlarge electron-lucent lipid. We have termed such lipid-filledlysosomes lipolysosomes. Recently, they have been describedin human hepatocytes of patients with Wilson's disease (38).judging from ultrastructure and from the early stages in theirdevelopment as described by Wake (39), the vitamin-A-con-taining "lipid droplets" in the "lipocytes" of rat liver are alsolipolysosomes.The studies of P. M. Novikoff et al. (40) also suggest a role

of GERL in lipid alteration within hepatocytes. A fatty liveris induced in rats by feeding a semisynthetic diet rich in oroticacid. The hypolipidemic drug clofibrate (ethyl chlorophe-noxyisobutyrate) produces a mild fatty liver on this semisyn-thetic diet. If the two agents are fed together a fatty liver failsto develop. In hepatocytes of the untreated rat GERL is notparticularly conspicuous, but in rats fed clofibrate with theorotate GERL is very much enlarged and swollen with materialinterpreted as lipid undergoing transformation. Stuffing ofGERL with lipid-like particles is even more prominent whena fatty liver is first induced in the rat by orotate and then clo-fibrate is added to the diet, causing the fatty liver to disappearand hepatocyte ultrastructure to revert to normal (P. M. No-vikoff and D. Edelstein, unpublished work). In the reversalstudies GERL was identified not only by characteristic ultra-structural features but also by AcPase cytochemistry (P. M.Novikoff, personal communication).

Biochemical, autoradiographic, and morphological evidence(summarized in ref. 40) has established the ER as the organellethrough which lipid is transported by the hepatocyte. In GERL,and in lysosomes derived from it, the lipid acquires a differentappearance. The biochemical events associated with thetransformation cannot be revealed by cytochemistry. Isolationprocedures are required.

GERL in melanocytesIn 1968 we reported that in the B-16 and Harding-Passey mousemelanomas GERL possesses tyrosinase ("dopaoxidase") activityas well as AcPase activity (6). Both enzyme activities are also

demonstrable in the premelanosomes and melanosomes (Fig.3). These findings have been confirmed, in general, in otherlaboratories for other melanomas and for normal melanocytes(4147). Since premelanosomes and melanosomes show AcPase(Fig. 2) as well as the typical tripartite delimiting membraneof residual bodies, P. Leuenberger and I (48) proposed the termmelanolysosome. Maul (41) and Maul and Romsdahl (42)suggest that AcPase may be involved in degrading enzymesafter melanization is completed or in degrading old melaningranules but they present no evidence for the proposals.

Microperoxisomes in mammalian cellsThe cytochemical 3,3'-diaminobenzidine procedure of Grahamand Karnovsky (49) has contributed to many areas of cellbiology. In our laboratory we modified the procedure to stainperoxisomes and other organelles (3, 7, 10). The alkaline 3,3'-diaminobenzidine procedure visualizing sites of peroxisomecatalase (7, 8) served to bridge morphology and biochemistryin studying peroxisomes (50). Unlike the nucleoid-containingperoxisomes found among mammalian cells only in liver andkidney, anucleoid peroxisomes, or microperoxisomes, were firstrevealed by the diaminobenzidine procedure (Figs. 14-16).Microperoxisomes are present in all cells save for matureerythrocytes and perhaps other specialized end-stage cells. Theydiffer from the nucleoid-containing peroxisomes in two otherrespects. They are generally smaller and they possess numerousslender connections to the ER. So frequent are the continuitiesof their delimiting membranes with the ER membrane thatmicroperoxisomes may well be considered as specialized regionsof ER in which catalase and other peroxisomal enzymes arecompartmentalized. The roles of these enzymes in animal cells,unlike those in plant cells (51), are still to be clarified. Theyappear in some manner to be involved in the metabolism,transport, and storage of lipid.

ER in thyroid epithelium cellsIf our interpretation (4) is correct there are two secretorypathways that by-pass the Golgi apparatus in the thyroid epi-thelial cells. One is like that in pancreas exocrine or endocrinecells considered above. So-called "B" granules (Fig. 4) areconsidered to arise from GERL and to carry noniodinatedthyroglobulin to the follicular lumen where it is released byexocytosis. The second pathway involves the ER generallywhich in this cell type has endogenous peroxidase activity.Peroxidase-positive "A" granules (Fig. 4), smaller than theperoxidase-negative "B" granules, appear to bud directly fromthe apical ER and to bring the peroxidase to the lumen whereit is secreted via exocytosis.

ConclusionWe have briefly reviewed how cytochemistry, by its demon-stration of specific enzyme localizations in situ, can help solveproblems in cell biology. We have emphasized a current focusof our laboratory, namely, the complexity of the ER. In additionto its well-known roles in protein synthesis, the ER appears tohave regions with specialized functions. For example, in vir-tually all mammalian cells, microperoxisomes attached to theER are sites where catalase and probably other peroxisomalenzymes are localized. In diverse cell types-including neurons,exocrine and endocrine cells, hepatocytes, and melanocytes-the region of ER near the Golgi apparatus known as GERLpossesses AcPase and morphological features in common withlysosomes (residual bodies) that appear to arise from it. Weinterpret our observations as revealing different transformations

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

Proc. Natl. Acad. Sci. USA 73 (1976)

ERr

,I, ..>>H

*,- ~ ~(s..14).., i.. a 9

z~~~~~~~~... F .

)~~~~~~~~~~~~~~~~~~~~st'P~~~~~~~s.<.....-Ezz~~

Q.)

ne:GN _6,

I< j9. -

::, itN

.'SJo_ t ., ..e X. t98 N,£t-

B

o' '' 5

t o'er

- R'2a^* * ;,b, . ,+ 0 {<Sb0R A-:o "

2

%q

v

FIGS. 14-16. Fig. 14. Section of rat thyroid gland incubated in the alkaline 3,3'-diaminobenzidine for visualizing catalase sites (8), fromNovikoff et al. (4). Only the microperoxisomes show reaction product. The arrows indicate two regions where microperoxisome membrane andmembrane of the endoplasmic reticulum (ER) are probably continuous. X35,000. Fig. 15. Section of guinea pig pancreas incubated as in Fig.14, from Novikoff et al. (52). Three reactive microperoxisomes are seen on two zymogen granule surfaces (Z). Note the proximity of endoplasmicreticulum (ER) to the upper left microperoxisome. X43,000. Fig. 16. Section of guinea pig ileum incubated as in Fig. 14, from Novikoff andNovikoff (53). Reactive microperoxisomes are numbered from 1 to 4. Ribosomes (R) are seen on some areas of ER. Arrowheads indicate continuitiesbetween microperoxisome membranes and membrane of the ER. X76,000.

within GERL-producing autophagic vacuoles and coatedvesicles as well as residual bodies, packaging secretory materials,changing the nature of lipid in transit within the ER, andforming premelanosomes. Yet another type of secretorymechanism is seen in thyroid epithelium cells, where some se-cretory granules appear to bud from the apical ER. As in themechanisms involving GERL, the Golgi apparatus would ap-pear to be bypassed. The functional relations of the Golgi ap-paratus to GERL remain to be elucidated in all cell types wehave studied.

I gratefully acknowledge the preparation of the final photographsby Mr. George Dominguez and the typing of successive versions byMs. Fay Grad. This work was supported by U.S. Public Health ServiceResearch Grant CA06576 and the John Polachek Foundation forMedical Research. The author is the recipient of the U.S. Public HealthService Research Career Award 5K6 CA14923 from the NationalCancer Institute. Needless to say, many colleagues have contributedto the work and ideas expressed in this review, but any failings are minealone.

1. Novikoff, A. B. & Heus, M. (1963) J. Biol. Chem. 238, 710-716.

2. Novikoff, A. B., Essner, E., Goldfischer, S. & Heus, M. (1962) inThe Interpretation of Ultrastructure, ed. Harris, R.J.C. (Aca-demic Press, New York), Vol. 1, pp. 149-192.

3. Novikoff, A. B., Beard, M. E., Albala, A., Sheid, B., Quintana, N.& Biempica, L. (1971) J. Microsc. (Paris) 12,381-404.

4. Novikoff, A. B., Novikoff, P. M., Ma, M., Shin, W.-Y. & Quintana,N. (1974) in Advances in Cytopharmacology, eds. Ceccarelli,B., Clementi, F. & Meldolesi, J. (Raven Press, New York), Vol.2, pp. 349-368.

5. Novikoff, P. M., Novikoff, A. B., Quintana, N. & Hauw, J.-J.(1971) J. Cell Biol. 50, 859-886.

6. Novikoff, A. B., Albala, A. & Biempica, L. (1968) J. Histochem.Cytochem. 16, 299-319.

7. Novikoff, A. B. & Goldfischer, S. (1969) J. Histochem. Cytochem.17,675-680.

8. Novikoff, A. B., Novikoff, P. M., Davis, C. & Quintana, N. (1972)J. Histochem. Cytochem. 20,1006-1023.

9. Novikoff, A. B. & Goldfischer, S. (1961) Proc. Nati. Acad. Sci.USA 47,802-810.

10. Novikoff, A. B. (1974) in The Proceedings of the Second Inter-national Symposium on Electron Microscopy and Cytochem-istry, eds. Daems, W.Th., Molenaar, I. & Van Duijn, P. (North-Holland Publishing Co., Amsterdam), pp. 89-109.

11. Novikoff, A. B., Shin, W.-Y. & Drucker, J. (1961) J. Biophys.Biochem. Cytol. 9,47-61.

12. Rosenthal, A. S., Moses, H. L., Beaver, D. L. & Schuffman, S. S.(1966) J. Histochem. Cytochem. 14,698-701.

13. Moses, H. L., Rosenthal, A. S., Beaver, D. L. & Schuffman, S. S.(1966) J. Histochem. Cytochem. 14,702-710.

14. Ganote, C. E., Rosenthal, A. S., Moses, H. L. & Tice, L. W. (1969)J. Histochem. Cytochem. 17,641-650.

15. Rosenthal, A. S., Moses, H. L. & Ganote, C. E. (1970) J. Histo-chem. Cytochem. 18,915.

16. Novikoff, A. B. (1967) J. Histochem. Cytochem. 15,353-354.17. Novikoff, A. B. (1970) J. Histochem. Cytochem. 18,366.18. Novikoff, A. B. (1970) J. Histochem. Cytochem. 18,916.19. de Duve, C. (1969) in Lysosomes in Biology and Pathology, eds.

Dingle, J. T. & Fell, H. B. (North-Holland Publishing Co., Am-sterdam), Vol. 1, pp. 3-40.

20. Novikoff, A. B. & Essner, E. (1962) Fed. Proc. 21, 1130-1142.21. Novikoff, A. B. (1963) in Ciba Foundation Symposium on Ly-

sosomes, eds. deReuck, A. V. S. & Cameron, M. P. (J. & A.Churchill, Ltd.-Little, Brown, & Co., Boston), pp. 36-73.

22. Novikoff, A. B. (1967) in The Neuron, ed. Hyden, H. (ElsevierPublishing Co., Amsterdam), chap. 6, pp. 255-318.

23. Novikoff, A. B. (1964) Biol. Bull. 127,358.24. Ehrenreich, J. H., Bergeron, J. J. M., Siekevitz, P. & Palade, G.

(1973) J. Cell Biol. 59, 45-72.25. Novikoff, A. B. (1973) in Lysosomes and Storage Diseases, eds.

Hers, G. & van Hoof, F. (Academic Press, New York), pp.1-41.

26. Boutry, J.-M. & Novikoff, A. B. (1975) Proc. Natl. Acad. Sci. USA72,508-512.

27. Nehemiah, J. L. & Novikoff, A. B. (1974) Exp. Mol. Pathol. 21,398-423.

2786 Cell Biology: Novikoff

.4'r4

t.

4

11,41 Ail k'"W, k, I-

A (

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021

Proc. Nati. Acad. Sci. USA 73 (1976)

28. Decker, R. S. (1974) J. Cell Biol. 61,599-612.29. Bentfeld, M. E. & Bainton, D. F. (1975) J. Clin. Invest. 56,

1635-1649.30. Palade, G. E. (1975) Science 189,347-358.31. Steiner, D. F., Kemmler, W., Clark, J. L., Oyer, P. E. & Rober-

stein, A. H. (1972) in Handbook of Physiology-Endocrine Pan-creas, eds. Steiner, D. F. & Freinkel, N. (Am. Physiol. Soc.,Washington, D.C.), Vol. I, pp. 175-198.

32. Novikoff, A. B., Yam, A. & Novikoff, P. M. (1975) Proc. Natl.Acad. Sci. USA 72,4501-4505.

33. Novikoff, A. B., Mori, M., Quintana, N. & Yam, A. (1976) J.Histochem. Cytochem. 24,612-613.

34. Claude, A. (1970) J. Cell Biol. 47, 745-766.35. Hand, A. & Oliver, C. (1975) J. Cell Blol. 68, 154a.36. Holtzman, E. & Dominitz, R. (1968) J. Histochem. Cytochem.

16,320-336.37. Essner, E. & Oliver, C. (1974) Lab. Invest. 30,596-607.38. Hayashi, H. & Sternlieb, I. (1975) Lab. Invest. 33, 1-7.39. Wake, K. (1974) J. Cell Biol. 63,683-691.40. Novikoff, P. M., Roheim, P. S., Novikoff, A. B. & Edelstein, D.

(1974) Lab. Invest. 30, 732-750.41. Maul, G. G. (1969) J. Ultrastruct. Res. 26, 163-176.

Endoplasmic reticulum (Review) 2787

42. Maul, G. G. & Romsdahl, M. M. (1970) Cancer Res. 30,2782-2790.

43. Maul, G. G. & Brumbaugh, J. A. (1971) J. Cell Biol. 48, 41-48.

44. Brumbaugh, J. A. (1971) Dev. Biol. 24,392-412.45. Brumbaugh, J. A. & Zeig, R. H. (1972) in Pigmentation: Its

Genesis and Biologic Control, ed. Riley, V. (Appleton-Cen-tury-Crofts, Meredith Corp., New York), pp. 107-123.

46. Ide, C. (1972) Z. Zellforsch. 131, 171-186.47. Eppig, J. J. & Dumont, J. N. (1972) J. Ultrastruct. Res. 39,

397-410.48. Leuenberger, P. M. & Novikoff, A. B. (1975) J. Cell Biol. 65,

324-334.49. Graham, R. C., Jr. & Karnovsky, M. J. (1966) J. Histochem. Cy-

tochem. 14, 291-302.50. Novikoff, A. B. & Novikoff, P. M. (1973) J. Histochem. Cyto-

chem. 21, 963-966.51. Vigil, E. (1973) Subcell. Biochem. 2,237-285.52. Novikoff, A. B., Novikoff, P. M., Davis, C. & Quintana, N. (1973)

J. Histochem. Cytochem. 21, 737-755.53. Novikoff, P. M. & Novikoff, A. B. (1972) J. Cell Biol. 53,532-

560.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

021