JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 17:2-14 (1991i
Perspectives on Golgi Apparatus Form and Function HILTON H. MOLLENHAUER AND D. JAMES MORRB pootl Animal Protection Research Laboratory, United States Department of Agriculture, Agricultural Research Service, College Station. Texas. and Department of Pathology and Laboratory Medicine. College of Medicine, Texas A & M Unzuersity, College Station, Tmczs 77840 (H.H.M.);Department of Medicinal Chemistry arid Pharmacognosy, Purdue University, West L a f a y t t e , Indiana 47907 ( D . J . M . )
KEY WORDS
Dictyosome, Endomembrane system, Terminology
In 1898, Camillio Golgi reported a new cellular constituent with the form of an ABSTRACT extensive intracellular network (the apparato reticolare interno),which now bears his name. However, the history of Golgi’s apparatus is replete with controversy regarding its reality, what components of the cell should be included under its aegis, and what terminology should be used when referring to it. Electron microscopy has resolved many of these controversies and it is appropriate that this volume emphasize that aspect of Golgi apparatus discovery. The principal structural component of the Golgi apparatus is the stack of cisternae, or dictyosome. As determined both biochemically and a t the level of electron microscopy, the dictyosome is a highly ordered and polarized structure. The maintenance c;f order within the stack is thought to result from either intercisternal bonding constituents, or filamentous structures (or both) that bridge the space between adjacent cisternae. Mechanisms proposed for movement of membrane and product into and out of the dictyosome (i.e.,the Golgi apparatus stack) include a serial mode which functions exclusively by the formation, displacement, and loss of cisternae from the stack, and a parallel mode which functions exclusively by the movement of membrane, product, or precursor molecules directly into the peripheral edges of the cisternae. In the parallel mode, all cisternae can be accessed either singly or simultaneously, a t least in theory, a t any position within the stack. It is probable that both the serial and the parallel modes function concomitantly and need not be mutually exclusive. Finally, the peripheral tubules of the cisternae represent a major membranous constituent of the cell with potentially unique functions. These tubules interconnect cisternae of adjacent stacks and may represent the major site of receptors for the shuttle (i.e., parallel) type of transfer among cisternae. Peripheral tubules as extensions of the cisternal lumina into the cytoplasm presumably have other functions, but these, like the tubules themselves, have only rarely been accommodated into functional models of Golgi apparatus dynamics in secretion or membrane flow.
INTRODUCTION descriptions (see Whaley, 1975, and references thereThe structure now known a s the Golgi apparatus ap- in). The basis of the controversy consisted of two issues: parently was observed by La Valette St. George and by 1) the impregnation methods were somewhat capriothers as early as 1865 (e.g., Platner, 1889 and Murry, cious and 2 ) the form of the structural entity that was 1898 as cited by Whaley, 1975). It was not popularly contrasted varied markedly among cell types and in recognized a s a potentially unique cellular entity until relation t o cell activity. The latter observations were the silver impregnation methods developed by Camillio contrary to what was expected for a universal cell orGolgi in 1873 were applied, first to nerve cells of the ganelle and it was not until many years later that Golbarn owl (Golgi, 1898) and then to other types of cells gi’s reticulum was recognized as a n extremely pleomor(see Bowen, 1926; Beams and Kessel, 1968; Whaley, phic cell component subject to changes in location 1975, for review). The effectiveness of Golgi’s procedure within the cell, intensity of activity, and functional was to stain various elements of the nervous system characteristics. This variability was (and still is) a mabrown or black and contrast this vividly against a n jor factor in establishing a realistic image of the strucalmost transparent background. With this impregna- ture of the organelle and determining how the comtion technique, Golgi observed a n external reticulum plexities of Golgi apparatus structure relate to its (which had been noted before) and an internal reticu- function. The metal impregnations introduced by Golgi caused lum (which had not previously been convincingly demthe internal reticulum to appear as a continuum of onstrated). It was the internal reticulum (i.e., the upparato reticolare interno), or a part of it, that was recognized first as the Golgi apparatus. The interpretation of Golgi’s observations generated Received April 20, 1989; accepted in r w i s e d form N ~ ~ v e m b e15. r 19x9 much debate. and numerous controversies regarding Address reprint requests to Hilton I I . Mollenhauer. USDA-AKS, Rt. 5, Box the reality of the reticulum ensued following the initial 810. College Station. TX 77840.
c
1991 WILEY-LISS, INC
GOLGI APPARATUS FORM AND FUNCTION
fibers in vertebrate somatic cells and as isolated scales in embryonic and germ cells (Golgi, 1898; Whaley, 1975).Thus, the total cellular complement of Golgi material may appear with the light microscope a s either a number of individual units (which we now recognize as dictyosomes or stacks), or a s highly elaborate (and continuous) Golgi zones of many interconnected units, or as any of a number of intermediate configurations between these two extremes. The microscopic image was complicated by the variability of the staining procedures and lack of all parts of the Golgi material t o respond to the stains in the same manner. Such problems were encountered by many of the early cytologists. For example, Baker (see Whaley, 1975) who was, perhaps, the most vociferous disbeliever in the reality of Golgi’s reticulum, suggested that the reticulum, if it existed a t all, included many cellular constituents to which Golgi’s name should not be attached. Certainly, the reticular part of Golgi’s apparatus (i.e., the filamentous part that Golgi observed between the scalelike units, i.e., between the stacks) may have included some elements of the endoplasmic reticulum. The latter have been shown by electron microscopy to precipitate Golgi stains in the same locations as those observed by the early cytologists (Kopsch, 1902, as cited by Inferrera and Carrozza, 1975). Even today, while the existence of a Golgi apparatus is no longer questioned, its boundaries, particularly in respect to its peripheral appendages (i.e., tubules, vesicles, and associated trans Golgi apparatus reticuli), are difficult, and perhaps impossible, to define precisely. In the classic electron microscope image, the Golgi apparatus appears as a structure consisting of stacks of flattened saccules (cisternae) which give rise to secretory vesicles (Mollenhauer and Morre, 1966a). The secretory vesicles are either attached directly to the cisternae (via short tubules; see Mollenhauer and Morre, 1966a) or in a region contiguous with the trans pole of the stack (Palade, 1969; Farquhar, 1985). Tubularivesicular networks are associated with the trans pole of each stack (e.g., Golgi Endoplasmic Reticulum Lysosome system (GERL)-Novikoff, 1976; Trans Golgi Network (TGN)-Griffiths and Simons, 1986; Compartment of Uncoupling of Ligand and Receptor (CURL)-Geuze et al., 1983, 1985; M6P-lgp enriched structure-Griffiths et al., 1988), which are implicated in membrane sorting and recycling (see Geuze and Morre; Griffing; and Beaudoin and Grondin, this volume). Another tubularivesicular network (i.e., the Partially Coated Reticulum (PCR)) of similar form, but less intimately associated with Golgi apparatus stacks has been reported that may also have membrane sorting and recycling functions (Pesacreta and Lucas, 1984, 1985). The numbers of cisternae per stack vary widely, although, in most animal cells and higher plant cells, there are 5-10. Each stack is polarized in the sense that the maturation (processing) of products and membranes appears to occur sequentially from a cis (forming) face on one side to a trans (maturing) face on the opposite side (Mollenhauer and Whaley, 1963). Each cisterna of the stack is unique in that it differs chemically, structurally, and hnctionally from the oth-
3
ers (Morre et al., 1971; Mollenhauer and Morre, 1980; Rothman, 1985). For simplicity, the stack may be divided into different regions (e.g., cis, medial, and trans), each representing a functionally different subset of cisternae (Rothman, 1985; Pfeffer and Rothman, 1987). In reality, however, it seems more likely that changes in function occur gradually across the stack of cisternae rather than in discrete steps. The dictyosomes of the Golgi apparatus most often occur side-by-side in mammalian cells to form a complex ribbon adjacent to the nucleus. In higher plants, fungi, and in some animal cells (e.g., invertebrates), however, the stacks or dictyosomes usually are sufficiently separated to appear as discrete units within the cytoplasm. Yet, even with the dispersed arrangement, the dictyosomes appear to be interconnected into a functional Golgi apparatus. The number of dictyosomes per cell in eukaryotes ranges from none (in certain fungi that lack stacked Golgi apparatus cisternae-Bracker, 1967) t o over 25,000 in some algal rhizoids (Severs, 1965).A typical plant or animal cell may contain 500 or more dictyosomes. The unstacked cisternae or tubular networks of fungi that lack dictyosomes carry out some Golgi apparatus functions. These Golgi apparatus equivalents (Bracker, 1967) are easily identified in electron micrographs even without the characteristic stacked configuration. If, as in a few algal cells, only one dictyosome is present, then this dictyosome is equivalent to the Golgi apparatus. While opinions may differ, multiple dictyosomes within a given cell function synchronously and appear sufficiently interconnected or interassociated in terms of their regulation to be regarded in totality as the Golgi apparatus. A single exception has been reported in Urodele sperm development by Werner (1970), where more than one Golgi apparatus appears to occur in a single cell. Each cisterna (pl. cisternae) consists of a lumen or central cavity that is surrounded by a membrane. The central flattened, or plate-like, region is most often referred to as the cisterna or saccule. At the periphery of some saccules, there may be fenestrae (openings) of varying diameter. The fenestrated margins of the saccules are usually continuous with a system of tubules that may extend for several microns from the edges of the saccules. Some tubules presumably connect with endoplasmic reticulum. Others may connect with cisternae of adjacent dictyosomes or, possibly, lysosomes through vesicular bridges. Cisternal tubules have been demonstrated by a variety of techniques both with isolated Golgi apparatus and in situ (e.g., Figs. 1,4-6,10, 11;Tandler and Morre, 1983) in both plant and animal cells (Mollenhauer e t al., 1967). In addition to changing from characteristics resembling endoplasmic reticulum to characteristics resembling plasma membrane, the differentiation of Golgi apparatus cisternae within the stack also involves changes in the arrangements of saccules, tubules, and vesicles. Toward the center of the stack, the plate-like, or saccular, regions are a dominant cisternal characteristic. Yet even these cisternae have fenestrated borders and tubules which emanate from their peripheries
4
H.H. MOLLENHAUER AND D.J. MORRfi
Fig. 1. Isolated rat germ cell Golgi apparatus negatively stained with phosphotungstic acid. These Golgi apparatus show clearly the stacks of plate-like cisternae interspersed with tubular intersaccular regions. Note large number of cisternal tubules, and differences in number of fenestrae and tubules on 0.5 km. individual cisternae. Bar
(Morre and Ovtracht, 1981).In some cells a t least, the trans cisternae may have extensive tubular networks without f'enestrae. Post-Golgi apparatus structures, for example, the trans Golgi network (e.g., TGN; Griffiths and Simons, 1986), or partially coated reticulum (PCR; Pesacreta and Lucas, 1985),lie adjacent t o the trans poles of the stacks. In plant cells, these structures appear t o be derived from trans cisternae as they are released from the stack (Mollenhauer et al., in press). The functions of the TGN and PCR may not be restricted only to the
sorting of exocytic and endocytic products (Tartakoff, 1982, 1983; Roth et al., 1985; Anderson and Pathak, 19851, but may participate also in the discharge of membrane a t the trans Golgi apparatus poles as well. Although the details of stack morphology and function are now well defined, questions remain a s to how the stack is formed and maintained, how secretory products andlor membrane move through the stack, and what is the role of the peripheral tubules. Some of these questions are addressed in the sections that follow.
GOLGI APPARATUS FORM AND FUNCTION
5
Fig. 2. A: Rat epididimal principal cell illustrating the form and interassociation of several dictyosomes (i.e., stacks of cisternae ID1). This cell has several hundred dictyosomes dispersed through much of the cytoplasm and each dictyosome has about 8-12 cisternae. These animal dictyosomes are similar, but not identical, to those of plants (see Griffng, this issue; Mollenhauer et al., 1967, Mollenhauer and Morre, 1978). B: Ryegrass rootcap cell illustrating the form and dis-
tribution of dictyosomes (D) typical of plant cells. Each dictyosome has 5-10 cisternac arranged in a polarized configuration. Cis (forming) and trans (maturing) faces are labeled “c” and “t”,respectively, and secretory vesicles (arrowheads) are illustrated. This cell probably contains between 400 and 600 dictyosomes distributed in groups or clusters throughout much of the cytoplasm. Bars = 0.5 km.
INTERCISTERNAL SUBSTANCES AND MAINTENANCE OF THE GOLGI APPARATUS STACK
al., 1973). Endoplasmic reticulum entering the Golgi apparatus zones of exclusion usually lack ribosomes. Mitochondria are excluded, but peroxisomes and lysosomes may be present. Free polysomes (i.e., the so-called Golgi apparatus polyribosomes; Mollenhauer and Morre, 1974; Elder and Morre, 1976) are associated closely with, but are not attached to, the Golgi apparatus membranes within the Golgi apparatus zone. Here, too, are concentrated the clathrin-coated vesicles associated predominantly with the mature faces of the Golgi apparatus stacks (Croze et al., 1982).
The cisternal stacks observed by electron microscopy (Fig. 2A,B) consist of cisternae that are separated physically from one another, but yet maintained in a precise order. This organization has been postulated to occur via substances within a specialized region of cytoplasm (the zone of exclusion) that surrounds elements of the Golgi apparatus (Franke et al., 1972; Mollenhauer and Morre, 1972, 1978a; Mollenhauer et
6
H.H. MOLLENHAUER AND D.J. MORRfi
When monovalent salts were added to the homogenization medium prior to glutaraldehyde stabilization, unstacking occurred within seconds to release intact cisternae into the homogenate (Fig. 5; compare with the nearly intact dictyosome of Fig. 4). These free cisternae often had plaques (Fig. 6A,B) and remnants of intercisternal elements (filaments) on the flattened parts of the cisternae (Fig. 6B,C). With continued exposure to monovalent salts, both the filaments and plaques disappeared followed by dissolution of the central parts of the cisternae (Fig. 6C,D). Breakdown of the cisternae began a t sites previously occupied by the plaque substances of the intercisternal region (Fig. 6B,C). While a n obvious function of intercisternal substances would be to stabilize the stack and maintain its organization, it seems likely that they may have other functions as well. For example, they might help control the timely release of vesicles and trans cisternae at maturity (Mollenhauer et al., 1973) as well as aid in the capture of precisternal vesicles a t the cis face of the stack.
MOVEMENT OF MEMBRANE AND PRODUCT THROUGH THE STACK Movement of membrane and product through the Golgi apparatus is accompanied by sequential (i.e., serial) formation, differentiation, and loss of Golgi apparatus elements (Mollenhauer and Whaley, 1963). The formation of new cisternae on one face of a Golgi apparatus stack would be required to compensate for loss Cisternae are separated from one another within of cisternae from the opposite face of the stack. The each dictyosome by a space of 7-15 nm. Because the source of these new cisternae was suggested early on t o stacks of cisternae can be isolated intact and subse- be a special region of the endoplasmic reticulum which quently unstacked by enzymatic, chemical, or mechan- gives rise to transition vesicles that move to, and conical means (Mollenhauer et al., 1973; Morre et al., dense on, the forming (cis) faces of the Golgi apparatus 19831,the intercisternal regions are thought to contain stacks where they would fuse together to form the new bonding substances that contribute to stack integrity. cisternae (Zeigel and Dalton, 1962; Morre and MollenIn invertebrate and plant cells, there are filamentous hauer, 1976). Alternatively, movement of materials elements in parallel arrays that lie midway between may occur via shuttle vesicles a t the peripheries of the adjacent cisternae (Mollenhauer, 1965a; Turner and cisternae to effect transfer from one cisterna to the Whaley, 1965; Fig. 3 ) . The numbers of these filaments next (Orci e t al., 1986).However, direct (i.e., nonvesicincrease toward the trans poles of the stacks and then ular) movement of substances into Golgi apparatus cisabruptly decrease just before release of the most trans ternae through the peripheral tubules of the cisternae cisterna. Dictyosomes of plant cells show a consistent must also be considered as yet another route for the and marked narrowing of the cisternal lumina from parallel delivery and transfer of substances in and out the proximal to the distal pole in almost direct propor- of the Golgi apparatus (Morre and Mollenhauer, 1974; tion to the number of intercisternal elements. Animals Mollenhauer et al., 1975; Mollenhauer and Morre, and fungi have Golgi apparatus that both lack the in- 1976; Harris and Oparka, 1983). Studies in plants have concentrated on the release of tercisternal elements and that fail to reveal the marked gradient in cisternal narrowing (Mollenhauer mature cisternae and contiguous secretory vesicles and Morre, 1978b). Although the function of intercis- from the trans faces of the stacks (Fig. 7) coupled with ternal elements is not known, there is circumstantial the formation of new, replacement cisternae on the cis evidence to suggest that they are associated with shap- faces of the stacks. Here, membrane and product are ing of secretory vesicles and/or cisternae (Mollenhauer considered to move serially through the stack as cisternae mature (Mollenhauer, 1971; Fig. 8).This mode and Morre, 1975; Kristen, 1978). One approach to relate the intercisternal region to of functioning is unequivocal since entire cisternae are organization of the stack has been to solubilize the in- separated from the stack (Brown, 1969; Mollenhauer, tercellular substances of isolated plant dictyosomes 1971; Mollenhauer et al., in press). Such separation and then stabilize what remains of the disintegrating must be accompanied by formation of new cisternae to stacks with glutaraldehyde (Mollenhauer et al., 1973). maintain stack integrity. This is, apparently, accomThe isolated stack constituents were negatively plished by a transport mechanism that shuttles vesistained with phosphotungstic acid and examined by cles from endoplasmic reticulum to the cis faces of the electron microscopy. stacks. The approximtely 60 nm vesicles that bleb from Fig. 3 . A maize root dictyosome with intercisternal filamentous elements viewed end-on (arrows,. The number of filaments increases toward the trans pole of the dictyosome (toward bottom of micrograph, but then decreasesjust before t h e most trans cisterna ( T Jis sloughed. Bar = 0.1 urn.
GOLGI APPARATUS FORM AND FUNCTION
Fig. 4. Plant dictyosome homogenized in medium containing glutaraldehyde and then isolated. The stacked configuration and much of’ the peripheral structure was retained. The dictyosome illustrated here has probably lost its most cis cisterna. Note that the central (flattened)parts of the cisternae are partially masked by tubules from overlaying cisternae (either cis or trans). Tubules of the type shown are also apparent in micrographs of fixed and embedded dictyosomes (see Figs. 10, 11). Coated vesicleibuds (arrowheads) were always present although they disappeared rapidly if exposed to PTA or monovalent salts before fixation. Bar = 0.1 p,m (ca).
7
Fig. 5. Plant cisternae isolated after treatment with 0.5 M monovalent salt (i.e., sodium chloride) for 90 seconds. Divalent salts (e.g., calcium chloride) tend to stabilize the stack. The smooth cisterna with fenestrae is probably from the center of the stack. The mass of tubules on the right appears to be an aggregate of two cisternae from either the cis or trans pole of a dictyosome. Bar = 0.1 km.
8
H.H. MOLLENHAUER AND D.J. MORRE
Fig. 6. This series of micrographs illustrates several stages in dictyosome disassembly by monovalent salts. Disassembly occurs very rapidly, first by cleavage of cisternae and then by dissolution of t h e central parts of the cleaved cisternae. Total time for dissassembly using 0.5 M monovalent salts a s t h e solubilization agent is usually less than 2 min. A: Plaques such a s this (arrow) were usually present on freshly cleaved cisternae. The peripheral tubules often appeared to be more turgid and stained differently t h a n the cisternae to which they were attached. B: Plaques and isolated bundles of intercisternal
filaments (arrowhead) were often visible a t t h e early stages ofdisassembly. The filaments and extensive tubular array identify this as a trans cisterna. C: When the plaques disappeared, the underlying membranes of the cisternae showed signs of breakdown (arrow). Filaments, although present in this micrograph, were usually absent a t this stage of disassembly. D: Eventually, t h e central parts of t h e cisternae were dissolved leaving only a ring and peripheral tubules. (Some of t h e micrographs a r e reproduced from Mollenhauer e t al., 1973.)Bars 0.1 wm.
nuclear envelope o r endoplasmic reticulum and form new cisternae are usually spherical and of uniform diameter. They appear to be coated by a nap-like covering (Mollenhauer e t al., 1976) distinct from the spiny coat of clathrin-coated vesicles (Croze et al., 1982) and from the coats described for budding profiles associated by Orci and collaborators (Orci et al., 1986) with intercisternal Golgi apparatus transport. The process is
completed when the most trans cisternae are released from the stacks. Similarly, in scale-forming algae (see Brown, 1969, and Melkonian e t al., this volume), whole cisternae containing newly synthesized scales are sloughed from the stack and secreted out of the cell in times as short a s one minute or less (Brown, 1969; Melkonian e t al., this volume). In some species, these scale secretion
~
GOLGI APPARATUS FORM AND FUNCTION
9
Fig. 7. Maize rootcap cell illustrating dictyosomes (D) and trans cisternae that have separated from the stack (arrows).Note the vesicular buds on one of the separated cisternae (arrowhead).Cell wall (W), vacuole (V), secretory vesicle (SV), starch in amyloplast (S). Bar = 0.5 km (ca.).
events can be witnessed in real time since the sizes of the scales are well within the range of resolution of light microscopes. Here one must conclude that both membrane and product move serially through the stack. In animals, studies of membrane and product movements through the stack have concentrated on vesicular shuttle mechanisms contiguous with the peripheral edges of the cisternae. Thus, Xothman and coworkers (Rothman et al., 1984) have suggested intercompartmental transport in the Golgi to be a dissociative (vesicular) process. Vesicular buds on the peripheral edges of cisternae (Orci et al., 1986) are thought to act as shuttles to move materials from one cisterna to another. These transfer vesicles would have access t o all cisternae simultaneously in parallel array (Fig. 8). Are the two models of membraneiproduct movement (i.e., serial differentiation of cisternae through the stack vs parallel differentiation by vesicle shuttle or direct endoplasmic reticulum transfer) mutually exclusive? Probably not. Vesicle buds of the type described by Orci and coworkers (1986) occur widely in both plants and animals (Cunningham et al., 1966; Mollenhauer and Morre, 1966a; Mollenhauer et al., 1973, in press; Fig. 3). Similarly, coated vesicles and post-Golgi reticuli are universally present a s well (Pavelka and Ellinger, 1983; Pesacreta and Lucas, 1984, 1985; Farquhar, 1985; Griffiths and Simons, 1986; Mollenhauer e t al., 1989, in press; Hubner et al., 1985; Griffing, 1988; Hillmer et al., 1988; Jones and Robinson, 1989). Thus, the structural bases for the parallel mode are present even in cells where the serial mode dominates. Similarly, cisternal maturation may also occur in animal cells (Morre et al., 1979) where the parallel mode is thought to dominate. What may differ from one cell type to another, however, is the degree to which each mode participates in membrane traffic within the stack. Furthermore, neither the serial or parallel mode
of functioning need dominate to the exclusion of the other. Both modes are probably capable of operating either separately or concomitantly to some extent in all species. An example of control of the mode of operation of the Golgi apparatus in plant cells is seen in a region of secretory vesicle transition in maize rootcap cells (Mollenhauer, 196513).In these cells, hybrid secretory vesicles form in a transition zone of the rootcap where cells from within the cap (which produce a spherical, epidermal-like, secretory vesicle) transform into cells of the outer rootcap (which produce a n elongate secretory vesicle; Fig. 9A-E). In the transition state, secretory vesicles are formed that contain both types of secretory product (Mollenhauer, 1965b; Fig. 9D,E). These hybrid vesicles form on the most trans cisterna of the stack as well as on cisternae within the stack. The hybrid vesicles clearly show that Golgi apparatus activities are developmentally regulated and that more than one pattern of activity can be expressed even simultaneously by the same basic Golgi apparatus stack. The data might support as well the presence of both serial and parallel forms of input into the Golgi apparatus stacks for the production of secretory vesicles as well as for cisternal maturation. In overall perspective, both serial and parallel routes through the stack have been proposed. Which route may predominate is unknown. However, the parallel route includes both vesicular and direct transfer of materials among all contiguous compartments within the pathway, whereas the serial path requires only the formation of new cisternae and displacement of mature cisternae.
GOLGI APPARATUS SHUTTLE VESICLES OF INTERCOMPARTMENT TRANSFER The formation, separation, and movement of vesicles that shuttle between cisternae has been proposed and
H.H. MOLLENHAUER AND D.J. MORRQ
10
is, potentially, a very important aspect of Golgi apparatus functioning. Evidence for the existence of a shuttle vesicle to transfer materials from one Golgi apparatus compartment to the next comes primarily from studies of Rothman and collaborators using the processing of vesicular stomatitus virus (VSV) glycoproteins as a model (Wattenberg, this volume). They employed a mutant BHK cell that lacked a processing enzyme UDP-N-acetylglucosamine N-acetylglucosamine:glycoprotein transferase unable to process VSV-G. When homogenates of infected cells containing radiolabeled viral precursors were mixed with homogenates of wild-type cells, processing of preformed VSV-G continued, suggestive of transfer from Golgi apparatus cisternae lacking the processing enzyme to Golgi apparatus cisternae containing the processing enzyme. Like the endoplasmic reticulum to Golgi apparatus Man,- cj t o Man, processing, the Golgi apparatus addition of the N-acetylglucosamine was sensitive to Nethylmaleimide and inhibited by GTP-yS. Coated buds associated with Golgi apparatus, which purportedly transport protein among cisternae, have been reported
.
. - . . I - . _ - rI
.
r
. .
I
1
.
,
,
-
_
.
,
-, - -',w-=- %
Perspectives on Golgi Apparatus Form and Function HILTON H. MOLLENHAUER AND D. JAMES MORRB pootl Animal Protection Research Laboratory, United States Department of Agriculture, Agricultural Research Service, College Station. Texas. and Department of Pathology and Laboratory Medicine. College of Medicine, Texas A & M Unzuersity, College Station, Tmczs 77840 (H.H.M.);Department of Medicinal Chemistry arid Pharmacognosy, Purdue University, West L a f a y t t e , Indiana 47907 ( D . J . M . )
KEY WORDS
Dictyosome, Endomembrane system, Terminology
In 1898, Camillio Golgi reported a new cellular constituent with the form of an ABSTRACT extensive intracellular network (the apparato reticolare interno),which now bears his name. However, the history of Golgi’s apparatus is replete with controversy regarding its reality, what components of the cell should be included under its aegis, and what terminology should be used when referring to it. Electron microscopy has resolved many of these controversies and it is appropriate that this volume emphasize that aspect of Golgi apparatus discovery. The principal structural component of the Golgi apparatus is the stack of cisternae, or dictyosome. As determined both biochemically and a t the level of electron microscopy, the dictyosome is a highly ordered and polarized structure. The maintenance c;f order within the stack is thought to result from either intercisternal bonding constituents, or filamentous structures (or both) that bridge the space between adjacent cisternae. Mechanisms proposed for movement of membrane and product into and out of the dictyosome (i.e.,the Golgi apparatus stack) include a serial mode which functions exclusively by the formation, displacement, and loss of cisternae from the stack, and a parallel mode which functions exclusively by the movement of membrane, product, or precursor molecules directly into the peripheral edges of the cisternae. In the parallel mode, all cisternae can be accessed either singly or simultaneously, a t least in theory, a t any position within the stack. It is probable that both the serial and the parallel modes function concomitantly and need not be mutually exclusive. Finally, the peripheral tubules of the cisternae represent a major membranous constituent of the cell with potentially unique functions. These tubules interconnect cisternae of adjacent stacks and may represent the major site of receptors for the shuttle (i.e., parallel) type of transfer among cisternae. Peripheral tubules as extensions of the cisternal lumina into the cytoplasm presumably have other functions, but these, like the tubules themselves, have only rarely been accommodated into functional models of Golgi apparatus dynamics in secretion or membrane flow.
INTRODUCTION descriptions (see Whaley, 1975, and references thereThe structure now known a s the Golgi apparatus ap- in). The basis of the controversy consisted of two issues: parently was observed by La Valette St. George and by 1) the impregnation methods were somewhat capriothers as early as 1865 (e.g., Platner, 1889 and Murry, cious and 2 ) the form of the structural entity that was 1898 as cited by Whaley, 1975). It was not popularly contrasted varied markedly among cell types and in recognized a s a potentially unique cellular entity until relation t o cell activity. The latter observations were the silver impregnation methods developed by Camillio contrary to what was expected for a universal cell orGolgi in 1873 were applied, first to nerve cells of the ganelle and it was not until many years later that Golbarn owl (Golgi, 1898) and then to other types of cells gi’s reticulum was recognized as a n extremely pleomor(see Bowen, 1926; Beams and Kessel, 1968; Whaley, phic cell component subject to changes in location 1975, for review). The effectiveness of Golgi’s procedure within the cell, intensity of activity, and functional was to stain various elements of the nervous system characteristics. This variability was (and still is) a mabrown or black and contrast this vividly against a n jor factor in establishing a realistic image of the strucalmost transparent background. With this impregna- ture of the organelle and determining how the comtion technique, Golgi observed a n external reticulum plexities of Golgi apparatus structure relate to its (which had been noted before) and an internal reticu- function. The metal impregnations introduced by Golgi caused lum (which had not previously been convincingly demthe internal reticulum to appear as a continuum of onstrated). It was the internal reticulum (i.e., the upparato reticolare interno), or a part of it, that was recognized first as the Golgi apparatus. The interpretation of Golgi’s observations generated Received April 20, 1989; accepted in r w i s e d form N ~ ~ v e m b e15. r 19x9 much debate. and numerous controversies regarding Address reprint requests to Hilton I I . Mollenhauer. USDA-AKS, Rt. 5, Box the reality of the reticulum ensued following the initial 810. College Station. TX 77840.
c
1991 WILEY-LISS, INC
GOLGI APPARATUS FORM AND FUNCTION
fibers in vertebrate somatic cells and as isolated scales in embryonic and germ cells (Golgi, 1898; Whaley, 1975).Thus, the total cellular complement of Golgi material may appear with the light microscope a s either a number of individual units (which we now recognize as dictyosomes or stacks), or a s highly elaborate (and continuous) Golgi zones of many interconnected units, or as any of a number of intermediate configurations between these two extremes. The microscopic image was complicated by the variability of the staining procedures and lack of all parts of the Golgi material t o respond to the stains in the same manner. Such problems were encountered by many of the early cytologists. For example, Baker (see Whaley, 1975) who was, perhaps, the most vociferous disbeliever in the reality of Golgi’s reticulum, suggested that the reticulum, if it existed a t all, included many cellular constituents to which Golgi’s name should not be attached. Certainly, the reticular part of Golgi’s apparatus (i.e., the filamentous part that Golgi observed between the scalelike units, i.e., between the stacks) may have included some elements of the endoplasmic reticulum. The latter have been shown by electron microscopy to precipitate Golgi stains in the same locations as those observed by the early cytologists (Kopsch, 1902, as cited by Inferrera and Carrozza, 1975). Even today, while the existence of a Golgi apparatus is no longer questioned, its boundaries, particularly in respect to its peripheral appendages (i.e., tubules, vesicles, and associated trans Golgi apparatus reticuli), are difficult, and perhaps impossible, to define precisely. In the classic electron microscope image, the Golgi apparatus appears as a structure consisting of stacks of flattened saccules (cisternae) which give rise to secretory vesicles (Mollenhauer and Morre, 1966a). The secretory vesicles are either attached directly to the cisternae (via short tubules; see Mollenhauer and Morre, 1966a) or in a region contiguous with the trans pole of the stack (Palade, 1969; Farquhar, 1985). Tubularivesicular networks are associated with the trans pole of each stack (e.g., Golgi Endoplasmic Reticulum Lysosome system (GERL)-Novikoff, 1976; Trans Golgi Network (TGN)-Griffiths and Simons, 1986; Compartment of Uncoupling of Ligand and Receptor (CURL)-Geuze et al., 1983, 1985; M6P-lgp enriched structure-Griffiths et al., 1988), which are implicated in membrane sorting and recycling (see Geuze and Morre; Griffing; and Beaudoin and Grondin, this volume). Another tubularivesicular network (i.e., the Partially Coated Reticulum (PCR)) of similar form, but less intimately associated with Golgi apparatus stacks has been reported that may also have membrane sorting and recycling functions (Pesacreta and Lucas, 1984, 1985). The numbers of cisternae per stack vary widely, although, in most animal cells and higher plant cells, there are 5-10. Each stack is polarized in the sense that the maturation (processing) of products and membranes appears to occur sequentially from a cis (forming) face on one side to a trans (maturing) face on the opposite side (Mollenhauer and Whaley, 1963). Each cisterna of the stack is unique in that it differs chemically, structurally, and hnctionally from the oth-
3
ers (Morre et al., 1971; Mollenhauer and Morre, 1980; Rothman, 1985). For simplicity, the stack may be divided into different regions (e.g., cis, medial, and trans), each representing a functionally different subset of cisternae (Rothman, 1985; Pfeffer and Rothman, 1987). In reality, however, it seems more likely that changes in function occur gradually across the stack of cisternae rather than in discrete steps. The dictyosomes of the Golgi apparatus most often occur side-by-side in mammalian cells to form a complex ribbon adjacent to the nucleus. In higher plants, fungi, and in some animal cells (e.g., invertebrates), however, the stacks or dictyosomes usually are sufficiently separated to appear as discrete units within the cytoplasm. Yet, even with the dispersed arrangement, the dictyosomes appear to be interconnected into a functional Golgi apparatus. The number of dictyosomes per cell in eukaryotes ranges from none (in certain fungi that lack stacked Golgi apparatus cisternae-Bracker, 1967) t o over 25,000 in some algal rhizoids (Severs, 1965).A typical plant or animal cell may contain 500 or more dictyosomes. The unstacked cisternae or tubular networks of fungi that lack dictyosomes carry out some Golgi apparatus functions. These Golgi apparatus equivalents (Bracker, 1967) are easily identified in electron micrographs even without the characteristic stacked configuration. If, as in a few algal cells, only one dictyosome is present, then this dictyosome is equivalent to the Golgi apparatus. While opinions may differ, multiple dictyosomes within a given cell function synchronously and appear sufficiently interconnected or interassociated in terms of their regulation to be regarded in totality as the Golgi apparatus. A single exception has been reported in Urodele sperm development by Werner (1970), where more than one Golgi apparatus appears to occur in a single cell. Each cisterna (pl. cisternae) consists of a lumen or central cavity that is surrounded by a membrane. The central flattened, or plate-like, region is most often referred to as the cisterna or saccule. At the periphery of some saccules, there may be fenestrae (openings) of varying diameter. The fenestrated margins of the saccules are usually continuous with a system of tubules that may extend for several microns from the edges of the saccules. Some tubules presumably connect with endoplasmic reticulum. Others may connect with cisternae of adjacent dictyosomes or, possibly, lysosomes through vesicular bridges. Cisternal tubules have been demonstrated by a variety of techniques both with isolated Golgi apparatus and in situ (e.g., Figs. 1,4-6,10, 11;Tandler and Morre, 1983) in both plant and animal cells (Mollenhauer e t al., 1967). In addition to changing from characteristics resembling endoplasmic reticulum to characteristics resembling plasma membrane, the differentiation of Golgi apparatus cisternae within the stack also involves changes in the arrangements of saccules, tubules, and vesicles. Toward the center of the stack, the plate-like, or saccular, regions are a dominant cisternal characteristic. Yet even these cisternae have fenestrated borders and tubules which emanate from their peripheries
4
H.H. MOLLENHAUER AND D.J. MORRfi
Fig. 1. Isolated rat germ cell Golgi apparatus negatively stained with phosphotungstic acid. These Golgi apparatus show clearly the stacks of plate-like cisternae interspersed with tubular intersaccular regions. Note large number of cisternal tubules, and differences in number of fenestrae and tubules on 0.5 km. individual cisternae. Bar
(Morre and Ovtracht, 1981).In some cells a t least, the trans cisternae may have extensive tubular networks without f'enestrae. Post-Golgi apparatus structures, for example, the trans Golgi network (e.g., TGN; Griffiths and Simons, 1986), or partially coated reticulum (PCR; Pesacreta and Lucas, 1985),lie adjacent t o the trans poles of the stacks. In plant cells, these structures appear t o be derived from trans cisternae as they are released from the stack (Mollenhauer et al., in press). The functions of the TGN and PCR may not be restricted only to the
sorting of exocytic and endocytic products (Tartakoff, 1982, 1983; Roth et al., 1985; Anderson and Pathak, 19851, but may participate also in the discharge of membrane a t the trans Golgi apparatus poles as well. Although the details of stack morphology and function are now well defined, questions remain a s to how the stack is formed and maintained, how secretory products andlor membrane move through the stack, and what is the role of the peripheral tubules. Some of these questions are addressed in the sections that follow.
GOLGI APPARATUS FORM AND FUNCTION
5
Fig. 2. A: Rat epididimal principal cell illustrating the form and interassociation of several dictyosomes (i.e., stacks of cisternae ID1). This cell has several hundred dictyosomes dispersed through much of the cytoplasm and each dictyosome has about 8-12 cisternae. These animal dictyosomes are similar, but not identical, to those of plants (see Griffng, this issue; Mollenhauer et al., 1967, Mollenhauer and Morre, 1978). B: Ryegrass rootcap cell illustrating the form and dis-
tribution of dictyosomes (D) typical of plant cells. Each dictyosome has 5-10 cisternac arranged in a polarized configuration. Cis (forming) and trans (maturing) faces are labeled “c” and “t”,respectively, and secretory vesicles (arrowheads) are illustrated. This cell probably contains between 400 and 600 dictyosomes distributed in groups or clusters throughout much of the cytoplasm. Bars = 0.5 km.
INTERCISTERNAL SUBSTANCES AND MAINTENANCE OF THE GOLGI APPARATUS STACK
al., 1973). Endoplasmic reticulum entering the Golgi apparatus zones of exclusion usually lack ribosomes. Mitochondria are excluded, but peroxisomes and lysosomes may be present. Free polysomes (i.e., the so-called Golgi apparatus polyribosomes; Mollenhauer and Morre, 1974; Elder and Morre, 1976) are associated closely with, but are not attached to, the Golgi apparatus membranes within the Golgi apparatus zone. Here, too, are concentrated the clathrin-coated vesicles associated predominantly with the mature faces of the Golgi apparatus stacks (Croze et al., 1982).
The cisternal stacks observed by electron microscopy (Fig. 2A,B) consist of cisternae that are separated physically from one another, but yet maintained in a precise order. This organization has been postulated to occur via substances within a specialized region of cytoplasm (the zone of exclusion) that surrounds elements of the Golgi apparatus (Franke et al., 1972; Mollenhauer and Morre, 1972, 1978a; Mollenhauer et
6
H.H. MOLLENHAUER AND D.J. MORRfi
When monovalent salts were added to the homogenization medium prior to glutaraldehyde stabilization, unstacking occurred within seconds to release intact cisternae into the homogenate (Fig. 5; compare with the nearly intact dictyosome of Fig. 4). These free cisternae often had plaques (Fig. 6A,B) and remnants of intercisternal elements (filaments) on the flattened parts of the cisternae (Fig. 6B,C). With continued exposure to monovalent salts, both the filaments and plaques disappeared followed by dissolution of the central parts of the cisternae (Fig. 6C,D). Breakdown of the cisternae began a t sites previously occupied by the plaque substances of the intercisternal region (Fig. 6B,C). While a n obvious function of intercisternal substances would be to stabilize the stack and maintain its organization, it seems likely that they may have other functions as well. For example, they might help control the timely release of vesicles and trans cisternae at maturity (Mollenhauer et al., 1973) as well as aid in the capture of precisternal vesicles a t the cis face of the stack.
MOVEMENT OF MEMBRANE AND PRODUCT THROUGH THE STACK Movement of membrane and product through the Golgi apparatus is accompanied by sequential (i.e., serial) formation, differentiation, and loss of Golgi apparatus elements (Mollenhauer and Whaley, 1963). The formation of new cisternae on one face of a Golgi apparatus stack would be required to compensate for loss Cisternae are separated from one another within of cisternae from the opposite face of the stack. The each dictyosome by a space of 7-15 nm. Because the source of these new cisternae was suggested early on t o stacks of cisternae can be isolated intact and subse- be a special region of the endoplasmic reticulum which quently unstacked by enzymatic, chemical, or mechan- gives rise to transition vesicles that move to, and conical means (Mollenhauer et al., 1973; Morre et al., dense on, the forming (cis) faces of the Golgi apparatus 19831,the intercisternal regions are thought to contain stacks where they would fuse together to form the new bonding substances that contribute to stack integrity. cisternae (Zeigel and Dalton, 1962; Morre and MollenIn invertebrate and plant cells, there are filamentous hauer, 1976). Alternatively, movement of materials elements in parallel arrays that lie midway between may occur via shuttle vesicles a t the peripheries of the adjacent cisternae (Mollenhauer, 1965a; Turner and cisternae to effect transfer from one cisterna to the Whaley, 1965; Fig. 3 ) . The numbers of these filaments next (Orci e t al., 1986).However, direct (i.e., nonvesicincrease toward the trans poles of the stacks and then ular) movement of substances into Golgi apparatus cisabruptly decrease just before release of the most trans ternae through the peripheral tubules of the cisternae cisterna. Dictyosomes of plant cells show a consistent must also be considered as yet another route for the and marked narrowing of the cisternal lumina from parallel delivery and transfer of substances in and out the proximal to the distal pole in almost direct propor- of the Golgi apparatus (Morre and Mollenhauer, 1974; tion to the number of intercisternal elements. Animals Mollenhauer et al., 1975; Mollenhauer and Morre, and fungi have Golgi apparatus that both lack the in- 1976; Harris and Oparka, 1983). Studies in plants have concentrated on the release of tercisternal elements and that fail to reveal the marked gradient in cisternal narrowing (Mollenhauer mature cisternae and contiguous secretory vesicles and Morre, 1978b). Although the function of intercis- from the trans faces of the stacks (Fig. 7) coupled with ternal elements is not known, there is circumstantial the formation of new, replacement cisternae on the cis evidence to suggest that they are associated with shap- faces of the stacks. Here, membrane and product are ing of secretory vesicles and/or cisternae (Mollenhauer considered to move serially through the stack as cisternae mature (Mollenhauer, 1971; Fig. 8).This mode and Morre, 1975; Kristen, 1978). One approach to relate the intercisternal region to of functioning is unequivocal since entire cisternae are organization of the stack has been to solubilize the in- separated from the stack (Brown, 1969; Mollenhauer, tercellular substances of isolated plant dictyosomes 1971; Mollenhauer et al., in press). Such separation and then stabilize what remains of the disintegrating must be accompanied by formation of new cisternae to stacks with glutaraldehyde (Mollenhauer et al., 1973). maintain stack integrity. This is, apparently, accomThe isolated stack constituents were negatively plished by a transport mechanism that shuttles vesistained with phosphotungstic acid and examined by cles from endoplasmic reticulum to the cis faces of the electron microscopy. stacks. The approximtely 60 nm vesicles that bleb from Fig. 3 . A maize root dictyosome with intercisternal filamentous elements viewed end-on (arrows,. The number of filaments increases toward the trans pole of the dictyosome (toward bottom of micrograph, but then decreasesjust before t h e most trans cisterna ( T Jis sloughed. Bar = 0.1 urn.
GOLGI APPARATUS FORM AND FUNCTION
Fig. 4. Plant dictyosome homogenized in medium containing glutaraldehyde and then isolated. The stacked configuration and much of’ the peripheral structure was retained. The dictyosome illustrated here has probably lost its most cis cisterna. Note that the central (flattened)parts of the cisternae are partially masked by tubules from overlaying cisternae (either cis or trans). Tubules of the type shown are also apparent in micrographs of fixed and embedded dictyosomes (see Figs. 10, 11). Coated vesicleibuds (arrowheads) were always present although they disappeared rapidly if exposed to PTA or monovalent salts before fixation. Bar = 0.1 p,m (ca).
7
Fig. 5. Plant cisternae isolated after treatment with 0.5 M monovalent salt (i.e., sodium chloride) for 90 seconds. Divalent salts (e.g., calcium chloride) tend to stabilize the stack. The smooth cisterna with fenestrae is probably from the center of the stack. The mass of tubules on the right appears to be an aggregate of two cisternae from either the cis or trans pole of a dictyosome. Bar = 0.1 km.
8
H.H. MOLLENHAUER AND D.J. MORRE
Fig. 6. This series of micrographs illustrates several stages in dictyosome disassembly by monovalent salts. Disassembly occurs very rapidly, first by cleavage of cisternae and then by dissolution of t h e central parts of the cleaved cisternae. Total time for dissassembly using 0.5 M monovalent salts a s t h e solubilization agent is usually less than 2 min. A: Plaques such a s this (arrow) were usually present on freshly cleaved cisternae. The peripheral tubules often appeared to be more turgid and stained differently t h a n the cisternae to which they were attached. B: Plaques and isolated bundles of intercisternal
filaments (arrowhead) were often visible a t t h e early stages ofdisassembly. The filaments and extensive tubular array identify this as a trans cisterna. C: When the plaques disappeared, the underlying membranes of the cisternae showed signs of breakdown (arrow). Filaments, although present in this micrograph, were usually absent a t this stage of disassembly. D: Eventually, t h e central parts of t h e cisternae were dissolved leaving only a ring and peripheral tubules. (Some of t h e micrographs a r e reproduced from Mollenhauer e t al., 1973.)Bars 0.1 wm.
nuclear envelope o r endoplasmic reticulum and form new cisternae are usually spherical and of uniform diameter. They appear to be coated by a nap-like covering (Mollenhauer e t al., 1976) distinct from the spiny coat of clathrin-coated vesicles (Croze et al., 1982) and from the coats described for budding profiles associated by Orci and collaborators (Orci et al., 1986) with intercisternal Golgi apparatus transport. The process is
completed when the most trans cisternae are released from the stacks. Similarly, in scale-forming algae (see Brown, 1969, and Melkonian e t al., this volume), whole cisternae containing newly synthesized scales are sloughed from the stack and secreted out of the cell in times as short a s one minute or less (Brown, 1969; Melkonian e t al., this volume). In some species, these scale secretion
~
GOLGI APPARATUS FORM AND FUNCTION
9
Fig. 7. Maize rootcap cell illustrating dictyosomes (D) and trans cisternae that have separated from the stack (arrows).Note the vesicular buds on one of the separated cisternae (arrowhead).Cell wall (W), vacuole (V), secretory vesicle (SV), starch in amyloplast (S). Bar = 0.5 km (ca.).
events can be witnessed in real time since the sizes of the scales are well within the range of resolution of light microscopes. Here one must conclude that both membrane and product move serially through the stack. In animals, studies of membrane and product movements through the stack have concentrated on vesicular shuttle mechanisms contiguous with the peripheral edges of the cisternae. Thus, Xothman and coworkers (Rothman et al., 1984) have suggested intercompartmental transport in the Golgi to be a dissociative (vesicular) process. Vesicular buds on the peripheral edges of cisternae (Orci et al., 1986) are thought to act as shuttles to move materials from one cisterna to another. These transfer vesicles would have access t o all cisternae simultaneously in parallel array (Fig. 8). Are the two models of membraneiproduct movement (i.e., serial differentiation of cisternae through the stack vs parallel differentiation by vesicle shuttle or direct endoplasmic reticulum transfer) mutually exclusive? Probably not. Vesicle buds of the type described by Orci and coworkers (1986) occur widely in both plants and animals (Cunningham et al., 1966; Mollenhauer and Morre, 1966a; Mollenhauer et al., 1973, in press; Fig. 3). Similarly, coated vesicles and post-Golgi reticuli are universally present a s well (Pavelka and Ellinger, 1983; Pesacreta and Lucas, 1984, 1985; Farquhar, 1985; Griffiths and Simons, 1986; Mollenhauer e t al., 1989, in press; Hubner et al., 1985; Griffing, 1988; Hillmer et al., 1988; Jones and Robinson, 1989). Thus, the structural bases for the parallel mode are present even in cells where the serial mode dominates. Similarly, cisternal maturation may also occur in animal cells (Morre et al., 1979) where the parallel mode is thought to dominate. What may differ from one cell type to another, however, is the degree to which each mode participates in membrane traffic within the stack. Furthermore, neither the serial or parallel mode
of functioning need dominate to the exclusion of the other. Both modes are probably capable of operating either separately or concomitantly to some extent in all species. An example of control of the mode of operation of the Golgi apparatus in plant cells is seen in a region of secretory vesicle transition in maize rootcap cells (Mollenhauer, 196513).In these cells, hybrid secretory vesicles form in a transition zone of the rootcap where cells from within the cap (which produce a spherical, epidermal-like, secretory vesicle) transform into cells of the outer rootcap (which produce a n elongate secretory vesicle; Fig. 9A-E). In the transition state, secretory vesicles are formed that contain both types of secretory product (Mollenhauer, 1965b; Fig. 9D,E). These hybrid vesicles form on the most trans cisterna of the stack as well as on cisternae within the stack. The hybrid vesicles clearly show that Golgi apparatus activities are developmentally regulated and that more than one pattern of activity can be expressed even simultaneously by the same basic Golgi apparatus stack. The data might support as well the presence of both serial and parallel forms of input into the Golgi apparatus stacks for the production of secretory vesicles as well as for cisternal maturation. In overall perspective, both serial and parallel routes through the stack have been proposed. Which route may predominate is unknown. However, the parallel route includes both vesicular and direct transfer of materials among all contiguous compartments within the pathway, whereas the serial path requires only the formation of new cisternae and displacement of mature cisternae.
GOLGI APPARATUS SHUTTLE VESICLES OF INTERCOMPARTMENT TRANSFER The formation, separation, and movement of vesicles that shuttle between cisternae has been proposed and
H.H. MOLLENHAUER AND D.J. MORRQ
10
is, potentially, a very important aspect of Golgi apparatus functioning. Evidence for the existence of a shuttle vesicle to transfer materials from one Golgi apparatus compartment to the next comes primarily from studies of Rothman and collaborators using the processing of vesicular stomatitus virus (VSV) glycoproteins as a model (Wattenberg, this volume). They employed a mutant BHK cell that lacked a processing enzyme UDP-N-acetylglucosamine N-acetylglucosamine:glycoprotein transferase unable to process VSV-G. When homogenates of infected cells containing radiolabeled viral precursors were mixed with homogenates of wild-type cells, processing of preformed VSV-G continued, suggestive of transfer from Golgi apparatus cisternae lacking the processing enzyme to Golgi apparatus cisternae containing the processing enzyme. Like the endoplasmic reticulum to Golgi apparatus Man,- cj t o Man, processing, the Golgi apparatus addition of the N-acetylglucosamine was sensitive to Nethylmaleimide and inhibited by GTP-yS. Coated buds associated with Golgi apparatus, which purportedly transport protein among cisternae, have been reported
.
. - . . I - . _ - rI
.
r
. .
I
1
.
,
,
-
_
.
,
-, - -',w-=- %
