key: cord-009371-ub4p4ngr
authors: Mollenhauer, Hilton H.; James Morré, D.; Rowe, Loyd D.
title: Alteration of intracellular traffic by monensin; mechanism, specificity and relationship to toxicity
date: 1990-05-07
journal: nan
DOI: 10.1016/0304-4157(90)90008-z
sha: 
doc_id: 9371
cord_uid: ub4p4ngr

Monensin, a monovalent ion-selective ionophore, facilitates the transmembrane exchange of principally sodium ions for protons. The outer surface of the ionophore-ion comples is composed largely of nonpolar hydrocarbon, which imparts a high solubility to the complexes in nonpolar solvents. In biological systems, these complexes are freely soluble in the lipid components of membranes and, presumably, diffuse or shuttle through the membranes from one aqueous membrane interface to the other. The net effect for monensin is a trans-membrane exchange of sodium ions for protons. However, the interaction of an ionophore with biological membranes, and its ionophoric expression, is highly dependent on the biochemical configuration of the membrane itself. One apparent consequence of this exchange is the neutralization of acidic intracellular compartments such as the trans Golgi apparatus cisternae and associated elements, lysosomes, and certain endosomes. This is accompanied by a disruption of trans Golgi apparatus cisternae and of lysosome and acidic endosome function. At the same time, Golgi apparatus cisternae appear to swell, presumably due to osmotic uptake of water resulting from the inward movement of ions. Monensin effects on Golgi apparatus are observed in cells from a wide range of plant and animal species. The action of monensin is most often exerted on the trans half of the stacked cisternae, often near the point of exit of secretory vesicles at the trans face of the stacked cisternae, or, especially at low monensin concentrations or short exposure times, near the middle of the stacked cisternae. The effects of monensin are quite rapid in both animal and plant cells; i.e., changes in Golgi apparatus may be observed after only 2–5 min of exposure. It is implicit in these observations that the uptake of osmotically active cations is accompanied by a concomitant efflux of H(+) and that a net influx of protons would be required to sustain the ionic exchange long enough to account for the swelling of cisternae observed in electron micrographs. In the Golgi apparatus, late processing events such as terminal glycosylation and proteolytic cleavages are most susceptible to inhibition by monensin. Yet, many incompletely processed molecules may still be secreted via yet poorly understood mechanisms that appear to bypass the Golgi apparatus. In endocytosis, monensin does not prevent internalization. However, intracellular degradation of internalized ligands may be prevented. It is becoming clear that endocytosis involves both acidic and non-acidic compartments and that monensin inhibits those processes that normally occur in acidic compartments. Thus, monensin, which is capable of collapsing Na(+) and H(+) gradients, has gained wide-spread acceptance as a tool for studying Golgi apparatus function and for localizing and identifying the molecular pathways of subcellular vesicular traffic involving acid compartments. Among its advantages are the low concentrations at which inhibitions are produced (0.01–1.0 μM), a minimum of troublesome side effects (e.g., little or no change of protein synthesis or ATP levels) and a reversible action. Because the affinity of monensin for Na(+) is ten times that for K(+), its nearest competitor, monensin mediates primarily a Na(+)-H(+) exchange. Monensin has little tendency to bind calcium. Not only is monensin of importance as an experimental tool, it is of great commercial value as a coccidiostat for poultry and to promote more efficient utilization of feed in cattle. The mechanisms by which monensin interact with coccidia and rumen microflora to achieved these benefits are reasonably well documented. However, the interactions between monensin and the tissues of the host animal are not well understood although the severe toxicological manifestations of monensin poisoning are well known. Equine species are particularly susceptible to monensin poisoning, and a common effect of monensin poisoning is vacuolization and/or swelling of mitochondria in striated muscle. Other pathological injuries to striated muscle, spleen, lung, liver and kidney also have been noted. A consistent observation is cardiac myocyte degeneration as well as vacuolization. Differences in cellular response resulting from exposure to monensin (i.e., Golgi apparatus swelling in cultured cells, isolated tissues, and plants vs.mitochondrial swelling in animals fed monensin) suggest that myocardial damage is due either to a monensin metabolite or is a secondary response to some other derivation. However, as pointed out by Bergen and Bates [26], the underlying mode of action of ionophores is on transmembrane ion fluxes which dissipate cation and proton gradients. Consequently, some or all of the observed monensin effects in vivo in animals could be secondary phenomena caused by disruption of normal membrane physiology resulting from altered ion fluxes.

Monensm, a Na ÷ ionophore capable of collapsing Na ÷ and H + gradients, has gamed wide-spread acceptance as a biochemical and biologacal investigative tool to study Golgi apparatus function and to localize and identify the molecular pathways of subcellular vesicular traffic. Among its advantages are the low concentrations at which lntubltions are produced (0.01-1.0 #M), a minimum of troublesome side effects (e.g., little or no change of protein synthesis or ATP levels), and a reversible action [1] . The purpose of this review is to examine the mechanism of action and specificity of monensln in Na+/H ÷ exchange and to attempt to reconcile tlus to the large body of structural and biochemical Information on monensin toxicity derived from animal studies

In 1964, Pressman and co-workers [2] reported a class of anUbiotics that induced alkali ion permeability in mitochondna and other membranous systems. These antibiotics functioned as lonophores (ion-carriers) to carry ions across hpid barriers as complexes soluble in the hpid phase of the membranes. The potential use of tonophores as probes of biological function, or as potential therapeutic agents, was recognized very early [2] [3] [4] [5] , but major economic Importance was not forthcoming until the discovery of monensin in 1967 and the recognlnon of tts potential in the poultry mdustrTy as a coccxdtostat [6] Subsequently, ~t was discovered that tonophores also could improve feed converston in rumtnants such as cattle [7, 8] , thus adding further to their commercial value Of the more than 100 lonophores that have been reported [9] , three, monensln, lasalocld and sahnomysin, have widespread commercial use Of those licensed, monensin is probably used most widely.

Several monensins have been tdentified [6, 10] _ Monenstn A, and specfftcally the sodium salt of monensin A (hereafter simply designated as monensm) ts derived from Streptomvces cmnamonensts [11] , and a crude mycehal preparation (Rumensm') containing about 6_6% monensin ts used commercially. The 1on spectfictty of monensm xs Ag > Na >> K > Rb > Cs > LI > Ca [6, 12] with approximately a 10-fold selectlvRy for sodtum over potassium [9] and little tendency to bind calcmm [13] i!1.

One of the original interests in ionophores was their percewed potential for directly modifying intracellular iomc gradients, pamcularly Ca 2+, winch would lead, hopefully, to the development of useful pharmacologic agents or, alternatively, provide a tool for studying cellular functtons medmted by changes m Ca 2+ [13, 14] . Of parttcular interest were divalent xonophores such as X-537A (normally considered a Ca 2+ ionophore -note, however, X-537A complexes Na ÷ and K + almost as well as Ca 2+) winch have been shown to mduce contractaon of aortic strips and increase the rate of contractility of tsolated perfused rabbit heart [5, 15] , and release Ca 2+ from energy-loaded vesicles derived from the sarcoplasrmc renculum of muscle [16] [17] [18] X-537A may also mcrease blood flow through coronary artenes and mcrease cardiac output [19] . The emphasis on heart physiology stemmed from this organ's strong dependence on calcmm for proper functioning [20] . Indeed, subsequent studies of X-537A used as a feed additive for poultry and cattle has shown that the heart ~s a primary target for lonophore toxicity [21] . X-537A affects many other cellular functions such as release of b~olog~cally active agents and the induction of sperm acrosome reactions of several species [19, 22] .

It was soon realized, however, that many of the ionotroplc effects of the divalent ionophores could be duplicated with even greater efficiency by monovalent ionophores such as monensm which complexes Na + but almost no Ca 2÷ Tins response apparently occurs because the movement of Na ÷ into a cellular compartment by monensin fac~htates the entry of Ca 2+ by a Na+-out/Ca2+-ln exchange [4, 9, 23, 24] . Thus, a Ca 2÷ shaft ts stall the pnmary factor medmtlng cellular responses although other factors may also play significant roles m monensm physiology For example, many tono-phores, either directly or as a result ol the lomc imbalance. may transport, promote uptake, or release effector substances such as serotonm, lustamme, prostaglandm and catacholamlne which, m turn. have profound effects on cellular function [4.19] Slmdarly. monensm. through alteration of the pH of mtracellular compartments may lnhlbR the release and/or transport of numerous agents and. in so doing, perturbate cellular function In cardiac tissues, both posmve and negative motroplsm has been observed sequentially (the variable factor being time) m tissues following exposure to monensin [4] Concentrahon of monensln may. also. cause similar posttlve/neganve ,notroptc responses [4,241. Finally. some monensln mgested b}¢ an animal is metabolized to other tonophores, the properties of which are largely unknown Thus, ionophores, in spite of many common characteristics, differ indtvtdually in their effects on cell~ Moreover. cells may respond to both direct tonophore interaction as well as secondary effects that develop from the initial ionophore reactions. The latter is partlcularly likely in the whole antmal where metabolites with unknown properties are produced from lonophore breakdown and where changes m the products of one organ can affect the function of other organs_

Monensin is an open chain molecule that is capable of ton complexation through a cyclic form stabilized by hydrogen bonding between the carboxyl and hydroxyl groups Charge transfer bonding Wltinn the cavity formed is responsible for 1on binding (Fig 1) _ Because the affinity of monensln for Na + is 10-ttmes that for K+. tts nearest competitor in biological systems. Bergan and Bates [26] ) The aniomc form of the lonophore is stabilized by the polar environment characteristic to the surface of a membrane_ The ionophore is capable of ion pamng with a metal cation either at the terrmnal carboxyhc acid moiety or at other internal sites The binding of a cation Initiates the formation of a hpophahc, cyclic catlon-lonophore complex that can diffuse through the interior of the blmolecular membrane structure_ After traversing the membrane, the complex is again subjected to a polar environment where the electrostatic forces that had stablhzed the complex are no longer greater than the unfavorable Gibbs free energy change of cyclzzation The ionophore then releases its enclosed cation and reverts to the low energy acyclic conformation Monensin, like other carboxyhc lonophores, binds metal ions through liganding s~tes such that the ions become centrally oriented ( Fig. 1) and masked from the extraceUular environment [10, 12, 19] . The outer surface of the ionophore-lon complex is composed largely of nonpolar hydrocarbon, which imparts a high solubility to the complexes m nonpolar solvents [10, 12, 19] In biological systems, these complexes are freely soluble in the lipid components of membranes and, presumably, diffuse through the membranes from one aqueous membrane interface to the other [10, 25] . Once the ion traverses the membrane as a monensin-lon complex, the ion is released, and the monensin molecule picks up a proton to form an undissociated molecule which then retraverses the membrane to release the proton to the outside of the cell, vesicle, organelle, or other subcellular compartment [19] (Fig. 2) . Thus, the net effect for monensin is a trans-membrane exchange of monovalent ions for protons. The 1on transfer rates may be very high and can approach, or even exceed, normal enzyme diffusion rates [19] , although the actual rate may be markedly altered depending on factors such as the concentration of K ÷ m the external medium and the 227 type and concentration of permeant ion that accompariles the accumulated K + [27] . However, effects of lonophores on cation transport and their distribution among different membrane-bounded compartments within the cell, will vary depending upon the physical and chemical properties of the different membranes. Membrane fluidity, thickness, curvature, charge and orientation of polar head groups of phosphohpids, cholesterol content, and protein content, all influence solubihty, penetration, and expression of the lonophore [28] . Moreover, in asymmetric membranes (i.e., biological membranes), ionophores generally exhibit asymmetric transport properties [29] . For example, Kovac et al [28] showed that both vahnomycm and nigericin (an lonophore similar to monensin) crossed the plasma membrane of Saccharomyces cereotslae at a rather low rate but then were preferentially located, and active as ~onophores, m the tuner mltochondrlal membrane. Thus, the physiological effects of monensln will depend on the membrane composition and functional characteristics of the different compartments involved.

Although mechanisms for 1on transfer through a blmolecular leaflet (membrane) have been proposed, questions still remain as to how this action is related to known effects of monensin on cell function and what relationships these may have, m turn, on blochermcal mechanisms leading to ammal toxaclty. While any consideration of ~onophore action must focus on the mechanism of lonophore interaction with biological membranes [26] , the complexity of the process and the multiplicity of potentml pathways involved suggests that a single causal mechanism cannot explain both the cellular responses and the clinical expressions of toxic-]ty in animals

Monensin is cost-effective in increasing the yield of meat from both fowl and ruminants [8] . In fowl, this increase m productivity is derived almost directly through the control of coccid]a that, if present, would adversely affect animal health [8] In ruminants, increased product]vity appears to result from several factors, the most obvious being an increase in the effectiveness of feed utlhzation [26, 30, 31] . These whole-animal effects (xe, systems effects) are well documented [26, [30] [31] [32] and will be paraphrased only briefly here.

The beneficial effects of monensm in cattle accrue, in part, through shifts m rumen rmcroflora population. For example, gram-positive bacteria (that are primarily acetate, butyrate, H2, and formate producers) are inhibited by monensin; whereas, gram-negaUve bacteria (many of which produce succinate) are less sensmve to monensm [26] . The outer layer of the multflayered well of the gram-negatxve bacterium may contribute to this resistance by acting as a barrier to the penetration of [ABLE I

Adapted from Ledger and Tanzer [1] In ~e~retlon Reduced secretion procollagen [34] [35] [36] [37] , fibronectm [34] [35] [36] 38] , proteoglycans [37, 39, 40] , prolactln [41] , alburmn [42] , transfernn [42] , promsuhn polypepudes [43] , larmmn [44] , a-amylase isoenzymes [45, 46] , newly synthesized proteins [47] , secretory proteins [48] , proteins for fast axonal transport [49, 50] , thyroxine-binding globuhn [51] , acetylchohnesterase [52, 53] , chononlc gonadotropin [54] , phytohemagglutmm [55] , very-low-density hpoprotem [56] , maize rootcap polysaccharldeS [57] (see however, Sticher and Jones [58] for lack of monensm effect), vesicular stomatltXS virus glycoprotein [59] , extracellular matrix [60] , type II collagen [37, 39] , reviews [19, 25, 61] Increased secreuon catecholarmne [62] , catheps'n D [63] _ Defective processing Pro-albumin to serum albumin [64] , receptors for msuhn and somatatomedin C [65] , pro-oplomelanocortm [66] Incomplete processing of ohgosacchandes (N-hnked and/or O-linked) myeloperoxldase [67] , PrENV glycoprotexn [68] ; fibronectm [69] , hCG subumts [54] , blocked formation of complex ohgo-~acchandes [63] , Herpes simplex glycoprotelns [70] , HLA-DR-as-SOClated mvanant chain [71] , coronavlrus glycoprotem [72] , review [73] Undersulphauon proteoglycans [37] , glycosammoglycan chains [39] , /3-D-xyloside glycosammoglycans [74, 75] In endocytosts and endosome actdzflcatton

Intubitlon of mternahzatlon arylsulfatase [63] , lmmunoglobulln [76] , a-2-macroglobuhn [77] , semlikl forest varus [78] , horseradish peroxldase [79] Inhibition of dlssocmtion of mternahzed hgand asmloglycoprotems [80] , asmlo-orosomucoJd [81] lnhlbxtlon of hgand transfer smb~s virus nuclear capsids to cytoplasm [82] , epidermal growth factor, /3-hexasamlmdase, low-density hpoprotein, lmmunoglobulIn, and proteoglycans to lysosomes [40, 76, 83] Inhibition of acldfflcauon endocytlc vesicles [84] [85] [86] , lysosomal and prelysosomal compartments [63] , interference with semllka forest virus genome penetraUon [78] , expression of diphtheria toxin [87] ; recycling of LDL receptors [88] , release of diphtheria toxin from endocyuc vesicles [40, 87] lnhlbltton of mtracellular degradation proteoglycans [40] , insuhn [89] , lysosomal (methylarmne-sensxtlve) protem degradation [90] Inhibition of contraction of contractile vacuoles Paramecmm aureha [911

Surface formation and growth Altered secretion of cell surface molecules proteoglycan [37, 39, 40] , type If collagen and/or procollagen [36, 37, 39] ; fibronectin [36, 38, 69] , lamlmn [44] , mcorporaUon of sulfaUdes into myelin [92] , incorporation of Po protein and myehn basic proteins into myelin [93] Inhibition of scale morphogenesis scales of the green alga Pyrarmmonas mconstans [94] Inhibition of cell spreading-cultured fibroblasts [95] ; mesoderm ceils [60] Stimulation of receptor capping mouse T-lymphoma cells [96] Inhtbltlon of growth rye seedhngs [97] , Pelhu 198] Transport of molecule.~ Recognition of independent secretory pathways acetylchohne receptor and acetylchohnesterase [53] , membrane glycoprotems/ assembly of Uukunlerm virus [99] , Ca2+-dependent and Ca2+-mdependent secretion of a-amylase [100] , proteoglycan and hyaluronate [101] , prolactln [41] ; galactosyl receptor [102] Maturation and/or transport of viral coat proteins vesicular stomatltis virus [42, 103] , herpes sxmplex varus [70, 104] , semhkJ forest wrus [105] , Uukumerm wrus [99] , alphavtrus [106] , coronawrus glycoprotean [72] , bovine herpes virus type 1 glycoprotelns [107] St~mulatlon of sugar and sugar nucleoude transport avian erythrocytes, Isolated rat and mouse dmphragm muscle, and red cells [108, 109, 110] ; mouse thymocytes [111] Redirection of secretory product plasmalemma to tonoplast [112] Inhibition of mtracellular transport: protein to rod outer segments [113] , myeloperoxtdase [67] , hCG subumts [54] , accumulation of lammln [44] ; gp70 glycoprotein [68] , procollagen [34] , fibronectm [34] Interactions wtth other toxm~ Enhancement of toxaclty tlamuhn m swine [114] , dlsulfide-hnked methotrexate-anti-transferrln receptor conjugate [115] , specific cytotoxlcxty of a breast cancer-associated antigen lmmunotoxan m humans [116] Reduction of toxicity selenmm and vitarmn E [114] monensm; although, a more direct influence revolving differenes in membrane energetlcs also has been implicated [26] Monensln also may decrease the degradation of dietary protein in the rumen and, thus, increase the amount of protein avaalable for dlgesuon and uptake in the small intestine [26] . Both a reduction in overall cell numbers m the rumen and a direct effect of monensln on bacterial protelnase and dearmnase activity have been suggested as contributing to this effect [26, 33] .

One of the first subcellular effects observed in relation to the topical apphcatlon of monensin was vacuolatlon of Golgi apparatus clsternae [34] . Subsequently, xn vitro studies clearly demonstrated that monensin altered or inhibited numerous membrane-located phenomena (Table I) . Among these were the transfer of a2-macroglobuhn from coated pits to receptosomes [77] , recycling of low-density hpoprotem receptors [88] , pinocytosis [79] , transfer of product from endoplasmac reticulum to Golgl apparatus [34] , maturation and/or transport of viral coat proteins [42, 70, 72, 99, 104, 105, 117, 118] , inhibition of transport of membrane proteins to rod outer segments [113] , xnhibltion of cholesterol transport from the Golgl apparatus to the mltochondnal site of steroidogenesis [119] , blockage of phytohemagglutinln transport out of Golgl apparatus and into protein bodies [55] , inhibition of procollagen and fibronectm secretion from cultured human fibroblasts [35] , inhibition of carbohydrate processing in cultured human flbroblasts [63] , and inhibition of processing and cellular secretion [34, 37, 41, [120] [121] [122] .

Additionally, cellular effects of monensm vary markedly depending upon the organism and the route of adrmnlstratlon. Cultured cells, cells of tissue shces or explants, and plant organs that have received a topical exposure to monensin sufficient to inhibit growth or some cellular processes, usually show deviations in Golgl apparatus structure and function. In contrast, ceils from animals poisoned by ingested monensin often exhibit gross mltochondrlal lesions without the corresponding Golgl apparatus modifications. The reasons for these fundamental differences in cellular responses to monensin provide one focus for the present review and illustrate the many complexities surrounding the use of monensln as a probe specific to a single metabolic process.

The primary functional unit of the Golgi apparatus is a stack of membranous compartments (i e., the Clsternae) each of which differs chemically, structurally, and functionally from the others [123] [124] [125] [126] . The number of cisternae per stack varies widely; although, in most animal cells and higher plant cells, there are about 5-10 cisternae per stack. Each stack is polarized in the sense that product and membrane maturation appear to occur sequentially from a ClS (fornung) face on one side to a trans (matunng) face on the other side [127] . For simplicity, the stack may be divided into umts (measured from cis to trans faces) each of which represents a known set of functions. Currently, there appear to be some 2-3 such units that make up each stack [128, 129] . In reahty, however, it seems more likely that these changes in function occur gradually across the stack of Clsternae rather than in discrete steps.

There is, also, one or more membraneous structures (e.g., the trans Golgl network [130] or TGN and the partially coated retlculum [131] or PCR) that he just off the trans poles of the stacks In plant cells, these structures appear to be derived from sloughed trans Clsternae [132] . TGN and PCR participate in the separation (I.e., sorting) of both secretion and endocytic products [61, [133] [134] [135] [136] and regulate the release of endocytosed substances through a pH-sensltlVe mechamsm [84] . The functions of these post-Golgi apparatus structures are rapidly affected by monensin.

Several mechanisms for the movement of membrane and product through the Golgi apparatus have been postulated. For example, movement may occur by sequential maturation of Golgi apparatus elements (i.e., formation, maturation and loss of cxsternae) through the Golgl apparatus stack [127] . This would require the formation of new Clsternae on one face of a Golgl apparatus sack and commensurate loss of Clsternae from the opposite face of the stack The source of these new Clsternae is a special region of endoplasmic retlculum which gives rise to transition vesicles that move and condense on the forming (ClS) face of the stack where they fuse together to form the new osternae [137] [138] [139] Product movement may also occur by shuttle vesicles at the peripheries of the Clsternae that move proteins from one cisterna to the next [140] . However, both direct (i e., nonvesicular) movement of substances into Golgi apparatus cisternae and an endoplasmlc reticulum-mediated movement of product through the penpheral tubules of the Clsternae must also be considered as viable options for the delivery and transfer of substances in and out of the Golgl apparatus [141] . The post-Golgl apparatus structures appear to move membrane and product almost entirely via shuttle vesicles, many of which are coated [142] [143] [144] Monensin exerts its most profound effects on the trans cxsternae of the Golgi apparatus stacks in those regions of the apparatus primarily associated with the final stages of secretory vesicle maturation and in post-Golgl apparatus structures primarily associated with endocytosis and membrane/product sorting. Because of its relative specificity, biologists have used monensln extensively as an inhibitor of trans Golgl apparatus function.

Incorporation of radiolabeled [35S]methlonine into secreted lmmunoglobulin M molecules in monensintreated cells was reduced as was slalylatlon of immunoglobulin M and lymphoid cell surface glycoprotelns [145] . These latter findings showed that the intracellular processing of N-asparagine-hnked oligosacchandes is altered in the presence of monensln with an effect primarily on those sugars (e.g., slalic acid, galactose, fucose) added late in the processing continuum [73] Flbronectin, secreted in human flbroblasts, was incompletely processed in the presence of monensln and exhibited a greater incorporation of mannose than did control protein molecules [69] Inhibition of fibronectln secretion in human melanoma also has been reported [38] . Similarly, when treated with monensln, rat astrocytes in primary culture accumulated lamimn, another matrix glycoprotem Involved in cell adhesion [44] .

Not only do the monovalent lonophores block transport and surface expression of several secretory glycoproteins in normal cell functioning [34, 35, 42] and the transport of membrane glycoproteins or enveloped viruses [70, 72, 99, 103, 104, 106, 117, 118] , they Inhibit formation of cell surfaces including assembly of peripheral myelin [92, 99] . In mouse thymocytes, monensin leads to 23 (1 stimulated incorporation of labeled sialyl-, galactosyl-, and N-acetyl glycosaminyl residues [111] . This enhanced accumulation was not due to a &rect effect of monensin on glycosyltransferase activities but, rather, as a consequence of a greater entry and accumulation of labeled sugar nucleotides in the swollen Clsternae_ Galactosyl transferase itself was translocated through the Golga apparatus at a slower rate with monensin_ However, the sialylatlon of the O-11nked ohgosaccharides of the enzyme was unaffected by monensln treatment [59] Effects of monensin on glycosyltransferases also may be indirect. Monensin has been reported to decrease galactosyltransferase activity m Golgl apparatus of rat embryo fibroblasts [79] although it had no effect on this activity in baby hamster kidney cells [146] Monensln is an especially useful inhibitor, since it blocks lntracellular transport of protein at the level of the Golgl apparatus without directly affecting protein synthesis [35, 52] The effect of monensin is considered to be on transport rather than on processing per se [121, 147] One argument is that oligosaccharide processing of those glycoprotelns that reach the appropriate site occurs normally even xn monensin-treated cells [35, 37, 68, 121] . However, these observations could be explained as well if processing of ollgosacchande chains of different secreted glycoprotelns occurred at different sites, only some of which were sensitive to monensln [53, 68] .

The ablhty of monensln to effectively 'freeze' processing of molecules at a particular stage had lead to its use in identifying transitory synthetic intermediates. Examples include the insulin receptor where several polypeptlde precursors have been described [65] , the intracellular accumulation of non-cleaved precursors of pituitary hormones that occur in the presence of monensin [66] , and dissection of the pathway for secretion of gonadotropin by cultured human trophoblastic cells [54] .

In some instances, the effect of monensin may be to redirect, rather than block, the movement of Golgl apparatus-derived product. For example, under normal conditions, proteins of developing seeds accumulate in a central vacuole which then partitions into smaller units of storage protein (I.e, the protein bodies). However, when treated with monensin, the Golgi apparatus-derived transport of the protein VlClhn in pea cotyledon was redirected from the vacuole to the plasmalemma and the newly synthesized viclhn was released from the cotyledon cells to accumulate between the plasmalemma and cell wall [112] Monensin lnhibmon of Golgi apparatus function ~s sufficiently well established [25, 42, 61, 121] that the phenomenon is used widely as one criterion for verifying the passage of a biochemical entity through the Golgi apparatus. Thus, based on partial monensm inhibition, Hammerschlag and co-workers [49, 50] concluded that passage through the Golgl apparatus was an obllgator~ step m the lntracellular routing of materials m ta,~t axonal transport_ Bartalena and Robblns [51] showed that rnonensin ~mpeded the exit of thyroxin-binding globulin from the GolgJ apparatus w~thout affecting the terminal glycosylatlon of the protein Yanagashito and Hascail [40] reported that monensin reduced and delayed transport of both secretory and membrane-associated forms of proteoglycans, suggesting passage through the Golgl apparatus of rat ovarian granulosa cells m culture Similarly, an involvement of the Golgi apparatus m the transport of sulfatldes to myelin [92] and phytohemaghitlnln to protein bodies m bean cotyledons [55] were deduced from monensin mhlbmon Fhckinger and co-workers [48] using [3H]leucine, showed that all, or nearly all, of the protein secretory product of mouse epididymis principal cells pass through the Golg~ apparatus in times approximately eqmvalent to those reported m other tissues. This transfer of product from Golgi apparatus to the cell surface was largely blocked by monensin

Swelling of Golgi apparatus cisternae observed in the electron microscope following fixation with glutaraldehyde is, perhaps, the most consistent visual in vitro demonstration of a monensin-mduced effect on a membranous cell compartment [34, 41, 61, 121, [148] [149] [150] The swollen Clsternae usually appear devoid of contents by electron microscopy (Figs. 3 and 4) but an electron-dense substance may be precipitated through the osmium tetroxlde-zmc iodide reaction (unpubhshed data) Although all Clsternae of the Golgi apparatus may swell in response to monensin ( Fig. 3 and 5A) , the major effect appears to be associated with the mature, or trans, parts of the Golgi apparatus stacks (Figs 4 and 5B) [148, 149, 152 ]_ Gnffiths and co-workers [118] showed that monensm inhibited the transport of viral membrane proteins from medml to trans Golg~ apparatus cisternae, thus indicating a monensln block between medial and trans cisternae Monensm also blocked tlamrmng of the high mannose bound to the viral membrane proteins and their conversion to complex ohgosacchartdes Similarly, Niemann and co-workers [72] found that monensm blocked glycosylation of E1 glycoproteln of corona virus in infected mouse cells. Srinlvas and co-workers [68] reported failure to process simple endo-H-sensitive to complex endo-H-reslstant ohgosacchandes and reduced efficiency of cleavage of the PrENV glycoprotein precursor to gp70 for Evehne mouse cells infected with Friend munne leukemia virus These findings indicate a block prior to entry into the Golgl apparatus Also, m cultured hepatoma cells, transport of vesicular stomatltis wrus (VSV) G protein was arrested prior to acquisition of endo-H resistance, suggesting a block early in the processing pathway [42] . Strous and co-workers [49] showed that monensin affects primarily the galactosyltransferase-containlng c~sternae of the Golgi apparatus based on studies of the metabohsm, localization, and biosynthesis of N-and O-linked oligosaccharides of galactosyltransferase in HeLz cells. The accumulation of incompletely processed glycoproteins indicates either an up-stream accumulation of secretory materials behind a Golgl apparatus blockage by monensm [44] or a monensin block near the exit site from endoplasmlc reticulum [68] . Monensin effects on Golgi apparatus have been observed in a wide range of plant and animal species and appear to be a universal response to the topxcal applicatton of monensln. As pointed out above, monensln action is exerted on the trans half of the stacked clsternae (Fig 4) , often near the point of exit of secretory vesicles [1, 34, 61, 57, 118, 149, 152] or, especially at low monensm concentrations or short exposure times, sometimes m the rmdreglon of the stacked cisternae [148, 149] . Intracellular transport may be blocked [35, 42, 49, 68, 117, 121] , often wlthxn rmnutes after exposure to monensln [88, 106, 125, 152, 154] . Swollen units usually accumulate near the Golgi apparatus [147, 152] 

A monensm effect is quite rapid in both animal and plant cells; 1 e, changes in Golgi apparatus have been observed after only 2-5 rmn of treatment [85, 149, 152, 155] . These early effects have been documented particularly well is suspension cultures of carrot (Daucus carota L ) [152] . When carrot cells were exposed to monensln at 10 -5 M (which IS approxamately the minimum concentration that will elicit a strong monensm response m plants), production of secretory vesicles ceased and, almost Immediately, an increased number of Clsternae in the dlctyosome stacks was observed. An average of one additional clsterna per stack was formed within the first 2-4 mln of monensin treatment and, in some experiments, a second Clsterna was formed within about 6 man. These effects occurred without significant swelling of cisternae. Thereafter, vacuoles, representing intact swollen Clsternae, began to accumulate in the cytoplasm at a rate of about one every 3-4 min (Fig. 6) . The mechanism postulated for this momentary increase of d~ctyosome clsternae was that monensln, acting on the trans pole of the dictyosome, blocked normal formation [125] for details of procedure) The two samples differed onl~¢ m that the one dlustrated m {A) was adjacent to a 'natural' (1 e_ uncut) surface of the hver lobe, whereat (B) was from a cell near the 'cut' surface of the tissue slice In both instances, come Golgx apparatus (GA) clsternae were sv~ollen, however, m (A) swelling was progresse~e from cls to trans pole (direction of arrow) whereas m (B) swelling was t.onfined to the trans c~sterna Note that m~tochondrm (M) were condensed of secretory vesicles but did not block the formation of new Clsternae at the cis face of the apparatus However, as the trans clsternae began to swell, the swollen Clsternae were eventually released as intact units that neither fragmented nor integrated (fused) with other cellular constituents (e.g., plasma membrane). With the scale producing green alga Pyramimonas lnconstans, exposures of 1 to several hours to monensln resulted m disorientation of the Golg~ apparatus and disruption of scale morphogenesls [94] . These effects were reversible More recent studies indicate that a similar pattern of swelling and accumulation of clsternae in the cytoplasm occurs in cultured animal cells [151] . When H-2 hepatoma cells were treated for varying times with 10 ~ to 10 5 M monensin, one swollen Clsterna per stack of clsternae was produced after 6-8 rain of treatment During tlus time, approximately one additional clsterna per stack was formed (Fig. 7, inset) . As the Clsternae veslculated, vacuoles began to appear m the cytoplasm These large swollen vacuoles were formed at the rate of one sin concentrations, the vacuoles were larger and appeared more rapidly than at low concentrations of monensln but the kinetics of vacuole formation were qualitatively similar. However, by 2 h following treatment with 10 -6 M momensin, all Clsternae of the Golgi apparatus appeared as vacuoles. The swollen trans compartments that accumulate in the Golgl apparatus region with monensln inhibition may contain regmns that are clathrin coated; e.g., condensing secretory material (prolnsuhn) in pancreatic cells [43] . The response of Golgi apparatus of hver slices to monensin was qualitatively similar to that with hepatoma cells in culture [154] . With liver shces, a fraction enriched in vacuoles was isolated and demonstrated to contain the trans Golgl apparatus markers, galactosyltransferase, and thiarmne pyrophosphatase, in ratios sirmlar to those of Golgl apparatus proper [154] . In barley aleurone layers, the a-amylase and acid phosphatase activities that accumulated within aleurone cells following treatment with monensin, were localized in cellular components with buoyant densities intermediate between endoplasrmc reticulum and mitochondria and cosedimented with latent inoslne diphos-phatase activity, a putative Golgi apparatus marker in plants [45] . Heupke and Robinson [46] reported a shift to higher density of Golgi apparatus membranes from monensin-treated barley cells, a response no obvious from work with mammalian cells. The Golgi apparatus Clsternae that accumulated behind a monensin block in Semlikl Forest virus-infected BHK cells bound viral nucleocapsids, and the resulting increase in density permltted their separation by gradient centnfugatlon from other Golgi apparatus elements [146] .

The effects of monensin on Golgl apparatus, at least up to several hours of exposure, appear to be fully reversible [83, 94, 152, 154] In carrot cells, normal secretory activity was resumed within 20 nun after transfer of cells to a monensln-free medium; although, in these cells, the vacuoles formed during the monensln block, remained in the vicinity of the Golgi apparatus for several hours or more, even after apparently normal secretory activity had resumed However, with the longer treatment times of several hours [98] or days [95] 

Monensin apparently causes swelhng of Golgi apparatus cisternae through a Na+-in/H+-out exchange across the membranes leading to a net uptake of Na ++ CI and entry of water [157, 158] Evidence in support of this concept was provided by studies with Isolated chromaffin granules which lysed readily after brief exposure to monensln in Na +-or K+-contalnlng ISOtOmc media For swelling to occur, the membrane must normally be lmpermeant to cations as is known for the chromaffin granule_ The chromaffln granule membrane contains a H+-ATPase which is electrogenexc and, in the presence of a permeant anion, acidifies the granule interior to pH 5 5_ Thus, the operation of this pump in the presence of monensin drives net salt uptake [158] . To test whether net salt uptake driven by the presence of a proton gradient also would explain the monensininduced swelling of Golgi apparatus cisternae, wild-carrot cells in suspension culture were treated with drugs and inhlbltors known to interfere with proton gradients [156] Monensin-induced swelling of Golgl apparatus in sJtu could be inhibited by the protonophore, carbonylcyanlde-p-trlfluoromethoxyphenylhydazone (FCCP), But was only little affected by the inhibitor of lysosomal acidification, quercetln, or by the lysosomotropic amines, chloroquine, and ammonia. Cyanide also dramatically decreased swelling, and arsenate (with prolonged treatments) reduced the number of swollen cisternae Organic acids, by providing a readily permeable counterlon, promoted monensln-lnduced swelhng These data imply that the monensin-induced swelling of Golgi apparatus cisternae involves a proton gradmnt at or near the mature poles of the Golgi apparatus Because monensin induces a 1 " 1 Na+/H + exchange, and since the Van't Hoft factors for H + and Na + are practically the same [159] , the osmolanty of the cell content should not increase to cause swelhng without a net proton influx One explanation would be that the pH of Golgi apparatus vesicles is highly regulated Proton translocatlng ATP hydrolyzing enzymes (H +-ATPases) are associated with several components of cells that develop acidic intermrs such as endosomes, coated vesicles, lysosomes, and trans Golgl apparatus Clsternae [160] . Ewdence for the presence of an H +-ATPase has been the demonstration of ATP-dependent vesicle acxdlflatlon in Golgl apparatus isolated from rodent liver [161] [162] [163] , corn coleoptlles [164] , and sycamore cells [165] Acldificatmn was demonstrated both by [lac]methylamlne uptake and by spectrophotometric assays of amd-quenched dye fluorescence (Acridine orange, Neutral red, or Quinacrme) Several lines of evidence confine the Golgi apparatus H +-ATPase to the trans cisternae. As emphasized in the preceding section, the early in SltU effects of monensm are frequently localized to the trans faces of the Golgl apparatus The resultant swelling, whmh is proposed Io be due to the accumulation of osmotically active ions in exchange for protons [156.166] , occurs predominantly m trans Clsternae Additionally, a basic congener o] dmltrophenol (3-(2,4-dlnitroanlhn o)-3'-amlno-Nmethyldlpropylamine, DAMP), which concentrates m acidic compartments as shown in fibroblasts by lmmunocytochemlstry is present onl) 15 Clsternae and vesicles associated with the trans faces of the Golgl apparatus [84] Moreover, DAMP rapidly leaves these compartments when cells are incubated with monensm, thus further indicating that accumulation of DAMP is due to the acid pH Some involvement of a low pH compartment is evidenced by the observatmn that some monensln effects on processing, ke, proteolytlc conversion of proalbumln [64] , are mimicked by amines However, in promyelocytlc leukemia cells, processing of myeloperoxldase, while blocked by both monensln and chloroqulne, was not affected by NH~ cations, thus indicating that processmg is not necessarily influenced by pH-dependent mechanisms [67] These results were interpreted as indicative of processing in Golgl apparatus based on inhibition of transport by monensin and chloroquine rather than processing m lysosomes and other late, acidic compartments involving a pH-dependent mechanism [671 Confirmatmn of a trans location of the H ~-ATPase has come from free-flow electrophoresis separations of Golgl apparatus yielding ClS, medial, and trans compartments m fractions of diffenng electrophoretic mobility [154, 167] _ In these separatmns, proton pumping activity was found exclusively in the most electronegative fractions coming from the trans-most Goigl apparatus re-gion_ Gnfflng and Ray [166] have offered the suggestion that acldificatmn of Clsternal lumlna may be part of an osmotic mechanism to compress and flatten the cisternae The latter, for example, might aid in the transfer of content into secretory vesmles. The inward pumping of protons would tend to favor the exit of Na + and K + out of the Clsternae Furthermore, as the pH falls, those monovalent cations remaining would tend to combine with acidm groups of the Golgx apparatus membranes, further reducing the osmolanty of the Clsternal content relative to that of the external cytoplasm. Thus, water would be driven osmotically out of the clsternae to both compress the Clsternae, as seen along a pronounced ClS to trans gradient for plant Golgl apparatus [168] , and, perhaps, to account for the condensation of secretory materials m condensing vacuoles and other trans Golgi apparatus compartments (e.g, ref 169 ).

Other cell compartments, including endocytic vesicles [170, 171] , lysosomes [172, 173] , multivesicular bodies [174] , and coated vesicles [175] [176] [177] have H+-ATPases. All use H+-ATPases to acidify their interiors and the enzymes responsible have been solubdized from lyso-somes and reconstituted into liposomes [178] . Vacuole (tonoplast) membranes [179] , and possibly also plasma membranes of fungi [180] , contain H÷-ATPase. Similarly, a gradient of acidification within the endocyUc pathway has been indicated from immuno-electron microscopy with protein A-colloidal gold and monospecific antibodies to the weak base pnmaquine [181] .

However, not all compartments with H+-ATPases (e.g., cells, vacuoles, lysosomes, coated vesicles) swell in response to monensin. Cell and/or vacuole swelhng may be hmlted due to the very large internal volumes revolved. The contracnon of contractile vacuoles of Paramectum was inhibited reversibly by monensin and in a manner dependent upon the presence of Na + [91] but marked swelling was not observed. Little is known about the swelhng response, ff any, of lysosomes, coated vesicles and other endocync compartments m response to monensin lntubition. Their functions, however, are inhabited by monensin as will be emphasized in the section that follows.

Carboxyhc lonophores strongly inhibit proton uptake by photosynthenc preparations [19] . In chloroplasts, swelling of thylakolds (inner membrane compartments) but not of the space between tuner and outer plastld membranes has been observed to result from monensln treatment [182] . Thylakold swelling, in contrast to swellmg of mitochondrial cristae and of Golgl apparatus Clsternae, was reduced upon mcubanon in darkness, again suggesting a relationship between swelling in the presence of monensln and the hght-driven proton gradient used for photophosphorylatlon [182] Mltochondrla have an outwardly directed energy-hnked proton pump and do not swell with monensln (rather, they tend to condense, see Figs. 3, 5A and 5B) while the light-driven proton pump of chloroplasts and chromatophores, and that of the Golgi apparatus pump, are directed reward causing the vesicles to swell.

Evidence for a secretory pathway bypassing the Golga apparatus in the monensln-blocked cells is provided by Kubo and Pigeon [183] who lnvesngated the effects of monensin on the synthesis and expression of membrane IgM of a human lymphoblastold hne. They found altered processing of both /L, K chains and incomplete terminal glycosylations. Yet, transport of the altered molecules was observed. That the aberrant processing did not influence markedly the membrane expression of the IgM is consistent with a secretory pathway bypassing the Golgl apparatus m monensin-blocked cells. Slmdarly, a dual secretory pathway, only one part of which was suscepnble to monensin, was deduced from studies of a-amylase secretion m rice seed scutellum [100] . In Zea maJze roots, monensln lnhib~ted secretion of aamylase but not polysaccharide slime [58] .

The blocked secretion that results m the intracellular accumulation of secretory products frequently is not absolute. Some portion of the material synthesized is released from the monensm-mhabited cells and this material frequently exhibits an abnormal type of posttranslational modification. For example, those proteoglycans from chicken embryo chondrocytes secreted m the presence of monensln are vastly undersulfated [37, 39, 74, 75] . Thus, membrane and secreted molecules leaving the cell following a monensin block appear to have been denied the full range of processing enzymes they would normally encounter during transit through the cell However, whether the incompletely processed molecules bypass one or more parncular lntracelhilar compartments (eg, the Golgi apparatus), or whether they pass through functionally incomplete compartments, remains to be determined. During maturation of Uukunieml virus m baby hamster kidney cells, monensin appeared to mhibit a terminal step of virus assembly, but not the expression of virus membrane glycoproteins G 1 and G 2 at the cell surface. These findings suggest that both G 1 and G 2 could enter a functional transport pathway in the presence of monensm that bypasses the trans Golgi apparatus compartment to become expressed at the cell surface [99] . Evidence for a Golgi apparatus bypass has been presented m liver where secretory hpoproteins may move directly from endoplasnuc reticulum to the cell surface without direct Golgl apparatus involvement [184, 185] .

At concentrations of 10 -7 M or higher, monensin inhibited secretion of albumin, transferrln, and VSV proteins G and X destined for delivery to the cell surface to the same extent m rat hepatoma cells [42] . This was taken as evidence that the same vesicles were used by all four proteins m their movement from Golgl apparatus to the plasma membrane However, the time required to move from ER to the Golgi apparatus, based on sensitivity to endoglycosidase H, differed for secretory and membrane proteins. An even more striking observation was that following the monensm block, secretory proteins accumulated in an endo-H-sensitive form, whereas, membrane proteins were already endo-H-resistant. This strongly implies that membrane and secretory proteins are not m the same compartment initially and would support the concept of peripheral input of secretory proteins into the secretory vesicles of the Golgi apparatus, at least in hver [42, 186, 187] Alonso and Compans [103] provided evidence for two distract pathways of glycoprotein transport in Madin-Darby canine kidney (MDCK) cells only one of which was blocked by Monensm However, in baby hamster kidney (BHK21) cells, both influenza virus and VSV maturation were sensitive to monensin. The VSV particles were synthesized m both MDCK and BHK21 cells, but transport to the cell surface was blocked only in the MDCK cells Thus, there appear to be two distinct pathways of transport of glycoproteins to the plasma membrane in MDCK cells, only one of which is blocked by monensln There is no information on the nature of the alternative transport vesicle that carries mfhienza virus to the cell surface of MDCK cells if, in fact, a vesicle is involved.

Melroy and Jones [45] reported accumulation of normally secreted a-amylase within barley aleurone layers after monensln treatment. However, only isozyme 2 was secreted normally whereas isozymes 1, 3 and 4 were not secreted Also, in the perfused rat hver, monensin treatment has less of an effect on blliary secretion than on secretion of plasma proteins [47] .

Sn-nllarly, in the transport of HLA-DR a-and /3chains [71] , processing of N-linked carbohydrate chains to full endo-H resistance occurs However, with the associated I-chain, processing of both O-and N-linked carbohydrate chains ~s inhibited, and carbohydrate chains remain predornlnantly endo-H susceptible. Here, the processing of membrane-associated proteins that occurs despite a monensln block may reside in intercalary Clsternae that constitute the Golgi apparatus region recently termed medial [118A46]

There is now considerable evidence for monenslnsusceptible compartments in the endocytotic pathway. Transfer of product to secondary lysosomes [83] as well as virus penetration into cultured cells [78] are impaired by monensm. Stein and co-workers [188] have shown that monensin blocks transfernn recycling by causing internahzed hgand to accumulate in the perinuclear regmn, primarily in multivesicular bodies, of the K562 cells used in the study. Based on studies of HRP uptake in rat fibroblasts, Wilcox and co-workers [79] suggested that inhibition of endocytic events may be the consequence of an inhibition of membrane recycling within the cell rather than a direct effect of monensin at the cell surface.

Maxfleld [85] reported that 6 ~tM monensln resulted m an mcrease in internal pH of endocytIc vesicles of cultured mouse fibroblasts from 5 0 to above 6 0 to account for its effects on receptor-mediated endocytosis Similarly, Marsh and co-workers [78] concluded that the lnhlbmon of Semhki Forest virus penetration into cultured cells was the result of this increase in pH of endocyUc vacuoles and lysosomes above pH 6 0, the threshold for fusion activity of viral membranes Monensln has also been shown to inhibit lysosomal degradation of protein by affecting lysosomal pH [90] and to abolish aslaloglycoprotein degradation in cultures of rat hepatocytes through a pH shift in prelysosomal endocytlc vesicles [80] . Using digttal image analysis, Tyco and co-workers [86] showed that monensin raised the pH of endocytic vesicles in cultured human hepatoma cells and caused a hgand-lndependent loss o! receptors

In other studies, monensln did not prevent lnternahzatlon of 35S-labeled proteoglycans by rat ovarian granulosa cells although their lntracellular degradation was completely inhibited [40] . Yet, degradation pathways Involving proteolysls of both dermatin and heparln sulfate and limited endoglycosldlc cleavage of heparln sulfate continued_ These findings, while consistent with an involvement of both acidic and nonacldlc compartments, show that monensm inhibition Is primarily on those processes that normally occur in acidic compartments such as endosomes or lysosomes by raising their pH. Sirmlarly, with isolated hepatocytes, Whittaker et al [89] found no effect of monensin on insulin internalization but, rather, an impairment of ~ts degradation once lnternahzed.

Rustan and co-workers [56] suggested that monensm inhibits both endo-and exo-cytosls by a similar mechanism, namely, disruption of proton gradients Their conclusions were based on studies of rat hepatocytes in which monensm inhibited both secretion of very-lowdensity hpoprotelns, and binding and degradation of aslalofetuin Both secretion and receptor binding were markedly decreased after only 15 rmn of monensin treatment although no effect on protein synthesis was observed. However, secretion was more sensitive [o monensln than endocytosis, suggesting that monensin independently intubits endocytlc and secretory functions although the mechanisms may be similar.

Marnell et al [87] explained the monensin block of the cytotoxic effect of &ptheria toxin on a sirmlar basis Following endocytosls of the toxin, the toxin was assumed to penetrate the membrane of the endosome and enter the cytoplasm in response to an acid environment By neutralizing the ability of endosomes to acidify their interiors, monensin, like the lysosomotroplc amines, was able to block the low pH-dependent dissociation of receptor-hgand complexes and subsequent release of ligands either to the cytoplasm (viruses and toxins) or to lysosomes (endocytosed proteins such as LDL). This in turn would prevent recycling of receptors and membrane and eventually bring endocytosls to a halt due, not necessardy to an inhibition of the uptake processes per se, but perhaps, to blockage of an internal step very similar to that believed to be blocked at the GolN apparatus.

Also consistent with sirmlar modes of monensin mhibition in processing both endocytlc and exocytIc vesicles are findings that a single mutatmn in Chinese hamster ovary cells impaired both Golg¢ apparatus and endosomal functions in parallel Included were the monensin sensitive steps of virus and toxin penetration from endosomes into the cytoplasm and of Golgl apparatus-associated maturation of Smbis virus [81] . The alterations correlated with losses of ATP-dependent vacuole acidification as if the ATPase of endosome and Golg~ apparatus shared a common genetically regulated subunit. Ono et al. [189] studied a monensln-reslstant mutant of mouse Balb/3T3 cells which also proved to be a poor host for either vesicular stomatltis virus or Semllkl Forest Virus multiplication. The mutant cells resistant to monensln, bound virus normally and contained acidic compartments. However, movement of virus from the cell surface to the endosome and lysosome compartments was extremely slow.

Thus, the ablhty of monensln to block processing of endocytic vesicles by making prelysosomal compartments less acidic, suggests a mechanism for perturbation of endocytosls based on its ionophorlc properties The mechanism could be slmxlar to the monensm-medlated exchange of monovalent alkah tons for protons that induces, by osmotic means, the observed swelhng of Golgl apparatus clsternae. Clearly, monensm does not interfere with the uptake and binding of particles at the cell surface. Monensm is ineffective against activities that occur at the cell surface. Hedm and Thyberg [76] showed that uptake of IgG prebound to the cell surface was unaffected by monensln Similar findings have been made in studies of receptor-medmted endocytosls of various other ligands [78, 80, 83] . However, monensm may secondarily affect mternahzation through depletion of monensln-sens~tive receptor sites at the cell surface. This would occur if the cell surface receptors are recycled back into the cell and then blocked in the post-Golgi region by monensin so that they could not return to the cell surface [63, 80, 88] . Thus, monensin inhibition of endocytlc events seems to be at the site of transfer from endocytlc vesicles to lysosomes [40, 63, 76, 93] or, in monensin-sensltive endosomes, inhibition of the dissociation of ligand-receptor complexes [81] 

Most animal cells show a dose-related response to monensm that falls off rapidly at monensm concentrations less than 10 -7 M Consequently, most studies of monensln effects use monensin concentrations of 10 -7 M or higher However, a few reports indicate a cellular response at monensm concentrations less than 10 -7 M For example, receptor capping by lymphocytes Is stimulated by low concentrations of monensin in the range 10 -7 to 10 -9 M and inhibited by monensm concentrations above 10 -7 M [96] Cultured adrenal chromaffin [62] and heart [190] cells also are stimulated by low concentrations of monensln. Though these effects occur at concentrations below the threshold effects for most Golgl apparatus responses, they still presumably result from Increased levels of cytoplasmic sodium due primarily to the ~onophore insertion at the plasma membrane Fig 8 (A) Outer cap cells from a control (nontreated) maize root txp preserved by freeze-subsUtutlon [192] . The form of dlctyosome (D) was normal and slrmlar to that following glutaraldehyde/osmlum tetroxade fixation (B) Same except that root tip was treated for 0 5 h with 10 -s M monen~m before being preserved by freeze-substitution. The trans Clsternae and/or secretory vesicles (arrowheads) were swollen, and mltochondna (M) were condensed (compare with mltochondna of A) W, cell wall, V, central vacuole Swelling of Golgl apparatus Clsternae occurs in a wide range of plant and animal cells (see subsection III-B 2 ) and may be a universal response to monensln poisoning. However, in plants, and to a lesser extent animals, swelhng is influenced by the fixative used to preserve the cells Specifically, morphological evidence of swelhng is less (in animal Golgl apparatus) or nonexistent (in plant Golgl apparatus) when the tissues are fixed in potassium permanganate as compared to fixation in ghitaraldehyde/osmium tetroxade [122] . These effects could be due either to fixation artifacts or to differences between plant and animal Golgl apparatus (e.g, Ref_ 168) This problem was evaluated by comparing the images of Golgl apparatus preserved by various chemical fixatives as well as preservation by freezing and low temperature substitution in acetone and osmium tetroxade. Presumably, the image following freeze sub-stitUtlOn would reflect the true ultrastructure more closely than the image following chemical fixation. The results of freeze substitution in both animal cells [191] and maize root (Figs 8A and B) , show swollen Golgi apparatus c~sternae following monensln exposure in a pattern similar to that observed after glutaraldehyde/ osmium tetroxide fixation [192] . However, using videoenhanced light microscopy and cultured bovine mammary epithelial cells, a marked swelling response to monensln was observed only after the addition of glutaraldehyde fixative to the monensln-incubated cells (Morr6, D J., Mollenhauer, H H., Spring, H., Trendlenberg, M, Morr6, D M. and Kartenbeck, J, unpubhshed data)_ Thus, whether monensin-lnduced swelhng occurs m VlVO or is in response to aldehyde fixations remains an important questiion.

No swelling was observed in Golgl apparatus of protoplasts of carrot cells freshly prepared by digestion of cell walls when exposed to monensin even though such a response was obtained in the same cultures prior to wall dissolution [193] . Similarly, in thin slices of hver incubated in monensin, the Golgl apparatus adjacent to cut edges of tissue slices showed a different swelling response than Golgl apparatus adjacent to the uncut natural surfaces of the lobe (unpublished data) Adjacent to a cut edge, fewer cisternae swelled and those that swelled were only in the most trans positions The basis for such differences is unknown but might, for example, indicate changes in cisternal proton pumping ability, or monensln uptake, in response to changes in the physiological state of the Golgi apparatus brought about by the tissue excision

The mechanism by which monensin interacts with coccldla and rumen rmcroflora is well documented [26,194 201] However, the interaction between monensm and the tissue of the host animal is less well understood even though the chnlcal manifestations of monensln poisoning are well known_ Most striking are differences between the in wtro monensln effect ohserved in cultured plant and animal cells, and plants, and the in vwo effects observed m animals When used at recommended levels, either as a coccidlostat for poultry [1, 202] or for cattle [8, 30, 202] , monensin seldom causes poisoning Nonetheless, misuse of the product, usually from improperly mixed or improperly distributed feed, may cause tOXlCOSiS and death [203] [204] [205] [206] [207] [208] [209] [210] [211] [212] [213] . Horses, ponies and other equine species are particularly sensitive to monensln poisoning [12,209. 214] . The median lethal dosage (LDso) for the horse Js 2-4 mg monensin per kg body weight compared to 50-80 mg per kg body weight for cattle and 200 mg per kg body weight for poultry [12, 214, 215] In mammals, the physical signs of monensln tOXlCOSlS commonly include anorexia, diarrhea, depression, sweating, ataxm, palpitations of the heart, and sudden death following exercise [204, 205, 209, 211, [214] [215] [216] [217] Stiffness of hindquarters and swollen gluteus muscles [210, 211] , elevated pulse rate [211, 217, 218] , and ECG abnormalities [207, 208, 214, 217] also have been reported In fowl, the outstanding signs of monensln tOXlCOSlS are drowsiness, excessive thirst, anorexia, depression and paralysis [205, 212, 213] Marked congestion in a variety of organs also has been noted [205] Severely poisoned birds may die in sternal recombency [205] . Routine clinical tests on serum from horses poisoned by monensm may show abnormally high values for blood urea nitrogen, total billrubin, creatlne kinase, lactate dehydrogenase and aspartate armnotransferase [204,207 209,211] However, chnlcal manifestations are often variable makang interpretation and diagnosis difficult. Moreover, serum levels of sodium, potassium, chlorine, calcium, phosphorus, and urea may remain at near-normal levels following monensin treatment [211] _ In poisoned mammals and fowl, generalized congestion, hemorrhage, and macroscopic injury to striated muscle [210, 211, 214, 219] , spleen [205, 212] , lung [212] , liver [209] , and kidney [209, 211, 212, 220, 221] have been noted. The most consistent macroscopic observation in ponies, cattle, pigs and fowl, has been cardiac myocyte degeneration and vacuollzatlon [150, 208, 217, [219] [220] [221] [222] [223] [224] [225] In animals poisoned with relatively high doses of monensin, an initial condensation of heart mltochondria is often seen (Fig 9) However, with longer exposure times, some mitochondria swell and vacuolate with an almost total loss of matnx substance (Fig 9) _ Intracnstae spaces generally remain unchanged with swelling being restricted to the mitochondrial matrix This is followed by loss or dilution of matnx components and a reduction in size of cnstae so that the rmtochondna appear as empty vacuoles with residual cristae Typically, only some mItochondrla In a particular fiber become swollen and appear as vacuoles, and Fig 9 Section of left ventricle from a rat treated wath a single mtraperitoneal mjecaon of 40 mg monensm per kg body weight Most rmtochondna (M) were either condensed or swollen In swollen rmtochondna, cnstae (arrowheads) were greatly reduced in extent but, otherwise, were of approxamately normal ttuckness Thas dosage level (e_g, 40 mg monensln per kg body wexght) is quite tugh for rats and often resulted m significant generahzed damage to the muscle fiber as dlustrated in the lower part of the rmcrograph However, rmtochondnal swelhng (but not condensauon) may occur as well at lower monensln dosages, even when no fiber damage can be identified ultrastructurally Swelling and vacuohzatlon of mltochondna are progresslve with time whereas rmtochondnal condensation was dose-dependent and often occurred rapidly , a few were normal (arrowhead), but most were rmldly condensed Swollen tmtochondna always appeared randomly distributed through a fiber although significant differences m rmtochondnal swelling were often noted between fibers (e g, see Fig_ 12) these are randomly distributed throughout the fiber (Fig. 10) . We have identified early stages of mitochondrial vacuolation and swelling in both the rat and pony (unpublished data), but these mitochondria were seldom plentiful, suggesting that transition to the vacuolated state, once started, was relatively rapid Non-swollen mitochondria may appear condensed, especially after exposure to high levels of monensin ( Fig. 9 and 10 ). Granulation of mitochondna ( Fig. 11 ) was observed only occasionally and those granules that were present appeared similar to the tncalclum phosphate granules often associated with normal mitochondria [226, 227] . However, whether monensln-lnduced granules contain Ca 2 ~ has not been determined. Granules like these have been observed in lschermc and reperfused hearts and are considered indicative of calcium overload [228] [229] [230] [231] _ Ca 2+ overload may be a potential cause of cell death or cellular dysfunction in ischerma [228, [232] [233] [234] [235] [236] [237] although the effects are reversible if the cell is not too severely damaged [228, 229] Exaggerated matochondrial swelling has also been observed immediately following reperfuslon of hearts rendered lschemlc by occluding blood flow for a few minutes [228, 229] Whether excess mitochondrial Ca 2+ occurs as a result of monensin is not clear Under appropriate condiuons, either monensin or Na + in excess can block Ca 2+ accumulation and promote its release from both mitochondrla and sarcoplasrmc reticulum over a broad range of monensin and Na + concentrations [20, 24, 238] It is probable that the release of Ca 2+ by monensln can occur as the result of an increase in cytoplasmac Na + from a monensin shuttle although monensin (at relatwely high concentrations) has been shown to release Ca 2+ directly m a Na+-lndependent manner from cardiac sarcoplasmic retlculum [24] In all probability, however, monensin is not present in myocytes at high concentrations since swelling of Golgl apparatus cxsternae (a characteristic response to monensin) ~s never observed in these cells. The picture is further comphcated by the fact that Ca 2+ release patterns may vary according to type of cell (e g., white vs red muscle cells [239] ) as well as the availability of extracellular Ca 2+ [110] . Although not related to this report, it may be of some interest to note that both mitochondrlal condensation and granulation are much more intense in X-537A-treated animals (and cultured cells as well) than in comparable animals (or cells) treated with monensin [21] _ The percentage of affected mltochondrla vaned markedly between muscle types and species of animal. Analyses of striated muscle m ponies showed that there was a much greater (50 100-times) hkehhood of finding altered mltochondrla in heart tissues than m diaphragm or appendicular muscle [224] A similar relationship existed in rats except that most swollen mitochondria were m the &aphragm [224] . Some antibioUcs (e g, tiamuhn and avoparcm) may act synergistically with monensln to induce shifts in the distribution of swollen mltochondrIa [240] and other cellular damage. Differences in distribution patterns of swollen mltochondrla also were observed between red and white muscle fibers of the rat diaphragm [224] . Red and white fibers were dffferentmted structurally by size, mltochondrlal content, and Z-band configuration [34, [241] [242] [243] Swollen matochondrm were present in all fiber types when the number of affected rmtochondna was small. However, when large numbers of swollen matochondrm were present, the dxsmbutlon pattern was heavily skewed toward the white muscle fibers (Fig_ 12). A differential effect of monensln was also noted by Van Vleet and co-workers [217] who observed in swine severe damage m the diaphragm, vastus lateralls, sem~tendlnosus, triceps and intercostal muscles: moderate damage m longissimus lumborum muscle, and little (or minimal) damage in tongue. Damage was greatest m muscles containing a high proportion of type 1 fibers These distribution patterns, coupled with the characteristic form of the degenerating rmtochondrla, are not common observations and, therefore, may be strong indtcatots of monensin poisoning in animals_ Swollen rmtochondrla have not been observed in any of the nonmuscle cells of the heart, diaphragm, or appendicular tissues or in hver, adrenal, or kidney cells Thus, monensm adrmnistered to mammals in VlVO tends to induce rmtochondrial changes only m selected tissues and/or types of muscle fibers The mechanisms for these rmtochondrlal changes and reasons for the specificity are not known. These problems are compounded by the fact that matochondnal condensation, but not subsequent swelling and degeneration, occurs in cells exposed topically to monensin.

Whale mitochondnal aberrations are the primary morpholog~c indicators of monensin poisomng in striated muscle, other aspects of muscle ultrastructure may show deterioration [217] or may remain relatively normal even with gross mltochondnal damage [224] . Generalized fiber and cell degeneration may occur following monensin poisoning [211, 217, 223] . The extent of monensin-induced injuries appears to be time and dosage dependent. If death occurs shortly after monensin exposure, there may be httle or no recognizable evidence of pathologtcal change, at least in liver, kidney, and striated muscle. Generalized necrosis appears to occur most often when monensln is administered over long periods of time. Even with single doses of monensin, structural aberrations in striated muscle develop progressively over several days and then regress if the animal survives the imtial insult (unpublished data). We have not observed permanent injury in either striated muscle or liver of monensin-treated rats although such effects have been noted in other animals [211] .

Many of the effects on striated muscle attributed to monensxn occur also with other lonophores Irrespective of their ion specxficlties [21, 244] This implies, again, that Na + is not the direct cause of muscle perturbation but, rather, that the iomc imbalance resulting from the intracellular influx of Na + triggers other cellular responses that lead to the observed perturbations. Monensin is known to both inhibit and promote Ca 2+ accumulation in myocytes depending on the absence or availability, respectively, of external Ca 2 + stores [109, 110] ; alter Na + gradient-dependent Ca 2+ transfer through the basolateral plasma membranes of rat small intestine [245] ; increase myocardial calcium activity [246] , inhibit Ca 2+ accumulation by cardiac macrosomes or cause release of accumulated Ca 2+ stores [24] ; and release Ca 2+ from macrosomes [147, 247] . Digitalis and other cardiac glycosides (which increase myocardial contractility) appear to act by altering intracellular Na + concentration through Inhibition of membrane-bound Na +-, K+-actlvated adenosine triphosphatase which secondarily results in an increase in mtracellular Ca 2+ [see 147] . Calcium ionophores such as lasalocld and A23187 have been suggested as potential probes for studying the effects of calcium imbalance on myocardial function [9] . In retrospect, monensin affects the myocardlum in much the same way as do the calcium lonophores and, in some instances, to an even greater extent [9] . These observations suggest that the xonotropic effects of monensln may be partially indirect; e g., through release of histamine and endogenous amines [4] , or stimulation of synthesis and/or release of prostaglandln [4, 25, 248] . Thus, Na + balance plays an indirect but critical role in modulating myocardial function However, a calcium-independent catecholamine depleting action of monensin in cultured rat pheochromocytoma cells [249] and m bovine adrenal medullary cells and chromaffln granules [250] suggests that monensln may also play a direct role in altering cellular function. Similarly, Sutko and co-wor~kers [251] showed that both monensxn and mgencin produce their effects in guinea pig atria by direct action as well as by releasing catecholamines from tissue stores.

Differences in subcellular responses to monensin, between the whole ammal and isolated cells or organs, have been noted by us as well as by others [252] . Thus, swelling of Golgl apparatus Clsternae observed an cultured cells, tissue shces, and plant roots and stems, is not an aberration characteristic of the cells of animals poisoned by monensln. Lack of Golga apparatus welhng in animals infers that cells from monensin-poisoned ammals are bathed in body fluids containing less than 10 -7 M monensln which is approximately the minimum effective dose of monensin that will cause sweUlng of Golgt apparatus Clsternae in cultured animal cells. Alternatively, lack of vacuolated and/or swollen mitochondna in cultured mammahan cells and plants generally imphes that the vacuolated and/or wollen rnttochondria observed in striated muscle from monensinpoisoned animals are secondary effects of monensin poisoning, perhaps caused by a metabohte of monensin [150, 224] . Alternatively, monensln could affect the synthesis and/or transport of humoral agents (e.g., catacholanunes) which, in turn, would alter muscle homeostasis and lead to the rmtochondnal aberrations observed in striated muscle.

For plants, at concentrations of 10 -s M, monensin was shown to inhibit gernunatlon and growth of ryegrass seedlings [97] . The effects were primarily associated with poor root development and significant reduction of root mass as compared to controls Leaf emergence and leaf mass was only slightly affected. At 10 -4 M monensln, roots often did not emerge from the seed during germination and root mass of seedlings was often near zero. Under these same conditions, shoot mass was reduced about 50% as compared to controls.

Monensin, a monovalent ion-selective ionophore, facilitates the transmembrane exchange of prinopally sodium ions for protons. The outer surface of the lonophore-ion complex as composed largely of nonpolar hydrocarbon, which tmparts a high solubility to the complexes in nonpolar solvents. In biological systems, these complexes are freely soluble In the lipid components of membranes and, presumably, diffuse or shuttle through the membranes from one aqueous membrane interface to the other. The net effect for monensin is a trans-membrane exchange of sodium ions for protons. However, the interaction of an ionophore with biologl-cal membranes, and ItS ionophorlc expression, is highly dependent on the blochemcial configuration of the membrane itself One apparent consequence of this exchange is the neutralization of acidic lntracellular compartments such as the trans Golgi apparatus Clsternae and associated elements, lysosomes, and certain endosomes. This is accompanied by a disruption of trans Golgi apparatus Clsternae and of lysosome and acidic endosome function. At the same time, Golgl apparatus Clsternae appear to swell, presumably due to osmotic uptake of water resulting from the inward movement of ions Monensin effects on Golga apparatus are observed in cells from a wide range of plant and animal species The action of monensin is most often exerted on the trans half of the stacked Clsternae, often near the point of exit of secretory vesicles at the trans face of the stacked cisternae, or, especially at low monensln concentrations or short exposure times, near the rmddle of the stacked cisternae. The effects of monensln are quite rapid in both animal and plant cells; l.e, changes in Golgl apparatus may be observed after only 2-5 min of exposure It is implicit in these observations that the uptake of osmotically active cations IS accompanied by a concornltant efflux of H + and that a net influx of protons would be required to sustain the ionic exchange long enough to account for the swelling of Clsternae observed in electron rmcrographs.

In the Golgi apparatus, late processing events such as terrmnal glycosylation and proteolytlc cleavages are most susceptible to inhibition by monensln. Yet, many incompletely processed molecules may still be secreted via yet poorly understood mechanisms that appear to bypass the Golgi apparatus In endocytosls, monensln does not prevent internalization However, intracellular degradation of internalized ligands may be prevented. It is becormng clear that endocytosls involves both acidic and non-acidic compartments and that monensln inhibits those processes that normally occur in acidic compartments.

Thus, monensln, which is capable of collapsing Na + and H + gradients, has gamed wide-spread acceptance as a tool for studying Golgi apparatus function and for locahzlng and identifying the molecular pathways of subcellular vesicular traffic involving acid compartments. Among its advantages are the low concentrations at which inhibitions are produced (001-1.0 /~M), a minimum of troublesome side effects (e g., little or no change of protein synthesis or ATP levels) and a reversible action. Because the affinity of monensIn for Na + is ten times that for K +, its nearest competitor, monensxn mediates primarily a Na+-H + exchange Monensin has little tendency to bind calcium.

Not only is monensin of importance as an experimental tool, it is of great commercial value as a coccidiostat for poultry and to promote more efficient utilization of feed m cattle The mechanisms by which monensln interact with cocctdla and rumen rmcroflora to achieve these benefits are reasonably well documented. However, the interactions between monensm and the tissues of the host animal are not well understood although the severe toxicological manifestations of monensln poisoning are well known Equine species are particularly susceptible to monensm poisoning, and a common effect of monensln poisoning is vacuolizatton and/or swelling of rmtochondna m striated muscle Other pathological injuries to striated muscle, spleen, lung, hver and kidney also have been noted A con-sIstent observation is cardiac myocyte degeneration as well as vacuohzation Differences m cellular response resulting from exposure to monensln (t e, Golgi apparatus swelling in cultured cells, isolated tissues, and plants vs. mltochondrial swelling m animals fed monensln) suggest that myocardial damage is due either to a monensln metabohte or is a secondary response to some other derivation However, as pointed out by Bergen and Bates [26] , the underlying mode of action of lonophores is on transmembrane ton fluxes which dissipate cation and proton gradients Consequently, some or all of the observed monensin effects m VlVO m animals could be secondary phenomena caused by disruption of normal membrane physiology resulting from altered ion fluxes.

The Role of Membranes in Metabohc Regulation

Polyether Antibiotics -Naturally Occurnng Acid Ionophores

Polyether Antibiotics -Naturally Occurnng Acid Ionopbores

Polyether AntlblOUCS -Naturally Occurnng Acid Ionophores

Polyether Antibiotics -Naturally Occumng Acid Ionophores (West-Icy

Polyether Antibiotics -Naturally Occurring Acid Ionophores

The Role of Membanes m Metabohc Regulalaon

Polyether Antlbtotlcs -Naturally Occurnng Acid Ionophores

Comp Blochem Phys-1o

Protoplasma

Clba Found Symp 92

Ongm and Continuity of Cell Organelles (Relnert

Endocytosls (Paston, I and Wdhngham

Proc. Natl Acad

24th Ann Proc

Equine Medacane and Surgery

Current Vetennary Therapy Food Ammal Practice (Howard

An Atlas of Fine Structure

Calcium Antagomsts and Cardiovascular Disease