INTRODUCTION

TOP ABSTRACT INTRODUCTION BACKGROUND EXPERIMENTAL TECHNIQUES TO... CLONED MAMMALIAN INTRACELLULAR... SECRETORY GRANULE ION CHANNELS... PERSPECTIVES REFERENCES

EXOCRINE AND ENDOCRINE CELLS, such as pancreatic acinar or -cells, are morphologically characterized by the presence of intracellular membrane-bound vesicles or granules that secrete their content into the extracellular milieu in a regulated manner. A key aspect of regulated secretion concerns exocytosis and release of macromolecular secretory products, for instance, digestive proenzymes (the "zymogens") in pancreatic acinar cells or insulin in pancreatic -cells. In this process, secretory granules gather beneath the cell membranes in clusters and lie there waiting until a signal reaching the cell membrane induces the granules to fuse with the plasma membrane (PM) and to discharge their stored macromolecules and/or solute molecules into the cell exterior. The sequence of events preceding and accompanying granule export has been described in great detail by ultrastructural techniques (143). A decisive element of this process is a "fusion-fission" reaction between the membrane of secretory granules and the PM of secretory cells, which may also involve the formation of proteinaceous pores between apposed membranes as the basis for fusion (128).

Up to about 1990, the term "exocytosis" described the export process of membrane-impermeant molecules stored within the secretory granules (6). Exocytosis was viewed as comprising several distinct stages, whose nomenclature had been established on the basis of ultrastructural studies (143). After granule assembly in the Golgi and "translocation" beneath the PM, a stimulus would lead to the "fusion" or juxtaposition of granules and PM, a process that would require an intracellular signal, such as an increase of intracellular Ca²⁺ concentration ([Ca²⁺]). This initial fusion occurred between the PM inner leaflet and granule outer leaflet (the "pentalaminar complex"). Finally, the barrier between granule interior and extracellular medium would break (undergoing "fission") and secretion would ensue.

The ultrastructural studies by Palade (143) led Pollard and coworkers (151) to propose a "chemiosmotic hypothesis of exocytosis," according to which H⁺ and Cl fluxes through granule channels and transporters were energetically coupled to granule fusion and fission via osmotic swelling of the secretory granules. This model was based on experiments in cell groups and isolated granules. Subsequent studies on isolated cells using techniques with a higher temporal and spatial resolution seem to have disproved the hypothesis forwarded by Pollard (see Refs. 40, 234).

The introduction of electrophysiological and optical methods with high temporal and spatial resolution, the identification of Ca²⁺-dependent and -independent proteins participating in the fusion-fission reaction, and the application of kinetic analysis to the process of granular secretion have contributed to a better understanding of regulated vesicular secretion. This allowed the detection of distinct stages of secretion occurring in a time frame of milliseconds. Exocytosis, by definition, now refers to a phase of secretion at which an electrical continuity of granule and PM prevails and the access of the granule content to the cell exterior is not impeded any more by a cellular membrane.

Before exocytosis, the granules undergo a "docking" reaction at specific release sites, which involves the binding of several granule- and PM-associated complementary proteins (see Ref. 101 for review). This is followed by a "priming" reaction, a process requiring metabolic energy and providing competence of docked granules for membrane fusion and exocytosis (41, 128, 215). A specific signal, e.g., an increase of cytosolic [Ca²⁺], stimulates exocytosis, which is then followed by the release of the granule matrix. The release of the granule matrix appears to be triggered by specific signaling molecules, temporally distinct from and also not obligatorily coupled to exocytosis (11, 23).

There is mounting evidence that cation and anion fluxes across granule membranes occur via specific ion channels. This review is not a survey of published literature on known properties and function of intracellular ion channels. It does, however, describe recently cloned ion channels with an intracellular localization, which are putative candidates for secretory granule ion channels and could play a role in exocytosis and release of secretory proteins. An overview is also given of sensitive electrophysiological, optical, and microscopic techniques with a high degree of temporal and spatial resolution and low signal-to-noise ratio that are currently in use to study exocytotic secretion. Finally, this review emphasizes recent biochemical and biophysical studies in pancreatic acinar and insulin-secreting -cells, which provide compelling evidence that ion flux through granule ion channels is required for secretion to occur. This happens by modulating the final steps of the secretory process, namely, exocytosis and/or the release of macromolecular secretory products.

	BACKGROUND

TOP ABSTRACT INTRODUCTION BACKGROUND EXPERIMENTAL TECHNIQUES TO... CLONED MAMMALIAN INTRACELLULAR... SECRETORY GRANULE ION CHANNELS... PERSPECTIVES REFERENCES

Mechanisms of Regulated Protein Secretion

Chemiosmotic Hypothesis of Exocytosis

This model was based on studies in chromaffin cells (146, 151). Reports in pancreatic islets supported Pollard and coworkers' chemiosmotic hypothesis in so far as they observed a Cl dependence of glucose-induced insulin release (140, 184). In subsequent work on permeabilized adrenal medullary cells, Knight and Baker (111) found that high concentrations of Cl could promote release of granular protein in the absence of Ca²⁺. Ca²⁺-dependent release, however, was inhibited by Cl. These data were therefore not consistent with the suggestion of Pollard and coworkers that entry of Cl into chromaffin granules promotes exocytosis (see Ref. 21). Unfortunately, the discrepancies regarding the role of Cl in secretion of chromaffin cells have not been followed up and resolved. The fact remains that chromaffin granules are acidified by the action of a V-ATPase (192), which necessitates a parallel conductance to shunt its electrical current for efficient operation and is mediated by either influx of anions or efflux of cations (12).

In exocrine glands, condensing vacuoles of pancreatic and parotid secretory granules were found to be acidic by quantification of the partition of a permeant weak base by immunoelectron microscopy (141). The internal pH of mature granules, however, may or may not be acidic. In one in vitro study, the intragranular pH of isolated pancreatic zymogen granules (ZG) was estimated to be ~6.5 with the pH-sensitive fluorescent dye 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF) (81). This was confirmed in intact cells with the pH-sensitive fluorescent dyes acridine orange (56, 83, 137), LysoTracker (83), or quinacrine (208). This is also supported by the fact that the 31- and 70-kDa subunits of the V-ATPase are expressed in the membranes of rat pancreatic ZG (167). In contrast, evidence for a V-ATPase activity in secretory granules from exocrine glands is lacking (14, 15). Moreover, other studies in intact cells using the weak base 3-(2,4-dinitroanilino)-3'-amino-N-methyldipropylamine (DAMP) or the pH-sensitive fluorescent dyes acridine orange and LysoTracker red came to the conclusion that the pH of ZG is neutral (141, 185). Consequently, ZG acidification as a prerequisite for a chemiosmotic mechanism of exocytosis is still an unresolved matter.

In 1985, Stanley and Ehrenstein (186) proposed that an initial step in the process of exocytosis of neurosecretory granules is activation of Ca²⁺-dependent K⁺ channels present in granule membranes. In this model, salt influx into the granule is controlled by a K⁺ channel that opens when intracellular [Ca²⁺] is elevated as a result of receptor activation. Electrical coupling to anion pathways results in salt and water influx and osmotic swelling, followed by fusion of the vesicle membrane with the luminal membrane and release of the vesicle contents. This model was used to explain protein secretion in exocrine cells with granules, which may lack an active V-ATPase, such as ZG (79).

According to the models proposed by Pollard and coworkers (151) and by Stanley and Ehrenstein (186), increasing the extragranular tonicity should inhibit secretion (72). Evidence for this effect of osmolarity was, for instance, provided by Knight and Baker (111) in chromaffin cells whose PM had been permeabilized by dielectric breakdown with intense electric fields but also by Fuller et al. (79) in protein secretion studies using isolated pancreatic acini permeabilized with digitonin.

However, subsequent pioneering studies in beige mouse mast cells with electrophysiological (patch clamp) membrane capacitance measurements (40, 234) demonstrated that fusion actually precedes swelling, which proved that osmotic swelling is not required for fusion. Swelling did occur, but it took place after fusion of granules with the PM (40, 234). Both reports speculated that granule swelling was necessary to stabilize and widen the exocytotic pore and that swelling occurred by movement of extracellular small solutes through the exocytotic pore into the granule matrix (40, 234). These observations were taken as strong evidence to fully dismiss the swelling (= chemiosmotic) hypothesis of exocytosis.

The studies by Zimmerberg et al. (234) and Breckenridge and Almers (40) were published at about the same time as interest was developing in the cellular and molecular biology of protein factors mediating membrane fusion in yeast. This shifted the focus toward other mechanisms that might explain the membrane fusion process in eukaryotic systems and led to the identification of ubiquitously expressed and obligatory protein components of cellular fusion events. These include soluble N-ethylmaleimide-sensitive factor (NSF) attachment protein receptor (SNARE) proteins, Sec1/Munc18 homologs (SM proteins), and Rab proteins. These three classes of Ca²⁺-dependent and -independent docking and fusion proteins appear to be universally involved in intracellular fusion reactions, and enormous progress has recently been made in the understanding of their role in vesicle fusion. A thorough and detailed discussion of their components and function is not the scope of this review and can be obtained elsewhere (see, for example, Ref. 101).

Chemiosmotic Hypothesis of Exocytosis: An Obsolete Mechanism?

CFTR Cl Channel

Thus recent studies have led to a more differentiated reevaluation of the chemiosmotic hypothesis of exocytosis. They indicate that a chemiosmotic process may be operative at particular stages during the sequence of events associated with protein secretion in exo- or endocrine cells. Support for the model forwarded by Stanley and Ehrenstein (186) has been obtained from studies on mucin granules, which provide evidence for Ca²⁺-activated K⁺ channels in the membrane of these granules (136). This study implies that Ca²⁺/K⁺ exchange in the granule matrix may induce disaggregation and swelling of granules, possibly because of K⁺ preventing protein condensation and binding of aggregates to granule membranes (51). Further evidence compatible with the importance of granule K⁺ channels for secretion as proposed by Stanley and Ehrenstein (186) has been obtained by Hoy et al. (97) in pancreatic -cells and by Jensen and collaborators (103) in renal juxtaglomerular cells. Data in pancreatic -cells (22, 24) are more in line with a critical role of granule acidification and concomitant Cl fluxes for the process of exocytosis and insulin secretion, as originally implied by Pollard et al. (151).

This new insight into a critical role of granule ion channels and ion fluxes for exocytosis and release of macromolecular secretory products has been made possible because of the recent availability of very sensitive techniques that allow us to study these complex processes in more detail. There is also increasing functional evidence for the presence of intracellular ion channels (for review, see Refs. 196, 227), which belong to several families of cloned ion channel proteins and also include putative intracellular ion channels.

	EXPERIMENTAL TECHNIQUES TO STUDY EXOCYTOSIS AND RELEASE OF SECRETORY PRODUCTS

TOP ABSTRACT INTRODUCTION BACKGROUND EXPERIMENTAL TECHNIQUES TO... CLONED MAMMALIAN INTRACELLULAR... SECRETORY GRANULE ION CHANNELS... PERSPECTIVES REFERENCES

A classic approach to study secretion involves the stimulation of a preparation of isolated cells or cell groups and measurements of the secreted proteins by radioactive metabolic labeling, ELISA, or enzymatic assays. This can also be done after permeabilization of the cells to gain access to the cytosol (79, 111, 226), thus providing some means of controlling intracellular processes and the cytosolic environment. However, these methods require large numbers of cells and offer little temporal resolution.

Three main approaches are currently used to study secretion from single cells: capacitance measurements that trace changes of the cell surface area due to membrane addition by exocytosis of secretory granules; electrochemical detection by oxidation or reduction of released transmitter molecules at the surface of a carbon-fiber electrode placed in close vicinity to the site of release; and optical techniques to monitor release, membrane addition, or pH changes.

Capacitance Measurements of Cell Surface Area

Any membrane that is added by exocytosis is detected as a capacitance increase; endocytosis on the other hand removes membrane from the cell surface and thus reduces the capacitance. This means that one problem associated with C_m measurements is that they only report net changes of membrane surface, which do not necessarily reflect reliable estimates of the rate of secretion. In -cells, this will not influence the measurement of exocytosis because endocytosis proceeds on a relatively slow time scale (66). However, C_m measurements may not be sufficient to independently resolve these processes in cells in which a close temporal coupling of exocytotic and endocytotic events prevails. This concern is of particular importance in pancreatic acinar cells, which secrete at a slower rate and for longer time periods than do neuroendocrine cells. Recently, optical measurements of membrane turnover have been developed, which use membrane-sensitive fluorescent probes and can provide real-time measurements of secretory dynamics (29, 30). This technique, in combination with C_m measurements, allows for the independent evaluation of exocytotic and endocytotic activity and reports spatial information about these processes (182). The most widely used probe is FM1-43, a fluorescent amphipathic styryl dye, which rapidly and reversibly partitions into the outer leaflet of biological membranes and becomes trapped in recycled vesicles on endocytosis. On stimulation, preloaded vesicles release dye into the bathing medium and fluorescence thereby declines (31). A study in rat pancreatic acinar cells selected this approach by monitoring changes of membrane surface area (C_m) in combination with measurements of the membrane turnover using FM1-43 (86). This study revealed that exo- and endocytosis arise within seconds after secretagogue-induced activation and coincide both temporally and spatially (86).

After incorporation of the granule membrane into the apical cell membrane, any granule conductance should contribute to the whole cell conductance (G_m). Studies have aimed at answering the question of whether fusion of individual granules with the PM inserts granule channels into the cell membrane. These studies combined capacitance with single-channel measurements in the whole cell configuration and attempted to correlate temporal changes in C_m and G_m induced by secretory stimuli. Studies carried out in rodent pancreatic acinar cells have yielded controversial results, either claiming a contribution of granular ion channels to increases in G_m after secretory stimuli (126) or not (172). However, the hypothesis tested by Schmid and Schulz (172) and Maruyama et al. (126) that granule exocytosis (C_m) and opening of putative granule ion channels (G_m) should temporally overlap is not valid. The studies by Zimmerberg et al. (234) and Breckenridge and Almers (40) demonstrated that granule swelling (and thus release of secretory products) lags significantly behind granule exocytosis. Similarly, significant delays between C_m and discharge of secreted macromolecules were recently reported in mast cells (225) and pancreatic -cells (23). If the assumption that activation of granule ion channels promotes the release of secreted macromolecules is correct, then an increase of G_m (induced by opening of granule ion channels) will occur after a change in C_m. This change of G_m may be difficult to resolve within the large increase of whole cell G_m (reflecting activities of Cl channels in the PM). One additional problem with the negative study (172) was also the choice of a nonphysiological stimulus [guanosine 5'-O-(3-thiotriphosphate); GTPS], when a physiological Ca²⁺-dependent agonist, such as acetylcholine (ACh) or cholecystokinin (CCK), would have been appropriate.

Amperometric Detection of Released Molecules

Although both capacitance and electrochemical techniques allow measurements of single-vesicle fusion to be performed with a time resolution of milliseconds, these techniques suffer from two major drawbacks. Their spatial resolution is low (as with whole cell C_m measurements or large-diameter carbon fiber electrodes with 5- to 8-µm tip diameter). Moreover, they do not provide information on the steps before membrane fusion or after endocytotic uptake of membrane, because they only measure membrane addition or release, respectively.

Evanescent Wave Microscopy

TIR is based on Snell's law of optics: If light traveling in a dense medium (e.g., a glass coverslip with a high refractive index, n₁) hits a less dense medium (e.g., an aqueous medium of lower refractive index, n₂), beyond a certain "critical angle," _c, the light will undergo TIR. This critical angle depends on the relative refractive indexes of the two media. If the n₂-to-n₁ ratio is very small, the critical angle is shallow (_c = 24.6°) and TIR is readily achieved.

In practice, cells are grown on glass coverslips or transparent materials of high refractive index, and a beam of light, usually from a laser, is optically coupled into the coverslip by a prism or the objective itself. If light approaches the aqueous medium at >_c, it totally reflects into the glass. However, at angles >_c, some of the energy slightly penetrates the aqueous medium as an "evanescent wave," propagating parallel to the interface, which can be derived from Maxwell's equations on the behavior of electromagnetic fields at a dielectric interface (18, 20). An important property of the evanescent wave is that the intensity falls off exponentially away from the coverslip. The "penetration depth" (the distance where the intensity I has decreased to I_o/e, where I_o is intensity at distance o and e is natural logarithm) depends on the incidence angle, wavelength, and polarization of light, as well as the refractive index of the coverslip and the medium. Penetration depths of <100 nm are achieved without difficulty. Thus only fluorophores near the coverslip are excited. TIR-FM illuminates vertical slices with the dimensions of a thin electron microscopy section (<100 nm) as opposed to slices of ~500-800 nm for one- and two-photon confocal systems, respectively. This thin optical sectioning means that the signal-to-noise ratio is much better than with confocal images and cellular photodamage and photobleaching are minimal.

TIR-FM is a complementary approach that can be combined with other microscopy techniques, such as brightfield, epifluorescence, confocal, or atomic force microscopy (AFM) (see Atomic Force Microscopy). TIR-FM applications in cell biology have been expanded to studies of secretion (94, 104, 119, 139, 165, 188, 189, 210) by the recent advent of green fluorescent protein (GFP) of the jellyfish Aequorea victoria for more specific staining of secretory organelles and its cyan, yellow, and red derivatives (CFP, YFP, and RFP, respectively). When linked to granular proteins and expressed in cells, it retains its fluorescence properties and can therefore be used as a granule marker (105, 119, 153). Like acidotropic dyes, e.g., acridine orange or dyes of the LysoTracker LysoSensor families, it is released with the granular contents after membrane fusion such that secretion is observed as a decrease in fluorescence. Recently, pH-dependent GFP mutants have become available (131). Further improvement of TIR-FM has also been accomplished by the recent introduction of new objective lenses and condensers (19).

Atomic Force Microscopy

Jena et al. (102) asked how the ZG morphology associated with kiss and run recycling could be maintained during secretion. They argued that ZG diameter should decrease during amylase secretion and therefore, as a consequence of Laplace's law, ZG surface tension should increase. An increase of ZG surface tension should further enhance enzyme release, which should therefore decrease ZG dimensions until total collapse of the granule. However, full fusion does not occur according to their observations (176). To clarify this issue, they used AMF in combination with confocal microscopy to study the three-dimensional dynamics of isolated ZG during stimulation (102). Exposure of ZG to GTP, but not GDP, increased their height by 15-25% as measured by AFM. A comparable increase in diameter was determined by confocal microscopy. Similar effects were observed with NaF. An active mastoparan analog (Mas7) known to stimulate G_i proteins increased ZG GTPase activity and also induced swelling. Mas7, NaF, and GTP increased vesicle swelling to a similar extent, suggesting that ZG-associated heterotrimeric GTP-binding proteins (G_i3 based on immunoblots of ZG membrane fractions) participate in regulating ZG size. Moreover, the effects of Mas7, NaF, and GTP on ZG size occurred in the presence of KCl, but not when KCl was replaced by cyclamide, suggesting that swelling of ZG may be mediated by ion flux through K⁺ and Cl channels in the granule membrane (102). Interestingly, 100 µM Ca²⁺ or 200 µM EGTA had no effect on ZG size [see Expression of a Ca²⁺-activated and bicarbonate-permeable anion channel (CLCA)]. On the basis of these AFM studies in pancreatic acinar cells and isolated ZG, Jena and coworkers (102) proposed that K⁺ and Cl channels in the granule membrane are required to induce granule swelling during secretion to prevent ZG collapse. Ion fluxes through K⁺ and Cl channels in the granule membrane and osmotic swelling thus appear to maintain granule integrity and morphology as a prerequisite for kiss and run recycling (102, 175, 176).

	CLONED MAMMALIAN INTRACELLULAR ION CHANNELS

TOP ABSTRACT INTRODUCTION BACKGROUND EXPERIMENTAL TECHNIQUES TO... CLONED MAMMALIAN INTRACELLULAR... SECRETORY GRANULE ION CHANNELS... PERSPECTIVES REFERENCES

ClC Cl Channels

ClC-2. A whole cell patch-clamp study on Cl channel activity in pig pancreatic acinar cells provided the first evidence that a ClC channel might function as an intracellular Cl channel (42). The Cl current was activated by strong hyperpolarization and cell swelling, which are biophysical properties reminiscent of ClC-2. This anion current also shared the Cl > I selectivity sequence with other ClC proteins. Immunohistochemical localization with a ClC-2 antibody revealed expression in both apical PM and ZG, suggesting that it may operate as a Cl efflux pathway in pancreatic acinar cells and remain functional in the PM after exocytosis, thus contributing to primary salt secretion. Subsequently, expression of ClC-2 in both rat pancreatic plasma and ZG membranes has been confirmed by immunoblotting (unpublished observations). Functional studies in isolated ZG have recently identified a largely Ca²⁺-independent Cl conductance in isolated ZG that could represent ClC-2 and may contribute to granule swelling [see Expression of a Ca²⁺-activated and bicarbonate-permeable anion channel (CLCA)].

ClC-3. Several studies have suggested that ClC-3 may be the major candidate for the volume-regulated Cl channel, I_Cl-swell (60, 61), but this has not remained uncontested (122). In contrast, a recent study provides compelling evidence that this channel plays a critical role in acidifying synaptic vesicles. Stobrawa et al. (190) reported a phenotype associated with targeted inactivation of the murine clcn3 gene. The mice exhibited postnatal growth retardation, blindness secondary to progressive retinal degeneration, and behavioral abnormalities associated with hippocampal degeneration. Further investigations revealed that ClC-3 is present on synaptic vesicles and participates in their acidification. This occurs through active proton pumping by an electrogenic V-type H⁺ pump. In the absence of a conductive pathway, the generation of an inside-positive potential difference across the vesicle membrane will limit the degree of acidification that can be achieved by the pump. The dissipation of this voltage by an electrical shunt allows for higher rates of proton transport. Indeed, biophysical studies have revealed Cl conductances in many intracellular organelles, including the endoplasmic reticulum (ER), Golgi, vesicles of the exocytotic and endocytotic pathways, lysosomes, and synaptic vesicles (3, 196, 227). Their acidification serves various purposes, such as modulation of enzymatic activities, processing of prohormones, differential sorting of receptors and ligands, endocytosis, and other vesicle trafficking. The electrochemical proton gradient also provides the driving force for the transport of other substances across the vesicle membrane as in the concentrative uptake of neurotransmitters into synaptic vesicles, e.g., of glutamate (127).

Expression of a Clc-3 Cl channel

ClC-5. The Cl channel ClC-5 shares the Cl > I selectivity with other ClC channels. Mutations encountered in Dent's disease, a rare familial disorder, reduce or abolish Cl currents associated with the functional ClC-5 cDNA in the Xenopus laevis oocyte expression system. This disease is associated with low-molecular-weight proteinuria and hypercalciuria; thus ClC-5 may mediate renal protein endocytosis. Indeed, ClC-5 knockout mice exhibiting a targeted disruption of the murine clcn5 gene showed impaired proximal tubule protein absorption and reduced receptor-mediated and fluid-phase endocytosis in proximal tubular cells (150, 220). After budding from the PM, endocytotic vesicles are progressively acidified on their way to the lysosomes, and ClC-5 may provide a rate-limiting anion conductance for efficient endosomal acidification in the proximal tubule (150, 220). However, ClC-5 expression in X. laevis oocytes or mammalian cells elicits strongly outwardly rectifying Cl currents that can be detected only in a voltage range more positive than +20 mV (75). However, such an inside-positive membrane potential is highly unlikely in intracellular compartments. Moreover, ClC-5 expressed in X. laevis oocytes or mammalian cells is inhibited by extracytosolic acidic pH (174). These properties cast doubt on ClC-5 being solely responsible for Cl conductance in renal endosomes because Cl movement into the acidifying compartment would be greatly limited. This suggests that endosomal anion conductance requires additional transport pathways and/or that ClC-5 displays different properties in situ because of the presence of regulatory subunits.

ClC-7. Osteoclastic bone resorption requires extracellular proton accumulation (198). Osteoclasts are multinucleated cells formed by the fusion of mononuclear hematopoietic stem cells belonging to the phagocyte series. When attached to bone, osteoclasts create lacunae, into which protons and acid hydrolases are translocated to digest mineralized bone matrix. Acid secretion is mediated by the fusion of internal vesicles containing V-type H⁺ pumps into the osteoclast surface membranes adjacent to the bone surface, creating the "ruffled border" where bone resorption occurs. The extracellular acidic lacunae require fluxes of counterions, e.g., of Cl through Cl channels for charge compensation. A recent report by Kornak et al. (113) defined the role of the ClC-7 channel protein in osteoclast-mediated bone resorption by studying clcn7-deficient mice. Inactivation of ClC-7 caused a severe phenotype with osteopetrosis (bone petrification) and retinal degeneration. The defects of skeletal morphogenesis could all be explained by impaired osteoclastic bone resorption. Osteoclasts in ClC-7-deficient mice exhibited poorly developed ruffled borders and did not form resorption lacunae. Immunohistochemical staining for ClC-7 in normal osteoclasts demonstrated intracellular staining of late endo-/lysosomal compartments as well as localization along ruffled border membranes similar to the distribution of the V-type H⁺-ATPase, which is consistent with a distribution of ClC-7 in subplasmalemmal secretory vesicles and in the PM. Further functional studies demonstrated a defect in extracellular acidification by osteoclasts at the PM, but no deficient late endo-/lysosomal or lysosomal acidification was observed (113). This indicates that ClC-7 is not solely responsible for the acidification of late endosomes and may also involve other anion channels, such as p62 (171) (see CLIC/p64 Cl Channels). Interestingly, a recent in situ hybridization study showed that ClC-6 and ClC-7 are strongly expressed in mouse pancreatic acinar cells but not in islets (107), suggesting that these channel proteins may also operate as intracellular anion channels in acinar cells, but their intracellular localization and physiological function is currently unknown.

CLIC/p64 Cl Channels

Recently, several related human genes that share homology with the COOH-terminal half of bovine p64 have been cloned. These genes constitute a protein family called CLIC (Cl intracellular channel). The CLIC/p64 superfamily has grown to include p64 itself, the p64-like protein parchorin (138) and five CLIC proteins. The first CLIC member to be identified, CLIC1/NCC27, was originally detected in cell nuclei (214) and was subsequently shown to be enriched in the brush border of proximal tubule cells (211). So far there is no evidence that CLIC2, CLIC3, and CLIC5 are expressed in secretory vesicles (28, 95, 154). CLIC4/huH1 was recently identified (64) as the human homolog of a rat brain protein termed p64H1 (47, 62). CLIC4 mRNA is widely expressed in neuronal and nonneuronal tissues. Although rat brain p64H1, which encodes for a 28.6-kDa protein, colocalized with markers for the ER of transiently transfected rat hippocampal HT-4 cells (62), in another study it was localized to the large dense-core vesicles of rat hippocampal neurons (47).

Indirect immunofluorescence, cell fractionation, and immunoblotting studies have localized native and recombinant CLIC4 proteins both to the cytosol and to intracellular membranes. A similar, very unusual, dual localization is characteristic of many CLIC and p64-related proteins (47, 64, 70, 138, 154, 160, 214). For example, human CLIC4 is enriched in the apical region of proximal tubule cells but partly colocalizes with caveolae in a pancreatic cell line (64). In addition, the mouse homolog of CLIC4, called mc3s5/mtCLIC, was recently localized to the mitochondria and cytoplasm of keratinocytes (70). Such a protein distribution could reflect a shuttling of CLIC4 between the cytoplasm and different endomembrane systems (including secretory vesicles, mitochondrial membranes, and caveolae) and an involvement of the protein in widespread cell biological processes such as membrane trafficking or vesicle transport.

Various electrophysiological studies have shown that several members of this gene family, including bovine p64 (65, 116, 117), CLIC1 (214), CLIC3 (154), and rat brain p64H1/CLIC4 (62), play a role in Cl transport. Cells overexpressing recombinant CLIC4 contain intracellular anion channel activity that is absent from mock-transfected cells (62). There are several alternative ways in which CLIC proteins could be linked to intracellular ion channel activity. They could be channel-forming proteins, they could activate endogenous ion channels (either directly or indirectly), or they could perform both functions. The "ion channel" hypothesis has been tested directly for CLIC1, which forms channels after being incorporated, as a pure recombinant protein, into planar bilayers (212). Although it has yet to be shown that this channel activity in vitro corresponds to an endogenous ion channel, it is clearly possible that CLIC1 and other CLIC proteins might have a direct role as intracellular ion channels. An alternative hypothesis is that some or all of the CLIC proteins are anion channel regulators, rather than (or as well as) ion channels. Several CLIC proteins are also expressed in endomembranes, which are not associated with acidic compartments (62, 64, 211, 214). This suggests that the role of CLIC proteins may not be restricted to its operation as a shunt pathway for electrogenic V-ATPases. However, as long as knockout animals for the various CLIC proteins are not available, the exact function of these proteins remains difficult to establish.

CFTR Cl Channel

Studies using isolated tissue from control and CF submandibular salivary glands (129) showed that -adrenergic agonists were able to stimulate the release of mucin and amylase from control tissue, but the secretory responses of CF tissues were reduced by ~60%. No differences were found between the total mucin pools of control and CF tissues. Similar observations were made in CFTR knockout mice (132). In these studies, release of [¹⁴C]glucosamine-labeled mucins from mouse submandibular glands was monitored in response to activation of the cAMP-mediated second messenger cascade. Although a large, sustained release of mucin was observed in control mouse tissue, the stimulated release of mucin from CFTR knockout mouse tissue was significantly reduced (132).

The regulation of mucin secretion by CFTR has also been assessed in airway epithelial cells. Mergey et al. (130) studied the release of mucin after incorporation of [¹⁴C]glucosamine into mucins. Stimulation of immortalized normal human airway epithelial cells, with either isoproterenol or forskolin, led to a 40% increase in mucin release. In contrast, stimulation of immortalized CF airway cells resulted only in a 1-3% increase in mucin release. The defective cAMP-mediated regulation of mucin secretion was restored after adenovirus-mediated gene transfer of wild-type CFTR to these cells. Granule exocytosis and mucin release in airway epithelial cells were also monitored by release of previously internalized fluorescein isothiocyanate (FITC)-dextran, a fluid-phase endocytosis marker, and by increases in C_m measured with whole cell patch clamp (179). Treatment of non-CF airway cells with a membrane-permeant cAMP analog resulted in a significant increase in C_m, consistent with net incorporation of membrane into the cell surface. Simultaneous release of FITC-dextran in response to the cAMP analog suggested that the cAMP-dependent increase in C_m is partly caused by fusion and incorporation of exocytotic vesicles into the PM. Under identical experimental conditions, stimulation with the cAMP analog had no effect on either C_m or discharge of internalized marker in airway cells derived from a CF patient (179).

Mucin secretion has also been studied in gallbladder epithelium, a model system for mucus secretion by columnar epithelial cells (115). Mucin granules released mucus by merocrine secretion in mouse gallbladder epithelium when examined by transmission electron microscopy. Immunofluorescence microscopic studies revealed intracellular colocalization of mucins and CFTR (115). The data in submandibular, airway, and gallbladder epithelia strongly suggest that CFTR Cl channels are present in mucin granule membranes and support a model according to which CFTR-mediated influx of Cl into the granule enhances secretion by modulating fusion, exocytosis, and/or release of mucus.

KCNQ1 (K_vLQT1) K+ Channel

	SECRETORY GRANULE ION CHANNELS AND REGULATED SECRETION

TOP ABSTRACT INTRODUCTION BACKGROUND EXPERIMENTAL TECHNIQUES TO... CLONED MAMMALIAN INTRACELLULAR... SECRETORY GRANULE ION CHANNELS... PERSPECTIVES REFERENCES

Pancreatic Zymogen Granules

Background. Secretion by the exocrine pancreas is carried out by two morphologically and functionally distinct epithelia, the acini and ducts. Acinar cells secrete a plasmalike primary fluid together with digestive enzymes, which are stored as proenzymes in ZG at the apex of the cells. The primary fluid is modified by the downstream duct cells, which generate the HCO-rich pancreatic juice.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Several candidate ion channel proteins have been identified, which are expressed in the membrane of pancreatic acinar zymogen granules (ZG). A Ca2+-activated anion channel (CLCA) is permeated by HCO and blocked by dithiothreitol (DTT) and H2-4,4'-diisothiocyanostilbene-2,2'-disulfonic acid (H2-DIDS). Activation of this channel may promote enzyme release by increasing the dissolution of the condensed granule contents. Luminal alkalinization may enhance exocytotic membrane retrieval. A DIDS-sensitive ClC-2 Cl channel and a 293B-inhibitable KCNQ1 (KvLQT1) K+ channel are also expressed in ZG membranes. Both channels could play a role in granule swelling to maintain ZG morphology required for "kiss and run" membrane dynamics and promote enzyme release. Two ZG membrane-associated regulatory proteins, ZG-16p, a high-affinity receptor for a dihydropyridine derivative (BZDC-DHP), and a 65-kDa mdr1 gene product, prevent opening of the K+ channel. The 65-kDa mdr1 protein also activates the Cl channel and mediates binding of ATP, sulfonylureas, quinidine and mdr1 antibodies (JSB-1, C219). A nonselective monovalent cation conductance, whose molecular identity is unknown, is blocked by flufenamate and also contributes to hormone-stimulated enzyme secretion. ZG may be acidified by a V-type H+-ATPase, which is expressed in ZG membranes. ZG also represent inositol trisphosphate (IP3)- and cADP ribose-sensitive Ca2+ stores, which may be involved in Ca2+ signaling and/or activation of ZG ion channels. The model leaves the alternatives open that the ion channels are localized in the same or distinct populations of ZG. See text for further details.

Expression of a Ca²+-activated and HCO-permeable anion channel (CLCA). A member of the CLCA channel family could be a potential candidate for a ZG anion conductive pathway, which is directly activated by micromolar concentrations of Ca²⁺. CLCA proteins currently include at least 10 isoforms from different species (76, 77). These putative anion channels are gated by Ca²⁺ and mostly expressed in the PM of epithelial tissues. When heterologously expressed in HEK 293 cells, all isoforms investigated (hCLCA1, hCLCA2, mCLCA1) are associated with the appearance of a Ca²⁺-sensitive anion conductance with an I > Cl selectivity profile, a linear current-voltage relationship, and a 25-pS single-channel conductance, which can be blocked by DIDS or the reducing agent dithiothreitol (DTT) (50). So far there had been no evidence to suggest an intracellular localization of CLCA channels, but using in situ hybridization techniques, Gruber et al. (90) demonstrated the expression of CLCA mRNA in mouse pancreatic acini.