Cross-talk between calpain and caspase-3/-7 in cisplatin-induced apoptosis of melanoma cells: a major role of calpain inhibition in cell death protection and p53 status.

The contribution of different proteolytic systems, in particular calpains and effector caspases, in apoptotic cell death is still controversial. In this paper, we show that during cisplatin-induced apoptosis of human metastatic melanoma cells, calpain activation, as measured in intact cells by two different fluorescent substrates, is an early event, taking place well before caspase-3/-7 activation, and progressively increasing during 48 h of treatment. Such activation appears to be independent from any intracellular calcium imbalance; in fact, an increase of cytosolic calcium along with emptying of the reticular stores occur only at very late stages, uniquely in frankly apoptotic, detached cells. Calpain activation proves to be an early and crucial event in the apoptotic machinery, as demonstrated by the significant protection of cell death in samples co-treated with the calpain inhibitors, MDL 28170, calpeptin and PD 150606, where a variable but significant reduction of both caspase-3/-7 activity and cell detachment is observed. Consistently, such a protective effect can be at least partially due to the impairment of cisplatin-induced p53 activation, occurring early in committed, preapoptotic cells. Furthermore, in late apoptotic cells, calpain activity is also responsible for the formation of a novel p53 proteolytic fragment (approximately 26 kDa), whose function is so far to be elucidated.

Inhibition of store-dependent capacitative Ca2+ influx by unsaturated fatty acids.

The effects of the unsaturated fatty acids, arachidonic and oleic acid, on the influx of Ca2+ activated by depletion of intracellular stores with thapsigargin were investigated in various cell types. By using a Ca2+ free/Ca2+ reintroduction protocol, we observed that arachidonic acid (2 to 5 microM) inhibited thapsigargin-induced rises in cytosolic free Ca2+ ([Ca2+]i) in Ehrlich tumor cells, Jurkat T lymphocytes, rat thymocytes, and Friend erythroleukemia and PC12 rat pheochromocytoma cells. This effect was attributed to the inhibition of Ca2+ entry, since arachidonate also inhibited thapsigargin-stimulated unidirectional entry of the Ca2+ surrogates Ba2+ and Mn2+. In Ehrlich cells, the IC50 for arachidonic and oleic acid was 1.2 and 1.8 microM, respectively. The inhibition appeared to depend on the ratio [fatty acid]/[cells] rather than on the absolute fatty acid concentration. Experiments with [3H]-oleic acid revealed that the inhibitory activity was not correlated with cell internalisation and metabolism of the fatty acid. The inhibition was reverted by removal of the fatty acid bound to cell membrane by fatty acid-free albumin treatment. The unsaturated fatty acids had no effect on ATP/ADP cell levels and plasma membrane potential. Pharmacological evidence indicated that cell phosphorylation/dephosphorylation events, and pertussis toxin-sensitive G proteins were not involved. Other amphipathic lipophilic compounds, i.e. 2-bromopalmitic acid, retinoic acid, sphingosine, and dihydrosphingosine, mimicked arachidonic/oleic acid as they inhibited thapsigargin-stimulated Ca2+ influx in an albumin-reversible fashion. These results suggest that physiologically relevant (unsaturated) fatty acids can inhibit capacitative Ca2+ influx possibly because they intercalate into the plasma membrane and directly affect the activity of the channels involved.

Measurement of mitochondrial and non-mitochondrial Ca2+ in isolated intact hepatocytes: a critical re-evaluation of the use of mitochondrial inhibitors.

Isolated rat hepatocytes treated with mitochondrial inhibitors FCCP or antimycin A release discrete amounts of Ca2+ in a Ca(2+)-free extracellular medium as revealed by changes in the absorbance of the Ca2+ indicator arsenazo III. The process is completed in 2 min and the amount of Ca2+ released is not affected by the type of the mitochondrial poison employed. The subsequent treatment with the cation ionophore A23187 causes a further release of Ca2+ that does not appear related to the specificity of the previous treatment with FCCP or antimycin A. Both FCCP and antimycin A cause a progressive loss of cellular ATP associated with a decrease in the ATP/ADP ratio from 6 to 2-1.5. However, this decrease does not significantly prevent 45Ca2+ accumulation in isolated liver microsomes. Moreover, the decrease of the ATP/ADP ratio to 1, does not promote a significant release of 45Ca2+ from 45Ca(2+)-preloaded microsomes. Finally, experiments with Fura-2-loaded hepatocytes reveal that agents specifically releasing Ca2+ from non-mitochondrial stores (vasopressin and 2,5-di-tert-butyl-1-4-benzohydroquinone) are still able to increase the cytosolic Ca2+ concentration in FCCP-treated cells. Taken together, these findings demonstrate that, in freshly isolated hepatocytes, FCCP specifically releases Ca2+ from mitochondrial stores without significantly affecting active Ca2+ sequestration in other cellular pools. For these reasons, FCCP can be used to release and quantitate mitochondrial Ca2+ in liver cells.

Biochemical and functional heterogeneity of rat cerebrum microsomal membranes in relation to SERCA Ca(2+)-ATPases and Ca2+ release channels.

Rat cerebrum microsomes were subfractionated on isopycnic linear sucrose (20-42%) density gradients. The Ca2+ loading/release properties and the distribution of intracellular Ca2+ store channels, inositol 1,4,5-trisphosphate (IP3) receptor and ryanodine (Ry) receptor, and SERCA pumps, were monitored in each subfraction by ligand binding and 45Ca2+ loading/release assays. Three different classes of vesicles were identified: (i) heavy density vesicles with high content of Ry receptors and Ca2+ pumps and high thapsigargin (TG)-sensitivity of Ca2+ loading; (ii) intermediate sucrose density vesicles with high content of IP3 receptor, high IP(S)3-sensitivity of Ca2+ loading and low content of Ry receptors; and (iii) light sucrose density vesicles with high content of Ry receptors, low content of IP3 receptors and low content of SERCA pumps highly sensitive to TG. Isolation of molecularly heterogeneous rat cerebrum microsomes and identification of specific Ca2+ loading/release properties support the presence of multiple, potentially active, heterogeneous rapidly exchanging Ca2+ stores in rat cerebrum.

Ca(2+)-dependent and independent mitochondrial damage in hepatocellular injury.

The alterations of mitochondrial membrane potential during the development of irreversible cell damage were investigated by measuring rhodamine-123 uptake and distribution in primary cultures as well as in suspensions of rat hepatocytes exposed to different toxic agents. Direct and indirect mechanisms of mitochondrial damage have been identified and a role for Ca2+ in the development of this type of injury by selected compounds was assessed by using extracellular as well as intracellular Ca2+ chelators. In addition, mitochondrial uncoupling by carbonylcyanide-m-chloro-phenylhydrazone (CCCP) resulted in a marked depletion of cellular ATP that was followed by an increase in cytosolic Ca2+ concentration, immediately preceding cell death. These results support the existence of a close relationship linking, in a sort of reverberating circuit, the occurrence of mitochondrial dysfunction and the alterations in cellular Ca2+ homeostasis during hepatocyte injury.

Calcium pools in Ehrlich carcinoma cells. A major, high affinity Ca2+ pool is sensitive to both inositol 1,4,5-trisphosphate and thapsigargin.

To investigate the presence and the size of different non-mitochondrial Ca2+ pools of Ehrlich ascites tumor cells (EATCs), digitonin-permeabilized cells were allowed to accumulate Ca2+ in the presence of mitochondrial inhibitors and treated with the reticular Ca(2+)-ATPase inhibitor thapsigargin, IP3 and the Ca2+ ionophore A23187. Emptying of thapsigargin-sensitive Ca2+ stores prevented any Ca2+ release by IP3, and, after IP3 addition, little or no Ca2+ was released by thapsigargin. In both instances, a further Ca2+ release was accomplished by A23187. The IP3-thapsigargin-sensitive pool and the residual A23187-sensitive one corresponded to approximately 60 and 37% of non-mitochondrial stored Ca2+, respectively. In intact EATCs, IP3-dependent agonists and thapsigargin discharged Ca2+ pools almost completely overlapping, and A32187 released a minor residual Ca2+ pool. The IP3-insensitive pool appeared to have a relatively low affinity for Ca2+ (below 600 nM). The high affinity, IP3-sensitive Ca2+ pool was discharged in a 'quantal' manner following step additions of sub maximal [IP3], and the IP3-induced fractional Ca2+ release was more marked at higher concentrations of stored (luminal) Ca2+, The IP3-sensitive Ca2+ pool appeared to be devoid of the Ca(2+)-activated Ca2+ release channel since caffeine did not released any Ca2+ in intact and permeabilized EATCs, and Western blot analyses of EATC microsomal membranes failed to detect any known ryanodine receptor isoform.

Fatty acyl-CoA esters induce calcium release from terminal cisternae of skeletal muscle.

The effect of palmitoyl-CoA (PCoA) on Ca2+ fluxes in unfractionated SR, longitudinal tubules (LSR) and terminal cisternae (TC) subfractions, obtained from rabbit fast-twitch skeletal muscles, was investigated. After MgATP-dependent Ca2+ preloading, PCoA released Ca2+ from unfractionated SR and TC, but not from LSR. Both the extent and the rate of PCoA-induced Ca2+ release from TC were increased in a dose-dependent manner, the half-maximal effect being attained at [PCoA] of approximately 6 microM. Ruthenium red, a Ca2+ release channel blocker, completely inhibited PCoA-induced Ca2+ release, whereas caffeine, a Ca2+ release channel agonist, depleted TC of Ca2+ and prevented the PCoA action. Scatchard plot analysis of [3H]-ryanodine binding showed that PCoA increased the affinity without affecting Bmax. The action of PCoA was mimicked by a nonhydrolysable analog. The present results indicate that PCoA interacts and opens the Ca2+ release channel (ryanodine receptor) of TC and that the mechanism of action involves binding rather than hydrolysis.

Caffeine releases a glucose-primed endoplasmic reticulum Ca2+ pool in the insulin secreting cell line INS-1.

Caffeine mobilized an intracellular Ca2+ pool in intact fura-2-loaded INS-1 cells in suspension exposed to high (16 mM) [glucose], while a minor effect was observed with low (2 mM) [glucose]. Cells were kept in a medium containing diaxozide or no Ca2+ to prevent the influx of extracellular Ca2+. The caffeine-sensitive intracellular Ca2+ pool was within the endoplasmic reticulum since it was depleted by the inhibitor of the reticular Ca2+ pumps thapsigargin and the InsP3-dependent agonist carbachol. No effect of caffeine was observed in the parent glucose-insensitive RINmF5 cells. In microsomes from INS-1 but not RINmF5 cells, the type 2 ryanodine receptor was present as revealed by Western blotting. It was concluded that the endoplasmic reticulum of INS-1 cells possesses caffeine-sensitive type 2 ryanodine receptors Ca2+ channels.

Inhibition of glucuronidation by an acyl-CoA-mediated indirect mechanism.

The mechanism of the inhibition of glucuronidation by long-chain fatty acyl-CoAs was studied in rat liver microsomal membranes and in isolated hepatocytes. Palmitoyl- and oleoyl-CoA did not affect p-nitrophenol UDP-glucuronosyltransferase activity in native microsomes but were inhibitory in permeabilised vesicles. The extent of inhibition was dependent on the effectiveness of permeabilisation and was constant in time in fully permeabilised microsomes. Fatty acyl-CoAs mobilised calcium from calcium-loaded microsomes. Elevation of the intracellular acyl-CoA level by the addition of palmitate or oleate inhibited the glucuronidation of p-nitrophenol in isolated hepatocytes. This effect could be abolished by emptying the intracellular calcium stores. Therefore, it is concluded that fatty acyl-CoAs inhibit glucuronidation indirectly, presumably via calcium mobilisation.

Alterations of hepatocyte Ca2+ homeostasis by triethylated lead (Et3Pb+): are they correlated with cytotoxicity?

Isolated rat hepatocytes were used to investigate the biochemical mechanisms of toxicity of triethyllead (Et3Pb+), a highly neurotoxic degradation product of the antiknocking petrol additive tetraethyllead. As early as 5 min from the addition of 50 microM Et3Pb+ to hepatocyte suspensions a decrease of mitochondrial membrane potential and of the capacity of mitochondria and microsomes to retain Ca2+ occurred. A dose-dependent release of mitochondrial Ca2+ as well as an inhibition of microsomal Ca(2+)-ATPase activity were also evident when Et3Pb+ (from 2.5 microM up to 50 microM) was added to, respectively, isolated liver mitochondria and microsomes. Further experiments using hepatocytes loaded with the Ca2+ indicator Fura-2AM demonstrate that 1 min from addition of Et3Pb+ the cytosolic free Ca2+ levels increased by about 3-fold. High affinity plasma membrane Ca(2+)-ATPase activity was also significantly inhibited in hepatocytes treated with Et3Pb+, suggesting that an impairement of the mechanisms controlling the efflux of extracellular Ca2+ was concomitantly involved in the rise in cytosolic Ca2+ concentration. The increase in the cytosolic Ca2+ levels caused by Et3Pb+ was followed by a rapid decline of cell viability. However, the addition of EGTA or of the intracellular Ca2+ chelator BAPTA/AM did not affect either the time-course or the extent of cytotoxicity. Conversely, fructose, a glycolytic substrate that was able to support ATP production, prevented hepatocyte death. Thus, the depletion of cellular energy stores rather than the increase in cytosolic Ca2+ appears to be the mechanism by which Et3Pb+ causes irreversible injury in isolated hepatocytes.

The KSR2-calcineurin complex regulates STIM1-ORAI1 dynamics and store-operated calcium entry (SOCE).

Store-operated calcium entry (SOCE) is the predominant Ca(2+) entry mechanism in nonexcitable cells and controls a variety of physiological and pathological processes. Although significant progress has been made in identifying the components required for SOCE, the molecular mechanisms underlying it are elusive. The present study provides evidence for a direct involvement of kinase suppressor of Ras 2 (KSR2) in SOCE. Using lymphocytes and fibroblasts from ksr2(-/-) mice and shKSR2-depleted cells, we find that KSR2 is critical for the elevation of cytosolic Ca(2+) concentration. Specifically, our results show that although it is dispensable for Ca(2+)-store depletion, KSR2 is required for optimal calcium entry. We observe that KSR2 deficiency affects stromal interaction molecule 1 (STIM1)/ORAI1 puncta formation, which is correlated with cytoskeleton disorganization. Of interest, we find that KSR2-associated calcineurin is crucial for SOCE. Blocking calcineurin activity impairs STIM1/ORAI1 puncta-like formation and cytoskeleton organization. In addition, we observe that calcineurin activity and its role in SOCE are both KSR2 dependent.

Endoplasmic reticulum stress underlying the pro-apoptotic effect of epigallocatechin gallate in mouse hepatoma cells.

It has been recently reported that tea flavanols, including epigallocatechin gallate (EGCG), efficiently inhibit glucosidase II in liver microsomes. Since glucosidase II plays a central role in glycoprotein processing and quality control in the endoplasmic reticulum we investigated the possible contribution of endoplasmic reticulum stress and unfolded protein response (UPR) to the pro-apoptotic activity of EGCG in mouse hepatoma cells. The enzyme activity measurements using 4-methylumbelliferyl-alpha-d-glucopyranoside substrate confirmed the inhibition of glucosidase II in intact and alamethicin-permeabilized cells. EGCG treatment caused a progressive elevation of apoptotic activity as assessed by annexin staining. The induction of CHOP/GADD153, the cleavage of procaspase-12 and the increasing phosphorylation of eIF2alpha were revealed in these cells by Western blot analysis while the induction of endoplasmic reticulum chaperones and foldases was not observed. Time- and concentration-dependent depletion of the endoplasmic reticulum calcium stores was also demonstrated in the EGCG-treated cells by single-cell fluorescent detection. The massive alterations in the endoplasmic reticulum morphology revealed by fluorescent microscopy further supported the development of UPR. Collectively, our results indicate that EGCG interferes with protein processing in the endoplasmic reticulum presumably due to inhibition of glucosidase II and that the stress induces an incomplete unfolded protein response with dominantly pro-apoptotic components.

Evidence that stimulation of plasma-membrane Ca2+ inflow is an early action of glucagon and dibutyryl cyclic AMP in rat hepatocytes.

The ability of glucagon (1 nM) and of dibutyryl cyclic AMP (50 microM) to increase cytosolic free Ca2+ concentration ([Ca2+]i) in Fura-loaded rat hepatocytes was examined in a system wherein Ca2+ inflow was induced by the re-admission of excess Ca2+ to a nominally Ca(2+)-free medium. An increase in [Ca2+]i did not occur in the absence of either agonist, but did so after co-addition of either agonist with Ca2+. Increasing the time between addition of dibutyryl cyclic AMP (or of glucagon) and Ca2+ led to increases in [Ca2+]i; half-maximal and maximal increases were observed at 0 s (i.e. at co-addition) and 5-7 s respectively. Dibutyryl cyclic AMP and Ca2+ each exhibited a concentration-dependence when their respective concentrations were changed for a fixed time interval between additions. Half-maximal and maximal effects were obtained with 30 microM and 50 microM dibutyryl cyclic AMP and with 0.5 mM and approx. 1 mM Ca2+ respectively. The data demonstrate an early action of glucagon and dibutyryl cyclic AMP on [Ca2+]i. It is argued that the agonist-induced rise in [Ca2+]i results from an increase in plasma-membrane Ca2+ inflow, an effect that appears to occur much earlier than that on mobilization of internal stores of Ca2+.

Liver glucose-6-phosphatase activity is not modulated by physiological intracellular Ca2+ concentrations.

1. In the presence of MgATP and increasing amounts of added Ca2+, isolated liver microsomal vesicles accumulate approx. 10 nmol of Ca2+/mg of protein and buffer ambient free Ca2+ at increasing concentrations (0.22-10.9 microM). Under these experimental conditions, microsomal glucose-6-phosphatase activity is unaffected by the concentration of extravesicular free Ca2+. 2. Different levels of intravesicular Ca2+ were obtained by treating microsomes with the Ca2+ ionophore A23187 and by stimulating active microsomal Ca2+ accumulation with Pi (3 mM). In both instances, microsomal glucose-6-phosphatase activity is unaffected by the level of intravesicular Ca2+.

Unsaturated fatty acids mobilize intracellular calcium independent of IP3 generation and VIA insertion at the plasma membrane.

Addition of oleic and arachidonic acids to Ehrlich ascites tumor cells mobilizes Ca2+ from the same intracellular pool as that mobilized by thapsigargin. Such mobilization occurs in the presence of the phospholipase C inhibitor U73122 as well as in cells treated with pertussis toxin. Co-addition of fatty acids and thapsigargin leads to initial rates of Ca2+ mobilization much greater than that induced by either compound alone. The responses induced by the fatty acids are observed also with other lipophiles like sphingosine, bromo-palmitate and the Ca2+ influx inhibitor econazole; all responses are rapidly reversed by addition of bovine serum albumin. Many of the above effects of fatty acids are observed also in Jurkat T lymphocytes and Friend erythroleukemia cells. The experiments provide evidence of lipid-induced plasma membrane perturbations that influence intracellular Ca2+ mobilization independent of the generation of currently known second messengers.

On the mechanism of action of econazole, the capacitative calcium inflow blocker.

The ability of bovine serum albumin to reverse the inhibitory action of econazole and the unsaturated fatty acid oleate on store-dependent Ca2+ inflow was examined in Ehrlich ascites tumour cells. We report that inhibition of Ca2+ inflow by both compounds is reversed immediately upon addition of bovine serum albumin. It is concluded that the inhibitory action of econazole resembles that of unsaturated fatty acids. The mechanism appears to be one pertaining to nonspecific events at the plasma membrane, possibly involving alterations in plasma membrane fluidity/structure.

MgATP-dependent accumulation of calcium ions and inorganic phosphate in a liver reticular pool.

1. MgATP-dependent Ca2+ uptake by rat liver microsomal preparations and permeabilized hepatocytes was measured in the presence or absence of Pi. 2. Monitoring of free Ca2+ in incubation systems with a Ca2+ electrode in the presence of Pi (2-7 mM) revealed a biphasic Ca2+ uptake, with the onset of a second, Pi-dependent, Ca2+ accumulation. 3. Increasing Pi concentrations (up to 10 mM) caused a progressive enlargement of 45Ca2(+)-loading capacity of microsomal fractions. 4. As a result of Pi stimulation of active Ca2+ uptake, [32P]Pi and 45Ca2+ were co-accumulated. 5. Experiments with permeabilized hepatocytes revealed that the amount of Ca2+ releasable by myo-inositol 1,4,5-trisphosphate is unaffected by Pi.

Physiological concentrations of inorganic phosphate affect MgATP-dependent Ca2+ storage and inositol trisphosphate-induced Ca2+ efflux in microsomal vesicles from non-hepatic cells.

1. MgATP-dependent 45Ca2+ uptake by microsomes obtained from various non-hepatic tissues, namely rat brain, rat solid Morris hepatoma 3924A and human platelets, was measured in the presence of P(i) at low, cytosol-like, concentrations. 2. Increasing P(i) concentrations (0.5-3 mM) caused a progressive enlargement of the 45Ca(2+)-storage capacity of all the microsomal fractions. 3. As a result of P(i) stimulation of Ca2+ uptake, 45Ca2+ and [32P]P(i) were co-accumulated by the three microsomal fractions. 4. The time course for 45Ca2+ and [32P]P(i) accumulation in brain microsomes revealed a biphasic 45Ca2+ uptake: a rapid phase was followed by a second, slower, phase, which depended on the presence of P(i). During the P(i)-dependent phase, the uptake of 45Ca2+ was paralleled by the uptake of [32P]Pi. 5. The passive efflux of Ca2+ was paralleled by the efflux of P(i) and vice versa. In fact, the inhibition of active Ca2+ uptake by excess EGTA, or lowering the P(i) concentration of the incubation system by dilution, caused the release of 45Ca2+ and [32P]P(i) from 45Ca2+ or [32P]P(i) pre-loaded brain microsomes. The Ca2+ ionophore A23187 also released 45Ca2+ and [32P]P(i). 6. Ca2+ efflux by A23187 was rapid (t 1/2 approx. 2 s) and independent of the extent of intravesicular Ca2+ loading, which indicates that Ca2+ and P(i) do not form intravesicular insoluble complexes. 7. The progressive increase in Ca2+ accumulation, depending on P(i) stimulation, resulted in a proportional increase in the amount of Ca2+ releasable by InsP3 in the three non-hepatic microsomal fractions and in digitonin-permeabilized platelets. 8. Concomitantly to Ca2+, microsomal P(i) was also released by InsP3.

Modulation of Ca2+ influx dependent on store depletion by intracellular adenine-guanine nucleotide levels.

The effects of a number of metabolic inhibitors on the influx of Ca2+ activated by stimulation of receptors coupled to inositol 1,4,5-trisphosphate generation or by depletion of intracellular Ca2+ stores with thapsigargin were investigated in four different cell types: Ehrlich ascites tumor cells, Jurkat and HeLa cell lines, and rat hepatocytes. Independently of their chemical structure and site of inhibition, all of these metabolic poisons markedly inhibited Ca2+ influx without significantly affecting Ca2+ release. This inhibition was not due to membrane potential depolarization or to alteration in cytosolic pH but appeared correlated to a drop in the cellular concentration of ATP. The decreases in cellular [ATP] were paralleled by decreases in [GTP] and by increases in [ADP] and [GDP]. The reduction in ATP level necessary to drastically reduce Ca2+ influx was quite small, e.g. a 50% inhibition for a 5% reduction in [ATP], thus within the range of fluctuation presumably occurring under physiological conditions. We suggest that changes in the adenine or guanine nucleotide concentrations may represent an important modulatory mechanism of Ca2+ influx activated by store depletion.

Evidence for the intraluminal positioning of p-nitrophenol UDP-glucuronosyltransferase activity in rat liver microsomal vesicles.

Addition of p-nitrophenol and UDP-glucuronic acid to rat hepatic microsomes enhanced the MgATP-stimulated Ca2+ sequestration. This stimulatory effect was more explicit in the presence of the activator of glucuronidation, UDP-N-acetylglucosamine. The stimulation of Ca2+ uptake was dependent on the p-nitrophenol concentration and showed a good correlation with the rate of p-nitrophenol glucuronidation. The stimulation of Ca2+ sequestration was probably due to its coaccumulation with the intraluminar Pi originated during glucuronidation. The increase in extravesicular osmolarity due to the addition of UDP-glucuronic acid to microsomes resuspended in an hyposmotic medium caused a rapid and prolonged shrinking as revealed by light-scattering measurements. This indicates a poor permeability of microsomal membrane to UDP-glucuronic acid. The subsequent addition of the pore-forming compound alamethicin resulted in an immediate swelling of vesicles indicating a rapid entry of UDP-glucuronic acid. Alamethicin also caused an about 15-fold increase in p-nitrophenol UDP-glucuronosyltransferase activity. These results support the hypothesis of the intravesicular compartmentation of the microsomal UDP-glucuronosyltransferase catalytic site.