Perfringolysin O, a cholesterol-binding cytolysin, as a probe for lipid rafts

Gaining an understanding of the structural and functional roles of cholesterol in membrane lipid rafts is a critical issue in studies on cellular signaling and because of the possible involvement of lipid rafts in various diseases. We have focused on the potential of perfringolysin O (theta-toxin), a cholesterol-binding cytolysin produced by Clostridium perfringens, as a probe for studies on membrane cholesterol. We prepared a protease-nicked and biotinylated derivative of perfringolysin O (BCtheta) that binds selectively to cholesterol in cholesterol-rich microdomains of cell membranes without causing membrane lesions. Since the domains fulfill the criteria of lipid rafts, BCtheta can be used to detect cholesterol-rich lipid rafts. This is in marked contrast to filipin, another cholesterol-binding reagent, which binds indiscriminately to cell cholesterol. Using BCtheta, we are now searching for molecules that localize specifically in cholesterol-rich lipid rafts. Recently, we demonstrated that the C-terminal domain of perfringolysin O, domain 4 (D4), possesses the same binding characteristics as BCtheta. BIAcore analysis showed that D4 binds specifically to cholesterol with the same binding affinity as the full-size toxin. Cell-bound D4 is recovered predominantly from detergent-insoluble, low-density membrane fractions where raft markers, such as cholesterol, flotillin and Src family kinases, are enriched, indicating that D4 also binds selectively to lipid rafts. Furthermore, a green fluorescent protein-D4 fusion protein (GFP-D4) was revealed to be useful for real-time monitoring of cholesterol in lipid rafts in the plasma membrane. In addition, the expression of GFP-D4 in the cytoplasm might allow the investigations of intracellular trafficking of lipid rafts. The simultaneous visualization of lipid rafts in plasma membranes and inside cells might help in gaining a total understanding of the dynamic behavior of lipid rafts.     

Cross-linking of plasmalemmal cholesterol in lymphocytes induces capping, membrane shedding, and endocytosis through coated pits.

By use of a nicked and biotinylated perfringolysin O (BCtheta), which binds to cholesterol specifically, we studied consequences of cross-linking cholesterol in lymphocytes. When bound with BCtheta and then with labeled avidin or streptavidin, capping occurred in most cells within 30 min at 37 degrees C. It was inhibited by cytochalasin D or NaN3, but not by nocodazole. When BCtheta-cholesterol was capped, Thy-1 and transferrin receptor, a GPI-anchored protein and a transmembrane protein, respectively, remained evenly distributed. By fluorescence and electron microscopy, a cluster of small vesicles bound with BCtheta were observed in the cap. They were then shed in the medium or internalized through coated pits. The result indicates that cross-linking of cholesterol in lymphocytes induces capping, but does not affect distribution of membrane proteins, and that the capped cholesterol molecules are either shed as vesicles or endocytosed.     

Alteration of human erythrocyte plasma membranes by perfringolysin O as revealed by freeze-fracture electron microscopy. Studies on Clostridium perfringens exotoxins V.

When human erythrocyte membranes were treated with perfringolysin O (Clostridium perfringens theta-toxin) and examined by electron microscopy after freeze-fracture, two ultrastructural alterations were observed in fracture faces of membrane. (1) A random aggregation of intramembranous particles was seen in the fracture face of the protoplasmic half (PF face) of all membranes treated with the toxin, even if at a low concentration (40 hemolytic units/ml). On the other hand, the aggregation in the fracture face of the exoplasmic half (EF face) was observed only in membranes treated with a high concentration (3300 hemolytic units/ml) for 2 h. (2) Round protrusions and "cavities" with 30 nm in diameter were visible in EF and PF faces of membranes treated with a high concentration, respectively. These structures were always protruded toward cytoplasmic side, but did not appear to form holes through the membrane. Ring and arc shaped structures with a dark center of 26 nm and a distinct border of 5 nm in width were observed when the toxin alone was negatively stained at a very high concentration (170,000 hemolytic units/ml). These structures were also produced in the presence of cholesterol even if the toxin concentration was low.     

Detection of cholesterol-rich microdomains in the inner leaflet of the plasma membrane

The C-terminal domain (D4) of perfringolysin O binds selectively to cholesterol in cholesterol-rich microdomains. To address the issue of whether cholesterol-rich microdomains exist in the inner leaflet of the plasma membrane, we expressed D4 as a fusion protein with EGFP in MEF cells. More than half of the EGFP-D4 expressed in stable cell clones was bound to membranes in raft fractions. Depletion of membrane cholesterol with beta-cyclodextrin reduced the amount of EGFP-D4 localized in raft fractions, confirming EGFP-D4 binding to cholesterol-rich microdomains. Subfractionation of the raft fractions showed most of the EGFP-D4 bound to the plasma membrane rather than to intracellular membranes. Taken together, these results strongly suggest the existence of cholesterol-rich microdomains in the inner leaflet of the plasma membrane.     

Alteration of human erythrocyte plasma membranes by perfringolysin O as revealed by freeze-fracture electron microscopy. Studies on Clostridium perfringens exotoxins V

When human erythrocyte membranes were treated with perfringolysin O (Clostridium perfringens θ-toxin) and examined by electron microscopy after freeze-fracture, two ultrastructural alterations were observed in fracture faces of membrane. (1) A random aggregation of intramembranous particles was seen in the fracture face of the protoplasmic half (PF face) of all membranes treated with the toxin, even if at a low concentration (40 hemolytic units/ml). On the other hand, the aggregation in the fracture face of the exoplasmic half (EF face) was observed only in membranes treated with a high concentration (3300 hemolytic units/ml) for 2 h. (2) Round protrusions and ‘cavities’ with 30 nm in diameter were visible in EF and PF faces of membranes treated with a high concentration, respectively. These structures were always protruded toward cytoplasmic side, but did not appear to form holes through the membrane.Ring and are shaped structures with a dark center of 26 nm and a distinct border of 5 nm in width were observed when the toxin alone was negatively stained at a very high concentration (170 000 hemolytic units/ml). These structures were also produced in the presence of cholesterol even if the toxin concentration was low.     

Subcellular distribution of enzymes determined by rapid digitonin fractionation of isolated hepatocytes

Conditions were determined for rapid separation of cytosolic and mitochondrial compartments by digitonin fractionation of rat hepatocytes. The minimum time required for separation of mitochondrial and cytosolic enzyme markers decreased rapidly with increasing temperature. Kyro EOB, a non-ionic detergent, increases the release of cytosolic enzymes, particularly at lower temperatures. Experimental procedures are described for greater than 90% release of cytosolic enzymes and less than 2% release of mitochondrial enzymes in 3s. By using appropriate concentrations of digitonin and Kyro EOB in a fractionation medium maintained at 1°C and a minimum time of exposure to the medium, nearly separate patterns of release were obtained for enzyme markers for the cytosol, mitochondrial matrix and mitochondrial intermembrane space. The distribution of enzymes that exist in more than one of these compartments was quantified by comparing their rates of release with those of marker enzymes. The cytosol/mitochondrial-matrix distributions for such enzymes in hepatocytes from starved rats were 16%/84% for aspartate aminotransferase, 34%/66% for fumarase and 77%/23% for ATP citrate lyase. In hepatocytes from rats that were induced to synthesize ATP citrate lyase by starvation and re-feeding, the ratio had increased to 95%/5%. The maximum cytosol/intermembrane-space ratio for adenylate kinase was 8%/92%. A procedure is also described for treating commercial digitonin that increases its solubility in water from about 1mg/ml to more than 800mg/ml.     

Crystal structure of cytotoxin protein suilysin from Streptococcus suis

Cholesterol-dependent cytolysins (CDC) are pore forming toxins. A prototype of the CDC family members is perfringolysin O (PFO), which directly binds to cholesterol rich cell membrane and lyses the cell. However, as an exception of this general observation, Streptococcus intermedius intermedilysin (ILY) requires human CD59 as its receptor in addition to cholesterol when exhibiting hemolytic activity. It was attempted to explain this functional difference based on a conformational variation in the C-terminal domain of the two toxin proteins, particularly a highly conserved undecapeptide termed tryptophan rich motif. Here, we present the crystal structure of suilysin, a CDC toxin from the swine infectious pathogen Streptococcus suis. Like PFO, suilysin does not require a host receptor for hemolytic activity; yet in the suilysin crystal it shares a similar conformation in the tryptophan rich motif with ILY. This observation suggests that current views of structure-function relationship of CDC proteins in membrane association are still far from complete.     

A basic model for the association of ligands with membrane cholesterol: application to cytolysin binding

Almost all the cholesterol in cellular membranes is associated with phospholipids in simple stoichiometric complexes. This limits the binding of sterol ligands such as filipin and perfringolysin O (PFO) to a small fraction of the total. We offer a simple mathematical model that characterizes this complexity. It posits that the cholesterol accessible to ligands has two forms: active cholesterol, which is that not complexed with phospholipids; and extractable cholesterol, that which ligands can capture competitively from the phospholipid complexes. Simulations based on the model match published data for the association of PFO oligomers with liposomes, plasma membranes, and the isolated endoplasmic reticulum. The model shows how the binding of a probe greatly underestimates cholesterol abundance when its affinity for the sterol is so weak that it competes poorly with the membrane phospholipids. Two examples are the understaining of plasma membranes by filipin and the failure of domain D4 of PFO to label their cytoplasmic leaflets. Conversely, the exaggerated staining of endolysosomes suggests that their cholesterol, being uncomplexed, is readily available. The model is also applicable to the association of cholesterol with intrinsic membrane proteins. For example, it supports the hypothesis that the sharp threshold in the regulation of homeostatic endoplasmic reticulum proteins by cholesterol derives from the cooperativity of their binding to the sterol weakly held by the phospholipids. Thus, the model explicates the complexity inherent in the binding of ligands like PFO and filipin to the small accessible fraction of membrane cholesterol.     

Helical crystallization on nickel-lipid nanotubes: perfringolysin O as a model protein

To facilitate purification and subsequent structural studies of recombinant proteins the most widely used genetically encoded tag is the histidine tag (His-tag) which specifically binds to N-nitrilotriacetic-acid-chelated nickel ions. Lipids derivatized with a nickel-chelating head group can be mixed with galactosylceramide glycolipids to prepare lipid nanotubes that bind His-tagged proteins. In this study, we use His-tagged perfringolysin O (PFO), a soluble toxin secreted by the bacterial pathogen Clostridium perfringens, as a model protein to test the utility of nickel-lipid nanotubes as a tool for structural studies of His-tagged proteins. PFO is a member of the cholesterol dependent cytolysin family (CDC) of oligomerizing, pore-forming toxins found in a variety of Gram-positive bacterial pathogens. CDC pores have been difficult to study by X-ray crystallography because they are membrane associated and vary in size. We demonstrate that both a wild-type and a mutant form of PFO form helical arrays on nickel-lipid containing nanotubes. Cryo-electron microscopy and image analysis of the helical arrays were used to reconstruct a 3D density map of wild-type PFO. This study suggests that the use of nickel-lipid nanotubes may offer a general approach for structural studies of recombinant proteins and may provide insights into the molecular interactions of proteins that have a natural affinity for a membrane surface.     

Effect of Copper Chloride Exposure on the Membrane Potential and Cytosolic Free Calcium in Primary Cultured Chicken Hepatocytes

This study was conducted to examine the effects of copper on membrane potential and cytosolic free calcium in isolated primary chicken hepatocytes which were exposed to different concentration of Cu(2+) (0, 10, 50, 100 μM) or a mixture of Cu(2+) and vitamin C (50 and 50 μM, respectively). Viability, membrane potential, and cytosolic free Ca(2+) of monolayer cultured hepatocytes were investigated at the indicated time point. Results showed that, among the different concentrations of Cu(2+) exposure, the viability of hepatocytes treated with 100 μM Cu(2+) was the worst at the 12th and 24th hours. The effects of Cu(2+) on viability and proliferation were time and dose dependent. Further investigation indicated that Cu(2+) exposure significantly enhanced cytosolic free Ca(2+) in hepatocytes, compared to that in control group, at the 24th hour. Meanwhile, membrane potential was noticeably reduced in hepatocytes increasing concentration of Cu(2+). Taking these results together, we have shown that Cu(2+) can cause toxicity to primary chicken hepatocytes in excessive dose and the effect of Cu(2+) exposure on membrane potential is not site specific, which is probably mediated by the changes of cytosolic free Ca(2+).     

Thawed human hepatocytes in primary culture.

In drug metabolism studies, isolated and cultured human hepatocytes provide a useful model for overcoming the difficulty of extrapolating from animal data. In vitro studies with human hepatocytes are scarce because of the lack of livers and suitable methods of storage. After developing a new method for cryopreservation of human hepatocytes, we evaluated the effects of deep freezing storage on their viability, morphology, and functional and toxicological capabilities in classical culture conditions. Freshly isolated human hepatocytes were cryopreserved in medium containing 10% Me2SO and 20% fetal calf serum, using a Nicool ST20 programmable freezer (-1.9 degrees C/min for 18 min and -30 degrees C/min for 4 min). Cells were stored in liquid nitrogen. Viability of thawed human hepatocytes was 50-65% as assessed by erythrosin exclusion test prior to purification on a Percoll density gradient. Morphological criteria showed that thawed human hepatocytes require an adaptation period to the medium after seeding. Functional assessments showed that human hepatocytes which survive freezing and thawing preserve their protein synthesis capabilities and are able to secrete a specific protein, anionic peptidic fraction, which is involved in the hepatic uptake of bile-destined cholesterol. We then studied Midazolam biotransformation to test metabolic functions, and erythromycin toxicity by Neutral Red test (cell viability) and 3-(4,5-dimethylthiazol-2-yl)-diphenyl tetrazolium bromide test (cell metabolism). All of these experiments indicated that thawed human hepatocytes should be used 38 h after seeding for optimum recovery of their functions: membrane integrity, protein synthesis, and stabilization of drug metabolism enzymes.     

High mobility group box 1 release from hepatocytes during ischemia and reperfusion injury is mediated by decreased histone deacetylase activity

The mobilization and extracellular release of nuclear high mobility group box-1 (HMGB1) by ischemic cells activates inflammatory pathways following liver ischemia/reperfusion (I/R) injury. In immune cells such as macrophages, post-translational modification by acetylation appears to be critical for active HMGB1 release. Hyperacetylation shifts its equilibrium from a predominant nuclear location toward cytosolic accumulation and subsequent release. However, mechanisms governing its release by parenchymal cells such as hepatocytes are unknown. In this study, we found that serum HMGB1 released following liver I/R in vivo is acetylated, and that hepatocytes exposed to oxidative stress in vitro also released acetylated HMGB1. Histone deacetylases (HDACs) are a family of enzymes that remove acetyl groups and control the acetylation status of histones and various intracellular proteins. Levels of acetylated HMGB1 increased with a concomitant decrease in total nuclear HDAC activity, suggesting that suppression in HDAC activity contributes to the increase in acetylated HMGB1 release after oxidative stress in hepatocytes. We identified the isoforms HDAC1 and HDAC4 as critical in regulating acetylated HMGB1 release. Activation of HDAC1 was decreased in the nucleus of hepatocytes undergoing oxidative stress. In addition, HDAC1 knockdown with siRNA promoted HMGB1 translocation and release. Furthermore, we demonstrate that HDAC4 is shuttled from the nucleus to cytoplasm in response to oxidative stress, resulting in decreased HDAC activity in the nucleus. Together, these findings suggest that decreased nuclear HDAC1 and HDAC4 activities in hepatocytes following liver I/R is a mechanism that promotes the hyperacetylation and subsequent release of HMGB1.     

Toxic injury from mercuric chloride in rat hepatocytes

The relationship between cytosolic free Ca2+, mitochondrial membrane potential, ATP depletion, pyridine nucleotide fluorescence, cell surface blebbing, and cell death was evaluated in rat hepatocytes exposed to HgCl2. In cell suspensions, 50 microM HgCl2 oxidized pyridine nucleotides between 1/2 and 2 min, caused ATP depletion between 2 and 5 min, and produced an 89% loss of cell viability after 20 min. Rates of cell killing were identical in high (1.2 mM) and low (2.6 microM) Ca2+ buffers. Cytosolic free Ca2+ was determined in 1-day cultured hepatocytes by ratio imaging of Fura-2 employing multiparameter digitized video microscopy. In high Ca2+ medium, HgCl2 caused a 3-4-fold increase of free Ca2+ beginning after 6-7 min, but free Ca2+ did not change in low Ca2+ medium. Bleb formation occurred after about 4-5 min in both buffers prior to any increase of free Ca2+. Subsequently, in high Ca2+ medium, blebs became hot spots of free Ca2+ (greater than 600 nM). After about 2 min of exposure to HgCl2, rhodamine 123 fluorescence redistributed from mitochondrial to cytosolic compartments signifying collapse of the mitochondrial membrane potential. The results taken together demonstrate that bleb formation, ATP depletion, and the onset of cell death are not dependent on an increase of cytosolic free Ca2+. HgCl2 toxicity appears to be a consequence of inhibition of oxidative phosphorylation leading to ATP depletion and cell death.     

Signal amplification between Gbetagamma release and PI3Kgamma-mediated PI(3,4,5)P3 formation monitored by a fluorescent Gbetagamma biosensor protein and repetitive two component total internal reflection/fluorescence redistribution after photobleaching analysis

Phosphoinositide 3-kinase gamma (PI3Kgamma) is activated by Gbetagamma release after stimulation of Galpha i -coupled receptors, involving a recruitment of the enzyme to the plasma membrane via interaction of the regulatory subunit p101 or p87 with Gbetagamma. The receptor-mediated release of Gbetagamma was, however, insufficient to elicit a translocation of p101 observable by classical fluorescence microscopy approaches. Since the mobilities of plasma membrane-associated and cytosolic proteins differ strongly, small changes in the amount of plasma membrane association should be detectable by an altered diffusional behavior. Here, changes in mobility were monitored by fluorescence redistribution after photobleaching (FRAP) which was repetitively applied before and after stimulation of cells. To combine the advantages of total internal reflection (TIR) illumination, which preferentially excites fluorophors located at or near the plasma membrane, with that provided by the mobility information, we developed a combined TIR/FRAP setup which enabled us to point bleach parts of an image that was observed under TIR illumination. For FRAP data analysis, we introduce a convolution-based method and a global two component model. Using this TIR/FRAP approach, an increased plasma membrane association of the fluorescent Gbetagamma-binding domain of p101 after Gbetagamma release by G protein-coupled receptor stimulation could be detected and quantified. By comparing the translocation efficiency of this domain with that of YFP-GRP1(PH), a biosensor for the PI3Kgamma product PI(3,4,5)P3, we evaluate the signal amplification between Gbetagamma release and PI(3,4,5)P3 formation after activation of Galpha i -coupled receptors.     

Effects of bile salts on the plasma membranes of isolated rat hepatocytes.

The conjugated trihydroxy bile salts glycocholate and taurocholate removed approx. 20--30% of the plasma-membrane enzymes 5'-nucleotidase, alkaline phosphatase and alkaline phosphodiesterase I from isolated hepatocytes before the onset of lysis, as judged by release of the cytosolic enzyme lactate dehydrogenase. The conjugated dihydroxy bile salt glycodeoxycholate similarly removed 10--20% of the 5'-nucleotidase and alkaline phosphatase activities, but not alkaline phosphodiesterase activity; this bile salt caused lysis of hepatocytes at approx. 10-fold lower concentrations (1.5--2.0mM) than either glycocholate or taurocholate (12--16mM). At low concentrations (7 mM), glycocholate released these enzymes in a predominantly particulate form, whereas at higher concentrations (15 mM) glycocholate further released these components in a predominantly 'soluble' form. Inclusion of 1% (w/v) bovine serum albumin in the incubations had a small protective effect on the release of enzymes from hepatocytes by glycodeoxycholate, but not by glycocholate. These observations are discussed in relation to the possible role of bile salts in the origin of some biliary proteins.     

Intracellular cholesterol-binding proteins enhance HDL-mediated cholesterol uptake in cultured primary mouse hepatocytes

A major gap in our knowledge of rapid hepatic HDL cholesterol clearance is the role of key intracellular factors that influence this process. Although the reverse cholesterol transport pathway targets HDL to the liver for net elimination of free cholesterol from the body, molecular details governing cholesterol uptake into hepatocytes are not completely understood. Therefore, the effects of sterol carrier protein (SCP)-2 and liver fatty acid-binding protein (L-FABP), high-affinity cholesterol-binding proteins present in hepatocyte cytosol, on HDL-mediated free cholesterol uptake were examined using gene-targeted mouse models, cultured primary hepatocytes, and 22-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-amino]-23,24-bisnor-5-cholen-3β-ol (NBD-cholesterol). While SCP-2 overexpression enhanced NBD-cholesterol uptake, counterintuitively, SCP-2/SCP-x gene ablation also 1) enhanced the rapid molecular phase of free sterol uptake detectable in <1 min and initial rate and maximal uptake of HDL free cholesterol and 2) differentially enhanced free cholesterol uptake mediated by the HDL3, rather than the HDL2, subfraction. The increased HDL free cholesterol uptake was not due to increased expression or distribution of the HDL receptor [scavenger receptor B1 (SRB1)], proteins regulating SRB1 [postsynaptic density protein (PSD-95)/Drosophila disk large tumor suppressor (dlg)/tight junction protein (ZO1) and 17-kDa membrane-associated protein], or other intracellular cholesterol trafficking proteins (steroidogenic acute response protein D, Niemann Pick C, and oxysterol-binding protein-related proteins). However, expression of L-FABP, the single most prevalent hepatic cytosolic protein that binds cholesterol, was upregulated twofold in SCP-2/SCP-x null hepatocytes. Double-immunogold electron microscopy detected L-FABP sufficiently close to SRB1 for direct interaction, similar to SCP-2. These data suggest a role for L-FABP in HDL cholesterol uptake, a finding confirmed with SCP-2/SCP-x/L-FABP null mice and hepatocytes. Taken together, these results suggest that L-FABP, particularly in the absence of SCP-2, plays a significant role in HDL-mediated cholesterol uptake in cultured primary hepatocytes.     

Effects of Lactic Acidosis on the Function of Cerebral Cortical Synaptosomes

Synaptosomes exposed to anoxic insult produce lactate at a slow rate (measured over 60 min). No measurable damaging effects were produced by prolonged depolarisation, anoxic insult, or exogenous lactate (2-32 mM) either on the synaptic plasma membrane (as judged by release of lactate dehydrogenase and soluble proteins), or on synaptosomal phospholipases (as judged by choline release from membrane phospholipids). Potassium-stimulated acetylcholine release was decreased by incubation in the presence of lactate (2-32 mM), as was potassium- and veratrine-stimulated calcium uptake and the calcium content of depolarised synaptosomes. The intrasynaptosomal pH was also reduced but there was no stimulation of oxygen radical production (as judged by H2O2 generation) by exogenous lactate. The role that lactic acidosis may play in giving rise to the altered calcium homeostasis and decreased acetylcholine release from synaptosomes exposed to anoxic insult is discussed.     

Drug metabolism and viability studies in cryopreserved rat hepatocytes

Rat hepatocytes were cryopreserved optimally by freezing them at 1 degrees C/min to -80 degrees C in cryoprotectant medium containing either 20% (v/v) dimethylsulfoxide (Me2SO) and 25% (v/v) fetal calf serum in Leibowitz L15 medium (Me2SO cryoprotectant) or 25% (v/v) vitrification solution (containing Me2SO, acetamide, propylene glycol and polyethylene glycol) in Leibowitz L15 medium (VS25). The VS25 solution was superior for maintaining viability during short-term storage (24-48 hr) but was slightly toxic during longer storage periods (7 days). Although thawed cells were 40-50% viable on ice after cryopreservation, their viability fell rapidly during incubation in suspension at 37 degrees C. This decline in viability occurred more rapidly after freezing in Me2SO cryoprotectant than in VS25 and was associated with extensive intracellular damage and cell swelling. The loss in viability at 37 degrees C does not appear to be due to ice-crystal damage as it occurred in cells stored at -10 degrees C (above the freezing point of the cryoprotectants) and it may be due to temperature/osmotic shock. Both cryoprotectant media were equally efficient at preserving enzyme activities in the hepatocytes over 7 days at -80 degrees C. Cytochrome P450 and reduced glutathione content and the activities of the microsomal enzymes responsible for aminopyrine N-demethylation and epoxide hydrolysis were well maintained over 7 days storage. In contrast, the cytosolic enzymes glutathione-S-transferase and glutathione reductase were markedly labile during cryopreservation. Cytosolic enzymes may be more susceptible to ice-crystal damage, whereas the microsomal membrane may protect the enzymes which are embedded in it.