Characterization of lipid insertion into monomolecular layers mediated by lung surfactant proteins SP-B and SP-C

Pulmonary surfactant proteins, SP-B and SP-C, if present in preformed monolayers can induce lipid insertion from lipid vesicles into the monolayer after the addition of (divalent) cations [Oosterlaken-Dijksterhuis, M. A., Haagsman, H. P., van Golde, L. M. G., & Demel, R. A. (1991) Biochemistry 30, 8276-8287]. This model system was used to study the mechanisms by which SP-B and SP-C induce monolayer formation from vesicles. Lipid insertion proceeds irrespectively of the molecular class, and PG is not required for this process. In addition to lipids that are immediately inserted from vesicles into the monolayer, large amounts of vesicles are bound to the monolayer and their lipids eventually inserted when the surface area is expanded. SP-B and SP-C are directly responsible for the binding of vesicles to the monolayer. By weight, the vesicle binding capacity of SP-B is approximately 4 times that of SP-C. For vesicle binding and insertion, the formation of close contacts between monolayer and vesicles is essential. SP-B and SP-C show very similar surface properties. Both proteins form extremely stable monolayers (collapse pressures 36-37 mN/m) of alpha-helical structures oriented parallel to the interface. In monolayers consisting of DPPC and SP-B or SP-C, an increase in mean molecular area is observed, which is mainly attributed to the phospholipid. This will greatly enhance the insertion of new lipid material into the monolayer. The results of this study suggest that the surface properties and the hydrophobic nature of SP-B and SP-C are important for the protein-mediated monolayer formation.     

Insertion of Alzheimer's A beta 40 peptide into lipid monolayers

The amyloid beta (A beta) peptide is the major component found in the amyloid deposits in the brains of Alzheimer's disease patients. In vitro studies have demonstrated that the aggregation of A beta can take place at three orders of magnitude lower concentrations in the presence of phospholipid molecules compared to bulk peptide studies, suggesting that membrane lipids may mediate A beta toxicity. To understand the interaction of A beta with lipid membranes, we have examined A beta 40 with anionic dipalmitoylphosphatidylglycerol (DPPG), zwitterionic dipalmitoylphosphatidylcholine (DPPC), and cationic dipalmitoyltrimethylammonium propane (DPTAP) monolayers under different subphase conditions. We have used a constant surface pressure insertion assay to assess the degree of peptide insertion into the lipids. Simultaneously, we monitored the surface morphology of the monolayers with fluorescence microscopy. We have also performed dual-probe fluorescence measurements where both the peptide and lipid are tagged with chromophores. Isotherm measurements show that A beta inserts into both DPTAP and DPPG monolayers under physiologically relevant conditions. Insertion into DPPC occurs at lipid densities below that found in a bilayer. The level of insertion is inversely proportional to the lipid packing density. Our results indicate that lipids need not be anionic to interact with A beta. Electrostatic effects involved in A beta 40-lipid interaction are discussed.     

The interaction of the B-ring of colchicine with alpha-tubulin: a novel footprinting approach

Tubulin, the major structural component of the microtubules, participates actively in mitotic spindle formation and chromosomal organization during cell division. Tubulin is the major target for a variety of anti-mitotic drugs. Some of the drugs, such as Vinca alkaloids and taxol, are routinely used for cancer chemotherapy. It is unfortunate that our knowledge of the binding sites on tubulin of these drugs is limited because of lack of a useful and appropriate tool. The photoaffinity labeling approach is the major technique available at present to detect the binding sites of drugs on tubulin. This method, however, has several limitations. First, only part of the binding site can be identified, namely, the residues which react with the photoaffinity label. Second, there are regions of tubulin which are not at the binding site but are affected by the binding of the drug; these regions can not be detected by the photoaffinity labeling approach. The third, and perhaps most serious, limitation is that the traditional approach can detect areas which have nothing to do with the binding of the ligand but which are within a certain distance of the binding site, that distance being less than the length of the photoreactive moiety attached to the ligand. There has been a great deal of controversy on the localization of the binding site of colchicine on tubulin, with some reports suggesting that the binding site is on alpha and some supporting a binding site on beta. Colchicine also has significant effects on tubulin conformation, but the regions which are affected have not been identified. We have attempted here to address these questions by a novel "footprinting" method by which the drug-binding sites and as well as the domain of tubulin affected by drug-induced conformational changes could be determined. Here, we report for the first time that the interaction of the B-ring of colchicine with the alpha-subunit affects a domain of tubulin which appears to be far from its binding site. This domain includes the cysteine residues at positions 295, 305, 315 and 316 on alpha-tubulin; these residues are located well away from the alpha/beta interface where colchicine appears to bind. This is correlated with the stabilizing effect of colchicine on the tubulin molecule. Furthermore, we also found that the B-ring of colchicine plays a major role in the stability of tubulin while the A and the C-rings have little effect on it. Our results therefore, support a model whereby colchicine binds at the alpha/beta interface of tubulin with the B-ring on the alpha-subunit and the A and the C-rings on the beta-subunit.     

Parameters modulating the maximum insertion pressure of proteins and peptides in lipid monolayers

The lipid monolayer model membrane is useful for studying the parameters responsible for protein and peptide membrane binding. Different approaches have been used to determine the extent of protein and peptide binding to lipid monolayers. This review focuses on the use of the “maximum insertion pressure” (MIP) to estimate the extent of protein and peptide penetration in lipid monolayers. The MIP data obtained with different proteins and peptides have been reviewed and discussed which allowed to draw conclusions on the parameters modulating the monolayer binding of proteins and peptides. In particular, secondary structure components such as amphipathic α-helices of proteins and peptides as well as electrostatic interactions play important roles in monolayer binding. The MIPs have been compared to the estimated lateral pressure of biomembranes which allowed to evaluate the possible association between proteins or peptides with natural membranes. For example, the MIP of a membrane-anchored protein with a glycosylphosphatidylinositol (GPI) was found to be far below the estimated lateral pressure of biomembranes. This allowed us to conclude that this protein is probably unable to penetrate the membrane and should thus be hanged at the membrane surface by use of its GPI lipid anchor. Moreover, the values of MIP obtained with myristoylated and non-myristoylated forms of calcineurin suggest that the myristoyl group does not contribute to monolayer binding. However, the acylation of a peptide resulted in a large increase of MIP. Finally, the physical state of lipid monolayers can have a strong effect on the values of MIP such that it is preferable to perform measurements with lipids showing a single physical state. Altogether the data show that the measurement of the maximum insertion pressure provides very useful information on the membrane binding properties of proteins and peptides although uncertainties must be provided to make sure the observed differences are significant.     

Insertion selectivity of antimicrobial peptide protegrin-1 into lipid monolayers: Effect of head group electrostatics and tail group packing

The ability to selectively target the harmful microbial membrane over that of the host cell is one of the most important characteristics of the antimicrobial peptides (AMPs). This selectivity strongly depends on the chemical and structural properties of the lipids that make up the cell membrane. A systematic study of the initial membrane selectivity of protegrin-1 (PG-1), a β-sheet AMP, was performed using Langmuir monolayers. Constant pressure insertion assay was used to quantify the amount of PG-1 insertion and fluorescence microscopy was employed to observe the effect of PG-1 on lipid ordering. Charge and packing properties of the monolayer were altered by using lipids with different head groups, substituting saturated with unsaturated lipid tail group(s) and incorporating spacer molecules. PG-1 inserted most readily into anionic films composed of phosphatidylglycerol (PG) and lipid A, consistent with its high selectivity for microbial membranes. It also discriminated between zwitteranionic phospholipids, inserting more readily into phosphatidylcholine (PC) monolayers than those composed of phosphatidylethanolamine, potentially explaining why PG-1 is hemolytic for PC-rich human erythrocytes and not for the PE-rich erythrocytes of ruminants. Increased packing density of the monolayer by increased surface pressure, increased tail group saturation or incorporation of dihydrocholesterol diminishes the insertion of PG-1. Fluorescence microscopy shows that lipid packing is disordered upon PG-1 insertion. However, the presence of PG-1 can still affect lipid morphology even with no observed PG-1 insertion. These results show the important role that lipid composition of the cell membrane plays in the activity of AMPs.     

Insertion selectivity of antimicrobial peptide protegrin-1 into lipid monolayers: effect of head group electrostatics and tail group packing

The ability to selectively target the harmful microbial membrane over that of the host cell is one of the most important characteristics of the antimicrobial peptides (AMPs). This selectivity strongly depends on the chemical and structural properties of the lipids that make up the cell membrane. A systematic study of the initial membrane selectivity of protegrin-1 (PG-1), a beta-sheet AMP, was performed using Langmuir monolayers. Constant pressure insertion assay was used to quantify the amount of PG-1 insertion and fluorescence microscopy was employed to observe the effect of PG-1 on lipid ordering. Charge and packing properties of the monolayer were altered by using lipids with different head groups, substituting saturated with unsaturated lipid tail group(s) and incorporating spacer molecules. PG-1 inserted most readily into anionic films composed of phosphatidylglycerol (PG) and lipid A, consistent with its high selectivity for microbial membranes. It also discriminated between zwitteranionic phospholipids, inserting more readily into phosphatidylcholine (PC) monolayers than those composed of phosphatidylethanolamine, potentially explaining why PG-1 is hemolytic for PC-rich human erythrocytes and not for the PE-rich erythrocytes of ruminants. Increased packing density of the monolayer by increased surface pressure, increased tail group saturation or incorporation of dihydrocholesterol diminishes the insertion of PG-1. Fluorescence microscopy shows that lipid packing is disordered upon PG-1 insertion. However, the presence of PG-1 can still affect lipid morphology even with no observed PG-1 insertion. These results show the important role that lipid composition of the cell membrane plays in the activity of AMPs.     

NBD-isocolcemid-tubulin interaction: a novel one-step reaction involving no conformational adjustment of reactants

Isocolcemid, a colcemid analogue in which the positions of the C-ring methoxy and carbonyl are exchanged, is virtually inactive in binding to tubulin and inhibiting the formation of microtubule assembly. We have found that the substitution of a NBD group in the side chain of the B-ring of isocolcemid can reverse the effect of these structural alterations (at the C-ring) and the newly synthesized NBD-isocolcemid restores the lost biological activity. It inhibits microtubule assembly with an IC(50) of 12 microM and competes efficiently with [(3)H]colchicine, for binding to tubulin. NBD-isocolcemid has two binding sites on tubulin; one is characterized by fast binding, whereas the binding to the other site is slow. These two sites are independent and unrelated to each other. Colchicine and its analogues compete with NBD-isocolcemid for the slow site. Association and dissociation rate constants for the fast site, obtained from the stopped-flow measurements, are (7.37 +/- 0. 70) x 10(5) M(-1) s(-1) and 7.82 +/- 2.74 s(-1), respectively. While the interaction of colchicine and its analogues with tubulin involves two steps, NBD-isocolcemid binding to tubulin at the slow site has been found to be a one-step reaction. This is evident from the linear dependence of the observed rate constant (k(obs)) with both NBD-isocolcemid and tubulin concentrations. The interaction of NBD-isocolcemid with tubulin does not involve the conformational change of NBD-isocolcemid, as is evident from the unchanged CD spectra of the drug. The absence of enhanced GTPase activity of tubulin and the native-like protease cleavage pattern of the NBD-isocolcemid-tubulin complex suggest an unaltered conformation of tubulin upon NBD-isocolcemid binding to it as well. Implications of this on the mechanism of polymerization inhibition have been discussed.     

Difference in lipid packing sensitivity of exchangeable apolipoproteins apoA-I and apoA-II: an important determinant for their distinctive role in lipid metabolism

Exchangeable apolipoproteins A-I and A-II play distinct roles in reverse cholesterol transport. ApoA-I interacts with phospholipids and cholesterol of the cell membrane to make high density lipoprotein particles whereas apolipoprotein A-II interacts with high density lipoprotein particles to release apolipoprotein A-I. The two proteins show a high activity at the aqueous solution/lipid interface and are characterized by a high content of amphipathic α-helices built upon repetition of the same structural motif. We set out to investigate to what extent the number of α-helix repeats of this structural motif modulates the affinity of the protein for lipids and the sensitivity to lipid packing. To this aim we have compared the insertion of apolipoproteins A-I and A-II in phospholipid monolayers formed on a Langmuir trough in conditions where lipid packing, surface pressure and charge were controlled. We also used atomic force microscopy to obtain high resolution topographic images of the surface at a resolution of several nanometers and performed statistical image analysis to calculate the spatial distribution and geometrical shape of apolipoproteins A-I and A-II clusters. Our data indicate that apolipoprotein A-I is sensitive to packing of zwitterionic lipids but insensitive to the packing of negatively charged lipids. Interestingly, apolipoprotein A-II proved to be insensitive to the packing of zwitterionic lipids. The different sensitivity to lipid packing provides clues as to why apolipoprotein A-II barely forms nascent high density lipoprotein particles while apolipoprotein A-I promotes their formation. We conclude that the different interfacial behaviors of apolipoprotein A-I and apolipoprotein A-II in lipidic monolayers are important determinants of their distinctive roles in lipid metabolism.     

Colchicine reversibly inhibits electrical activity in arthropod mechanoreceptors

Colchicine and some other microtubule-active agents inhibit the electrical responses of cockroach tibial spine mechanoreceptors. Lumicolchicine, a colchicine analog which does not bind to microtubule protein, does not inhibit mechanoreceptive responses. Colchicine inhibition of peripheral mechanoreceptive responses is fully reversible and dose dependent, but colchicine has no effect on conduction in leg nerve axons. Colchicine inhibition is therefore an effect on the sensory dendrites or soma. The inhibition produced by colchicine could be produced by several effects. Colchicine may inhibit because it (1) disrupts the numerous intracellular microtubules which are a part of this sensory receptor's dendrite, (2) blocks axoplasmic transport of essential materials to the sensory dendrite, or (3) binds to tubulin or other proteins in the dendritic membrane.     

Molecular packing of cord factor and its interaction with phosphatidylinositol in mixed monolayers.

Cord factor (trehalose 6,6'-dimycolate, CF) is a glycolipid located in the outer mycobacterial cell wall that is implicated in the pathogenesis of mycobacteria. Furthermore, CF is a convenient model for studying mycolic acid residues, the major lipid constituents of the mycobacterial cell wall that are believed to form a barrier against drug penetration. The surface properties of CF and its interactions with phosphatidylinositol (PI) have been investigated using the monolayer technique. During compression/expansion/recompression cycles, CF monolayers switch from a loosely packed to a more tightly packed structure. The change in surface properties suggests a molecular rearrangement, perhaps involving interdigitation of long and short chains of the CF molecules. In CF-PI monolayers, maximal lateral packing density occurs between 0.5 and 0.7 mole fraction CF, which is close to the relative composition of mycolic acid residues and shorter-chain lipids in the mycobacterial cell wall. Low concentrations of CF increase the order in PI monolayers, consistent with CF toxicity involving rigidification of cell membranes.     

The natural antioxidant rosmarinic acid spontaneously penetrates membranes to inhibit lipid peroxidation in situ

Exogenous molecules from dietary sources such as polyphenols are very efficient in preventing the alteration of lipid membranes by oxidative stress. Among the polyphenols, we have chosen to study rosmarinic acid (RA). We investigated the efficiency of RA in preventing lipid peroxidation and in interacting with lipids. We used liposomes of 1,2-dilinoleoyl-sn-glycero-3-phosphocholine (DLPC) to show that RA was an efficient antioxidant. By HPLC, we determined that the maximum amount of RA associated with the lipids was ~1 mol%. Moreover, by using Langmuir monolayers, we evidenced that cholesterol decreases the penetration of RA. The investigation of transferred lipid/RA monolayers by atomic force microscopy revealed that 1 mol% of RA in the membrane was not sufficient to alter the membrane structure at the nanoscale. By fluorescence, we observed no significant modification of membrane permeability and fluidity caused by the interaction with RA. We also deduced that RA molecules were mainly located among the polar headgroups of the lipids. Finally, we prepared DLPC/RA vesicles to evidence for the first time that up to 1 mol% of RA inserts spontaneously in the membrane, which is high enough to fully prevent lipid peroxidation without any noticeable alteration of the membrane structure due to RA insertion.     

Colchicine binding to an oligomer of tubulin

An oligomeric form of tubulin present in microtubule protein prepared from mammalian brain, the 36S double ring containing tau protein, is reported to bind colchicine. Colchicine binds to each individual 6S tubulin subunit in the 36S ring without apparent effect on quarternary structure. The colchicine-oligomer complex forms by colchicine binding directly to the tubulin ring; alternatively, complexes formed by colchicine with 6S tubulin subunits associate in the presence of tau protein to form the colchicine-oligomer complex.     

Effect of chemical modifications of myelin basic protein on its interaction with lipid interfaces and cell fusion ability

Summary The ability of native and chemically modified myelin basic protein to induce fusion of chicken erythrocytes and to interact with lipids in monolayers at the air-water interface and liposomes was studied. Chemical modifications of myelin basic protein were performed by acetylation and succinylation: the positive charges of the native protein were blocked to an extent of about 90–95%.Cellular aggregation and fusion of erythrocytes into multinucleated cells was induced by the native myelin basic protein. This effect was diminished for both acetylated and succinylated myelin basic protein. Native myelin basic protein penetrated appreciably in sulphatide-containing lipid monolayers while lower penetration occurred in monolayers of neutral lipids. Contrary to this, both chemically modified myelin basic proteins did not show any selectivity to penetrate into interfaces of neutral or negatively charged lipids. The intrinsic fluorescence of the native and chemically modified myelin basic proteins upon interacting with liposomes constituted by dipalmitoylphosphatidycholine, glycosphingolipids, egg phosphatidic acid or dipalmitoylphosphatidyl glycerol was studied. The interaction with liposomes of anionic lipids is accompanied by a blue shift of the maximum of the native protein emission fluorescence spectrum from 346 nm to 335 nm; no shift was observed with liposomes containing neutral lipids. The acetylated and succinylated myelin basic proteins did not show changes of their emission spectra upon interacting with any of the lipids studied. The results obtained in monolayers and the fluorescence shifts indicate a lack of correlation between the ability of the modified proteins to penetrate lipid interfaces and the microenvironment sensed by the tryptophan-containing domain.Abbreviations MBP myelin basic protein - DPPC dipalmitoyl phosphatidylcholine - DPPG dipalmitoyl phosphatidylglycerol - PA phosphatidic acid     

Differential interaction of equinatoxin II with model membranes in response to lipid composition

Equinatoxin II is a 179-amino-acid pore-forming protein isolated from the venom of the sea anemone Actinia equina. Large unilamellar vesicles and lipid monolayers of different lipid compositions have been used to study its interaction with membranes. The critical pressure for insertion is the same in monolayers made of phosphatidylcholine or sphingomyelin (approximately 26 mN m(-1)) and explains why the permeabilization of large unilamellar vesicles by equinatoxin II with these lipid compositions is null or moderate. In phosphatidylcholine-sphingomyelin (1:1) monolayers, the critical pressure is higher (approximately 33 mN m(-1)), thus permitting the insertion of equinatoxin II in large unilamellar vesicles, a process that is accompanied by major conformational changes. In the presence of vesicles made of phosphatidylcholine, a fraction of the protein molecules remains associated with the membranes. This interaction is fully reversible, does not involve major conformational changes, and is governed by the high affinity for membrane interfaces of the protein region comprising amino acids 101-120. We conclude that although the presence of sphingomyelin within the membrane creates conditions for irreversible insertion and pore formation, this lipid is not essential for the initial partitioning event, and its role as a specific receptor for the toxin is not so clear-cut.     

Measuring protein insertion areas in lipid monolayers by fluorescence correlation spectroscopy

Membrane insertion of protein domains is an important step in many membrane remodeling processes, for example, in vesicular transport. The membrane area taken up by the protein insertion influences the protein binding affinity as well as the mechanical stress induced in the membrane and thereby its curvature. To our knowledge, this is the first optical measurement of this quantity on a system in equilibrium with direct determination of the number of inserted protein and no further assumptions concerning the binding thermodynamics. Whereas macroscopic total area changes in lipid monolayers are typically measured on a Langmuir film balance, finding the number of inserted proteins without perturbing the system and quantitating any small area changes has posed a challenge. Here, we address both issues by performing two-color fluorescence correlation spectroscopy directly on the monolayer. With a fraction of the protein being fluorescently labeled, the number of inserted proteins is determined in situ without resorting to invasive techniques such as collecting the monolayer by aspiration. The second color channel is exploited to monitor a small fraction of labeled lipids to determine the total area increase. Here, we use this method to determine the insertion area per molecule of Sar1, a protein of the COPII complex, which is involved in transport vesicle formation. Sar1 has an N-terminal amphipathic helix, which is responsible for membrane binding and curvature generation. An insertion area of (3.4 ± 0.8) nm² was obtained for Sar1 in monolayers from a lipid mixture typically used in COPII reconstitution experiments, in good agreement with the expected insertion area of the Sar1 amphipathic helix. By using the two-color approach, determining insertion areas relies only on local fluorescence measurements. No macroscopic area measurements are needed, giving the method the potential to also be applied to laterally heterogeneous monolayers and bilayers.     

In situ characterization of lipid A interaction with antimicrobial peptides using surface X-ray scattering

Lipid A structure at the air-aqueous interface has been studied using pressure-area isotherm methods coupled with the surface X-ray scattering techniques of X-ray reflectivity (XR) and grazing incidence X-ray diffraction (GIXD). Lipid A monolayers were formed at the air-aqueous interface to represent the lipid moiety of the outer membrane of Gram-negative bacteria. Lipid A structure was characterized at surface pressures between 10 and 35 mN/m. Interactions of α-helical antimicrobial peptides LL-37, SMAP-29 and D2A22 with lipid A monolayers were subsequently studied. Although insertion into the lipid A monolayers was observed with the α-helical peptides, little change was seen from the X-ray data, suggesting that the lipid A hydrocarbon chains are involved in reorientation during insertion and that the hydrocarbon chains have a relatively rigid structure.     

Effect of nonbilayer lipids on membrane binding and insertion of the catalytic domain of leader peptidase

Biological membranes contain a substantial amount of "nonbilayer lipids", which have a tendency to form nonlamellar phases. In this study the hypothesis was tested that the presence of nonbilayer lipids in a membrane, due to their overall small headgroup, results in a lower packing density in the headgroup region, which might facilitate the interfacial insertion of proteins. Using the catalytic domain of leader peptidase (delta2-75) from Escherichia coli as a model protein, we studied the lipid class dependence of its insertion and binding. In both lipid monolayers and vesicles, the membrane binding of (catalytically active) delta2-75 was much higher for the nonbilayer lipid DOPE compared to the bilayer lipid DOPC. For the nonbilayer lipids DOG and MGDG a similar effect was observed as for DOPE, strongly suggesting that no specific interactions are involved but that the small headgroups create hydrophobic interfacial insertion sites. On the basis of the results of the monolayer experiments, calculations were performed to estimate the space between the lipid headgroups accessible to the protein. We estimate a maximal size of the insertion sites of 15 +/- 7 A2/lipid molecule for DOPE, relative to DOPC. The size of the insertion sites decreases with an increase in headgroup size. These results show that nonbilayer lipids stimulate the membrane insertion of delta2-75 and support the idea that such lipids create insertion sites by reducing the packing density at the membrane-water interface. It is suggested that PE in the bacterial membrane facilitates membrane insertion of the catalytic domain of leader peptidase, allowing the protein to reach the cleavage site in preproteins.     

Colchicine binding in the free-living nematode Caenorhabditis elegans

The [3H]colchicine-binding activity of a crude supernatant of the free-living nematode Caenorhabditis elegans was resolved into a non-saturable component and a tubulin-specific component after partial purification of tubulin by polylysine affinity chromatography. The two fractions displayed opposing thermal dependencies of [3H]colchicine binding, with non-saturable binding increasing, and tubulin binding decreasing, at 4 degrees C. Binding of [3H]colchicine to C.elegans tubulin at 37 degrees C is a pseudo-first-order rate process with a long equilibration time. The affinity of C. elegans tubulin for [3H]colchicine is relatively low (Ka = 1.7 x 10(5) M(-1)) and is characteristic of the colchicine binding affinities observed for tubulins derived from parasitic nematodes. [3H]Colchicine binding to C. elegans tubulin was inhibited by unlabelled colchicine, podophyllotoxin and mebendazole, and was enhanced by vinblastine. The inhibition of [3H]colchicine binding by mebendazole was 10-fold greater for C. elegans tubulin than for ovine brain tubulin. The inhibition of [3H]colchicine binding to C. elegans tubulin by mebendazole is consistent with the recognised anthelmintic action of the benzimidazole carbamates. These data indicate that C. elegans is a useful model for examining the interactions between microtubule inhibitors and the colchicine binding site of nematode tubulin.     

Phosphorylation of voltage-dependent anion channel by serine/threonine kinases governs its interaction with tubulin

Tubulin was recently found to be a uniquely potent regulator of the voltage-dependent anion channel (VDAC), the most abundant channel of the mitochondrial outer membrane, which constitutes a major pathway for ATP/ADP and other metabolites across this membrane. Dimeric tubulin induces reversible blockage of VDAC reconstituted into a planar lipid membrane and dramatically reduces respiration of isolated mitochondria. Here we show that VDAC phosphorylation is an important determinant of its interaction with dimeric tubulin. We demonstrate that in vitro phosphorylation of VDAC by either glycogen synthase kinase-3β (GSK3β) or cAMP-dependent protein kinase A (PKA), increases the on-rate of tubulin binding to the reconstituted channel by orders of magnitude, but only for tubulin at the cis side of the membrane. This and the fact the basic properties of VDAC, such as single-channel conductance and selectivity, remained unaltered by phosphorylation allowed us to suggest the phosphorylation regions positioned on the cytosolic loops of VDAC and establish channel orientation in our reconstitution experiments. Experiments on human hepatoma cells HepG2 support our conjecture that VDAC permeability for the mitochondrial respiratory substrates is regulated by dimeric tubulin and channel phosphorylation. Treatment of HepG2 cells with colchicine prevents microtubule polymerization, thus increasing dimeric tubulin availability in the cytosol. Accordingly, this leads to a decrease of mitochondrial potential measured by assessing mitochondrial tetramethylrhodamine methyester uptake with confocal microscopy. Inhibition of PKA activity blocks and reverses mitochondrial depolarization induced by colchicine. Our findings suggest a novel functional link between serine/threonine kinase signaling pathways, mitochondrial respiration, and the highly dynamic microtubule network which is characteristic of cancerogenesis and cell proliferation.     

Microtubule Integrity Is Necessary for the Epithelial Barrier Function of Cultured Thyroid Cell Monolayers

Vectorial transport in the thyroid epithelium requires an efficient barrier against passive paracellular flux, a role which is principally performed by the tight junction (zonula occludens). There is increasing evidence that tight junction integrity is determined by integral and peripheral membrane proteins which interact with the cell cytoskeleton. Although the contribution of the actin cytoskeleton to tight junction physiology has been intensively studied, less is known about possible interactions with microtubules. In the present study we used electrophysiological and immunohistochemical approaches to investigate the contribution of microtubules to the paracellular barrier in cultured thyroid cell monolayers which displayed a high transepithelial electrical resistance (6000-9000 ohm · cm²). Colchicine (1 μM) caused a progressive fall in electrical resistance to <10% of baseline after 6 h and depolarization of the transepithelial electrical potential difference consistent with a significant increase in paracellular permeability. The effect of colchicine on TER was not affected by agents which inhibit the major apical conductances of thyroid cells but was reversed upon removal of the drug. Immunofluorescent staining for tubulin combined with confocal laser scanning microscopy demonstrated that thyroid cells possessed a dense microtubule network extending throughout the cytoplasm which was destroyed by colchicine. Colchicine also produced changes in the localization of the tight junction-associated protein, ZO-1: its normally continuous junctional distribution was disrupted by striking discontinuities and the appearance of many fine strands which extended into the cytoplasm. A similar disruption in E-cadherin staining was also observed, but colchicine did not affect the distribution of vinculin associated with adherens junctions nor the integrity of the perijunctional actin ring. We conclude that microtubules are necessary for the functional and structural integrity of tight junctions in this electrically tight, transporting epithelium.