Effects of reductive methylation on microtubule assembly. Evidence for an essential amino group in the alpha-chain

Handling Difficult Cryo-ET Samples: A Study with Primary Neurons from Drosophila melanogaster

Cellular neurobiology has benefited from recent advances in the field of cryo-electron tomography (cryo-ET). Numerous structural and ultrastructural insights have been obtained from plunge-frozen primary neurons cultured on electron microscopy grids. With most primary neurons having been derived from rodent sources, we sought to expand the breadth of sample availability by using primary neurons derived from 3rd instar Drosophila melanogaster larval brains. Ultrastructural abnormalities were encountered while establishing this model system for cryo-ET, which were exemplified by excessive membrane blebbing and cellular fragmentation. To optimize neuronal samples, we integrated substrate selection, micropatterning, montage data collection, and chemical fixation. Efforts to address difficulties in establishing Drosophila neurons for future cryo-ET studies in cellular neurobiology also provided insights that future practitioners can use when attempting to establish other cell-based model systems.

Handling difficult cryo-ET samples: A study with primary neurons from Drosophila melanogaster

Cellular neurobiology has benefited from recent advances in the field of cryo-electron tomography (cryo-ET). Numerous structural and ultrastructural insights have been obtained from plunge-frozen primary neurons cultured on electron microscopy grids. With most primary neurons been derived from rodent sources, we sought to expand the breadth of sample availability by using primary neurons derived from 3rd instar Drosophila melanogaster larval brains. Ultrastructural abnormalities were encountered while establishing this model system for cryo-ET, which were exemplified by excessive membrane blebbing and cellular fragmentation. To optimize neuronal samples, we integrated substrate selection, micropatterning, montage data collection, and chemical fixation. Efforts to address difficulties in establishing Drosophila neurons for future cryo-ET studies in cellular neurobiology also provided insights that future practitioners can use when attempting to establish other cell-based model systems.

Alcohol-induced tubulin post-translational modifications directly alter hepatic protein trafficking

BACKGROUND

Chronic ethanol exposure leads to enhanced protein acetylation and acetaldehyde adduction. Of the multitude of proteins that are modified on ethanol administration, tubulin is among the best studied. However, an open question is whether these modifications are observed in patient samples. Both modifications have also been implicated in promoting alcohol-induced defects in protein trafficking, but whether they do so directly is also unanswered.

METHODS AND RESULTS

We first confirmed that tubulin was hyperacetylated and acetaldehyde-adducted in the livers from ethanol-exposed individuals to a similar extent as observed in the livers from ethanol-fed animals and hepatic cells. Livers from individuals with nonalcohol-associated fatty liver showed modest increases in tubulin acetylation, whereas nonalcohol-associated fibrotic human and mouse livers showed virtually no tubulin modifications. We also asked whether tubulin acetylation or acetaldehyde adduction can directly explain the known alcohol-induced defects in protein trafficking. Acetylation was induced by overexpressing the α-tubulin-specific acetyltransferase, αTAT1, whereas adduction was induced by directly adding acetaldehyde to cells. Both αTAT1 overexpression and acetaldehyde treatment significantly impaired plus-end (secretion) and minus-end (transcytosis)-directed microtubule-dependent trafficking and clathrin-mediated endocytosis. Each modification led to similar levels of impairment as observed in ethanol-treated cells. The levels of impairment by either modification showed no dose dependence or no additive effects suggesting that substoichiometric tubulin modifications lead to altered protein trafficking and that lysines are not selectively modified.

CONCLUSIONS

These results not only confirm that enhanced tubulin acetylation is observed in human livers but that it is most relevant to alcohol-induced injury. Because these tubulin modifications are associated with altered protein trafficking that alters proper hepatic function, we propose that changing the cellular acetylation levels or scavenging free aldehydes are feasible strategies for treating alcohol-associated liver disease.

Spontaneous Alignment of Microtubules Via Tubulin Polymerization in a Narrow Space Under a Temperature Gradient

In this chapter, protocols for spontaneous alignment of microtubules (MTs), such as helices and spherulites, via tubulin polymerization in a narrow space and under a temperature gradient are presented for tubulin solutions and tubulin-polymer mixtures. These protocols provide an easy route for hierarchical MT assembly and may extend our current understanding of cytoskeletal protein self-assembly under dissipative conditions.

The Altered Hepatic Tubulin Code in Alcoholic Liver Disease

The molecular mechanisms that lead to the progression of alcoholic liver disease have been actively examined for decades. Because the hepatic microtubule cytoskeleton supports innumerable cellular processes, it has been the focus of many such mechanistic studies. It has long been appreciated that α-tubulin is a major target for modification by highly reactive ethanol metabolites and reactive oxygen species. It is also now apparent that alcohol exposure induces post-translational modifications that are part of the natural repertoire, mainly acetylation. In this review, the modifications of the "tubulin code" are described as well as those adducts by ethanol metabolites. The potential cellular consequences of microtubule modification are described with a focus on alcohol-induced defects in protein trafficking and enhanced steatosis. Possible mechanisms that can explain hepatic dysfunction are described and how this relates to the onset of liver injury is discussed. Finally, we propose that agents that alter the cellular acetylation state may represent a novel therapeutic strategy for treating liver disease.

Thermoresponsive synthetic polymer-microtubule hybrids

Thermoresponsive hybrids consisting of synthetic polymers and microtubules (MTs), i.e., assemblies of tubulins, were prepared by bonding MTs covalently to a few reactive units in a macromolecular strand. The hybrids exhibited the gel/sol transition because of the "assembling of tubulins to MTs/disintegrating of MTs to tubulins" by the temperature change between 37 and 4 °C, respectively. The viscoelastic behaviors of the hybrid gels depended upon the quantity of polymer feed and the amount of resulting covalent bonds between the polymers and tubulin units. Furthermore, in a confined space of a thin and long rectangular cell with the temperature gradient from 4 °C (cold terminal) to 37 °C (warm terminal), the sol state hybrid turned to the gel state that propagated from the warm terminal toward the cold terminal to form uniaxially oriented MT arrays. Upon changing the temperature of the whole system between 37 and 4 °C, the uniaxial arrays appeared/disappeared reversibly.

Molecular dynamics modeling of tubulin C-terminal tail interactions with the microtubule surface

Tubulin, an α/β heterodimer, has had most of its 3D structure analyzed; however, the carboxy (C)-termini remain elusive. Importantly, the C-termini play critical roles in regulating microtubule structure and function. They are sites of most of the post-translational modifications of tubulin and interaction sites with molecular motors and microtubule-associated proteins. Simulated annealing was used in our molecular dynamics modeling to predict the interactions of the C-terminal tails with the tubulin dimer. We examined differences in their flexibility, interactions with the body of tubulin, and the existence of structural motifs. We found that the α-tubulin tail interacts with the H11 helix of β-tubulin, and the β-tubulin tail interacts with the H11 helix of α-tubulin. Tail domains and H10/B9 loops interact with each other and compete for interactions with positively-charged residues of the H11 helix on the neighboring monomer. In a simulation in which α-tubulin's H10/B9 loop switches on sub-nanosecond intervals between interactions with the C-terminal tail of α-tubulin and the H11 helix of β-tubulin, the intermediate domain of α-tubulin showed more fluctuations compared to those in the other simulations, indicating that tail domains may cause shifts in the position of this domain. This suggests that C-termini may affect the conformation of the tubulin dimer which may explain their essential function in microtubule formation and effects on ligand binding to microtubules. Our modeling also provides evidence for a disordered-helical/helical double-state system of the T3/H3 region of the microtubule, which could be linked to depolymerization following GTP hydrolysis.

Alcohol consumption impairs hepatic protein trafficking: mechanisms and consequences

Alcoholic liver disease is a major biomedical health concern in the United States. Despite considerable research efforts aimed at understanding the progression of the disease, the specific mechanisms leading to alcohol-induced damage remain elusive. Numerous proteins are known to have alcohol-induced alterations in their dynamics. Defining these defects in protein trafficking is an active area of research. In general, two trafficking pathways are affected: transport of newly synthesized secretory or membrane glycoproteins from the Golgi to the basolateral membrane and clathrin-mediated endocytosis from the sinusoidal surface. Both impaired secretion and internalization require ethanol metabolism and are likely mediated by acetaldehyde. Although the mechanisms by which ethanol exposure impairs protein trafficking are not fully understood, recent work implicates alcohol-induced modifications on tubulin or components of the clathrin machinery as potential mediators. Furthermore, the physiological ramifications of impaired protein trafficking are not fully understood. In this review, we will list and discuss the proteins whose trafficking patterns are known to be impaired by ethanol exposure. We will then describe what is known about the possible mechanisms leading to impaired protein trafficking and how disrupted protein trafficking alters liver function and may explain clinical features of the alcoholic patient.

Residue-specific adduction of tubulin by 4-hydroxynonenal and 4-oxononenal causes cross-linking and inhibits polymerization

The modification of proteins by lipid aldehydes produced in cells undergoing oxidative stress has been proposed as an important event that contributes to the pathology of numerous diseases. In this context, the alpha,beta-unsaturated aldehydes 4-hydroxynonenal (4-HNE) and 4-oxononenal (4-ONE) generated during membrane lipid peroxidation have been shown to adduct and inactivate numerous proteins. We report here that purified bovine brain tubulin modified with physiologically relevant concentrations of 4-HNE or 4-ONE results in significant protein cross-linking and marked inhibition of the functional capacity of tubulin polymerization. Comparative analysis demonstrated that 4-ONE is a much more potent cross-linker and inhibitor of tubulin assembly than 4-HNE. Additional experiments revealed the unique property of 4-ONE, initiation of depolymerization of intact microtubules. LC-MS/MS analysis demonstrated that Cys 347alpha, Cys 376alpha, and Cys 303beta are consistently modified by 4-HNE. The identification of target residues within tubulin modified by 4-ONE was not successful, and this was attributed to the marked tubulin cross-linking that occurred immediately after addition of 4-ONE. The modification of Lys residues by reductive propylation demonstrated that the majority of 4-HNE and 4-ONE adducts involve Lys residues, suggesting that tubulin cross-links are Lys-dependent. Taken together, these data suggest a mechanistic basis for the impairment of tubulin function by 4-HNE and 4-ONE produced as a consequence of diseases associated with chronic oxidative stress.

A proteomic survey of rat cerebral cortical synaptosomes

Previous findings from our laboratory and others indicate that two-dimensional gel electrophoresis (2-DE) can be used to study protein expression in defined brain regions, but mainly the proteins which are present in high abundance in glia are readily detected. The current study was undertaken to determine the protein profile in a synaptosomal subcellular fraction isolated from the cerebral cortex of the rat. Both 2-DE and liquid chromatography - tandem mass spectrometry (LC-MS/MS) procedures were used to isolate and identify proteins in the synaptosomal fraction and accordingly >900 proteins were detected using 2-DE; the 167 most intense gel spots were isolated and identified with matrix-assisted laser desorption/ionization - time of flight peptide mass fingerprinting or LC-MS/MS. In addition, over 200 proteins were separated and identified with the LC-MS/MS "shotgun proteomics" technique, some in post-translationally modified form. The following classes of proteins associated with synaptic function were detected: (a) proteins involved in synaptic vesicle trafficking-docking (e.g., SNAP-25, synapsin I and II, synaptotagmin I, II, and V, VAMP-2, syntaxin 1A and 1B, etc.); (b) proteins that function as transporters or receptors (e.g., excitatory amino acid transporters 1 and 2, GABA transporter 1); (c) proteins that are associated with the synaptic plasma membrane (e.g., post-synaptic density-95/synapse-associated protein-90 complex, neuromodulin (GAP-43), voltage-dependent anion-selective channel protein (VDACs), sodium-potassium ATPase subunits, alpha 2 spectrin, septin 7, etc.); and (d) proteins that mediate intracellular signaling cascades that modulate synaptic function (e.g., calmodulin, calcium-calmodulin-dependent protein kinase subunits, etc.). Other identified proteins are associated with mitochondrial or general cytosolic function. Of the two proteins identified as endoplasmic reticular, both interact with the synaptic SNARE complex to regulate vesicle trafficking. Taken together, these results suggest that the integrity of the synaptosomes was maintained during the isolation procedure and that this subcellular fractionation technique enables the enrichment of proteins associated with synaptic function. The results also suggest that this experimental approach can be used to study the differential expression of multiple proteins involved in alterations of synaptic function.

Shift in FTIR spectrum patterns in methomyl-exposed rat spleen cells

Methomyl is a highly toxic carbamate insecticide which is widely used in many agricultural countries. We have applied the Fourier-transformed infrared (FTIR) spectroscopic method to study the toxicity of methomyl on cytoskeletal protein and the nucleic acid of rat spleen cells. Rats were given methomyl by gavage at 2, 6 and 8 mg/kg in single doses. Colchicine, a microtubule-disrupting agent, was given to rats at 2, 4, and 6 mg/kg in single doses and mitomycin C, an alkylating agent which acts as a DNA-cross-linking agent, was given by an intraperitoneal route to rats at 1 mg/kg. It was shown that the wavenumber of FTIR spectra at amide I and amide II in both methomyl- and colchicine-exposed rats shifted in dose response manner when compared with the control (P < 0.05). The amide I and II shifts in these regions have been proposed to be the result of an alpha-helix protein conformational change. Toxic doses of mitomycin C, a DNA-cross-linking agent, did not result in this pattern. Moreover, all exposed rats showed an increase in the absorbance ratios that were related to the vibrational mode of the phosphodiester group in nucleic acid (P < 0.05).

Effects of methionine on the cytoplasmic distribution of actin and tubulin during neural tube closure in rat embryos

Research has previously shown that, without methionine supplements, neural tube proteins of rat embryos cultured on bovine sera were hypomethylated and neural tubes failed to close. In the present study, to identify the proteins that became methylated during neurulation, rat embryos were first cultured on methionine-deficient bovine serum for 40 hr, then incubated with puromycin for 1 hr, and, finally, incubated with [methyl-14C]methionine and puromycin for 5 hr. On the basis of molecular weights, isoelectric points, and Western immunoblots, the methyl-14C-labeled proteins were identified as actin, alpha beta-tubulin, and neurofilament L. Indirect immunofluorescence studies indicated that without the addition of methionine to the culture, localization of actin and alpha beta-tubulin in the basal cytoplasm did not occur and these neuroepithelial cells lost their columnar morphology.

Substoichiometric inhibition of microtubule formation by acetaldehyde-tubulin adducts

We have shown previously that acetaldehyde forms stable covalent adducts with tubulin, resulting in impaired microtubule formation. The present study explored the mechanism responsible for impaired microtubule formation caused by the substoichiometric stable binding of acetaldehyde to tubulin. The free tubulin dimer was much more reactive with acetaldehyde than microtubules, binding more than twice as much aldehyde. The dimer also formed nearly twice as many stable adducts on its alpha-chain as on its beta-chain, whereas microtubules exhibited an equal distribution of adducts between the two subunits. These data confirm that the alpha-chain of free tubulin, but not microtubules, has an accessible highly reactive lysine (HRL) residue that is a preferential target of acetaldehyde binding. Adduct formation with the HRL residue also correlated with impaired tubulin polymerization, and only 0.08 moles of acetaldehyde bound per mole of HRL was required for complete inhibition; however, adducts with other lysine residues (bulk adducts) did not affect assembly. Adducts to microtubule-associated proteins (MAPs) also impaired the assembly of tubulin, but were much less effective than HRL adducts. In a copolymerization assay, HRL-adducted tubulin, in addition to being itself assembly incompetent, also interfered with polymerization of normal (unadducted) tubulin. Bulk adducts did not alter assembly and were incorporated normally into the growing polymer. When tubulin was cleaved by the proteolytic enzyme, subtilisin, microtubule formation could readily take place in the absence of MAPs. In this polymerization system, HRL adducts, but not bulk adducts, still markedly inhibited assembly. When low concentrations of acetaldehyde (50 microM) were used to generate HRL adducts, an adduct on only 1 out of 20 tubulin molecules was sufficient to totally block polymerization. These findings indicate that substoichiometric amounts of acetaldehyde bound to HRL of tubulin can markedly inhibit microtubule formation via direct interference of dimer-dimer interactions, and further suggest that low concentrations of acetaldehyde could generate sufficient amounts of HRL adducts in cellular systems to alter microtubule formation and function.

Posttranslational modifications of nerve cytoskeletal proteins in experimental diabetes

Axonal transport is known to be impaired in peripheral nerve of experimentally diabetic rats. As axonal transport is dependent on the integrity of the neuronal cytoskeleton, we have studied the way in which rat brain and nerve cytoskeletal proteins are altered in experimental diabetes. Rats were made diabetic by injection of streptozotocin (STZ). Up to six weeks later, sciatic nerves, spinal cords, and brains were removed and used to prepare neurofilaments, microtubules, and a crude preparation of cytoskeletal proteins. The extent of nonenzymatic glycation of brain microtubule proteins and peripheral nerve tubulin was assessed by incubation with 3H-sodium borohydride followed by separation on two-dimensional polyacrylamide gels and affinity chromatography of the separated proteins. There was no difference in the nonenzymatic glycation of brain microtubule proteins from two-week diabetic and nondiabetic rats. Nor was the assembly of microtubule proteins into microtubules affected by the diabetic state. On the other hand, there was a significant increase in nonenzymatic glycation of sciatic nerve tubulin after 2 weeks of diabetes. We also identified an altered electrophoretic mobility of brain actin from a cytoskeletal protein preparation from brain of 2 week and 6 week diabetic rats. An additional novel polypeptide was demonstrated with a slightly more acidic isoelectric point than actin that could be immunostained with anti-actin antibodies. The same polypeptide could be produced by incubation of purified actin with glucose in vitro, thus identifying it as a product of nonenzymatic glycation. These results are discussed in relation to data from a clinical study of diabetic patients in which we identified increased glycation of platelet actin. STZ-diabetes also led to an increase in the phosphorylation of spinal cord neurofilament proteins in vivo during 6 weeks of diabetes. This hyperphosphorylation along with a reduced activity of a neurofilament-associated protein kinase led to a reduced incorporation of 32P into purified neurofilament proteins when they were incubated with 32P-ATP in vitro. Our combined data show a number of posttranslation modifications of neuronal cytoskeletal proteins that may contribute to the altered axonal transport and subsequent nerve dysfunction in experimental diabetes.

Effect of reductive alkylation on thermal stability of ribonuclease A and chymotrypsinogen A

In order to probe changes in the structural stability induced by the introduction of hydrophobic groups into proteins, the amino groups of ribonuclease A and chymotrypsinogen A were reductively alkylated by reaction with various aliphatic aldehydes, formaldehyde, acetaldehyde, n-butylaldehyde and n-hexylaldehyde, and their thermal stabilities were investigated by differential scanning calorimetry (DSC) at different acidic pH values. Ribonuclease A was thermally unstabilized by reductive alkylation, while chymotrypsinogen A was slightly stabilized, depending on both the size of the introduced alkyl groups and the extent of modification. These observations suggest that the effects induced by alkylation involve not only steric hindrance due to the entering bulky groups but also certain other factors such as the participation of the chemically introduced alkyl groups in hydrophobic interactions.

Glycation of rat sciatic nerve tubulin in experimental diabetes mellitus

Diabetic neuropathy is associated with some early defects of axonal transport in experimental animals. Axonal transport is dependent on intact microtubules, and unsubstituted lysine residues of tubulin are essential for microtubule polymerization. As lysine residues are the major target for the non-enzymatic attachment of glucose, the effect of diabetes on the extent of glycation of tubulin was investigated. There was a more than four-fold increase in the extent of glycation of tubulin in the sciatic nerve of rats with streptozotocin-induced diabetes of 2 weeks duration compared with control rats. In contrast, no such increase in glycation was observed in brain microtubule protein from diabetic rats at that stage of diabetes. Incubation of brain microtubule protein with glucose prior to in vitro polymerization showed that the early stages of glycation were not associated with inhibition of microtubule assembly. The observed glycation of peripheral nerve tubulin in early experimental diabetes may nevertheless contribute to axonal transport abnormalities through an as yet undetermined impairment of microtubule function.

Acetaldehyde and microtubules

Acetaldehyde covalently binds to tubulin to form stable and unstable adducts. Although tubulin has numerous lysine residues available to react with acetaldehyde, a key highly reactive lysine (HRL) on the alpha chain appears to be a preferential target for stable binding. The HRL residue is available for selective binding when tubulin is in the free (dimer) state but not when it is in the polymerized (microtubule) state. Stable binding of acetaldehyde to the HRL residue markedly inhibits tubulin assembly into microtubules, whereas stable binding to other residues (bulk adducts) has little influence on assembly. Substoichiometric stable binding of acetaldehyde to the HRL is sufficient to inhibit polymerization, via direct interference of tubulin dimer-dimer interactions, and an HRL adduct on only one out of 20 tubulin molecules can totally inhibit polymerization. These findings, along with our previous studies demonstrating impaired microtubule-dependent protein trafficking pathways in livers of ethanol-fed animals, indicate that low acetaldehyde concentrations, formed during ethanol oxidation in vivo, could generate sufficient amounts of HRL adducts on the alpha chain of tubulin in cellular systems to alter microtubule formation and function. In addition to alpha-tubulin, calmodulin and actin have also been found to have enhanced reactivity toward acetaldehyde. Thus, a general hypothesis to describe cellular injury induced by acetaldehyde adducts can be formulated: during ethanol oxidation, acetaldehyde forms stable adducts via binding to reactive lysine residues of preferential target proteins, resulting in selective functional impairment of these proteins and ultimately leading to cellular injury.

Levuglandin E2 inhibits mitosis and microtubule assembly

Levuglandin E2 (LGE2) is a gamma-keto aldehyde produced by rearrangement of the prostaglandin endoperoxide PGH2 under the aqueous conditions of its biosynthesis. We show that exogenous LGE2 enters cells and efficiently inhibits the first synchronous cell division of fertilized sea urchin eggs. We attribute this inhibition to covalent modification of tubulin and thereby to inhibition of microtubule assembly.

The active site composition of porcine pancreatic lipase: possible involvement of lysine

A mechanism is proposed wherein an essential lysine in porcine pancreatic lipase is the acylable residue in the catalytic mechanism of the enzyme. This mechanism involves an initial interfacial activation step were acylation first takes place in a rate-limiting step on a serine residue assisted by histidine and a carboxyl-containing residue, aspartic acid or glutamic acid, and then in a fast subsequent step the acyl group is transferred to the essential lysine residue at the catalytic site. Indirect support for the mechanism is presented. When the essential lysine is made inactive by reductive methylation, the lipase is functionally converted to a proteinase, as predicted by the mechanism.

Ethoxyformylation of tubulin with [3H]diethyl pyrocarbonate: a reexamination of the mechanism of assembly inhibition

In this study we reexamined the basis for the profound inhibitory effects of low concentrations of diethyl pyrocarbonate (DEP) on tubulin's ability to assemble into microtubules [cf. Lee, Y. C., Houston, L. I., & Himes, R. H. (1976) Biochem. Biophys. Res. Commun. 70, 50-56]. Assembly inhibition at low DEP concentrations can be resolved into two components: a component reversible with hydroxylamine (attributed to monoethoxyformylation of histidyl residues) that contributes approximately 40% of the inhibition and a hydroxylamine-resistant component (attributed to ethoxyformylation of non-histidyl residues) that contributes approximately 60% of the inhibition. Comparisons between the extent of assembly inhibition associated with each component and the degree of residue modification argue for the involvement of a small number of highly reactive residues in the inhibition process. To identify these residues, tubulin was reacted with limiting concentrations of [3H]DEP and subjected to tryptic digestion and HPLC analysis. Only one moderately reactive histidyl residue was detected. This residue (approximately 2-3-fold more reactive than the bulk histidyl residues) eluted in an apparently large, hydrophobic fragment. We failed to detect any non-histidyl residues that were exceptionally reactive to [3H]DEP. However, we did observe that the N-terminal methionyl residues in native protein were ethoxyformylated at rates comparable to that of the bulk histidyl residues. In denatured protein these methionyl residues were ethoxyformylated to a much larger extent (approximately 3-4-fold) than the bulk histidyl residues. We suggest that the N-terminal methionyl residues in tubulin are partly buried or are in a salt-bridge interaction in native protein and that ethoxyformylation of these residues disrupts tubulin structure and interferes with microtubule assembly.

Preferential covalent binding of acetaldehyde to the alpha-chain of purified rat liver tubulin

Hepatic ethanol oxidation generates the reactive intermediate acetaldehyde, which binds to proteins. Previous work, using bovine brain tubulin as a model protein, has shown that acetaldehyde preferentially formed stable adducts on the alpha-chain of the heterodimeric molecule. This binding resulted in functional impairment of the tubulin/microtubule system as evidenced by a decreased ability of adducted tubulin to form microtubules. Since tubulin/microtubules are believed to be very important cytoskeletal components of the hepatocyte and results with brain tubulin were interesting, our goal was to extend these studies to liver tubulin. We purified tubulin from rat liver by a polymerization-based cycle method followed by phosphocellulose chromatography. We then characterized the covalent binding of [14C]acetaldehyde to liver tubulin. Naturally forming and cyanoborohydride-stimulated stable adducts formed linearly with liver tubulin in a manner almost identical to that with brain tubulin. We also found that the alpha-chain of the native heterodimeric liver tubulin molecule was the preferred site of adduct formation at low acetaldehyde to protein ratios. These results confirm and extend our previous findings with the brain tubulin model and further suggest that the alpha-chain of tubulin may be a preferential site for acetaldehyde-adduct formation during ethanol oxidation in the liver.

The reaction of acetaldehyde with brain microtubular proteins: formation of stable adducts and inhibition of polymerization

Stable adducts were formed by treatment of bovine brain microtubular proteins (MTP) with acetaldehyde, followed by gel filtration to remove excess acetaldehyde. The extent of stable adduct formation was determined using [14C]acetaldehyde and was correlated with acetaldehyde concentration and reaction time. Significant inhibition of MTP polymerization was observed at adduct concentrations of 0.6 mol acetaldehyde/mol tubulin dimer. The data suggest that repeated exposure of MTP to low concentrations of acetaldehyde, as would occur in the brain and other tissues of alcoholics, may inhibit MTP polymerization with neurological consequences.

Effects of acetaldehyde on polymerization of microtubule proteins

The in vitro effects of ethanol and acetaldehyde on polymerization of calf brain microtubular proteins (MTP) were examined. While ethanol up to 100 mM had no effect on the polymerization of MTP, acetaldehyde above 0.5 mM had an inhibitory effect. This effect was not dependent on the presence of microtubule-associated proteins (MAPs), since acetaldehyde had a similar effect on the polymerization of highly purified tubulin. Electron microscopy revealed that the number and the length of microtubules at equilibrium was reduced by the presence of acetaldehyde. Acetaldehyde raised the critical concentration for tubulin assembly and caused greater inhibition at lower tubulin concentrations. Acetaldehyde augmented the depolymerizing effects of Ca2+ on preassembled microtubules. In addition, acetaldehyde itself caused depolymerization of microtubules but only in the absence of MAPs. Long-term (19.5 h) incubation of MTP with acetaldehyde led to significant loss of polymerization ability which could not be reversed by removal of acetaldehyde. This loss of activity was apparently independent of the observed formation of reducible adducts between acetaldehyde and MTP.

Covalent binding of acetaldehyde to tubulin: evidence for preferential binding to the alpha-chain

The covalent binding of [14C]acetaldehyde to purified beef brain tubulin was characterized. As we have found for several other proteins, tubulin bound acetaldehyde to form both stable and unstable adducts. Unstable adducts (Schiff bases) were stabilized, and rendered detectable, by treating incubated reaction mixtures with the reducing agent sodium borohydride. In short-term incubations, the majority of the adducts formed were unstable, but the percentage of total adducts that were stable gradually increased with time. Stable adduct formation was greatly increased by the inclusion of sodium cyanoborohydride in reaction mixtures (reductive ethylation). When reaction mixtures were submitted to sodium dodecyl sulfate-polyacrylamide gel electrophoresis to separate the alpha- and beta-chains of the heterodimeric tubulin molecule, the alpha-chain of free tubulin, but not intact microtubules, was the preferential site of stable adduct formation under both reductive and nonreductive conditions. Denaturation studies showed that the native tubulin conformation was necessary for the alpha-chain to show enhanced reactivity toward acetaldehyde. Competition binding studies showed that alpha-tubulin could effectively compete with beta-tubulin and bovine serum albumin for a limited amount of acetaldehyde. Unstable acetaldehyde adducts with free tubulin or microtubules did not exhibit alpha-chain selectivity. Analysis of reaction mixtures indicates that lysine residues are the major group of the protein participating in adduct formation. These data indicate that the alpha-chain of free tubulin is the preferential site of stable acetaldehyde-tubulin adduct formation. Further, these data raise the possibility that alpha-tubulin may be a selective target for acetaldehyde adduct formation in cellular systems.

The interaction of acetaldehyde with tubulin

Acetaldehyde covalently binds to purified tubulin in vitro to form both stable and unstable adducts. The formation of stable adducts can be greatly facilitated by the inclusion of the relatively gentle and Schiff base specific reducing agent, sodium cyanoborohydride. Although the tubulin molecule has multiple lysine resides available to react with acetaldehyde, certain key lysine residues on the alpha-chain appear to be selective targets for adduct formation. The formation of alpha-chain specific stable acetaldehyde-tubulin adducts results in functional impairment of the ability of tubulin to polymerize. Under relatively physiologic conditions where acetaldehyde-to-protein ratios are low, alpha-chain specific binding is prominent. These results, coupled with the studies presented in another report in this volume, raise the possibility that low levels of adduct formation may be detrimental to the structure or function of certain proteins (e.g. tubulin) in the liver. The alteration of this or other biologically important proteins by sustained low levels of adduct formation may contribute to the pathogenesis of alcoholic liver injury.

Microtubule assembly is dependent on a cluster of basic residues in alpha-tubulin

Previous studies have shown that tubulin, a major protein component of the microtubule, is rendered assembly incompetent when a highly reactive lysine residue (HRL) in the alpha polypeptide of tubulin dimer is reductively methylated [cf. Sherman, G., Rosenberry, T. L., & Sternlicht, H. (1983) J. Biol. Chem. 258, 2148-2156]. In this study we demonstrate that the HRL in bovine brain tubulin is Lys-394, a residue proximal in the alpha-tubulin sequence to the highly negatively charged carboxy-terminus region (residues 412-450) previously implicated in assembly. pH studies were undertaken to probe the local environment of Lys-394. These studies indicated that Lys-394 reactivity toward HCHO is sensitive to the titration of a pKa 6.3 group presumed to be a histidine residue. This assignment is supported by our finding that histidine modification via diethyl pyrocarbonate strongly affects Lys-394 reactivity toward HCHO as well as microtubule assembly. We propose on the basis of secondary structure considerations and published sequence data for a variety of tubulins that Lys-394 is part of an evolutionarily conserved cluster of basic residues (effective charge: 2+ to 2.5+ at neutral pH) composed of Lys-394, His-393, and Arg-390, which is important for tubulin function and which renders Lys-394 reactive as a nucleophile.