Crystal structure of human L-isoaspartyl methyltransferase.

The enzyme l-isoaspartyl methyltransferase initiates the repair of damaged proteins by recognizing and methylating isomerized and racemized aspartyl residues in aging proteins. The crystal structure of the human enzyme containing a bound S-adenosyl-l-homocysteine cofactor is reported here at a resolution of 2.1 A. A comparison of the human enzyme to homologs from two other species reveals several significant differences among otherwise similar structures. In all three structures, we find that three conserved charged residues are buried in the protein interior near the active site. Electrostatics calculations suggest that these buried charges might make significant contributions to the energetics of binding the charged S-adenosyl-l-methionine cofactor and to catalysis. We suggest a possible structural explanation for the observed differences in reactivity toward the structurally similar l-isoaspartyl and d-aspartyl residues in the human, archael, and eubacterial enzymes. Finally, the human structure reveals that the known genetic polymorphism at residue 119 (Val/Ile) maps to an exposed region away from the active site.     

Widespread phylogenetic distribution of a protein methyltransferase that modifies L-isoaspartyl residues.

Protein L-isoaspartyl methyltransferase is implicated in the repair or degradation of age-damaged proteins that contain atypical, L-isoaspartyl residues. The enzyme has previously been demonstrated in a variety of vertebrates and in the bacterium S. typhimurium (O'Connor, C.M. and Clarke, S. (1985) Biochem. Biophys. Res. Commun. 132, 1144-1150). We report here that the enzyme is present in a mollusc (great slug), a crustacean (pill woodlouse), a fungus (mushroom), and a plant (wheat germ). Using mushroom as an example, we show that the enzyme activity may, in some instances, require a partial purification before its presence is clearly detectable. Our findings significantly extend the known phylogenetic distribution of this enzyme and suggest that it may play an indispensable role in protein metabolism.     

Structural elements affecting the recognition of L-isoaspartyl residues by the L-isoaspartyl/D-aspartyl protein methyltransferase. Implications for the repair hypothesis.

We have synthesized a series of L-isoaspartyl-containing (isoD) peptides and characterized their interaction with the human erythrocyte L-isoaspartyl/D-aspartyl protein methyltransferase (EC 2.1.1.77). Our findings indicate that this enzyme interacts with 6 residues extending from the isoD-2 to isoD+3 positions in peptide substrates. Although peptides as simple as G-isoD-G are methylated with low affinity (Km = 17.8 mM), a wide variety of L-isoaspartyl-containing sequences in larger peptides are recognized with high affinity (Km less than 20 microM), the best yet discovered being VYP-isoD-HA, with a Km of 0.29 microM. Only two sequence elements have been found that can interfere with the high affinity binding of peptides of 4 or more residues, these being a prolyl residue in the isoD+1 position and negatively charged residues in the isoD+1, isoD+2, and/or isoD+3 positions. We investigated the effect of higher order structure on binding affinity using several L-isoaspartyl-containing proteins. Although conformation did, in some cases, lower the affinity of the methyltransferase for L-isoaspartyl residues, the range of kinetic constants for the methylation of these proteins was similar to that observed with the synthetic peptides. The L-isoaspartyl/D-aspartyl methyltransferase has been proposed to function in vivo to prevent the accumulation of L-isoaspartyl residues that arise spontaneously as proteins age. To examine whether such a mechanism is feasible given the wide range of substrate Km values observed in vitro, we set up a computer simulation to model the degradation and methylation reactions in aging human erythrocytes. Our results suggest that enough methyltransferase activity exists in these cells to significantly lower the expected number of L-isoaspartyl residues, even when these residues have millimolar Km values for methylation.     

Catalytic implications from the Drosophila protein L-isoaspartyl methyltransferase structure and site-directed mutagenesis

Protein L-isoaspartyl methyltransferases (PIMT; EC 2.1.1.77) catalyze the S-adenosylmethionine-dependent methylation of L-isoaspartyl residues that arise spontaneously in proteins with age, thereby initiating a repair process that restores the normal backbone configuration to the damaged polypeptide. In Drosophila melanogaster, overexpression of PIMT in transgenic flies extends the normal life span, suggesting that protein damage can be a limiting factor in longevity. To understand structural features of the Drosophila PIMT (dPIMT) important for catalysis, the crystal structure of dPIMT was determined at a resolution of 2.2 A, and site-directed mutagenesis was used to identify the role of Ser-60 in catalysis. The core structure of dPIMT is similar to the modified nucleotide-binding fold observed in PIMTs from extreme thermophiles and humans. A striking difference of the dPIMT structure is the rotation of the C-terminal residues by 90 degrees relative to the homologous structures. Effectively, this displacement generates a more open conformation that allows greater solvent access to S-adenosylhomocysteine, which is almost completely buried in other PIMT structures. The enzyme may alternate between the open conformation found for dPIMT and the more closed conformations described for other PIMTs during its catalytic cycle, thereby allowing the exchange of substrates and products. Catalysis by dPIMT requires the side chain of the conserved, active site residue Ser-60, since substitution of this residue with Thr, Gln, or Ala reduces or abolishes the methylation of both protein and isoaspartyl peptide substrates.     

Crystal structure of a protein repair methyltransferase from Pyrococcus furiosus with its L-isoaspartyl peptide substrate.

Protein L-isoaspartyl (D-aspartyl) methyltransferases (EC 2.1.1.77) are found in almost all organisms. These enzymes catalyze the S-adenosylmethionine (AdoMet)-dependent methylation of isomerized and racemized aspartyl residues in age-damaged proteins as part of an essential protein repair process. Here, we report crystal structures of the repair methyltransferase at resolutions up to 1.2 A from the hyperthermophilic archaeon Pyrococcus furiosus. Refined structures include binary complexes with the active cofactor AdoMet, its reaction product S-adenosylhomocysteine (AdoHcy), and adenosine. The enzyme places the methyl-donating cofactor in a deep, electrostatically negative pocket that is shielded from solvent. Across the multiple crystal structures visualized, the presence or absence of the methyl group on the cofactor correlates with a significant conformational change in the enzyme in a loop bordering the active site, suggesting a role for motion in catalysis or cofactor exchange. We also report the structure of a ternary complex of the enzyme with adenosine and the methyl-accepting polypeptide substrate VYP(L-isoAsp)HA at 2.1 A. The substrate binds in a narrow active site cleft with three of its residues in an extended conformation, suggesting that damaged proteins may be locally denatured during the repair process in cells. Manual and computer-based docking studies on different isomers help explain how the enzyme uses steric effects to make the critical distinction between normal L-aspartyl and age-damaged L-isoaspartyl and D-aspartyl residues.     

Purification, gene cloning, and sequence analysis of an L-isoaspartyl protein carboxyl methyltransferase from Escherichia coli

Mammalian tissues contain protein carboxyl methyltransferases that catalyze the transfer of methyl groups from S-adenosylmethionine to the free carboxyl groups of D-aspartyl or L-isoaspartyl residues (EC 2.1.1.77). These enzymes have been postulated to play a role in the repair and/or degradation of spontaneously damaged proteins. We have now characterized a similar activity from Escherichia coli that recognizes L-isoaspartyl-containing peptides as well as protein substrates such as ovalbumin. The enzyme was purified by DEAE-cellulose, hydroxylapatite, Sephadex G-100, polyaspartate, and reversed-phase chromatography and was shown to consist of a single 24-kDa polypeptide chain. The sequence determined for the N-terminal 39 residues was used to design an oligonucleotide probe that allowed the precise localization of its structural gene (pcm) on the physical map of the E. coli chromosome at 59 min. Transformation of E. coli cells with a plasmid containing DNA from this region results in a 3-4-fold overproduction of enzyme activity. The nucleotide sequence determined for the pcm gene and its flanking regions was used to deduce a mature amino acid sequence of 207 residues with a calculated molecular weight of 23,128. This sequence shows 30.8% sequence identity with the human L-isoaspartyl/D-aspartyl methyltransferase and suggests that this enzyme catalyzes a fundamental reaction in both procaryotic and eucaryotic cells.     

Synapsin I is a major endogenous substrate for protein L-isoaspartyl methyltransferase in mammalian brain

The accumulation of potentially deleterious L-isoaspartyl linkages in proteins is prevented by the action of protein L-isoaspartyl O-methyltransferase, a widely distributed enzyme that is particularly active in mammalian brain. Methyltransferase-deficient (knock-out) mice exhibit greatly increased levels of isoaspartate and typically succumb to fatal epileptic seizures at 4-10 weeks of age. The link between isoaspartate accumulation and the neurological abnormalities of these mice is poorly understood. Here, we demonstrate that synapsin I from knock-out mice contains 0.9 +/- 0.3 mol of isoaspartate/mol of synapsin, whereas the levels in wild-type and heterozygous mice are undetectable. Transgenic mice that selectively express methyltransferase only in neurons show reduced levels of synapsin damage, and the degree of reduction correlates with the phenotype of these mice. Isoaspartate levels in synapsin from the knock-out mice are five to seven times greater than those in the average protein from brain cytosol or from a synaptic vesicle-enriched fraction. The isoaspartyl sites in synapsin from knock-out mice are efficiently repaired in vitro by incubation with purified methyltransferase and S-adenosyl-L-methionine. These findings demonstrate that synapsin I is a major substrate for the isoaspartyl methyltransferase in neurons and suggest that isoaspartate-related alterations in the function of presynaptic proteins may contribute to the neurological abnormalities of mice deficient in this enzyme.     

Down-regulation of protein L-isoaspartyl methyltransferase in human epileptic hippocampus contributes to generation of damaged tubulin

Protein L-isoaspartyl methyltransferase (PIMT) repairs the damaged proteins which have accumulated abnormal aspartyl residues during cell aging. Gene targeting has elucidated a physiological role for PIMT by showing that mice lacking PIMT died prematurely from fatal epileptic seizures. Here we investigated the role of PIMT in human mesial temporal lobe epilepsy. Using surgical specimens of hippocampus and neocortex from controls and epileptic patients, we showed that PIMT activity and expression were 50% lower in epileptic hippocampus than in controls but were unchanged in neocortex. Although the protein was down-regulated, PIMT mRNA expression was unchanged in epileptic hippocampus, suggesting post-translational regulation of the PIMT level. Moreover, several proteins with abnormal aspartyl residues accumulate in epileptic hippocampus. Microtubules component beta-tubulin, one of the major PIMT substrates, had an increased amount (two-fold) of L-isoaspartyl residues in the epileptic hippocampus. These results demonstrate that the down-regulation of PIMT in epileptic hippocampus leads to a significant accumulation of damaged tubulin that could contribute to neuron dysfunction in human mesial temporal lobe epilepsy.     

Distinct patterns of expression but similar biochemical properties of protein L-isoaspartyl methyltransferase in higher plants

Protein L-isoaspartyl methyltransferase is a widely distributed repair enzyme that initiates the conversion of abnormal L-isoaspartyl residues to their normal L-aspartyl forms. Here we show that this activity is expressed in developing corn (Zea mays) and carrot (Daucus carota var. Danvers Half Long) plants in patterns distinct from those previously seen in winter wheat (Triticum aestivum cv Augusta) and thale cress (Arabidopsis thaliana), whereas the pattern of expression observed in rice (Oryza sativa) is similar to that of winter wheat. Although high levels of activity are found in the seeds of all of these plants, relatively high levels of activity in vegetative tissues are only found in corn and carrot. The activity in leaves was found to decrease with aging, an unexpected finding given the postulated role of this enzyme in repairing age-damaged proteins. In contrast with the situation in wheat and Arabidopsis, we found that osmotic or salt stress could increase the methyltransferase activity in newly germinated seeds (but not in seeds or seedlings), whereas abscisic acid had no effect. We found that the corn, rice, and carrot enzymes have comparable affinity for methyl-accepting substrates and similar optimal temperatures for activity of 45 degrees C to 55 degrees C as the wheat and Arabidopsis enzymes. These experiments suggest that this enzyme may have specific roles in different plant tissues despite a common catalytic function.     

Reactive oxygen species generated by thiol-modifying phenylarsine oxide stimulate the expression of protein L-isoaspartyl methyltransferase

Expression of the repair enzyme protein l-isoaspartyl methyltransferase (PIMT) has been reported to play important roles in brain. However, little is known about the regulation of PIMT expression following protein damage by oxidation in brain. Phenylarsine oxide (PAO) is an arsenical compound that alters proteins by forming disulfide bond with vicinal cysteinyl residues. Here we report that PIMT was rapidly up-regulated by PAO in U-87 human astroglioma cells. We also confirmed that PIMT up-regulation by PAO was mediated by the reaction with vicinal cysteines. Furthermore, we showed that PIMT induction by PAO was dependent on formation of reactive oxygen species (ROS). Crucially, both ROS formation and PIMT induction by PAO were inhibited by antioxidant N-acetyl-l-cysteine and NADPH oxidase inhibitor diphenyleneiodonium chloride. Importantly, down-regulation of PIMT by siRNA strikingly enhanced PAO-induced ROS. Together, these results highlight that PIMT expression is regulated by ROS and could primarily act as an antioxidant enzyme.     

The type II isoform of bovine brain protein L-isoaspartyl methyltransferase has an endoplasmic reticulum retention signal (...RDEL) at its C-terminus.

Bovine brain is known to contain two major isoforms of protein L-isoaspartyl methyltransferase (PIMT), an enzyme that facilitates repair of atypical L-isoaspartyl peptide bonds in proteins. Although the two isoforms can be separated by anion-exchange chromatography, they appear to have similar, if not identical, substrate specificities in vitro. The more basic type I isoform has been extensively characterized, and its complete sequence has been reported. The present study was undertaken in an attempt to understand the structural and functional uniqueness of the more acidic type II isoform. Electrospray mass spectrometry of the intact enzymes revealed that the type II isoform is approximately 43 amu heavier than the type I isoform. Cyanogen bromide cleavage followed by HPLC with on-line mass analysis revealed that the type II isoform contains a unique C-terminal fragment which is 43 amu heavier than the corresponding fragment from the type I isoform. Amino acid composition analysis and direct sequencing of this fragment indicate that the type II isoform ends in the sequence ...RDEL, while the type I is known to end in ...RWK. Since ...RDEL, like ...KDEL, serves as an effective endoplasmic reticulum retention signal, we propose that the type II isoform serves to repair damaged proteins within the endoplasmic reticulum or, perhaps, within some other specialized compartment of the cell. Comparison of the protein sequences of the two bovine brain isoforms to DNA sequences for rodent PIMT reported by others suggests that the type II isoform may be produced by splicing within the codon for Arg224.     

Optimal conditions for the use of protein L-isoaspartyl methyltransferase in assessing the isoaspartate content of peptides and proteins.

Protein L-isoaspartyl methyltransferase provides a basis for enzymatic measurement of atypical, isoaspartyl linkages which make a major contribution to protein microheterogeneity. The low Vmax of the methyltransferase reaction and the instability of the methyl ester can hinder accurate determinations, and different laboratories using different conditions have achieved discrepant values for the isoaspartate content of the same proteins. To investigate the effects of these conditions, and to optimize the assay, isoaspartyl delta sleep-inducing peptide was methylated under a variety of conditions. We found that 1 microM methyltransferase was required to obtain stoichiometric modification of 2 microM peptide in 40-min reactions at pH 6.2 and 30 degrees C. A computer model utilizing kinetic constants obtained from studies on initial rates of methylation predicted the same requirement for enzyme concentration. Carrier protein was necessary for optimal methyltransferase activity at enzyme concentrations below 0.4 microM. Stoichiometric methylation required concentrations of S-adenosylmethionine to be in substantial excess over those of peptide; 50 microM S-adenosylmethionine is the minimum needed for complete modification of 10 microM peptide. Spontaneous demethylation was significant under all conditions tested, so that the methyl ester itself never reached a ratio of 1 mol/mol of total peptide. These results demonstrate that the most accurate measurements of isoaspartate are obtained when reactions are carried out at low peptide concentrations, high S-adenosylmethionine concentrations, and high enzyme concentrations.(ABSTRACT TRUNCATED AT 250 WORDS)     

Considerations in the identification of endogenous substrates for protein L-isoaspartyl methyltransferase: the case of synuclein

Protein L-isoaspartyl methyltransferase (PIMT) repairs abnormal isoaspartyl peptide bonds in age-damaged proteins. It has been reported that synuclein, a protein implicated in neurodegenerative diseases, is a major target of PIMT in mouse brain. To extend this finding and explore its possible relevance to neurodegenerative diseases, we attempted to determine the stoichiometry of isoaspartate accumulation in synuclein in vivo and in vitro. Brain proteins from PIMT knockout mice were separated by 2D electrophoresis followed by on-blot [(3)H]-methylation to label isoaspartyl proteins, and by immunoblotting to confirm the coincident presence of synuclein. On-blot (3)H-methylation revealed numerous isoaspartyl proteins, but no signal in the position of synuclein. This finding was corroborated by immunoprecipitation of synuclein followed by on-blot (3)H-methylation. To assess the propensity of synuclein to form isoaspartyl sites in vitro, samples of recombinant mouse and human α-synucleins were aged for two weeks by incubation at pH 7.5 and 37°C. The stoichiometries of isoaspartate accumulation were extremely low at 0.02 and 0.07 mol of isoaspartate per mol of protein respectively. Using a simple mathematical model based on the first order kinetics of isoaspartyl protein methyl ester hydrolysis, we ascribe the discrepancy between our results and the previous report to methodological limitations of the latter stemming from an inherent, and somewhat counterintuitive, relationship between the propensity of proteins to form isoaspartyl sites and the instability of the (3)H-methyl esters used to tag them. The results presented here indicate that synuclein is not a major target of PIMT in vivo, and emphasize the need to minimize methyl ester hydrolysis when using methylation to assess the abundance of isoaspartyl sites in proteins.     

Distribution of 10-formyltetrahydrofolate synthetase in eubacteria.

The distribution of 10-formyltetrahydrofolate synthetase, which activates formate for use as a one-carbon donor in a variety of biosynthetic reactions, was determined for a variety of eubacteria. Organisms from several genera were found to lack detectable synthetase activity; however, all organisms tested were found to contain 5,10-methylenetetrahydrofolate dehydrogenase activity.     

Recognition of D-aspartyl residues in polypeptides by the erythrocyte L-isoaspartyl/D-aspartyl protein methyltransferase. Implications for the repair hypothesis.

We provide here the first direct evidence that D-aspartyl residues in peptides are substrates for the L-isoaspartyl/D-aspartyl protein carboxyl methyltransferase (EC 2.1.1.77). We do this by showing that D-aspartic acid beta-methyl ester can be isolated from carboxypeptidase Y digests of enzymatically methylated D-aspartyl-containing synthetic peptides. The specificity of this reaction is supported by the lack of methylation of L-aspartyl-containing peptides under similar conditions. Methylation of D-aspartyl residues in synthetic peptides was not observed previously because with Km values ranging from 2.5 to 4.8 mM, these peptides are recognized by the methyltransferase with 700-10,000-fold lower affinity than are their L-isoaspartyl-containing counterparts. The physiological significance of D-aspartyl methylation was investigated in two ways. First, analysis of in situ methylated human erythrocyte proteins showed that at least 22% of the methyl groups associated with the proteins ankyrin and band 4.1 are on D-aspartyl residues, suggesting that D-aspartyl methylation is an important function of the methyltransferase in vivo. Second, mathematical modeling of the protein aging and methylation reactions occurring in intact erythrocytes indicated that the accumulation of D-aspartyl residues can be reduced as much as 2-5-fold by the methyltransferase activity. Although this reduction is much less than that predicted for L-isoaspartyl residues, it may be significant in maintaining functional proteins throughout the 120-day life span of these cells.     

The primary structure of a protein carboxyl methyltransferase from bovine brain that selectively methylates L-isoaspartyl sites

Protein L-isoaspartyl methyltransferase (PIMT) transfers the methyl group of S-adenosyl-L-methionine to free alpha-carboxyl groups of atypical L-isoaspartyl residues in proteins. The complete primary structure of the type I isoform of bovine brain PIMT was determined by sequence analysis of peptides generated by endoprotease Lys-C, trypsin, cyanogen bromide, and endoprotease Asp-N digests. The correct composition of every peptide was verified by fast atom bombardment mass spectrometry. The efficiency of sequencing by tandem mass spectrometry was examined for several peptides by comparing its speed and accuracy with automated Edman degradation. Tandem mass spectrometry was used to determine the structure of the NH2-terminal blocked peptide derived from a hydroxylamine cleavage. PIMT is 226 residues with Mr = 24,500 and contains acetyl alanine as the amino-terminal residue. The partial sequence (141 residues from 8 tryptic peptides) of a homologous human red cell PIMT (Gilbert, J. M., Fowler, A., Bleibaum, J., and Clarke, S. (1988) Biochemistry 27, 5227-5233) shows a 97% identity with the corresponding peptides of the bovine brain enzyme. The complete brain enzyme sequence reported here bears no significant homology to any other known class of methyltransferase including those which methylate the side chain gamma-carboxyl group of receptor proteins involved in bacterial chemotaxis.     

Protein repair in the brain, proteomic analysis of endogenous substrates for protein L-isoaspartyl methyltransferase in mouse brain

Protein L-isoaspartyl methyltransferase (PIMT) catalyzes repair of L-isoaspartyl peptide bonds, a major source of protein damage under physiological conditions. PIMT knock-out (KO) mice exhibit brain enlargement and fatal epileptic seizures. All organs accumulate isoaspartyl proteins, but only the brain manifests an overt pathology. To further explore the role of PIMT in brain function, we undertook a global analysis of endogenous substrates for PIMT in mouse brain. Extracts from PIMT-KO mice were subjected to two-dimensional gel electrophoresis and blotted onto membranes. Isoaspartyl proteins were radiolabeled on-blot using [methyl-(3)H]S-adenosyl-L-methionine and recombinant PIMT. Fluorography of the blot revealed 30-35 (3)H-labeled proteins, 22 of which were identified by peptide mass fingerprinting. These isoaspartate-prone proteins represent a wide range of cellular functions, including neuronal development, synaptic transmission, cytoskeletal structure and dynamics, energy metabolism, nitrogen metabolism, pH homeostasis, and protein folding. The following five proteins, all of which are rich in neurons, accumulated exceptional levels of isoaspartate: collapsin response mediator protein 2 (CRMP2/ULIP2/DRP-2), dynamin 1, synapsin I, synapsin II, and tubulin. Several of the proteins identified here are prone to age-dependent oxidation in vivo, and many have been identified as autoimmune antigens, of particular interest because isoaspartate can greatly enhance the antigenicity of self-peptides. We propose that the PIMT-KO phenotype results from the cumulative effect of isoaspartate-related damage to a number of the neuron-rich proteins detected in this study. Further study of the isoaspartate-prone proteins identified here may help elucidate the molecular basis of one or more developmental and/or age-related neurological diseases.     

Distinct C-terminal sequences of isozymes I and II of the human erythrocyte L-isoaspartyl/D-aspartyl protein methyltransferase

We have purified the more acidic major isozyme (II) of the human erythrocyte L-isoaspartyl/D-aspartyl methyltransferase and compared its structure to that of the previously sequenced isozyme I. These isozymes are both monomers of 25,000 molecular weight polypeptides and have similar enzymatic properties, but have isoelectric points that differ by one pH unit. Analysis of 16 tryptic peptides of isozyme II accounting for 89% of the sequence of isozyme I revealed no differences between these enzyme forms. However, analysis of a Staphylococcal V8 protease C-terminal fragment revealed that the last two residues of these proteins differed. The Trp-Lys-COOH terminus of isozyme I is replaced by a Asp-Asp-COOH terminus in isozyme II. Southern blot analysis of genomic DNA suggests that the human genome [corrected] may contain only a single gene encoding the enzyme. We propose that the distinct C-termini of isozymes I and II can arise from the generation of multiple mRNA's by alternative splicing.