Determinants of substrate specificity for saccharopine dehydrogenase from Saccharomyces cerevisiae

A survey of NADH, alpha-Kg, and lysine analogues has been undertaken in an attempt to define the substrate specificity of saccharopine dehydrogenase and to identify functional groups on all substrates and dinucleotides important for substrate binding. A number of NAD analogues, including NADP, 3-acetylpyridine adenine dinucleotide (3-APAD), 3-pyridinealdehyde adenine dinucleotide (3-PAAD), and thionicotinamide adenine dinucleotide (thio-NAD), can serve as a substrate in the oxidative deamination reaction, as can a number of alpha-keto analogues, including glyoxylate, pyruvate, alpha-ketobutyrate, alpha-ketovalerate, alpha-ketomalonate, and alpha-ketoadipate. Inhibition studies using nucleotide analogues suggest that the majority of the binding energy of the dinucleotides comes from the AMP portion and that distinctly different conformations are generated upon binding of the oxidized and reduced dinucleotides. Addition of the 2'-phosphate as in NADPH causes poor binding of subsequent substrates but has little effect on coenzyme binding and catalysis. In addition, the 10-fold decrease in affinity of 3-APAD in comparison to NAD suggests that the nicotinamide ring binding pocket is hydrophilic. Extensive inhibition studies using aliphatic and aromatic keto acid analogues have been carried out to gain insight into the keto acid binding pocket. Data suggest that a side chain with three carbons (from the alpha-keto group up to and including the side chain carboxylate) is optimal. In addition, the distance between the C1-C2 unit and the C5 carboxylate of the alpha-keto acid is also important for binding; the alpha-oxo group contributes a factor of 10 to affinity. The keto acid binding pocket is relatively large and flexible and can accommodate the bulky aromatic ring of a pyridine dicarboxylic acid and a negative charge at the C3 but not the C4 position. However, the amino acid binding site is hydrophobic, and the optimal length of the hydrophobic portion of the amino acid carbon side chain is three or four carbons. In addition, the amino acid binding pocket can accommodate a branch at the gamma-carbon, but not at the beta-carbon.     

Val216 decides the substrate specificity of alpha-glucosidase in Saccharomyces cerevisiae.

Differences in the substrate specificity of alpha-glucosidases should be due to the differences in the substrate binding and the catalytic domains of the enzymes. To elucidate such differences of enzymes hydrolyzing alpha-1,4- and alpha-1,6-glucosidic linkages, two alpha-glucosidases, maltase and isomaltase, from Saccharomyces cerevisiae were cloned and analyzed. The cloned yeast isomaltase and maltase consisted of 589 and 584 amino acid residues, respectively. There was 72.1% sequence identity with 165 amino acid alterations between the two alpha-glucosidases. These two alpha-glucosidase genes were subcloned into the pKP1500 expression vector and expressed in Escherichia coli. The purified alpha-glucosidases showed the same substrate specificities as those of their parent native glucosidases. Chimeric enzymes constructed from isomaltase by exchanging with maltase fragments were characterized by their substrate specificities. When the consensus region II, which is one of the four regions conserved in family 13 (alpha-amylase family), is replaced with the maltase type, the chimeric enzymes alter to hydrolyze maltose. Three amino acid residues in consensus region II were different in the two alpha-glucosidases. Thus, we modified Val216, Gly217, and Ser218 of isomaltase to the maltase-type amino acids by site-directed mutagenesis. The Val216 mutant was altered to hydrolyze both maltose and isomaltose but neither the Gly217 nor the Ser218 mutant changed their substrate specificity, indicating that Val216 is an important residue discriminating the alpha-1,4- and 1,6-glucosidic linkages of substrates.     

Mitochondrial methionyl-tRNAfMet formyltransferase from Saccharomyces cerevisiae: gene disruption and tRNA substrate specificity

Initiation of protein synthesis in bacteria, mitochondria, and chloroplasts involves a formylated methionyl-tRNA species. Formylation of this tRNA is catalyzed by a methionyl-tRNA(f)(Met) formyltransferase (formylase). Upon inactivation of the gene encoding formylase, the growth rate of Escherichia coli is severely decreased. This behavior underlines the importance of formylation to give tRNA(Met) an initiator identity. Surprisingly, however, recent data [Li, Y., Holmes, W. B., Appling, D. R., and RajBhandary, U. L. (2000) J. Bacteriol. 182, 2886-2892] showed that the respiratory growth of Saccharomyces cerevisiaewas not sensitive to deprivation of the mitochondrial formylase. In the present study, we report conditions of temperature or of growth medium composition in which inactivation of the formylase gene indeed impairs the growth of a S. cerevisiae haploid strain. Therefore, some selective advantage can eventually be associated to the existence of a formylating activity in the fungal mitochondrion under severe growth conditions. Finally, the specificity toward tRNA of S. cerevisiae mitochondrial formylase was studied using E. coli initiator tRNA and mutants derived from it. Like its bacterial counterpart, this formylase recognizes nucleotidic features in the acceptor stem of mitochondrial initiator tRNA. This behavior markedly distinguishes the mitochondrial formylase of yeast from that of animals. Indeed, it was shown that bovine mitochondrial formylase mainly recognizes the side chain of the esterified methionine plus a purine-pyrimidine base pair in the D-stem of tRNA [Takeuchi, N., Vial, L., Panvert, M., Schmitt, E., Watanabe, K., Mechulam, Y., and Blanquet, S. (2001) J. Biol. Chem. 276, 20064-20068]. Distinct tRNA recognition mechanisms adopted by the formylases of prokaryotic, fungal, or mammalian origins are likely to reflect coevolution of these enzymes with their tRNA substrate. Each mechanism appears well suited to an efficient selection of the substrate within the pool of all tRNAs.     

Saccharomyces cerevisiae nucleoside-diphosphate kinase: purification, characterization, and substrate specificity.

Nucleoside-diphosphate kinase is an enzyme which catalyzes the phosphorylation of nucleoside diphosphates into the corresponding triphosphates for nucleic acid biosynthesis. In this communication, we describe the purification and characterization of nucleoside-diphosphate kinase from yeast. The purified protein appears to be homogeneous by sodium dodecyl sulfate-polyacrylamide gel analysis, with a molecular weight of about 17,000-18,000. An estimate from the fast protein liquid chromatography Superose 12 gel filtration shows a native molecular weight of about 68,000 to 70,000. The results suggest that yeast nucleoside-diphosphate kinase is composed of four subunits. Substrate specificity studies show that the relative activity of nucleoside diphosphates (NDP) as phosphate acceptors is in the order of dTDP greater than CDP greater than UDP greater than dUDP greater than GDP greater than or equal to dGDP greater than dCDP greater than dADP greater than ADP; and the relative activity of triphosphate donors is in the order of UTP greater than dTTP greater than CTP greater than dCTP greater than dATP greater than ATP greater than or equal to dGTP greater than GTP. The Km and Vm of dTDP, dGDP, dCDP, dUDP, CDP, and UDP have been determined. The rate constant studies indicate that the purified NDP kinase prefers using, to a slight extent, dTDP (approximately 800 min-1) as the substrate rather than other tested deoxyribo- and ribonucleotides (350-450 min-1). The broad substrate specificity and kinetic data suggest that the enzyme is involved in both DNA and RNA metabolism.     

Induction and substrate specificity of the lanosterol 14 alpha-demethylase from Saccharomyces cerevisiae Y222.

The potential inducibility of the lanosterol 14 alpha-demethylase (P-45014DM) from Saccharomyces cerevisiae Y222 by xenobiotics was investigated. This enzyme and NADPH-cytochrome P-450 reductase were unaffected by a number of compounds known to induce mammalian and some yeast cytochrome P-450 monooxygenases. Furthermore, dibutyryl cyclic AMP did not affect P-45014DM or P-450 reductase levels, while growth at 37 degrees C resulted in a slight decrease. P-45014DM was found to be specific for lanosterol and did not metabolize a number of P-450 substrates including benzo[a]pyrene.     

A 5'----3' exoribonuclease of Saccharomyces cerevisiae: size and novel substrate specificity

The purification scheme for a 5'----3' exoribonuclease of Saccharomyces cerevisiae has been modified to facilitate purification of larger amounts of enzyme and further extended to yield highly purified enzyme by use of poly(A)-agarose chromatography. As determined by either sodium dodecyl sulfate-polyacrylamide gel electrophoresis or physical characterization, the enzyme has a molecular weight of about 160,000. Further studies of its substrate specificity show that poly(C) and poly(U) preparations require 5' phosphorylation for activity and that poly(A) with a 5'-triphosphate end group is hydrolyzed at only 12% of the rate of poly(A) with a 5'-monophosphate end group. DNA is not hydrolyzed, but synthetic polydeoxyribonucleotides are strong competitive inhibitors of the hydrolysis of noncomplementary ribopolymers. Poly(A).poly(U) and poly(A).poly(dT) are hydrolyzed at 60 and 50%, respectively, of the rate of poly(A) at 37 degrees C. The RNase H activity of the enzyme can also be demonstrated using an RNA X M13 DNA hybrid as a substrate. When poly(dT).poly(dA) with a 5'-terminal poly(A) segment on the poly(dA) is used as a substrate, the enzyme hydrolyzes the poly(A) "tail," removing the last ribonucleotide, but does not hydrolyze the poly(dA).     

The crystal structure of the cytosolic exopolyphosphatase from Saccharomyces cerevisiae reveals the basis for substrate specificity

Inorganic long-chain polyphosphate is a ubiquitous linear polymer in biology, consisting of many phosphate moieties linked by phosphoanhydride bonds. It is synthesized by polyphosphate kinase, and metabolised by a number of enzymes, including exo- and endopolyphosphatases. The Saccharomyces cerevisiae gene PPX1 encodes for a 45 kDa, metal-dependent, cytosolic exopolyphosphatase that processively cleaves the terminal phosphate group from the polyphosphate chain, until inorganic pyrophosphate is all that remains. PPX1 belongs to the DHH family of phosphoesterases, which includes: family-2 inorganic pyrophosphatases, found in Gram-positive bacteria; prune, a cyclic AMPase; and RecJ, a single-stranded DNA exonuclease. We describe the high-resolution X-ray structures of yeast PPX1, solved using the multiple isomorphous replacement with anomalous scattering (MIRAS) technique, and its complexes with phosphate (1.6 A), sulphate (1.8 A) and ATP (1.9 A). Yeast PPX1 folds into two domains, and the structures reveal a strong similarity to the family-2 inorganic pyrophosphatases, particularly in the active-site region. A large, extended channel formed at the interface of the N and C-terminal domains is lined with positively charged amino acids and represents a conduit for polyphosphate and the site of phosphate hydrolysis. Structural comparisons with the inorganic pyrophosphatases and analysis of the ligand-bound complexes lead us to propose a hydrolysis mechanism. Finally, we discuss a structural basis for substrate selectivity and processivity.     

Mitochondrial proteinases of the yeast Saccharomyces cerevisiae: electrophoretic detection, substrate specificity and sensitivity to inhibitors

The proteins of submitochondrial particles solubilized with 0.1% Triton X-100 were separated by polyacrylamide gel electrophoresis. Hydrolysis of several proteinase substrates was registered directly in the gel after completion of electrophoresis. According to the data obtained the inner mitochondrial membrane contains one or two enzymes which catalyze hydrolysis of cytochrome c as well as one or two enzymes splitting synthetic substrate of trypsin-like proteinases, e. g. N-alpha-benzoyl-L-arginine-p-nitroanilide (BAPA) and N-alpha-benzoyl-L-arginine-beta-naphthylamide (BANA). Submitochondrial particles were shown to catalyze hydrolysis of 3H-labelled cytochrome c. This activity is suppressed by the same inhibitors as the hydrolysis of mitochondrial translation products, i. e. phenyl-methylsulfonylfluoride, p-chloromercuribenzosulfonate, leupeptin and antipain. Presumably these two processes are catalyzed by the same enzyme localized in the inner mitochondrial membrane. Physiological functions of BAPA- and BANA-hydrolyzing enzyme(s) are still unclear.     

Cleavage specificity of Saccharomyces cerevisiae flap endonuclease 1 suggests a double-flap structure as the cellular substrate

Flap endonuclease 1 (FEN1) is a structure-specific nuclease that cleaves substrates containing unannealed 5'-flaps during Okazaki fragment processing. Cleavage removes the flap at or near the point of annealing. The preferred substrate for archaeal FEN1 or the 5'-nuclease domains of bacterial DNA polymerases is a double-flap structure containing a 3'-tail on the upstream primer adjacent to the 5'-flap. We report that FEN1 in Saccharomyces cerevisiae (Rad27p) exhibits a similar specificity. Cleavage was most efficient when the upstream primer contained a 1-nucleotide 3'-tail as compared with the fully annealed upstream primer traditionally tested. The site of cleavage was exclusively at a position one nucleotide into the annealed region, allowing human DNA ligase I to seal all resulting nicks. In contrast, a portion of the products from traditional flap substrates is not ligated. The 3'-OH of the upstream primer is not critical for double-flap recognition, because Rad27p is tolerant of modifications. However, the positioning of the 3'-nucleotide defines the site of cleavage. We have tested substrates having complementary tails that equilibrate to many structures by branch migration. FEN1 only cleaved those containing a 1-nucleotide 3'-tail. Equilibrating substrates containing 12-ribonucleotides at the end of the 5'-flap simulates the situation in vivo. Rad27p cleaves this substrate in the expected 1-nucleotide 3'-tail configuration. Overall, these results suggest that the double-flap substrate is formed and cleaved during eukaryotic DNA replication in vivo.     

Stereochemical specificity for sterols in Saccharomyces cerevisiae

When sterol biosynthesis in oxygen-deprived wild type Saccharomyces cerevisiae was prevented by the presence of 2,3-iminosqualene, an inhibitor of 2,3-oxidosqualene cyclase, an absolute requirement for a sterol with a 24 beta-methyl group was found. Neither the configuration nor the size of the alkyl group at C-24 could be altered. For instance, while 24 beta-methylcholesterol (22-dihydrobrassicasterol) permitted good growth, contrary to earlier work without the inhibitor no growth at all resulted from the presence of cholesterol or its 24 alpha-methyl-, 24 alpha-ethyl-, or 24 beta-ethyl derivatives (campesterol, sitosterol, and clionasterol, respectively). The only sterol lacking a 24 beta-methyl group which allowed growth was desmosterol (24-dehydro-cholesterol), but desmosterol was metabolized to 24 beta-methylcholesterol by C1-transfer and reduction. When cholesterol supported growth in the absence of the inhibitor, small amounts of endogenously synthesized 24 beta-methylsterols (ergosterol and 22-dihydroergosterol) were identified. This previously unrecognized absolute specificity for both chirality and bulk at C-24 suggests the involvement of protein binding in at least one of the roles which sterol plays in this single-celled eukaryote.     

Saccharomyces cerevisiae structural cell wall mannoprotein

A novel mannoprotein fraction with an average molecular weight of 180 000 has been isolated from Saccharomyces cerevisiae mnn9 mutant cell wall that was solubilized by beta-glucanase digestion. The same material could be extracted from purified wall fragments with 1% sodium dodecyl sulfate. The protein component, 12% by weight, is rich in proline, whereas the carbohydrate, mainly mannose, is about evenly distributed between asparagine and hydroxyamino acids. Endoglucosaminidase H digestion of the isolated mannoprotein reduced its average molecular weight to 150 000, but the mannoprotein, while still embedded in the cell wall, was inaccessible to the enzyme. Biosynthesis and translocation of the mannoprotein were investigated by following incorporation of [3H]proline into this fraction. In the presence of tunicamycin, both mnn9 and wild-type X2180 cells made a mannoprotein fraction with an average molecular weight of 140 000, whereas in the absence of the glycosylation inhibitor, the mnn9 mutant made material with a molecular weight of 180 000 and the mannoprotein made by wild-type cells was too large to penetrate the polyacrylamide gel. Although the cell wall mannoprotein was resistant to heat and proteolytic enzymes, attempts to isolate the carbohydrate-free component failed to yield any characteristic peptide material.     

Identification and substrate specificity of a ferrichrome-type siderophore transporter (Arn1p) in Saccharomyces cerevisiae

Genes encoding transporters for heterologous siderophores have been identified in Saccharomyces cerevisiae, of which SIT1, TAF1, and ENB1 encode the transporters for ferrioxamines, ferric triacetylfusarinine C and ferric enterobactin, respectively. In the present communication we have shown that a further gene encoding a member of the major facilitator superfamily, ARN1 (YHL040c), is involved in the transport of a specific class of ferrichromes, possessing anhydromevalonyl residues linked to N(delta)-ornithine (ARN). Ferrirubin and ferrirhodin, which both are produced by filamentous fungi, are the most common representatives of this class of ferrichromes. A strain possessing a disruption in the ARN1 gene was unable to transport ferrirubin, ferrirhodin and also ferrichrome A, indicating that the encoded transporter recognizes anhydromevalonyl and the structurally-related methylglutaconyl side-chains surrounding the iron center. Ferrichromes possessing short-chain ornithine-N(delta)-acetyl residues such as ferrichrome, ferricrocin and ferrichrysin, were excluded by the Arn1 transporter. Substitution of the iron-surrounding N-acyl chains of ferrichromes by propionyl residues had no effect, whereas substitution by butyryl residues led to recognition by the Arn1 transporter. This would indicate that a chain length of four C-atoms is sufficient to allow binding. Using different asperchromes (B1, D1) we also found that a minimal number of two anhydromevalonyl residues is sufficient for recognition by Arn1p. Contrary to the iron-surrounding N-acyl residues, the peptide backbone of ferrichromes was not an important determinant for the Arn1 transporter.     

Molecular specificity of 2-aminopurine in Saccharomyces cerevisiae

The rates of DNA, RNA and protein synthesis were investigated by incorporation of radioactive precursors into the excised root tips of V. faba. 2-h exposure to 0.1% caffeine resulted in inhibition of protein synthesis to about 60% of the control rate. RNA synthesis was reduced in the range of 20–30%. The same concentration of caffeine did not affect the rate of DNA synthesis even during 12-h incubation, but concentrations higher than 1% caused a significant decrease in [³H]thymidine incorporation.     

Comprehensive X-ray structural studies of the quinolinate phosphoribosyl transferase (BNA6) from Saccharomyces cerevisiae

Quinolinic acid phosphoribosyl transferase (QAPRTase, EC 2.4.2.19) is a 32 kDa enzyme encoded by the BNA6 gene in yeast and catalyzes the formation of nicotinate mononucleotide from quinolinate and 5-phosphoribosyl-1-pyrophosphate (PRPP). QAPRTase plays a key role in the tryptophan degradation pathway via kynurenine, leading to the de novo biosynthesis of NAD (+) and clearing the neurotoxin quinolinate. To improve our understanding of the specificity of the eukaryotic enzyme and the course of events associated with catalysis, we have determined the crystal structures of the apo and singly bound forms with the substrates quinolinate and PRPP. This reveals that the enzyme folds in a manner similar to that of various prokaryotic forms which are approximately 30% identical in sequence. In addition, the structure of the Michaelis complex is approximated by PRPP and the quinolinate analogue phthalate bound to the active site. These results allow insight into the kinetic mechanism of QAPRTase and provide an understanding of structural diversity in the active site of the Saccharomyces cerevisiae enzyme when compared to prokaryotic homologues.     

A region within a lumenal loop of Saccharomyces cerevisiae Ycf1p directs proteolytic processing and substrate specificity

Ycf1p, a member of the yeast multidrug resistance-associated protein (MRP) subfamily of ATP-binding cassette proteins, is a vacuolar membrane transporter that confers resistance to a variety of toxic substances such as cadmium and arsenite. Ycf1p undergoes a PEP4-dependent processing event to yield N- and C-terminal cleavage products that remain associated with one another. In the present study, we sought to determine whether proteolytic cleavage is required for Ycf1p activity. We have identified a unique region within lumenal loop 6 of Ycf1p, designated the loop 6 insertion (L6_ins), which appears to be necessary and sufficient for proteolytic cleavage, since L6_ins can promote processing when moved to new locations in Ycf1p or into a related transporter, Bpt1p. Surprisingly, mutational results indicate that proteolytic processing is not essential for Ycf1p transport activity. Instead, the L6_ins appears to regulate substrate specificity of Ycf1p, since certain mutations in this region lower cellular cadmium resistance with a concomitant gain in arsenite resistance. Although some of these L6_ins mutations block processing, there is no correlation between processing and substrate specificity. The activity profiles of the Ycf1p L6_ins mutants are dramatically affected by the strain background in which they are expressed, raising the possibility that another cellular component may functionally impact Ycf1p activity. A candidate component may be a new full-length MRP-type transporter (NFT1), reported in the Saccharomyces Genome Database as two adjacent open reading frames, YKR103w and YKR104w, but which we show here is present in most Saccharomyces strains as a single open reading frame.     

Substrate specificity of the Saccharomyces cerevisiae Mus81-Mms4 endonuclease

Mus81-Mms4/Eme1 is a conserved structure-specific endonuclease that functions in mitotic and meiotic recombination. It has been difficult to identify a single preferred substrate of this nuclease because it is active on a variety of DNA structures. In addition, it has been suggested that the specificity of the recombinant protein may differ from that of the native enzyme. Here, we addressed these issues with respect to Mus81-Mms4 from S. cerevisiae. At low substrate concentrations, Mus81-Mms4 was active on any substrate containing a free end adjacent to the branchpoint. This includes 3'-flap (3'F), regressed leading strand replication fork (RLe), regressed lagging strand replication fork (RLa), and nicked Holliday junction (nHJ) substrates. Kinetic analysis was used to quantitate differences between substrates. High Kcat/Km values were obtained only for substrates with a 5'-end near the branchpoint (i.e., 3'F, RLe, and nHJ); 10-fold lower values were obtained for nicked duplex (nD) and RLa substrates. Substrates lacking any free ends at the branch point generated Kcat/Km values that were four orders of magnitude lower than those of the preferred substrates. Native Mus81-Mms4 was partially purified from yeast cells and found to retain its preference for 3'F over intact HJ substrates. Taken together, these results narrow the range of optimal substrates for Mus81-Mms4 and indicate that, at least for S. cerevisae, the native and recombinant enzymes display similar substrate specificities.     

Saccharomyces cerevisiae Ste5 is important for induction and substrate specificity of Fus3 MAP kinase in the pheromone signaling pathway

The pheromone pathway is one of the mitogen activated protein kinase (MAPK) signaling pathways identified in Saccharomyces cerevisiae and is involved in both G1 cell cycle arrest and mating of cells. Fus3 functions at a branching point for G1 cell cycle arrest and mating responses in the signaling cascade, and the Fus3 MAPK uses components of both G1 arrest and mating routes as substrates. The Ste5 is a scaffold protein of the MAPK module and is essential for the activation of Fus3. However, it is not known how Ste5 is involved in the specific activation of Fus3 in G1 arrest and mating. In this study, we characterized several G1 arrest defective Ste5 mutants to better understand the roles of Ste5 in the regulation of Fus3. The level of Fus3 increased by treatment with alpha-factor. However, the alpha-factor effects were not readily apparent in the observation of yeast cells containing G1 arrest defective ste5 mutant. This suggests that Ste5 plays an essential role in Fus3 induction. Fus3 immune kinase assay of G1 arrest defective ste5 transformants revealed that Ste5 is important for substrate specificity of Fus3 for G1 arrest and/or mating.     

Molecular specificity of x-radiation and its repair in Saccharomyces cerevisiae.

Molecular specificity of soft X-radiation has been studied in yeast by analyzing the transitions UAA in equilibrium UAG and nonsense leads to sense mutations in the codon tyr7-1. Synchronized cell populations in the most radiosensitive and radioresistant stages were compared: they did not show any qualitative or quantitative differences in their sensitivities to the mutagenic action of X-rays. We conclude that repair mechanisms, which remain unexpressed in the sensitive cells, do not affect point mutations of the base-substitution type.     

Biochemical and genetic classification of riboflavine deficient mutants of Saccharomyces cerevisiae

Summary Twenty six riboflavine deficient mutants of Saccharomyces cerevisiae were isolated. They can be divided by biochemical methods into four classes, accumulating (i) no specific product (S), (ii) 2,5-diamino-6-hydroxy-4-ribitylaminopyrimidine (AP), (iii) 5-amino-2,6-dihydroxy-4-ribitylaminopyrimidine (HP), (iv) 6,7-dimethyl-8-ribityllumazine (LU).In genetic experiments six complementation groups were found. Complementation group I is congruent with class S, group II with AP, III and IV with HP and V and VI with LU. In the crosses tested so far, no linkage between complementation groups could be detected by tetrad analysis. Intragenic complementation was found within complementation group V (mutants accumulating LU).Genetic experiments were performed at the Forstbotanisches Institut der Universität Freiburg. The senior author (O.O.) wishes to thank Prof. Dr. Dr. Hans Marquardt for generously providing the facilities of the Forstbotanisches Institut. This work was supported by a grant of the Deutsche Forschungsgemeinschaft to O.O.