Halogenated pyruvate derivatives as substrates of transketolase from Saccharomyces cerevisiae.

Pyruvate derivatives halogenated at C3 were shown to be donor substrates in the transketolase reaction. No drastic differences between the derivatives were observed in the value of the catalytic constant, whereas the Michaelis constant increased in the following order: Br-pyruvate < Cl-pyruvate < Cl2-pyruvate < F-pyruvate < Br2-pyruvate. The presence of the halogenated pyruvate derivatives increased the affinity of apotransketolase for the coenzyme; of note, the extent of this effect was equal with both of the active centers of the enzyme. In contrast, the presence of any other substrate known to date, including hydroxypyruvate (i.e. pyruvate hydroxylated at C3), induced nonequivalence of the active centers in that they differed in the extent to which the affinity for the coenzyme increased. Consequently, the beta-hydroxyl of dihydroxyethylthiamine diphosphate (an intermediate of the transketolase reaction) played an important role in the phenomenon of nonequivalence of the active centers associated with the coenzyme binding. The fundamental possibility was demonstrated of using halogenated pyruvate derivatives as donors of the halogen-hydroxyethyl group in organic synthesis of halogenated carbohydrates involving transketolase.

New function of the amino group of thiamine diphosphate in thiamine catalysis.

In this work, we investigated the rate of formation of the central intermediate of the transketolase reaction with thiamine diphosphate (ThDP) or 4'-methylamino-ThDP as cofactors and its stability using stopped-flow spectroscopy and circular dichroism (CD) spectroscopy. The intermediates of the transketolase reaction were analyzed by NMR spectroscopy. The kinetic stability of the intermediate was shown to be dependent on the state of the amino group of the coenzyme. The rates of the intermediate formation were the same in the case of the native and methylated ThDP, but the rates of the protonation or oxidation of the complex in the ferricyanide reaction were significantly higher in the complex with methylated ThDP. A new negative band was detected in the CD spectrum of the complex transketolase--4'-methylamino-ThDP corresponding to the protonated dihydroxyethyl-4'-methylamino-ThDP released from the active sites of the enzyme. These data suggest that transketolase in the complex with the NH2-methylated ThDP exhibits dihydroxyethyl-4'-methylamino-ThDP-synthase activity. Thus, the 4'-amino group of the coenzyme provides kinetic stability of the central intermediate of the transketolase reaction, dihydroxyethyl-ThDP.

Which stage of the process of apotransketolase interaction with thiamine diphosphate is affected by the regulatory activity of the donor substrate?

The interaction of thiamine diphosphate (ThDP) with transketolase (TK) involves at least two stages: [formula: see text] During the first stage, an inactive intermediate complex (TK...ThDP) is formed, which is then transformed into a catalytically active holoenzyme (TK* - ThDP). The second stage is related to conformational changes of the protein. In the preceding publication (Esakova, O. A., Meshalkina, L. E., Golbik, R., Hübner, G., and Kochetov, G. A. Eur. J. Biochem. 2004, 271, 4189 - 4194) we reported that the affinity of ThDP for TK considerably increases in the presence of the donor substrate, which may be a mechanism whereby the activity of the enzyme is regulated under the conditions of the coenzyme deficiency. Here, we demonstrate that the substrate affects the stage of the reverse conformational transition, characterized by the constant k(-1): in the presence of the substrate, its value is decreased several fold, whereas K(d) and k(+1) remain unchanged.

Effect of transketolase substrates on holoenzyme reconstitution and stability.

The influence of transketolase substrates on the interaction of apotransketolase with its coenzyme thiamine diphosphate (TDP) and on the stability of the reconstituted holoenzyme was studied. Donor substrates increased the affinity of the coenzyme for transketolase, whereas acceptor substrate did not. In the presence of magnesium ions, the active centers of transketolase initially identical in TDP binding lose their equivalence in the presence of donor substrates. The stability of transketolase depended on the cation type used during its reconstitution--the holoenzyme reconstituted in the presence of calcium ions was more stable than the holoenzyme produced in the presence of magnesium ions. In the presence of donor substrate, the holoenzyme stability increased without depending on the cation used during the reconstitution. Donor substrate did not influence the interaction of apotransketolase with the inactive analog of the coenzyme N3'-pyridyl thiamine diphosphate and did not stabilize the transketolase complex with this analog. The findings suggest that the effect of the substrate on the interaction of the coenzyme with apotransketolase and on stability of the reconstituted holoenzyme is caused by generation of 2-(alpha,beta-dihydroxyethyl)thiamine diphosphate (an intermediate product of the transketolase reaction), which has higher affinity for apotransketolase than TDP.

Effect of environmental conditions on aggregation and fibril formation of barstar.

The dependence on environmental conditions of the assembly of barstar into amyloid fibrils was investigated starting from the nonnative, partially folded state at low pH (A-state). The kinetics of this process was monitored by CD spectroscopy and static and dynamic light scattering. The morphology of the fibrils was visualized by electron microscopy, while the existence of the typical cross-beta structure substantiated by solution X-ray scattering. At room temperature, barstar in the A-state is unable to form amyloid fibrils, instead amorphous aggregation is observed at high ionic strength. Further destabilization of the structure is required to transform the polypeptide chain into an ensemble of conformations capable of forming amyloid fibrils. At moderate ionic strength (75 mM NaCl), the onset and the rate of fibril formation can be sensitively tuned by increasing the temperature. Two types of fibrils can be detected differing in their morphology, length distribution and characteristic far UV CD spectrum. The formation of the different types depends on the particular environmental conditions. The sequence of conversion: A-state-->fibril type I-->fibril type II appears to be irreversible. The transition into fibrils is most effective when the protein chain fulfills particular requirements concerning secondary structure, structural flexibility and tendency to cluster.

Binding and activation of thiamin diphosphate in acetohydroxyacid synthase.

Acetohydroxyacid synthases (AHASs) are biosynthetic thiamin diphosphate- (ThDP) and FAD-dependent enzymes. They are homologous to pyruvate oxidase and other members of a family of ThDP-dependent enzymes which catalyze reactions in which the first step is decarboxylation of a 2-ketoacid. AHAS catalyzes the condensation of the 2-carbon moiety, derived from the decarboxylation of pyruvate, with a second 2-ketoacid, to form acetolactate or acetohydroxybutyrate. A structural model for AHAS isozyme II (AHAS II) from Escherichia coli has been constructed on the basis of its homology with pyruvate oxidase from Lactobacillus plantarum (LpPOX). We describe here experiments which further test the model, and test whether the binding and activation of ThDP in AHAS involve the same structural elements and mechanism identified for homologous enzymes. Interaction of a conserved glutamate with the N1' of the ThDP aminopyrimidine moiety is involved in activation of the cofactor for proton exchange in several ThDP-dependent enzymes. In accord with this, the analogue N3'-pyridyl thiamin diphosphate does not support AHAS activity. Mutagenesis of Glu47, the putative conserved glutamate, decreases the rate of proton exchange at C-2 of bound ThDP by nearly 2 orders of magnitude and decreases the turnover rate for the mutants by about 10-fold. Mutant E47A also has altered substrate specificity, pH dependence, and other changes in properties. Mutagenesis of Asp428, presumed on the basis of the model to be the crucial carboxylate ligand to Mg(2+) in the "ThDP motif", leads to a decrease in the affinity of AHAS II for Mg(2+). While mutant D428N shows ThDP affinity close to that of the wild-type on saturation with Mg(2+), D428E has a decreased affinity for ThDP. These mutations also lead to dependence of the enzyme on K(+). These experiments demonstrate that AHAS binds and activates ThDP in the same way as do pyruvate decarboxylase, transketolase, and other ThDP-dependent enzymes. The biosynthetic activity of AHAS also involves many other factors beyond the binding and deprotonation of ThDP; changes in the ligands to ThDP can have interesting and unexpected effects on the reaction.

Proteins with beta-(thienopyrrolyl)alanines as alternative chromophores and pharmaceutically active amino acids.

L-beta-(Thieno[3,2-b]pyrrolyl)alanine and L-beta-(thieno[2,3-b]pyrrolyl)alanine are mutually isosteric and pharmaceutically active amino acids that mimic tryptophan with the benzene ring in the indole moiety replaced by thiophene. Sulfur as a heteroatom causes physicochemical changes in these tryptophan surrogates that bring about completely new properties not found in the indole moiety. These synthetic amino acids were incorporated into recombinant proteins in response to the Trp UGG codons by fermentation in a Trp-auxotrophic Escherichia coli host strain using the selective pressure incorporation method. Related protein mutants expectedly retain the secondary structure of the native proteins but show significantly changed optical and thermodynamic properties. In this way, new spectral windows, fluorescence, polarity, thermodynamics, or pharmacological properties are inserted into proteins. Such an engineering approach by translational integration of synthetic amino acids with a priori defined properties, as shown in this study, proved to be a novel and useful tool for protein rational design.

Consequences of a modified putative substrate-activation site on catalysis by yeast pyruvate decarboxylase.

Earlier, it had been proposed in the laboratories at Halle that a cysteine residue is responsible for the hysteretic substrate activation behavior of yeast pyruvate decarboxylase. More recently, this idea has received support in a series of studies from Rutgers with the identification of residue C221 as the site where substrate is bound to transmit the information to H92, to E91, to W412, and finally to the active center thiamin diphosphate. According to steady-state kinetic assays, the C221A/C222A variant is no longer subject to substrate activation yet is still a well-functioning enzyme. Several further experiments are reported on this variant: (1) The variant exhibits lag phases in the product formation progress curves, which can be attributed to a unimolecular step in the pre-steady-state stage of catalysis. (2) The rate of exchange with solvent deuterium of the thiamin diphosphate C2H atom is slowed by a factor of 2 compared to the wild-type enzyme, suggesting that the reduced activity that results from the substitutions some 20 A from the active center is also seen in the first key step of the reaction. (3) The solvent (deuterium oxide) kinetic isotope effect was found to be inverse on V(max)/K(m) (0.62), and small but normal on V(max) (1.26), virtually ruling out residue C221 as being responsible for the inverse effects reported for the wild-type enzyme at low substrate concentrations. The solvent kinetic isotope effects are compared to those on two related enzymes not subject to substrate activation, Zymomonas mobilis pyruvate decarboxylase and benzoylformate decarboxylase.

Active oligomeric states of pyruvate decarboxylase and their functional characterization.

Homomeric pyruvate decarboxylase (E.C 4.1.1.1) from yeast consists of dimers and tetramers under physiological conditions, a K(d) value of 8.1 microM was determined by analytical ultracentrifugation. Dimers and monomers of the enzyme could be populated by equilibrium denaturation using urea as denaturant at defined concentrations and monitored by a combination of optical (fluorescence and circular dichroism) and hydrodynamic methods (analytical ultracentrifugation). Dimers occur after treatment with 0.5 M urea, monomers with 2.0 M urea independent of the protein concentration. The structured monomers are catalytically inactive. At even higher denaturant concentrations (6 M urea) the monomers unfold. The contact sites of two monomers in forming a dimer as the smallest enzymatically active unit are mainly determined by aromatic amino acids. Their interactions have been quantified both by structure-theoretical calculations on the basis of the X-ray crystallography structure, and experimentally by binding of the fluorescent dye bis-ANS. The contact sites of two dimers in tetramer formation, however, are mainly determined by electrostatic interactions. Homomeric pyruvate decarboxylase (PDC) is activated by its substrate pyruvate. There was no difference in the steady-state activity (specific activity) between dimers and tetramers. The activation kinetics of the two oligomeric states, however, revealed differences in the dissociation constant of the regulatory substrate (K(a)) by one order of magnitude. The tetramer formation is related to structural consequences of the interaction transfer in the activation process causing an improved substrate utilization.

Examination of donor substrate conversion in yeast transketolase.

The cleavage of the donor substrate d-xylulose 5-phosphate by wild-type and H263A mutant yeast transketolase was studied using enzyme kinetics and circular dichroism spectroscopy. The enzymes are able to catalyze the cleavage of donor substrates, the first half-reaction, even in the absence of any acceptor substrate yielding d-glyceraldehyde 3-phosphate as measured in the coupled optical test according to Kochetov (Kochetov, G. A. (1982) Methods Enzymol. 90, 209-223) and compared with the H263A variant. Overall, the H263A mutant enzyme is less active than the wild-type. However, an increase in the rate constant of the release of the enzyme-bound glycolyl moiety was observed and related to a stabilization of the "active glycolaldehyde" (alpha-carbanion) by histidine 263. Chemically synthesized dl-(alpha,beta-dihydroxyethyl)thiamin diphosphate is bound to wild-type transketolase with an apparent K(D) of 4.3 +/- 0.8 microm (racemate) calculated from titration experiments using circular dichroism spectroscopy. Both enantiomers are cleaved by the enzyme at different rates. In contrast to the enzyme-generated alpha-carbanion of (alpha,beta-dihydroxyethyl)thiamin diphosphate formed by decarboxylation of hydroxylactylthiamin diphosphate after incubation of transketolase with beta-hydroxypyruvate, the synthesized dl-(alpha,beta-dihydroxyethyl)thiamin diphosphate did not work as donor substrate when erythrose 4-phosphate is used as acceptor substrate in the coupled enzymatic test according to Sprenger (Sprenger, G. A., Schörken, U., Sprenger, G., and Sahm, H. (1995) Eur. J. Biochem. 230, 525-532).

Regulation of phosphotransferase activity of hexokinase 2 from Saccharomyces cerevisiae by modification at serine-14.

Isoenzyme 2 of hexokinase functions in sugar sensing and glucose repression in Saccharomyces cerevisiae. The degree of in vivo phosphorylation of hexokinase 2 at serine-14 is inversely related to the extracellular glucose concentration [Vojtek, A. B., and Fraenkel, D. G. (1990) Eur. J. Biochem. 190, 371-375]; however, a physiological role of the modification causing the dissociation of the dimeric enzyme in vitro [as effected by a serine-glutamate exchange at position 14; Behlke et al. (1998) Biochemistry 37, 11989-11995] is unclear. This paper describes a comparative stopped-flow kinetic and sedimentation equilibrium analysis performed with native unphosphorylated hexokinase 2 and a permanently pseudophosphorylated glutamate-14 mutant enzyme to determine the functional consequences of phosphorylation-induced enzyme dissociation. The use of a dye-linked hexokinase assay monitoring proton generation allowed the investigation of the kinetics of glucose phosphorylation over a wide range of enzyme concentrations. The kinetic data indicated that monomeric hexokinase represents the high-affinity form of isoenzyme 2 for both glycolytic substrates. Inhibition of glucose phosphorylation by ATP [Moreno et al. (1986) Eur. J. Biochem. 161, 565-569] was only observed at a low enzyme concentration, whereas no inhibition was detected at the high concentration of hexokinase 2 presumed to occur in the cell. Pseudophosphorylation by glutamate substitution for serine-14 increased substrate affinity at high enzyme concentration and stimulated the autophosphorylation of isoenzyme 2. The possible role of hexokinase 2 in vivo phosphorylation at serine-14 in glucose signaling is discussed.

Mechanism of elementary catalytic steps of pyruvate oxidase from Lactobacillus plantarum.

Single steps in the catalytic cycle of pyruvate oxidase from Lactobacillus plantarum have been characterized kinetically and mechanistically by stopped-flow in combination with kinetic solvent isotope effect studies. Reversible substrate binding of pyruvate occurs with an on-rate of 6.5 x 10(4) M(-1) s(-1) and an off-rate of pyruvate of 20 s(-1). Decarboxylation of the intermediate lactyl-ThDP and the reduction of FAD which consists of two consecutive single electron-transfer steps from HEThDP to FAD occur with rates of about k(dec) = 112 s(-1) and k(red) = 422 s(-1). Flavin radical intermediates are not observed during reduction, and kinetic solvent isotope effects are absent, indicating that electron transfer and protonation processes are not rate limiting in the overall reduction process. Reoxidation of FADH(2) by O(2) to yield H(2)O(2) takes place at a pseudo-first-order rate of about 35 s(-1) in air-saturated buffer. A comparable value of about 35 s(-1) was estimated for the phosphorolysis of the acetyl-ThDP intermediate at phosphate saturation. In competition with phosphorolysis, enzyme-bound acetyl-ThDP is hydrolyzed with a rate k = 0.03 s(-1). This is the first report in which the reaction of enzyme-bound acetyl-ThDP with phosphate and OH(-) is monitored directly by FAD absorbance changes using the sequential stopped-flow technique.

Denaturation of phosphofructokinase-1 from Saccharomyces cerevisiae by guanidinium chloride and reconstitution of the unfolded subunits to their catalytically active form.

Unfolding and refolding of heterooctameric phosphofructokinase-1 from Saccharomyces cerevisiae were investigated by application of kinetic, hydrodynamic, and spectroscopic methods and by use of guanidinium chloride (GdmCl) as denaturant. Inactivation of the enzyme starts at about 0.3 M GdmCl and undergoes a sharp unfolding transition in a narrow range of the denaturant concentration. The inactivation is accompanied by a dissociation of the enzyme into dimers (at 0.6 M GdmCl), which could be detected by changes of the circular dichroism and intrinsic fluorescence. Protein aggregates were observed from 0.7 to 1.5 M GdmCl that unfold at higher denaturant concentrations. Refolding of chemically denatured phosphofructokinase proceeds as a stepwise process via the generation of elements of secondary structure, the formation of assembly-competent monomers that associate to heterodimers and the assembly of dimers to heterotetramers and heterooctamers. The assembly reactions seem to be rate-limiting. Recovery of the enzyme activity (maximum 65%) competes with an nonproductive aggregation of the subunits. alpha-Cyclodextrin functions as an artificial chaperone by preventing aggregation of the subunits, whereas ATP is suggested to support the generation of heterodimers that are competent to a further assembly.

The spatial orientation of the essential amino acid residues arginine and aspartate within the alpha1beta1 integrin recognition site of collagen IV has been resolved using fluorescence resonance energy transfer.

The interaction of collagen IV with cells is mediated mainly by the integrin alpha1beta1. The recognition site has been located to a segment of the triple-helical domain 100 nm away from the N terminus of the collagen molecule. The three essential amino acid residues of the alpha1beta1 binding site, arginine alpha2(IV)461 and the two aspartate residues alpha1(IV)461, are all located on different chains. Since the spatial array of the three residues depends on the stagger of the chains within the triple helix, the stagger has been elucidated using fluorescence resonance energy transfer with phenylalanine alpha1(IV)473 and tryptophan alpha2(IV)479 as the fluorescent donor/acceptor pair. The distance R between phenylalanine and tryptophan was determined by analysis of the energy transfer efficiency, E, and the orientation factor, kappa(2). In parallel, distance R and orientation factor, kappa(2 )were also calculated from the coordinates of the triple helix. Comparison of the calculated and empirically determined values unequivocally showed the stagger to be alpha1'alpha1alpha2. This arrangement of the three alpha chains describes the conformation of the alpha1beta1 integrin recognition site, that is the distinct orientation of the side-chains of the essential residues aspartate and arginine in respect to the helix axis.

The Janus face of the archaeal Cdc48/p97 homologue VAT: protein folding versus unfolding.

Members of the AAA family of ATPases have been implicated in chaperone-like activities. We used the archaeal Cdc48/p97 homologue VAT as a model system to investigate the effect of an AAA protein on the folding and unfolding of two well-studied, heterologous substrates, cyclophilin and penicillinase. We found that, depending on the Mg2+ concentration, VAT assumes two states with maximum rates of ATP hydrolysis that differ by an order of magnitude. In the low-activity state, VAT accelerated the refolding of penicillinase, whereas in the high-activity state, it accelerated its unfolding. Both reactions were ATP-dependent. In its interaction with cyclophilin, VAT was ATP-independent and only promoted refolding. The N-terminal domain of VAT, which lacks ATPase activity, also accelerated the refolding of cyclophilin but showed no effect on penicillinase. VAT appears to be structurally equivalent over its entire length to Sec18/NSF, suggesting that these results apply more broadly to group II AAA proteins.

Folding of barstar C40A/C82A/P27A and catalysis of the peptidyl-prolyl cis/trans isomerization by human cytosolic cyclophilin (Cyp18).

Refolding of b*C40A/C82A/P27A is comprised of several kinetically detectable folding phases. The slowest phase in refolding originates from trans-->cis isomerization of the Tyr47-Pro48 peptide bond being in cis conformation in the native state. This refolding phase can be accelerated by the peptidyl-prolyl cis/trans isomerase human cytosolic cyclophilin (Cyp18) with a kcat/K(M) of 254,000 M(-1) s(-1). The fast refolding phase is not influenced by the enzyme.

Novel molecular architecture of the multimeric archaeal PEP-synthase homologue (MAPS) from Staphylothermus marinus.

The phosphoenolpyruvate (PEP)-synthases belong to the family of structurally and functionally related PEP-utilizing enzymes. The only archaeal member of this family characterized thus far is the Multimeric Archaeal PEP-Synthase homologue from Staphylothermus marinus (MAPS). This protein complex differs from the bacterial and eukaryotic representatives characterized to date in its homomultimeric, as opposed to dimeric or tetrameric, structure. We have probed the molecular architecture of MAPS using limited proteolytic digestion in conjunction with electron microscopic, biochemical, and biophysical techniques. The 2.2 MDa particle was found to be organized in a concentric fashion. The 93.7 kDa monomers possess a pronounced tripartite domain structure and are arranged such that the N-terminal domains form an outer shell, the intermediate domains form an inner shell, and the C-terminal domains form a core structure responsible for the assembly into a multimeric complex. The core domain was shown to be capable of assembling into the native multimer by recombinant expression in Escherichia coli. Deletion mutants as well as a synthetic peptide were investigated for their state of oligomerization using native polyacrylamide gel electrophoresis, molecular sieve chromatography, analytical ultracentrifugation, circular dichroism (CD) spectroscopy, and chemical cross-linking. Our data confirmed the existence of a short C-terminal, alpha-helical oligomerization motif that had been suggested by multiple sequence alignments and secondary structure predictions. We propose that this motif bundles the monomers into six groups of four. An additional formation of 12 dimers between globular domains from different bundles leads to the multimeric assembly. According to our model, each of the six bundles of globular domains is positioned at the corners of an imaginary octahedron, and the helical C-terminal segments are oriented towards the centre of the particle. The edges of the octahedron represent the dimeric contacts. Phylogenetic analysis suggests that the ancient predecessor of this family of enzymes contained the C-terminal oligomerization motif as a feature that was preserved in some hyperthermophiles.

Atomic mutations in annexin V--thermodynamic studies of isomorphous protein variants.

We have recently developed methods for specific and high-level replacement of methionine with 2-aminohexanoic acid, selenomethionine and telluromethionine as isosteric and atomic analogues for structural investigations of human recombinant annexin V. The variants formed isomorphic crystals and retained the parent three-dimensional fold and bioactivities. Folding parameters were determined from thermal and chemical unfolding to partially denatured states. Stabilities estimated from guanidinium chloride unfolding equilibria are not changed significantly for the atomic mutants (S-->Se-->Te) while the denaturation midpoint is shifted toward lower values with an increase of the m values at the increase of hydrophobicity. In contrast, stabilities in urea are considerably affected by the atomic substitutions, decreasing together with the m and [D]1/2 values. The methylene and selenium variants are identical within the limits of error of all measurements performed here. The physical parameters of the amino acid analogues and the values derived from the slopes of the unfolding data are highly correlated. This approach demonstrates how systematic variations in atomic number at the site of replacement (atomic mutations) can provide a method to probe specific folding properties of proteins.

Thermodynamic stability and folding of GroEL minichaperones.

The apical domain of GroEL (residues 191 to 376) and its C-terminally truncated fragment GroEL(191-345) are expressed with high yield in Escherichia coli to give functional monomeric minichaperones. Owing to the reversible folding behaviour of the minichaperones we can analyse the folding of the polypeptide binding domain of the multidomain GroEL protein, the folding of which is known to be irreversible. The apical domain shows two reversible temperature transitions with transition midpoints at 35 degrees C and at 67 degrees C that can be attributed to the unfolding of the C-terminal helices and the domain core, respectively. The native state of the domain core is stabilized by 5.5 kcal mol-1 relative to the unfolded state. The rate constant of folding of the apical domain core is independent of the minichaperone concentration and the presence of the C-terminal alpha-helices. A folding intermediate on the folding pathway is destabilized relative to the native state by 1.6 kcal mol-1, which is also detected by equilibrium and kinetic binding of the dye bis-ANS. Reversible folding of the polypeptide domain of GroEL guarantees highly efficient chaperonin activity within the GroEL toroid.

Dissecting the assembly pathway of the 20S proteasome.

Proteasomes reach their mature active state via a complex cascade of folding, assembly and processing events. The Rhodococcus proteasome offers a means to dissect the assembly pathway and to characterize intermediates; its four subunits (alpha1, alpha2, beta1, beta2) assemble efficiently in vitro with any combination of alpha and beta. Assembly studies with wild-type and N-terminally truncated beta-subunits in conjunction with refolding studies allowed to define the role of the propeptide which is two-fold: It supports the initial folding of the beta-subunits and it promotes the maturation of the holoproteasomes.