The complete primary structure of mouse 20S proteasomes

 The proteasome is a large multicatalytic proteinase that plays a role in the generation of peptides for presentation by major histocompatibility complex class I molecules. The 20S proteolytic core of mammalian proteasomes is assembled from a group of 17 protein subunits that generate a distinctive pattern of spots upon two-dimensional gel electrophoresis. The genes for most of these subunits have been cloned from humans and rats. We isolated cDNA clones for the mouse orthologues of ten of the subunits [PSMA1 (C2), PSMA2 (C3), PSMA3 (C8), PSMA4 (C9), PSMA5 (ZETA), PSMA6 (IOTA), PSMA7 (C6-I), PSMB2 (C7-I), PSMB3 (C10-II), and PSMB5 (X)] to complete the cloning of all of the mouse subunits. Using antisera raised against these subunits or their orthologues, we verified the identity of these proteins by two-dimensional NEPHGE-PAGE. Received: 8 March 1999 / Accepted: 8 April 1999     

Structure determination of the constitutive 20S proteasome from bovine liver at 2.75 A resolution

The crystal structure of the 20S proteasome from bovine liver was determined by the molecular replacement method using the structure of the 20S proteasome from the yeast Sacccharomyces cerevisiae. The initial phases were refined by density modification coupled with non-crystallographic symmetry averaging. The structural model was refined with the program CNS. The final R-factor and R(free) were 0.25 and 0.29, respectively. The constitutive proteasome without any contamination by the immunoproteasome was identified in the crystal structure.     

Subunit topology of two 20S proteasomes from Haloferax volcanii

Haloferax volcanii, a halophilic archaeon, synthesizes three different proteins (alpha1, alpha2, and beta) which are classified in the 20S proteasome superfamily. The alpha1 and beta proteins alone form active 20S proteasomes; the role of alpha2, however, is not clear. To address this, alpha2 was synthesized with an epitope tag and purified by affinity chromatography from recombinant H. volcanii. The alpha2 protein copurified with alpha1 and beta in a complex with an overall structure and peptide-hydrolyzing activity comparable to those of the previously described alpha1-beta proteasome. Supplementing buffers with 10 mM CaCl(2) stabilized the halophilic proteasomes in the absence of salt and enabled them to be separated by native gel electrophoresis. This facilitated the discovery that wild-type H. volcanii synthesizes more than one type of 20S proteasome. Two 20S proteasomes, the alpha1-beta and alpha1-alpha2-beta proteasomes, were identified during stationary phase. Cross-linking of these enzymes, coupled with available structural information, suggested that the alpha1-beta proteasome was a symmetrical cylinder with alpha1 rings on each end. In contrast, the alpha1-alpha2-beta proteasome appeared to be asymmetrical with homo-oligomeric alpha1 and alpha2 rings positioned on separate ends. Inter-alpha-subunit contacts were only detected when the ratio of alpha1 to alpha2 was perturbed in the cell using recombinant technology. These results support a model that the ratio of alpha proteins may modulate the composition and subunit topology of 20S proteasomes in the cell.     

Isolation and characterization of two 20S proteasomes from the endoplasmic reticulum of rat liver microsomes.

Two new forms of proteasomes, designated as the endoplasmic reticulum (ER) membrane-associated proteasome (ERa proteasome) and ER membrane-bound proteasome (ERb proteasome), were purified to homogeneity from 0.0125 and 2.5% sodium cholate extracts, respectively, of a rat liver microsomal fraction. SDS-PAGE analysis revealed that the purified ERa and ERb proteasomes were composed of multiple subunits similar to the cytosolic 20S proteasome. However, electrophoretic, structural and immunochemical differences between the ERa, ERb and cytosolic 20S proteasomes were observed on native PAGE, two-dimensional (2D) PAGE, and immunoblot analyses. Purification of ERb from a 2.5% sodium cholate extract of the trypsin-treated microsomal fraction yielded a trypsin-modified form of ERb (tERb), which lacked the C2 subunit at least. On the other hand, no ERa proteasome was obtained from the 0.0125% sodium cholate extract of the trypsin-treated microsomes, suggesting that ERa and ERb are ER membrane-associated and -bound proteasomes, respectively. The ERa, ERb, and cytosolic 20S proteasomes exhibited similar specificities as to peptide hydrolyzing activity, although differences in their activities were noted in the presence of SDS and phospholipid. With respect to the proteolysis of protein substrates, only the ERb proteasome cleaved beta-casein, and it also degraded reduced and carboxymethylated lysozyme considerably faster than the cytosolic 20S and ERa proteasomes. Collectively our results suggest that the ERa and ERb proteasomes may play roles in intracellular proteolysis distinct from that of the cytosolic 20S proteasome.     

Multiple chaperone-assisted formation of mammalian 20S proteasomes

Protein degradation is essential for maintenance of cellular homeostasis. The majority of proteins are selectively degraded in eukaryotic cells by the ubiquitin-proteasome system. The 26S proteasome selects target proteins that are covalently modified with polyubiquitin chains. The 26S proteasome is a multisubunit protease responsible for regulated proteolysis in eukaryotic cells. The catalytic activities are carried out by the core 20S proteasome. The eukaryotic 20S proteasome is composed of 28 subunits arranged in a cylindrical particle as four heteroheptameric rings, alpha1-7beta1-7beta1-7alpha1-7. Recent studies have revealed the mechanism responsible for the assembly of such a complex structure. This article recounts the observations that disclosed the biogenesis of 20S proteasomes and discusses the difference in the mechanism of assembly between archael, yeast, and mammalian 20S proteasomes.     

Development of lipopeptides for inhibiting 20S proteasomes

Proteasomes are responsible for the cytoplasmic turnover of the vast majority of proteins including regulatory proteins. We have synthesized lipopeptides a new class of non-covalent inhibitors of the 20S proteasome and assayed their inhibitory capacities. Their ability to inhibit at micromolar concentrations chymotrypsin-like and post-acid activities depends on peptide length (3 or 6 amino acids), sequence (presence of a positively or negatively charged amino acid), and alkyl chain length (C6-C18). These structural features could be varied to selectively inhibit one or more of the three proteasome activities.     

Native structure of rat liver immune proteasomes

Native structure of active forms of rat liver immune proteasomes has been studied by two-dimensional electrophoresis method modified for analysis of unpurified protein fractions. The developed method allowed revealing the proteasome immune subunits LMP7 and LMP2 in 20S subparticles and in the structures bound to one or two PA28αβ activators, but not to the PA700 activator, which is involved in the hydrolysis of ubiquitinated proteins. The results obtained indicate the participation of the immune proteasomes in delicate regulatory mechanisms based on the production of biologically active peptides and exclude their participation in processes of crude degradation of “rotated” ubiquitinated proteins.     

Substrate size selectivity of 20S proteasomes: analysis with variable-sized synthetic substrates

Proteasomes are tubular complexes with proteolytic activities on their lumenal surfaces so that large substrates should be sterically hindered from reaching the catalytic sites. Here we examine effects of substrate size on rates of cleavage by 20S proteasomes of Methanosarcina thermophila. Synthetic chromogenic substrates of variable size were prepared by linking a constant substrate group (Ala-Ala-Phe-p-nitroanilide) to a linear polymer (methoxypolyethylene glycol) with variable chain length. The smallest macromolecular substrates were cleaved more efficiently than free tripeptide substrate, and cleavage of macromolecular substrates was saturable, whereas cleavage of free tripeptide substrate was not, indicating mechanistic differences between the cleavage of large and small substrates. Rates of macromolecular substrate cleavage decreased progressively up to 10-fold as the size of the polymeric component of substrates increased. Macromolecular synthetic substrates appear to be better models of proteasome action on natural protein substrates and demonstrate substrate size selectivity of proteasomes.     

Identical subunit topographies of human and yeast 20S proteasomes

The arrangement of subunits in human 20S proteasomes was recently determined by us by immunoelectron microscopy and chemical cross-linking. The positions of 4 of the 14 subunits differed from those found in the yeast proteasome by X-ray crystallography. Double labeling of human 20S proteasomes with antibodies to subunits C2 and C5 has now shown that these subunits are nearest neighbors. The result contradicts our published model for the human proteasome but is in accordance with the subunit arrangement in yeast proteasomes, suggesting that yeast and human proteasomes most probably have identical subunit arrangements. Immunoelectron microscopy also showed that the C-terminal extension at the human C2 subunit is flexible but takes up a well-defined position in the proteasome.     

Late events in the assembly of 20S proteasomes

Electron microscopy and STEM mass measurements have been used to characterize late intermediates in the assembly pathway of wildtype and mutant Rhodococcus proteasomes. A proteolytically inactive and processing-incompetent mutant, betaK33A, allowed a short-lived late intermediate of the pathway to be captured, the preholoproteasome. In this fully assembled 20S complex the 14 propeptides with an aggregate mass of 100 kDa fill the whole central cavity and most of the two antechambers. It is further shown that in wildtype Rhodococcus proteasomes the propeptides are degraded in a processive manner undergoing multiple cleavages before the products are discharged and the inner cavities are cleared. It appears that the docking of two half-proteasomes, i.e., preholoproteasome formation, is sufficient to trigger autocleavage of the Gly-1/Thr1 bond necessary for active site formation and the subsequent degradation of the propeptides.     

High resolution crystal structure of bovine mitochondrial EF-Tu in complex with GDP

The crystal structure of bovine mitochondrial elongation factor Tu (EF-Tu) in complex with GDP has been determined at a resolution of 1. 94 A. The structure is similar to that of EF-Tu:GDP from Escherichia coli and Thermus aquaticus, but the orientation of the GDP-binding domain 1 is changed relative to domains 2 and 3. Sixteen conserved water molecules common to EF-Tu and other G-proteins in the GDP-binding site are described. These water molecules create a network linking separated parts of the binding pocket. Mitochondrial EF-Tu binds nucleotides less tightly than prokaryotic EF-Tu possibly due to an increased mobility in regions close to the GDP-binding site. The C-terminal extension of mitochondrial EF-Tu has structural similarities with DNA recognising zinc fingers suggesting that the extension may be involved in recognition of RNA.     

The effect of heat shock on 20S/26S proteasomes

We have studied the consequences of heat shock on 20S/26S proteasome activity and activation, the proteasomal subunit composition, proteasome assembly, subunit mRNA stability as well as on the intracellular distribution of proteasomes. Our data show that heat shock locks 20S proteasomes in their latent inactive state and impairs further activation of the 26S proteasome by ATP. Proteasome mRNA levels are decreased after heat shock and the assembly of the proteasome complex is inhibited. Heat shock also induces a rapid reorganisation of the cellular distribution of the proteasome which appears to be connected with proteasome activity and the change of the cellular architecture after heat shock.     

Dehydration converts DsbG crystal diffraction from low to high resolution

Diffraction quality crystals are essential for crystallographic studies of protein structure, and the production of poorly diffracting crystals is often regarded as a dead end in the process. Here we show a dramatic improvement of poorly diffracting DsbG crystals allowing high-resolution diffraction data measurement. Before dehydration, the crystals are fragile and the diffraction pattern is streaky, extending to 10 A resolution. After dehydration, there is a spectacular improvement, with the diffraction pattern extending to 2 A resolution. This and other recent results show that dehydration is a simple, rapid, and inexpensive approach to convert poor quality crystals into diffraction quality crystals.     

The structure of the mammalian 20S proteasome at 2.75 A resolution

The 20S proteasome is the catalytic portion of the 26S proteasome. Constitutively expressed mammalian 20S proteasomes have three active subunits, beta 1, beta 2, and beta 5, which are replaced in the immunoproteasome by interferon-gamma-inducible subunits beta 1i, beta 2i, and beta 5i, respectively. Here we determined the crystal structure of the bovine 20S proteasome at 2.75 A resolution. The structures of alpha 2, beta 1, beta 5, beta 6, and beta 7 subunits of the bovine enzyme were different from the yeast enzyme but enabled the bovine proteasome to accommodate either the constitutive or the inducible subunits. A novel N-terminal nucleophile hydrolase activity was proposed for the beta 7 subunit. We also determined the site of the nuclear localization signals in the molecule. A model of the immunoproteasome was predicted from this constitutive structure.     

Yeast tRNA(Asp)-aspartyl-tRNA synthetase complex: low resolution crystal structure

Yeast aspartyl-tRNA synthetase, a dimer of molecular weight 125,000, and two molecules of its cognate tRNA (Mr = 24160) cocrystallize in the cubic space group I432 (a = 354 A). The crystal structure was solved to low resolution using neutron and X-ray diffraction data. Neutron single crystal diffraction data were collected in five solvents differing by their D2O content in order to use the contrast variation method to distinguish between the protein and tRNA. The synthetase was first located at 40 A resolution using the 65% D2O neutron data (tRNA matched) tRNA molecules were found at 20 A resolution using both neutron and X-ray data. The resulting model was refined against 10 A resolution X-ray data, using density modification and least-squares refinement of the tRNA positions. The crystal structure solved without a priori phase knowledge, was confirmed later by isomorphous replacement. The molecular model of the complex is in good agreement with results obtained in solution by probing the protected part of the tRNA by chemical reagents.     

Primary structure of hepatoredoxin from bovine liver mitochondria

The primary structure of hepatoredoxin from bovine liver mitochondria was established. It consists of 117 amino acid residues. The identity of the amino acid sequences of bovine hepatoredoxin and adrenodoxin from bovine adrenal cortex mitochondria was shown. It is assumed that the role of a ferredoxin component in mitochondrial steroid-hydroxylating systems from different organs is played by the same [2Fe-2S]-protein.     

Phosphorylation of proteasomes and alpha-RNP from rat liver cells

In eukaryotic cells the population of proteasomes is heterogeneous. Here we have shown that proteasomes from nuclei and cytoplasm of rat liver cells differ in their subunit patterns. The subunit pattern of alpha-RNP differs from that of proteasomes, however, alpha-RNP particles contain the number of 26S proteasome subunits. Moreover, the proteasomes contain subunits of alpha-RNP. We have shown for the first time that nuclear proteasomes and alpha-RNP are hyperphosphorylated on threonine residues. Differences in phosphorylation state of subunits of nuclear and cytoplasmic proteasomes and alpha-RNP on threonine and tyrosine residues have been revealed. A suggestion is put forward that hyperphosphorylation of subunits may determine nuclear localization of these complexes in rat liver cells. The results obtained suggest that a highly specialized system of protein kinases and phosphatases may be involved in the regulation of phosphorylation state of different populations of proteasomes and alpha-RNP in rat liver cells.